223031012
/
DAMASK_EICMD
mirror of https://github.com/achalhp/DAMASK_EICMD.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
DAMASK_EICMD/code/constitutive_nonlocal.f90

3814 lines
194 KiB
Fortran
Raw Normal View History

updated copyright header
2013-03-22 23:05:05 +05:30
! Copyright 2011-13 Max-Planck-Institut für Eisenforschung GmbH
added copyright text to all f90 (free) format files
2011-04-04 19:39:54 +05:30
!
! This file is part of DAMASK,
changed enconding of all source files to UTF-8 without BOM (signature) Codepage 65001
2011-04-07 12:50:28 +05:30
! the Düsseldorf Advanced MAterial Simulation Kit.
added copyright text to all f90 (free) format files
2011-04-04 19:39:54 +05:30
!
! DAMASK is free software: you can redistribute it and/or modify
! it under the terms of the GNU General Public License as published by
! the Free Software Foundation, either version 3 of the License, or
! (at your option) any later version.
!
! DAMASK is distributed in the hope that it will be useful,
! but WITHOUT ANY WARRANTY; without even the implied warranty of
! MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
! GNU General Public License for more details.
!
! You should have received a copy of the GNU General Public License
! along with DAMASK. If not, see <http://www.gnu.org/licenses/>.
!
!##############################################################
added version information to all files do NOT edit text like this: $Id: constitutive_phenopowerlaw.f90 406 2009-08-31 14:13:10Z MPIE\f.roters $
2009-08-31 20:39:15 +05:30
!* $Id$
no change to previous version, just forget to mention one important change of the last revision: * fluxes are now again calculated and distributed only! by the originating material point. this means that the central MP might change the dotState of its neighbor. have to see whether locks slow down parallel computation
2010-10-26 19:19:58 +05:30
!************************************
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!* Module: CONSTITUTIVE_NONLOCAL *
!************************************
!* contains: *
!* - constitutive equations *
!* - parameters definition *
!************************************
MODULE constitutive_nonlocal
!* Include other modules
polishing
2013-05-23 17:55:56 +05:30
use prec, only: &
pReal, &
pInt, &
p_vec
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
implicit none
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
private
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!* Definition of parameters
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
character (len=*), parameter, public :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
CONSTITUTIVE_NONLOCAL_LABEL = 'nonlocal'
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
improved accuracy of "accumulated shear": now added as a state variable, which facilitates integration (dotstate=shearrate); needs absolute tolerance value for state integration in material.config
2013-04-04 19:07:14 +05:30
character(len=22), dimension(11), parameter, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
BASICSTATES = (/'rhoSglEdgePosMobile ', &
'rhoSglEdgeNegMobile ', &
'rhoSglScrewPosMobile ', &
'rhoSglScrewNegMobile ', &
'rhoSglEdgePosImmobile ', &
'rhoSglEdgeNegImmobile ', &
'rhoSglScrewPosImmobile', &
'rhoSglScrewNegImmobile', &
'rhoDipEdge ', &
'rhoDipScrew ', &
'accumulatedshear ' /) !< list of "basic" microstructural state variables that are independent from other state variables
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
character(len=16), dimension(3), parameter, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
DEPENDENTSTATES = (/'rhoForest ', &
'tauThreshold ', &
'tauBack ' /) !< list of microstructural state variables that depend on other state variables
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
character(len=20), dimension(6), parameter, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
OTHERSTATES = (/'velocityEdgePos ', &
'velocityEdgeNeg ', &
'velocityScrewPos ', &
'velocityScrewNeg ', &
'maxDipoleHeightEdge ', &
'maxDipoleHeightScrew' /) !< list of other dependent state variables that are not updated by microstructure
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
real(pReal), parameter, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
KB = 1.38e-23_pReal !< Physical parameter, Boltzmann constant in J/Kelvin
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!* Definition of global variables
* internal stress calculation now considers dead dislocations always at the interface, not at the center of the IP volume; used to merge them together with "normal" dislocations for stress calculation. * dislocation flux is blocked if we encounter a sign change in the resolved shear stress from the central ip to the neighbor * do not set density to zero if below certain threshold; this creates an artificial sink term
2011-08-02 16:47:45 +05:30
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
integer(pInt), dimension(:), allocatable, public :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_sizeDotState, & !< number of dotStates = number of basic state variables
constitutive_nonlocal_sizeDependentState, & !< number of dependent state variables
constitutive_nonlocal_sizeState, & !< total number of state variables
constitutive_nonlocal_sizePostResults !< cumulative size of post results
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
integer(pInt), dimension(:,:), allocatable, target, public :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_sizePostResult !< size of each post result output
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
character(len=64), dimension(:,:), allocatable, target, public :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_output !< name of each post result output
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
integer(pInt), dimension(:), allocatable, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Noutput !< number of outputs per instance of this plasticity
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
integer(pInt), dimension(:,:), allocatable, private :: &
iRhoEPU, & !< state indices for density of Unblocked Positive Edges
iRhoENU, & !< state indices for density of Unblocked Negative Edges
iRhoSPU, & !< state indices for density of Unblocked Positive Screws
iRhoSNU, & !< state indices for density of Unblocked Negative Screws
iRhoEPB, & !< state indices for density of Blocked Positive Edges
iRhoENB, & !< state indices for density of Blocked Negative Edges
iRhoSPB, & !< state indices for density of Blocked Positive Screws
iRhoSNB, & !< state indices for density of Blocked Negative Screws
iRhoED, & !< state indices for density of Edge Dipoles
iRhoSD, & !< state indices for density of Screw Dipoles
iGamma, & !< state indices for accumulated shear
iRhoF, & !< state indices for forest density
rename state variable index for critical resolved shear stress from "iTau" to "iTauF"
2013-05-24 03:02:36 +05:30
iTauF, & !< state indices for critical resolved shear stress
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
iTauB, & !< state indices for backstress
iVEP, & !< state indices for velocity of Positive Edges
iVEN, & !< state indices for velocity of Negative Edges
iVSP, & !< state indices for velocity of Positive Screws
iVSN, & !< state indices for velocity of Negative Screws
iDE, & !< state indices for stable edge dipole height
iDS !< state indices for stable screw dipole height
moved stiffness tensor calculation to lattice introduced 'isotropic' and 'orthorhombic' lattice types to use corresponding symmetries in stiffness tensor. intended to be used with non-crystal plasticity models (j2, constitutive_none with isotropic, cubic or orthotropic elasticity).
2013-01-22 03:27:26 +05:30
character(len=32), dimension(:), allocatable, public :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_structureName !< name of the lattice structure
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
integer(pInt), dimension(:), allocatable, public :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_structure !< number representing the kind of lattice structure
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
integer(pInt), dimension(:), allocatable, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
totalNslip !< total number of active slip systems for each instance
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
integer(pInt), dimension(:,:), allocatable, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Nslip, & !< number of active slip systems for each family and instance
slipFamily, & !< lookup table relating active slip system to slip family for each instance
slipSystemLattice, & !< lookup table relating active slip system index to lattice slip system index for each instance
colinearSystem !< colinear system to the active slip system (only valid for fcc!)
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
real(pReal), dimension(:), allocatable, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
CoverA, & !< c/a ratio for hex type lattice
mu, & !< shear modulus
nu, & !< poisson's ratio
atomicVolume, & !< atomic volume
Dsd0, & !< prefactor for self-diffusion coefficient
selfDiffusionEnergy, & !< activation enthalpy for diffusion
aTolRho, & !< absolute tolerance for dislocation density in state integration
aTolShear, & !< absolute tolerance for accumulated shear in state integration
significantRho, & !< density considered significant
significantN, & !< number of dislocations considered significant
cutoffRadius, & !< cutoff radius for dislocation stress
doublekinkwidth, & !< width of a doubkle kink in multiples of the burgers vector length b
solidSolutionEnergy, & !< activation energy for solid solution in J
solidSolutionSize, & !< solid solution obstacle size in multiples of the burgers vector length
solidSolutionConcentration, & !< concentration of solid solution in atomic parts
pParam, & !< parameter for kinetic law (Kocks,Argon,Ashby)
qParam, & !< parameter for kinetic law (Kocks,Argon,Ashby)
viscosity, & !< viscosity for dislocation glide in Pa s
fattack, & !< attack frequency in Hz
rhoSglScatter, & !< standard deviation of scatter in initial dislocation density
surfaceTransmissivity, & !< transmissivity at free surface
grainboundaryTransmissivity, & !< transmissivity at grain boundary (identified by different texture)
CFLfactor, & !< safety factor for CFL flux condition
fEdgeMultiplication, & !< factor that determines how much edge dislocations contribute to multiplication (0...1)
rhoSglRandom, &
rhoSglRandomBinning, &
linetensionEffect, &
edgeJogFactor
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
real(pReal), dimension(:,:), allocatable, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoSglEdgePos0, & !< initial edge_pos dislocation density per slip system for each family and instance
rhoSglEdgeNeg0, & !< initial edge_neg dislocation density per slip system for each family and instance
rhoSglScrewPos0, & !< initial screw_pos dislocation density per slip system for each family and instance
rhoSglScrewNeg0, & !< initial screw_neg dislocation density per slip system for each family and instance
rhoDipEdge0, & !< initial edge dipole dislocation density per slip system for each family and instance
rhoDipScrew0, & !< initial screw dipole dislocation density per slip system for each family and instance
lambda0PerSlipFamily, & !< mean free path prefactor for each family and instance
lambda0, & !< mean free path prefactor for each slip system and instance
burgersPerSlipFamily, & !< absolute length of burgers vector [m] for each family and instance
burgers, & !< absolute length of burgers vector [m] for each slip system and instance
interactionSlipSlip !< coefficients for slip-slip interaction for each interaction type and instance
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
real(pReal), dimension(:,:,:), allocatable, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Cslip66, & !< elasticity matrix in Mandel notation for each instance
minDipoleHeightPerSlipFamily, & !< minimum stable edge/screw dipole height for each family and instance
minDipoleHeight, & !< minimum stable edge/screw dipole height for each slip system and instance
peierlsStressPerSlipFamily, & !< Peierls stress (edge and screw)
peierlsStress, & !< Peierls stress (edge and screw)
forestProjectionEdge, & !< matrix of forest projections of edge dislocations for each instance
forestProjectionScrew, & !< matrix of forest projections of screw dislocations for each instance
interactionMatrixSlipSlip !< interaction matrix of the different slip systems for each instance
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
real(pReal), dimension(:,:,:,:), allocatable, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
lattice2slip, & !< orthogonal transformation matrix from lattice coordinate system to slip coordinate system (passive rotation !!!)
rhoDotEdgeJogsOutput, &
sourceProbability, &
shearrate
rates in constitutive_results are taken directly from last converged step and not calculated again; added output of rates for annihilation of screws and deposition of edge jogs on collinear system
2012-08-16 16:33:22 +05:30
real(pReal), dimension(:,:,:,:,:), allocatable, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Cslip3333, & !< elasticity matrix for each instance
rhoDotFluxOutput, &
rhoDotMultiplicationOutput, &
rhoDotSingle2DipoleGlideOutput, &
rhoDotAthermalAnnihilationOutput, &
rhoDotThermalAnnihilationOutput
rates in constitutive_results are taken directly from last converged step and not calculated again; added output of rates for annihilation of screws and deposition of edge jogs on collinear system
2012-08-16 16:33:22 +05:30
real(pReal), dimension(:,:,:,:,:,:), allocatable, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
compatibility !< slip system compatibility between me and my neighbors
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
added non-schmid structure
2013-01-22 05:20:28 +05:30
real(pReal), dimension(:,:), allocatable, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
nonSchmidCoeff
added non-schmid structure
2013-01-22 05:20:28 +05:30
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
logical, dimension(:), allocatable, private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
shortRangeStressCorrection, & !< flag indicating the use of the short range stress correction by a excess density gradient term
deadZoneScaling, &
probabilisticMultiplication
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
public :: &
constitutive_nonlocal_init, &
constitutive_nonlocal_stateInit, &
constitutive_nonlocal_aTolState, &
constitutive_nonlocal_homogenizedC, &
constitutive_nonlocal_microstructure, &
constitutive_nonlocal_LpAndItsTangent, &
constitutive_nonlocal_dotState, &
constitutive_nonlocal_deltaState, &
constitutive_nonlocal_dotTemperature, &
constitutive_nonlocal_updateCompatibility, &
constitutive_nonlocal_postResults
private :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_kinetics, &
constitutive_nonlocal_dislocationstress
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
CONTAINS
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!**************************************
!* Module initialization *
!**************************************
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
subroutine constitutive_nonlocal_init(myFile)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
polishing, adding _pInt etc. where applicable post_results now handels zero increment different (like FEM, it is always there even if it is not counted)
2012-02-13 19:48:07 +05:30
use, intrinsic :: iso_fortran_env ! to get compiler_version and compiler_options (at least for gfortran 4.6 at the moment)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
use math, only: math_Mandel3333to66, &
math_Voigt66to3333, &
* internal stress calculation now considers dead dislocations always at the interface, not at the center of the IP volume; used to merge them together with "normal" dislocations for stress calculation. * dislocation flux is blocked if we encounter a sign change in the resolved shear stress from the central ip to the neighbor * do not set density to zero if below certain threshold; this creates an artificial sink term
2011-08-02 16:47:45 +05:30
math_mul3x3, &
added new math function "math_deviatoric33" which returns the deviatoric part of a 3x3 tensor renamed some math functions, so that we have a universal naming scheme: for matrix multiplications use an "x" (e.g. math_mul33x3); don't use the "x" to describe the shape of the tensor that the function is applied to (e.g. math_invert33 instead of math_invert3x3)
2012-01-26 19:20:00 +05:30
math_transpose33
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
use IO, only: IO_lc, &
IO_getTag, &
IO_isBlank, &
IO_stringPos, &
IO_stringValue, &
IO_floatValue, &
IO_intValue, &
reporting current time for most module inits
2013-02-25 22:04:59 +05:30
IO_error, &
IO_timeStamp
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
use debug, only: debug_level, &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
debug_constitutive, &
debug_levelBasic
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
use mesh, only: mesh_NcpElems, &
* in nonlocal model: dislocation flux now with valid description also for large angle grain boundaries; transfer of dislocations from one slip system to the other according to new variable "constitutive_nonlocal_compatibility" which depends on the angle between slip plane normals and slip directions and is updated once per cycle in function "crystallite_orientations" * reactivated debugging functionality for "non-standard" integration methods
2010-10-12 18:38:54 +05:30
mesh_maxNips, &
replaced FE_maxNipNeighbors with mesh_maxNipNeighbors, which is sufficient for memory allocation, since this gives the maximum of neighbors in all cp elements
2013-05-23 23:16:21 +05:30
mesh_maxNipNeighbors
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
use material, only: homogenization_maxNgrains, &
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
phase_plasticity, &
phase_plasticityInstance, &
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
phase_Noutput
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
use lattice
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*** output variables
!*** input variables
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
integer(pInt), intent(in) :: myFile
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*** local variables
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
integer(pInt), parameter :: maxNchunks = 21_pInt
integer(pInt), &
dimension(1_pInt+2_pInt*maxNchunks) :: positions
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
integer(pInt), dimension(6) :: configNchunks
removed unused variables, added pure statements, and declared external functions as external (using gfortrans debug option)
2013-02-11 16:13:45 +05:30
integer(pInt) :: section, &
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
maxNinstance, &
maxTotalNslip, &
myStructure, &
f, & ! index of my slip family
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
i, & ! index of my instance of this plasticity
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
l, &
convergence of state in crystallite is now tested as follows: (state < relevant state) or (residuum < relative tolerance * state) since the relevant value for the state variables depend on their nature and can vary by large scales (e.g. volume fraction: 1e-10, dislocation density: 1e5) it is not possible to set a unique value. instead the constitutive law has to decide what is relevant. therefore, all constitutive laws now read in parameters from the material.config that determine the values for relevantState [@luc: in dislobased law relevant State is for the moment generally set to 1e-200, so no additional parameters necessary in material.config. if you also want this feature, we can still implement it, no big deal] - added sanity checks in constitutive_nonlocal.f90 - corrected coordinate transformation for backstress calculation in constitutive_nonlocal.f90 - corrected equations for evolution of dipole dislocation densities (athermal annihilation and formation by glide)
2009-09-18 21:07:14 +05:30
ns, & ! short notation for total number of active slip systems for the current instance
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
o, & ! index of my output
s, & ! index of my slip system
s1, & ! index of my slip system
s2, & ! index of my slip system
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
it, & ! index of my interaction type
total density used in backstress calculation now based on average of neighborhood fixed small bug in state_init: random distribution of density was probably not working correctly, as some variables were not properly initialized
2013-03-27 18:34:01 +05:30
Nchunks_SlipSlip = 0_pInt, &
Nchunks_SlipFamilies = 0_pInt, &
removed unused variables, added pure statements, and declared external functions as external (using gfortrans debug option)
2013-02-11 16:13:45 +05:30
mySize = 0_pInt ! to suppress warnings, safe as init is called only once
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
character(len=64) tag
remove omp statements during initialization, "line" for reading in from material.config now has a default (empty) value to prevent conditional jump depend on this value
2013-01-09 03:41:59 +05:30
character(len=1024) :: line = '' ! to start initialized
write(6,*)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
write(6,*) '<<<+- constitutive_',trim(CONSTITUTIVE_NONLOCAL_LABEL),' init -+>>>'
remove omp statements during initialization, "line" for reading in from material.config now has a default (empty) value to prevent conditional jump depend on this value
2013-01-09 03:41:59 +05:30
write(6,*) '$Id$'
reporting current time for most module inits
2013-02-25 22:04:59 +05:30
write(6,'(a16,a)') ' Current time : ',IO_timeStamp()
made code standard conform to Fortran 2008 (ignoring warning concerning comments beyond character 132). Basically, changing "x" format specifier to "Nx" ("1x") plus removing $ format specifier added compiler switches for gfortran and ifort to check for standard conformity old gnu compilers <4.4 are not longer supported because they don't provide the c binding for fftw
2012-02-01 00:48:55 +05:30
#include "compilation_info.f90"
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
maxNinstance = int(count(phase_plasticity == CONSTITUTIVE_NONLOCAL_LABEL),pInt)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
if (maxNinstance == 0) return ! we don't have to do anything if there's no instance for this constitutive law
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
if (iand(debug_level(debug_constitutive),debug_levelBasic) /= 0_pInt) then
remove omp statements during initialization, "line" for reading in from material.config now has a default (empty) value to prevent conditional jump depend on this value
2013-01-09 03:41:59 +05:30
write(6,'(a16,1x,i5)') '# instances:',maxNinstance
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
write(6,*)
debugging output is now controlled by the "verbosity" parameter in the debug.config ranging from 0 (=almost no output) to 8 (=very detailed output) 0 : only version infos and all from "hypela2"/"umat" 1 : basic outputs from "CPFEM.f90", basic output from initialization routines, debug_info 2 : extensive outputs from "CPFEM.f90", extensive output from initialization routines 3 : basic outputs from "homogenization.f90" 4 : extensive outputs from "homogenization.f90" 5 : basic outputs from "crystallite.f90" 6 : extensive outputs from "crystallite.f90" 7 : basic outputs from the constitutive files 8 : extensive outputs from the constitutive files If verbosity is equal to zero, all counters in debug are not set during calculation (e.g. debug_StressLoopDistribution or debug_cumDotStateTicks). This might speed up parallel calculation, because all these need critical statements which extremely slow down parallel computation.
2011-03-21 16:01:17 +05:30
endif
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
!*** memory allocation for global variables
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
allocate(constitutive_nonlocal_sizeDotState(maxNinstance))
constitutive_nonlocal: dislocation velocities are stored in the state, so we actually now have three "parts" of the state, the basic states that are updated by "constitutive_dotState" come first, then the dependent states that are calculated by "constitutive_microstructure" follow, and finally we have a last part reserved for other variables that just use the memory reserved by the state array and are updated somewhere else. constitutive: LpAndItsTangent does not need the full state, but only the local state, so changed that at least for the nonlocal constitutive law
2011-11-04 18:42:17 +05:30
allocate(constitutive_nonlocal_sizeDependentState(maxNinstance))
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
allocate(constitutive_nonlocal_sizeState(maxNinstance))
allocate(constitutive_nonlocal_sizePostResults(maxNinstance))
allocate(constitutive_nonlocal_sizePostResult(maxval(phase_Noutput), maxNinstance))
allocate(constitutive_nonlocal_output(maxval(phase_Noutput), maxNinstance))
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(Noutput(maxNinstance))
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
constitutive_nonlocal_sizeDotState = 0_pInt
constitutive_nonlocal: dislocation velocities are stored in the state, so we actually now have three "parts" of the state, the basic states that are updated by "constitutive_dotState" come first, then the dependent states that are calculated by "constitutive_microstructure" follow, and finally we have a last part reserved for other variables that just use the memory reserved by the state array and are updated somewhere else. constitutive: LpAndItsTangent does not need the full state, but only the local state, so changed that at least for the nonlocal constitutive law
2011-11-04 18:42:17 +05:30
constitutive_nonlocal_sizeDependentState = 0_pInt
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
constitutive_nonlocal_sizeState = 0_pInt
constitutive_nonlocal_sizePostResults = 0_pInt
constitutive_nonlocal_sizePostResult = 0_pInt
constitutive_nonlocal_output = ''
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Noutput = 0_pInt
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
allocate(constitutive_nonlocal_structureName(maxNinstance))
allocate(constitutive_nonlocal_structure(maxNinstance))
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(Nslip(lattice_maxNslipFamily, maxNinstance))
allocate(slipFamily(lattice_maxNslip, maxNinstance))
allocate(slipSystemLattice(lattice_maxNslip, maxNinstance))
allocate(totalNslip(maxNinstance))
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
constitutive_nonlocal_structureName = ''
constitutive_nonlocal_structure = 0_pInt
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Nslip = 0_pInt
slipFamily = 0_pInt
slipSystemLattice = 0_pInt
totalNslip = 0_pInt
allocate(CoverA(maxNinstance))
allocate(mu(maxNinstance))
allocate(nu(maxNinstance))
allocate(atomicVolume(maxNinstance))
allocate(Dsd0(maxNinstance))
allocate(selfDiffusionEnergy(maxNinstance))
allocate(aTolRho(maxNinstance))
allocate(aTolShear(maxNinstance))
allocate(significantRho(maxNinstance))
allocate(significantN(maxNinstance))
allocate(Cslip66(6,6,maxNinstance))
allocate(Cslip3333(3,3,3,3,maxNinstance))
allocate(cutoffRadius(maxNinstance))
allocate(doublekinkwidth(maxNinstance))
allocate(solidSolutionEnergy(maxNinstance))
allocate(solidSolutionSize(maxNinstance))
allocate(solidSolutionConcentration(maxNinstance))
allocate(pParam(maxNinstance))
allocate(qParam(maxNinstance))
allocate(viscosity(maxNinstance))
allocate(fattack(maxNinstance))
allocate(rhoSglScatter(maxNinstance))
allocate(rhoSglRandom(maxNinstance))
allocate(rhoSglRandomBinning(maxNinstance))
allocate(surfaceTransmissivity(maxNinstance))
allocate(grainboundaryTransmissivity(maxNinstance))
allocate(shortRangeStressCorrection(maxNinstance))
allocate(deadZoneScaling(maxNinstance))
allocate(probabilisticMultiplication(maxNinstance))
allocate(CFLfactor(maxNinstance))
allocate(fEdgeMultiplication(maxNinstance))
allocate(linetensionEffect(maxNinstance))
allocate(edgeJogFactor(maxNinstance))
CoverA = 0.0_pReal
mu = 0.0_pReal
atomicVolume = 0.0_pReal
Dsd0 = -1.0_pReal
selfDiffusionEnergy = 0.0_pReal
aTolRho = 0.0_pReal
aTolShear = 0.0_pReal
significantRho = 0.0_pReal
significantN = 0.0_pReal
nu = 0.0_pReal
Cslip66 = 0.0_pReal
Cslip3333 = 0.0_pReal
cutoffRadius = -1.0_pReal
doublekinkwidth = 0.0_pReal
solidSolutionEnergy = 0.0_pReal
solidSolutionSize = 0.0_pReal
solidSolutionConcentration = 0.0_pReal
pParam = 1.0_pReal
qParam = 1.0_pReal
viscosity = 0.0_pReal
fattack = 0.0_pReal
rhoSglScatter = 0.0_pReal
rhoSglRandom = 0.0_pReal
rhoSglRandomBinning = 1.0_pReal
surfaceTransmissivity = 1.0_pReal
grainboundaryTransmissivity = -1.0_pReal
CFLfactor = 2.0_pReal
fEdgeMultiplication = 0.0_pReal
linetensionEffect = 0.0_pReal
edgeJogFactor = 1.0_pReal
shortRangeStressCorrection = .false.
deadZoneScaling = .false.
probabilisticMultiplication = .false.
allocate(rhoSglEdgePos0(lattice_maxNslipFamily,maxNinstance))
allocate(rhoSglEdgeNeg0(lattice_maxNslipFamily,maxNinstance))
allocate(rhoSglScrewPos0(lattice_maxNslipFamily,maxNinstance))
allocate(rhoSglScrewNeg0(lattice_maxNslipFamily,maxNinstance))
allocate(rhoDipEdge0(lattice_maxNslipFamily,maxNinstance))
allocate(rhoDipScrew0(lattice_maxNslipFamily,maxNinstance))
allocate(burgersPerSlipFamily(lattice_maxNslipFamily,maxNinstance))
allocate(lambda0PerSlipFamily(lattice_maxNslipFamily,maxNinstance))
allocate(interactionSlipSlip(lattice_maxNinteraction,maxNinstance))
rhoSglEdgePos0 = -1.0_pReal
rhoSglEdgeNeg0 = -1.0_pReal
rhoSglScrewPos0 = -1.0_pReal
rhoSglScrewNeg0 = -1.0_pReal
rhoDipEdge0 = -1.0_pReal
rhoDipScrew0 = -1.0_pReal
burgersPerSlipFamily = 0.0_pReal
lambda0PerSlipFamily = 0.0_pReal
interactionSlipSlip = 0.0_pReal
allocate(minDipoleHeightPerSlipFamily(lattice_maxNslipFamily,2,maxNinstance))
allocate(peierlsStressPerSlipFamily(lattice_maxNslipFamily,2,maxNinstance))
minDipoleHeightPerSlipFamily = -1.0_pReal
peierlsStressPerSlipFamily = 0.0_pReal
allocate(nonSchmidCoeff(lattice_maxNonSchmid,maxNinstance))
nonSchmidCoeff = 0.0_pReal
added non-schmid structure
2013-01-22 05:20:28 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*** readout data from material.config file
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
rewind(myFile)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
line = ''
polishing, adding _pInt etc. where applicable post_results now handels zero increment different (like FEM, it is always there even if it is not counted)
2012-02-13 19:48:07 +05:30
section = 0_pInt
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
removed parameter check complaining about unknown keyword as default case. there are just too many keywords that do not trigger a variable assignment to cope with them (e.g. /nonlocal/, constitution, etc.) I suggest to kill similar logic from other files if present somewhere...
2012-02-14 05:00:59 +05:30
do while (IO_lc(IO_getTag(line,'<','>')) /= 'phase') ! wind forward to <phase>
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
read(myFile,'(a1024)',END=100) line
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
enddo
removed parameter check complaining about unknown keyword as default case. there are just too many keywords that do not trigger a variable assignment to cope with them (e.g. /nonlocal/, constitution, etc.) I suggest to kill similar logic from other files if present somewhere...
2012-02-14 05:00:59 +05:30
do ! read thru sections of phase part
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
read(myFile,'(a1024)',END=100) line
removed parameter check complaining about unknown keyword as default case. there are just too many keywords that do not trigger a variable assignment to cope with them (e.g. /nonlocal/, constitution, etc.) I suggest to kill similar logic from other files if present somewhere...
2012-02-14 05:00:59 +05:30
if (IO_isBlank(line)) cycle ! skip empty lines
if (IO_getTag(line,'<','>') /= '') exit ! stop at next part
if (IO_getTag(line,'[',']') /= '') then ! next section
section = section + 1_pInt ! advance section counter
* added a new material parameter "surfaceTransmissivity" (default value 1.0) which allows to change the transmissivity of the material surface between 0 and 1 * now complaining when encountering an unknown nonlocal parameter in material.config * use same error ID for all material parameters out of bounds * symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity) * switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
2011-02-16 22:05:38 +05:30
cycle
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
endif
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (section > 0_pInt .and. phase_plasticity(section) == CONSTITUTIVE_NONLOCAL_LABEL) then ! one of my sections
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
i = phase_plasticityInstance(section) ! which instance of my plasticity is present phase
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
positions = IO_stringPos(line,maxNchunks)
removed parameter check complaining about unknown keyword as default case. there are just too many keywords that do not trigger a variable assignment to cope with them (e.g. /nonlocal/, constitution, etc.) I suggest to kill similar logic from other files if present somewhere...
2012-02-14 05:00:59 +05:30
tag = IO_lc(IO_stringValue(line,positions,1_pInt)) ! extract key
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
select case(tag)
constitutive models need to skip new keyword elasticity when reading their arameters
2012-03-15 14:52:24 +05:30
case('plasticity','elasticity','/nonlocal/')
* added a new material parameter "surfaceTransmissivity" (default value 1.0) which allows to change the transmissivity of the material surface between 0 and 1 * now complaining when encountering an unknown nonlocal parameter in material.config * use same error ID for all material parameters out of bounds * symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity) * switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
2011-02-16 22:05:38 +05:30
cycle
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
case ('(output)')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Noutput(i) = Noutput(i) + 1_pInt
constitutive_nonlocal_output(Noutput(i),i) = IO_lc(IO_stringValue(line,positions,2_pInt))
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
case ('lattice_structure')
polishing, adding _pInt etc. where applicable post_results now handels zero increment different (like FEM, it is always there even if it is not counted)
2012-02-13 19:48:07 +05:30
constitutive_nonlocal_structureName(i) = IO_lc(IO_stringValue(line,positions,2_pInt))
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
configNchunks = lattice_configNchunks(constitutive_nonlocal_structureName(i))
Nchunks_SlipFamilies = configNchunks(1)
Nchunks_SlipSlip = configNchunks(3)
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
case ('c/a_ratio','covera_ratio')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
CoverA(i) = IO_floatValue(line,positions,2_pInt)
moved stiffness tensor calculation to lattice introduced 'isotropic' and 'orthorhombic' lattice types to use corresponding symmetries in stiffness tensor. intended to be used with non-crystal plasticity models (j2, constitutive_none with isotropic, cubic or orthotropic elasticity).
2013-01-22 03:27:26 +05:30
case ('c11')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Cslip66(1,1,i) = IO_floatValue(line,positions,2_pInt)
moved stiffness tensor calculation to lattice introduced 'isotropic' and 'orthorhombic' lattice types to use corresponding symmetries in stiffness tensor. intended to be used with non-crystal plasticity models (j2, constitutive_none with isotropic, cubic or orthotropic elasticity).
2013-01-22 03:27:26 +05:30
case ('c12')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Cslip66(1,2,i) = IO_floatValue(line,positions,2_pInt)
moved stiffness tensor calculation to lattice introduced 'isotropic' and 'orthorhombic' lattice types to use corresponding symmetries in stiffness tensor. intended to be used with non-crystal plasticity models (j2, constitutive_none with isotropic, cubic or orthotropic elasticity).
2013-01-22 03:27:26 +05:30
case ('c13')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Cslip66(1,3,i) = IO_floatValue(line,positions,2_pInt)
moved stiffness tensor calculation to lattice introduced 'isotropic' and 'orthorhombic' lattice types to use corresponding symmetries in stiffness tensor. intended to be used with non-crystal plasticity models (j2, constitutive_none with isotropic, cubic or orthotropic elasticity).
2013-01-22 03:27:26 +05:30
case ('c22')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Cslip66(2,2,i) = IO_floatValue(line,positions,2_pInt)
moved stiffness tensor calculation to lattice introduced 'isotropic' and 'orthorhombic' lattice types to use corresponding symmetries in stiffness tensor. intended to be used with non-crystal plasticity models (j2, constitutive_none with isotropic, cubic or orthotropic elasticity).
2013-01-22 03:27:26 +05:30
case ('c23')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Cslip66(2,3,i) = IO_floatValue(line,positions,2_pInt)
moved stiffness tensor calculation to lattice introduced 'isotropic' and 'orthorhombic' lattice types to use corresponding symmetries in stiffness tensor. intended to be used with non-crystal plasticity models (j2, constitutive_none with isotropic, cubic or orthotropic elasticity).
2013-01-22 03:27:26 +05:30
case ('c33')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Cslip66(3,3,i) = IO_floatValue(line,positions,2_pInt)
moved stiffness tensor calculation to lattice introduced 'isotropic' and 'orthorhombic' lattice types to use corresponding symmetries in stiffness tensor. intended to be used with non-crystal plasticity models (j2, constitutive_none with isotropic, cubic or orthotropic elasticity).
2013-01-22 03:27:26 +05:30
case ('c44')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Cslip66(4,4,i) = IO_floatValue(line,positions,2_pInt)
moved stiffness tensor calculation to lattice introduced 'isotropic' and 'orthorhombic' lattice types to use corresponding symmetries in stiffness tensor. intended to be used with non-crystal plasticity models (j2, constitutive_none with isotropic, cubic or orthotropic elasticity).
2013-01-22 03:27:26 +05:30
case ('c55')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Cslip66(5,5,i) = IO_floatValue(line,positions,2_pInt)
moved stiffness tensor calculation to lattice introduced 'isotropic' and 'orthorhombic' lattice types to use corresponding symmetries in stiffness tensor. intended to be used with non-crystal plasticity models (j2, constitutive_none with isotropic, cubic or orthotropic elasticity).
2013-01-22 03:27:26 +05:30
case ('c66')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Cslip66(6,6,i) = IO_floatValue(line,positions,2_pInt)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
case ('nslip')
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
do f = 1_pInt, Nchunks_SlipFamilies
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Nslip(f,i) = IO_intValue(line,positions,1_pInt+f)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
case ('rhosgledgepos0')
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
do f = 1_pInt, Nchunks_SlipFamilies
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoSglEdgePos0(f,i) = IO_floatValue(line,positions,1_pInt+f)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
case ('rhosgledgeneg0')
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
do f = 1_pInt, Nchunks_SlipFamilies
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoSglEdgeNeg0(f,i) = IO_floatValue(line,positions,1_pInt+f)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
case ('rhosglscrewpos0')
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
do f = 1_pInt, Nchunks_SlipFamilies
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoSglScrewPos0(f,i) = IO_floatValue(line,positions,1_pInt+f)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
case ('rhosglscrewneg0')
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
do f = 1_pInt, Nchunks_SlipFamilies
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoSglScrewNeg0(f,i) = IO_floatValue(line,positions,1_pInt+f)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
case ('rhodipedge0')
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
do f = 1_pInt, Nchunks_SlipFamilies
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoDipEdge0(f,i) = IO_floatValue(line,positions,1_pInt+f)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
case ('rhodipscrew0')
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
do f = 1_pInt, Nchunks_SlipFamilies
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoDipScrew0(f,i) = IO_floatValue(line,positions,1_pInt+f)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
added dislocation multiplication to dotState in constitutive_nonlocal.f90 cleaned up debugging output statements to *.out file
2009-08-12 16:52:02 +05:30
case ('lambda0')
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
do f = 1_pInt, Nchunks_SlipFamilies
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
lambda0PerSlipFamily(f,i) = IO_floatValue(line,positions,1_pInt+f)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
case ('burgers')
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
do f = 1_pInt, Nchunks_SlipFamilies
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
burgersPerSlipFamily(f,i) = IO_floatValue(line,positions,1_pInt+f)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
case('cutoffradius','r')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
cutoffRadius(i) = IO_floatValue(line,positions,2_pInt)
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
case('minimumdipoleheightedge','ddipminedge')
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
do f = 1_pInt, Nchunks_SlipFamilies
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
minDipoleHeightPerSlipFamily(f,1_pInt,i) = IO_floatValue(line,positions,1_pInt+f)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
case('minimumdipoleheightscrew','ddipminscrew')
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
do f = 1_pInt, Nchunks_SlipFamilies
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
minDipoleHeightPerSlipFamily(f,2_pInt,i) = IO_floatValue(line,positions,1_pInt+f)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
case('atomicvolume')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
atomicVolume(i) = IO_floatValue(line,positions,2_pInt)
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
case('selfdiffusionprefactor','dsd0')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Dsd0(i) = IO_floatValue(line,positions,2_pInt)
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
case('selfdiffusionenergy','qsd')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
selfDiffusionEnergy(i) = IO_floatValue(line,positions,2_pInt)
improved accuracy of "accumulated shear": now added as a state variable, which facilitates integration (dotstate=shearrate); needs absolute tolerance value for state integration in material.config
2013-04-04 19:07:14 +05:30
case('atol_rho','atol_density','absolutetolerancedensity','absolutetolerance_density')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
aTolRho(i) = IO_floatValue(line,positions,2_pInt)
improved accuracy of "accumulated shear": now added as a state variable, which facilitates integration (dotstate=shearrate); needs absolute tolerance value for state integration in material.config
2013-04-04 19:07:14 +05:30
case('atol_shear','atol_plasticshear','atol_accumulatedshear','absolutetoleranceshear','absolutetolerance_shear')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
aTolShear(i) = IO_floatValue(line,positions,2_pInt)
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
case('significantrho','significant_rho','significantdensity','significant_density')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
significantRho(i) = IO_floatValue(line,positions,2_pInt)
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
case('significantn','significant_n','significantdislocations','significant_dislcations')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
significantN(i) = IO_floatValue(line,positions,2_pInt)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
case ('interaction_slipslip')
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
do it = 1_pInt, Nchunks_SlipSlip
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
interactionSlipSlip(it,i) = IO_floatValue(line,positions,1_pInt+it)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
introduced material parameter "linetension", which indicates how strong the effect of the line tension on the hardening coefficients is
2012-10-05 20:12:41 +05:30
case('linetension','linetensioneffect','linetension_effect')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
linetensionEffect(i) = IO_floatValue(line,positions,2_pInt)
new material parameter "egdejog": fraction of annihilated screw dipoles that forms edge jogs
2012-12-23 18:26:15 +05:30
case('edgejog','edgejogs','edgejogeffect','edgejog_effect')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
edgeJogFactor(i) = IO_floatValue(line,positions,2_pInt)
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
case('peierlsstressedge','peierlsstress_edge')
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
do f = 1_pInt, Nchunks_SlipFamilies
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
peierlsStressPerSlipFamily(f,1_pInt,i) = IO_floatValue(line,positions,1_pInt+f)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
case('peierlsstressscrew','peierlsstress_screw')
parsing of material.config file now expects adequate number of (slip/twin) family entries according to given lattice type. used to read up to maxN, but that caused unnecessary IO_warnings...
2013-02-15 03:54:55 +05:30
do f = 1_pInt, Nchunks_SlipFamilies
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
peierlsStressPerSlipFamily(f,2_pInt,i) = IO_floatValue(line,positions,1_pInt+f)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
corrected kinetics law and changed parameters. For solid solution hardening there are 3 parameters: the activation energy, the concentration of obstacles that determines the activation length and meanfreepath, and the obstacle size that determines the activation volume. For the Peierls mechanism there is: the width of doublekinks that determines the activation volume and the Peierls stress for edge and screw. Still testing needed to check whether the current formulation makes sense or not.
2012-02-03 18:20:54 +05:30
case('doublekinkwidth')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
doublekinkwidth(i) = IO_floatValue(line,positions,2_pInt)
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
case('solidsolutionenergy')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
solidSolutionEnergy(i) = IO_floatValue(line,positions,2_pInt)
corrected kinetics law and changed parameters. For solid solution hardening there are 3 parameters: the activation energy, the concentration of obstacles that determines the activation length and meanfreepath, and the obstacle size that determines the activation volume. For the Peierls mechanism there is: the width of doublekinks that determines the activation volume and the Peierls stress for edge and screw. Still testing needed to check whether the current formulation makes sense or not.
2012-02-03 18:20:54 +05:30
case('solidsolutionsize')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
solidSolutionSize(i) = IO_floatValue(line,positions,2_pInt)
corrected kinetics law and changed parameters. For solid solution hardening there are 3 parameters: the activation energy, the concentration of obstacles that determines the activation length and meanfreepath, and the obstacle size that determines the activation volume. For the Peierls mechanism there is: the width of doublekinks that determines the activation volume and the Peierls stress for edge and screw. Still testing needed to check whether the current formulation makes sense or not.
