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

3810 lines
202 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/>.
!
introduced named constants for homogenization_type, phase_plasticitiy, and phase_elasticity to replace expensive string comparison. Labels for different schemes need to be stored in material.f90 because they are needed when parsing the config file
2013-11-27 13:34:05 +05:30
!--------------------------------------------------------------------------------------------------
! $Id$
!--------------------------------------------------------------------------------------------------
!> @author Christoph Kords, Max-Planck-Institut für Eisenforschung GmbH
!> @author Franz Roters, Max-Planck-Institut für Eisenforschung GmbH
!> @author Philip Eisenlohr, Max-Planck-Institut für Eisenforschung GmbH
!> @brief material subroutine for plasticity including dislocation flux
!--------------------------------------------------------------------------------------------------
module constitutive_nonlocal
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
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 temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +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 temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +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 temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +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
added doxygen documentation and unified variable names and some common parts of the code
2013-10-08 21:57:26 +05:30
integer(pInt), dimension(:), allocatable, public, protected :: &
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 :: &
iGamma, & !< state indices for accumulated shear
iRhoF, & !< state indices for forest density
combined some state indices to an array with a more generic name
2013-05-24 14:32:30 +05:30
iTauF, & !< state indices for critical resolved shear stress
iTauB !< state indices for backstress
integer(pInt), dimension(:,:,:), allocatable, private :: &
iRhoU, & !< state indices for unblocked density
iRhoB, & !< state indices for blocked density
iRhoD, & !< state indices for dipole density
iV, & !< state indices for dislcation velocities
iD !< state indices for stable dipole height
introduced named indices for state variables; however, not yet used
2013-05-24 01:45:23 +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_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
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
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, &
fixed bug in deltaState; used to lead to unpredictable values in accumulated shear, since deltaState was not defined at this state position
2013-06-12 02:33:23 +05:30
sourceProbability
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
rhoDotFluxOutput, &
rhoDotMultiplicationOutput, &
rhoDotSingle2DipoleGlideOutput, &
rhoDotAthermalAnnihilationOutput, &
nonSchmid behavior now ready to use for bcc with phenopowerlaw and nonlocal constitutive model. nonSchmid tensors according to Koester,Ma,Hartmaier,2012. extended lattice_Sslip with non-Schmid tensors (the full non-symmetric tensors are required for the tangent and cannot be retrieved from the symmetrized Mandel notation)
2013-08-05 14:53:21 +05:30
rhoDotThermalAnnihilationOutput, &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
nonSchmidProjection !< combined projection of Schmid and non-Schmid contributions to the resolved shear stress (only for screws)
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
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
enum, bind(c)
adapted changes from rev 2776 for constitutive_nonlocal: improved usage of "enum", fixed bug in file reading, usage of "source" specifier for "allocate"
2013-12-12 15:46:50 +05:30
enumerator :: undefined_ID, &
rho_ID, &
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
delta_ID, &
rho_edge_ID, &
rho_screw_ID, &
rho_sgl_ID, &
delta_sgl_ID, &
rho_sgl_edge_ID, &
rho_sgl_edge_pos_ID, &
rho_sgl_edge_neg_ID, &
rho_sgl_screw_ID, &
rho_sgl_screw_pos_ID, &
rho_sgl_screw_neg_ID, &
rho_sgl_mobile_ID, &
rho_sgl_edge_mobile_ID, &
rho_sgl_edge_pos_mobile_ID, &
rho_sgl_edge_neg_mobile_ID, &
rho_sgl_screw_mobile_ID, &
rho_sgl_screw_pos_mobile_ID, &
rho_sgl_screw_neg_mobile_ID, &
rho_sgl_immobile_ID, &
rho_sgl_edge_immobile_ID, &
rho_sgl_edge_pos_immobile_ID, &
rho_sgl_edge_neg_immobile_ID, &
rho_sgl_screw_immobile_ID, &
rho_sgl_screw_pos_immobile_ID, &
rho_sgl_screw_neg_immobile_ID, &
rho_dip_ID, &
delta_dip_ID, &
rho_dip_edge_ID, &
rho_dip_screw_ID, &
excess_rho_ID, &
excess_rho_edge_ID, &
excess_rho_screw_ID, &
rho_forest_ID, &
shearrate_ID, &
resolvedstress_ID, &
resolvedstress_external_ID, &
resolvedstress_back_ID, &
resistance_ID, &
rho_dot_ID, &
rho_dot_sgl_ID, &
added two new outputs: "rho_dot_flux_mobile" and "rho_dot_sgl_mobile" corrected outputs of dislocation density rates: take into account that blocked dislocation density is a signed quantity
2014-01-07 21:25:28 +05:30
rho_dot_sgl_mobile_ID, &
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
rho_dot_dip_ID, &
rho_dot_gen_ID, &
rho_dot_gen_edge_ID, &
rho_dot_gen_screw_ID, &
rho_dot_sgl2dip_ID, &
rho_dot_sgl2dip_edge_ID, &
rho_dot_sgl2dip_screw_ID, &
rho_dot_ann_ath_ID, &
rho_dot_ann_the_ID, &
rho_dot_ann_the_edge_ID, &