2012-02-03 18:20:54 +05:30
case('solidsolutionconcentration')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
solidSolutionConcentration(i) = IO_floatValue(line,positions,2_pInt)
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
case('p')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
pParam(i) = IO_floatValue(line,positions,2_pInt)
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
case('q')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
qParam(i) = IO_floatValue(line,positions,2_pInt)
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
case('viscosity','glideviscosity')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
viscosity(i) = IO_floatValue(line,positions,2_pInt)
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
case('attackfrequency','fattack')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
fattack(i) = IO_floatValue(line,positions,2_pInt)
* now able to introduce some scatter in the initial single dislocation density (only for nonlocal model!) ; setting the parameter "rhoSglScatter" to a positive value generates a gauss distribution for the dislocation density with standard deviation equal to "rhoSglScatter" * dislocation stress calculation is only done for nonlocal constitution ("/nonlocal/" keyword is present in material.config)
2011-02-04 21:11:32 +05:30
case('rhosglscatter')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoSglScatter(i) = IO_floatValue(line,positions,2_pInt)
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
case('rhosglrandom')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoSglRandom(i) = IO_floatValue(line,positions,2_pInt)
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
case('rhosglrandombinning')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoSglRandomBinning(i) = IO_floatValue(line,positions,2_pInt)
* added a new material parameter "surfaceTransmissivity" (default value 1.0) which allows to change the transmissivity of the material surface between 0 and 1 * now complaining when encountering an unknown nonlocal parameter in material.config * use same error ID for all material parameters out of bounds * symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity) * switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
2011-02-16 22:05:38 +05:30
case('surfacetransmissivity')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
surfaceTransmissivity(i) = IO_floatValue(line,positions,2_pInt)
added possibility to assign the transmissivity of grain boundaries a fixed value by use of the keyword "grainboundaryTransmissivity" in the material.config; grain boundaries are identified as interfaces between areas of different material_texture
2012-08-03 20:02:49 +05:30
case('grainboundarytransmissivity')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
grainboundaryTransmissivity(i) = IO_floatValue(line,positions,2_pInt)
corrected debug output of evolution rates; added possibility to change the safety factor for the CFL flux condition via the material.config file; updated the example material.config due to recent changes in nonlocal constitutive
2012-08-16 14:43:38 +05:30
case('cflfactor')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
CFLfactor(i) = IO_floatValue(line,positions,2_pInt)
introduced factor to control edge contribution to multiplication; dislocation density below a single dislocation per IP considered not significant
2012-09-04 22:26:37 +05:30
case('fedgemultiplication','edgemultiplicationfactor','edgemultiplication')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
fEdgeMultiplication(i) = IO_floatValue(line,positions,2_pInt)
adding switch in material.config for short range stress correction
2012-03-15 20:28:12 +05:30
case('shortrangestresscorrection')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
shortRangeStressCorrection(i) = IO_floatValue(line,positions,2_pInt) > 0.0_pReal
added non-schmid structure
2013-01-22 05:20:28 +05:30
case ('nonschmid_coefficients')
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
do f = 1_pInt, lattice_maxNonSchmid
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
nonSchmidCoeff(f,i) = IO_floatValue(line,positions,1_pInt+f)
Made corresponding changes in the constitutive files and DAMASK_spectral_driver and interface files with respect to the changes made to IO.f90
2013-02-06 22:15:34 +05:30
enddo
added possibility to scale the plastic shearrate by the ratio of mobile to total density; keyword in material.config is "deadZone"
2012-10-29 18:19:28 +05:30
case('deadzonescaling','deadzone','deadscaling')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
deadZoneScaling(i) = IO_floatValue(line,positions,2_pInt) > 0.0_pReal
made random multiplication switchable via material.config
2012-12-03 18:29:38 +05:30
case('probabilisticmultiplication','randomsources','randommultiplication','discretesources')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
probabilisticMultiplication(i) = IO_floatValue(line,positions,2_pInt) > 0.0_pReal
added check for unknown keywords again (in agreement with Philip ;-) indeed 'constitution' and '/nonlocal/' are the only keywords which have to be ignored, since they are parsed elsewhere
2012-02-14 14:52:37 +05:30
case default
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
call IO_error(210_pInt,ext_msg=tag//' ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
end select
endif
enddo
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
100 do i = 1_pInt,maxNinstance
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
constitutive_nonlocal_structure(i) = &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
lattice_initializeStructure(constitutive_nonlocal_structureName(i), CoverA(i)) ! our lattice structure is defined in the material.config file by the structureName (and the c/a ratio)
convergence of state in crystallite is now tested as follows: (state < relevant state) or (residuum < relative tolerance * state) since the relevant value for the state variables depend on their nature and can vary by large scales (e.g. volume fraction: 1e-10, dislocation density: 1e5) it is not possible to set a unique value. instead the constitutive law has to decide what is relevant. therefore, all constitutive laws now read in parameters from the material.config that determine the values for relevantState [@luc: in dislobased law relevant State is for the moment generally set to 1e-200, so no additional parameters necessary in material.config. if you also want this feature, we can still implement it, no big deal] - added sanity checks in constitutive_nonlocal.f90 - corrected coordinate transformation for backstress calculation in constitutive_nonlocal.f90 - corrected equations for evolution of dipole dislocation densities (athermal annihilation and formation by glide)
2009-09-18 21:07:14 +05:30
myStructure = constitutive_nonlocal_structure(i)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
!*** sanity checks
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (myStructure < 1_pInt) &
call IO_error(205_pInt,e=i)
if (sum(Nslip(:,i)) <= 0_pInt) &
call IO_error(211_pInt,ext_msg='Nslip ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do o = 1_pInt,maxval(phase_Noutput)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if(len(constitutive_nonlocal_output(o,i)) > 64_pInt) &
call IO_error(666_pInt)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
enddo
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do f = 1_pInt,lattice_maxNslipFamily
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (Nslip(f,i) > 0_pInt) then
if (rhoSglEdgePos0(f,i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoSglEdgePos0 ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (rhoSglEdgeNeg0(f,i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoSglEdgeNeg0 ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (rhoSglScrewPos0(f,i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoSglScrewPos0 ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (rhoSglScrewNeg0(f,i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoSglScrewNeg0 ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (rhoDipEdge0(f,i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoDipEdge0 ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (rhoDipScrew0(f,i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoDipScrew0 ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (burgersPerSlipFamily(f,i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='Burgers ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (lambda0PerSlipFamily(f,i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='lambda0 ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (minDipoleHeightPerSlipFamily(f,1,i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='minimumDipoleHeightEdge ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (minDipoleHeightPerSlipFamily(f,2,i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='minimumDipoleHeightScrew ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (peierlsStressPerSlipFamily(f,1,i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='peierlsStressEdge ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (peierlsStressPerSlipFamily(f,2,i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='peierlsStressScrew ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
added dislocation multiplication to dotState in constitutive_nonlocal.f90 cleaned up debugging output statements to *.out file
2009-08-12 16:52:02 +05:30
endif
enddo
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (any(interactionSlipSlip(1:maxval(lattice_interactionSlipSlip(:,:,myStructure)),i) < 0.0_pReal)) &
call IO_error(211_pInt,ext_msg='interaction_SlipSlip ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (linetensionEffect(i) < 0.0_pReal .or. linetensionEffect(i) > 1.0_pReal) &
call IO_error(211_pInt,ext_msg='linetension ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (edgeJogFactor(i) < 0.0_pReal .or. edgeJogFactor(i) > 1.0_pReal) &
call IO_error(211_pInt,ext_msg='edgejog ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (cutoffRadius(i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='r ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (atomicVolume(i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='atomicVolume ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (Dsd0(i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='selfDiffusionPrefactor ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (selfDiffusionEnergy(i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='selfDiffusionEnergy ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (aTolRho(i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='aTol_rho ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (aTolShear(i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='aTol_shear ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (significantRho(i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='significantRho ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (significantN(i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='significantN ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (doublekinkwidth(i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='doublekinkwidth ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (solidSolutionEnergy(i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='solidSolutionEnergy ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (solidSolutionSize(i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='solidSolutionSize ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (solidSolutionConcentration(i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='solidSolutionConcentration ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (pParam(i) <= 0.0_pReal .or. pParam(i) > 1.0_pReal) &
call IO_error(211_pInt,ext_msg='p ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (qParam(i) < 1.0_pReal .or. qParam(i) > 2.0_pReal) &
call IO_error(211_pInt,ext_msg='q ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (viscosity(i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='viscosity ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (fattack(i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='attackFrequency ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (rhoSglScatter(i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoSglScatter ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (rhoSglRandom(i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoSglRandom ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (rhoSglRandomBinning(i) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoSglRandomBinning ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (surfaceTransmissivity(i) < 0.0_pReal .or. surfaceTransmissivity(i) > 1.0_pReal) &
call IO_error(211_pInt,ext_msg='surfaceTransmissivity ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (grainboundaryTransmissivity(i) > 1.0_pReal) &
call IO_error(211_pInt,ext_msg='grainboundaryTransmissivity ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (CFLfactor(i) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='CFLfactor ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
if (fEdgeMultiplication(i) < 0.0_pReal .or. fEdgeMultiplication(i) > 1.0_pReal) &
call IO_error(211_pInt,ext_msg='edgemultiplicationfactor ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
* kinetics now according to Mohles; enables solid solution hardening * changed example in material.config accordingly
2011-01-26 15:47:42 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
!*** determine total number of active slip systems
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Nslip(1:lattice_maxNslipFamily,i) = min(lattice_NslipSystem(1:lattice_maxNslipFamily,myStructure), &
Nslip(1:lattice_maxNslipFamily,i) ) ! we can't use more slip systems per family than specified in lattice
totalNslip(i) = sum(Nslip(1:lattice_maxNslipFamily,i))
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
enddo
!*** allocation of variables whose size depends on the total number of active slip systems
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
maxTotalNslip = maxval(totalNslip)
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
allocate(iRhoEPU(maxTotalNslip, maxNinstance))
allocate(iRhoENU(maxTotalNslip, maxNinstance))
allocate(iRhoSPU(maxTotalNslip, maxNinstance))
allocate(iRhoSNU(maxTotalNslip, maxNinstance))
allocate(iRhoEPB(maxTotalNslip, maxNinstance))
allocate(iRhoENB(maxTotalNslip, maxNinstance))
allocate(iRhoSPB(maxTotalNslip, maxNinstance))
allocate(iRhoSNB(maxTotalNslip, maxNinstance))
allocate(iRhoED(maxTotalNslip, maxNinstance))
allocate(iRhoSD(maxTotalNslip, maxNinstance))
allocate(iGamma(maxTotalNslip, maxNinstance))
allocate(iRhoF(maxTotalNslip, maxNinstance))
rename state variable index for critical resolved shear stress from "iTau" to "iTauF"
2013-05-24 03:02:36 +05:30
allocate(iTauF(maxTotalNslip, maxNinstance))
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
allocate(iTauB(maxTotalNslip, maxNinstance))
allocate(iVEP(maxTotalNslip, maxNinstance))
allocate(iVEN(maxTotalNslip, maxNinstance))
allocate(iVSP(maxTotalNslip, maxNinstance))
allocate(iVSN(maxTotalNslip, maxNinstance))
allocate(iDE(maxTotalNslip, maxNinstance))
allocate(iDS(maxTotalNslip, maxNinstance))
iRhoEPU = 0_pInt
iRhoENU = 0_pInt
iRhoSPU = 0_pInt
iRhoSNU = 0_pInt
iRhoEPB = 0_pInt
iRhoENB = 0_pInt
iRhoSPB = 0_pInt
iRhoSNB = 0_pInt
iRhoED = 0_pInt
iRhoSD = 0_pInt
iGamma = 0_pInt
iRhoF = 0_pInt
rename state variable index for critical resolved shear stress from "iTau" to "iTauF"
2013-05-24 03:02:36 +05:30
iTauF = 0_pInt
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
iTauB = 0_pInt
iVEP = 0_pInt
iVEN = 0_pInt
iVSP = 0_pInt
iVSN = 0_pInt
iDE = 0_pInt
iDS = 0_pInt
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(burgers(maxTotalNslip, maxNinstance))
burgers = 0.0_pReal
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(lambda0(maxTotalNslip, maxNinstance))
lambda0 = 0.0_pReal
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(minDipoleHeight(maxTotalNslip,2,maxNinstance))
minDipoleHeight = -1.0_pReal
* kinetics now according to Mohles; enables solid solution hardening * changed example in material.config accordingly
2011-01-26 15:47:42 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(forestProjectionEdge(maxTotalNslip, maxTotalNslip, maxNinstance))
forestProjectionEdge = 0.0_pReal
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(forestProjectionScrew(maxTotalNslip, maxTotalNslip, maxNinstance))
forestProjectionScrew = 0.0_pReal
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(interactionMatrixSlipSlip(maxTotalNslip, maxTotalNslip, maxNinstance))
interactionMatrixSlipSlip = 0.0_pReal
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(lattice2slip(1:3, 1:3, maxTotalNslip, maxNinstance))
lattice2slip = 0.0_pReal
* internal stress calculation now considers dead dislocations always at the interface, not at the center of the IP volume; used to merge them together with "normal" dislocations for stress calculation. * dislocation flux is blocked if we encounter a sign change in the resolved shear stress from the central ip to the neighbor * do not set density to zero if below certain threshold; this creates an artificial sink term
2011-08-02 16:47:45 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(sourceProbability(maxTotalNslip, homogenization_maxNgrains, mesh_maxNips, mesh_NcpElems))
sourceProbability = 2.0_pReal
modified multiplication term in order to mimic a discrete distribution of dislocation sources; namely, at high resolution, when there is (on average) less than one source per integration point, a random generator decides whether there is a source in the finite volume or not; if there is one, then the multiplication turns independent of the density and depends only on the velocity/volume times a scalar factor; recovers density*velocity/lambda for large finite volumes
2012-10-19 17:10:17 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(shearrate(maxTotalNslip,homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems))
shearrate = 0.0_pReal
improved accuracy of accumulated shear output: take shear rate from LpAndItsTangent for time integration
2013-05-21 15:34:52 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(rhoDotFluxOutput(maxTotalNslip, 8, homogenization_maxNgrains, mesh_maxNips, mesh_NcpElems))
allocate(rhoDotMultiplicationOutput(maxTotalNslip, 2, homogenization_maxNgrains, mesh_maxNips, mesh_NcpElems))
allocate(rhoDotSingle2DipoleGlideOutput(maxTotalNslip, 2, homogenization_maxNgrains, mesh_maxNips, mesh_NcpElems))
allocate(rhoDotAthermalAnnihilationOutput(maxTotalNslip, 2, homogenization_maxNgrains, mesh_maxNips, mesh_NcpElems))
allocate(rhoDotThermalAnnihilationOutput(maxTotalNslip, 2, homogenization_maxNgrains, mesh_maxNips, mesh_NcpElems))
allocate(rhoDotEdgeJogsOutput(maxTotalNslip, homogenization_maxNgrains, mesh_maxNips, mesh_NcpElems))
rhoDotFluxOutput = 0.0_pReal
rhoDotMultiplicationOutput = 0.0_pReal
rhoDotSingle2DipoleGlideOutput = 0.0_pReal
rhoDotAthermalAnnihilationOutput = 0.0_pReal
rhoDotThermalAnnihilationOutput = 0.0_pReal
rhoDotEdgeJogsOutput = 0.0_pReal
added output for dislocation evolution rate associated with flux: "rho_dot_flux"
2010-03-10 15:19:40 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(compatibility(2,maxTotalNslip, maxTotalNslip, mesh_maxNipNeighbors, mesh_maxNips, mesh_NcpElems))
compatibility = 0.0_pReal
* in nonlocal model: dislocation flux now with valid description also for large angle grain boundaries; transfer of dislocations from one slip system to the other according to new variable "constitutive_nonlocal_compatibility" which depends on the angle between slip plane normals and slip directions and is updated once per cycle in function "crystallite_orientations" * reactivated debugging functionality for "non-standard" integration methods
2010-10-12 18:38:54 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(peierlsStress(maxTotalNslip,2,maxNinstance))
peierlsStress = 0.0_pReal
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
allocate(colinearSystem(maxTotalNslip,maxNinstance))
colinearSystem = 0_pInt
introduced annihilation of screw dipoles; so far no thermal activation, but immediate annihilation; each annihilated screw dipole leaves behind an edge jog of same length as the dipole height on the collinear system
2012-08-14 17:56:20 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
do i = 1,maxNinstance
myStructure = constitutive_nonlocal_structure(i) ! lattice structure of this instance
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
!*** Inverse lookup of my slip system family and the slip system in lattice
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
l = 0_pInt
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do f = 1_pInt,lattice_maxNslipFamily
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
do s = 1_pInt,Nslip(f,i)
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
l = l + 1_pInt
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
slipFamily(l,i) = f
slipSystemLattice(l,i) = sum(lattice_NslipSystem(1:f-1_pInt, myStructure)) + s
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
enddo; enddo
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
!*** determine size of state array
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
ns = totalNslip(i)
constitutive_nonlocal_sizeDotState(i) = int(size(BASICSTATES),pInt) * ns
constitutive_nonlocal_sizeDependentState(i) = int(size(DEPENDENTSTATES),pInt) * ns
constitutive_nonlocal: dislocation velocities are stored in the state, so we actually now have three "parts" of the state, the basic states that are updated by "constitutive_dotState" come first, then the dependent states that are calculated by "constitutive_microstructure" follow, and finally we have a last part reserved for other variables that just use the memory reserved by the state array and are updated somewhere else. constitutive: LpAndItsTangent does not need the full state, but only the local state, so changed that at least for the nonlocal constitutive law
2011-11-04 18:42:17 +05:30
constitutive_nonlocal_sizeState(i) = constitutive_nonlocal_sizeDotState(i) &
+ constitutive_nonlocal_sizeDependentState(i) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
+ int(size(OTHERSTATES),pInt) * ns
constitutive_nonlocal: dislocation velocities are stored in the state, so we actually now have three "parts" of the state, the basic states that are updated by "constitutive_dotState" come first, then the dependent states that are calculated by "constitutive_microstructure" follow, and finally we have a last part reserved for other variables that just use the memory reserved by the state array and are updated somewhere else. constitutive: LpAndItsTangent does not need the full state, but only the local state, so changed that at least for the nonlocal constitutive law
2011-11-04 18:42:17 +05:30
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
!*** determine indices to state array
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iRhoEPU(s,i) = s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iRhoENU(s,i) = iRhoEPU(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iRhoSPU(s,i) = iRhoENU(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iRhoSNU(s,i) = iRhoSPU(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iRhoEPB(s,i) = iRhoSNU(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iRhoENB(s,i) = iRhoEPB(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iRhoSPB(s,i) = iRhoENB(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iRhoSNB(s,i) = iRhoSPB(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iRhoED(s,i) = iRhoSNB(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iRhoSD(s,i) = iRhoED(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iGamma(s,i) = iRhoSD(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iRhoF(s,i) = iGamma(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
rename state variable index for critical resolved shear stress from "iTau" to "iTauF"
2013-05-24 03:02:36 +05:30
iTauF(s,i) = iRhoF(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
rename state variable index for critical resolved shear stress from "iTau" to "iTauF"
2013-05-24 03:02:36 +05:30
iTauB(s,i) = iTauF(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iVEP(s,i) = iTauB(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iVEN(s,i) = iVEP(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iVSP(s,i) = iVEN(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iVSN(s,i) = iVSP(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iDE(s,i) = iVSN(ns,i) + s
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
forall (s = 1:ns) &
corrected named indices for state variables; however, not yet used
2013-05-24 02:00:06 +05:30
iDS(s,i) = iDE(ns,i) + s
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
if (iDS(ns,i) /= constitutive_nonlocal_sizeState(i)) & ! check if last index is equal to size of state
call IO_error(0_pInt, ext_msg = 'state indices not properly set ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +05:30
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
!*** determine size of postResults array
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
do o = 1_pInt,Noutput(i)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
select case(constitutive_nonlocal_output(o,i))
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
case( 'rho', &
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
'delta', &
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
'rho_edge', &
'rho_screw', &
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
'rho_sgl', &
'delta_sgl', &
'rho_sgl_edge', &
'rho_sgl_edge_pos', &
'rho_sgl_edge_neg', &
'rho_sgl_screw', &
'rho_sgl_screw_pos', &
'rho_sgl_screw_neg', &
'rho_sgl_mobile', &
'rho_sgl_edge_mobile', &
'rho_sgl_edge_pos_mobile', &
'rho_sgl_edge_neg_mobile', &
'rho_sgl_screw_mobile', &
'rho_sgl_screw_pos_mobile', &
'rho_sgl_screw_neg_mobile', &
'rho_sgl_immobile', &
'rho_sgl_edge_immobile', &
'rho_sgl_edge_pos_immobile', &
'rho_sgl_edge_neg_immobile', &
'rho_sgl_screw_immobile', &
'rho_sgl_screw_pos_immobile', &
'rho_sgl_screw_neg_immobile', &
'rho_dip', &
'delta_dip', &
'rho_dip_edge', &
'rho_dip_screw', &
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
'excess_rho', &
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
'excess_rho_edge', &
'excess_rho_screw', &
'rho_forest', &
'shearrate', &
'resolvedstress', &
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
'resolvedstress_external', &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
'resolvedstress_back', &
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
'resistance', &
'rho_dot', &
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
'rho_dot_sgl', &
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
'rho_dot_dip', &
'rho_dot_gen', &
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
'rho_dot_gen_edge', &
'rho_dot_gen_screw', &
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
'rho_dot_sgl2dip', &
added separate output of dipole formation rate for edge and screw
2012-08-27 21:27:31 +05:30
'rho_dot_sgl2dip_edge', &
'rho_dot_sgl2dip_screw', &
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
'rho_dot_ann_ath', &
'rho_dot_ann_the', &
added missing labels to list of available outputs
2012-08-27 18:11:57 +05:30
'rho_dot_ann_the_edge', &
'rho_dot_ann_the_screw', &
'rho_dot_edgejogs', &
added output for dislocation evolution rate associated with flux: "rho_dot_flux"
2010-03-10 15:19:40 +05:30
'rho_dot_flux', &
in nonlocal_microstructure: at the model surface the excess density is now extrapolated from opposite neighbor instead of assuming zero density; this results in better modeling of the dislocation stress at the surface restructured nonlocal_dotstate, to be able to easily switch on and off particular effects in the microstructure evolution nonlocal_dotstate now enforces a cutback when single density runs the risk of becoming negative; in the case of a state already below the relevantState dotState is set to zero introduced two new output variables: rho_dot_edge and rho_dot_screw
2010-05-21 14:21:15 +05:30
'rho_dot_flux_edge', &
'rho_dot_flux_screw', &
dislocation velocity output for each dislocation type (edge+,edge-,screw+,screw-) for fcc constitutive_nonlocal_kinetics is called only once and the result is used for each dislocation type, for all other lattice structures we have one call for each dislocation type
2012-01-26 13:13:36 +05:30
'velocity_edge_pos', &
'velocity_edge_neg', &
'velocity_screw_pos', &
'velocity_screw_neg', &
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
'slipdirection.x', &
'slipdirection.y', &
'slipdirection.z', &
'slipnormal.x', &
'slipnormal.y', &
'slipnormal.z', &
'fluxdensity_edge_pos.x', &
'fluxdensity_edge_pos.y', &
'fluxdensity_edge_pos.z', &
'fluxdensity_edge_neg.x', &
'fluxdensity_edge_neg.y', &
'fluxdensity_edge_neg.z', &
'fluxdensity_screw_pos.x', &
'fluxdensity_screw_pos.y', &
'fluxdensity_screw_pos.z', &
'fluxdensity_screw_neg.x', &
'fluxdensity_screw_neg.y', &
'fluxdensity_screw_neg.z', &
corrected typos
2012-01-26 18:20:04 +05:30
'maximumdipoleheight_edge', &
'maximumdipoleheight_screw', &
added possibility to scale the plastic shearrate by the ratio of mobile to total density; keyword in material.config is "deadZone"
2012-10-29 18:19:28 +05:30
'accumulatedshear', &
'boundarylayer' )
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
mySize = totalNslip(i)
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
case('dislocationstress')
* new output variable "internalstress", which gives the internal stress tensor * use "math_invert3x3" instead of "math_inv3x3" for inversion of Fe * for dislocation stress calculation: first regular case, then special case of dead dislocations in central ip * "dv_dtau" now given for each dislocation type, so is a (ns,4) array * deleted unused variables in "_LpAndItsTangent" * corrected contribution of deads in "_LpAndItsTangent" * the NaN variables defined in math did not give a proper NaN value, so use 0.0/0.0 again * neighbors with nonlocal constitution but local properties (i.e. /nonlocal/ flag not set) are also considered for incoming fluxes
2011-09-07 17:00:28 +05:30
mySize = 6_pInt
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
case default
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
call IO_error(212_pInt,ext_msg=constitutive_nonlocal_output(o,i)//'&
('//CONSTITUTIVE_NONLOCAL_LABEL//')')
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
end select
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (mySize > 0_pInt) then ! any meaningful output found
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
constitutive_nonlocal_sizePostResult(o,i) = mySize
constitutive_nonlocal_sizePostResults(i) = constitutive_nonlocal_sizePostResults(i) + mySize
endif
enddo
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
!*** elasticity matrix and shear modulus according to material.config
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
Cslip66(:,:,i) = lattice_symmetrizeC66(constitutive_nonlocal_structureName(i), Cslip66(:,:,i))
mu(i) = 0.2_pReal * ( Cslip66(1,1,i) - Cslip66(1,2,i) + 3.0_pReal*Cslip66(4,4,i)) ! (C11iso-C12iso)/2 with C11iso=(3*C11+2*C12+4*C44)/5 and C12iso=(C11+4*C12-2*C44)/5
nu(i) = (Cslip66(1,1,i) + 4.0_pReal*Cslip66(1,2,i) - 2.0_pReal*Cslip66(1,2,i)) &
/ (4.0_pReal*Cslip66(1,1,i) + 6.0_pReal*Cslip66(1,2,i) + 2.0_pReal*Cslip66(4,4,i)) ! C12iso/(C11iso+C12iso) with C11iso=(3*C11+2*C12+4*C44)/5 and C12iso=(C11+4*C12-2*C44)/5
Cslip66(1:6,1:6,i) = math_Mandel3333to66(math_Voigt66to3333(Cslip66(1:6,1:6,i)))
Cslip3333(1:3,1:3,1:3,1:3,i) = math_Voigt66to3333(Cslip66(1:6,1:6,i))
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
polishing, adding _pInt etc. where applicable post_results now handels zero increment different (like FEM, it is always there even if it is not counted)
2012-02-13 19:48:07 +05:30
do s1 = 1_pInt,ns
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
f = slipFamily(s1,i)
added dislocation multiplication to dotState in constitutive_nonlocal.f90 cleaned up debugging output statements to *.out file
2009-08-12 16:52:02 +05:30
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
!*** burgers vector, mean free path prefactor and minimum dipole distance for each slip system
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
burgers(s1,i) = burgersPerSlipFamily(f,i)
lambda0(s1,i) = lambda0PerSlipFamily(f,i)
minDipoleHeight(s1,1:2,i) = minDipoleHeightPerSlipFamily(f,1:2,i)
peierlsStress(s1,1:2,i) = peierlsStressPerSlipFamily(f,1:2,i)
constitutive_nonlocal - take orientation gradients into account when calculating dislocation stress and dislocation fluxes - hard coded value for nu - changed kinetics (parameter G0 is currently defined as a parameter, needs to be read from material.config) - added some output statements constitutive: - some functions and subroutines needed additional input variables for passing to constitutive_nonlocal crystallite: - some functions and subroutines needed additional input variables for passing to constitutive - call microstructure with current temperature, Fp, Fe, not "sub0" values - show number of IPs, that are "onTrack" instead of those not "onTrack" - calculate Fe at beginning of substep, since we need it for state preguess
2009-10-07 21:01:52 +05:30
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do s2 = 1_pInt,ns
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
!*** calculation of forest projections for edge and screw dislocations. s2 acts as forest for s1
constitutive_nonlocal - take orientation gradients into account when calculating dislocation stress and dislocation fluxes - hard coded value for nu - changed kinetics (parameter G0 is currently defined as a parameter, needs to be read from material.config) - added some output statements constitutive: - some functions and subroutines needed additional input variables for passing to constitutive_nonlocal crystallite: - some functions and subroutines needed additional input variables for passing to constitutive - call microstructure with current temperature, Fp, Fe, not "sub0" values - show number of IPs, that are "onTrack" instead of those not "onTrack" - calculate Fe at beginning of substep, since we need it for state preguess
2009-10-07 21:01:52 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
forestProjectionEdge(s1,s2,i) &
= abs(math_mul3x3(lattice_sn(1:3,slipSystemLattice(s1,i),myStructure), &
lattice_st(1:3,slipSystemLattice(s2,i),myStructure))) ! forest projection of edge dislocations is the projection of (t = b x n) onto the slip normal of the respective slip plane
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
forestProjectionScrew(s1,s2,i) &
= abs(math_mul3x3(lattice_sn(1:3,slipSystemLattice(s1,i),myStructure), &
lattice_sd(1:3,slipSystemLattice(s2,i),myStructure))) ! forest projection of screw dislocations is the projection of b onto the slip normal of the respective splip plane
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
!*** calculation of interaction matrices
constitutive_nonlocal - take orientation gradients into account when calculating dislocation stress and dislocation fluxes - hard coded value for nu - changed kinetics (parameter G0 is currently defined as a parameter, needs to be read from material.config) - added some output statements constitutive: - some functions and subroutines needed additional input variables for passing to constitutive_nonlocal crystallite: - some functions and subroutines needed additional input variables for passing to constitutive - call microstructure with current temperature, Fp, Fe, not "sub0" values - show number of IPs, that are "onTrack" instead of those not "onTrack" - calculate Fe at beginning of substep, since we need it for state preguess
2009-10-07 21:01:52 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
interactionMatrixSlipSlip(s1,s2,i) &
= interactionSlipSlip(lattice_interactionSlipSlip(slipSystemLattice(s1,i), &
slipSystemLattice(s2,i), &
myStructure), i)
introduced annihilation of screw dipoles; so far no thermal activation, but immediate annihilation; each annihilated screw dipole leaves behind an edge jog of same length as the dipole height on the collinear system
2012-08-14 17:56:20 +05:30
!*** colinear slip system (only makes sense for fcc like it is defined here)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (lattice_interactionSlipSlip(slipSystemLattice(s1,i), &
slipSystemLattice(s2,i), &
introduced annihilation of screw dipoles; so far no thermal activation, but immediate annihilation; each annihilated screw dipole leaves behind an edge jog of same length as the dipole height on the collinear system
2012-08-14 17:56:20 +05:30
myStructure) == 3_pInt) then
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
colinearSystem(s1,i) = s2
introduced annihilation of screw dipoles; so far no thermal activation, but immediate annihilation; each annihilated screw dipole leaves behind an edge jog of same length as the dipole height on the collinear system
2012-08-14 17:56:20 +05:30
endif
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
enddo
* internal stress calculation now considers dead dislocations always at the interface, not at the center of the IP volume; used to merge them together with "normal" dislocations for stress calculation. * dislocation flux is blocked if we encounter a sign change in the resolved shear stress from the central ip to the neighbor * do not set density to zero if below certain threshold; this creates an artificial sink term
2011-08-02 16:47:45 +05:30
!*** rotation matrix from lattice configuration to slip system
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
lattice2slip(1:3,1:3,s1,i) &
= math_transpose33( reshape([ lattice_sd(1:3, slipSystemLattice(s1,i), myStructure), &
-lattice_st(1:3, slipSystemLattice(s1,i), myStructure), &
lattice_sn(1:3, slipSystemLattice(s1,i), myStructure)], [3,3]))
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
enddo
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
enddo
endsubroutine
!*********************************************************************
!* initial microstructural state (just the "basic" states) *
!*********************************************************************
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
subroutine constitutive_nonlocal_stateInit(state)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
use IO, only: IO_error
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
use lattice, only: lattice_maxNslipFamily
* now able to introduce some scatter in the initial single dislocation density (only for nonlocal model!) ; setting the parameter "rhoSglScatter" to a positive value generates a gauss distribution for the dislocation density with standard deviation equal to "rhoSglScatter" * dislocation stress calculation is only done for nonlocal constitution ("/nonlocal/" keyword is present in material.config)
2011-02-04 21:11:32 +05:30
use math, only: math_sampleGaussVar
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
use mesh, only: mesh_ipVolume, &
mesh_NcpElems, &
mesh_maxNips, &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
mesh_element, &
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
FE_Nips, &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
FE_geomtype
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
use material, only: material_phase, &
phase_plasticityInstance, &
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
phase_plasticity, &
homogenization_Ngrains
* now able to introduce some scatter in the initial single dislocation density (only for nonlocal model!) ; setting the parameter "rhoSglScatter" to a positive value generates a gauss distribution for the dislocation density with standard deviation equal to "rhoSglScatter" * dislocation stress calculation is only done for nonlocal constitution ("/nonlocal/" keyword is present in material.config)
2011-02-04 21:11:32 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
implicit none
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
!*** input/output variables
type(p_vec), dimension(1,mesh_maxNips,mesh_NcpElems), intent(inout) :: &
state ! microstructural state
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*** local variables
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
integer(pInt) el, &
ip, &
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
e, &
i, &
g, &
idx, &
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
ns, & ! short notation for total number of active slip systems
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
f, & ! index of lattice family
constitutive_nonlocal - take orientation gradients into account when calculating dislocation stress and dislocation fluxes - hard coded value for nu - changed kinetics (parameter G0 is currently defined as a parameter, needs to be read from material.config) - added some output statements constitutive: - some functions and subroutines needed additional input variables for passing to constitutive_nonlocal crystallite: - some functions and subroutines needed additional input variables for passing to constitutive - call microstructure with current temperature, Fp, Fe, not "sub0" values - show number of IPs, that are "onTrack" instead of those not "onTrack" - calculate Fe at beginning of substep, since we need it for state preguess
2009-10-07 21:01:52 +05:30
from, &
upto, &
* now able to introduce some scatter in the initial single dislocation density (only for nonlocal model!) ; setting the parameter "rhoSglScatter" to a positive value generates a gauss distribution for the dislocation density with standard deviation equal to "rhoSglScatter" * dislocation stress calculation is only done for nonlocal constitution ("/nonlocal/" keyword is present in material.config)
2011-02-04 21:11:32 +05:30
s, & ! index of slip system
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
t, &
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
j, &
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
myInstance, &
maxNinstance
* now able to introduce some scatter in the initial single dislocation density (only for nonlocal model!) ; setting the parameter "rhoSglScatter" to a positive value generates a gauss distribution for the dislocation density with standard deviation equal to "rhoSglScatter" * dislocation stress calculation is only done for nonlocal constitution ("/nonlocal/" keyword is present in material.config)
2011-02-04 21:11:32 +05:30
real(pReal), dimension(2) :: noise
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
real(pReal), dimension(4) :: rnd
real(pReal) meanDensity, &
totalVolume, &
random distribution of initial dislocation density now really independent of meshsize
2012-10-02 20:56:58 +05:30
densityBinning, &
random distribution of initial dislocation density now independent of mesh size
2012-10-02 19:05:34 +05:30
minimumIpVolume
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
maxNinstance = int(count(phase_plasticity == CONSTITUTIVE_NONLOCAL_LABEL),pInt)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
! ititalize all states to zero
do e = 1_pInt,mesh_NcpElems
do i = 1_pInt,FE_Nips(FE_geomtype(mesh_element(2,e)))
do g = 1_pInt,homogenization_Ngrains(mesh_element(3,e))
state(g,i,e)%p = 0.0_pReal
enddo
enddo
enddo
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
do myInstance = 1_pInt,maxNinstance
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
ns = totalNslip(myInstance)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
! randomly distribute dislocation segments on random slip system and of random type in the volume
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (rhoSglRandom(myInstance) > 0.0_pReal) then
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
! get the total volume of the instance
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
random distribution of initial dislocation density now independent of mesh size
2012-10-02 19:05:34 +05:30
minimumIpVolume = 1e99_pReal
total density used in backstress calculation now based on average of neighborhood fixed small bug in state_init: random distribution of density was probably not working correctly, as some variables were not properly initialized
2013-03-27 18:34:01 +05:30
totalVolume = 0.0_pReal
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
do e = 1_pInt,mesh_NcpElems
do i = 1_pInt,FE_Nips(FE_geomtype(mesh_element(2,e)))
if (CONSTITUTIVE_NONLOCAL_LABEL == phase_plasticity(material_phase(1,i,e)) &
.and. myInstance == phase_plasticityInstance(material_phase(1,i,e))) then
totalVolume = totalVolume + mesh_ipVolume(i,e)
minimumIpVolume = min(minimumIpVolume, mesh_ipVolume(i,e))
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
endif
enddo
enddo
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
densityBinning = rhoSglRandomBinning(myInstance) / minimumIpVolume ** (2.0_pReal / 3.0_pReal)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
! subsequently fill random ips with dislocation segments until we reach the desired overall density
meanDensity = 0.0_pReal
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
do while(meanDensity < rhoSglRandom(myInstance))
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
call random_number(rnd)
el = nint(rnd(1)*real(mesh_NcpElems,pReal)+0.5_pReal,pInt)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
ip = nint(rnd(2)*real(FE_Nips(FE_geomtype(mesh_element(2,el))),pReal)+0.5_pReal,pInt)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (CONSTITUTIVE_NONLOCAL_LABEL == phase_plasticity(material_phase(1,ip,el)) &
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
.and. myInstance == phase_plasticityInstance(material_phase(1,ip,el))) then
s = nint(rnd(3)*real(ns,pReal)+0.5_pReal,pInt)
t = nint(rnd(4)*4.0_pReal+0.5_pReal,pInt)
random distribution of initial dislocation density now really independent of meshsize
2012-10-02 20:56:58 +05:30
meanDensity = meanDensity + densityBinning * mesh_ipVolume(ip,el) / totalVolume
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
if (t==1_pInt) then
idx = iRhoEPU(s,myInstance)
elseif (t==2_pInt) then
idx = iRhoENU(s,myInstance)
elseif (t==3_pInt) then
idx = iRhoSPU(s,myInstance)
elseif (t==4_pInt) then
idx = iRhoSNU(s,myInstance)
else
call IO_error(-1,ext_msg='state init failed ('//CONSTITUTIVE_NONLOCAL_LABEL//')')
endif
state(1,ip,el)%p(idx) = state(1,ip,el)%p(idx) + densityBinning
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
endif
enddo
! homogeneous distribution of density with some noise
else
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
do e = 1_pInt,mesh_NcpElems
do i = 1_pInt,FE_Nips(FE_geomtype(mesh_element(2,e)))
if (CONSTITUTIVE_NONLOCAL_LABEL == phase_plasticity(material_phase(1,i,e)) &
.and. myInstance == phase_plasticityInstance(material_phase(1,i,e))) then
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
do f = 1_pInt,lattice_maxNslipFamily
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
from = 1_pInt + sum(Nslip(1:f-1_pInt,myInstance))
upto = sum(Nslip(1:f,myInstance))
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
do s = from,upto
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
do j = 1_pInt,2_pInt
noise(j) = math_sampleGaussVar(0.0_pReal, rhoSglScatter(myInstance))
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
enddo
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
state(1,i,e)%p(iRhoEPU(s,myInstance)) = rhoSglEdgePos0(f, myInstance) + noise(1)
state(1,i,e)%p(iRhoENU(s,myInstance)) = rhoSglEdgeNeg0(f, myInstance) + noise(1)
state(1,i,e)%p(iRhoSPU(s,myInstance)) = rhoSglScrewPos0(f, myInstance) + noise(2)
state(1,i,e)%p(iRhoSNU(s,myInstance)) = rhoSglScrewNeg0(f, myInstance) + noise(2)
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
enddo
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
state(1,i,e)%p(iRhoED(from:upto,myInstance)) = rhoDipEdge0(f, myInstance)
state(1,i,e)%p(iRhoSD(from:upto,myInstance)) = rhoDipScrew0(f, myInstance)
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
enddo
endif
enddo
enddo
endif
enddo
endsubroutine
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
convergence of state in crystallite is now tested as follows: (state < relevant state) or (residuum < relative tolerance * state) since the relevant value for the state variables depend on their nature and can vary by large scales (e.g. volume fraction: 1e-10, dislocation density: 1e5) it is not possible to set a unique value. instead the constitutive law has to decide what is relevant. therefore, all constitutive laws now read in parameters from the material.config that determine the values for relevantState [@luc: in dislobased law relevant State is for the moment generally set to 1e-200, so no additional parameters necessary in material.config. if you also want this feature, we can still implement it, no big deal] - added sanity checks in constitutive_nonlocal.f90 - corrected coordinate transformation for backstress calculation in constitutive_nonlocal.f90 - corrected equations for evolution of dipole dislocation densities (athermal annihilation and formation by glide)
2009-09-18 21:07:14 +05:30
!*********************************************************************
* former "relevantRho" and "relevantResistance" is renamed to "atol_rho" and "atol_resistance" and now used as an absolute tolerance for the state residuum. before it was used rather as significant state, so whenever the state dropped below that value it was considered converged. (In dislotwin and titanmod constitutive law there is only one value atol_rho which is used for all the states, though the state array consists not only of densities. We need further parameters here!) * somewhat simplified convergence check for adaptive euler and runge-kutta
2010-10-26 18:46:37 +05:30
!* absolute state tolerance *
convergence of state in crystallite is now tested as follows: (state < relevant state) or (residuum < relative tolerance * state) since the relevant value for the state variables depend on their nature and can vary by large scales (e.g. volume fraction: 1e-10, dislocation density: 1e5) it is not possible to set a unique value. instead the constitutive law has to decide what is relevant. therefore, all constitutive laws now read in parameters from the material.config that determine the values for relevantState [@luc: in dislobased law relevant State is for the moment generally set to 1e-200, so no additional parameters necessary in material.config. if you also want this feature, we can still implement it, no big deal] - added sanity checks in constitutive_nonlocal.f90 - corrected coordinate transformation for backstress calculation in constitutive_nonlocal.f90 - corrected equations for evolution of dipole dislocation densities (athermal annihilation and formation by glide)
2009-09-18 21:07:14 +05:30