rho_dot_ann_the_screw_ID, &
rho_dot_edgejogs_ID, &
rho_dot_flux_ID, &
added two new outputs: "rho_dot_flux_mobile" and "rho_dot_sgl_mobile" corrected outputs of dislocation density rates: take into account that blocked dislocation density is a signed quantity
2014-01-07 21:25:28 +05:30
rho_dot_flux_mobile_ID, &
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
rho_dot_flux_edge_ID, &
rho_dot_flux_screw_ID, &
velocity_edge_pos_ID, &
velocity_edge_neg_ID, &
velocity_screw_pos_ID, &
velocity_screw_neg_ID, &
slipdirectionx_ID, &
slipdirectiony_ID, &
slipdirectionz_ID, &
slipnormalx_ID, &
slipnormaly_ID, &
slipnormalz_ID, &
fluxdensity_edge_posx_ID, &
fluxdensity_edge_posy_ID, &
fluxdensity_edge_posz_ID, &
fluxdensity_edge_negx_ID, &
fluxdensity_edge_negy_ID, &
fluxdensity_edge_negz_ID, &
fluxdensity_screw_posx_ID, &
fluxdensity_screw_posy_ID, &
fluxdensity_screw_posz_ID, &
fluxdensity_screw_negx_ID, &
fluxdensity_screw_negy_ID, &
fluxdensity_screw_negz_ID, &
maximumdipoleheight_edge_ID, &
maximumdipoleheight_screw_ID, &
accumulatedshear_ID, &
dislocationstress_ID
end enum
adapted changes from rev 2776 for constitutive_nonlocal: improved usage of "enum", fixed bug in file reading, usage of "source" specifier for "allocate"
2013-12-12 15:46:50 +05:30
integer(kind(undefined_ID)), dimension(:,:), allocatable, private :: &
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
constitutive_nonlocal_outputID !< ID 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
public :: &
constitutive_nonlocal_init, &
constitutive_nonlocal_stateInit, &
constitutive_nonlocal_aTolState, &
constitutive_nonlocal_microstructure, &
constitutive_nonlocal_LpAndItsTangent, &
constitutive_nonlocal_dotState, &
constitutive_nonlocal_deltaState, &
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
introduced named constants for homogenization_type, phase_plasticitiy, and phase_elasticity to replace expensive string comparison. Labels for different schemes need to be stored in material.f90 because they are needed when parsing the config file
2013-11-27 13:34:05 +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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief module initialization
!> @details reads in material parameters, allocates arrays, and does sanity checks
!--------------------------------------------------------------------------------------------------
adapted changes from rev 2776 for constitutive_nonlocal: improved usage of "enum", fixed bug in file reading, usage of "source" specifier for "allocate"
2013-12-12 15:46:50 +05:30
subroutine constitutive_nonlocal_init(fileUnit)
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
added possibility for multi-level inclusion of files in *.config and loadcase files. include subfiles by stating {path/to/include}
2013-06-27 00:49:00 +05:30
use IO, only: IO_read, &
IO_lc, &
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
IO_getTag, &
IO_isBlank, &
IO_stringPos, &
IO_stringValue, &
IO_floatValue, &
IO_intValue, &
require all hardening coefficients/ non schmid coefficients to be specified in material.config, do not silently ignore any warning message now more meaningful (?) when using less than maximum number of slip families
2013-09-12 20:17:09 +05:30
IO_warning, &
fixed over-sensitive error in homogeniztion and ifort option in makefile
2013-12-19 14:19:47 +05:30
IO_error, &
adapted changes from rev 2776 for constitutive_nonlocal: improved usage of "enum", fixed bug in file reading, usage of "source" specifier for "allocate"
2013-12-12 15:46:50 +05:30
IO_timeStamp, &
IO_EOF
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, &
introduced named constants for homogenization_type, phase_plasticitiy, and phase_elasticity to replace expensive string comparison. Labels for different schemes need to be stored in material.f90 because they are needed when parsing the config file
2013-11-27 13:34:05 +05:30
phase_Noutput, &
PLASTICITY_NONLOCAL_label, &
fixed over-sensitive error in homogeniztion and ifort option in makefile
2013-12-19 14:19:47 +05:30
PLASTICITY_NONLOCAL_ID, &
MATERIAL_partPhase
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
adapted changes from rev 2776 for constitutive_nonlocal: improved usage of "enum", fixed bug in file reading, usage of "source" specifier for "allocate"
2013-12-12 15:46:50 +05:30
integer(pInt), intent(in) :: fileUnit
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
integer(pInt), parameter :: MAXNCHUNKS = LATTICE_maxNinteraction + 1_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
integer(pInt), &
require all hardening coefficients/ non schmid coefficients to be specified in material.config, do not silently ignore any warning message now more meaningful (?) when using less than maximum number of slip families
2013-09-12 20:17:09 +05:30
dimension(1_pInt+2_pInt*MAXNCHUNKS) :: positions
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
integer(pInt) :: phase = 0_pInt, &
renamed "maxNmatIDs" accordingly to "maxNinstances"
2014-03-04 19:17:04 +05:30
maxNinstances, &
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
maxTotalNslip, &
f, & ! index of my slip family
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
instance, & ! 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
combined some state indices to an array with a more generic name