!*********************************************************************
* former "relevantRho" and "relevantResistance" is renamed to "atol_rho" and "atol_resistance" and now used as an absolute tolerance for the state residuum. before it was used rather as significant state, so whenever the state dropped below that value it was considered converged. (In dislotwin and titanmod constitutive law there is only one value atol_rho which is used for all the states, though the state array consists not only of densities. We need further parameters here!) * somewhat simplified convergence check for adaptive euler and runge-kutta
2010-10-26 18:46:37 +05:30
pure function constitutive_nonlocal_aTolState(myInstance)
convergence of state in crystallite is now tested as follows: (state < relevant state) or (residuum < relative tolerance * state) since the relevant value for the state variables depend on their nature and can vary by large scales (e.g. volume fraction: 1e-10, dislocation density: 1e5) it is not possible to set a unique value. instead the constitutive law has to decide what is relevant. therefore, all constitutive laws now read in parameters from the material.config that determine the values for relevantState [@luc: in dislobased law relevant State is for the moment generally set to 1e-200, so no additional parameters necessary in material.config. if you also want this feature, we can still implement it, no big deal] - added sanity checks in constitutive_nonlocal.f90 - corrected coordinate transformation for backstress calculation in constitutive_nonlocal.f90 - corrected equations for evolution of dipole dislocation densities (athermal annihilation and formation by glide)
2009-09-18 21:07:14 +05:30
implicit none
!*** input variables
improved accuracy of "accumulated shear": now added as a state variable, which facilitates integration (dotstate=shearrate); needs absolute tolerance value for state integration in material.config
2013-04-04 19:07:14 +05:30
integer(pInt), intent(in) :: myInstance ! number specifying the current instance of the plasticity
convergence of state in crystallite is now tested as follows: (state < relevant state) or (residuum < relative tolerance * state) since the relevant value for the state variables depend on their nature and can vary by large scales (e.g. volume fraction: 1e-10, dislocation density: 1e5) it is not possible to set a unique value. instead the constitutive law has to decide what is relevant. therefore, all constitutive laws now read in parameters from the material.config that determine the values for relevantState [@luc: in dislobased law relevant State is for the moment generally set to 1e-200, so no additional parameters necessary in material.config. if you also want this feature, we can still implement it, no big deal] - added sanity checks in constitutive_nonlocal.f90 - corrected coordinate transformation for backstress calculation in constitutive_nonlocal.f90 - corrected equations for evolution of dipole dislocation densities (athermal annihilation and formation by glide)
2009-09-18 21:07:14 +05:30
!*** output variables
real(pReal), dimension(constitutive_nonlocal_sizeState(myInstance)) :: &
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
constitutive_nonlocal_aTolState ! absolute state tolerance for the current instance of this plasticity
convergence of state in crystallite is now tested as follows: (state < relevant state) or (residuum < relative tolerance * state) since the relevant value for the state variables depend on their nature and can vary by large scales (e.g. volume fraction: 1e-10, dislocation density: 1e5) it is not possible to set a unique value. instead the constitutive law has to decide what is relevant. therefore, all constitutive laws now read in parameters from the material.config that determine the values for relevantState [@luc: in dislobased law relevant State is for the moment generally set to 1e-200, so no additional parameters necessary in material.config. if you also want this feature, we can still implement it, no big deal] - added sanity checks in constitutive_nonlocal.f90 - corrected coordinate transformation for backstress calculation in constitutive_nonlocal.f90 - corrected equations for evolution of dipole dislocation densities (athermal annihilation and formation by glide)
2009-09-18 21:07:14 +05:30
!*** local variables
improved accuracy of "accumulated shear": now added as a state variable, which facilitates integration (dotstate=shearrate); needs absolute tolerance value for state integration in material.config
2013-04-04 19:07:14 +05:30
integer(pInt) :: ns
convergence of state in crystallite is now tested as follows: (state < relevant state) or (residuum < relative tolerance * state) since the relevant value for the state variables depend on their nature and can vary by large scales (e.g. volume fraction: 1e-10, dislocation density: 1e5) it is not possible to set a unique value. instead the constitutive law has to decide what is relevant. therefore, all constitutive laws now read in parameters from the material.config that determine the values for relevantState [@luc: in dislobased law relevant State is for the moment generally set to 1e-200, so no additional parameters necessary in material.config. if you also want this feature, we can still implement it, no big deal] - added sanity checks in constitutive_nonlocal.f90 - corrected coordinate transformation for backstress calculation in constitutive_nonlocal.f90 - corrected equations for evolution of dipole dislocation densities (athermal annihilation and formation by glide)
2009-09-18 21:07:14 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
ns = totalNslip(myInstance)
improved accuracy of "accumulated shear": now added as a state variable, which facilitates integration (dotstate=shearrate); needs absolute tolerance value for state integration in material.config
2013-04-04 19:07:14 +05:30
constitutive_nonlocal_aTolState = 0.0_pReal
use named state variable indices in "aTolState"
2013-05-24 02:43:56 +05:30
constitutive_nonlocal_aTolState(iRhoEPU(1:ns,myInstance)) = aTolRho(myInstance)
constitutive_nonlocal_aTolState(iRhoENU(1:ns,myInstance)) = aTolRho(myInstance)
constitutive_nonlocal_aTolState(iRhoSPU(1:ns,myInstance)) = aTolRho(myInstance)
constitutive_nonlocal_aTolState(iRhoSNU(1:ns,myInstance)) = aTolRho(myInstance)
constitutive_nonlocal_aTolState(iRhoEPB(1:ns,myInstance)) = aTolRho(myInstance)
constitutive_nonlocal_aTolState(iRhoENB(1:ns,myInstance)) = aTolRho(myInstance)
constitutive_nonlocal_aTolState(iRhoSPB(1:ns,myInstance)) = aTolRho(myInstance)
constitutive_nonlocal_aTolState(iRhoSNB(1:ns,myInstance)) = aTolRho(myInstance)
constitutive_nonlocal_aTolState(iRhoED(1:ns,myInstance)) = aTolRho(myInstance)
constitutive_nonlocal_aTolState(iRhoSD(1:ns,myInstance)) = aTolRho(myInstance)
constitutive_nonlocal_aTolState(iGamma(1:ns,myInstance)) = aTolShear(myInstance)
convergence of state in crystallite is now tested as follows: (state < relevant state) or (residuum < relative tolerance * state) since the relevant value for the state variables depend on their nature and can vary by large scales (e.g. volume fraction: 1e-10, dislocation density: 1e5) it is not possible to set a unique value. instead the constitutive law has to decide what is relevant. therefore, all constitutive laws now read in parameters from the material.config that determine the values for relevantState [@luc: in dislobased law relevant State is for the moment generally set to 1e-200, so no additional parameters necessary in material.config. if you also want this feature, we can still implement it, no big deal] - added sanity checks in constitutive_nonlocal.f90 - corrected coordinate transformation for backstress calculation in constitutive_nonlocal.f90 - corrected equations for evolution of dipole dislocation densities (athermal annihilation and formation by glide)
2009-09-18 21:07:14 +05:30
endfunction
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*********************************************************************
!* calculates homogenized elacticity matrix *
!*********************************************************************
pure function constitutive_nonlocal_homogenizedC(state,g,ip,el)
use mesh, only: mesh_NcpElems, &
mesh_maxNips
use material, only: homogenization_maxNgrains, &
material_phase, &
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
phase_plasticityInstance
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
implicit none
!*** input variables
integer(pInt), intent(in) :: g, & ! current grain ID
ip, & ! current integration point
el ! current element
type(p_vec), dimension(homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), intent(in) :: state ! microstructural state
!*** output variables
real(pReal), dimension(6,6) :: constitutive_nonlocal_homogenizedC ! homogenized elasticity matrix
!*** local variables
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
integer(pInt) myInstance ! current instance of this plasticity
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
myInstance = phase_plasticityInstance(material_phase(g,ip,el))
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_homogenizedC = Cslip66(1:6,1:6,myInstance)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
endfunction
!*********************************************************************
!* calculates quantities characterizing the microstructure *
!*********************************************************************
polishing
2013-05-23 17:55:56 +05:30
subroutine constitutive_nonlocal_microstructure(state, Temperature, Fe, Fp, gr, ip, el)
use IO, only: &
IO_error
use math, only: &
pi, &
math_mul33x3, &
math_mul3x3, &
math_norm3, &
math_invert33, &
math_transpose33
use debug, only: &
debug_level, &
debug_constitutive, &
debug_levelBasic, &
debug_levelExtensive, &
debug_levelSelective, &
debug_g, &
debug_i, &
debug_e
use mesh, only: &
mesh_NcpElems, &
mesh_maxNips, &
mesh_element, &
mesh_ipNeighborhood, &
mesh_ipCoordinates, &
mesh_ipVolume, &
mesh_ipAreaNormal, &
mesh_ipArea, &
FE_NipNeighbors, &
replaced FE_maxNipNeighbors with mesh_maxNipNeighbors, which is sufficient for memory allocation, since this gives the maximum of neighbors in all cp elements
2013-05-23 23:16:21 +05:30
mesh_maxNipNeighbors, &
polishing
2013-05-23 17:55:56 +05:30
FE_geomtype, &
FE_celltype
use material, only: &
homogenization_maxNgrains, &
material_phase, &
phase_localPlasticity, &
phase_plasticityInstance
use lattice, only: &
lattice_sd, &
lattice_st, &
lattice_interactionSlipSlip
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
implicit none
!*** input variables
polishing
2013-05-23 17:55:56 +05:30
integer(pInt), intent(in) :: gr, & ! current grain ID
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
ip, & ! current integration point
el ! current element
real(pReal), intent(in) :: Temperature ! temperature
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
real(pReal), dimension(3,3), intent(in) :: &
Fe, & ! elastic deformation gradient
Fp ! elastic deformation gradient
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
constitutive_nonlocal: dislocation velocities are stored in the state, so we actually now have three "parts" of the state, the basic states that are updated by "constitutive_dotState" come first, then the dependent states that are calculated by "constitutive_microstructure" follow, and finally we have a last part reserved for other variables that just use the memory reserved by the state array and are updated somewhere else. constitutive: LpAndItsTangent does not need the full state, but only the local state, so changed that at least for the nonlocal constitutive law
2011-11-04 18:42:17 +05:30
!*** input/output variables
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
type(p_vec), dimension(homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), intent(inout) :: &
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
state ! microstructural state
constitutive_nonlocal: dislocation velocities are stored in the state, so we actually now have three "parts" of the state, the basic states that are updated by "constitutive_dotState" come first, then the dependent states that are calculated by "constitutive_microstructure" follow, and finally we have a last part reserved for other variables that just use the memory reserved by the state array and are updated somewhere else. constitutive: LpAndItsTangent does not need the full state, but only the local state, so changed that at least for the nonlocal constitutive law
2011-11-04 18:42:17 +05:30
!*** output variables
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
constitutive_nonlocal: dislocation velocities are stored in the state, so we actually now have three "parts" of the state, the basic states that are updated by "constitutive_dotState" come first, then the dependent states that are calculated by "constitutive_microstructure" follow, and finally we have a last part reserved for other variables that just use the memory reserved by the state array and are updated somewhere else. constitutive: LpAndItsTangent does not need the full state, but only the local state, so changed that at least for the nonlocal constitutive law
2011-11-04 18:42:17 +05:30
!*** local variables
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
integer(pInt) neighboring_el, & ! element number of neighboring material point
neighboring_ip, & ! integration point of neighboring material point
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
instance, & ! my instance of this plasticity
neighboring_instance, & ! instance of this plasticity of neighboring material point
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
latticeStruct, & ! my lattice structure
neighboring_latticeStruct, & ! lattice structure of neighboring material point
phase, &
neighboring_phase, &
ns, & ! total number of active slip systems at my material point
neighboring_ns, & ! total number of active slip systems at neighboring material point
c, & ! index of dilsocation character (edge, screw)
s, & ! slip system index
hardening coefficients for fcc are density dependent (line tension effect, Devincre, Kubin, Hoc; 2006; Physical analyses of crystal plasticity by DD simulations) fixed bug with thermal annihilation producing negative density in case of very small dipole density and high climb velocity
2012-10-04 23:38:40 +05:30
s2, & ! slip system index
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
t, & ! index of dilsocation type (e+, e-, s+, s-, used e+, used e-, used s+, used s-)
dir, &
hardening coefficients for fcc are density dependent (line tension effect, Devincre, Kubin, Hoc; 2006; Physical analyses of crystal plasticity by DD simulations) fixed bug with thermal annihilation producing negative density in case of very small dipole density and high climb velocity
2012-10-04 23:38:40 +05:30
n, &
total density used in backstress calculation now based on average of neighborhood fixed small bug in state_init: random distribution of density was probably not working correctly, as some variables were not properly initialized
2013-03-27 18:34:01 +05:30
nRealNeighbors, & ! number of really existing neighbors
hardening coefficients for fcc are density dependent (line tension effect, Devincre, Kubin, Hoc; 2006; Physical analyses of crystal plasticity by DD simulations) fixed bug with thermal annihilation producing negative density in case of very small dipole density and high climb velocity
2012-10-04 23:38:40 +05:30
interactionCoefficient
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
integer(pInt), dimension(2) :: neighbor
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal) detFe, &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
detFp, &
adding switch in material.config for short range stress correction
2012-03-15 20:28:12 +05:30
FVsize, &
hardening coefficients for fcc are density dependent (line tension effect, Devincre, Kubin, Hoc; 2006; Physical analyses of crystal plasticity by DD simulations) fixed bug with thermal annihilation producing negative density in case of very small dipole density and high climb velocity
2012-10-04 23:38:40 +05:30
temp, &
correction, &
myRhoForest
local stress correction of gradient in dead dislocations and neighboring excess density is additive
2012-03-14 20:54:19 +05:30
real(pReal), dimension(2) :: rhoExcessGradient, &
rhoExcessGradient_over_rho, &
rhoTotal
fixed bug in backstress calculation for elements with periodic neighbors
2013-05-23 18:06:48 +05:30
real(pReal), dimension(3) :: rhoExcessDifferences, &
normal_latticeConf
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(gr,ip,el)))) :: &
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
rhoForest, & ! forest dislocation density
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
tauBack, & ! back stress from pileup on same slip system
removed unused variables
2011-04-13 19:46:22 +05:30
tauThreshold ! threshold shear stress
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
real(pReal), dimension(3,3) :: invFe, & ! inverse of elastic deformation gradient
local stress correction of gradient in dead dislocations and neighboring excess density is additive
2012-03-14 20:54:19 +05:30
invFp, & ! inverse of plastic deformation gradient
adding switch in material.config for short range stress correction
2012-03-15 20:28:12 +05:30
connections, &
invConnections
replaced FE_maxNipNeighbors with mesh_maxNipNeighbors, which is sufficient for memory allocation, since this gives the maximum of neighbors in all cp elements
2013-05-23 23:16:21 +05:30
real(pReal), dimension(3,mesh_maxNipNeighbors) :: &
corrected special treatment of periodic images in calculation of stress correction
2012-03-15 15:38:08 +05:30
connection_latticeConf
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(2,totalNslip(phase_plasticityInstance(material_phase(gr,ip,el)))) :: &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
rhoExcess
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(gr,ip,el))),2) :: &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
rhoDip ! dipole dislocation density (edge, screw)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(gr,ip,el))),8) :: &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
rhoSgl ! single dislocation density (edge+, edge-, screw+, screw-, used edge+, used edge-, used screw+, used screw-)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(gr,ip,el))), &
totalNslip(phase_plasticityInstance(material_phase(gr,ip,el)))) :: &
make correction of hardening coefficients thread safe
2012-10-05 21:35:51 +05:30
myInteractionMatrix ! corrected slip interaction matrix
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(2,maxval(totalNslip),mesh_maxNipNeighbors) :: &
total density used in backstress calculation now based on average of neighborhood fixed small bug in state_init: random distribution of density was probably not working correctly, as some variables were not properly initialized
2013-03-27 18:34:01 +05:30
neighboring_rhoExcess, & ! excess density at neighboring material point
neighboring_rhoTotal ! total density at neighboring material point
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(3,totalNslip(phase_plasticityInstance(material_phase(gr,ip,el))),2) :: &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
m ! direction of dislocation motion
* new output variable "internalstress", which gives the internal stress tensor * use "math_invert3x3" instead of "math_inv3x3" for inversion of Fe * for dislocation stress calculation: first regular case, then special case of dead dislocations in central ip * "dv_dtau" now given for each dislocation type, so is a (ns,4) array * deleted unused variables in "_LpAndItsTangent" * corrected contribution of deads in "_LpAndItsTangent" * the NaN variables defined in math did not give a proper NaN value, so use 0.0/0.0 again * neighbors with nonlocal constitution but local properties (i.e. /nonlocal/ flag not set) are also considered for incoming fluxes
2011-09-07 17:00:28 +05:30
logical inversionError
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
polishing
2013-05-23 17:55:56 +05:30
phase = material_phase(gr,ip,el)
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
instance = phase_plasticityInstance(phase)
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
latticeStruct = constitutive_nonlocal_structure(instance)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
ns = totalNslip(instance)
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*** get basic states
use named state variable indices in "microstructure"
2013-05-24 02:58:45 +05:30
forall (s = 1_pInt:ns)
rhoSgl(s,1) = max(state(gr,ip,el)%p(iRhoEPU(s,instance)), 0.0_pReal) ! ensure positive single mobile densities
rhoSgl(s,2) = max(state(gr,ip,el)%p(iRhoENU(s,instance)), 0.0_pReal) ! ensure positive single mobile densities
rhoSgl(s,3) = max(state(gr,ip,el)%p(iRhoSPU(s,instance)), 0.0_pReal) ! ensure positive single mobile densities
rhoSgl(s,4) = max(state(gr,ip,el)%p(iRhoSNU(s,instance)), 0.0_pReal) ! ensure positive single mobile densities
endforall
rhoSgl(1:ns,5) = state(gr,ip,el)%p(iRhoEPB(1:ns,instance))
rhoSgl(1:ns,6) = state(gr,ip,el)%p(iRhoENB(1:ns,instance))
rhoSgl(1:ns,7) = state(gr,ip,el)%p(iRhoSPB(1:ns,instance))
rhoSgl(1:ns,8) = state(gr,ip,el)%p(iRhoSNB(1:ns,instance))
forall (s = 1_pInt:ns)
rhoDip(s,1) = max(state(gr,ip,el)%p(iRhoED(s,instance)), 0.0_pReal) ! ensure positive dipole densities
rhoDip(s,2) = max(state(gr,ip,el)%p(iRhoSD(s,instance)), 0.0_pReal) ! ensure positive dipole densities
endforall
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
where (abs(rhoSgl) * mesh_ipVolume(ip,el) ** 0.667_pReal < significantN(instance) &
.or. abs(rhoSgl) < significantRho(instance)) &
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
rhoSgl = 0.0_pReal
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
where (abs(rhoDip) * mesh_ipVolume(ip,el) ** 0.667_pReal < significantN(instance) &
.or. abs(rhoDip) < significantRho(instance)) &
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
rhoDip = 0.0_pReal
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*** calculate the forest dislocation density
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
!*** (= projection of screw and edge dislocations)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
viscosity term in kinetics law proportional to resolvedstress minus forest term
2012-03-14 20:48:36 +05:30
forall (s = 1_pInt:ns) &
corrected special treatment of periodic images in calculation of stress correction
2012-03-15 15:38:08 +05:30
rhoForest(s) = dot_product((sum(abs(rhoSgl(1:ns,[1,2,5,6])),2) + rhoDip(1:ns,1)), &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
forestProjectionEdge(s,1:ns,instance)) &
corrected special treatment of periodic images in calculation of stress correction
2012-03-15 15:38:08 +05:30
+ dot_product((sum(abs(rhoSgl(1:ns,[3,4,7,8])),2) + rhoDip(1:ns,2)), &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
forestProjectionScrew(s,1:ns,instance))
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
constitutive_nonlocal - take orientation gradients into account when calculating dislocation stress and dislocation fluxes - hard coded value for nu - changed kinetics (parameter G0 is currently defined as a parameter, needs to be read from material.config) - added some output statements constitutive: - some functions and subroutines needed additional input variables for passing to constitutive_nonlocal crystallite: - some functions and subroutines needed additional input variables for passing to constitutive - call microstructure with current temperature, Fp, Fe, not "sub0" values - show number of IPs, that are "onTrack" instead of those not "onTrack" - calculate Fe at beginning of substep, since we need it for state preguess
2009-10-07 21:01:52 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*** calculate the threshold shear stress for dislocation slip
make correction of hardening coefficients thread safe
2012-10-05 21:35:51 +05:30
myInteractionMatrix = 0.0_pReal
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
myInteractionMatrix(1:ns,1:ns) = interactionMatrixSlipSlip(1:ns,1:ns,instance)
polishing
2013-05-23 17:55:56 +05:30
if (latticeStruct == 1_pInt) then ! in case of fcc: coefficients are corrected for the line tension effect (see Kubin,Devincre,Hoc; 2008; Modeling dislocation storage rates and mean free paths in face-centered cubic crystals)
hardening coefficients for fcc are density dependent (line tension effect, Devincre, Kubin, Hoc; 2006; Physical analyses of crystal plasticity by DD simulations) fixed bug with thermal annihilation producing negative density in case of very small dipole density and high climb velocity
2012-10-04 23:38:40 +05:30
do s = 1_pInt,ns
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
myRhoForest = max(rhoForest(s),significantRho(instance))
correction = ( 1.0_pReal - linetensionEffect(instance) &
+ linetensionEffect(instance) &
* log(0.35_pReal * burgers(s,instance) * sqrt(myRhoForest)) &
/ log(0.35_pReal * burgers(s,instance) * 1e6_pReal)) ** 2.0_pReal
hardening coefficients for fcc are density dependent (line tension effect, Devincre, Kubin, Hoc; 2006; Physical analyses of crystal plasticity by DD simulations) fixed bug with thermal annihilation producing negative density in case of very small dipole density and high climb velocity
2012-10-04 23:38:40 +05:30
do s2 = 1_pInt,ns
polishing
2013-05-23 17:55:56 +05:30
interactionCoefficient = &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
lattice_interactionSlipSlip(slipSystemLattice(s,instance), &
slipSystemLattice(s2,instance), &
polishing
2013-05-23 17:55:56 +05:30
latticeStruct)
hardening coefficients for fcc are density dependent (line tension effect, Devincre, Kubin, Hoc; 2006; Physical analyses of crystal plasticity by DD simulations) fixed bug with thermal annihilation producing negative density in case of very small dipole density and high climb velocity
2012-10-04 23:38:40 +05:30
select case(interactionCoefficient)
polishing
2013-05-23 17:55:56 +05:30
case(4_pInt,5_pInt,6_pInt) ! only correct junction forming interactions (4,5,6)
make correction of hardening coefficients thread safe
2012-10-05 21:35:51 +05:30
myInteractionMatrix(s,s2) = correction * myInteractionMatrix(s,s2)
hardening coefficients for fcc are density dependent (line tension effect, Devincre, Kubin, Hoc; 2006; Physical analyses of crystal plasticity by DD simulations) fixed bug with thermal annihilation producing negative density in case of very small dipole density and high climb velocity
2012-10-04 23:38:40 +05:30
endselect
enddo
enddo
endif
viscosity term in kinetics law proportional to resolvedstress minus forest term
2012-03-14 20:48:36 +05:30
forall (s = 1_pInt:ns) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
tauThreshold(s) = mu(instance) * burgers(s,instance) &
make correction of hardening coefficients thread safe
2012-10-05 21:35:51 +05:30
* sqrt(dot_product((sum(abs(rhoSgl),2) + sum(abs(rhoDip),2)), myInteractionMatrix(s,1:ns)))
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
!*** calculate the dislocation stress of the neighboring excess dislocation densities
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
!*** zero for material points of local plasticity
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
tauBack = 0.0_pReal
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (.not. phase_localPlasticity(phase) .and. shortRangeStressCorrection(instance)) then
added new math function "math_deviatoric33" which returns the deviatoric part of a 3x3 tensor renamed some math functions, so that we have a universal naming scheme: for matrix multiplications use an "x" (e.g. math_mul33x3); don't use the "x" to describe the shape of the tensor that the function is applied to (e.g. math_invert33 instead of math_invert3x3)
2012-01-26 19:20:00 +05:30
call math_invert33(Fe, invFe, detFe, inversionError)
call math_invert33(Fp, invFp, detFp, inversionError)
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
rhoExcess(1,1:ns) = rhoSgl(1:ns,1) - rhoSgl(1:ns,2)
rhoExcess(2,1:ns) = rhoSgl(1:ns,3) - rhoSgl(1:ns,4)
FVsize = mesh_ipVolume(ip,el) ** (1.0_pReal/3.0_pReal)
* now able to introduce some scatter in the initial single dislocation density (only for nonlocal model!) ; setting the parameter "rhoSglScatter" to a positive value generates a gauss distribution for the dislocation density with standard deviation equal to "rhoSglScatter" * dislocation stress calculation is only done for nonlocal constitution ("/nonlocal/" keyword is present in material.config)
2011-02-04 21:11:32 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
!* loop through my neighborhood and get the connection vectors (in lattice frame) and the excess densities
total density used in backstress calculation now based on average of neighborhood fixed small bug in state_init: random distribution of density was probably not working correctly, as some variables were not properly initialized
2013-03-27 18:34:01 +05:30
nRealNeighbors = 0_pInt
neighboring_rhoTotal = 0.0_pReal
neighborhood relations were screwed up for some elementtypes as of version 2318, might have caused segmentation faults with nonlocal model, now fixed
2013-04-22 19:05:35 +05:30
do n = 1_pInt,FE_NipNeighbors(FE_celltype(FE_geomtype(mesh_element(2,el))))
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
neighboring_el = mesh_ipNeighborhood(1,n,ip,el)
neighboring_ip = mesh_ipNeighborhood(2,n,ip,el)
if (neighboring_el > 0 .and. neighboring_ip > 0) then
polishing
2013-05-23 17:55:56 +05:30
neighboring_phase = material_phase(gr,neighboring_ip,neighboring_el)
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
neighboring_instance = phase_plasticityInstance(neighboring_phase)
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
neighboring_latticeStruct = constitutive_nonlocal_structure(neighboring_instance)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
neighboring_ns = totalNslip(neighboring_instance)
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
if (.not. phase_localPlasticity(neighboring_phase) &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
.and. neighboring_latticeStruct == latticeStruct &
.and. neighboring_instance == instance) then
if (neighboring_ns == ns) then
fixed bug in backstress calculation for elements with periodic neighbors
2013-05-23 18:06:48 +05:30
nRealNeighbors = nRealNeighbors + 1_pInt
use named state variable indices in "microstructure"
2013-05-24 02:58:45 +05:30
forall (s = 1_pInt:ns)
neighboring_rhoExcess(1,s,n) = &
max(state(gr,neighboring_ip,neighboring_el)%p(iRhoEPU(s,neighboring_instance)), 0.0_pReal) &! positive mobiles
- max(state(gr,neighboring_ip,neighboring_el)%p(iRhoENU(s,neighboring_instance)), 0.0_pReal) ! negative mobiles
neighboring_rhoExcess(2,s,n) = &
max(state(gr,neighboring_ip,neighboring_el)%p(iRhoSPU(s,neighboring_instance)), 0.0_pReal) &! positive mobiles
- max(state(gr,neighboring_ip,neighboring_el)%p(iRhoSNU(s,neighboring_instance)), 0.0_pReal) ! negative mobiles
neighboring_rhoTotal(1,s,n) = &
max(state(gr,neighboring_ip,neighboring_el)%p(iRhoEPU(s,neighboring_instance)), 0.0_pReal) &! positive mobiles
+ max(state(gr,neighboring_ip,neighboring_el)%p(iRhoENU(s,neighboring_instance)), 0.0_pReal) &! negative mobiles
+ abs(state(gr,neighboring_ip,neighboring_el)%p(iRhoEPB(s,neighboring_instance))) & ! positive deads
+ abs(state(gr,neighboring_ip,neighboring_el)%p(iRhoENB(s,neighboring_instance))) & ! negative deads
+ max(state(gr,neighboring_ip,neighboring_el)%p(iRhoED(s,neighboring_instance)), 0.0_pReal) ! dipoles
neighboring_rhoTotal(2,s,n) = &
max(state(gr,neighboring_ip,neighboring_el)%p(iRhoSPU(s,neighboring_instance)), 0.0_pReal) &! positive mobiles
+ max(state(gr,neighboring_ip,neighboring_el)%p(iRhoSNU(s,neighboring_instance)), 0.0_pReal) &! negative mobiles
+ abs(state(gr,neighboring_ip,neighboring_el)%p(iRhoSPB(s,neighboring_instance))) & ! positive deads
+ abs(state(gr,neighboring_ip,neighboring_el)%p(iRhoSNB(s,neighboring_instance))) & ! negative deads
+ max(state(gr,neighboring_ip,neighboring_el)%p(iRhoSD(s,neighboring_instance)), 0.0_pReal) ! dipoles
fixed bug in backstress calculation for elements with periodic neighbors
2013-05-23 18:06:48 +05:30
endforall
connection_latticeConf(1:3,n) = &
math_mul33x3(invFe, mesh_ipCoordinates(1:3,neighboring_ip,neighboring_el) &
- mesh_ipCoordinates(1:3,ip,el))
normal_latticeConf = math_mul33x3(math_transpose33(invFp), mesh_ipAreaNormal(1:3,n,ip,el))
if (math_mul3x3(normal_latticeConf,connection_latticeConf(1:3,n)) < 0.0_pReal) then ! neighbor connection points in opposite direction to face normal: must be periodic image
connection_latticeConf(1:3,n) = normal_latticeConf * mesh_ipVolume(ip,el) &
/ mesh_ipArea(n,ip,el) ! instead take the surface normal scaled with the diameter of the cell
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
endif
else
! different number of active slip systems
polishing, adding _pInt etc. where applicable post_results now handels zero increment different (like FEM, it is always there even if it is not counted)
2012-02-13 19:48:07 +05:30
call IO_error(-1_pInt,ext_msg='different number of active slip systems in neighboring IPs of same crystal structure')
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
endif
else
corrected special treatment of periodic images in calculation of stress correction
2012-03-15 15:38:08 +05:30
! local neighbor or different lattice structure or different constitution instance -> use central values instead
connection_latticeConf(1:3,n) = 0.0_pReal
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
neighboring_rhoExcess(1:2,1:ns,n) = rhoExcess
endif
else
corrected special treatment of periodic images in calculation of stress correction
2012-03-15 15:38:08 +05:30
! free surface -> use central values instead
connection_latticeConf(1:3,n) = 0.0_pReal
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
neighboring_rhoExcess(1:2,1:ns,n) = rhoExcess
endif
enddo
local stress correction of gradient in dead dislocations and neighboring excess density is additive
2012-03-14 20:54:19 +05:30
!* loop through the slip systems and calculate the dislocation gradient by
!* 1. interpolation of the excess density in the neighorhood
!* 2. interpolation of the dead dislocation density in the central volume
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
m(1:3,1:ns,1) = lattice_sd(1:3,slipSystemLattice(1:ns,instance),latticeStruct)
m(1:3,1:ns,2) = -lattice_st(1:3,slipSystemLattice(1:ns,instance),latticeStruct)
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do s = 1_pInt,ns
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
local stress correction of gradient in dead dislocations and neighboring excess density is additive
2012-03-14 20:54:19 +05:30
!* gradient from interpolation of neighboring excess density
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
local stress correction of gradient in dead dislocations and neighboring excess density is additive
2012-03-14 20:54:19 +05:30
do c = 1_pInt,2_pInt
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do dir = 1_pInt,3_pInt
local stress correction of gradient in dead dislocations and neighboring excess density is additive
2012-03-14 20:54:19 +05:30
neighbor(1) = 2_pInt * dir - 1_pInt
neighbor(2) = 2_pInt * dir
fixed bug in backstress calculation for elements with periodic neighbors
2013-05-23 18:06:48 +05:30
connections(dir,1:3) = connection_latticeConf(1:3,neighbor(1)) &
- connection_latticeConf(1:3,neighbor(2))
rhoExcessDifferences(dir) = neighboring_rhoExcess(c,s,neighbor(1)) &
- neighboring_rhoExcess(c,s,neighbor(2))
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
enddo
adding switch in material.config for short range stress correction
2012-03-15 20:28:12 +05:30
call math_invert33(connections,invConnections,temp,inversionError)
if (inversionError) then
call IO_error(-1_pInt,ext_msg='back stress calculation: inversion error')
endif
polishing
2013-05-23 17:55:56 +05:30
rhoExcessGradient(c) = math_mul3x3(m(1:3,s,c), &
math_mul33x3(invConnections,rhoExcessDifferences))
local stress correction of gradient in dead dislocations and neighboring excess density is additive
2012-03-14 20:54:19 +05:30
enddo
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
local stress correction of gradient in dead dislocations and neighboring excess density is additive
2012-03-14 20:54:19 +05:30
!* plus gradient from deads
do t = 1_pInt,4_pInt
c = (t - 1_pInt) / 2_pInt + 1_pInt
rhoExcessGradient(c) = rhoExcessGradient(c) + rhoSgl(s,t+4_pInt) / FVsize
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
enddo
local stress correction of gradient in dead dislocations and neighboring excess density is additive
2012-03-14 20:54:19 +05:30
!* normalized with the total density
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
local stress correction of gradient in dead dislocations and neighboring excess density is additive
2012-03-14 20:54:19 +05:30
rhoExcessGradient_over_rho = 0.0_pReal
total density used in backstress calculation now based on average of neighborhood fixed small bug in state_init: random distribution of density was probably not working correctly, as some variables were not properly initialized
2013-03-27 18:34:01 +05:30
forall (c = 1_pInt:2_pInt) &
rhoTotal(c) = (sum(abs(rhoSgl(s,[2*c-1,2*c,2*c+3,2*c+4]))) + rhoDip(s,c) + sum(neighboring_rhoTotal(c,s,:))) &
fixed some warnings issued by gfortran (type conversion, not needed use statements)
2013-05-08 14:53:47 +05:30
/ real(1_pInt + nRealNeighbors,pReal)
local stress correction of gradient in dead dislocations and neighboring excess density is additive
2012-03-14 20:54:19 +05:30
forall (c = 1_pInt:2_pInt, rhoTotal(c) > 0.0_pReal) &
rhoExcessGradient_over_rho(c) = rhoExcessGradient(c) / rhoTotal(c)
!* gives the local stress correction when multiplied with a factor
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
tauBack(s) = - mu(instance) * burgers(s,instance) / (2.0_pReal * pi) &
* (rhoExcessGradient_over_rho(1) / (1.0_pReal - nu(instance)) + rhoExcessGradient_over_rho(2))
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
enddo
endif
Internal stress calculation in nonlocal model: instead of integration of excess gradients (->Bayley) we now sum up contributions from adjacent superdislocations within a certain radius R. When periodicity is used, also periodic images are considered in stress evaluation.
2011-03-29 13:04:33 +05:30
dead dislocations now exert a backstress on their "home" MP
2011-05-26 15:05:42 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
!*** set dependent states
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
use named state variable indices in "microstructure"
2013-05-24 02:58:45 +05:30
state(gr,ip,el)%p(iRhoF(1:ns,instance)) = rhoForest
rename state variable index for critical resolved shear stress from "iTau" to "iTauF"
2013-05-24 03:02:36 +05:30
state(gr,ip,el)%p(iTauF(1:ns,instance)) = tauThreshold
use named state variable indices in "microstructure"
2013-05-24 02:58:45 +05:30
state(gr,ip,el)%p(iTauB(1:ns,instance)) = tauBack
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
#ifndef _OPENMP
In crystallite: for state integrators 2 and 3 the deltaState always has to be done at the beginning of the integration, not at the end; otherwise the stiffness calculation can be completely wrong, when the state is changed after the stress integration for the central solution. Also changed some debugging levels in crystallite and mainly in constitutive_nonlocal
2012-10-22 13:29:35 +05:30
if (iand(debug_level(debug_constitutive),debug_levelExtensive) /= 0_pInt &
polishing
2013-05-23 17:55:56 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == gr)&
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
.or. .not. iand(debug_level(debug_constitutive),debug_levelSelective) /= 0_pInt)) then
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
write(6,*)
polishing
2013-05-23 17:55:56 +05:30
write(6,'(a,i8,1x,i2,1x,i1)') '<< CONST >> nonlocal_microstructure at el ip g',el,ip,gr
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
write(6,*)
some f2008 standard related correction i didn't check in last time
2012-02-02 01:50:05 +05:30
write(6,'(a,/,12x,12(e10.3,1x))') '<< CONST >> rhoForest', rhoForest
write(6,'(a,/,12x,12(f10.5,1x))') '<< CONST >> tauThreshold / MPa', tauThreshold/1e6
write(6,'(a,/,12x,12(f10.5,1x))') '<< CONST >> tauBack / MPa', tauBack/1e6
modified multiplication term in order to mimic a discrete distribution of dislocation sources; namely, at high resolution, when there is (on average) less than one source per integration point, a random generator decides whether there is a source in the finite volume or not; if there is one, then the multiplication turns independent of the density and depends only on the velocity/volume times a scalar factor; recovers density*velocity/lambda for large finite volumes
2012-10-19 17:10:17 +05:30
write(6,*)
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
endif
#endif
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
endsubroutine
!*********************************************************************
!* calculates kinetics *
!*********************************************************************
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
subroutine constitutive_nonlocal_kinetics(v, tau, c, Temperature, state, g, ip, el, dv_dtau)
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
use debug, only: debug_level, &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
debug_constitutive, &
debug_levelBasic, &
In crystallite: for state integrators 2 and 3 the deltaState always has to be done at the beginning of the integration, not at the end; otherwise the stiffness calculation can be completely wrong, when the state is changed after the stress integration for the central solution. Also changed some debugging levels in crystallite and mainly in constitutive_nonlocal
2012-10-22 13:29:35 +05:30
debug_levelExtensive, &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
debug_levelSelective, &
* checked and corrected parallelization of code. compiles now successfully, but simply aborts computation with first parallel directive without any comment :-((( * also put a call to constitutive_microstructure at the start of each crystallite_integration subroutine like it was before. need that for nonlocal model in case of crystallite cutback
2010-11-03 22:52:48 +05:30
debug_g, &
debug_i, &
debug_e
mixed things up, did not want to commit constitutive_nonlocal.f90 but rather DAMASK_spectral.f90!!!! did a lot of polishing on constitutive_nonlocal.f90 (mostly checked in at last commit), for changes on DAMASK_spectral.f90 see previous message
2012-02-23 22:50:57 +05:30
use material, only: material_phase, &
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
phase_plasticityInstance
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
implicit none
!*** input variables
integer(pInt), intent(in) :: g, & ! current grain number
ip, & ! current integration point
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
el, & ! current element number
c ! dislocation character (1:edge, 2:screw)
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
real(pReal), intent(in) :: Temperature ! temperature
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el)))), &
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
intent(in) :: tau ! resolved external shear stress (for bcc this already contains non Schmid effects)
type(p_vec), intent(in) :: state ! microstructural state
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
constitutive_nonlocal: dislocation velocities are stored in the state, so we actually now have three "parts" of the state, the basic states that are updated by "constitutive_dotState" come first, then the dependent states that are calculated by "constitutive_microstructure" follow, and finally we have a last part reserved for other variables that just use the memory reserved by the state array and are updated somewhere else. constitutive: LpAndItsTangent does not need the full state, but only the local state, so changed that at least for the nonlocal constitutive law
2011-11-04 18:42:17 +05:30
!*** input/output variables
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
!*** output variables
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el)))), &
constitutive_nonlocal: dislocation velocities are stored in the state, so we actually now have three "parts" of the state, the basic states that are updated by "constitutive_dotState" come first, then the dependent states that are calculated by "constitutive_microstructure" follow, and finally we have a last part reserved for other variables that just use the memory reserved by the state array and are updated somewhere else. constitutive: LpAndItsTangent does not need the full state, but only the local state, so changed that at least for the nonlocal constitutive law
2011-11-04 18:42:17 +05:30
intent(out) :: v ! velocity
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el)))), &
* now able to choose method for state integration (integrator and integratorStiffness in numerics.config); standard is "Fixed Point Iteration", which is basically the same as before; others available are "Explicit Euler", "AdaptiveEuler", "Classical Runge-Kutta" and "Runge-Kutta Cash-Karp" * now remembering stiffness similar to how we do it for Lp etc.; avoids undefined stiffness values for nonconverged stiffness calculation * non-local stuff: * changed non-local kinetics (Gilman2002) * enforce zero shearrate for overall carrrier density below relevant density * enforce zero density for those states that become negative and were below relevant density before * dislocation velocity is not limited by V^(1/3) / dt anymore
2010-10-01 17:48:49 +05:30
intent(out), optional :: dv_dtau ! velocity derivative with respect to resolved shear stress
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
!*** local variables
removed unused variables, added pure statements, and declared external functions as external (using gfortrans debug option)
2013-02-11 16:13:45 +05:30
integer(pInt) :: instance, & ! current instance of this plasticity
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
ns, & ! short notation for the total number of active slip systems
removed unused variables, added pure statements, and declared external functions as external (using gfortrans debug option)
2013-02-11 16:13:45 +05:30
s ! index of my current slip system
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el)))) :: &
viscosity term in kinetics law proportional to resolvedstress minus forest term
2012-03-14 20:48:36 +05:30
tauThreshold, & ! threshold shear stress
tauEff ! effective shear stress
corrected kinetics law and changed parameters. For solid solution hardening there are 3 parameters: the activation energy, the concentration of obstacles that determines the activation length and meanfreepath, and the obstacle size that determines the activation volume. For the Peierls mechanism there is: the width of doublekinks that determines the activation volume and the Peierls stress for edge and screw. Still testing needed to check whether the current formulation makes sense or not.
2012-02-03 18:20:54 +05:30
real(pReal) tauRel_P, &
tauRel_S, &
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
tPeierls, & ! waiting time in front of a peierls barriers
corrected kinetics law and changed parameters. For solid solution hardening there are 3 parameters: the activation energy, the concentration of obstacles that determines the activation length and meanfreepath, and the obstacle size that determines the activation volume. For the Peierls mechanism there is: the width of doublekinks that determines the activation volume and the Peierls stress for edge and screw. Still testing needed to check whether the current formulation makes sense or not.
2012-02-03 18:20:54 +05:30
tSolidSolution, & ! waiting time in front of a solid solution obstacle
vViscous, & ! viscous glide velocity
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
dtPeierls_dtau, & ! derivative with respect to resolved shear stress
corrected kinetics law and changed parameters. For solid solution hardening there are 3 parameters: the activation energy, the concentration of obstacles that determines the activation length and meanfreepath, and the obstacle size that determines the activation volume. For the Peierls mechanism there is: the width of doublekinks that determines the activation volume and the Peierls stress for edge and screw. Still testing needed to check whether the current formulation makes sense or not.
2012-02-03 18:20:54 +05:30
dtSolidSolution_dtau, & ! derivative with respect to resolved shear stress
meanfreepath_S, & ! mean free travel distance for dislocations between two solid solution obstacles
meanfreepath_P, & ! mean free travel distance for dislocations between two Peierls barriers
jumpWidth_P, & ! depth of activated area
jumpWidth_S, & ! depth of activated area
activationLength_P, & ! length of activated dislocation line
activationLength_S, & ! length of activated dislocation line
activationVolume_P, & ! volume that needs to be activated to overcome barrier
activationVolume_S, & ! volume that needs to be activated to overcome barrier
activationEnergy_P, & ! energy that is needed to overcome barrier
activationEnergy_S, & ! energy that is needed to overcome barrier
criticalStress_P, & ! maximum obstacle strength
criticalStress_S, & ! maximum obstacle strength
mobility ! dislocation mobility
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
instance = phase_plasticityInstance(material_phase(g,ip,el))
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
ns = totalNslip(instance)
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
use named state variable indices in "kinetics" and "microstructure"
2013-05-24 03:10:00 +05:30
tauThreshold = state%p(iTauF(1:ns,instance))
viscosity term in kinetics law proportional to resolvedstress minus forest term
2012-03-14 20:48:36 +05:30
tauEff = abs(tau) - tauThreshold
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
openmp parallelization working again (at least for j2 and nonlocal constitutive model). In order to keep it like that, please follow these simple rules: DON'T use implicit array subscripts: example: real, dimension(3,3) :: A,B A(:,2) = B(:,1) <--- DON'T USE A(1:3,2) = B(1:3,1) <--- BETTER USE In many cases the use of explicit array subscripts is inevitable for parallelization. Additionally, it is an easy means to prevent memory leaks. Enclose all write statements with the following: !$OMP CRITICAL (write2out) <your write statement> !$OMP END CRITICAL (write2out) Whenever you change something in the code and are not sure if it affects parallelization and leads to nonconforming behavior, please ask me and/or Franz to check this.