2013-05-24 14:32:30 +05:30
t, & ! index of dislocation type
c, & ! index of dislocation character
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, &
require all hardening coefficients/ non schmid coefficients to be specified in material.config, do not silently ignore any warning message now more meaningful (?) when using less than maximum number of slip families
2013-09-12 20:17:09 +05:30
Nchunks_nonSchmid = 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
fixed small bug in detecting the lattice structure. only first 3 letters count (e.g. 'ort' and 'orthography' will give the orthorombic lattice type)
2013-11-28 14:26:02 +05:30
character(len=65536) :: &
tag = '', &
line = ''
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
introduced named constants for homogenization_type, phase_plasticitiy, and phase_elasticity to replace expensive string comparison. Labels for different schemes need to be stored in material.f90 because they are needed when parsing the config file
2013-11-27 13:34:05 +05:30
write(6,'(/,a)') ' <<<+- constitutive_'//PLASTICITY_NONLOCAL_label//' init -+>>>'
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
write(6,'(a)') ' $Id$'
write(6,'(a15,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
renamed "maxNmatIDs" accordingly to "maxNinstances"
2014-03-04 19:17:04 +05:30
maxNinstances = int(count(phase_plasticity == PLASTICITY_NONLOCAL_ID),pInt)
if (maxNinstances == 0) return ! we don't have to do anything if there's no instance for this constitutive law
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
if (iand(debug_level(debug_constitutive),debug_levelBasic) /= 0_pInt) &
renamed "maxNmatIDs" accordingly to "maxNinstances"
2014-03-04 19:17:04 +05:30
write(6,'(a16,1x,i5,/)') '# instances:',maxNinstances
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
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
renamed "maxNmatIDs" accordingly to "maxNinstances"
2014-03-04 19:17:04 +05:30
allocate(constitutive_nonlocal_sizeDotState(maxNinstances), source=0_pInt)
allocate(constitutive_nonlocal_sizeDependentState(maxNinstances), source=0_pInt)
allocate(constitutive_nonlocal_sizeState(maxNinstances), source=0_pInt)
allocate(constitutive_nonlocal_sizePostResults(maxNinstances), source=0_pInt)
allocate(constitutive_nonlocal_sizePostResult(maxval(phase_Noutput), maxNinstances), source=0_pInt)
allocate(Noutput(maxNinstances), source=0_pInt)
allocate(constitutive_nonlocal_output(maxval(phase_Noutput), maxNinstances))
adapted changes from rev 2776 for constitutive_nonlocal: improved usage of "enum", fixed bug in file reading, usage of "source" specifier for "allocate"
2013-12-12 15:46:50 +05:30
constitutive_nonlocal_output = ''
renamed "maxNmatIDs" accordingly to "maxNinstances"
2014-03-04 19:17:04 +05:30
allocate(constitutive_nonlocal_outputID(maxval(phase_Noutput), maxNinstances), source=undefined_ID)
allocate(Nslip(lattice_maxNslipFamily,maxNinstances), source=0_pInt)
allocate(slipFamily(lattice_maxNslip,maxNinstances), source=0_pInt)
allocate(slipSystemLattice(lattice_maxNslip,maxNinstances), source=0_pInt)
allocate(totalNslip(maxNinstances), source=0_pInt)
allocate(atomicVolume(maxNinstances), source=0.0_pReal)
allocate(Dsd0(maxNinstances), source=-1.0_pReal)
allocate(selfDiffusionEnergy(maxNinstances), source=0.0_pReal)
allocate(aTolRho(maxNinstances), source=0.0_pReal)
allocate(aTolShear(maxNinstances), source=0.0_pReal)
allocate(significantRho(maxNinstances), source=0.0_pReal)
allocate(significantN(maxNinstances), source=0.0_pReal)
allocate(cutoffRadius(maxNinstances), source=-1.0_pReal)
allocate(doublekinkwidth(maxNinstances), source=0.0_pReal)
allocate(solidSolutionEnergy(maxNinstances), source=0.0_pReal)
allocate(solidSolutionSize(maxNinstances), source=0.0_pReal)
allocate(solidSolutionConcentration(maxNinstances), source=0.0_pReal)
allocate(pParam(maxNinstances), source=1.0_pReal)
allocate(qParam(maxNinstances), source=1.0_pReal)
allocate(viscosity(maxNinstances), source=0.0_pReal)
allocate(fattack(maxNinstances), source=0.0_pReal)
allocate(rhoSglScatter(maxNinstances), source=0.0_pReal)
allocate(rhoSglRandom(maxNinstances), source=0.0_pReal)
allocate(rhoSglRandomBinning(maxNinstances), source=1.0_pReal)
allocate(surfaceTransmissivity(maxNinstances), source=1.0_pReal)
allocate(grainboundaryTransmissivity(maxNinstances), source=-1.0_pReal)
allocate(CFLfactor(maxNinstances), source=2.0_pReal)
allocate(fEdgeMultiplication(maxNinstances), source=0.0_pReal)
allocate(linetensionEffect(maxNinstances), source=0.0_pReal)
allocate(edgeJogFactor(maxNinstances), source=1.0_pReal)
allocate(shortRangeStressCorrection(maxNinstances), source=.false.)
allocate(probabilisticMultiplication(maxNinstances), source=.false.)
allocate(rhoSglEdgePos0(lattice_maxNslipFamily,maxNinstances), source=-1.0_pReal)
allocate(rhoSglEdgeNeg0(lattice_maxNslipFamily,maxNinstances), source=-1.0_pReal)
allocate(rhoSglScrewPos0(lattice_maxNslipFamily,maxNinstances), source=-1.0_pReal)
allocate(rhoSglScrewNeg0(lattice_maxNslipFamily,maxNinstances), source=-1.0_pReal)
allocate(rhoDipEdge0(lattice_maxNslipFamily,maxNinstances), source=-1.0_pReal)
allocate(rhoDipScrew0(lattice_maxNslipFamily,maxNinstances), source=-1.0_pReal)
allocate(burgersPerSlipFamily(lattice_maxNslipFamily,maxNinstances), source=0.0_pReal)
allocate(lambda0PerSlipFamily(lattice_maxNslipFamily,maxNinstances), source=0.0_pReal)
allocate(interactionSlipSlip(lattice_maxNinteraction,maxNinstances), source=0.0_pReal)
allocate(minDipoleHeightPerSlipFamily(lattice_maxNslipFamily,2,maxNinstances), source=-1.0_pReal)
allocate(peierlsStressPerSlipFamily(lattice_maxNslipFamily,2,maxNinstances), source=0.0_pReal)
allocate(nonSchmidCoeff(lattice_maxNnonSchmid,maxNinstances), source=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
added non-schmid structure
2013-01-22 05:20:28 +05:30
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
rewind(fileUnit)
phase = 0_pInt
do while (trim(line) /= IO_EOF .and. IO_lc(IO_getTag(line,'<','>')) /= MATERIAL_partPhase) ! wind forward to <phase>
line = IO_read(fileUnit)
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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
parsingFile: do while (trim(line) /= IO_EOF) ! read through phases of phase part
line = IO_read(fileUnit)
if (IO_isBlank(line)) cycle ! skip empty lines
if (IO_getTag(line,'<','>') /= '') then ! stop at next part
line = IO_read(fileUnit, .true.) ! reset IO_read
exit
endif
if (IO_getTag(line,'[',']') /= '') then ! next phase
phase = phase + 1_pInt ! advance phase section counter
if (phase_plasticity(phase) == PLASTICITY_NONLOCAL_ID) then
Nchunks_SlipFamilies = count(lattice_NslipSystem(:,phase) > 0_pInt)
Nchunks_SlipSlip = maxval(lattice_InteractionSlipSlip(:,:,phase))
Nchunks_nonSchmid = lattice_NnonSchmid(phase)
endif
cycle
endif
if (phase > 0_pInt ) then; if (phase_plasticity(phase) == PLASTICITY_NONLOCAL_ID) then ! one of my phases. do not short-circuit here (.and. with next if statement). It's not safe in Fortran
instance = phase_plasticityInstance(phase) ! which instance of my plasticity is present phase
positions = IO_stringPos(line,MAXNCHUNKS)
tag = IO_lc(IO_stringValue(line,positions,1_pInt)) ! extract key
select case(tag)
case ('plasticity','elasticity','lattice_structure', &
'covera_ratio','c/a_ratio','c/a', &
'c11','c12','c13','c22','c23','c33','c44','c55','c66',&
'/nonlocal/')
cycle
case ('(output)')
Noutput(instance) = Noutput(instance) + 1_pInt
constitutive_nonlocal_output(Noutput(instance),instance) = IO_lc(IO_stringValue(line,positions,2_pInt))
select case(IO_lc(IO_stringValue(line,positions,2_pInt)))
case('rho')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_ID
case('delta')
constitutive_nonlocal_outputID(Noutput(instance),instance) = delta_ID
case('rho_edge')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_edge_ID
case('rho_screw')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_screw_ID
case('rho_sgl')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_ID
case('delta_sgl')
constitutive_nonlocal_outputID(Noutput(instance),instance) = delta_sgl_ID
case('rho_sgl_edge')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_edge_ID
case('rho_sgl_edge_pos')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_edge_pos_ID
case('rho_sgl_edge_neg')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_edge_neg_ID
case('rho_sgl_screw')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_screw_ID
case('rho_sgl_screw_pos')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_screw_pos_ID
case('rho_sgl_screw_neg')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_screw_neg_ID
case('rho_sgl_mobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_mobile_ID
case('rho_sgl_edge_mobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_edge_mobile_ID
case('rho_sgl_edge_pos_mobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_edge_pos_mobile_ID
case('rho_sgl_edge_neg_mobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_edge_neg_mobile_ID
case('rho_sgl_screw_mobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_screw_mobile_ID
case('rho_sgl_screw_pos_mobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_screw_pos_mobile_ID
case('rho_sgl_screw_neg_mobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_screw_neg_mobile_ID
case('rho_sgl_immobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_immobile_ID
case('rho_sgl_edge_immobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_edge_immobile_ID
case('rho_sgl_edge_pos_immobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_edge_pos_immobile_ID
case('rho_sgl_edge_neg_immobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_edge_neg_immobile_ID
case('rho_sgl_screw_immobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_screw_immobile_ID
case('rho_sgl_screw_pos_immobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_screw_pos_immobile_ID
case('rho_sgl_screw_neg_immobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_sgl_screw_neg_immobile_ID
case('rho_dip')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dip_ID
case('delta_dip')
constitutive_nonlocal_outputID(Noutput(instance),instance) = delta_dip_ID
case('rho_dip_edge')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dip_edge_ID
case('rho_dip_screw')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dip_screw_ID
case('excess_rho')
constitutive_nonlocal_outputID(Noutput(instance),instance) = excess_rho_ID
case('excess_rho_edge')
constitutive_nonlocal_outputID(Noutput(instance),instance) = excess_rho_edge_ID
case('excess_rho_screw')
constitutive_nonlocal_outputID(Noutput(instance),instance) = excess_rho_screw_ID