2011-03-17 16:16:17 +05:30
v = 0.0_pReal
* now able to choose method for state integration (integrator and integratorStiffness in numerics.config); standard is "Fixed Point Iteration", which is basically the same as before; others available are "Explicit Euler", "AdaptiveEuler", "Classical Runge-Kutta" and "Runge-Kutta Cash-Karp" * now remembering stiffness similar to how we do it for Lp etc.; avoids undefined stiffness values for nonconverged stiffness calculation * non-local stuff: * changed non-local kinetics (Gilman2002) * enforce zero shearrate for overall carrrier density below relevant density * enforce zero density for those states that become negative and were below relevant density before * dislocation velocity is not limited by V^(1/3) / dt anymore
2010-10-01 17:48:49 +05:30
if (present(dv_dtau)) dv_dtau = 0.0_pReal
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
* kinetics now according to Mohles; enables solid solution hardening * changed example in material.config accordingly
2011-01-26 15:47:42 +05:30
if (Temperature > 0.0_pReal) then
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do s = 1_pInt,ns
viscosity term in kinetics law proportional to resolvedstress minus forest term
2012-03-14 20:48:36 +05:30
if (tauEff(s) > 0.0_pReal) then
* corrected compatibility for screws (always positive) * detection of grain boundary in constitutive_nonlocal_microstructure with the help of transmissivity * enforce positive densities in constitutive_nonlocal_microstructure (needed because dotState does not create cutbacks for negative densities anymore) * reset single mobile densities below certain threshold to zero (also done in constitutive_nonlocal_microstructure) * constitutive_nonlocal_kinetics only gets local state variable as input, no need for the entire array here * dv_dtau is always positive * multiplication is only active when there is already some initial density of the respective type
2010-10-26 19:12:18 +05:30
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
!* Peierls contribution
!* The derivative only gives absolute values; the correct sign is taken care of in the formula for the derivative of the velocity
corrected kinetics law and changed parameters. For solid solution hardening there are 3 parameters: the activation energy, the concentration of obstacles that determines the activation length and meanfreepath, and the obstacle size that determines the activation volume. For the Peierls mechanism there is: the width of doublekinks that determines the activation volume and the Peierls stress for edge and screw. Still testing needed to check whether the current formulation makes sense or not.
2012-02-03 18:20:54 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
meanfreepath_P = burgers(s,instance)
jumpWidth_P = burgers(s,instance)
activationLength_P = doublekinkwidth(instance) * burgers(s,instance)
activationVolume_P = activationLength_P * jumpWidth_P * burgers(s,instance)
criticalStress_P = peierlsStress(s,c,instance)
corrected kinetics law and changed parameters. For solid solution hardening there are 3 parameters: the activation energy, the concentration of obstacles that determines the activation length and meanfreepath, and the obstacle size that determines the activation volume. For the Peierls mechanism there is: the width of doublekinks that determines the activation volume and the Peierls stress for edge and screw. Still testing needed to check whether the current formulation makes sense or not.
2012-02-03 18:20:54 +05:30
activationEnergy_P = criticalStress_P * activationVolume_P
added relativistic correction of velocity (maximum velocity given by parameter in material.config); probability of thermal activation in velocity law not allowed to be greater than one. output of creation rate for edge jogs now for own system, not collinear. check for negative densities in dotState now uses absoluteToleranceRho instead of significantRho.
2012-08-30 13:03:13 +05:30
tauRel_P = min(1.0_pReal, tauEff(s) / criticalStress_P) ! ensure that the activation probability cannot become greater than one
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
tPeierls = 1.0_pReal / fattack(instance) &
* exp(activationEnergy_P / (KB * Temperature) &
* (1.0_pReal - tauRel_P**pParam(instance))**qParam(instance))
* kinetics now according to Mohles; enables solid solution hardening * changed example in material.config accordingly
2011-01-26 15:47:42 +05:30
if (present(dv_dtau)) then
added relativistic correction of velocity (maximum velocity given by parameter in material.config); probability of thermal activation in velocity law not allowed to be greater than one. output of creation rate for edge jogs now for own system, not collinear. check for negative densities in dotState now uses absoluteToleranceRho instead of significantRho.
2012-08-30 13:03:13 +05:30
if (tauEff(s) < criticalStress_P) then
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
dtPeierls_dtau = tPeierls * pParam(instance) * qParam(instance) * activationVolume_P / (KB * Temperature) &
* (1.0_pReal - tauRel_P**pParam(instance))**(qParam(instance)-1.0_pReal) &
* tauRel_P**(pParam(instance)-1.0_pReal)
added relativistic correction of velocity (maximum velocity given by parameter in material.config); probability of thermal activation in velocity law not allowed to be greater than one. output of creation rate for edge jogs now for own system, not collinear. check for negative densities in dotState now uses absoluteToleranceRho instead of significantRho.
2012-08-30 13:03:13 +05:30
else
dtPeierls_dtau = 0.0_pReal
endif
* kinetics now according to Mohles; enables solid solution hardening * changed example in material.config accordingly
2011-01-26 15:47:42 +05:30
endif
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
!* Contribution from solid solution strengthening
!* The derivative only gives absolute values; the correct sign is taken care of in the formula for the derivative of the velocity
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
meanfreepath_S = burgers(s,instance) / sqrt(solidSolutionConcentration(instance))
jumpWidth_S = solidSolutionSize(instance) * burgers(s,instance)
activationLength_S = burgers(s,instance) / sqrt(solidSolutionConcentration(instance))
activationVolume_S = activationLength_S * jumpWidth_S * burgers(s,instance)
activationEnergy_S = solidSolutionEnergy(instance)
corrected kinetics law and changed parameters. For solid solution hardening there are 3 parameters: the activation energy, the concentration of obstacles that determines the activation length and meanfreepath, and the obstacle size that determines the activation volume. For the Peierls mechanism there is: the width of doublekinks that determines the activation volume and the Peierls stress for edge and screw. Still testing needed to check whether the current formulation makes sense or not.
2012-02-03 18:20:54 +05:30
criticalStress_S = activationEnergy_S / activationVolume_S
added relativistic correction of velocity (maximum velocity given by parameter in material.config); probability of thermal activation in velocity law not allowed to be greater than one. output of creation rate for edge jogs now for own system, not collinear. check for negative densities in dotState now uses absoluteToleranceRho instead of significantRho.
2012-08-30 13:03:13 +05:30
tauRel_S = min(1.0_pReal, tauEff(s) / criticalStress_S) ! ensure that the activation probability cannot become greater than one
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
tSolidSolution = 1.0_pReal / fattack(instance) &
* exp(activationEnergy_S / (KB * Temperature) &
* (1.0_pReal - tauRel_S**pParam(instance))**qParam(instance))
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
if (present(dv_dtau)) then
added relativistic correction of velocity (maximum velocity given by parameter in material.config); probability of thermal activation in velocity law not allowed to be greater than one. output of creation rate for edge jogs now for own system, not collinear. check for negative densities in dotState now uses absoluteToleranceRho instead of significantRho.
2012-08-30 13:03:13 +05:30
if (tauEff(s) < criticalStress_S) then
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
dtSolidSolution_dtau = tSolidSolution * pParam(instance) * qParam(instance) &
* activationVolume_S / (KB * Temperature) &
* (1.0_pReal - tauRel_S**pParam(instance))**(qParam(instance)-1.0_pReal) &
* tauRel_S**(pParam(instance)-1.0_pReal)
added relativistic correction of velocity (maximum velocity given by parameter in material.config); probability of thermal activation in velocity law not allowed to be greater than one. output of creation rate for edge jogs now for own system, not collinear. check for negative densities in dotState now uses absoluteToleranceRho instead of significantRho.
2012-08-30 13:03:13 +05:30
else
dtSolidSolution_dtau = 0.0_pReal
endif
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
endif
corrected kinetics law and changed parameters. For solid solution hardening there are 3 parameters: the activation energy, the concentration of obstacles that determines the activation length and meanfreepath, and the obstacle size that determines the activation volume. For the Peierls mechanism there is: the width of doublekinks that determines the activation volume and the Peierls stress for edge and screw. Still testing needed to check whether the current formulation makes sense or not.
2012-02-03 18:20:54 +05:30
!* viscous glide velocity
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
mobility = burgers(s,instance) / viscosity(instance)
viscosity term in kinetics law proportional to resolvedstress minus forest term
2012-03-14 20:48:36 +05:30
vViscous = mobility * tauEff(s)
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
corrected kinetics law and changed parameters. For solid solution hardening there are 3 parameters: the activation energy, the concentration of obstacles that determines the activation length and meanfreepath, and the obstacle size that determines the activation volume. For the Peierls mechanism there is: the width of doublekinks that determines the activation volume and the Peierls stress for edge and screw. Still testing needed to check whether the current formulation makes sense or not.
2012-02-03 18:20:54 +05:30
!* Mean velocity results from waiting time at peierls barriers and solid solution obstacles with respective meanfreepath of
simplified dislocation kinetics: dropped relativistic correction of velocity, since the infinite attack frequency already sets a limit for the velocity; dropped second; also dropped the correction term for backward jumps, since it was probably not meaningful
2013-03-21 18:22:29 +05:30
!* free flight at glide velocity in between.
!* adopt sign from resolved stress
added relativistic correction of velocity (maximum velocity given by parameter in material.config); probability of thermal activation in velocity law not allowed to be greater than one. output of creation rate for edge jogs now for own system, not collinear. check for negative densities in dotState now uses absoluteToleranceRho instead of significantRho.
2012-08-30 13:03:13 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
v(s) = sign(1.0_pReal,tau(s)) &
/ (tPeierls / meanfreepath_P + tSolidSolution / meanfreepath_S + 1.0_pReal / vViscous)
added relativistic correction of velocity (maximum velocity given by parameter in material.config); probability of thermal activation in velocity law not allowed to be greater than one. output of creation rate for edge jogs now for own system, not collinear. check for negative densities in dotState now uses absoluteToleranceRho instead of significantRho.
2012-08-30 13:03:13 +05:30
if (present(dv_dtau)) then
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
dv_dtau(s) = v(s) * v(s) &
* (dtPeierls_dtau / meanfreepath_P &
+ dtSolidSolution_dtau / meanfreepath_S &
+ 1.0_pReal / (mobility * tauEff(s)*tauEff(s)))
added relativistic correction of velocity (maximum velocity given by parameter in material.config); probability of thermal activation in velocity law not allowed to be greater than one. output of creation rate for edge jogs now for own system, not collinear. check for negative densities in dotState now uses absoluteToleranceRho instead of significantRho.
2012-08-30 13:03:13 +05:30
endif
* now able to choose method for state integration (integrator and integratorStiffness in numerics.config); standard is "Fixed Point Iteration", which is basically the same as before; others available are "Explicit Euler", "AdaptiveEuler", "Classical Runge-Kutta" and "Runge-Kutta Cash-Karp" * now remembering stiffness similar to how we do it for Lp etc.; avoids undefined stiffness values for nonconverged stiffness calculation * non-local stuff: * changed non-local kinetics (Gilman2002) * enforce zero shearrate for overall carrrier density below relevant density * enforce zero density for those states that become negative and were below relevant density before * dislocation velocity is not limited by V^(1/3) / dt anymore
2010-10-01 17:48:49 +05:30
endif
enddo
endif
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
openmp parallelization working again (at least for j2 and nonlocal constitutive model). In order to keep it like that, please follow these simple rules: DON'T use implicit array subscripts: example: real, dimension(3,3) :: A,B A(:,2) = B(:,1) <--- DON'T USE A(1:3,2) = B(1:3,1) <--- BETTER USE In many cases the use of explicit array subscripts is inevitable for parallelization. Additionally, it is an easy means to prevent memory leaks. Enclose all write statements with the following: !$OMP CRITICAL (write2out) <your write statement> !$OMP END CRITICAL (write2out) Whenever you change something in the code and are not sure if it affects parallelization and leads to nonconforming behavior, please ask me and/or Franz to check this.
2011-03-17 16:16:17 +05:30
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
#ifndef _OPENMP
In crystallite: for state integrators 2 and 3 the deltaState always has to be done at the beginning of the integration, not at the end; otherwise the stiffness calculation can be completely wrong, when the state is changed after the stress integration for the central solution. Also changed some debugging levels in crystallite and mainly in constitutive_nonlocal
2012-10-22 13:29:35 +05:30
if (iand(debug_level(debug_constitutive),debug_levelExtensive) /= 0_pInt &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == g)&
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
.or. .not. iand(debug_level(debug_constitutive),debug_levelSelective) /= 0_pInt)) then
debugging output is now controlled by the "verbosity" parameter in the debug.config ranging from 0 (=almost no output) to 8 (=very detailed output) 0 : only version infos and all from "hypela2"/"umat" 1 : basic outputs from "CPFEM.f90", basic output from initialization routines, debug_info 2 : extensive outputs from "CPFEM.f90", extensive output from initialization routines 3 : basic outputs from "homogenization.f90" 4 : extensive outputs from "homogenization.f90" 5 : basic outputs from "crystallite.f90" 6 : extensive outputs from "crystallite.f90" 7 : basic outputs from the constitutive files 8 : extensive outputs from the constitutive files If verbosity is equal to zero, all counters in debug are not set during calculation (e.g. debug_StressLoopDistribution or debug_cumDotStateTicks). This might speed up parallel calculation, because all these need critical statements which extremely slow down parallel computation.
2011-03-21 16:01:17 +05:30
write(6,*)
some f2008 standard related correction i didn't check in last time
2012-02-02 01:50:05 +05:30
write(6,'(a,i8,1x,i2,1x,i1)') '<< CONST >> nonlocal_kinetics at el ip g',el,ip,g
debugging output is now controlled by the "verbosity" parameter in the debug.config ranging from 0 (=almost no output) to 8 (=very detailed output) 0 : only version infos and all from "hypela2"/"umat" 1 : basic outputs from "CPFEM.f90", basic output from initialization routines, debug_info 2 : extensive outputs from "CPFEM.f90", extensive output from initialization routines 3 : basic outputs from "homogenization.f90" 4 : extensive outputs from "homogenization.f90" 5 : basic outputs from "crystallite.f90" 6 : extensive outputs from "crystallite.f90" 7 : basic outputs from the constitutive files 8 : extensive outputs from the constitutive files If verbosity is equal to zero, all counters in debug are not set during calculation (e.g. debug_StressLoopDistribution or debug_cumDotStateTicks). This might speed up parallel calculation, because all these need critical statements which extremely slow down parallel computation.
2011-03-21 16:01:17 +05:30
write(6,*)
some f2008 standard related correction i didn't check in last time
2012-02-02 01:50:05 +05:30
write(6,'(a,/,12x,12(f12.5,1x))') '<< CONST >> tau / MPa', tau / 1e6_pReal
viscosity term in kinetics law proportional to resolvedstress minus forest term
2012-03-14 20:48:36 +05:30
write(6,'(a,/,12x,12(f12.5,1x))') '<< CONST >> tauEff / MPa', tauEff / 1e6_pReal
corrected kinetics law and changed parameters. For solid solution hardening there are 3 parameters: the activation energy, the concentration of obstacles that determines the activation length and meanfreepath, and the obstacle size that determines the activation volume. For the Peierls mechanism there is: the width of doublekinks that determines the activation volume and the Peierls stress for edge and screw. Still testing needed to check whether the current formulation makes sense or not.
2012-02-03 18:20:54 +05:30
write(6,'(a,/,12x,12(f12.5,1x))') '<< CONST >> v / 1e-3m/s', v * 1e3
derivative of velocity with respect to resolved stress was flawed; now much better convergence of stress integration for nonlocal constitutive law
2012-08-21 20:10:01 +05:30
if (present(dv_dtau)) then
write(6,'(a,/,12x,12(e12.5,1x))') '<< CONST >> dv_dtau', dv_dtau
endif
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
endif
#endif
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
endsubroutine
!*********************************************************************
!* calculates plastic velocity gradient and its tangent *
!*********************************************************************
* former "relevantRho" and "relevantResistance" is renamed to "atol_rho" and "atol_resistance" and now used as an absolute tolerance for the state residuum. before it was used rather as significant state, so whenever the state dropped below that value it was considered converged. (In dislotwin and titanmod constitutive law there is only one value atol_rho which is used for all the states, though the state array consists not only of densities. We need further parameters here!) * somewhat simplified convergence check for adaptive euler and runge-kutta
2010-10-26 18:46:37 +05:30
subroutine constitutive_nonlocal_LpAndItsTangent(Lp, dLp_dTstar99, Tstar_v, Temperature, state, g, ip, el)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
use math, only: math_Plain3333to99, &
added non-schmid structure
2013-01-22 05:20:28 +05:30
math_mul6x6, &
math_Mandel6to33
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
use debug, only: debug_level, &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
debug_constitutive, &
debug_levelBasic, &
In crystallite: for state integrators 2 and 3 the deltaState always has to be done at the beginning of the integration, not at the end; otherwise the stiffness calculation can be completely wrong, when the state is changed after the stress integration for the central solution. Also changed some debugging levels in crystallite and mainly in constitutive_nonlocal
2012-10-22 13:29:35 +05:30
debug_levelExtensive, &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
debug_levelSelective, &
* checked and corrected parallelization of code. compiles now successfully, but simply aborts computation with first parallel directive without any comment :-((( * also put a call to constitutive_microstructure at the start of each crystallite_integration subroutine like it was before. need that for nonlocal model in case of crystallite cutback
2010-11-03 22:52:48 +05:30
debug_g, &
debug_i, &
debug_e
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
use material, only: homogenization_maxNgrains, &
material_phase, &
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
phase_plasticityInstance
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
use lattice, only: lattice_Sslip, &
added non-schmid structure
2013-01-22 05:20:28 +05:30
lattice_Sslip_v, &
NnonSchmid
introduced factor to control edge contribution to multiplication; dislocation density below a single dislocation per IP considered not significant
2012-09-04 22:26:37 +05:30
use mesh, only: mesh_ipVolume
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
implicit none
!*** input variables
integer(pInt), intent(in) :: g, & ! current grain number
ip, & ! current integration point
el ! current element number
real(pReal), intent(in) :: Temperature ! temperature
real(pReal), dimension(6), intent(in) :: Tstar_v ! 2nd Piola-Kirchhoff stress in Mandel notation
constitutive_nonlocal: dislocation velocities are stored in the state, so we actually now have three "parts" of the state, the basic states that are updated by "constitutive_dotState" come first, then the dependent states that are calculated by "constitutive_microstructure" follow, and finally we have a last part reserved for other variables that just use the memory reserved by the state array and are updated somewhere else. constitutive: LpAndItsTangent does not need the full state, but only the local state, so changed that at least for the nonlocal constitutive law
2011-11-04 18:42:17 +05:30
!*** input/output variables
type(p_vec), intent(inout) :: state ! microstructural state
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*** output variables
real(pReal), dimension(3,3), intent(out) :: Lp ! plastic velocity gradient
real(pReal), dimension(9,9), intent(out) :: dLp_dTstar99 ! derivative of Lp with respect to Tstar (9x9 matrix)
!*** local variables
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
integer(pInt) myInstance, & ! current instance of this plasticity
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
myStructure, & ! current lattice structure
ns, & ! short notation for the total number of active slip systems
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
c, &
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
i, &
j, &
k, &
l, &
t, & ! dislocation type
s, & ! index of my current slip system
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
sLattice ! index of my current slip system according to lattice order
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
real(pReal), dimension(3,3,3,3) :: dLp_dTstar3333 ! derivative of Lp with respect to Tstar (3x3x3x3 matrix)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(3,3,2,totalNslip(phase_plasticityInstance(material_phase(g,ip,el)))) :: &
fixed bug in constitutive_nonlocal_kinetics that was introduced in rev2088 along with the non-Schmid behavior: nonSchmid matrix of last slip system was used for all slip systems; led to extremely bad convergence due to flawed dLp_dT
2013-05-11 03:59:12 +05:30
nonSchmidTensor
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),8) :: &
- make sure that formerly blocked density is not able to move to neighbor until it is remobilized - added some explicit indices instead of ":"
2013-04-03 21:59:48 +05:30
rhoSgl ! single dislocation densities (including blocked)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),4) :: &
constitutive_nonlocal: dislocation velocities are stored in the state, so we actually now have three "parts" of the state, the basic states that are updated by "constitutive_dotState" come first, then the dependent states that are calculated by "constitutive_microstructure" follow, and finally we have a last part reserved for other variables that just use the memory reserved by the state array and are updated somewhere else. constitutive: LpAndItsTangent does not need the full state, but only the local state, so changed that at least for the nonlocal constitutive law
2011-11-04 18:42:17 +05:30
v, & ! velocity
added non-schmid structure
2013-01-22 05:20:28 +05:30
tau, & ! resolved shear stress including non Schmid and backstress terms
- make sure that formerly blocked density is not able to move to neighbor until it is remobilized - added some explicit indices instead of ":"
2013-04-03 21:59:48 +05:30
dgdot_dtau, & ! derivative of the shear rate with respect to the shear stress
* new output variable "internalstress", which gives the internal stress tensor * use "math_invert3x3" instead of "math_inv3x3" for inversion of Fe * for dislocation stress calculation: first regular case, then special case of dead dislocations in central ip * "dv_dtau" now given for each dislocation type, so is a (ns,4) array * deleted unused variables in "_LpAndItsTangent" * corrected contribution of deads in "_LpAndItsTangent" * the NaN variables defined in math did not give a proper NaN value, so use 0.0/0.0 again * neighbors with nonlocal constitution but local properties (i.e. /nonlocal/ flag not set) are also considered for incoming fluxes
2011-09-07 17:00:28 +05:30
dv_dtau ! velocity derivative with respect to the shear stress
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el)))) :: &
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
gdotTotal, & ! shear rate
added possibility to scale the plastic shearrate by the ratio of mobile to total density; keyword in material.config is "deadZone"
2012-10-29 18:19:28 +05:30
tauBack, & ! back stress from dislocation gradients on same slip system
deadZoneSize
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*** initialize local variables
Lp = 0.0_pReal
dLp_dTstar3333 = 0.0_pReal
fixed bug in constitutive_nonlocal_kinetics that was introduced in rev2088 along with the non-Schmid behavior: nonSchmid matrix of last slip system was used for all slip systems; led to extremely bad convergence due to flawed dLp_dT
2013-05-11 03:59:12 +05:30
nonSchmidTensor = 0.0_pReal
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
myInstance = phase_plasticityInstance(material_phase(g,ip,el))
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
myStructure = constitutive_nonlocal_structure(myInstance)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
ns = totalNslip(myInstance)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
openmp parallelization working again (at least for j2 and nonlocal constitutive model). In order to keep it like that, please follow these simple rules: DON'T use implicit array subscripts: example: real, dimension(3,3) :: A,B A(:,2) = B(:,1) <--- DON'T USE A(1:3,2) = B(1:3,1) <--- BETTER USE In many cases the use of explicit array subscripts is inevitable for parallelization. Additionally, it is an easy means to prevent memory leaks. Enclose all write statements with the following: !$OMP CRITICAL (write2out) <your write statement> !$OMP END CRITICAL (write2out) Whenever you change something in the code and are not sure if it affects parallelization and leads to nonconforming behavior, please ask me and/or Franz to check this.
2011-03-17 16:16:17 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*** shortcut to state variables
use named state variable indices in "kinetics" and "microstructure"
2013-05-24 03:10:00 +05:30
forall (s = 1_pInt:ns)
rhoSgl(s,1) = max(state%p(iRhoEPU(s,myInstance)), 0.0_pReal) ! ensure positive single mobile densities
rhoSgl(s,2) = max(state%p(iRhoENU(s,myInstance)), 0.0_pReal) ! ensure positive single mobile densities
rhoSgl(s,3) = max(state%p(iRhoSPU(s,myInstance)), 0.0_pReal) ! ensure positive single mobile densities
rhoSgl(s,4) = max(state%p(iRhoSNU(s,myInstance)), 0.0_pReal) ! ensure positive single mobile densities
endforall
rhoSgl(1:ns,5) = state%p(iRhoEPB(1:ns,myInstance))
rhoSgl(1:ns,6) = state%p(iRhoENB(1:ns,myInstance))
rhoSgl(1:ns,7) = state%p(iRhoSPB(1:ns,myInstance))
rhoSgl(1:ns,8) = state%p(iRhoSNB(1:ns,myInstance))
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
where (abs(rhoSgl) * mesh_ipVolume(ip,el) ** 0.667_pReal < significantN(myInstance) &
use named state variable indices in "kinetics" and "microstructure"
2013-05-24 03:10:00 +05:30
.or. abs(rhoSgl) < significantRho(myInstance)) &
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
rhoSgl = 0.0_pReal
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
use named state variable indices in "kinetics" and "microstructure"
2013-05-24 03:10:00 +05:30
tauBack = state%p(iTauB(1:ns,myInstance))
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
!*** get effective resolved shear stress
fixed bug in constitutive_nonlocal_kinetics that was introduced in rev2088 along with the non-Schmid behavior: nonSchmid matrix of last slip system was used for all slip systems; led to extremely bad convergence due to flawed dLp_dT
2013-05-11 03:59:12 +05:30
!*** add non schmid contributions to ONLY screw components if present (i.e. if NnonSchmid(myStructure) > 0)
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do s = 1_pInt,ns
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sLattice = slipSystemLattice(s,myInstance)
fixed bug in constitutive_nonlocal_kinetics that was introduced in rev2088 along with the non-Schmid behavior: nonSchmid matrix of last slip system was used for all slip systems; led to extremely bad convergence due to flawed dLp_dT
2013-05-11 03:59:12 +05:30
tau(s,1:4) = math_mul6x6(Tstar_v, lattice_Sslip_v(1:6,1,sLattice,myStructure)) + tauBack(s)
nonSchmidTensor(1:3,1:3,1,s) = lattice_Sslip(1:3,1:3,sLattice,myStructure)
nonSchmidTensor(1:3,1:3,2,s) = nonSchmidTensor(1:3,1:3,1,s)
added non-schmid structure
2013-01-22 05:20:28 +05:30
do k = 1_pInt, NnonSchmid(myStructure)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
tau(s,3) = tau(s,3) + nonSchmidCoeff(k,myInstance) &
fixed bug in constitutive_nonlocal_kinetics that was introduced in rev2088 along with the non-Schmid behavior: nonSchmid matrix of last slip system was used for all slip systems; led to extremely bad convergence due to flawed dLp_dT
2013-05-11 03:59:12 +05:30
* math_mul6x6(Tstar_v, lattice_Sslip_v(1:6,2*k,sLattice,myStructure))
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
tau(s,4) = tau(s,4) + nonSchmidCoeff(k,myInstance) &
fixed bug in constitutive_nonlocal_kinetics that was introduced in rev2088 along with the non-Schmid behavior: nonSchmid matrix of last slip system was used for all slip systems; led to extremely bad convergence due to flawed dLp_dT
2013-05-11 03:59:12 +05:30
* math_mul6x6(Tstar_v, lattice_Sslip_v(1:6,2*k+1,sLattice,myStructure))
nonSchmidTensor(1:3,1:3,1,s) = nonSchmidTensor(1:3,1:3,1,s) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
+ nonSchmidCoeff(k,myInstance) &
fixed bug in constitutive_nonlocal_kinetics that was introduced in rev2088 along with the non-Schmid behavior: nonSchmid matrix of last slip system was used for all slip systems; led to extremely bad convergence due to flawed dLp_dT
2013-05-11 03:59:12 +05:30
* math_Mandel6to33(lattice_Sslip_v(1:6,2*k,sLattice,myStructure))
nonSchmidTensor(1:3,1:3,2,s) = nonSchmidTensor(1:3,1:3,2,s) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
+ nonSchmidCoeff(k,myInstance) &
fixed bug in constitutive_nonlocal_kinetics that was introduced in rev2088 along with the non-Schmid behavior: nonSchmid matrix of last slip system was used for all slip systems; led to extremely bad convergence due to flawed dLp_dT
2013-05-11 03:59:12 +05:30
* math_Mandel6to33(lattice_Sslip_v(1:6,2*k+1,sLattice,myStructure))
added non-schmid structure
2013-01-22 05:20:28 +05:30
enddo
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
enddo
!*** get dislocation velocity and its tangent and store the velocity in the state array
fixed bug in constitutive_nonlocal_kinetics that was introduced in rev2088 along with the non-Schmid behavior: nonSchmid matrix of last slip system was used for all slip systems; led to extremely bad convergence due to flawed dLp_dT
2013-05-11 03:59:12 +05:30
if (myStructure == 1_pInt .and. NnonSchmid(myStructure) == 0_pInt) then ! for fcc all velcities are equal
call constitutive_nonlocal_kinetics(v(1:ns,1), tau(1:ns,1), 1_pInt, Temperature, state, &
g, ip, el, dv_dtau(1:ns,1))
polishing, adding _pInt etc. where applicable post_results now handels zero increment different (like FEM, it is always there even if it is not counted)
2012-02-13 19:48:07 +05:30
do t = 1_pInt,4_pInt
dislocation velocity output for each dislocation type (edge+,edge-,screw+,screw-) for fcc constitutive_nonlocal_kinetics is called only once and the result is used for each dislocation type, for all other lattice structures we have one call for each dislocation type
2012-01-26 13:13:36 +05:30
v(1:ns,t) = v(1:ns,1)
dv_dtau(1:ns,t) = dv_dtau(1:ns,1)
enddo
use named state variable indices in "kinetics" and "microstructure"
2013-05-24 03:10:00 +05:30
state%p(iVEP(1:ns,myInstance)) = v(1:ns,1)
state%p(iVEN(1:ns,myInstance)) = v(1:ns,1)
state%p(iVSP(1:ns,myInstance)) = v(1:ns,1)
state%p(iVSN(1:ns,myInstance)) = v(1:ns,1)
fixed bug in constitutive_nonlocal_kinetics that was introduced in rev2088 along with the non-Schmid behavior: nonSchmid matrix of last slip system was used for all slip systems; led to extremely bad convergence due to flawed dLp_dT
2013-05-11 03:59:12 +05:30
else ! for all other lattice structures the velocities may vary with character and sign
polishing, adding _pInt etc. where applicable post_results now handels zero increment different (like FEM, it is always there even if it is not counted)
2012-02-13 19:48:07 +05:30
do t = 1_pInt,4_pInt
c = (t-1_pInt)/2_pInt+1_pInt
fixed bug in constitutive_nonlocal_kinetics that was introduced in rev2088 along with the non-Schmid behavior: nonSchmid matrix of last slip system was used for all slip systems; led to extremely bad convergence due to flawed dLp_dT
2013-05-11 03:59:12 +05:30
call constitutive_nonlocal_kinetics(v(1:ns,t), tau(1:ns,t), c, Temperature, state, &
g, ip, el, dv_dtau(1:ns,t))
dislocation velocity output for each dislocation type (edge+,edge-,screw+,screw-) for fcc constitutive_nonlocal_kinetics is called only once and the result is used for each dislocation type, for all other lattice structures we have one call for each dislocation type
2012-01-26 13:13:36 +05:30
enddo
use named state variable indices in "kinetics" and "microstructure"
2013-05-24 03:10:00 +05:30
state%p(iVEP(1:ns,myInstance)) = v(1:ns,1)
state%p(iVEN(1:ns,myInstance)) = v(1:ns,2)
state%p(iVSP(1:ns,myInstance)) = v(1:ns,3)
state%p(iVSN(1:ns,myInstance)) = v(1:ns,4)
dislocation velocity output for each dislocation type (edge+,edge-,screw+,screw-) for fcc constitutive_nonlocal_kinetics is called only once and the result is used for each dislocation type, for all other lattice structures we have one call for each dislocation type
2012-01-26 13:13:36 +05:30
endif
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
!*** Bauschinger effect
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
forall (s = 1_pInt:ns, t = 5_pInt:8_pInt, rhoSgl(s,t) * v(s,t-4_pInt) < 0.0_pReal) &
rhoSgl(s,t-4_pInt) = rhoSgl(s,t-4_pInt) + abs(rhoSgl(s,t))
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
!*** Calculation of gdot and its tangent
added possibility to scale the plastic shearrate by the ratio of mobile to total density; keyword in material.config is "deadZone"
2012-10-29 18:19:28 +05:30
deadZoneSize = 0.0_pReal
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (deadZoneScaling(myInstance)) then
added possibility to scale the plastic shearrate by the ratio of mobile to total density; keyword in material.config is "deadZone"
2012-10-29 18:19:28 +05:30
forall(s = 1_pInt:ns, sum(abs(rhoSgl(s,1:8))) > 0.0_pReal) &
deadZoneSize(s) = maxval(abs(rhoSgl(s,5:8)) / (rhoSgl(s,1:4) + abs(rhoSgl(s,5:8))))
endif
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
gdotTotal = sum(rhoSgl(1:ns,1:4) * v, 2) * burgers(1:ns,myInstance) * (1.0_pReal - deadZoneSize)
added non-schmid structure
2013-01-22 05:20:28 +05:30
do t = 1_pInt,4_pInt
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
dgdot_dtau(:,t) = rhoSgl(1:ns,t) * dv_dtau(1:ns,t) * burgers(1:ns,myInstance) * (1.0_pReal - deadZoneSize)
added non-schmid structure
2013-01-22 05:20:28 +05:30
enddo
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
shearrate(1:ns,g,ip,el) = gdotTotal
openmp parallelization working again (at least for j2 and nonlocal constitutive model). In order to keep it like that, please follow these simple rules: DON'T use implicit array subscripts: example: real, dimension(3,3) :: A,B A(:,2) = B(:,1) <--- DON'T USE A(1:3,2) = B(1:3,1) <--- BETTER USE In many cases the use of explicit array subscripts is inevitable for parallelization. Additionally, it is an easy means to prevent memory leaks. Enclose all write statements with the following: !$OMP CRITICAL (write2out) <your write statement> !$OMP END CRITICAL (write2out) Whenever you change something in the code and are not sure if it affects parallelization and leads to nonconforming behavior, please ask me and/or Franz to check this.
2011-03-17 16:16:17 +05:30
fixed bug in constitutive_nonlocal_kinetics that was introduced in rev2088 along with the non-Schmid behavior: nonSchmid matrix of last slip system was used for all slip systems; led to extremely bad convergence due to flawed dLp_dT
2013-05-11 03:59:12 +05:30
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
!*** Calculation of Lp and its tangent
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do s = 1_pInt,ns
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sLattice = slipSystemLattice(s,myInstance)
In order to help preventing further memory leaks all array sections now have an explicit instead of assumed shape, e.g. Fe(1:3,1:3) instead of Fe(:,:).
2011-02-09 18:42:46 +05:30
Lp = Lp + gdotTotal(s) * lattice_Sslip(1:3,1:3,sLattice,myStructure)
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
forall (i=1_pInt:3_pInt,j=1_pInt:3_pInt,k=1_pInt:3_pInt,l=1_pInt:3_pInt) &
fixed bug in constitutive_nonlocal_kinetics that was introduced in rev2088 along with the non-Schmid behavior: nonSchmid matrix of last slip system was used for all slip systems; led to extremely bad convergence due to flawed dLp_dT
2013-05-11 03:59:12 +05:30
dLp_dTstar3333(i,j,k,l) = dLp_dTstar3333(i,j,k,l) &
+ dgdot_dtau(s,1) * lattice_Sslip(i,j,sLattice,myStructure) * lattice_Sslip(k,l,sLattice,myStructure) &
+ dgdot_dtau(s,2) * lattice_Sslip(i,j,sLattice,myStructure) * lattice_Sslip(k,l,sLattice,myStructure) &
+ dgdot_dtau(s,3) * lattice_Sslip(i,j,sLattice,myStructure) * nonSchmidTensor(k,l,1,s) &
+ dgdot_dtau(s,4) * lattice_Sslip(i,j,sLattice,myStructure) * nonSchmidTensor(k,l,2,s)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
enddo
added dislocation multiplication to dotState in constitutive_nonlocal.f90 cleaned up debugging output statements to *.out file
2009-08-12 16:52:02 +05:30
dLp_dTstar99 = math_Plain3333to99(dLp_dTstar3333)
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
#ifndef _OPENMP
In crystallite: for state integrators 2 and 3 the deltaState always has to be done at the beginning of the integration, not at the end; otherwise the stiffness calculation can be completely wrong, when the state is changed after the stress integration for the central solution. Also changed some debugging levels in crystallite and mainly in constitutive_nonlocal
2012-10-22 13:29:35 +05:30
if (iand(debug_level(debug_constitutive),debug_levelExtensive) /= 0_pInt &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == g)&
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
.or. .not. iand(debug_level(debug_constitutive),debug_levelSelective) /= 0_pInt )) then
debugging output is now controlled by the "verbosity" parameter in the debug.config ranging from 0 (=almost no output) to 8 (=very detailed output) 0 : only version infos and all from "hypela2"/"umat" 1 : basic outputs from "CPFEM.f90", basic output from initialization routines, debug_info 2 : extensive outputs from "CPFEM.f90", extensive output from initialization routines 3 : basic outputs from "homogenization.f90" 4 : extensive outputs from "homogenization.f90" 5 : basic outputs from "crystallite.f90" 6 : extensive outputs from "crystallite.f90" 7 : basic outputs from the constitutive files 8 : extensive outputs from the constitutive files If verbosity is equal to zero, all counters in debug are not set during calculation (e.g. debug_StressLoopDistribution or debug_cumDotStateTicks). This might speed up parallel calculation, because all these need critical statements which extremely slow down parallel computation.
2011-03-21 16:01:17 +05:30
write(6,*)
some f2008 standard related correction i didn't check in last time
2012-02-02 01:50:05 +05:30
write(6,'(a,i8,1x,i2,1x,i1)') '<< CONST >> nonlocal_LpandItsTangent at el ip g ',el,ip,g
debugging output is now controlled by the "verbosity" parameter in the debug.config ranging from 0 (=almost no output) to 8 (=very detailed output) 0 : only version infos and all from "hypela2"/"umat" 1 : basic outputs from "CPFEM.f90", basic output from initialization routines, debug_info 2 : extensive outputs from "CPFEM.f90", extensive output from initialization routines 3 : basic outputs from "homogenization.f90" 4 : extensive outputs from "homogenization.f90" 5 : basic outputs from "crystallite.f90" 6 : extensive outputs from "crystallite.f90" 7 : basic outputs from the constitutive files 8 : extensive outputs from the constitutive files If verbosity is equal to zero, all counters in debug are not set during calculation (e.g. debug_StressLoopDistribution or debug_cumDotStateTicks). This might speed up parallel calculation, because all these need critical statements which extremely slow down parallel computation.