case('rho_forest')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_forest_ID
case('shearrate')
constitutive_nonlocal_outputID(Noutput(instance),instance) = shearrate_ID
case('resolvedstress')
constitutive_nonlocal_outputID(Noutput(instance),instance) = resolvedstress_ID
case('resolvedstress_external')
constitutive_nonlocal_outputID(Noutput(instance),instance) = resolvedstress_external_ID
case('resolvedstress_back')
constitutive_nonlocal_outputID(Noutput(instance),instance) = resolvedstress_back_ID
case('resistance')
constitutive_nonlocal_outputID(Noutput(instance),instance) = resistance_ID
case('rho_dot')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_ID
case('rho_dot_sgl')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_sgl_ID
case('rho_dot_sgl_mobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_sgl_mobile_ID
case('rho_dot_dip')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_dip_ID
case('rho_dot_gen')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_gen_ID
case('rho_dot_gen_edge')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_gen_edge_ID
case('rho_dot_gen_screw')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_gen_screw_ID
case('rho_dot_sgl2dip')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_sgl2dip_ID
case('rho_dot_sgl2dip_edge')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_sgl2dip_edge_ID
case('rho_dot_sgl2dip_screw')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_sgl2dip_screw_ID
case('rho_dot_ann_ath')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_ann_ath_ID
case('rho_dot_ann_the')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_ann_the_ID
case('rho_dot_ann_the_edge')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_ann_the_edge_ID
case('rho_dot_ann_the_screw')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_ann_the_screw_ID
case('rho_dot_edgejogs')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_edgejogs_ID
case('rho_dot_flux')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_flux_ID
case('rho_dot_flux_mobile')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_flux_mobile_ID
case('rho_dot_flux_edge')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_flux_edge_ID
case('rho_dot_flux_screw')
constitutive_nonlocal_outputID(Noutput(instance),instance) = rho_dot_flux_screw_ID
case('velocity_edge_pos')
constitutive_nonlocal_outputID(Noutput(instance),instance) = velocity_edge_pos_ID
case('velocity_edge_neg')
constitutive_nonlocal_outputID(Noutput(instance),instance) = velocity_edge_neg_ID
case('velocity_screw_pos')
constitutive_nonlocal_outputID(Noutput(instance),instance) = velocity_screw_pos_ID
case('velocity_screw_neg')
constitutive_nonlocal_outputID(Noutput(instance),instance) = velocity_screw_neg_ID
case('slipdirection.x')
constitutive_nonlocal_outputID(Noutput(instance),instance) = slipdirectionx_ID
case('slipdirection.y')
constitutive_nonlocal_outputID(Noutput(instance),instance) = slipdirectiony_ID
case('slipdirection.z')
constitutive_nonlocal_outputID(Noutput(instance),instance) = slipdirectionz_ID
case('slipnormal.x')
constitutive_nonlocal_outputID(Noutput(instance),instance) = slipnormalx_ID
case('slipnormal.y')
constitutive_nonlocal_outputID(Noutput(instance),instance) = slipnormaly_ID
case('slipnormal.z')
constitutive_nonlocal_outputID(Noutput(instance),instance) = slipnormalz_ID
case('fluxdensity_edge_pos.x')
constitutive_nonlocal_outputID(Noutput(instance),instance) = fluxdensity_edge_posx_ID
case('fluxdensity_edge_pos.y')
constitutive_nonlocal_outputID(Noutput(instance),instance) = fluxdensity_edge_posy_ID
case('fluxdensity_edge_pos.z')
constitutive_nonlocal_outputID(Noutput(instance),instance) = fluxdensity_edge_posz_ID
case('fluxdensity_edge_neg.x')
constitutive_nonlocal_outputID(Noutput(instance),instance) = fluxdensity_edge_negx_ID
case('fluxdensity_edge_neg.y')
constitutive_nonlocal_outputID(Noutput(instance),instance) = fluxdensity_edge_negy_ID
case('fluxdensity_edge_neg.z')
constitutive_nonlocal_outputID(Noutput(instance),instance) = fluxdensity_edge_negz_ID
case('fluxdensity_screw_pos.x')
constitutive_nonlocal_outputID(Noutput(instance),instance) = fluxdensity_screw_posx_ID
case('fluxdensity_screw_pos.y')
constitutive_nonlocal_outputID(Noutput(instance),instance) = fluxdensity_screw_posy_ID
case('fluxdensity_screw_pos.z')
constitutive_nonlocal_outputID(Noutput(instance),instance) = fluxdensity_screw_posz_ID
case('fluxdensity_screw_neg.x')
constitutive_nonlocal_outputID(Noutput(instance),instance) = fluxdensity_screw_negx_ID
case('fluxdensity_screw_neg.y')
constitutive_nonlocal_outputID(Noutput(instance),instance) = fluxdensity_screw_negy_ID
case('fluxdensity_screw_neg.z')
constitutive_nonlocal_outputID(Noutput(instance),instance) = fluxdensity_screw_negz_ID
case('maximumdipoleheight_edge')
constitutive_nonlocal_outputID(Noutput(instance),instance) = maximumdipoleheight_edge_ID
case('maximumdipoleheight_screw')
constitutive_nonlocal_outputID(Noutput(instance),instance) = maximumdipoleheight_screw_ID
case('accumulatedshear')
constitutive_nonlocal_outputID(Noutput(instance),instance) = accumulatedshear_ID
case('dislocationstress')
constitutive_nonlocal_outputID(Noutput(instance),instance) = dislocationstress_ID
case default
call IO_error(105_pInt,ext_msg=IO_stringValue(line,positions,2_pInt)//' ('//PLASTICITY_NONLOCAL_label//')')
end select