2011-03-21 16:01:17 +05:30
write(6,*)
some f2008 standard related correction i didn't check in last time
2012-02-02 01:50:05 +05:30
write(6,'(a,/,12x,12(f12.5,1x))') '<< CONST >> gdot total / 1e-3',gdotTotal*1e3_pReal
fixed bug in constitutive_nonlocal_kinetics that was introduced in rev2088 along with the non-Schmid behavior: nonSchmid matrix of last slip system was used for all slip systems; led to extremely bad convergence due to flawed dLp_dT
2013-05-11 03:59:12 +05:30
write(6,'(a,/,3(12x,3(f12.7,1x),/))') '<< CONST >> Lp',transpose(Lp)
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
endif
#endif
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
endsubroutine
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
!*********************************************************************
!* incremental change of microstructure *
!*********************************************************************
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
subroutine constitutive_nonlocal_deltaState(deltaState, state, Tstar_v, Temperature, g,ip,el)
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
use debug, only: debug_level, &
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
debug_constitutive, &
debug_levelBasic, &
In crystallite: for state integrators 2 and 3 the deltaState always has to be done at the beginning of the integration, not at the end; otherwise the stiffness calculation can be completely wrong, when the state is changed after the stress integration for the central solution. Also changed some debugging levels in crystallite and mainly in constitutive_nonlocal
2012-10-22 13:29:35 +05:30
debug_levelExtensive, &
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
debug_levelSelective, &
debug_g, &
debug_i, &
debug_e
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
use math, only: pi, &
math_mul6x6
use lattice, only: lattice_Sslip_v
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
use mesh, only: mesh_NcpElems, &
introduced factor to control edge contribution to multiplication; dislocation density below a single dislocation per IP considered not significant
2012-09-04 22:26:37 +05:30
mesh_maxNips, &
mesh_ipVolume
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
use material, only: homogenization_maxNgrains, &
material_phase, &
phase_plasticityInstance
implicit none
!*** input variables
integer(pInt), intent(in) :: g, & ! current grain number
ip, & ! current integration point
el ! current element number
real(pReal), intent(in) :: Temperature ! temperature
real(pReal), dimension(6), intent(in) :: Tstar_v ! current 2nd Piola-Kirchhoff stress in Mandel notation
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
!*** input/output variables
type(p_vec), dimension(homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), intent(inout) :: &
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
state ! current microstructural state
!*** output variables
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
type(p_vec), intent(out) :: deltaState ! change of state variables / microstructure
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
!*** local variables
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
integer(pInt) myInstance, & ! current instance of this plasticity
myStructure, & ! current lattice structure
ns, & ! short notation for the total number of active slip systems
c, & ! character of dislocation
t, & ! type of dislocation
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
s, & ! index of my current slip system
sLattice ! index of my current slip system according to lattice order
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),10) :: &
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
deltaRho, & ! density increment
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
deltaRhoRemobilization, & ! density increment by remobilization
annihilate only part of the screw dipoles (specified by minimumDipoleHeight), not all; moved annihilation of screws from deltaState back to dotState
2012-11-28 17:39:48 +05:30
deltaRhoDipole2SingleStress ! density increment by dipole dissociation (by stress change)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),8) :: &
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
rhoSgl ! current single dislocation densities (positive/negative screw and edge without dipoles)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),4) :: &
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
v ! dislocation glide velocity
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el)))) :: &
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
tau, & ! current resolved shear stress
annihilate only part of the screw dipoles (specified by minimumDipoleHeight), not all; moved annihilation of screws from deltaState back to dotState
2012-11-28 17:39:48 +05:30
tauBack ! current back stress from pileups on same slip system
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),2) :: &
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
rhoDip, & ! current dipole dislocation densities (screw and edge dipoles)
dLower, & ! minimum stable dipole distance for edges and screws
dUpper, & ! current maximum stable dipole distance for edges and screws
dUpperOld, & ! old maximum stable dipole distance for edges and screws
deltaDUpper ! change in maximum stable dipole distance for edges and screws
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
#ifndef _OPENMP
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
if (iand(debug_level(debug_constitutive),debug_levelBasic) /= 0_pInt &
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == g)&
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
.or. .not. iand(debug_level(debug_constitutive),debug_levelSelective) /= 0_pInt)) then
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
write(6,*)
small correction in debug output
2012-05-30 13:11:22 +05:30
write(6,'(a,i8,1x,i2,1x,i1)') '<< CONST >> nonlocal_deltaState at el ip g ',el,ip,g
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
write(6,*)
endif
#endif
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
myInstance = phase_plasticityInstance(material_phase(g,ip,el))
myStructure = constitutive_nonlocal_structure(myInstance)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
ns = totalNslip(myInstance)
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
!*** shortcut to state variables
use named state variable indices in "deltaState"
2013-05-24 03:16:15 +05:30
forall (s = 1_pInt:ns)
rhoSgl(s,1) = max(state(g,ip,el)%p(iRhoEPU(s,myInstance)), 0.0_pReal) ! ensure positive single mobile densities
rhoSgl(s,2) = max(state(g,ip,el)%p(iRhoENU(s,myInstance)), 0.0_pReal) ! ensure positive single mobile densities
rhoSgl(s,3) = max(state(g,ip,el)%p(iRhoSPU(s,myInstance)), 0.0_pReal) ! ensure positive single mobile densities
rhoSgl(s,4) = max(state(g,ip,el)%p(iRhoSNU(s,myInstance)), 0.0_pReal) ! ensure positive single mobile densities
endforall
rhoSgl(1:ns,5) = state(g,ip,el)%p(iRhoEPB(1:ns,myInstance))
rhoSgl(1:ns,6) = state(g,ip,el)%p(iRhoENB(1:ns,myInstance))
rhoSgl(1:ns,7) = state(g,ip,el)%p(iRhoSPB(1:ns,myInstance))
rhoSgl(1:ns,8) = state(g,ip,el)%p(iRhoSNB(1:ns,myInstance))
forall (s = 1_pInt:ns)
rhoDip(s,1) = max(state(g,ip,el)%p(iRhoED(s,myInstance)), 0.0_pReal) ! ensure positive dipole densities
rhoDip(s,2) = max(state(g,ip,el)%p(iRhoSD(s,myInstance)), 0.0_pReal) ! ensure positive dipole densities
endforall
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
where (abs(rhoSgl) * mesh_ipVolume(ip,el) ** 0.667_pReal < significantN(myInstance) &
use named state variable indices in "deltaState"
2013-05-24 03:16:15 +05:30
.or. abs(rhoSgl) < significantRho(myInstance)) &
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
rhoSgl = 0.0_pReal
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
where (abs(rhoDip) * mesh_ipVolume(ip,el) ** 0.667_pReal < significantN(myInstance) &
use named state variable indices in "deltaState"
2013-05-24 03:16:15 +05:30
.or. abs(rhoDip) < significantRho(myInstance)) &
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
rhoDip = 0.0_pReal
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
use named state variable indices in "deltaState"
2013-05-24 03:16:15 +05:30
tauBack = state(g,ip,el)%p(iTauB(1:ns,myInstance))
v(1:ns,1) = state(g,ip,el)%p(iVEP(1:ns,myInstance))
v(1:ns,2) = state(g,ip,el)%p(iVEN(1:ns,myInstance))
v(1:ns,3) = state(g,ip,el)%p(iVSP(1:ns,myInstance))
v(1:ns,4) = state(g,ip,el)%p(iVSN(1:ns,myInstance))
dUpperOld(1:ns,1) = state(g,ip,el)%p(iDE(1:ns,myInstance))
dUpperOld(1:ns,2) = state(g,ip,el)%p(iDS(1:ns,myInstance))
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
!****************************************************************************
!*** dislocation remobilization (bauschinger effect)
deltaRhoRemobilization = 0.0_pReal
do t = 1_pInt,4_pInt
do s = 1_pInt,ns
if (rhoSgl(s,t+4_pInt) * v(s,t) < 0.0_pReal) then
deltaRhoRemobilization(s,t) = abs(rhoSgl(s,t+4_pInt))
rhoSgl(s,t) = rhoSgl(s,t) + abs(rhoSgl(s,t+4_pInt))
deltaRhoRemobilization(s,t+4_pInt) = - rhoSgl(s,t+4_pInt)
rhoSgl(s,t+4_pInt) = 0.0_pReal
endif
enddo
enddo
!****************************************************************************
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
!*** calculate dipole formation and dissociation by stress change
!*** calculate limits for stable dipole height
do s = 1_pInt,ns
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sLattice = slipSystemLattice(s,myInstance)
added code structure for non-schmid mechanics. work in progress…
2013-01-22 04:41:16 +05:30
tau(s) = math_mul6x6(Tstar_v, lattice_Sslip_v(1:6,1,sLattice,myStructure)) + tauBack(s)
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
if (abs(tau(s)) < 1.0e-15_pReal) tau(s) = 1.0e-15_pReal
enddo
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
dLower = minDipoleHeight(1:ns,1:2,myInstance)
dUpper(1:ns,1) = mu(myInstance) * burgers(1:ns,myInstance) &
/ (8.0_pReal * pi * (1.0_pReal - nu(myInstance)) * abs(tau))
dUpper(1:ns,2) = mu(myInstance) * burgers(1:ns,myInstance) / (4.0_pReal * pi * abs(tau))
corrected formulation of maximum height of screw dislocation dipoles
2012-07-24 20:20:11 +05:30
forall (c = 1_pInt:2_pInt) &
corrected limit for upper stable dipole distance; fixed a bug in the postResults where division by zero might happen
2012-10-29 18:32:01 +05:30
dUpper(1:ns,c) = min(1.0_pReal / sqrt(rhoSgl(1:ns,2*c-1) + rhoSgl(1:ns,2*c) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
+ abs(rhoSgl(1:ns,2*c+3)) + abs(rhoSgl(1:ns,2*c+4)) + rhoDip(1:ns,c)), &
corrected limit for upper stable dipole distance; fixed a bug in the postResults where division by zero might happen
2012-10-29 18:32:01 +05:30
dUpper(1:ns,c))
allow minimum stable dipole height equal to zero; ensure that maximum stable dipole height is not smaller than the minimum value
2012-05-20 19:27:35 +05:30
dUpper = max(dUpper,dLower)
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
deltaDUpper = dUpper - dUpperOld
annihilate only part of the screw dipoles (specified by minimumDipoleHeight), not all; moved annihilation of screws from deltaState back to dotState
2012-11-28 17:39:48 +05:30
!*** dissociation by stress increase
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
deltaRhoDipole2SingleStress = 0.0_pReal
annihilate only part of the screw dipoles (specified by minimumDipoleHeight), not all; moved annihilation of screws from deltaState back to dotState
2012-11-28 17:39:48 +05:30
forall (c=1_pInt:2_pInt, s=1_pInt:ns, deltaDUpper(s,c) < 0.0_pReal) &
deltaRhoDipole2SingleStress(s,8_pInt+c) = rhoDip(s,c) * deltaDUpper(s,c) / (dUpperOld(s,c) - dLower(s,c))
forall (t=1_pInt:4_pInt) &
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
deltaRhoDipole2SingleStress(1_pInt:ns,t) = -0.5_pReal * deltaRhoDipole2SingleStress(1_pInt:ns,(t-1_pInt)/2_pInt+9_pInt)
annihilate only part of the screw dipoles (specified by minimumDipoleHeight), not all; moved annihilation of screws from deltaState back to dotState
2012-11-28 17:39:48 +05:30
no dipole formation by stress decrease; dipole dissociation rate proportional to "deltaDipoleLimit / oldDipoleLimit", not "deltaDipoleLimit / newDipoleLimit"
2012-07-24 12:27:37 +05:30
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
!*** store new maximum dipole height in state
use named state variable indices in "deltaState"
2013-05-24 03:16:15 +05:30
state(g,ip,el)%p(iDE(1:ns,myInstance)) = dUpper(1:ns,1)
state(g,ip,el)%p(iDS(1:ns,myInstance)) = dUpper(1:ns,2)
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
!****************************************************************************
introduced annihilation of screw dipoles; so far no thermal activation, but immediate annihilation; each annihilated screw dipole leaves behind an edge jog of same length as the dipole height on the collinear system
2012-08-14 17:56:20 +05:30
!*** assign the changes in the dislocation densities to deltaState
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
deltaRho = 0.0_pReal
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
deltaRho = deltaRhoRemobilization &
annihilate only part of the screw dipoles (specified by minimumDipoleHeight), not all; moved annihilation of screws from deltaState back to dotState
2012-11-28 17:39:48 +05:30
+ deltaRhoDipole2SingleStress
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
deltaState%p = reshape(deltaRho,(/10_pInt*ns/))
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
#ifndef _OPENMP
In crystallite: for state integrators 2 and 3 the deltaState always has to be done at the beginning of the integration, not at the end; otherwise the stiffness calculation can be completely wrong, when the state is changed after the stress integration for the central solution. Also changed some debugging levels in crystallite and mainly in constitutive_nonlocal
2012-10-22 13:29:35 +05:30
if (iand(debug_level(debug_constitutive),debug_levelExtensive) /= 0_pInt &
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == g)&
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
.or. .not. iand(debug_level(debug_constitutive),debug_levelSelective) /= 0_pInt )) then
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
write(6,'(a,/,8(12x,12(e12.5,1x),/))') '<< CONST >> dislocation remobilization', deltaRhoRemobilization(1:ns,1:8)
added some debug output
2012-05-30 13:35:36 +05:30
write(6,'(a,/,10(12x,12(e12.5,1x),/))') '<< CONST >> dipole dissociation by stress increase', deltaRhoDipole2SingleStress
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
write(6,*)
endif
#endif
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
dipole dissociation and formation by change in stress as new mechanism in deltaState; had to add dipole height to state variables, which is however updated by deltaState instead of microstructure; alternatively microstructure had to know the current stress state
2012-05-18 20:05:52 +05:30
endsubroutine
all constitutive modules now contain a new function "deltaState", which in the future allows to have an instantaneous and incremental change of the state (additional to the rate based evolution with dotState).
2012-05-16 20:13:26 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*********************************************************************
!* rate of change of microstructure *
!*********************************************************************
removed some unused variables
2012-11-30 00:14:00 +05:30
function constitutive_nonlocal_dotState(Tstar_v, Fe, Fp, Temperature, state, state0, timestep, subfrac, g,ip,el)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
polishing
2013-05-23 17:55:56 +05:30
use prec, only: DAMASK_NaN
Introduced an alternative cutback scheme for the nonlocal model, that allows to keep the results of most of the ips that immediately converged, and only do cutbacks in regions where some ips are in big trouble. Only works for nonlocal model and state integrator 2.
2012-11-28 00:06:55 +05:30
use numerics, only: numerics_integrationMode, &
numerics_timeSyncing
- flux calculation is now also compatible to neighborhood of local constitution - flux density interpolation now depends on the position of the interface between ttwo neighboring material points - simplified flux calculation scheme - introduced sanity check for dislocation velocity to ensure v*dt< cellsize
2010-03-04 22:44:47 +05:30
use IO, only: IO_error
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
use debug, only: debug_level, &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
debug_constitutive, &
debug_levelBasic, &
modified multiplication term in order to mimic a discrete distribution of dislocation sources; namely, at high resolution, when there is (on average) less than one source per integration point, a random generator decides whether there is a source in the finite volume or not; if there is one, then the multiplication turns independent of the density and depends only on the velocity/volume times a scalar factor; recovers density*velocity/lambda for large finite volumes
2012-10-19 17:10:17 +05:30
debug_levelExtensive, &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
debug_levelSelective, &
* checked and corrected parallelization of code. compiles now successfully, but simply aborts computation with first parallel directive without any comment :-((( * also put a call to constitutive_microstructure at the start of each crystallite_integration subroutine like it was before. need that for nonlocal model in case of crystallite cutback
2010-11-03 22:52:48 +05:30
debug_g, &
debug_i, &
debugging output is now controlled by the "verbosity" parameter in the debug.config ranging from 0 (=almost no output) to 8 (=very detailed output) 0 : only version infos and all from "hypela2"/"umat" 1 : basic outputs from "CPFEM.f90", basic output from initialization routines, debug_info 2 : extensive outputs from "CPFEM.f90", extensive output from initialization routines 3 : basic outputs from "homogenization.f90" 4 : extensive outputs from "homogenization.f90" 5 : basic outputs from "crystallite.f90" 6 : extensive outputs from "crystallite.f90" 7 : basic outputs from the constitutive files 8 : extensive outputs from the constitutive files If verbosity is equal to zero, all counters in debug are not set during calculation (e.g. debug_StressLoopDistribution or debug_cumDotStateTicks). This might speed up parallel calculation, because all these need critical statements which extremely slow down parallel computation.
2011-03-21 16:01:17 +05:30
debug_e
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
use math, only: math_norm3, &
math_mul6x6, &
math_mul3x3, &
math_mul33x3, &
constitutive_nonlocal - take orientation gradients into account when calculating dislocation stress and dislocation fluxes - hard coded value for nu - changed kinetics (parameter G0 is currently defined as a parameter, needs to be read from material.config) - added some output statements constitutive: - some functions and subroutines needed additional input variables for passing to constitutive_nonlocal crystallite: - some functions and subroutines needed additional input variables for passing to constitutive - call microstructure with current temperature, Fp, Fe, not "sub0" values - show number of IPs, that are "onTrack" instead of those not "onTrack" - calculate Fe at beginning of substep, since we need it for state preguess
2009-10-07 21:01:52 +05:30
math_mul33x33, &
added new math function "math_deviatoric33" which returns the deviatoric part of a 3x3 tensor renamed some math functions, so that we have a universal naming scheme: for matrix multiplications use an "x" (e.g. math_mul33x3); don't use the "x" to describe the shape of the tensor that the function is applied to (e.g. math_invert33 instead of math_invert3x3)
2012-01-26 19:20:00 +05:30
math_inv33, &
math_det33, &
math_transpose33, &
improved accuracy of "accumulated shear": now added as a state variable, which facilitates integration (dotstate=shearrate); needs absolute tolerance value for state integration in material.config
2013-04-04 19:07:14 +05:30
pi
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
use mesh, only: mesh_NcpElems, &
mesh_maxNips, &
mesh_element, &
mesh_ipNeighborhood, &
mesh_ipVolume, &
mesh_ipArea, &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
mesh_ipAreaNormal, &
FE_NipNeighbors, &
neighborhood relations were screwed up for some elementtypes as of version 2318, might have caused segmentation faults with nonlocal model, now fixed
2013-04-22 19:05:35 +05:30
FE_geomtype, &
FE_celltype
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
use material, only: homogenization_maxNgrains, &
material_phase, &
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
phase_plasticityInstance, &
phase_localPlasticity, &
phase_plasticity
mixed things up, did not want to commit constitutive_nonlocal.f90 but rather DAMASK_spectral.f90!!!! did a lot of polishing on constitutive_nonlocal.f90 (mostly checked in at last commit), for changes on DAMASK_spectral.f90 see previous message
2012-02-23 22:50:57 +05:30
use lattice, only: lattice_Sslip_v, &
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
lattice_sd, &
mixed things up, did not want to commit constitutive_nonlocal.f90 but rather DAMASK_spectral.f90!!!! did a lot of polishing on constitutive_nonlocal.f90 (mostly checked in at last commit), for changes on DAMASK_spectral.f90 see previous message
2012-02-23 22:50:57 +05:30
lattice_st
* corrected compatibility for screws (always positive) * detection of grain boundary in constitutive_nonlocal_microstructure with the help of transmissivity * enforce positive densities in constitutive_nonlocal_microstructure (needed because dotState does not create cutbacks for negative densities anymore) * reset single mobile densities below certain threshold to zero (also done in constitutive_nonlocal_microstructure) * constitutive_nonlocal_kinetics only gets local state variable as input, no need for the entire array here * dv_dtau is always positive * multiplication is only active when there is already some initial density of the respective type
2010-10-26 19:12:18 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
implicit none
!*** input variables
line break in lines > 132
2013-01-22 16:36:39 +05:30
integer(pInt), intent(in) :: g, & !< current grain number
ip, & !< current integration point
el !< current element number
real(pReal), intent(in) :: Temperature, & !< temperature
timestep !< substepped crystallite time increment
real(pReal), dimension(6), intent(in) :: Tstar_v !< current 2nd Piola-Kirchhoff stress in Mandel notation
Introduced an alternative cutback scheme for the nonlocal model, that allows to keep the results of most of the ips that immediately converged, and only do cutbacks in regions where some ips are in big trouble. Only works for nonlocal model and state integrator 2.
2012-11-28 00:06:55 +05:30
real(pReal), dimension(homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), intent(in) :: &
line break in lines > 132
2013-01-22 16:36:39 +05:30
subfrac !< fraction of timestep at the beginning of the substepped crystallite time increment
constitutive_nonlocal - take orientation gradients into account when calculating dislocation stress and dislocation fluxes - hard coded value for nu - changed kinetics (parameter G0 is currently defined as a parameter, needs to be read from material.config) - added some output statements constitutive: - some functions and subroutines needed additional input variables for passing to constitutive_nonlocal crystallite: - some functions and subroutines needed additional input variables for passing to constitutive - call microstructure with current temperature, Fp, Fe, not "sub0" values - show number of IPs, that are "onTrack" instead of those not "onTrack" - calculate Fe at beginning of substep, since we need it for state preguess
2009-10-07 21:01:52 +05:30
real(pReal), dimension(3,3,homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), intent(in) :: &
line break in lines > 132
2013-01-22 16:36:39 +05:30
Fe, & !< elastic deformation gradient
Fp !< plastic deformation gradient
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
type(p_vec), dimension(homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), intent(in) :: &
line break in lines > 132
2013-01-22 16:36:39 +05:30
state, & !< current microstructural state
state0 !< microstructural state at beginning of crystallite increment
Introduced an alternative cutback scheme for the nonlocal model, that allows to keep the results of most of the ips that immediately converged, and only do cutbacks in regions where some ips are in big trouble. Only works for nonlocal model and state integrator 2.
2012-11-28 00:06:55 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*** input/output variables
!*** output variables
dotState does not have to be reset to zero. this is a remnant from older versions when the dotState for the nonlocal model used to be updated by the neighboring integration point
2012-05-16 21:05:14 +05:30
real(pReal), dimension(constitutive_nonlocal_sizeDotState(phase_plasticityInstance(material_phase(g,ip,el)))) :: &
line break in lines > 132
2013-01-22 16:36:39 +05:30
constitutive_nonlocal_dotState !< evolution of state variables / microstructure
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*** local variables
line break in lines > 132
2013-01-22 16:36:39 +05:30
integer(pInt) myInstance, & !< current instance of this plasticity
myStructure, & !< current lattice structure
ns, & !< short notation for the total number of active slip systems
c, & !< character of dislocation
n, & !< index of my current neighbor
neighboring_el, & !< element number of my neighbor
neighboring_ip, & !< integration point of my neighbor
neighboring_n, & !< neighbor index pointing to me when looking from my neighbor
opposite_neighbor, & !< index of my opposite neighbor
opposite_ip, & !< ip of my opposite neighbor
opposite_el, & !< element index of my opposite neighbor
opposite_n, & !< neighbor index pointing to me when looking from my opposite neighbor
t, & !< type of dislocation
topp, & !< type of dislocation with opposite sign to t
s, & !< index of my current slip system
sLattice, & !< index of my current slip system according to lattice order
Dead dislocations are treated the same as mobile dislocations for the flux part though they do not contribute to slip. By that the pileup of dead dislocations is able to diminish with time (kinetics are the same as for the glide part; this is not correct but gives valuable results). Also remobilization of dead dislocations at neighbor is taken into account for flux.
2012-03-14 21:02:50 +05:30
deads
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),10) :: &
line break in lines > 132
2013-01-22 16:36:39 +05:30
rhoDot, & !< density evolution
rhoDotMultiplication, & !< density evolution by multiplication
rhoDotFlux, & !< density evolution by flux
rhoDotSingle2DipoleGlide, & !< density evolution by dipole formation (by glide)
rhoDotAthermalAnnihilation, & !< density evolution by athermal annihilation
rhoDotThermalAnnihilation !< density evolution by thermal annihilation
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),8) :: &
line break in lines > 132
2013-01-22 16:36:39 +05:30
rhoSgl, & !< current single dislocation densities (positive/negative screw and edge without dipoles)
slightly changed the multiplication term in the nonlocal model for the starvation case
2012-11-30 00:20:25 +05:30
rhoSglOriginal, &
reverted changes made in rev2384
2013-05-17 18:24:47 +05:30
neighboring_rhoSgl, & !< current single dislocation densities of neighboring ip (positive/negative screw and edge without dipoles)
line break in lines > 132
2013-01-22 16:36:39 +05:30
rhoSgl0, & !< single dislocation densities at start of cryst inc (positive/negative screw and edge without dipoles)
better not allow for deads to contribute to flux, since not clear at which interface they got stuck
2013-05-08 00:17:17 +05:30
rhoSglMe !< single dislocation densities of central ip (positive/negative screw and edge without dipoles)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),4) :: &
line break in lines > 132
2013-01-22 16:36:39 +05:30
v, & !< current dislocation glide velocity
v0, & !< dislocation glide velocity at start of cryst inc
vMe, & !< dislocation glide velocity of central ip
neighboring_v, & !< dislocation glide velocity of enighboring ip
gdot !< shear rates
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el)))) :: &
line break in lines > 132
2013-01-22 16:36:39 +05:30
rhoForest, & !< forest dislocation density
tauThreshold, & !< threshold shear stress
tau, & !< current resolved shear stress
tauBack, & !< current back stress from pileups on same slip system
vClimb, & !< climb velocity of edge dipoles
slightly changed the multiplication term in the nonlocal model for the starvation case
2012-11-30 00:20:25 +05:30
nSources
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),2) :: &
line break in lines > 132
2013-01-22 16:36:39 +05:30
rhoDip, & !< current dipole dislocation densities (screw and edge dipoles)
use non-corrected densities for rate check in dotState
2012-12-06 22:44:35 +05:30
rhoDipOriginal, &
line break in lines > 132
2013-01-22 16:36:39 +05:30
dLower, & !< minimum stable dipole distance for edges and screws
dUpper !< current maximum stable dipole distance for edges and screws
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(3,totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),4) :: &
line break in lines > 132
2013-01-22 16:36:39 +05:30
m !< direction of dislocation motion
real(pReal), dimension(3,3) :: my_F, & !< my total deformation gradient
neighboring_F, & !< total deformation gradient of my neighbor
my_Fe, & !< my elastic deformation gradient
neighboring_Fe, & !< elastic deformation gradient of my neighbor
Favg !< average total deformation gradient of me and my neighbor
real(pReal), dimension(3) :: normal_neighbor2me, & !< interface normal pointing from my neighbor to me in neighbor's lattice configuration
normal_neighbor2me_defConf, & !< interface normal pointing from my neighbor to me in shared deformed configuration
normal_me2neighbor, & !< interface normal pointing from me to my neighbor in my lattice configuration
normal_me2neighbor_defConf !< interface normal pointing from me to my neighbor in shared deformed configuration
real(pReal) area, & !< area of the current interface
transmissivity, & !< overall transmissivity of dislocation flux to neighboring material point
lineLength, & !< dislocation line length leaving the current interface
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
selfDiffusion, & !< self diffusion
slightly changed the multiplication term in the nonlocal model for the starvation case
2012-11-30 00:20:25 +05:30
rnd, &
meshlength
* added a new material parameter "surfaceTransmissivity" (default value 1.0) which allows to change the transmissivity of the material surface between 0 and 1 * now complaining when encountering an unknown nonlocal parameter in material.config * use same error ID for all material parameters out of bounds * symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity) * switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
2011-02-16 22:05:38 +05:30
logical considerEnteringFlux, &
slightly changed the multiplication term in the nonlocal model for the starvation case
2012-11-30 00:20:25 +05:30
considerLeavingFlux
- flux calculation is now also compatible to neighborhood of local constitution - flux density interpolation now depends on the position of the interface between ttwo neighboring material points - simplified flux calculation scheme - introduced sanity check for dislocation velocity to ensure v*dt< cellsize
2010-03-04 22:44:47 +05:30
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
#ifndef _OPENMP
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
if (iand(debug_level(debug_constitutive),debug_levelBasic) /= 0_pInt &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == g)&
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
.or. .not. iand(debug_level(debug_constitutive),debug_levelSelective) /= 0_pInt)) then
debugging output is now controlled by the "verbosity" parameter in the debug.config ranging from 0 (=almost no output) to 8 (=very detailed output) 0 : only version infos and all from "hypela2"/"umat" 1 : basic outputs from "CPFEM.f90", basic output from initialization routines, debug_info 2 : extensive outputs from "CPFEM.f90", extensive output from initialization routines 3 : basic outputs from "homogenization.f90" 4 : extensive outputs from "homogenization.f90" 5 : basic outputs from "crystallite.f90" 6 : extensive outputs from "crystallite.f90" 7 : basic outputs from the constitutive files 8 : extensive outputs from the constitutive files If verbosity is equal to zero, all counters in debug are not set during calculation (e.g. debug_StressLoopDistribution or debug_cumDotStateTicks). This might speed up parallel calculation, because all these need critical statements which extremely slow down parallel computation.
2011-03-21 16:01:17 +05:30
write(6,*)
some f2008 standard related correction i didn't check in last time
2012-02-02 01:50:05 +05:30
write(6,'(a,i8,1x,i2,1x,i1)') '<< CONST >> nonlocal_dotState at el ip g ',el,ip,g
- flux calculation is now also compatible to neighborhood of local constitution - flux density interpolation now depends on the position of the interface between ttwo neighboring material points - simplified flux calculation scheme - introduced sanity check for dislocation velocity to ensure v*dt< cellsize
2010-03-04 22:44:47 +05:30
write(6,*)
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
endif
#endif
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
myInstance = phase_plasticityInstance(material_phase(g,ip,el))
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
myStructure = constitutive_nonlocal_structure(myInstance)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
ns = totalNslip(myInstance)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
tau = 0.0_pReal
added dislocation multiplication to dotState in constitutive_nonlocal.f90 cleaned up debugging output statements to *.out file
2009-08-12 16:52:02 +05:30
gdot = 0.0_pReal
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
openmp parallelization working again (at least for j2 and nonlocal constitutive model). In order to keep it like that, please follow these simple rules: DON'T use implicit array subscripts: example: real, dimension(3,3) :: A,B A(:,2) = B(:,1) <--- DON'T USE A(1:3,2) = B(1:3,1) <--- BETTER USE In many cases the use of explicit array subscripts is inevitable for parallelization. Additionally, it is an easy means to prevent memory leaks. Enclose all write statements with the following: !$OMP CRITICAL (write2out) <your write statement> !$OMP END CRITICAL (write2out) Whenever you change something in the code and are not sure if it affects parallelization and leads to nonconforming behavior, please ask me and/or Franz to check this.
2011-03-17 16:16:17 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
!*** shortcut to state variables
polishing, adding _pInt etc. where applicable post_results now handels zero increment different (like FEM, it is always there even if it is not counted)
2012-02-13 19:48:07 +05:30
forall (s = 1_pInt:ns, t = 1_pInt:4_pInt) &
rhoSgl(s,t) = max(state(g,ip,el)%p((t-1_pInt)*ns+s), 0.0_pReal)
forall (s = 1_pInt:ns, t = 5_pInt:8_pInt) &
rhoSgl(s,t) = state(g,ip,el)%p((t-1_pInt)*ns+s)
forall (s = 1_pInt:ns, c = 1_pInt:2_pInt) &
rhoDip(s,c) = max(state(g,ip,el)%p((7_pInt+c)*ns+s), 0.0_pReal)
improved accuracy of "accumulated shear": now added as a state variable, which facilitates integration (dotstate=shearrate); needs absolute tolerance value for state integration in material.config
2013-04-04 19:07:14 +05:30
rhoForest = state(g,ip,el)%p(11_pInt*ns+1:12_pInt*ns)
tauThreshold = state(g,ip,el)%p(12_pInt*ns+1_pInt:13_pInt*ns)
tauBack = state(g,ip,el)%p(13_pInt*ns+1:14_pInt*ns)
polishing, adding _pInt etc. where applicable post_results now handels zero increment different (like FEM, it is always there even if it is not counted)
2012-02-13 19:48:07 +05:30
forall (t = 1_pInt:4_pInt) &
improved accuracy of "accumulated shear": now added as a state variable, which facilitates integration (dotstate=shearrate); needs absolute tolerance value for state integration in material.config
2013-04-04 19:07:14 +05:30
v(1_pInt:ns,t) = state(g,ip,el)%p((13_pInt+t)*ns+1_pInt:(14_pInt+t)*ns)
slightly changed the multiplication term in the nonlocal model for the starvation case
2012-11-30 00:20:25 +05:30
rhoSglOriginal = rhoSgl
use non-corrected densities for rate check in dotState
2012-12-06 22:44:35 +05:30
rhoDipOriginal = rhoDip
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
where (abs(rhoSgl) * mesh_ipVolume(ip,el) ** 0.667_pReal < significantN(myInstance) &
.or. abs(rhoSgl) < significantRho(myInstance)) &
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
rhoSgl = 0.0_pReal
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
where (abs(rhoDip) * mesh_ipVolume(ip,el) ** 0.667_pReal < significantN(myInstance) &
.or. abs(rhoDip) < significantRho(myInstance)) &
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
rhoDip = 0.0_pReal
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
openmp parallelization working again (at least for j2 and nonlocal constitutive model). In order to keep it like that, please follow these simple rules: DON'T use implicit array subscripts: example: real, dimension(3,3) :: A,B A(:,2) = B(:,1) <--- DON'T USE A(1:3,2) = B(1:3,1) <--- BETTER USE In many cases the use of explicit array subscripts is inevitable for parallelization. Additionally, it is an easy means to prevent memory leaks. Enclose all write statements with the following: !$OMP CRITICAL (write2out) <your write statement> !$OMP END CRITICAL (write2out) Whenever you change something in the code and are not sure if it affects parallelization and leads to nonconforming behavior, please ask me and/or Franz to check this.
2011-03-17 16:16:17 +05:30
- flux calculation is now also compatible to neighborhood of local constitution - flux density interpolation now depends on the position of the interface between ttwo neighboring material points - simplified flux calculation scheme - introduced sanity check for dislocation velocity to ensure v*dt< cellsize
2010-03-04 22:44:47 +05:30
!*** sanity check for timestep
if (timestep <= 0.0_pReal) then ! if illegal timestep...
dotState does not have to be reset to zero. this is a remnant from older versions when the dotState for the nonlocal model used to be updated by the neighboring integration point
2012-05-16 21:05:14 +05:30
constitutive_nonlocal_dotState = 0.0_pReal ! ...return without doing anything (-> zero dotState)
- flux calculation is now also compatible to neighborhood of local constitution - flux density interpolation now depends on the position of the interface between ttwo neighboring material points - simplified flux calculation scheme - introduced sanity check for dislocation velocity to ensure v*dt< cellsize
2010-03-04 22:44:47 +05:30
return
endif
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
openmp parallelization working again (at least for j2 and nonlocal constitutive model). In order to keep it like that, please follow these simple rules: DON'T use implicit array subscripts: example: real, dimension(3,3) :: A,B A(:,2) = B(:,1) <--- DON'T USE A(1:3,2) = B(1:3,1) <--- BETTER USE In many cases the use of explicit array subscripts is inevitable for parallelization. Additionally, it is an easy means to prevent memory leaks. Enclose all write statements with the following: !$OMP CRITICAL (write2out) <your write statement> !$OMP END CRITICAL (write2out) Whenever you change something in the code and are not sure if it affects parallelization and leads to nonconforming behavior, please ask me and/or Franz to check this.
2011-03-17 16:16:17 +05:30
added dislocation multiplication to dotState in constitutive_nonlocal.f90 cleaned up debugging output statements to *.out file
2009-08-12 16:52:02 +05:30
!****************************************************************************
!*** Calculate shear rate
polishing, adding _pInt etc. where applicable post_results now handels zero increment different (like FEM, it is always there even if it is not counted)
2012-02-13 19:48:07 +05:30
forall (t = 1_pInt:4_pInt) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
gdot(1_pInt:ns,t) = rhoSgl(1_pInt:ns,t) * burgers(1:ns,myInstance) * v(1:ns,t)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
#ifndef _OPENMP
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
if (iand(debug_level(debug_constitutive),debug_levelBasic) /= 0_pInt &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == g)&
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
.or. .not. iand(debug_level(debug_constitutive),debug_levelSelective) /= 0_pInt )) then
some f2008 standard related correction i didn't check in last time
2012-02-02 01:50:05 +05:30
write(6,'(a,/,10(12x,12(e12.5,1x),/))') '<< CONST >> rho / 1/m^2', rhoSgl, rhoDip
write(6,'(a,/,4(12x,12(e12.5,1x),/))') '<< CONST >> gdot / 1/s',gdot
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
endif
#endif
* corrected compatibility for screws (always positive) * detection of grain boundary in constitutive_nonlocal_microstructure with the help of transmissivity * enforce positive densities in constitutive_nonlocal_microstructure (needed because dotState does not create cutbacks for negative densities anymore) * reset single mobile densities below certain threshold to zero (also done in constitutive_nonlocal_microstructure) * constitutive_nonlocal_kinetics only gets local state variable as input, no need for the entire array here * dv_dtau is always positive * multiplication is only active when there is already some initial density of the respective type
2010-10-26 19:12:18 +05:30
openmp parallelization working again (at least for j2 and nonlocal constitutive model). In order to keep it like that, please follow these simple rules: DON'T use implicit array subscripts: example: real, dimension(3,3) :: A,B A(:,2) = B(:,1) <--- DON'T USE A(1:3,2) = B(1:3,1) <--- BETTER USE In many cases the use of explicit array subscripts is inevitable for parallelization. Additionally, it is an easy means to prevent memory leaks. Enclose all write statements with the following: !$OMP CRITICAL (write2out) <your write statement> !$OMP END CRITICAL (write2out) Whenever you change something in the code and are not sure if it affects parallelization and leads to nonconforming behavior, please ask me and/or Franz to check this.
2011-03-17 16:16:17 +05:30
added dislocation multiplication to dotState in constitutive_nonlocal.f90 cleaned up debugging output statements to *.out file
2009-08-12 16:52:02 +05:30
!****************************************************************************
* removed calculations for dipole formation/dissociation by stress change, since it is not used anyways; also removed associated constitutive outputs from material.config * removed input variables in constitutive_collectDotState and constitutive_postResults that are not needed anymore (because of recent changes in constitutive_nonlocal)
2011-02-25 15:23:20 +05:30
!*** calculate limits for stable dipole height
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do s = 1_pInt,ns ! loop over slip systems
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sLattice = slipSystemLattice(s,myInstance)
added code structure for non-schmid mechanics. work in progress…
2013-01-22 04:41:16 +05:30
tau(s) = math_mul6x6(Tstar_v, lattice_Sslip_v(1:6,1,sLattice,myStructure)) + tauBack(s)
fixed potential division by zero error discovered by our American friends
2012-05-08 12:46:00 +05:30
if (abs(tau(s)) < 1.0e-15_pReal) tau(s) = 1.0e-15_pReal
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
enddo
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
dLower = minDipoleHeight(1:ns,1:2,myInstance)
dUpper(1:ns,1) = mu(myInstance) * burgers(1:ns,myInstance) &
/ (8.0_pReal * pi * (1.0_pReal - nu(myInstance)) * abs(tau))
dUpper(1:ns,2) = mu(myInstance) * burgers(1:ns,myInstance) &
corrected formulation of maximum height of screw dislocation dipoles
2012-07-24 20:20:11 +05:30
/ (4.0_pReal * pi * abs(tau))
forall (c = 1_pInt:2_pInt) &
corrected limit for upper stable dipole distance; fixed a bug in the postResults where division by zero might happen
2012-10-29 18:32:01 +05:30
dUpper(1:ns,c) = min(1.0_pReal / sqrt(rhoSgl(1:ns,2*c-1) + rhoSgl(1:ns,2*c) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
+ abs(rhoSgl(1:ns,2*c+3)) + abs(rhoSgl(1:ns,2*c+4)) + rhoDip(1:ns,c)), &
corrected limit for upper stable dipole distance; fixed a bug in the postResults where division by zero might happen
2012-10-29 18:32:01 +05:30
dUpper(1:ns,c))
allow minimum stable dipole height equal to zero; ensure that maximum stable dipole height is not smaller than the minimum value
2012-05-20 19:27:35 +05:30
dUpper = max(dUpper,dLower)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
added check for CFL condition in constitutive_nonlocal_dotState
2011-04-06 14:37:36 +05:30
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
!****************************************************************************
!*** calculate dislocation multiplication
constitutive_nonlocal - take orientation gradients into account when calculating dislocation stress and dislocation fluxes - hard coded value for nu - changed kinetics (parameter G0 is currently defined as a parameter, needs to be read from material.config) - added some output statements constitutive: - some functions and subroutines needed additional input variables for passing to constitutive_nonlocal crystallite: - some functions and subroutines needed additional input variables for passing to constitutive - call microstructure with current temperature, Fp, Fe, not "sub0" values - show number of IPs, that are "onTrack" instead of those not "onTrack" - calculate Fe at beginning of substep, since we need it for state preguess
2009-10-07 21:01:52 +05:30
* corrected compatibility for screws (always positive) * detection of grain boundary in constitutive_nonlocal_microstructure with the help of transmissivity * enforce positive densities in constitutive_nonlocal_microstructure (needed because dotState does not create cutbacks for negative densities anymore) * reset single mobile densities below certain threshold to zero (also done in constitutive_nonlocal_microstructure) * constitutive_nonlocal_kinetics only gets local state variable as input, no need for the entire array here * dv_dtau is always positive * multiplication is only active when there is already some initial density of the respective type
2010-10-26 19:12:18 +05:30
rhoDotMultiplication = 0.0_pReal
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (probabilisticMultiplication(myInstance)) then
made random multiplication switchable via material.config
2012-12-03 18:29:38 +05:30
meshlength = mesh_ipVolume(ip,el)**0.333_pReal
where(sum(rhoSgl(1:ns,1:4),2) > 0.0_pReal)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
nSources = (sum(rhoSgl(1:ns,1:2),2) * fEdgeMultiplication(myInstance) + sum(rhoSgl(1:ns,3:4),2)) &
/ sum(rhoSgl(1:ns,1:4),2) * meshlength / lambda0(1:ns,myInstance) * sqrt(rhoForest(1:ns))
made random multiplication switchable via material.config
2012-12-03 18:29:38 +05:30
elsewhere
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
nSources = meshlength / lambda0(1:ns,myInstance) * sqrt(rhoForest(1:ns))
made random multiplication switchable via material.config
2012-12-03 18:29:38 +05:30
endwhere
do s = 1_pInt,ns
if (nSources(s) < 1.0_pReal) then
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (sourceProbability(s,g,ip,el) > 1.0_pReal) then
made random multiplication switchable via material.config
2012-12-03 18:29:38 +05:30
call random_number(rnd)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sourceProbability(s,g,ip,el) = rnd
!$OMP FLUSH(sourceProbability)
made random multiplication switchable via material.config
2012-12-03 18:29:38 +05:30
endif
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (sourceProbability(s,g,ip,el) > 1.0_pReal - nSources(s)) then
corrected typo in random multiplication term
2012-12-06 19:34:18 +05:30
rhoDotMultiplication(s,1:4) = sum(rhoSglOriginal(s,1:4) * abs(v(s,1:4))) / meshlength
made random multiplication switchable via material.config
2012-12-03 18:29:38 +05:30
endif
else
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sourceProbability(s,g,ip,el) = 2.0_pReal
made random multiplication switchable via material.config
2012-12-03 18:29:38 +05:30
rhoDotMultiplication(s,1:4) = &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
(sum(abs(gdot(s,1:2))) * fEdgeMultiplication(myInstance) + sum(abs(gdot(s,3:4)))) &
/ burgers(s,myInstance) * sqrt(rhoForest(s)) / lambda0(s,myInstance)
slightly changed the multiplication term in the nonlocal model for the starvation case
2012-11-30 00:20:25 +05:30
endif
made random multiplication switchable via material.config
2012-12-03 18:29:38 +05:30
enddo
modified multiplication term in order to mimic a discrete distribution of dislocation sources; namely, at high resolution, when there is (on average) less than one source per integration point, a random generator decides whether there is a source in the finite volume or not; if there is one, then the multiplication turns independent of the density and depends only on the velocity/volume times a scalar factor; recovers density*velocity/lambda for large finite volumes
2012-10-19 17:10:17 +05:30
#ifndef _OPENMP
if (iand(debug_level(debug_constitutive),debug_levelExtensive) /= 0_pInt &
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == g)&
.or. .not. iand(debug_level(debug_constitutive),debug_levelSelective) /= 0_pInt )) then
slightly changed the multiplication term in the nonlocal model for the starvation case
2012-11-30 00:20:25 +05:30
write(6,'(a,/,4(12x,12(f12.5,1x),/))') '<< CONST >> sources', nSources
modified multiplication term in order to mimic a discrete distribution of dislocation sources; namely, at high resolution, when there is (on average) less than one source per integration point, a random generator decides whether there is a source in the finite volume or not; if there is one, then the multiplication turns independent of the density and depends only on the velocity/volume times a scalar factor; recovers density*velocity/lambda for large finite volumes
2012-10-19 17:10:17 +05:30
write(6,*)
endif
#endif
made random multiplication switchable via material.config
2012-12-03 18:29:38 +05:30
else
rhoDotMultiplication(1:ns,1:4) = spread( &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
(sum(abs(gdot(1:ns,1:2)),2) * fEdgeMultiplication(myInstance) + sum(abs(gdot(1:ns,3:4)),2)) &
* sqrt(rhoForest(1:ns)) / lambda0(1:ns,myInstance) / burgers(1:ns,myInstance), 2, 4)
made random multiplication switchable via material.config
2012-12-03 18:29:38 +05:30
endif
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
openmp parallelization working again (at least for j2 and nonlocal constitutive model). In order to keep it like that, please follow these simple rules: DON'T use implicit array subscripts: example: real, dimension(3,3) :: A,B A(:,2) = B(:,1) <--- DON'T USE A(1:3,2) = B(1:3,1) <--- BETTER USE In many cases the use of explicit array subscripts is inevitable for parallelization. Additionally, it is an easy means to prevent memory leaks. Enclose all write statements with the following: !$OMP CRITICAL (write2out) <your write statement> !$OMP END CRITICAL (write2out) Whenever you change something in the code and are not sure if it affects parallelization and leads to nonconforming behavior, please ask me and/or Franz to check this.
2011-03-17 16:16:17 +05:30
added dislocation multiplication to dotState in constitutive_nonlocal.f90 cleaned up debugging output statements to *.out file
2009-08-12 16:52:02 +05:30
!****************************************************************************
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
!*** calculate dislocation fluxes (only for nonlocal plasticity)
added dislocation multiplication to dotState in constitutive_nonlocal.f90 cleaned up debugging output statements to *.out file
2009-08-12 16:52:02 +05:30
in nonlocal_microstructure: at the model surface the excess density is now extrapolated from opposite neighbor instead of assuming zero density; this results in better modeling of the dislocation stress at the surface restructured nonlocal_dotstate, to be able to easily switch on and off particular effects in the microstructure evolution nonlocal_dotstate now enforces a cutback when single density runs the risk of becoming negative; in the case of a state already below the relevantState dotState is set to zero introduced two new output variables: rho_dot_edge and rho_dot_screw
2010-05-21 14:21:15 +05:30
rhoDotFlux = 0.0_pReal
constitutive_nonlocal - take orientation gradients into account when calculating dislocation stress and dislocation fluxes - hard coded value for nu - changed kinetics (parameter G0 is currently defined as a parameter, needs to be read from material.config) - added some output statements constitutive: - some functions and subroutines needed additional input variables for passing to constitutive_nonlocal crystallite: - some functions and subroutines needed additional input variables for passing to constitutive - call microstructure with current temperature, Fp, Fe, not "sub0" values - show number of IPs, that are "onTrack" instead of those not "onTrack" - calculate Fe at beginning of substep, since we need it for state preguess
2009-10-07 21:01:52 +05:30
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
if (.not. phase_localPlasticity(material_phase(g,ip,el))) then ! only for nonlocal plasticity
* "constitutive_nonlocal_flux" is set to zero in "constitutive_nonlocal_dotState", not in "constitutive_nonlocal_microstructure" * dislocation flux and internal stress calculation now consistent with new definition of slip system lattice according to paper (polarity of screws inverted)
2011-02-23 13:38:06 +05:30
corrected debug output of evolution rates; added possibility to change the safety factor for the CFL flux condition via the material.config file; updated the example material.config due to recent changes in nonlocal constitutive
2012-08-16 14:43:38 +05:30
!*** check CFL (Courant-Friedrichs-Lewy) condition for flux
if (any( abs(gdot) > 0.0_pReal & ! any active slip system ...
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
.and. CFLfactor(myInstance) * abs(v) * timestep &
corrected debug output of evolution rates; added possibility to change the safety factor for the CFL flux condition via the material.config file; updated the example material.config due to recent changes in nonlocal constitutive
2012-08-16 14:43:38 +05:30
> mesh_ipVolume(ip,el) / maxval(mesh_ipArea(:,ip,el)))) then ! ...with velocity above critical value (we use the reference volume and area for simplicity here)
#ifndef _OPENMP
made random multiplication switchable via material.config
2012-12-03 18:29:38 +05:30
if (iand(debug_level(debug_constitutive),debug_levelExtensive) /= 0_pInt) then
corrected debug output of evolution rates; added possibility to change the safety factor for the CFL flux condition via the material.config file; updated the example material.config due to recent changes in nonlocal constitutive
2012-08-16 14:43:38 +05:30
write(6,'(a,i5,a,i2)') '<< CONST >> CFL condition not fullfilled at el ',el,' ip ',ip
minor correction of debugging output
2012-09-05 16:49:46 +05:30
write(6,'(a,e10.3,a,e10.3)') '<< CONST >> velocity is at ', &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
maxval(abs(v), abs(gdot) > 0.0_pReal .and. CFLfactor(myInstance) * abs(v) * timestep &
minor correction of debugging output
2012-09-05 16:49:46 +05:30
> mesh_ipVolume(ip,el) / maxval(mesh_ipArea(:,ip,el))), &
' at a timestep of ',timestep
corrected debug output of evolution rates; added possibility to change the safety factor for the CFL flux condition via the material.config file; updated the example material.config due to recent changes in nonlocal constitutive
2012-08-16 14:43:38 +05:30
write(6,'(a)') '<< CONST >> enforcing cutback !!!'
endif
#endif
constitutive_nonlocal_dotState = DAMASK_NaN ! -> return NaN and, hence, enforce cutback
return
endif
Introduced an alternative cutback scheme for the nonlocal model, that allows to keep the results of most of the ips that immediately converged, and only do cutbacks in regions where some ips are in big trouble. Only works for nonlocal model and state integrator 2.
2012-11-28 00:06:55 +05:30
if (numerics_timeSyncing) then
forall (t = 1_pInt:4_pInt) &
v0(1_pInt:ns,t) = state0(g,ip,el)%p((12_pInt+t)*ns+1_pInt:(13_pInt+t)*ns)
forall (t = 1_pInt:8_pInt) &
rhoSgl0(1_pInt:ns,t) = state0(g,ip,el)%p((t-1_pInt)*ns+1_pInt:t*ns)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
where (abs(rhoSgl0) * mesh_ipVolume(ip,el) ** 0.667_pReal < significantN(myInstance) &
.or. abs(rhoSgl0) < significantRho(myInstance)) &
Introduced an alternative cutback scheme for the nonlocal model, that allows to keep the results of most of the ips that immediately converged, and only do cutbacks in regions where some ips are in big trouble. Only works for nonlocal model and state integrator 2.
2012-11-28 00:06:55 +05:30
rhoSgl0 = 0.0_pReal
endif
corrected debug output of evolution rates; added possibility to change the safety factor for the CFL flux condition via the material.config file; updated the example material.config due to recent changes in nonlocal constitutive
2012-08-16 14:43:38 +05:30
!*** be aware of the definition of lattice_st = lattice_sd x lattice_sn !!!
* "constitutive_nonlocal_flux" is set to zero in "constitutive_nonlocal_dotState", not in "constitutive_nonlocal_microstructure" * dislocation flux and internal stress calculation now consistent with new definition of slip system lattice according to paper (polarity of screws inverted)
2011-02-23 13:38:06 +05:30
!*** opposite sign to our p vector in the (s,p,n) triplet !!!
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
m(1:3,1:ns,1) = lattice_sd(1:3, slipSystemLattice(1:ns,myInstance), myStructure)
m(1:3,1:ns,2) = -lattice_sd(1:3, slipSystemLattice(1:ns,myInstance), myStructure)
m(1:3,1:ns,3) = -lattice_st(1:3, slipSystemLattice(1:ns,myInstance), myStructure)
m(1:3,1:ns,4) = lattice_st(1:3, slipSystemLattice(1:ns,myInstance), myStructure)
* had to correct nonlocal slip system compatibility and flux calculation. last update gave wrong results.
2010-10-15 18:49:26 +05:30
* added a new material parameter "surfaceTransmissivity" (default value 1.0) which allows to change the transmissivity of the material surface between 0 and 1 * now complaining when encountering an unknown nonlocal parameter in material.config * use same error ID for all material parameters out of bounds * symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity) * switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
2011-02-16 22:05:38 +05:30
my_Fe = Fe(1:3,1:3,g,ip,el)
my_F = math_mul33x33(my_Fe, Fp(1:3,1:3,g,ip,el))
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
neighborhood relations were screwed up for some elementtypes as of version 2318, might have caused segmentation faults with nonlocal model, now fixed
2013-04-22 19:05:35 +05:30
do n = 1_pInt,FE_NipNeighbors(FE_celltype(FE_geomtype(mesh_element(2,el)))) ! loop through my neighbors
fluxes are disabled when keyword "/nonlocal/" is omitted in the material.config (this is a way to use the constitutive_nonlocal model as a local dislocation based model)
2011-01-11 20:25:36 +05:30
neighboring_el = mesh_ipNeighborhood(1,n,ip,el)
neighboring_ip = mesh_ipNeighborhood(2,n,ip,el)
added possibility to scale the plastic shearrate by the ratio of mobile to total density; keyword in material.config is "deadZone"
2012-10-29 18:19:28 +05:30
neighboring_n = mesh_ipNeighborhood(3,n,ip,el)
* added a new material parameter "surfaceTransmissivity" (default value 1.0) which allows to change the transmissivity of the material surface between 0 and 1 * now complaining when encountering an unknown nonlocal parameter in material.config * use same error ID for all material parameters out of bounds * symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity) * switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
2011-02-16 22:05:38 +05:30
added possibility to scale the plastic shearrate by the ratio of mobile to total density; keyword in material.config is "deadZone"
2012-10-29 18:19:28 +05:30
opposite_neighbor = n + mod(n,2_pInt) - mod(n+1_pInt,2_pInt)
opposite_el = mesh_ipNeighborhood(1,opposite_neighbor,ip,el)
opposite_ip = mesh_ipNeighborhood(2,opposite_neighbor,ip,el)
opposite_n = mesh_ipNeighborhood(3,opposite_neighbor,ip,el)
fluxes are disabled when keyword "/nonlocal/" is omitted in the material.config (this is a way to use the constitutive_nonlocal model as a local dislocation based model)
2011-01-11 20:25:36 +05:30
added possibility to scale the plastic shearrate by the ratio of mobile to total density; keyword in material.config is "deadZone"
2012-10-29 18:19:28 +05:30
if (neighboring_n > 0_pInt) then ! if neighbor exists, average deformation gradient
* added a new material parameter "surfaceTransmissivity" (default value 1.0) which allows to change the transmissivity of the material surface between 0 and 1 * now complaining when encountering an unknown nonlocal parameter in material.config * use same error ID for all material parameters out of bounds * symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity) * switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
2011-02-16 22:05:38 +05:30
neighboring_Fe = Fe(1:3,1:3,g,neighboring_ip,neighboring_el)
neighboring_F = math_mul33x33(neighboring_Fe, Fp(1:3,1:3,g,neighboring_ip,neighboring_el))
Favg = 0.5_pReal * (my_F + neighboring_F)
fluxes are disabled when keyword "/nonlocal/" is omitted in the material.config (this is a way to use the constitutive_nonlocal model as a local dislocation based model)
2011-01-11 20:25:36 +05:30
else ! if no neighbor, take my value as average
* added a new material parameter "surfaceTransmissivity" (default value 1.0) which allows to change the transmissivity of the material surface between 0 and 1 * now complaining when encountering an unknown nonlocal parameter in material.config * use same error ID for all material parameters out of bounds * symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity) * switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
2011-02-16 22:05:38 +05:30
Favg = my_F
fluxes are disabled when keyword "/nonlocal/" is omitted in the material.config (this is a way to use the constitutive_nonlocal model as a local dislocation based model)
2011-01-11 20:25:36 +05:30
endif
* added a new material parameter "surfaceTransmissivity" (default value 1.0) which allows to change the transmissivity of the material surface between 0 and 1 * now complaining when encountering an unknown nonlocal parameter in material.config * use same error ID for all material parameters out of bounds * symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity) * switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
2011-02-16 22:05:38 +05:30
!* FLUX FROM MY NEIGHBOR TO ME
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
!* This is only considered, if I have a neighbor of nonlocal plasticity (also nonlocal constitutive law with local properties) that is at least a little bit compatible.
* added a new material parameter "surfaceTransmissivity" (default value 1.0) which allows to change the transmissivity of the material surface between 0 and 1 * now complaining when encountering an unknown nonlocal parameter in material.config * use same error ID for all material parameters out of bounds * symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity) * switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
2011-02-16 22:05:38 +05:30
!* If it's not at all compatible, no flux is arriving, because everything is dammed in front of my neighbor's interface.
!* The entering flux from my neighbor will be distributed on my slip systems according to the compatibility
considerEnteringFlux = .false.
Dead dislocations are treated the same as mobile dislocations for the flux part though they do not contribute to slip. By that the pileup of dead dislocations is able to diminish with time (kinetics are the same as for the glide part; this is not correct but gives valuable results). Also remobilization of dead dislocations at neighbor is taken into account for flux.
2012-03-14 21:02:50 +05:30
neighboring_v = 0.0_pReal ! needed for check of sign change in flux density below
neighboring_rhoSgl = 0.0_pReal
added possibility to scale the plastic shearrate by the ratio of mobile to total density; keyword in material.config is "deadZone"
2012-10-29 18:19:28 +05:30
if (neighboring_n > 0_pInt) then
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (phase_plasticity(material_phase(1,neighboring_ip,neighboring_el)) == CONSTITUTIVE_NONLOCAL_LABEL &
.and. any(compatibility(:,:,:,n,ip,el) > 0.0_pReal)) &
* internal stress calculation now considers dead dislocations always at the interface, not at the center of the IP volume; used to merge them together with "normal" dislocations for stress calculation. * dislocation flux is blocked if we encounter a sign change in the resolved shear stress from the central ip to the neighbor * do not set density to zero if below certain threshold; this creates an artificial sink term
2011-08-02 16:47:45 +05:30
considerEnteringFlux = .true.
* added a new material parameter "surfaceTransmissivity" (default value 1.0) which allows to change the transmissivity of the material surface between 0 and 1 * now complaining when encountering an unknown nonlocal parameter in material.config * use same error ID for all material parameters out of bounds * symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity) * switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
2011-02-16 22:05:38 +05:30
endif
if (considerEnteringFlux) then
fixed timesyncing for special case of ip that synchronizes its time step, but does not have a neighbor
2012-12-11 19:08:36 +05:30
if(numerics_timeSyncing .and. (subfrac(g,neighboring_ip,neighboring_el) /= subfrac(g,ip,el))) then ! for timesyncing: in case of a timestep at the interface we have to use "state0" to make sure that fluxes n both sides are equal
small changes in nonlocal constitutive model: - make sure neighboring density is always greater equal zero - no extra Bauschinger in postResults
2013-04-03 21:52:55 +05:30
forall (t = 1_pInt:4_pInt)
improved accuracy of "accumulated shear": now added as a state variable, which facilitates integration (dotstate=shearrate); needs absolute tolerance value for state integration in material.config
2013-04-04 19:07:14 +05:30
neighboring_v(1_pInt:ns,t) = state0(g,neighboring_ip,neighboring_el)%p((13_pInt+t)*ns+1_pInt:(14_pInt+t)*ns)
small changes in nonlocal constitutive model: - make sure neighboring density is always greater equal zero - no extra Bauschinger in postResults
2013-04-03 21:52:55 +05:30
neighboring_rhoSgl(1_pInt:ns,t) = max(state0(g,neighboring_ip,neighboring_el)%p((t-1_pInt)*ns+1_pInt:t*ns), 0.0_pReal)
endforall
forgot to set neighboring_rho (also reverting changes made in rev 2384)
2013-05-20 00:53:31 +05:30
forall (t = 5_pInt:8_pInt) &
neighboring_rhoSgl(1_pInt:ns,t) = state0(g,neighboring_ip,neighboring_el)%p((t-1_pInt)*ns+1_pInt:t*ns)
Introduced an alternative cutback scheme for the nonlocal model, that allows to keep the results of most of the ips that immediately converged, and only do cutbacks in regions where some ips are in big trouble. Only works for nonlocal model and state integrator 2.
2012-11-28 00:06:55 +05:30
else
small changes in nonlocal constitutive model: - make sure neighboring density is always greater equal zero - no extra Bauschinger in postResults
2013-04-03 21:52:55 +05:30
forall (t = 1_pInt:4_pInt)
improved accuracy of "accumulated shear": now added as a state variable, which facilitates integration (dotstate=shearrate); needs absolute tolerance value for state integration in material.config
2013-04-04 19:07:14 +05:30
neighboring_v(1_pInt:ns,t) = state(g,neighboring_ip,neighboring_el)%p((13_pInt+t)*ns+1_pInt:(14_pInt+t)*ns)
small changes in nonlocal constitutive model: - make sure neighboring density is always greater equal zero - no extra Bauschinger in postResults
2013-04-03 21:52:55 +05:30
neighboring_rhoSgl(1_pInt:ns,t) = max(state(g,neighboring_ip,neighboring_el)%p((t-1_pInt)*ns+1_pInt:t*ns), 0.0_pReal)
endforall
forgot to set neighboring_rho (also reverting changes made in rev 2384)
2013-05-20 00:53:31 +05:30
forall (t = 5_pInt:8_pInt) &
neighboring_rhoSgl(1_pInt:ns,t) = state(g,neighboring_ip,neighboring_el)%p((t-1_pInt)*ns+1_pInt:t*ns)
Introduced an alternative cutback scheme for the nonlocal model, that allows to keep the results of most of the ips that immediately converged, and only do cutbacks in regions where some ips are in big trouble. Only works for nonlocal model and state integrator 2.
2012-11-28 00:06:55 +05:30
endif
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
where (abs(neighboring_rhoSgl) * mesh_ipVolume(neighboring_ip,neighboring_el) ** 0.667_pReal &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
< significantN(myInstance) &
.or. abs(neighboring_rhoSgl) < significantRho(myInstance)) &
introduced factor to control edge contribution to multiplication; dislocation density below a single dislocation per IP considered not significant
2012-09-04 22:26:37 +05:30
neighboring_rhoSgl = 0.0_pReal
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
normal_neighbor2me_defConf = math_det33(Favg) * math_mul33x3(math_inv33(transpose(Favg)), &
mesh_ipAreaNormal(1:3,neighboring_n,neighboring_ip,neighboring_el)) ! calculate the normal of the interface in (average) deformed configuration (now pointing from my neighbor to me!!!)
added new math function "math_deviatoric33" which returns the deviatoric part of a 3x3 tensor renamed some math functions, so that we have a universal naming scheme: for matrix multiplications use an "x" (e.g. math_mul33x3); don't use the "x" to describe the shape of the tensor that the function is applied to (e.g. math_invert33 instead of math_invert3x3)
2012-01-26 19:20:00 +05:30
normal_neighbor2me = math_mul33x3(transpose(neighboring_Fe), normal_neighbor2me_defConf) / math_det33(neighboring_Fe) ! interface normal in the lattice configuration of my neighbor
* added a new material parameter "surfaceTransmissivity" (default value 1.0) which allows to change the transmissivity of the material surface between 0 and 1 * now complaining when encountering an unknown nonlocal parameter in material.config * use same error ID for all material parameters out of bounds * symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity) * switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
2011-02-16 22:05:38 +05:30
area = mesh_ipArea(neighboring_n,neighboring_ip,neighboring_el) * math_norm3(normal_neighbor2me)
normal_neighbor2me = normal_neighbor2me / math_norm3(normal_neighbor2me) ! normalize the surface normal to unit length
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do s = 1_pInt,ns
do t = 1_pInt,4_pInt
c = (t + 1_pInt) / 2
topp = t + mod(t,2_pInt) - mod(t+1_pInt,2_pInt)
Dead dislocations are treated the same as mobile dislocations for the flux part though they do not contribute to slip. By that the pileup of dead dislocations is able to diminish with time (kinetics are the same as for the glide part; this is not correct but gives valuable results). Also remobilization of dead dislocations at neighbor is taken into account for flux.
2012-03-14 21:02:50 +05:30
if (neighboring_v(s,t) * math_mul3x3(m(1:3,s,t), normal_neighbor2me) > 0.0_pReal & ! flux from my neighbor to me == entering flux for me
slight change in flux formulation: no flux to neighbor, not only if there is a sign change in the velocity, but also if the neighboring velocity is exactly zero
2013-05-10 14:12:42 +05:30
.and. v(s,t) * neighboring_v(s,t) > 0.0_pReal ) then ! ... only if no sign change in flux density
reverted changes made in rev2384
2013-05-17 18:24:47 +05:30
do deads = 0_pInt,4_pInt,4_pInt
lineLength = abs(neighboring_rhoSgl(s,t+deads)) * neighboring_v(s,t) &
* math_mul3x3(m(1:3,s,t), normal_neighbor2me) * area ! positive line length that wants to enter through this interface
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
where (compatibility(c,1_pInt:ns,s,n,ip,el) > 0.0_pReal) & ! positive compatibility...
reverted changes made in rev2384
2013-05-17 18:24:47 +05:30
rhoDotFlux(1_pInt:ns,t) = rhoDotFlux(1_pInt:ns,t) + lineLength / mesh_ipVolume(ip,el) & ! ... transferring to equally signed mobile dislocation type
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
* compatibility(c,1_pInt:ns,s,n,ip,el) ** 2.0_pReal
where (compatibility(c,1_pInt:ns,s,n,ip,el) < 0.0_pReal) & ! ..negative compatibility...
reverted changes made in rev2384
2013-05-17 18:24:47 +05:30
rhoDotFlux(1_pInt:ns,topp) = rhoDotFlux(1_pInt:ns,topp) + lineLength / mesh_ipVolume(ip,el) & ! ... transferring to opposite signed mobile dislocation type
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
* compatibility(c,1_pInt:ns,s,n,ip,el) ** 2.0_pReal
reverted changes made in rev2384
2013-05-17 18:24:47 +05:30
enddo
fluxes are disabled when keyword "/nonlocal/" is omitted in the material.config (this is a way to use the constitutive_nonlocal model as a local dislocation based model)
2011-01-11 20:25:36 +05:30
endif
* added a new material parameter "surfaceTransmissivity" (default value 1.0) which allows to change the transmissivity of the material surface between 0 and 1 * now complaining when encountering an unknown nonlocal parameter in material.config * use same error ID for all material parameters out of bounds * symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity) * switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
2011-02-16 22:05:38 +05:30
enddo
enddo
fluxes are disabled when keyword "/nonlocal/" is omitted in the material.config (this is a way to use the constitutive_nonlocal model as a local dislocation based model)
2011-01-11 20:25:36 +05:30
endif
* internal stress calculation now considers dead dislocations always at the interface, not at the center of the IP volume; used to merge them together with "normal" dislocations for stress calculation. * dislocation flux is blocked if we encounter a sign change in the resolved shear stress from the central ip to the neighbor * do not set density to zero if below certain threshold; this creates an artificial sink term
2011-08-02 16:47:45 +05:30
!* FLUX FROM ME TO MY NEIGHBOR
* new output variable "internalstress", which gives the internal stress tensor * use "math_invert3x3" instead of "math_inv3x3" for inversion of Fe * for dislocation stress calculation: first regular case, then special case of dead dislocations in central ip * "dv_dtau" now given for each dislocation type, so is a (ns,4) array * deleted unused variables in "_LpAndItsTangent" * corrected contribution of deads in "_LpAndItsTangent" * the NaN variables defined in math did not give a proper NaN value, so use 0.0/0.0 again * neighbors with nonlocal constitution but local properties (i.e. /nonlocal/ flag not set) are also considered for incoming fluxes
2011-09-07 17:00:28 +05:30
!* This is not considered, if my opposite neighbor has a different constitutive law than nonlocal (still considered for nonlocal law with lcal properties).
* internal stress calculation now considers dead dislocations always at the interface, not at the center of the IP volume; used to merge them together with "normal" dislocations for stress calculation. * dislocation flux is blocked if we encounter a sign change in the resolved shear stress from the central ip to the neighbor * do not set density to zero if below certain threshold; this creates an artificial sink term
2011-08-02 16:47:45 +05:30
!* Then, we assume, that the opposite(!) neighbor sends an equal amount of dislocations to me.
!* So the net flux in the direction of my neighbor is equal to zero:
!* leaving flux to neighbor == entering flux from opposite neighbor
!* In case of reduced transmissivity, part of the leaving flux is stored as dead dislocation density.
!* That means for an interface of zero transmissivity the leaving flux is fully converted to dead dislocations.
considerLeavingFlux = .true.
added possibility to scale the plastic shearrate by the ratio of mobile to total density; keyword in material.config is "deadZone"
2012-10-29 18:19:28 +05:30
if (opposite_n > 0_pInt) then
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (phase_plasticity(material_phase(1,opposite_ip,opposite_el)) /= CONSTITUTIVE_NONLOCAL_LABEL) &
* internal stress calculation now considers dead dislocations always at the interface, not at the center of the IP volume; used to merge them together with "normal" dislocations for stress calculation. * dislocation flux is blocked if we encounter a sign change in the resolved shear stress from the central ip to the neighbor * do not set density to zero if below certain threshold; this creates an artificial sink term
2011-08-02 16:47:45 +05:30
considerLeavingFlux = .false.
endif
if (considerLeavingFlux) then
fixed timesyncing for special case of ip that synchronizes its time step, but does not have a neighbor
2012-12-11 19:08:36 +05:30
!* timeSyncing mode: If the central ip has zero subfraction, always use "state0". This is needed in case of
!* a synchronization step for the central ip, because then "state" contains the values at the end of the
!* previously converged full time step. Also, if either me or my neighbor has zero subfraction, we have to
!* use "state0" to make sure that fluxes on both sides of the (potential) timestep are equal.
rhoSglMe = rhoSgl
vMe = v
if(numerics_timeSyncing) then
if (subfrac(g,ip,el) == 0.0_pReal) then
always use the current state for leaving flux calculation when no neighbor is present
2012-12-09 17:54:32 +05:30
rhoSglMe = rhoSgl0
vMe = v0
fixed timesyncing for special case of ip that synchronizes its time step, but does not have a neighbor
2012-12-11 19:08:36 +05:30
elseif (neighboring_n > 0_pInt) then
if (subfrac(g,neighboring_ip,neighboring_el) == 0.0_pReal) then
rhoSglMe = rhoSgl0
vMe = v0
endif
always use the current state for leaving flux calculation when no neighbor is present
2012-12-09 17:54:32 +05:30
endif
Introduced an alternative cutback scheme for the nonlocal model, that allows to keep the results of most of the ips that immediately converged, and only do cutbacks in regions where some ips are in big trouble. Only works for nonlocal model and state integrator 2.
2012-11-28 00:06:55 +05:30
endif
fixed timesyncing for special case of ip that synchronizes its time step, but does not have a neighbor
2012-12-11 19:08:36 +05:30
corrected kinetics law and changed parameters. For solid solution hardening there are 3 parameters: the activation energy, the concentration of obstacles that determines the activation length and meanfreepath, and the obstacle size that determines the activation volume. For the Peierls mechanism there is: the width of doublekinks that determines the activation volume and the Peierls stress for edge and screw. Still testing needed to check whether the current formulation makes sense or not.
2012-02-03 18:20:54 +05:30
normal_me2neighbor_defConf = math_det33(Favg) * math_mul33x3(math_inv33(math_transpose33(Favg)), &
mesh_ipAreaNormal(1:3,n,ip,el)) ! calculate the normal of the interface in (average) deformed configuration (pointing from me to my neighbor!!!)
N^2 initialization loop (former spectralPictureMode) rewritten in material.f90 additional output in DAMASK_spectral_interface.f90 132 character cut off in constitutive_nonlocal.f90 rounding error in math.f90 complex number initialization (1.0_pReal)*2.0_pReal*pi new $DAMASK_NUM_THREADS warning in numerics.f90 / IO.f90 polishing in DAMASK_spectral.f90
2012-01-30 19:22:41 +05:30
normal_me2neighbor = math_mul33x3(math_transpose33(my_Fe), normal_me2neighbor_defConf) / math_det33(my_Fe) ! interface normal in my lattice configuration
* internal stress calculation now considers dead dislocations always at the interface, not at the center of the IP volume; used to merge them together with "normal" dislocations for stress calculation. * dislocation flux is blocked if we encounter a sign change in the resolved shear stress from the central ip to the neighbor * do not set density to zero if below certain threshold; this creates an artificial sink term
2011-08-02 16:47:45 +05:30
area = mesh_ipArea(n,ip,el) * math_norm3(normal_me2neighbor)
normal_me2neighbor = normal_me2neighbor / math_norm3(normal_me2neighbor) ! normalize the surface normal to unit length
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do s = 1_pInt,ns
do t = 1_pInt,4_pInt
Introduced an alternative cutback scheme for the nonlocal model, that allows to keep the results of most of the ips that immediately converged, and only do cutbacks in regions where some ips are in big trouble. Only works for nonlocal model and state integrator 2.
2012-11-28 00:06:55 +05:30
c = (t + 1_pInt) / 2_pInt
if (vMe(s,t) * math_mul3x3(m(1:3,s,t), normal_me2neighbor) > 0.0_pReal ) then ! flux from me to my neighbor == leaving flux for me (might also be a pure flux from my mobile density to dead density if interface not at all transmissive)
slight change in flux formulation: no flux to neighbor, not only if there is a sign change in the velocity, but also if the neighboring velocity is exactly zero
2013-05-10 14:12:42 +05:30
if (vMe(s,t) * neighboring_v(s,t) > 0.0_pReal) then ! no sign change in flux density
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
transmissivity = sum(compatibility(c,1_pInt:ns,s,n,ip,el)**2.0_pReal) ! overall transmissivity from this slip system to my neighbor
* internal stress calculation now considers dead dislocations always at the interface, not at the center of the IP volume; used to merge them together with "normal" dislocations for stress calculation. * dislocation flux is blocked if we encounter a sign change in the resolved shear stress from the central ip to the neighbor * do not set density to zero if below certain threshold; this creates an artificial sink term
2011-08-02 16:47:45 +05:30
else ! sign change in flux density means sign change in stress which does not allow for dislocations to arive at the neighbor
transmissivity = 0.0_pReal
endif
Introduced an alternative cutback scheme for the nonlocal model, that allows to keep the results of most of the ips that immediately converged, and only do cutbacks in regions where some ips are in big trouble. Only works for nonlocal model and state integrator 2.
2012-11-28 00:06:55 +05:30
lineLength = rhoSglMe(s,t) * vMe(s,t) * math_mul3x3(m(1:3,s,t), normal_me2neighbor) * area ! positive line length of mobiles that wants to leave through this interface
Dead dislocations are treated the same as mobile dislocations for the flux part though they do not contribute to slip. By that the pileup of dead dislocations is able to diminish with time (kinetics are the same as for the glide part; this is not correct but gives valuable results). Also remobilization of dead dislocations at neighbor is taken into account for flux.
2012-03-14 21:02:50 +05:30
rhoDotFlux(s,t) = rhoDotFlux(s,t) - lineLength / mesh_ipVolume(ip,el) ! subtract dislocation flux from current type
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
rhoDotFlux(s,t+4_pInt) = rhoDotFlux(s,t+4_pInt) + lineLength / mesh_ipVolume(ip,el) * (1.0_pReal - transmissivity) &
Introduced an alternative cutback scheme for the nonlocal model, that allows to keep the results of most of the ips that immediately converged, and only do cutbacks in regions where some ips are in big trouble. Only works for nonlocal model and state integrator 2.
2012-11-28 00:06:55 +05:30
* sign(1.0_pReal, vMe(s,t)) ! dislocation flux that is not able to leave through interface (because of low transmissivity) will remain as immobile single density at the material point
reverted changes made in rev2384
2013-05-17 18:24:47 +05:30
lineLength = rhoSglMe(s,t+4_pInt) * vMe(s,t) * math_mul3x3(m(1:3,s,t), normal_me2neighbor) * area ! positive line length of deads that wants to leave through this interface
rhoDotFlux(s,t+4_pInt) = rhoDotFlux(s,t+4_pInt) - lineLength / mesh_ipVolume(ip,el) * transmissivity ! dead dislocations leaving through this interface
* internal stress calculation now considers dead dislocations always at the interface, not at the center of the IP volume; used to merge them together with "normal" dislocations for stress calculation. * dislocation flux is blocked if we encounter a sign change in the resolved shear stress from the central ip to the neighbor * do not set density to zero if below certain threshold; this creates an artificial sink term
2011-08-02 16:47:45 +05:30
endif
enddo
enddo
endif
* added a new material parameter "surfaceTransmissivity" (default value 1.0) which allows to change the transmissivity of the material surface between 0 and 1 * now complaining when encountering an unknown nonlocal parameter in material.config * use same error ID for all material parameters out of bounds * symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity) * switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
2011-02-16 22:05:38 +05:30
* in flux calculation: take care of special case for single element model * do not overwrite "constitutive_nonlocal_rhoDotFlux" when doing a state perturbation, so we can use it for output
2011-02-24 15:31:41 +05:30
enddo ! neighbor loop
* corrected compatibility for screws (always positive) * detection of grain boundary in constitutive_nonlocal_microstructure with the help of transmissivity * enforce positive densities in constitutive_nonlocal_microstructure (needed because dotState does not create cutbacks for negative densities anymore) * reset single mobile densities below certain threshold to zero (also done in constitutive_nonlocal_microstructure) * constitutive_nonlocal_kinetics only gets local state variable as input, no need for the entire array here * dv_dtau is always positive * multiplication is only active when there is already some initial density of the respective type
2010-10-26 19:12:18 +05:30
endif
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
openmp parallelization working again (at least for j2 and nonlocal constitutive model). In order to keep it like that, please follow these simple rules: DON'T use implicit array subscripts: example: real, dimension(3,3) :: A,B A(:,2) = B(:,1) <--- DON'T USE A(1:3,2) = B(1:3,1) <--- BETTER USE In many cases the use of explicit array subscripts is inevitable for parallelization. Additionally, it is an easy means to prevent memory leaks. Enclose all write statements with the following: !$OMP CRITICAL (write2out) <your write statement> !$OMP END CRITICAL (write2out) Whenever you change something in the code and are not sure if it affects parallelization and leads to nonconforming behavior, please ask me and/or Franz to check this.
2011-03-17 16:16:17 +05:30
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
!****************************************************************************
!*** calculate dipole formation and annihilation
!*** formation by glide
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do c = 1_pInt,2_pInt
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoDotSingle2DipoleGlide(1:ns,2*c-1) = -2.0_pReal * dUpper(1:ns,c) / burgers(1:ns,myInstance) &
In order to help preventing further memory leaks all array sections now have an explicit instead of assumed shape, e.g. Fe(1:3,1:3) instead of Fe(:,:).
2011-02-09 18:42:46 +05:30
* (rhoSgl(1:ns,2*c-1) * abs(gdot(1:ns,2*c)) & ! negative mobile --> positive mobile
+ rhoSgl(1:ns,2*c) * abs(gdot(1:ns,2*c-1)) & ! positive mobile --> negative mobile
+ abs(rhoSgl(1:ns,2*c+4)) * abs(gdot(1:ns,2*c-1))) ! positive mobile --> negative immobile
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoDotSingle2DipoleGlide(1:ns,2*c) = -2.0_pReal * dUpper(1:ns,c) / burgers(1:ns,myInstance) &
In order to help preventing further memory leaks all array sections now have an explicit instead of assumed shape, e.g. Fe(1:3,1:3) instead of Fe(:,:).
2011-02-09 18:42:46 +05:30
* (rhoSgl(1:ns,2*c-1) * abs(gdot(1:ns,2*c)) & ! negative mobile --> positive mobile
+ rhoSgl(1:ns,2*c) * abs(gdot(1:ns,2*c-1)) & ! positive mobile --> negative mobile
+ abs(rhoSgl(1:ns,2*c+3)) * abs(gdot(1:ns,2*c))) ! negative mobile --> positive immobile
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoDotSingle2DipoleGlide(1:ns,2*c+3) = -2.0_pReal * dUpper(1:ns,c) / burgers(1:ns,myInstance) &
In order to help preventing further memory leaks all array sections now have an explicit instead of assumed shape, e.g. Fe(1:3,1:3) instead of Fe(:,:).
2011-02-09 18:42:46 +05:30
* rhoSgl(1:ns,2*c+3) * abs(gdot(1:ns,2*c)) ! negative mobile --> positive immobile
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoDotSingle2DipoleGlide(1:ns,2*c+4) = -2.0_pReal * dUpper(1:ns,c) / burgers(1:ns,myInstance) &
In order to help preventing further memory leaks all array sections now have an explicit instead of assumed shape, e.g. Fe(1:3,1:3) instead of Fe(:,:).
2011-02-09 18:42:46 +05:30
* rhoSgl(1:ns,2*c+4) * abs(gdot(1:ns,2*c-1)) ! positive mobile --> negative immobile
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
In order to help preventing further memory leaks all array sections now have an explicit instead of assumed shape, e.g. Fe(1:3,1:3) instead of Fe(:,:).
2011-02-09 18:42:46 +05:30
rhoDotSingle2DipoleGlide(1:ns,c+8) = - rhoDotSingle2DipoleGlide(1:ns,2*c-1) - rhoDotSingle2DipoleGlide(1:ns,2*c) &
+ abs(rhoDotSingle2DipoleGlide(1:ns,2*c+3)) + abs(rhoDotSingle2DipoleGlide(1:ns,2*c+4))
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
enddo
constitutive_nonlocal - take orientation gradients into account when calculating dislocation stress and dislocation fluxes - hard coded value for nu - changed kinetics (parameter G0 is currently defined as a parameter, needs to be read from material.config) - added some output statements constitutive: - some functions and subroutines needed additional input variables for passing to constitutive_nonlocal crystallite: - some functions and subroutines needed additional input variables for passing to constitutive - call microstructure with current temperature, Fp, Fe, not "sub0" values - show number of IPs, that are "onTrack" instead of those not "onTrack" - calculate Fe at beginning of substep, since we need it for state preguess
2009-10-07 21:01:52 +05:30
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
annihilate only part of the screw dipoles (specified by minimumDipoleHeight), not all; moved annihilation of screws from deltaState back to dotState
2012-11-28 17:39:48 +05:30
!*** athermal annihilation
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
in nonlocal_microstructure: at the model surface the excess density is now extrapolated from opposite neighbor instead of assuming zero density; this results in better modeling of the dislocation stress at the surface restructured nonlocal_dotstate, to be able to easily switch on and off particular effects in the microstructure evolution nonlocal_dotstate now enforces a cutback when single density runs the risk of becoming negative; in the case of a state already below the relevantState dotState is set to zero introduced two new output variables: rho_dot_edge and rho_dot_screw
2010-05-21 14:21:15 +05:30
rhoDotAthermalAnnihilation = 0.0_pReal
annihilate only part of the screw dipoles (specified by minimumDipoleHeight), not all; moved annihilation of screws from deltaState back to dotState
2012-11-28 17:39:48 +05:30
forall (c=1_pInt:2_pInt) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoDotAthermalAnnihilation(1:ns,c+8_pInt) = -2.0_pReal * dLower(1:ns,c) / burgers(1:ns,myInstance) &
annihilate only part of the screw dipoles (specified by minimumDipoleHeight), not all; moved annihilation of screws from deltaState back to dotState
2012-11-28 17:39:48 +05:30
* ( 2.0_pReal * (rhoSgl(1:ns,2*c-1) * abs(gdot(1:ns,2*c)) + rhoSgl(1:ns,2*c) * abs(gdot(1:ns,2*c-1))) & ! was single hitting single
+ 2.0_pReal * (abs(rhoSgl(1:ns,2*c+3)) * abs(gdot(1:ns,2*c)) + abs(rhoSgl(1:ns,2*c+4)) * abs(gdot(1:ns,2*c-1))) & ! was single hitting immobile single or was immobile single hit by single
+ rhoDip(1:ns,c) * (abs(gdot(1:ns,2*c-1)) + abs(gdot(1:ns,2*c)))) ! single knocks dipole constituent
! annihilated screw dipoles leave edge jogs behind on the colinear system
if (myStructure == 1_pInt) then ! only fcc
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
forall (s = 1:ns, colinearSystem(s,myInstance) > 0_pInt) &
rhoDotAthermalAnnihilation(colinearSystem(s,myInstance),1:2) = - rhoDotAthermalAnnihilation(s,10) &
* 0.25_pReal * sqrt(rhoForest(s)) * (dUpper(s,2) + dLower(s,2)) * edgeJogFactor(myInstance)
annihilate only part of the screw dipoles (specified by minimumDipoleHeight), not all; moved annihilation of screws from deltaState back to dotState
2012-11-28 17:39:48 +05:30
endif
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
annihilation of screws
2012-11-17 19:20:20 +05:30
!*** thermally activated annihilation of edge dipoles by climb
constitutive_nonlocal - take orientation gradients into account when calculating dislocation stress and dislocation fluxes - hard coded value for nu - changed kinetics (parameter G0 is currently defined as a parameter, needs to be read from material.config) - added some output statements constitutive: - some functions and subroutines needed additional input variables for passing to constitutive_nonlocal crystallite: - some functions and subroutines needed additional input variables for passing to constitutive - call microstructure with current temperature, Fp, Fe, not "sub0" values - show number of IPs, that are "onTrack" instead of those not "onTrack" - calculate Fe at beginning of substep, since we need it for state preguess
2009-10-07 21:01:52 +05:30
in nonlocal_microstructure: at the model surface the excess density is now extrapolated from opposite neighbor instead of assuming zero density; this results in better modeling of the dislocation stress at the surface restructured nonlocal_dotstate, to be able to easily switch on and off particular effects in the microstructure evolution nonlocal_dotstate now enforces a cutback when single density runs the risk of becoming negative; in the case of a state already below the relevantState dotState is set to zero introduced two new output variables: rho_dot_edge and rho_dot_screw
2010-05-21 14:21:15 +05:30
rhoDotThermalAnnihilation = 0.0_pReal
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
selfDiffusion = Dsd0(myInstance) * exp(-selfDiffusionEnergy(myInstance) / (KB * Temperature))
vClimb = atomicVolume(myInstance) * selfDiffusion / ( KB * Temperature ) &
* mu(myInstance) / ( 2.0_pReal * PI * (1.0_pReal-nu(myInstance)) ) &
In order to help preventing further memory leaks all array sections now have an explicit instead of assumed shape, e.g. Fe(1:3,1:3) instead of Fe(:,:).