case ('nslip')
if (positions(1) < 1_pInt + Nchunks_SlipFamilies) &
call IO_warning(50_pInt,ext_msg=trim(tag)//' ('//PLASTICITY_NONLOCAL_LABEL//')')
Nchunks_SlipFamilies = positions(1) - 1_pInt
do f = 1_pInt, Nchunks_SlipFamilies
Nslip(f,instance) = IO_intValue(line,positions,1_pInt+f)
enddo
case ('rhosgledgepos0')
do f = 1_pInt, Nchunks_SlipFamilies
rhoSglEdgePos0(f,instance) = IO_floatValue(line,positions,1_pInt+f)
enddo
case ('rhosgledgeneg0')
do f = 1_pInt, Nchunks_SlipFamilies
rhoSglEdgeNeg0(f,instance) = IO_floatValue(line,positions,1_pInt+f)
enddo
case ('rhosglscrewpos0')
do f = 1_pInt, Nchunks_SlipFamilies
rhoSglScrewPos0(f,instance) = IO_floatValue(line,positions,1_pInt+f)
enddo
case ('rhosglscrewneg0')
do f = 1_pInt, Nchunks_SlipFamilies
rhoSglScrewNeg0(f,instance) = IO_floatValue(line,positions,1_pInt+f)
enddo
case ('rhodipedge0')
do f = 1_pInt, Nchunks_SlipFamilies
rhoDipEdge0(f,instance) = IO_floatValue(line,positions,1_pInt+f)
enddo
case ('rhodipscrew0')
do f = 1_pInt, Nchunks_SlipFamilies
rhoDipScrew0(f,instance) = IO_floatValue(line,positions,1_pInt+f)
enddo
case ('lambda0')
do f = 1_pInt, Nchunks_SlipFamilies
lambda0PerSlipFamily(f,instance) = IO_floatValue(line,positions,1_pInt+f)
enddo
case ('burgers')
do f = 1_pInt, Nchunks_SlipFamilies
burgersPerSlipFamily(f,instance) = IO_floatValue(line,positions,1_pInt+f)
enddo
case('cutoffradius','r')
cutoffRadius(instance) = IO_floatValue(line,positions,2_pInt)
case('minimumdipoleheightedge','ddipminedge')
do f = 1_pInt, Nchunks_SlipFamilies
minDipoleHeightPerSlipFamily(f,1_pInt,instance) = IO_floatValue(line,positions,1_pInt+f)
enddo
case('minimumdipoleheightscrew','ddipminscrew')
do f = 1_pInt, Nchunks_SlipFamilies
minDipoleHeightPerSlipFamily(f,2_pInt,instance) = IO_floatValue(line,positions,1_pInt+f)
enddo
case('atomicvolume')
atomicVolume(instance) = IO_floatValue(line,positions,2_pInt)
case('selfdiffusionprefactor','dsd0')
Dsd0(instance) = IO_floatValue(line,positions,2_pInt)
case('selfdiffusionenergy','qsd')
selfDiffusionEnergy(instance) = IO_floatValue(line,positions,2_pInt)
case('atol_rho','atol_density','absolutetolerancedensity','absolutetolerance_density')
aTolRho(instance) = IO_floatValue(line,positions,2_pInt)
case('atol_shear','atol_plasticshear','atol_accumulatedshear','absolutetoleranceshear','absolutetolerance_shear')
aTolShear(instance) = IO_floatValue(line,positions,2_pInt)
case('significantrho','significant_rho','significantdensity','significant_density')
significantRho(instance) = IO_floatValue(line,positions,2_pInt)
case('significantn','significant_n','significantdislocations','significant_dislcations')
significantN(instance) = IO_floatValue(line,positions,2_pInt)
case ('interaction_slipslip')
if (positions(1) < 1_pInt + Nchunks_SlipSlip) &
call IO_warning(52_pInt,ext_msg=trim(tag)//' ('//PLASTICITY_NONLOCAL_LABEL//')')
do it = 1_pInt,Nchunks_SlipSlip
interactionSlipSlip(it,instance) = IO_floatValue(line,positions,1_pInt+it)
enddo
case('linetension','linetensioneffect','linetension_effect')
linetensionEffect(instance) = IO_floatValue(line,positions,2_pInt)
case('edgejog','edgejogs','edgejogeffect','edgejog_effect')
edgeJogFactor(instance) = IO_floatValue(line,positions,2_pInt)
case('peierlsstressedge','peierlsstress_edge')
do f = 1_pInt, Nchunks_SlipFamilies
peierlsStressPerSlipFamily(f,1_pInt,instance) = IO_floatValue(line,positions,1_pInt+f)
enddo
case('peierlsstressscrew','peierlsstress_screw')
do f = 1_pInt, Nchunks_SlipFamilies
peierlsStressPerSlipFamily(f,2_pInt,instance) = IO_floatValue(line,positions,1_pInt+f)
enddo
case('doublekinkwidth')
doublekinkwidth(instance) = IO_floatValue(line,positions,2_pInt)
case('solidsolutionenergy')
solidSolutionEnergy(instance) = IO_floatValue(line,positions,2_pInt)
case('solidsolutionsize')
solidSolutionSize(instance) = IO_floatValue(line,positions,2_pInt)
case('solidsolutionconcentration')
solidSolutionConcentration(instance) = IO_floatValue(line,positions,2_pInt)
case('p')
pParam(instance) = IO_floatValue(line,positions,2_pInt)
case('q')
qParam(instance) = IO_floatValue(line,positions,2_pInt)
case('viscosity','glideviscosity')
viscosity(instance) = IO_floatValue(line,positions,2_pInt)
case('attackfrequency','fattack')
fattack(instance) = IO_floatValue(line,positions,2_pInt)
case('rhosglscatter')
rhoSglScatter(instance) = IO_floatValue(line,positions,2_pInt)
case('rhosglrandom')
rhoSglRandom(instance) = IO_floatValue(line,positions,2_pInt)
case('rhosglrandombinning')
rhoSglRandomBinning(instance) = IO_floatValue(line,positions,2_pInt)
case('surfacetransmissivity')
surfaceTransmissivity(instance) = IO_floatValue(line,positions,2_pInt)
case('grainboundarytransmissivity')
grainboundaryTransmissivity(instance) = IO_floatValue(line,positions,2_pInt)
case('cflfactor')
CFLfactor(instance) = IO_floatValue(line,positions,2_pInt)
case('fedgemultiplication','edgemultiplicationfactor','edgemultiplication')
fEdgeMultiplication(instance) = IO_floatValue(line,positions,2_pInt)
case('shortrangestresscorrection')
shortRangeStressCorrection(instance) = IO_floatValue(line,positions,2_pInt) > 0.0_pReal
case ('nonschmid_coefficients')
if (positions(1) < 1_pInt + Nchunks_nonSchmid) &
call IO_warning(52_pInt,ext_msg=trim(tag)//' ('//PLASTICITY_NONLOCAL_label//')')
do f = 1_pInt,Nchunks_nonSchmid
nonSchmidCoeff(f,instance) = IO_floatValue(line,positions,1_pInt+f)
enddo
case('probabilisticmultiplication','randomsources','randommultiplication','discretesources')
probabilisticMultiplication(instance) = IO_floatValue(line,positions,2_pInt) > 0.0_pReal
case default
call IO_error(210_pInt,ext_msg=trim(tag)//' ('//PLASTICITY_NONLOCAL_label//')')
end select
endif; endif
enddo parsingFile
sanityChecks: do phase = 1_pInt, size(phase_plasticity)
myPhase: if (phase_plasticity(phase) == PLASTICITY_NONLOCAL_ID) then
if (sum(Nslip(:,instance)) <= 0_pInt) &
call IO_error(211_pInt,ext_msg='Nslip ('//PLASTICITY_NONLOCAL_label//')')
do o = 1_pInt,maxval(phase_Noutput)
if(len(constitutive_nonlocal_output(o,instance)) > 64_pInt) &
call IO_error(666_pInt)
enddo
do f = 1_pInt,lattice_maxNslipFamily
if (Nslip(f,instance) > 0_pInt) then
if (rhoSglEdgePos0(f,instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoSglEdgePos0 ('//PLASTICITY_NONLOCAL_label//')')
if (rhoSglEdgeNeg0(f,instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoSglEdgeNeg0 ('//PLASTICITY_NONLOCAL_label//')')
if (rhoSglScrewPos0(f,instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoSglScrewPos0 ('//PLASTICITY_NONLOCAL_label//')')
if (rhoSglScrewNeg0(f,instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoSglScrewNeg0 ('//PLASTICITY_NONLOCAL_label//')')
if (rhoDipEdge0(f,instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoDipEdge0 ('//PLASTICITY_NONLOCAL_label//')')
if (rhoDipScrew0(f,instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoDipScrew0 ('//PLASTICITY_NONLOCAL_label//')')
if (burgersPerSlipFamily(f,instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='Burgers ('//PLASTICITY_NONLOCAL_label//')')
if (lambda0PerSlipFamily(f,instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='lambda0 ('//PLASTICITY_NONLOCAL_label//')')
if (minDipoleHeightPerSlipFamily(f,1,instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='minimumDipoleHeightEdge ('//PLASTICITY_NONLOCAL_label//')')
if (minDipoleHeightPerSlipFamily(f,2,instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='minimumDipoleHeightScrew ('//PLASTICITY_NONLOCAL_label//')')
if (peierlsStressPerSlipFamily(f,1,instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='peierlsStressEdge ('//PLASTICITY_NONLOCAL_label//')')
if (peierlsStressPerSlipFamily(f,2,instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='peierlsStressScrew ('//PLASTICITY_NONLOCAL_label//')')
endif
enddo
if (any(interactionSlipSlip(1:maxval(lattice_interactionSlipSlip(:,:,phase)),instance) < 0.0_pReal)) &
call IO_error(211_pInt,ext_msg='interaction_SlipSlip ('//PLASTICITY_NONLOCAL_label//')')
if (linetensionEffect(instance) < 0.0_pReal .or. linetensionEffect(instance) > 1.0_pReal) &
call IO_error(211_pInt,ext_msg='linetension ('//PLASTICITY_NONLOCAL_label//')')
if (edgeJogFactor(instance) < 0.0_pReal .or. edgeJogFactor(instance) > 1.0_pReal) &
call IO_error(211_pInt,ext_msg='edgejog ('//PLASTICITY_NONLOCAL_label//')')
if (cutoffRadius(instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='r ('//PLASTICITY_NONLOCAL_label//')')
if (atomicVolume(instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='atomicVolume ('//PLASTICITY_NONLOCAL_label//')')
if (Dsd0(instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='selfDiffusionPrefactor ('//PLASTICITY_NONLOCAL_label//')')
if (selfDiffusionEnergy(instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='selfDiffusionEnergy ('//PLASTICITY_NONLOCAL_label//')')
if (aTolRho(instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='aTol_rho ('//PLASTICITY_NONLOCAL_label//')')
if (aTolShear(instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='aTol_shear ('//PLASTICITY_NONLOCAL_label//')')
if (significantRho(instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='significantRho ('//PLASTICITY_NONLOCAL_label//')')
if (significantN(instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='significantN ('//PLASTICITY_NONLOCAL_label//')')
if (doublekinkwidth(instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='doublekinkwidth ('//PLASTICITY_NONLOCAL_label//')')
if (solidSolutionEnergy(instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='solidSolutionEnergy ('//PLASTICITY_NONLOCAL_label//')')
if (solidSolutionSize(instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='solidSolutionSize ('//PLASTICITY_NONLOCAL_label//')')
if (solidSolutionConcentration(instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='solidSolutionConcentration ('//PLASTICITY_NONLOCAL_label//')')
if (pParam(instance) <= 0.0_pReal .or. pParam(instance) > 1.0_pReal) &
call IO_error(211_pInt,ext_msg='p ('//PLASTICITY_NONLOCAL_label//')')
if (qParam(instance) < 1.0_pReal .or. qParam(instance) > 2.0_pReal) &
call IO_error(211_pInt,ext_msg='q ('//PLASTICITY_NONLOCAL_label//')')
if (viscosity(instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='viscosity ('//PLASTICITY_NONLOCAL_label//')')
if (fattack(instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='attackFrequency ('//PLASTICITY_NONLOCAL_label//')')
if (rhoSglScatter(instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoSglScatter ('//PLASTICITY_NONLOCAL_label//')')
if (rhoSglRandom(instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoSglRandom ('//PLASTICITY_NONLOCAL_label//')')
if (rhoSglRandomBinning(instance) <= 0.0_pReal) &