2011-02-09 18:42:46 +05:30
* 2.0_pReal / ( dUpper(1:ns,1) + dLower(1:ns,1) )
hardening coefficients for fcc are density dependent (line tension effect, Devincre, Kubin, Hoc; 2006; Physical analyses of crystal plasticity by DD simulations) fixed bug with thermal annihilation producing negative density in case of very small dipole density and high climb velocity
2012-10-04 23:38:40 +05:30
forall (s = 1_pInt:ns, dUpper(s,1) > dLower(s,1)) &
rhoDotThermalAnnihilation(s,9) = max(- 4.0_pReal * rhoDip(s,1) * vClimb(s) / (dUpper(s,1) - dLower(s,1)), &
- rhoDip(s,1) / timestep - rhoDotAthermalAnnihilation(s,9) - rhoDotSingle2DipoleGlide(s,9)) ! make sure that we do not annihilate more dipoles than we have
introduced annihilation of screw dipoles; so far no thermal activation, but immediate annihilation; each annihilated screw dipole leaves behind an edge jog of same length as the dipole height on the collinear system
2012-08-14 17:56:20 +05:30
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
!****************************************************************************
!*** assign the rates of dislocation densities to my dotState
dotState=NaN enforces cutback, if evolution rates lead to negative densities
2012-02-22 21:38:22 +05:30
!*** if evolution rates lead to negative densities, a cutback is enforced
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
in nonlocal_microstructure: at the model surface the excess density is now extrapolated from opposite neighbor instead of assuming zero density; this results in better modeling of the dislocation stress at the surface restructured nonlocal_dotstate, to be able to easily switch on and off particular effects in the microstructure evolution nonlocal_dotstate now enforces a cutback when single density runs the risk of becoming negative; in the case of a state already below the relevantState dotState is set to zero introduced two new output variables: rho_dot_edge and rho_dot_screw
2010-05-21 14:21:15 +05:30
rhoDot = 0.0_pReal
* internal stress calculation now considers dead dislocations always at the interface, not at the center of the IP volume; used to merge them together with "normal" dislocations for stress calculation. * dislocation flux is blocked if we encounter a sign change in the resolved shear stress from the central ip to the neighbor * do not set density to zero if below certain threshold; this creates an artificial sink term
2011-08-02 16:47:45 +05:30
rhoDot = rhoDotFlux &
+ rhoDotMultiplication &
+ rhoDotSingle2DipoleGlide &
+ rhoDotAthermalAnnihilation &
+ rhoDotThermalAnnihilation
in nonlocal_microstructure: at the model surface the excess density is now extrapolated from opposite neighbor instead of assuming zero density; this results in better modeling of the dislocation stress at the surface restructured nonlocal_dotstate, to be able to easily switch on and off particular effects in the microstructure evolution nonlocal_dotstate now enforces a cutback when single density runs the risk of becoming negative; in the case of a state already below the relevantState dotState is set to zero introduced two new output variables: rho_dot_edge and rho_dot_screw
2010-05-21 14:21:15 +05:30
rates in constitutive_results are taken directly from last converged step and not calculated again; added output of rates for annihilation of screws and deposition of edge jogs on collinear system
2012-08-16 16:33:22 +05:30
if (numerics_integrationMode == 1_pInt) then ! save rates for output if in central integration mode
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
rhoDotFluxOutput(1:ns,1:8,g,ip,el) = rhoDotFlux(1:ns,1:8)
rhoDotMultiplicationOutput(1:ns,1:2,g,ip,el) = rhoDotMultiplication(1:ns,[1,3])
rhoDotSingle2DipoleGlideOutput(1:ns,1:2,g,ip,el) = rhoDotSingle2DipoleGlide(1:ns,9:10)
rhoDotAthermalAnnihilationOutput(1:ns,1:2,g,ip,el) = rhoDotAthermalAnnihilation(1:ns,9:10)
rhoDotThermalAnnihilationOutput(1:ns,1:2,g,ip,el) = rhoDotThermalAnnihilation(1:ns,9:10)
rhoDotEdgeJogsOutput(1:ns,g,ip,el) = 2.0_pReal * rhoDotThermalAnnihilation(1:ns,1)
rates in constitutive_results are taken directly from last converged step and not calculated again; added output of rates for annihilation of screws and deposition of edge jogs on collinear system
2012-08-16 16:33:22 +05:30
endif
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
#ifndef _OPENMP
In crystallite: for state integrators 2 and 3 the deltaState always has to be done at the beginning of the integration, not at the end; otherwise the stiffness calculation can be completely wrong, when the state is changed after the stress integration for the central solution. Also changed some debugging levels in crystallite and mainly in constitutive_nonlocal
2012-10-22 13:29:35 +05:30
if (iand(debug_level(debug_constitutive),debug_levelExtensive) /= 0_pInt &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == g)&
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
.or. .not. iand(debug_level(debug_constitutive),debug_levelSelective) /= 0_pInt )) then
some f2008 standard related correction i didn't check in last time
2012-02-02 01:50:05 +05:30
write(6,'(a,/,4(12x,12(e12.5,1x),/))') '<< CONST >> dislocation multiplication', rhoDotMultiplication(1:ns,1:4) * timestep
write(6,'(a,/,8(12x,12(e12.5,1x),/))') '<< CONST >> dislocation flux', rhoDotFlux(1:ns,1:8) * timestep
write(6,'(a,/,10(12x,12(e12.5,1x),/))') '<< CONST >> dipole formation by glide', rhoDotSingle2DipoleGlide * timestep
annihilate only part of the screw dipoles (specified by minimumDipoleHeight), not all; moved annihilation of screws from deltaState back to dotState
2012-11-28 17:39:48 +05:30
write(6,'(a,/,10(12x,12(e12.5,1x),/))') '<< CONST >> athermal dipole annihilation', &
rhoDotAthermalAnnihilation * timestep
annihilation of screws
2012-11-17 19:20:20 +05:30
write(6,'(a,/,2(12x,12(e12.5,1x),/))') '<< CONST >> thermally activated dipole annihilation', &
rhoDotThermalAnnihilation(1:ns,9:10) * timestep
some f2008 standard related correction i didn't check in last time
2012-02-02 01:50:05 +05:30
write(6,'(a,/,10(12x,12(e12.5,1x),/))') '<< CONST >> total density change', rhoDot * timestep
always use the current state for leaving flux calculation when no neighbor is present
2012-12-09 17:54:32 +05:30
write(6,'(a,/,10(12x,12(f12.5,1x),/))') '<< CONST >> relative density change', &
rhoDot(1:ns,1:8) * timestep / (abs(rhoSglOriginal)+1.0e-10), &
rhoDot(1:ns,9:10) * timestep / (rhoDipOriginal+1.0e-10)
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
write(6,*)
* Introduced preprocessor directives in order to suppress compilation of most write statements when using openmp. This tremendously improves efficiency of parallelization. * Also added some more openmp directives to increase percentage of parallelized code. * "implicit none" was missing in two subroutines of homogenization and constitutive.
2011-03-29 12:57:19 +05:30
endif
#endif
* corrected compatibility for screws (always positive) * detection of grain boundary in constitutive_nonlocal_microstructure with the help of transmissivity * enforce positive densities in constitutive_nonlocal_microstructure (needed because dotState does not create cutbacks for negative densities anymore) * reset single mobile densities below certain threshold to zero (also done in constitutive_nonlocal_microstructure) * constitutive_nonlocal_kinetics only gets local state variable as input, no need for the entire array here * dv_dtau is always positive * multiplication is only active when there is already some initial density of the respective type
2010-10-26 19:12:18 +05:30
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if ( any(rhoSglOriginal(1:ns,1:4) + rhoDot(1:ns,1:4) * timestep < -aTolRho(myInstance)) &
.or. any(rhoDipOriginal(1:ns,1:2) + rhoDot(1:ns,9:10) * timestep < -aTolRho(myInstance))) then
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
#ifndef _OPENMP
made random multiplication switchable via material.config
2012-12-03 18:29:38 +05:30
if (iand(debug_level(debug_constitutive),debug_levelExtensive) /= 0_pInt) then
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
write(6,'(a,i5,a,i2)') '<< CONST >> evolution rate leads to negative density at el ',el,' ip ',ip
write(6,'(a)') '<< CONST >> enforcing cutback !!!'
endif
#endif
constitutive_nonlocal_dotState = DAMASK_NaN
return
else
constitutive_nonlocal_dotState(1:10_pInt*ns) = reshape(rhoDot,(/10_pInt*ns/))
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_dotState(10_pInt*ns+1:11_pInt*ns) = shearrate(1:ns,g,ip,el)
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
endif
dotState does not have to be reset to zero. this is a remnant from older versions when the dotState for the nonlocal model used to be updated by the neighboring integration point
2012-05-16 21:05:14 +05:30
endfunction
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
!*********************************************************************
* had to correct nonlocal slip system compatibility and flux calculation. last update gave wrong results.
2010-10-15 18:49:26 +05:30
!* COMPATIBILITY UPDATE *
!* Compatibility is defined as normalized product of signed cosine *
!* of the angle between the slip plane normals and signed cosine of *
!* the angle between the slip directions. Only the largest values *
!* that sum up to a total of 1 are considered, all others are set to *
!* zero. *
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
!*********************************************************************
* in nonlocal model: dislocation flux now with valid description also for large angle grain boundaries; transfer of dislocations from one slip system to the other according to new variable "constitutive_nonlocal_compatibility" which depends on the angle between slip plane normals and slip directions and is updated once per cycle in function "crystallite_orientations" * reactivated debugging functionality for "non-standard" integration methods
2010-10-12 18:38:54 +05:30
subroutine constitutive_nonlocal_updateCompatibility(orientation,i,e)
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
use math, only: math_qDisorientation, &
* in nonlocal model: dislocation flux now with valid description also for large angle grain boundaries; transfer of dislocations from one slip system to the other according to new variable "constitutive_nonlocal_compatibility" which depends on the angle between slip plane normals and slip directions and is updated once per cycle in function "crystallite_orientations" * reactivated debugging functionality for "non-standard" integration methods
2010-10-12 18:38:54 +05:30
math_mul3x3, &
math_qRot
use material, only: material_phase, &
added possibility to assign the transmissivity of grain boundaries a fixed value by use of the keyword "grainboundaryTransmissivity" in the material.config; grain boundaries are identified as interfaces between areas of different material_texture
2012-08-03 20:02:49 +05:30
material_texture, &
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
phase_localPlasticity, &
phase_plasticityInstance, &
* in nonlocal model: dislocation flux now with valid description also for large angle grain boundaries; transfer of dislocations from one slip system to the other according to new variable "constitutive_nonlocal_compatibility" which depends on the angle between slip plane normals and slip directions and is updated once per cycle in function "crystallite_orientations" * reactivated debugging functionality for "non-standard" integration methods
2010-10-12 18:38:54 +05:30
homogenization_maxNgrains
use mesh, only: mesh_element, &
mesh_ipNeighborhood, &
mesh_maxNips, &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
mesh_NcpElems, &
FE_NipNeighbors, &
neighborhood relations were screwed up for some elementtypes as of version 2318, might have caused segmentation faults with nonlocal model, now fixed
2013-04-22 19:05:35 +05:30
FE_geomtype, &
FE_celltype
* in nonlocal model: dislocation flux now with valid description also for large angle grain boundaries; transfer of dislocations from one slip system to the other according to new variable "constitutive_nonlocal_compatibility" which depends on the angle between slip plane normals and slip directions and is updated once per cycle in function "crystallite_orientations" * reactivated debugging functionality for "non-standard" integration methods
2010-10-12 18:38:54 +05:30
use lattice, only: lattice_sn, &
mixed things up, did not want to commit constitutive_nonlocal.f90 but rather DAMASK_spectral.f90!!!! did a lot of polishing on constitutive_nonlocal.f90 (mostly checked in at last commit), for changes on DAMASK_spectral.f90 see previous message
2012-02-23 22:50:57 +05:30
lattice_sd
new subroutine crystallite_orientations: calculates the orientation matrix and the euler angles and - in case of single grain ips - the misorientation for each neighboring ip. The calculation of the misorientation is now done once in crystallite init and at the end of every FE increment. This saves a lot of time compared to doing it in dotState for every crystallite subinc.
2009-12-18 21:16:33 +05:30
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
implicit none
!* input variables
* in nonlocal model: dislocation flux now with valid description also for large angle grain boundaries; transfer of dislocations from one slip system to the other according to new variable "constitutive_nonlocal_compatibility" which depends on the angle between slip plane normals and slip directions and is updated once per cycle in function "crystallite_orientations" * reactivated debugging functionality for "non-standard" integration methods
2010-10-12 18:38:54 +05:30
integer(pInt), intent(in) :: i, & ! ip index
e ! element index
real(pReal), dimension(4,homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), intent(in) :: &
orientation ! crystal orientation in quaternions
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
!* output variables
!* local variables
restructured constitutive_nonlocal_compatibility and incorporated the "surfaceTransmissivity" in the compatibility calculation
2011-02-16 22:08:18 +05:30
integer(pInt) Nneighbors, & ! number of neighbors
n, & ! neighbor index
* in nonlocal model: dislocation flux now with valid description also for large angle grain boundaries; transfer of dislocations from one slip system to the other according to new variable "constitutive_nonlocal_compatibility" which depends on the angle between slip plane normals and slip directions and is updated once per cycle in function "crystallite_orientations" * reactivated debugging functionality for "non-standard" integration methods
2010-10-12 18:38:54 +05:30
neighboring_e, & ! element index of my neighbor
neighboring_i, & ! integration point index of my neighbor
restructured constitutive_nonlocal_compatibility and incorporated the "surfaceTransmissivity" in the compatibility calculation
2011-02-16 22:08:18 +05:30
my_phase, &
neighboring_phase, &
added possibility to assign the transmissivity of grain boundaries a fixed value by use of the keyword "grainboundaryTransmissivity" in the material.config; grain boundaries are identified as interfaces between areas of different material_texture
2012-08-03 20:02:49 +05:30
my_texture, &
neighboring_texture, &
restructured constitutive_nonlocal_compatibility and incorporated the "surfaceTransmissivity" in the compatibility calculation
2011-02-16 22:08:18 +05:30
my_structure, & ! lattice structure
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
my_instance, & ! instance of plasticity
restructured constitutive_nonlocal_compatibility and incorporated the "surfaceTransmissivity" in the compatibility calculation
2011-02-16 22:08:18 +05:30
ns, & ! number of active slip systems
* in nonlocal model: dislocation flux now with valid description also for large angle grain boundaries; transfer of dislocations from one slip system to the other according to new variable "constitutive_nonlocal_compatibility" which depends on the angle between slip plane normals and slip directions and is updated once per cycle in function "crystallite_orientations" * reactivated debugging functionality for "non-standard" integration methods
2010-10-12 18:38:54 +05:30
s1, & ! slip system index (me)
s2 ! slip system index (my neighbor)
real(pReal), dimension(4) :: absoluteMisorientation ! absolute misorientation (without symmetry) between me and my neighbor
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(2,totalNslip(phase_plasticityInstance(material_phase(1,i,e))),&
totalNslip(phase_plasticityInstance(material_phase(1,i,e))),&
neighborhood relations were screwed up for some elementtypes as of version 2318, might have caused segmentation faults with nonlocal model, now fixed
2013-04-22 19:05:35 +05:30
FE_NipNeighbors(FE_celltype(FE_geomtype(mesh_element(2,e))))) :: &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
myCompatibility ! myCompatibility for current element and ip
real(pReal), dimension(3,totalNslip(phase_plasticityInstance(material_phase(1,i,e)))) :: &
restructured constitutive_nonlocal_compatibility and incorporated the "surfaceTransmissivity" in the compatibility calculation
2011-02-16 22:08:18 +05:30
slipNormal, &
slipDirection
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal) myCompatibilitySum, &
restructured constitutive_nonlocal_compatibility and incorporated the "surfaceTransmissivity" in the compatibility calculation
2011-02-16 22:08:18 +05:30
thresholdValue, &
nThresholdValues
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
logical, dimension(totalNslip(phase_plasticityInstance(material_phase(1,i,e)))) :: &
restructured constitutive_nonlocal_compatibility and incorporated the "surfaceTransmissivity" in the compatibility calculation
2011-02-16 22:08:18 +05:30
belowThreshold
* in nonlocal model: dislocation flux now with valid description also for large angle grain boundaries; transfer of dislocations from one slip system to the other according to new variable "constitutive_nonlocal_compatibility" which depends on the angle between slip plane normals and slip directions and is updated once per cycle in function "crystallite_orientations" * reactivated debugging functionality for "non-standard" integration methods
2010-10-12 18:38:54 +05:30
neighborhood relations were screwed up for some elementtypes as of version 2318, might have caused segmentation faults with nonlocal model, now fixed
2013-04-22 19:05:35 +05:30
Nneighbors = FE_NipNeighbors(FE_celltype(FE_geomtype(mesh_element(2,e))))
message of last commit was incomplete: moved annihilation of screw dipoles from dotstate to deltaState, since it is not a rate
2012-11-17 19:24:22 +05:30
my_phase = material_phase(1,i,e)
my_texture = material_texture(1,i,e)
my_instance = phase_plasticityInstance(my_phase)
restructured constitutive_nonlocal_compatibility and incorporated the "surfaceTransmissivity" in the compatibility calculation
2011-02-16 22:08:18 +05:30
my_structure = constitutive_nonlocal_structure(my_instance)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
ns = totalNslip(my_instance)
slipNormal(1:3,1:ns) = lattice_sn(1:3, slipSystemLattice(1:ns,my_instance), my_structure)
slipDirection(1:3,1:ns) = lattice_sd(1:3, slipSystemLattice(1:ns,my_instance), my_structure)
restructured constitutive_nonlocal_compatibility and incorporated the "surfaceTransmissivity" in the compatibility calculation
2011-02-16 22:08:18 +05:30
!*** start out fully compatible
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
myCompatibility = 0.0_pReal
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
forall(s1 = 1_pInt:ns) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
myCompatibility(1:2,s1,s1,1:Nneighbors) = 1.0_pReal
restructured constitutive_nonlocal_compatibility and incorporated the "surfaceTransmissivity" in the compatibility calculation
2011-02-16 22:08:18 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
!*** Loop thrugh neighbors and check whether there is any myCompatibility.
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do n = 1_pInt,Nneighbors
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
neighboring_e = mesh_ipNeighborhood(1,n,i,e)
neighboring_i = mesh_ipNeighborhood(2,n,i,e)
!* FREE SURFACE
!* Set surface transmissivity to the value specified in the material.config
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
if (neighboring_e <= 0_pInt .or. neighboring_i <= 0_pInt) then
forall(s1 = 1_pInt:ns) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
myCompatibility(1:2,s1,s1,n) = sqrt(surfaceTransmissivity(my_instance))
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
cycle
endif
!* PHASE BOUNDARY
!* If we encounter a different nonlocal "cpfem" phase at the neighbor,
!* we consider this to be a real "physical" phase boundary, so completely incompatible.
introduced factor to control edge contribution to multiplication; dislocation density below a single dislocation per IP considered not significant
2012-09-04 22:26:37 +05:30
!* If one of the two "CPFEM" phases has a local plasticity law,
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
!* we do not consider this to be a phase boundary, so completely compatible.
neighboring_phase = material_phase(1,neighboring_i,neighboring_e)
if (neighboring_phase /= my_phase) then
introduced factor to control edge contribution to multiplication; dislocation density below a single dislocation per IP considered not significant
2012-09-04 22:26:37 +05:30
if (.not. phase_localPlasticity(neighboring_phase) .and. .not. phase_localPlasticity(my_phase)) then
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
forall(s1 = 1_pInt:ns) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
myCompatibility(1:2,s1,s1,n) = 0.0_pReal ! = sqrt(0.0)
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
endif
cycle
endif
added possibility to assign the transmissivity of grain boundaries a fixed value by use of the keyword "grainboundaryTransmissivity" in the material.config; grain boundaries are identified as interfaces between areas of different material_texture
2012-08-03 20:02:49 +05:30
!* GRAIN BOUNDARY !
!* fixed transmissivity for adjacent ips with different texture (only if explicitly given in material.config)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (grainboundaryTransmissivity(my_instance) >= 0.0_pReal) then
added possibility to assign the transmissivity of grain boundaries a fixed value by use of the keyword "grainboundaryTransmissivity" in the material.config; grain boundaries are identified as interfaces between areas of different material_texture
2012-08-03 20:02:49 +05:30
neighboring_texture = material_texture(1,neighboring_i,neighboring_e)
if (neighboring_texture /= my_texture) then
if (.not. phase_localPlasticity(neighboring_phase)) then
forall(s1 = 1_pInt:ns) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
myCompatibility(1:2,s1,s1,n) = sqrt(grainboundaryTransmissivity(my_instance))
added possibility to assign the transmissivity of grain boundaries a fixed value by use of the keyword "grainboundaryTransmissivity" in the material.config; grain boundaries are identified as interfaces between areas of different material_texture
2012-08-03 20:02:49 +05:30
endif
cycle
endif
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
!* GRAIN BOUNDARY ?
added possibility to assign the transmissivity of grain boundaries a fixed value by use of the keyword "grainboundaryTransmissivity" in the material.config; grain boundaries are identified as interfaces between areas of different material_texture
2012-08-03 20:02:49 +05:30
!* Compatibility defined by relative orientation of slip systems:
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
!* The myCompatibility value is defined as the product of the slip normal projection and the slip direction projection.
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
!* Its sign is always positive for screws, for edges it has the same sign as the slip normal projection.
!* Since the sum for each slip system can easily exceed one (which would result in a transmissivity larger than one),
!* only values above or equal to a certain threshold value are considered. This threshold value is chosen, such that
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
!* the number of compatible slip systems is minimized with the sum of the original myCompatibility values exceeding one.
!* Finally the smallest myCompatibility value is decreased until the sum is exactly equal to one.
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
!* All values below the threshold are set to zero.
added possibility to assign the transmissivity of grain boundaries a fixed value by use of the keyword "grainboundaryTransmissivity" in the material.config; grain boundaries are identified as interfaces between areas of different material_texture
2012-08-03 20:02:49 +05:30
else
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
absoluteMisorientation = math_qDisorientation(orientation(1:4,1,i,e), &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
orientation(1:4,1,neighboring_i,neighboring_e), &
0_pInt) ! no symmetry
added possibility to assign the transmissivity of grain boundaries a fixed value by use of the keyword "grainboundaryTransmissivity" in the material.config; grain boundaries are identified as interfaces between areas of different material_texture
2012-08-03 20:02:49 +05:30
do s1 = 1_pInt,ns ! my slip systems
do s2 = 1_pInt,ns ! my neighbor's slip systems
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
myCompatibility(1,s2,s1,n) = math_mul3x3(slipNormal(1:3,s1), math_qRot(absoluteMisorientation, slipNormal(1:3,s2))) &
* abs(math_mul3x3(slipDirection(1:3,s1), math_qRot(absoluteMisorientation, slipDirection(1:3,s2))))
myCompatibility(2,s2,s1,n) = abs(math_mul3x3(slipNormal(1:3,s1), math_qRot(absoluteMisorientation, slipNormal(1:3,s2)))) &
* abs(math_mul3x3(slipDirection(1:3,s1), math_qRot(absoluteMisorientation, slipDirection(1:3,s2))))
added possibility to assign the transmissivity of grain boundaries a fixed value by use of the keyword "grainboundaryTransmissivity" in the material.config; grain boundaries are identified as interfaces between areas of different material_texture
2012-08-03 20:02:49 +05:30
enddo
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
myCompatibilitySum = 0.0_pReal
added possibility to assign the transmissivity of grain boundaries a fixed value by use of the keyword "grainboundaryTransmissivity" in the material.config; grain boundaries are identified as interfaces between areas of different material_texture
2012-08-03 20:02:49 +05:30
belowThreshold = .true.
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
do while (myCompatibilitySum < 1.0_pReal .and. any(belowThreshold(1:ns)))
thresholdValue = maxval(myCompatibility(2,1:ns,s1,n), belowThreshold(1:ns)) ! screws always positive
nThresholdValues = real(count(myCompatibility(2,1:ns,s1,n) == thresholdValue),pReal)
where (myCompatibility(2,1:ns,s1,n) >= thresholdValue) &
added possibility to assign the transmissivity of grain boundaries a fixed value by use of the keyword "grainboundaryTransmissivity" in the material.config; grain boundaries are identified as interfaces between areas of different material_texture
2012-08-03 20:02:49 +05:30
belowThreshold(1:ns) = .false.
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (myCompatibilitySum + thresholdValue * nThresholdValues > 1.0_pReal) &
where (abs(myCompatibility(1:2,1:ns,s1,n)) == thresholdValue) &
myCompatibility(1:2,1:ns,s1,n) = sign((1.0_pReal - myCompatibilitySum) &
/ nThresholdValues, myCompatibility(1:2,1:ns,s1,n))
myCompatibilitySum = myCompatibilitySum + nThresholdValues * thresholdValue
added possibility to assign the transmissivity of grain boundaries a fixed value by use of the keyword "grainboundaryTransmissivity" in the material.config; grain boundaries are identified as interfaces between areas of different material_texture
2012-08-03 20:02:49 +05:30
enddo
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
where (belowThreshold(1:ns)) myCompatibility(1,1:ns,s1,n) = 0.0_pReal
where (belowThreshold(1:ns)) myCompatibility(2,1:ns,s1,n) = 0.0_pReal
added possibility to assign the transmissivity of grain boundaries a fixed value by use of the keyword "grainboundaryTransmissivity" in the material.config; grain boundaries are identified as interfaces between areas of different material_texture
2012-08-03 20:02:49 +05:30
enddo ! my slip systems cycle
endif
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
enddo ! neighbor cycle
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
compatibility(1:2,1:ns,1:ns,1:Nneighbors,i,e) = myCompatibility
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
endsubroutine
!*********************************************************************
!* rate of change of temperature *
!*********************************************************************
pure function constitutive_nonlocal_dotTemperature(Tstar_v,Temperature,state,g,ip,el)
use mesh, only: mesh_NcpElems, &
mesh_maxNips
use material, only: homogenization_maxNgrains
implicit none
!* input variables
integer(pInt), intent(in) :: g, & ! current grain ID
ip, & ! current integration point
el ! current element
real(pReal), intent(in) :: Temperature ! temperature
real(pReal), dimension(6), intent(in) :: Tstar_v ! 2nd Piola-Kirchhoff stress in Mandel notation
type(p_vec), dimension(homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), intent(in) :: &
state ! microstructural state
!* output variables
real(pReal) constitutive_nonlocal_dotTemperature ! evolution of Temperature
!* local variables
constitutive_nonlocal_dotTemperature = 0.0_pReal
endfunction
!*********************************************************************
!* calculates quantities characterizing the microstructure *
!*********************************************************************
function constitutive_nonlocal_dislocationstress(state, Fe, g, ip, el)
use math, only: math_mul33x33, &
math_mul33x3, &
added new math function "math_deviatoric33" which returns the deviatoric part of a 3x3 tensor renamed some math functions, so that we have a universal naming scheme: for matrix multiplications use an "x" (e.g. math_mul33x3); don't use the "x" to describe the shape of the tensor that the function is applied to (e.g. math_invert33 instead of math_invert3x3)
2012-01-26 19:20:00 +05:30
math_invert33, &
math_transpose33, &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
pi
use mesh, only: mesh_NcpElems, &
mesh_maxNips, &
mesh_element, &
mesh_node0, &
introduced distinction between ip coordinates and cell center coordinates; the former can be globally calculated by subroutine mesh_build_ipCoordinates, the latter locally by the function mesh_cellCenterCoordinates; renamed mesh_ipCenterOfGravity to mesh_ipCordinates
2012-11-06 20:07:13 +05:30
mesh_cellCenterCoordinates, &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
mesh_ipVolume, &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
mesh_periodicSurface, &
FE_Nips, &
FE_geomtype
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
use material, only: homogenization_maxNgrains, &
material_phase, &
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
phase_localPlasticity, &
phase_plasticityInstance
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
implicit none
!*** input variables
integer(pInt), intent(in) :: g, & ! current grain ID
ip, & ! current integration point
el ! current element
real(pReal), dimension(3,3,homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), intent(in) :: &
Fe ! elastic deformation gradient
type(p_vec), dimension(homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), intent(in) :: &
state ! microstructural state
!*** input/output variables
!*** output variables
real(pReal), dimension(3,3) :: constitutive_nonlocal_dislocationstress
!*** local variables
integer(pInt) neighboring_el, & ! element number of neighboring material point
neighboring_ip, & ! integration point of neighboring material point
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
instance, & ! my instance of this plasticity
neighboring_instance, & ! instance of this plasticity of neighboring material point
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
latticeStruct, & ! my lattice structure
neighboring_latticeStruct, & ! lattice structure of neighboring material point
phase, &
neighboring_phase, &
ns, & ! total number of active slip systems at my material point
neighboring_ns, & ! total number of active slip systems at neighboring material point
c, & ! index of dilsocation character (edge, screw)
s, & ! slip system index
t, & ! index of dilsocation type (e+, e-, s+, s-, used e+, used e-, used s+, used s-)
dir, &
deltaX, deltaY, deltaZ, &
side, &
j
integer(pInt), dimension(2,3) :: periodicImages
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal) x, y, z, & ! coordinates of connection vector in neighboring lattice frame
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
xsquare, ysquare, zsquare, & ! squares of respective coordinates
distance, & ! length of connection vector
segmentLength, & ! segment length of dislocations
lambda, &
R, Rsquare, Rcube, &
denominator, &
flipSign, &
neighboring_ipVolumeSideLength, &
detFe
real(pReal), dimension(3) :: connection, & ! connection vector between me and my neighbor in the deformed configuration
connection_neighboringLattice, & ! connection vector between me and my neighbor in the lattice configuration of my neighbor
connection_neighboringSlip, & ! connection vector between me and my neighbor in the slip system frame of my neighbor
maxCoord, minCoord, &
meshSize, &
introduced distinction between ip coordinates and cell center coordinates; the former can be globally calculated by subroutine mesh_build_ipCoordinates, the latter locally by the function mesh_cellCenterCoordinates; renamed mesh_ipCenterOfGravity to mesh_ipCordinates
2012-11-06 20:07:13 +05:30
coords, & ! x,y,z coordinates of cell center of ip volume
neighboring_coords ! x,y,z coordinates of cell center of neighboring ip volume
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
real(pReal), dimension(3,3) :: sigma, & ! dislocation stress for one slip system in neighboring material point's slip system frame
Tdislo_neighboringLattice, & ! dislocation stress as 2nd Piola-Kirchhoff stress at neighboring material point
invFe, & ! inverse of my elastic deformation gradient
neighboring_invFe, &
neighboringLattice2myLattice ! mapping from neighboring MPs lattice configuration to my lattice configuration
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(2,2,maxval(totalNslip)) :: &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
neighboring_rhoExcess ! excess density at neighboring material point (edge/screw,mobile/dead,slipsystem)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(2,maxval(totalNslip)) :: &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
rhoExcessDead
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),8) :: &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
rhoSgl ! single dislocation density (edge+, edge-, screw+, screw-, used edge+, used edge-, used screw+, used screw-)
logical inversionError
phase = material_phase(g,ip,el)
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
instance = phase_plasticityInstance(phase)
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
latticeStruct = constitutive_nonlocal_structure(instance)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
ns = totalNslip(instance)
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
!*** get basic states
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
forall (s = 1_pInt:ns, t = 1_pInt:4_pInt) &
rhoSgl(s,t) = max(state(g,ip,el)%p((t-1_pInt)*ns+s), 0.0_pReal) ! ensure positive single mobile densities
forall (t = 5_pInt:8_pInt) &
rhoSgl(1:ns,t) = state(g,ip,el)%p((t-1_pInt)*ns+1_pInt:t*ns)
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
!*** calculate the dislocation stress of the neighboring excess dislocation densities
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
!*** zero for material points of local plasticity
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
constitutive_nonlocal_dislocationstress = 0.0_pReal
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
if (.not. phase_localPlasticity(phase)) then
corrected kinetics law and changed parameters. For solid solution hardening there are 3 parameters: the activation energy, the concentration of obstacles that determines the activation length and meanfreepath, and the obstacle size that determines the activation volume. For the Peierls mechanism there is: the width of doublekinks that determines the activation volume and the Peierls stress for edge and screw. Still testing needed to check whether the current formulation makes sense or not.
2012-02-03 18:20:54 +05:30
call math_invert33(Fe(1:3,1:3,g,ip,el), invFe, detFe, inversionError)
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
!* in case of periodic surfaces we have to find out how many periodic images in each direction we need
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do dir = 1_pInt,3_pInt
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
maxCoord(dir) = maxval(mesh_node0(dir,:))
minCoord(dir) = minval(mesh_node0(dir,:))
enddo
meshSize = maxCoord - minCoord
introduced distinction between ip coordinates and cell center coordinates; the former can be globally calculated by subroutine mesh_build_ipCoordinates, the latter locally by the function mesh_cellCenterCoordinates; renamed mesh_ipCenterOfGravity to mesh_ipCordinates
2012-11-06 20:07:13 +05:30
coords = mesh_cellCenterCoordinates(ip,el)
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
periodicImages = 0_pInt
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do dir = 1_pInt,3_pInt
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
if (mesh_periodicSurface(dir)) then
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
periodicImages(1,dir) = floor((coords(dir) - cutoffRadius(instance) - minCoord(dir)) / meshSize(dir), pInt)
periodicImages(2,dir) = ceiling((coords(dir) + cutoffRadius(instance) - maxCoord(dir)) / meshSize(dir), pInt)
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
endif
enddo
!* loop through all material points (also through their periodic images if present),
!* but only consider nonlocal neighbors within a certain cutoff radius R
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do neighboring_el = 1_pInt,mesh_NcpElems
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
ipLoop: do neighboring_ip = 1_pInt,FE_Nips(FE_geomtype(mesh_element(2,neighboring_el)))
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
neighboring_phase = material_phase(g,neighboring_ip,neighboring_el)
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
if (phase_localPlasticity(neighboring_phase)) then
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
cycle
endif
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
neighboring_instance = phase_plasticityInstance(neighboring_phase)
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
neighboring_latticeStruct = constitutive_nonlocal_structure(neighboring_instance)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
neighboring_ns = totalNslip(neighboring_instance)
added new math function "math_deviatoric33" which returns the deviatoric part of a 3x3 tensor renamed some math functions, so that we have a universal naming scheme: for matrix multiplications use an "x" (e.g. math_mul33x3); don't use the "x" to describe the shape of the tensor that the function is applied to (e.g. math_invert33 instead of math_invert3x3)
2012-01-26 19:20:00 +05:30
call math_invert33(Fe(1:3,1:3,1,neighboring_ip,neighboring_el), neighboring_invFe, detFe, inversionError)
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
neighboring_ipVolumeSideLength = mesh_ipVolume(neighboring_ip,neighboring_el) ** (1.0_pReal/3.0_pReal) ! reference volume used here
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
forall (s = 1_pInt:neighboring_ns, c = 1_pInt:2_pInt) &
neighboring_rhoExcess(c,1,s) = state(g,neighboring_ip,neighboring_el)%p((2_pInt*c-2_pInt)*neighboring_ns+s) & ! positive mobiles
- state(g,neighboring_ip,neighboring_el)%p((2_pInt*c-1_pInt)*neighboring_ns+s) ! negative mobiles
forall (s = 1_pInt:neighboring_ns, c = 1_pInt:2_pInt) &
neighboring_rhoExcess(c,2,s) = abs(state(g,neighboring_ip,neighboring_el)%p((2_pInt*c+2_pInt)*neighboring_ns+s)) & ! positive deads
- abs(state(g,neighboring_ip,neighboring_el)%p((2_pInt*c+3_pInt)*neighboring_ns+s)) ! negative deads
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
Tdislo_neighboringLattice = 0.0_pReal
do deltaX = periodicImages(1,1),periodicImages(2,1)
do deltaY = periodicImages(1,2),periodicImages(2,2)
do deltaZ = periodicImages(1,3),periodicImages(2,3)
!* regular case
if (neighboring_el /= el .or. neighboring_ip /= ip &
.or. deltaX /= 0_pInt .or. deltaY /= 0_pInt .or. deltaZ /= 0_pInt) then
introduced distinction between ip coordinates and cell center coordinates; the former can be globally calculated by subroutine mesh_build_ipCoordinates, the latter locally by the function mesh_cellCenterCoordinates; renamed mesh_ipCenterOfGravity to mesh_ipCordinates
2012-11-06 20:07:13 +05:30
neighboring_coords = mesh_cellCenterCoordinates(neighboring_ip,neighboring_el) &
+ (/real(deltaX,pReal), real(deltaY,pReal), real(deltaZ,pReal)/) * meshSize
connection = neighboring_coords - coords
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
distance = sqrt(sum(connection * connection))
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (distance > cutoffRadius(instance)) then
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
cycle
endif
!* the segment length is the minimum of the third root of the control volume and the ip distance
!* this ensures, that the central MP never sits on a neighboring dislocation segment
connection_neighboringLattice = math_mul33x3(neighboring_invFe, connection)
segmentLength = min(neighboring_ipVolumeSideLength, distance)
* in nonlocal model: dislocation flux now with valid description also for large angle grain boundaries; transfer of dislocations from one slip system to the other according to new variable "constitutive_nonlocal_compatibility" which depends on the angle between slip plane normals and slip directions and is updated once per cycle in function "crystallite_orientations" * reactivated debugging functionality for "non-standard" integration methods
2010-10-12 18:38:54 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
!* loop through all slip systems of the neighboring material point
!* and add up the stress contributions from egde and screw excess on these slip systems (if significant)
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do s = 1_pInt,neighboring_ns
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (all(abs(neighboring_rhoExcess(:,:,s)) < significantRho(instance))) then
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
cycle ! not significant
endif
!* map the connection vector from the lattice into the slip system frame
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
connection_neighboringSlip = math_mul33x3(lattice2slip(1:3,1:3,s,neighboring_instance), &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
connection_neighboringLattice)
!* edge contribution to stress
sigma = 0.0_pReal
x = connection_neighboringSlip(1)
y = connection_neighboringSlip(2)
z = connection_neighboringSlip(3)
xsquare = x * x
ysquare = y * y
zsquare = z * z
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do j = 1_pInt,2_pInt
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (abs(neighboring_rhoExcess(1,j,s)) < significantRho(instance)) then
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
cycle
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
elseif (j > 1_pInt) then
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
x = connection_neighboringSlip(1) + sign(0.5_pReal * segmentLength, &
state(g,neighboring_ip,neighboring_el)%p(4*neighboring_ns+s) &
- state(g,neighboring_ip,neighboring_el)%p(5*neighboring_ns+s))
xsquare = x * x
endif
flipSign = sign(1.0_pReal, -y)
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do side = 1_pInt,-1_pInt,-2_pInt
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
lambda = real(side,pReal) * 0.5_pReal * segmentLength - y
R = sqrt(xsquare + zsquare + lambda * lambda)
Rsquare = R * R
Rcube = Rsquare * R
denominator = R * (R + flipSign * lambda)
if (denominator == 0.0_pReal) then
exit ipLoop
endif
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sigma(1,1) = sigma(1,1) - real(side,pReal) &
* flipSign * z / denominator &
* (1.0_pReal + xsquare / Rsquare + xsquare / denominator) &
* neighboring_rhoExcess(1,j,s)
sigma(2,2) = sigma(2,2) - real(side,pReal) &
* (flipSign * 2.0_pReal * nu(instance) * z / denominator + z * lambda / Rcube) &
* neighboring_rhoExcess(1,j,s)
sigma(3,3) = sigma(3,3) + real(side,pReal) &
* flipSign * z / denominator &
* (1.0_pReal - zsquare / Rsquare - zsquare / denominator) &
* neighboring_rhoExcess(1,j,s)
sigma(1,2) = sigma(1,2) + real(side,pReal) &
* x * z / Rcube * neighboring_rhoExcess(1,j,s)
sigma(1,3) = sigma(1,3) + real(side,pReal) &
* flipSign * x / denominator &
* (1.0_pReal - zsquare / Rsquare - zsquare / denominator) &
* neighboring_rhoExcess(1,j,s)
sigma(2,3) = sigma(2,3) - real(side,pReal) &
* (nu(instance) / R - zsquare / Rcube) * neighboring_rhoExcess(1,j,s)
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
enddo
enddo
!* screw contribution to stress
x = connection_neighboringSlip(1) ! have to restore this value, because position might have been adapted for edge deads before
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do j = 1_pInt,2_pInt
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (abs(neighboring_rhoExcess(2,j,s)) < significantRho(instance)) then
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
cycle
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
elseif (j > 1_pInt) then
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
y = connection_neighboringSlip(2) + sign(0.5_pReal * segmentLength, &
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
state(g,neighboring_ip,neighboring_el)%p(6_pInt*neighboring_ns+s) &
- state(g,neighboring_ip,neighboring_el)%p(7_pInt*neighboring_ns+s))
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
ysquare = y * y
endif
changed internal representation of orientation and misorientation from euler angles to quaternions (this should also fix some bugs in the math_misorientation subroutine). includes a couple of new functions in math.f90 and some changes in crystallite.f90. beware that crystallite output "orientation" now by default returns the orientation as quaternion. if you want euler angles instead, you have to add "eulerangles" as a crystallite output in your material.config file (see material.config template). for input of orientations in the texture block of the material.config you still have to specify the rotation in terms of euler angles, quaternions are not yet supported for input.
2010-03-18 17:53:17 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
flipSign = sign(1.0_pReal, x)
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do side = 1_pInt,-1_pInt,-2_pInt
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
lambda = x + real(side,pReal) * 0.5_pReal * segmentLength
R = sqrt(ysquare + zsquare + lambda * lambda)
Rsquare = R * R
Rcube = Rsquare * R
denominator = R * (R + flipSign * lambda)
if (denominator == 0.0_pReal) then
exit ipLoop
endif
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sigma(1,2) = sigma(1,2) - real(side,pReal) * flipSign * z * (1.0_pReal - nu(instance)) / denominator &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
* neighboring_rhoExcess(2,j,s)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sigma(1,3) = sigma(1,3) + real(side,pReal) * flipSign * y * (1.0_pReal - nu(instance)) / denominator &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
* neighboring_rhoExcess(2,j,s)
enddo
enddo
if (all(abs(sigma) < 1.0e-10_pReal)) then ! SIGMA IS NOT A REAL STRESS, THATS WHY WE NEED A REALLY SMALL VALUE HERE
cycle
endif
openmp parallelization working again (at least for j2 and nonlocal constitutive model). In order to keep it like that, please follow these simple rules: DON'T use implicit array subscripts: example: real, dimension(3,3) :: A,B A(:,2) = B(:,1) <--- DON'T USE A(1:3,2) = B(1:3,1) <--- BETTER USE In many cases the use of explicit array subscripts is inevitable for parallelization. Additionally, it is an easy means to prevent memory leaks. Enclose all write statements with the following: !$OMP CRITICAL (write2out) <your write statement> !$OMP END CRITICAL (write2out) Whenever you change something in the code and are not sure if it affects parallelization and leads to nonconforming behavior, please ask me and/or Franz to check this.
2011-03-17 16:16:17 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
!* copy symmetric parts
sigma(2,1) = sigma(1,2)
sigma(3,1) = sigma(1,3)
sigma(3,2) = sigma(2,3)
openmp parallelization working again (at least for j2 and nonlocal constitutive model). In order to keep it like that, please follow these simple rules: DON'T use implicit array subscripts: example: real, dimension(3,3) :: A,B A(:,2) = B(:,1) <--- DON'T USE A(1:3,2) = B(1:3,1) <--- BETTER USE In many cases the use of explicit array subscripts is inevitable for parallelization. Additionally, it is an easy means to prevent memory leaks. Enclose all write statements with the following: !$OMP CRITICAL (write2out) <your write statement> !$OMP END CRITICAL (write2out) Whenever you change something in the code and are not sure if it affects parallelization and leads to nonconforming behavior, please ask me and/or Franz to check this.
2011-03-17 16:16:17 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
!* scale stresses and map them into the neighboring material point's lattice configuration
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sigma = sigma * mu(neighboring_instance) * burgers(s,neighboring_instance) &
/ (4.0_pReal * pi * (1.0_pReal - nu(instance))) &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
* mesh_ipVolume(neighboring_ip,neighboring_el) / segmentLength ! reference volume is used here (according to the segment length calculation)
Tdislo_neighboringLattice = Tdislo_neighboringLattice &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
+ math_mul33x33(math_transpose33(lattice2slip(1:3,1:3,s,neighboring_instance)), &
math_mul33x33(sigma, lattice2slip(1:3,1:3,s,neighboring_instance)))
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
enddo ! slip system loop
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
!* special case of central ip volume
!* only consider dead dislocations
!* we assume that they all sit at a distance equal to half the third root of V
!* in direction of the according slip direction
else
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
forall (s = 1_pInt:ns, c = 1_pInt:2_pInt) &
rhoExcessDead(c,s) = state(g,ip,el)%p((2_pInt*c+2_pInt)*ns+s) & ! positive deads (here we use symmetry: if this has negative sign it is treated as negative density at positive position instead of positive density at negative position)
+ state(g,ip,el)%p((2_pInt*c+3_pInt)*ns+s) ! negative deads (here we use symmetry: if this has negative sign it is treated as positive density at positive position instead of negative density at negative position)
do s = 1_pInt,ns
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
if (all(abs(rhoExcessDead(:,s)) < significantRho(instance))) then
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
cycle ! not significant
endif
sigma = 0.0_pReal ! all components except for sigma13 are zero
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sigma(1,3) = - (rhoExcessDead(1,s) + rhoExcessDead(2,s) * (1.0_pReal - nu(instance))) &
* neighboring_ipVolumeSideLength * mu(instance) * burgers(s,instance) &
/ (sqrt(2.0_pReal) * pi * (1.0_pReal - nu(instance)))
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
sigma(3,1) = sigma(1,3)
Tdislo_neighboringLattice = Tdislo_neighboringLattice &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
+ math_mul33x33(math_transpose33(lattice2slip(1:3,1:3,s,instance)), &
math_mul33x33(sigma, lattice2slip(1:3,1:3,s,instance)))
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
enddo ! slip system loop
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
endif
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
enddo ! deltaZ loop
enddo ! deltaY loop
enddo ! deltaX loop
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
!* map the stress from the neighboring MP's lattice configuration into the deformed configuration
!* and back into my lattice configuration
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
neighboringLattice2myLattice = math_mul33x33(invFe, Fe(1:3,1:3,1,neighboring_ip,neighboring_el))
constitutive_nonlocal_dislocationstress = constitutive_nonlocal_dislocationstress &
+ math_mul33x33(neighboringLattice2myLattice, &
math_mul33x33(Tdislo_neighboringLattice, &
added new math function "math_deviatoric33" which returns the deviatoric part of a 3x3 tensor renamed some math functions, so that we have a universal naming scheme: for matrix multiplications use an "x" (e.g. math_mul33x3); don't use the "x" to describe the shape of the tensor that the function is applied to (e.g. math_invert33 instead of math_invert3x3)
2012-01-26 19:20:00 +05:30
math_transpose33(neighboringLattice2myLattice)))
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
enddo ipLoop
enddo ! element loop
endif
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
endfunction
!*********************************************************************
!* return array of constitutive results *
!*********************************************************************
* removed calculations for dipole formation/dissociation by stress change, since it is not used anyways; also removed associated constitutive outputs from material.config * removed input variables in constitutive_collectDotState and constitutive_postResults that are not needed anymore (because of recent changes in constitutive_nonlocal)
2011-02-25 15:23:20 +05:30
function constitutive_nonlocal_postResults(Tstar_v, Fe, Temperature, dt, state, dotState, g,ip,el)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
added new math function "math_deviatoric33" which returns the deviatoric part of a 3x3 tensor renamed some math functions, so that we have a universal naming scheme: for matrix multiplications use an "x" (e.g. math_mul33x3); don't use the "x" to describe the shape of the tensor that the function is applied to (e.g. math_invert33 instead of math_invert3x3)
2012-01-26 19:20:00 +05:30
use math, only: math_mul6x6, &
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
math_mul33x3, &
math_mul33x33, &
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
pi
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
use mesh, only: mesh_NcpElems, &
introduced factor to control edge contribution to multiplication; dislocation density below a single dislocation per IP considered not significant
2012-09-04 22:26:37 +05:30
mesh_maxNips, &
mesh_ipVolume
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
use material, only: homogenization_maxNgrains, &
material_phase, &
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
phase_plasticityInstance, &
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
phase_Noutput
mixed things up, did not want to commit constitutive_nonlocal.f90 but rather DAMASK_spectral.f90!!!! did a lot of polishing on constitutive_nonlocal.f90 (mostly checked in at last commit), for changes on DAMASK_spectral.f90 see previous message
2012-02-23 22:50:57 +05:30
use lattice, only: lattice_Sslip_v, &
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
lattice_sd, &
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
lattice_st, &
lattice_sn
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
implicit none
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
!*** input variables
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
integer(pInt), intent(in) :: g, & ! current grain number
ip, & ! current integration point
el ! current element number
real(pReal), intent(in) :: Temperature, & ! temperature
* removed calculations for dipole formation/dissociation by stress change, since it is not used anyways; also removed associated constitutive outputs from material.config * removed input variables in constitutive_collectDotState and constitutive_postResults that are not needed anymore (because of recent changes in constitutive_nonlocal)
2011-02-25 15:23:20 +05:30
dt ! time increment
real(pReal), dimension(6), intent(in) :: Tstar_v ! current 2nd Piola-Kirchhoff stress in Mandel notation
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
real(pReal), dimension(3,3,homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), intent(in) :: &
Fe ! elastic deformation gradient
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
type(p_vec), dimension(homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), intent(in) :: &
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
state ! current microstructural state
type(p_vec), intent(in) :: dotState ! evolution rate of microstructural state
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
!*** output variables
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
real(pReal), dimension(constitutive_nonlocal_sizePostResults(phase_plasticityInstance(material_phase(g,ip,el)))) :: &
new subroutine crystallite_orientations: calculates the orientation matrix and the euler angles and - in case of single grain ips - the misorientation for each neighboring ip. The calculation of the misorientation is now done once in crystallite init and at the end of every FE increment. This saves a lot of time compared to doing it in dotState for every crystallite subinc.
2009-12-18 21:16:33 +05:30
constitutive_nonlocal_postResults
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
!*** local variables
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
integer(pInt) myInstance, & ! current instance of this plasticity
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
myStructure, & ! current lattice structure
ns, & ! short notation for the total number of active slip systems
c, & ! character of dislocation
cs, & ! constitutive result index
o, & ! index of current output
t, & ! type of dislocation
s, & ! index of my current slip system
sLattice ! index of my current slip system according to lattice order
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),8) :: &
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
rhoSgl, & ! current single dislocation densities (positive/negative screw and edge without dipoles)
rhoDotSgl ! evolution rate of single dislocation densities (positive/negative screw and edge without dipoles)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),4) :: &
constitutive_nonlocal: dislocation velocities are stored in the state, so we actually now have three "parts" of the state, the basic states that are updated by "constitutive_dotState" come first, then the dependent states that are calculated by "constitutive_microstructure" follow, and finally we have a last part reserved for other variables that just use the memory reserved by the state array and are updated somewhere else. constitutive: LpAndItsTangent does not need the full state, but only the local state, so changed that at least for the nonlocal constitutive law
2011-11-04 18:42:17 +05:30
gdot, & ! shear rates
v ! velocities
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el)))) :: &
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
rhoForest, & ! forest dislocation density
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
tauThreshold, & ! threshold shear stress
tau, & ! current resolved shear stress
removed some unused variables
2012-11-30 00:14:00 +05:30
tauBack ! back stress from pileups on same slip system
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),2) :: &
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
rhoDip, & ! current dipole dislocation densities (screw and edge dipoles)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
rhoDotDip, & ! evolution rate of dipole dislocation densities (screw and edge dipoles)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
dLower, & ! minimum stable dipole distance for edges and screws
* removed calculations for dipole formation/dissociation by stress change, since it is not used anyways; also removed associated constitutive outputs from material.config * removed input variables in constitutive_collectDotState and constitutive_postResults that are not needed anymore (because of recent changes in constitutive_nonlocal)
2011-02-25 15:23:20 +05:30
dUpper ! current maximum stable dipole distance for edges and screws
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(3,totalNslip(phase_plasticityInstance(material_phase(g,ip,el))),2) :: &
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
m, & ! direction of dislocation motion for edge and screw (unit vector)
m_currentconf ! direction of dislocation motion for edge and screw (unit vector) in current configuration
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
real(pReal), dimension(3,totalNslip(phase_plasticityInstance(material_phase(g,ip,el)))) :: &
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
n_currentconf ! slip system normal (unit vector) in current configuration
* new output variable "internalstress", which gives the internal stress tensor * use "math_invert3x3" instead of "math_inv3x3" for inversion of Fe * for dislocation stress calculation: first regular case, then special case of dead dislocations in central ip * "dv_dtau" now given for each dislocation type, so is a (ns,4) array * deleted unused variables in "_LpAndItsTangent" * corrected contribution of deads in "_LpAndItsTangent" * the NaN variables defined in math did not give a proper NaN value, so use 0.0/0.0 again * neighbors with nonlocal constitution but local properties (i.e. /nonlocal/ flag not set) are also considered for incoming fluxes
2011-09-07 17:00:28 +05:30
real(pReal), dimension(3,3) :: sigma
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
replaced keyword constitution with plasticity in preparation of upcomming generalized elasticity
2012-03-12 19:39:37 +05:30
myInstance = phase_plasticityInstance(material_phase(g,ip,el))
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
myStructure = constitutive_nonlocal_structure(myInstance)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
ns = totalNslip(myInstance)
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = 0_pInt
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
constitutive_nonlocal_postResults = 0.0_pReal
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
- corrected contribution of immobile singles to dipole formation and annihilation rates - remobilization of immobile singles immediately increases the mobile single dislocation density and therefore directly affects all other mechanisms in dotState
2010-01-06 15:24:00 +05:30
!* short hand notations for state variables
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
forall (s = 1_pInt:ns, t = 1_pInt:4_pInt) &
rhoSgl(s,t) = max(state(g,ip,el)%p((t-1_pInt)*ns+s), 0.0_pReal)
forall (s = 1_pInt:ns, t = 5_pInt:8_pInt) &
rhoSgl(s,t) = state(g,ip,el)%p((t-1_pInt)*ns+s)
forall (c = 1_pInt:2_pInt) &
rhoDip(1:ns,c) = max(state(g,ip,el)%p((7_pInt+c)*ns+1_pInt:(8_pInt+c)*ns), 0.0_pReal)
improved accuracy of "accumulated shear": now added as a state variable, which facilitates integration (dotstate=shearrate); needs absolute tolerance value for state integration in material.config
2013-04-04 19:07:14 +05:30
rhoForest = state(g,ip,el)%p(11_pInt*ns+1:12_pInt*ns)
tauThreshold = state(g,ip,el)%p(12_pInt*ns+1:13_pInt*ns)
tauBack = state(g,ip,el)%p(13_pInt*ns+1:14_pInt*ns)
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
forall (t = 1_pInt:8_pInt) rhoDotSgl(1:ns,t) = dotState%p((t-1_pInt)*ns+1_pInt:t*ns)
forall (c = 1_pInt:2_pInt) rhoDotDip(1:ns,c) = dotState%p((7_pInt+c)*ns+1_pInt:(8_pInt+c)*ns)
improved accuracy of "accumulated shear": now added as a state variable, which facilitates integration (dotstate=shearrate); needs absolute tolerance value for state integration in material.config
2013-04-04 19:07:14 +05:30
forall (t = 1_pInt:4_pInt) v(1:ns,t) = state(g,ip,el)%p((13_pInt+t)*ns+1_pInt:(14_pInt+t)*ns)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
where (abs(rhoSgl) * mesh_ipVolume(ip,el) ** 0.667_pReal < significantN(myInstance) &
.or. abs(rhoSgl) < significantRho(myInstance)) &
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
rhoSgl = 0.0_pReal
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
where (abs(rhoDip) * mesh_ipVolume(ip,el) ** 0.667_pReal < significantN(myInstance) &
.or. abs(rhoDip) < significantRho(myInstance)) &
added possibility to start with an initial random distribution of dislocation segments with specific overall density
2012-10-02 18:27:24 +05:30
rhoDip = 0.0_pReal
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
- corrected contribution of immobile singles to dipole formation and annihilation rates - remobilization of immobile singles immediately increases the mobile single dislocation density and therefore directly affects all other mechanisms in dotState
2010-01-06 15:24:00 +05:30
!* Calculate shear rate
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
forall (t = 1_pInt:4_pInt) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
gdot(1:ns,t) = rhoSgl(1:ns,t) * burgers(1:ns,myInstance) * v(1:ns,t)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
constitutive_nonlocal: dislocation velocities are stored in the state, so we actually now have three "parts" of the state, the basic states that are updated by "constitutive_dotState" come first, then the dependent states that are calculated by "constitutive_microstructure" follow, and finally we have a last part reserved for other variables that just use the memory reserved by the state array and are updated somewhere else. constitutive: LpAndItsTangent does not need the full state, but only the local state, so changed that at least for the nonlocal constitutive law
2011-11-04 18:42:17 +05:30
* removed calculations for dipole formation/dissociation by stress change, since it is not used anyways; also removed associated constitutive outputs from material.config * removed input variables in constitutive_collectDotState and constitutive_postResults that are not needed anymore (because of recent changes in constitutive_nonlocal)
2011-02-25 15:23:20 +05:30
!* calculate limits for stable dipole height
- corrected contribution of immobile singles to dipole formation and annihilation rates - remobilization of immobile singles immediately increases the mobile single dislocation density and therefore directly affects all other mechanisms in dotState
2010-01-06 15:24:00 +05:30
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do s = 1_pInt,ns
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sLattice = slipSystemLattice(s,myInstance)
added code structure for non-schmid mechanics. work in progress…
2013-01-22 04:41:16 +05:30
tau(s) = math_mul6x6(Tstar_v, lattice_Sslip_v(1:6,1,sLattice,myStructure)) + tauBack(s)
corrected limit for upper stable dipole distance; fixed a bug in the postResults where division by zero might happen
2012-10-29 18:32:01 +05:30
if (abs(tau(s)) < 1.0e-15_pReal) tau(s) = 1.0e-15_pReal
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
enddo
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
dLower = minDipoleHeight(1:ns,1:2,myInstance)
dUpper(1:ns,1) = mu(myInstance) * burgers(1:ns,myInstance) &
/ (8.0_pReal * pi * (1.0_pReal - nu(myInstance)) * abs(tau))
dUpper(1:ns,2) = mu(myInstance) * burgers(1:ns,myInstance) &
corrected formulation of maximum height of screw dislocation dipoles
2012-07-24 20:20:11 +05:30
/ (4.0_pReal * pi * abs(tau))
forall (c = 1_pInt:2_pInt) &
corrected limit for upper stable dipole distance; fixed a bug in the postResults where division by zero might happen
2012-10-29 18:32:01 +05:30
dUpper(1:ns,c) = min(1.0_pReal / sqrt(rhoSgl(1:ns,2*c-1) + rhoSgl(1:ns,2*c) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
+ abs(rhoSgl(1:ns,2*c+3)) + abs(rhoSgl(1:ns,2*c+4)) + rhoDip(1:ns,c)), &
corrected limit for upper stable dipole distance; fixed a bug in the postResults where division by zero might happen
2012-10-29 18:32:01 +05:30
dUpper(1:ns,c))
rates in constitutive_results are taken directly from last converged step and not calculated again; added output of rates for annihilation of screws and deposition of edge jogs on collinear system
2012-08-16 16:33:22 +05:30
dUpper = max(dUpper,dLower)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
!*** dislocation motion
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
m(1:3,1:ns,1) = lattice_sd(1:3,slipSystemLattice(1:ns,myInstance),myStructure)
m(1:3,1:ns,2) = -lattice_st(1:3,slipSystemLattice(1:ns,myInstance),myStructure)
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
forall (c = 1_pInt:2_pInt, s = 1_pInt:ns) &
bug fix for calculation of fluxdensity and slipnormal/direction in constitutive_nonlocal_postResults: indices for Fe were missing
2013-05-23 20:22:57 +05:30
m_currentconf(1:3,s,c) = math_mul33x3(Fe(1:3,1:3,g,ip,el), m(1:3,s,c))
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
forall (s = 1_pInt:ns) &
bug fix for calculation of fluxdensity and slipnormal/direction in constitutive_nonlocal_postResults: indices for Fe were missing
2013-05-23 20:22:57 +05:30
n_currentconf(1:3,s) = math_mul33x3(Fe(1:3,1:3,g,ip,el), &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
lattice_sn(1:3,slipSystemLattice(s,myInstance),myStructure))
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do o = 1_pInt,phase_Noutput(material_phase(g,ip,el))
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
select case(constitutive_nonlocal_output(o,myInstance))
case ('rho')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(abs(rhoSgl),2) + sum(rhoDip,2)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
case ('rho_sgl')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(abs(rhoSgl),2)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_mobile')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(abs(rhoSgl(1:ns,1:4)),2)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_immobile')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoSgl(1:ns,5:8),2)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_dip')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoDip,2)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
case ('rho_edge')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(abs(rhoSgl(1:ns,(/1,2,5,6/))),2) + rhoDip(1:ns,1)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_edge')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(abs(rhoSgl(1:ns,(/1,2,5,6/))),2)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_edge_mobile')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoSgl(1:ns,1:2),2)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
case ('rho_sgl_edge_immobile')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoSgl(1:ns,5:6),2)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
case ('rho_sgl_edge_pos')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(1:ns,1) + abs(rhoSgl(1:ns,5))
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_edge_pos_mobile')
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = state(g,ip,el)%p(1:ns)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_edge_pos_immobile')
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = state(g,ip,el)%p(4*ns+1:5*ns)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_edge_neg')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(1:ns,2) + abs(rhoSgl(1:ns,6))
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_edge_neg_mobile')
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = state(g,ip,el)%p(ns+1:2*ns)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_edge_neg_immobile')
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = state(g,ip,el)%p(5*ns+1:6*ns)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_dip_edge')
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = state(g,ip,el)%p(8*ns+1:9*ns)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
case ('rho_screw')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(abs(rhoSgl(1:ns,(/3,4,7,8/))),2) + rhoDip(1:ns,2)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_screw')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(abs(rhoSgl(1:ns,(/3,4,7,8/))),2)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_screw_mobile')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoSgl(1:ns,3:4),2)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_screw_immobile')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoSgl(1:ns,7:8),2)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_screw_pos')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(1:ns,3) + abs(rhoSgl(1:ns,7))
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
case ('rho_sgl_screw_pos_mobile')
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = state(g,ip,el)%p(2*ns+1:3*ns)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
case ('rho_sgl_screw_pos_immobile')
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = state(g,ip,el)%p(6*ns+1:7*ns)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_screw_neg')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(1:ns,4) + abs(rhoSgl(1:ns,8))
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_screw_neg_mobile')
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = state(g,ip,el)%p(3*ns+1:4*ns)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_sgl_screw_neg_immobile')
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = state(g,ip,el)%p(7*ns+1:8*ns)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_dip_screw')
new parameter "significantRho" for nonlocal constitutive law ; density below this value will hardly contribute to dislocation glide or any dislocation reaction ; old parameter "absoluteToleranceRho" is now used only for its initial purpose, namely the absolute tolerance in the state integration
2012-08-23 11:18:21 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = state(g,ip,el)%p(9*ns+1:10*ns)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
case ('excess_rho')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = (rhoSgl(1:ns,1) + abs(rhoSgl(1:ns,5))) &
- (rhoSgl(1:ns,2) + abs(rhoSgl(1:ns,6))) &
+ (rhoSgl(1:ns,3) + abs(rhoSgl(1:ns,7))) &
- (rhoSgl(1:ns,4) + abs(rhoSgl(1:ns,8)))
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
in nonlocal_microstructure: at the model surface the excess density is now extrapolated from opposite neighbor instead of assuming zero density; this results in better modeling of the dislocation stress at the surface restructured nonlocal_dotstate, to be able to easily switch on and off particular effects in the microstructure evolution nonlocal_dotstate now enforces a cutback when single density runs the risk of becoming negative; in the case of a state already below the relevantState dotState is set to zero introduced two new output variables: rho_dot_edge and rho_dot_screw
2010-05-21 14:21:15 +05:30
case ('excess_rho_edge')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = (rhoSgl(1:ns,1) + abs(rhoSgl(1:ns,5))) &
- (rhoSgl(1:ns,2) + abs(rhoSgl(1:ns,6)))
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
in nonlocal_microstructure: at the model surface the excess density is now extrapolated from opposite neighbor instead of assuming zero density; this results in better modeling of the dislocation stress at the surface restructured nonlocal_dotstate, to be able to easily switch on and off particular effects in the microstructure evolution nonlocal_dotstate now enforces a cutback when single density runs the risk of becoming negative; in the case of a state already below the relevantState dotState is set to zero introduced two new output variables: rho_dot_edge and rho_dot_screw
2010-05-21 14:21:15 +05:30
case ('excess_rho_screw')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = (rhoSgl(1:ns,3) + abs(rhoSgl(1:ns,7))) &
- (rhoSgl(1:ns,4) + abs(rhoSgl(1:ns,8)))
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
case ('rho_forest')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoForest
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
case ('delta')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = 1.0_pReal / sqrt(sum(abs(rhoSgl),2) + sum(rhoDip,2))
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
case ('delta_sgl')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = 1.0_pReal / sqrt(sum(abs(rhoSgl),2))
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
case ('delta_dip')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = 1.0_pReal / sqrt(sum(rhoDip,2))
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
case ('shearrate')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(gdot,2)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
case ('resolvedstress')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = tau
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
case ('resolvedstress_back')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = tauBack
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
cs = cs + ns
case ('resolvedstress_external')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
do s = 1_pInt,ns
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sLattice = slipSystemLattice(s,myInstance)
added code structure for non-schmid mechanics. work in progress…
2013-01-22 04:41:16 +05:30
constitutive_nonlocal_postResults(cs+s) = math_mul6x6(Tstar_v, lattice_Sslip_v(1:6,1,sLattice,myStructure))
constitutive_nonlocal: - reworked contribution of immobile dislocation density for rate equations - flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState - dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables - better use flux density vector as output variable instead of scalar flux values for each interface - added output variables internal and external resolved stress crystallite: - added flag to force local stiffness calculation in case of nonlocal model - misorientation angle is explicitly set to zero when no neighbor can be found debug: - added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g - debugger can now be used in its original sense
2010-02-17 18:51:36 +05:30
enddo
cs = cs + ns
constitutive_nonlocal.f90 - completed postResults output function - connecting vector of neighboring material points is mapped to intermediate configuration of my neighbor crystallite.f90 - zero out dotState only when crystallite is non-finished - set nonfinished flag to false if crystallite is not on Track after state update - in updateState: set onTrack flag to false if encounter NaN - removed some old debugging outputs and added others homogenization.f90 - in debugging mode now telling when a cutback happens
2009-08-24 13:46:01 +05:30
case ('resistance')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = tauThreshold
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
case ('rho_dot')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoDotSgl,2) + sum(rhoDotDip,2)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_dot_sgl')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoDotSgl,2)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
case ('rho_dot_dip')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoDotDip,2)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
case ('rho_dot_gen')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotMultiplicationOutput(1:ns,1,g,ip,el) &
+ rhoDotMultiplicationOutput(1:ns,2,g,ip,el)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_dot_gen_edge')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotMultiplicationOutput(1:ns,1,g,ip,el)
- introduction of immobile (used) single dislocation densities as new state variables - changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density - changed material.config accordingly
2010-01-05 21:37:24 +05:30
cs = cs + ns
case ('rho_dot_gen_screw')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotMultiplicationOutput(1:ns,2,g,ip,el)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
case ('rho_dot_sgl2dip')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotSingle2DipoleGlideOutput(1:ns,1,g,ip,el) &
+ rhoDotSingle2DipoleGlideOutput(1:ns,2,g,ip,el)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
added separate output of dipole formation rate for edge and screw
2012-08-27 21:27:31 +05:30
case ('rho_dot_sgl2dip_edge')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotSingle2DipoleGlideOutput(1:ns,1,g,ip,el)
added separate output of dipole formation rate for edge and screw
2012-08-27 21:27:31 +05:30
cs = cs + ns
case ('rho_dot_sgl2dip_screw')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotSingle2DipoleGlideOutput(1:ns,2,g,ip,el)
added separate output of dipole formation rate for edge and screw
2012-08-27 21:27:31 +05:30
cs = cs + ns
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
case ('rho_dot_ann_ath')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotAthermalAnnihilationOutput(1:ns,1,g,ip,el) &
+ rhoDotAthermalAnnihilationOutput(1:ns,2,g,ip,el)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
case ('rho_dot_ann_the')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotThermalAnnihilationOutput(1:ns,1,g,ip,el) &
+ rhoDotThermalAnnihilationOutput(1:ns,2,g,ip,el)
rates in constitutive_results are taken directly from last converged step and not calculated again; added output of rates for annihilation of screws and deposition of edge jogs on collinear system
2012-08-16 16:33:22 +05:30
cs = cs + ns
case ('rho_dot_ann_the_edge')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotThermalAnnihilationOutput(1:ns,1,g,ip,el)
rates in constitutive_results are taken directly from last converged step and not calculated again; added output of rates for annihilation of screws and deposition of edge jogs on collinear system
2012-08-16 16:33:22 +05:30
cs = cs + ns
case ('rho_dot_ann_the_screw')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotThermalAnnihilationOutput(1:ns,2,g,ip,el)
rates in constitutive_results are taken directly from last converged step and not calculated again; added output of rates for annihilation of screws and deposition of edge jogs on collinear system
2012-08-16 16:33:22 +05:30
cs = cs + ns
annihilate only part of the screw dipoles (specified by minimumDipoleHeight), not all; moved annihilation of screws from deltaState back to dotState
2012-11-28 17:39:48 +05:30
case ('rho_dot_edgejogs')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotEdgeJogsOutput(1:ns,g,ip,el)
annihilate only part of the screw dipoles (specified by minimumDipoleHeight), not all; moved annihilation of screws from deltaState back to dotState
2012-11-28 17:39:48 +05:30
cs = cs + ns
added output for dislocation evolution rate associated with flux: "rho_dot_flux"
2010-03-10 15:19:40 +05:30
case ('rho_dot_flux')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoDotFluxOutput(1:ns,1:4,g,ip,el),2) &
+ sum(abs(rhoDotFluxOutput(1:ns,5:8,g,ip,el)),2)
added output for dislocation evolution rate associated with flux: "rho_dot_flux"
2010-03-10 15:19:40 +05:30
cs = cs + ns
in nonlocal_microstructure: at the model surface the excess density is now extrapolated from opposite neighbor instead of assuming zero density; this results in better modeling of the dislocation stress at the surface restructured nonlocal_dotstate, to be able to easily switch on and off particular effects in the microstructure evolution nonlocal_dotstate now enforces a cutback when single density runs the risk of becoming negative; in the case of a state already below the relevantState dotState is set to zero introduced two new output variables: rho_dot_edge and rho_dot_screw
2010-05-21 14:21:15 +05:30
case ('rho_dot_flux_edge')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoDotFluxOutput(1:ns,1:2,g,ip,el),2) &
+ sum(abs(rhoDotFluxOutput(1:ns,5:6,g,ip,el)),2)
in nonlocal_microstructure: at the model surface the excess density is now extrapolated from opposite neighbor instead of assuming zero density; this results in better modeling of the dislocation stress at the surface restructured nonlocal_dotstate, to be able to easily switch on and off particular effects in the microstructure evolution nonlocal_dotstate now enforces a cutback when single density runs the risk of becoming negative; in the case of a state already below the relevantState dotState is set to zero introduced two new output variables: rho_dot_edge and rho_dot_screw
2010-05-21 14:21:15 +05:30
cs = cs + ns
added output for dislocation evolution rate associated with flux: "rho_dot_flux"
2010-03-10 15:19:40 +05:30
in nonlocal_microstructure: at the model surface the excess density is now extrapolated from opposite neighbor instead of assuming zero density; this results in better modeling of the dislocation stress at the surface restructured nonlocal_dotstate, to be able to easily switch on and off particular effects in the microstructure evolution nonlocal_dotstate now enforces a cutback when single density runs the risk of becoming negative; in the case of a state already below the relevantState dotState is set to zero introduced two new output variables: rho_dot_edge and rho_dot_screw
2010-05-21 14:21:15 +05:30
case ('rho_dot_flux_screw')
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoDotFluxOutput(1:ns,3:4,g,ip,el),2) &
+ sum(abs(rhoDotFluxOutput(1:ns,7:8,g,ip,el)),2)
in nonlocal_microstructure: at the model surface the excess density is now extrapolated from opposite neighbor instead of assuming zero density; this results in better modeling of the dislocation stress at the surface restructured nonlocal_dotstate, to be able to easily switch on and off particular effects in the microstructure evolution nonlocal_dotstate now enforces a cutback when single density runs the risk of becoming negative; in the case of a state already below the relevantState dotState is set to zero introduced two new output variables: rho_dot_edge and rho_dot_screw
2010-05-21 14:21:15 +05:30
cs = cs + ns
dislocation velocity output for each dislocation type (edge+,edge-,screw+,screw-) for fcc constitutive_nonlocal_kinetics is called only once and the result is used for each dislocation type, for all other lattice structures we have one call for each dislocation type
2012-01-26 13:13:36 +05:30
case ('velocity_edge_pos')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = v(1:ns,1)
new output "dislocationvelocity"
2010-02-23 22:53:07 +05:30
cs = cs + ns
dislocation velocity output for each dislocation type (edge+,edge-,screw+,screw-) for fcc constitutive_nonlocal_kinetics is called only once and the result is used for each dislocation type, for all other lattice structures we have one call for each dislocation type
2012-01-26 13:13:36 +05:30
case ('velocity_edge_neg')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = v(1:ns,2)
dislocation velocity output for each dislocation type (edge+,edge-,screw+,screw-) for fcc constitutive_nonlocal_kinetics is called only once and the result is used for each dislocation type, for all other lattice structures we have one call for each dislocation type
2012-01-26 13:13:36 +05:30
cs = cs + ns
case ('velocity_screw_pos')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = v(1:ns,3)
new outputs "accumulatedshear", "velocity_edge", and "velocity_screw" (the latter two replace the output "dislocationvelocity")
2011-11-09 14:52:52 +05:30
cs = cs + ns
dislocation velocity output for each dislocation type (edge+,edge-,screw+,screw-) for fcc constitutive_nonlocal_kinetics is called only once and the result is used for each dislocation type, for all other lattice structures we have one call for each dislocation type
2012-01-26 13:13:36 +05:30
case ('velocity_screw_neg')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = v(1:ns,4)
dislocation velocity output for each dislocation type (edge+,edge-,screw+,screw-) for fcc constitutive_nonlocal_kinetics is called only once and the result is used for each dislocation type, for all other lattice structures we have one call for each dislocation type
2012-01-26 13:13:36 +05:30
cs = cs + ns
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
case ('slipdirection.x')
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = m_currentconf(1,1:ns,1)
cs = cs + ns
case ('slipdirection.y')
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = m_currentconf(2,1:ns,1)
cs = cs + ns
case ('slipdirection.z')
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = m_currentconf(3,1:ns,1)
cs = cs + ns
case ('slipnormal.x')
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = n_currentconf(1,1:ns)
cs = cs + ns
case ('slipnormal.y')
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = n_currentconf(2,1:ns)
cs = cs + ns
case ('slipnormal.z')
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = n_currentconf(3,1:ns)
cs = cs + ns
case ('fluxdensity_edge_pos.x')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(1:ns,1) * v(1:ns,1) * m_currentconf(1,1:ns,1)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
case ('fluxdensity_edge_pos.y')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(1:ns,1) * v(1:ns,1) * m_currentconf(2,1:ns,1)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
case ('fluxdensity_edge_pos.z')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(1:ns,1) * v(1:ns,1) * m_currentconf(3,1:ns,1)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
case ('fluxdensity_edge_neg.x')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = - rhoSgl(1:ns,2) * v(1:ns,2) * m_currentconf(1,1:ns,1)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
case ('fluxdensity_edge_neg.y')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = - rhoSgl(1:ns,2) * v(1:ns,2) * m_currentconf(2,1:ns,1)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
case ('fluxdensity_edge_neg.z')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = - rhoSgl(1:ns,2) * v(1:ns,2) * m_currentconf(3,1:ns,1)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
case ('fluxdensity_screw_pos.x')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(1:ns,3) * v(1:ns,3) * m_currentconf(1,1:ns,2)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
case ('fluxdensity_screw_pos.y')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(1:ns,3) * v(1:ns,3) * m_currentconf(2,1:ns,2)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
case ('fluxdensity_screw_pos.z')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(1:ns,3) * v(1:ns,3) * m_currentconf(3,1:ns,2)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
case ('fluxdensity_screw_neg.x')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = - rhoSgl(1:ns,4) * v(1:ns,4) * m_currentconf(1,1:ns,2)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
case ('fluxdensity_screw_neg.y')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = - rhoSgl(1:ns,4) * v(1:ns,4) * m_currentconf(2,1:ns,2)
constitutive_nonlocal: - corrected flux term - multiplication is now aware of dislocation type - corrected change rate for "dipole size" dupper - corrected term for dipole dissociation by stress change - added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle) - added more output variables constitutive: - 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration) - timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore) crystallite: - convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge - need call to microstructure before first call to collect dotState for dependent states - stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step. - updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState IO: - additional warning message for unknown crystal symmetry
2009-12-15 13:50:31 +05:30
cs = cs + ns
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
case ('fluxdensity_screw_neg.z')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = - rhoSgl(1:ns,4) * v(1:ns,4) * m_currentconf(3,1:ns,2)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
case ('maximumdipoleheight_edge')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = dUpper(1:ns,1)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
Reworked kinetics. Now there is a universal law which can be used for both fcc and bcc (and probably even hcp). It mainly consists of an athermal forest cutting part , solid solution and peierls as thermally activated processes, and viscous glide. Therefore, lots of new parameters in the material.config. Not sure yet if the values make sense though. Also renamed some of the old parameters. Unknown constitutive output raises error.
2012-01-25 22:34:37 +05:30
case ('maximumdipoleheight_screw')
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = dUpper(1:ns,2)
constitutive_nonlocal: - read in activation energy for dislocation glide from material.config - changed naming of dDipMin/Max to dLower/dUpper - added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw - poisson's ratio is now calculated from elastic constants - microstrucutre has state as first argument, since this is our output variable - periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal. - corrected calculation for dipole formation by glide - added terms for dipole formation/annihilation by stress decrease/increase constitutive: - passing of arguments is adapted for constitutive_nonlocal model crystallite: - in stiffness calculation: call to collect_dotState used wrong arguments - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults homogenization: - crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults IO: - changed error message 229 material.config: - changed example for nonlocal constitution according to constitutive_nonlocal all: - added some flush statements
2009-10-20 20:06:03 +05:30
cs = cs + ns
* new output variable "internalstress", which gives the internal stress tensor * use "math_invert3x3" instead of "math_inv3x3" for inversion of Fe * for dislocation stress calculation: first regular case, then special case of dead dislocations in central ip * "dv_dtau" now given for each dislocation type, so is a (ns,4) array * deleted unused variables in "_LpAndItsTangent" * corrected contribution of deads in "_LpAndItsTangent" * the NaN variables defined in math did not give a proper NaN value, so use 0.0/0.0 again * neighbors with nonlocal constitution but local properties (i.e. /nonlocal/ flag not set) are also considered for incoming fluxes
2011-09-07 17:00:28 +05:30
dislocation stress does not add to stress field anymore, but is only available as constitutive output for purposes of postprocessing. instead, we now use a backstress term (which is added to the resolved shear stress) that depends on the gradient of excess density on the same slip system.
2012-01-17 15:56:57 +05:30
case('dislocationstress')
sigma = constitutive_nonlocal_dislocationstress(state, Fe, g, ip, el)
introduced itmin parameter for spectral code for defining minimum number of cycles removed simplified_algorthim flag because the basic scheme using the polarization field will not be implemented introduced divergence_correction flag for making divergence criterion resolution-independent (still experimental and not set by default) corrected output and restart frequency (now modulo on incs of current load case)
2012-02-23 22:13:17 +05:30
constitutive_nonlocal_postResults(cs+1_pInt) = sigma(1,1)
constitutive_nonlocal_postResults(cs+2_pInt) = sigma(2,2)
constitutive_nonlocal_postResults(cs+3_pInt) = sigma(3,3)
constitutive_nonlocal_postResults(cs+4_pInt) = sigma(1,2)
constitutive_nonlocal_postResults(cs+5_pInt) = sigma(2,3)
constitutive_nonlocal_postResults(cs+6_pInt) = sigma(3,1)
* new output variable "internalstress", which gives the internal stress tensor * use "math_invert3x3" instead of "math_inv3x3" for inversion of Fe * for dislocation stress calculation: first regular case, then special case of dead dislocations in central ip * "dv_dtau" now given for each dislocation type, so is a (ns,4) array * deleted unused variables in "_LpAndItsTangent" * corrected contribution of deads in "_LpAndItsTangent" * the NaN variables defined in math did not give a proper NaN value, so use 0.0/0.0 again * neighbors with nonlocal constitution but local properties (i.e. /nonlocal/ flag not set) are also considered for incoming fluxes
2011-09-07 17:00:28 +05:30
cs = cs + 6_pInt
new outputs "accumulatedshear", "velocity_edge", and "velocity_screw" (the latter two replace the output "dislocationvelocity")
2011-11-09 14:52:52 +05:30
case('accumulatedshear')
improved accuracy of "accumulated shear": now added as a state variable, which facilitates integration (dotstate=shearrate); needs absolute tolerance value for state integration in material.config
2013-04-04 19:07:14 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = state(g,ip,el)%p(10*ns+1:11*ns)
new outputs "accumulatedshear", "velocity_edge", and "velocity_screw" (the latter two replace the output "dislocationvelocity")
2011-11-09 14:52:52 +05:30
cs = cs + ns
added possibility to scale the plastic shearrate by the ratio of mobile to total density; keyword in material.config is "deadZone"
2012-10-29 18:19:28 +05:30
case('boundarylayer')
do s = 1_pInt,ns
if (sum(abs(rhoSgl(s,1:8))) > 0.0_pReal) then
constitutive_nonlocal_postResults(cs+s) = maxval(abs(rhoSgl(s,5:8))/(rhoSgl(s,1:4)+abs(rhoSgl(s,5:8))))
else
constitutive_nonlocal_postResults(cs+s) = 0.0_pReal
endif
enddo
cs = cs + ns
new outputs "accumulatedshear", "velocity_edge", and "velocity_screw" (the latter two replace the output "dislocationvelocity")
2011-11-09 14:52:52 +05:30
now with first draft of nonlocal constitutive law debugging memory leak closed debugging counters corrected center of gravity stored in mesh state updated is now split into a collecting loop and an execution updateState and updateTemperature fill sequentially separate logicals and evaluate afterwards to converged added 3x3 transposition function, norm for 3x1 matrix and 33x3 matrix multiplication in math non-converged crystallite triggers materialpoint cutback (used to respond elastically) non-converged materialpoint raises terminal illness which in turn renders whole FE increment useless by means of odd stress/stiffness and thus waits for FE cutback
2009-08-11 22:01:57 +05:30
end select
enddo
endfunction
openmp parallelization working again (at least for j2 and nonlocal constitutive model). In order to keep it like that, please follow these simple rules: DON'T use implicit array subscripts: example: real, dimension(3,3) :: A,B A(:,2) = B(:,1) <--- DON'T USE A(1:3,2) = B(1:3,1) <--- BETTER USE In many cases the use of explicit array subscripts is inevitable for parallelization. Additionally, it is an easy means to prevent memory leaks. Enclose all write statements with the following: !$OMP CRITICAL (write2out) <your write statement> !$OMP END CRITICAL (write2out) Whenever you change something in the code and are not sure if it affects parallelization and leads to nonconforming behavior, please ask me and/or Franz to check this.
2011-03-17 16:16:17 +05:30
constitutive_nonlocal: - dipole dislocations with evolution crystallite.f90: - collect state uses subdt and subTstar0_v - in nonlocal modus: set all crystallites to broken if one is not on track anymore after either stress integration or state update - constitutive_microstructure is now called inside state update and not in integrate_stress anymore material.config: - new parameter for nonlocal constitution CPFEM.f90: - age Tstar after increment was finished
2009-08-28 19:20:47 +05:30
END MODULE