call IO_error(211_pInt,ext_msg='rhoSglRandomBinning ('//PLASTICITY_NONLOCAL_label//')')
if (surfaceTransmissivity(instance) < 0.0_pReal .or. surfaceTransmissivity(instance) > 1.0_pReal) &
call IO_error(211_pInt,ext_msg='surfaceTransmissivity ('//PLASTICITY_NONLOCAL_label//')')
if (grainboundaryTransmissivity(instance) > 1.0_pReal) &
call IO_error(211_pInt,ext_msg='grainboundaryTransmissivity ('//PLASTICITY_NONLOCAL_label//')')
if (CFLfactor(instance) < 0.0_pReal) &
call IO_error(211_pInt,ext_msg='CFLfactor ('//PLASTICITY_NONLOCAL_label//')')
if (fEdgeMultiplication(instance) < 0.0_pReal .or. fEdgeMultiplication(instance) > 1.0_pReal) &
call IO_error(211_pInt,ext_msg='edgemultiplicationfactor ('//PLASTICITY_NONLOCAL_label//')')
!*** determine total number of active slip systems
Nslip(1:lattice_maxNslipFamily,instance) = min(lattice_NslipSystem(1:lattice_maxNslipFamily,phase), &
Nslip(1:lattice_maxNslipFamily,instance) ) ! we can't use more slip systems per family than specified in lattice
totalNslip(instance) = sum(Nslip(1:lattice_maxNslipFamily,instance))
endif myPhase
enddo sanityChecks
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
!*** 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
renamed "maxNmatIDs" accordingly to "maxNinstances"
2014-03-04 19:17:04 +05:30
allocate(iRhoU(maxTotalNslip,4,maxNinstances), source=0_pInt)
allocate(iRhoB(maxTotalNslip,4,maxNinstances), source=0_pInt)
allocate(iRhoD(maxTotalNslip,2,maxNinstances), source=0_pInt)
allocate(iV(maxTotalNslip,4,maxNinstances), source=0_pInt)
allocate(iD(maxTotalNslip,2,maxNinstances), source=0_pInt)
allocate(iGamma(maxTotalNslip,maxNinstances), source=0_pInt)
allocate(iRhoF(maxTotalNslip,maxNinstances), source=0_pInt)
allocate(iTauF(maxTotalNslip,maxNinstances), source=0_pInt)
allocate(iTauB(maxTotalNslip,maxNinstances), source=0_pInt)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
allocate(burgers(maxTotalNslip,maxNinstances), source=0.0_pReal)
allocate(lambda0(maxTotalNslip,maxNinstances), source=0.0_pReal)
allocate(minDipoleHeight(maxTotalNslip,2,maxNinstances), source=-1.0_pReal)
allocate(forestProjectionEdge(maxTotalNslip,maxTotalNslip,maxNinstances), source=0.0_pReal)
allocate(forestProjectionScrew(maxTotalNslip,maxTotalNslip,maxNinstances), source=0.0_pReal)
allocate(interactionMatrixSlipSlip(maxTotalNslip,maxTotalNslip,maxNinstances), source=0.0_pReal)
allocate(lattice2slip(1:3, 1:3, maxTotalNslip,maxNinstances), source=0.0_pReal)
adapted changes from rev 2776 for constitutive_nonlocal: improved usage of "enum", fixed bug in file reading, usage of "source" specifier for "allocate"
2013-12-12 15:46:50 +05:30
allocate(sourceProbability(maxTotalNslip,homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), &
source=2.0_pReal)
allocate(rhoDotFluxOutput(maxTotalNslip,8,homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), &
source=0.0_pReal)
allocate(rhoDotMultiplicationOutput(maxTotalNslip,2,homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), &
source=0.0_pReal)
allocate(rhoDotSingle2DipoleGlideOutput(maxTotalNslip,2,homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), &
source=0.0_pReal)
allocate(rhoDotAthermalAnnihilationOutput(maxTotalNslip,2,homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), &
source=0.0_pReal)
allocate(rhoDotThermalAnnihilationOutput(maxTotalNslip,2,homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), &
source=0.0_pReal)
allocate(rhoDotEdgeJogsOutput(maxTotalNslip,homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), &
source=0.0_pReal)
allocate(compatibility(2,maxTotalNslip,maxTotalNslip,mesh_maxNipNeighbors,mesh_maxNips,mesh_NcpElems), &
source=0.0_pReal)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
allocate(peierlsStress(maxTotalNslip,2,maxNinstances), source=0.0_pReal)
allocate(colinearSystem(maxTotalNslip,maxNinstances), source=0_pInt)
allocate(nonSchmidProjection(3,3,4,maxTotalNslip,maxNinstances), source=0.0_pReal)
nonSchmid behavior now ready to use for bcc with phenopowerlaw and nonlocal constitutive model. nonSchmid tensors according to Koester,Ma,Hartmaier,2012. extended lattice_Sslip with non-Schmid tensors (the full non-symmetric tensors are required for the tangent and cannot be retrieved from the symmetrized Mandel notation)
2013-08-05 14:53:21 +05:30
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
initializeInstances: do phase = 1_pInt, size(phase_plasticity)
if (phase_plasticity(phase) == PLASTICITY_NONLOCAL_ID) then
!*** Inverse lookup of my slip system family and the slip system in lattice
l = 0_pInt
do f = 1_pInt,lattice_maxNslipFamily
do s = 1_pInt,Nslip(f,instance)
l = l + 1_pInt
slipFamily(l,instance) = f
slipSystemLattice(l,instance) = sum(lattice_NslipSystem(1:f-1_pInt, phase)) + s
enddo; enddo
!*** determine size of state array
ns = totalNslip(instance)
constitutive_nonlocal_sizeDotState(instance) = int(size(BASICSTATES),pInt) * ns
constitutive_nonlocal_sizeDependentState(instance) = int(size(DEPENDENTSTATES),pInt) * ns
constitutive_nonlocal_sizeState(instance) = constitutive_nonlocal_sizeDotState(instance) &
+ constitutive_nonlocal_sizeDependentState(instance) &
+ int(size(OTHERSTATES),pInt) * ns
!*** determine indices to state array
l = 0_pInt
do t = 1_pInt,4_pInt
do s = 1_pInt,ns
l = l + 1_pInt
iRhoU(s,t,instance) = l
enddo
enddo
do t = 1_pInt,4_pInt
do s = 1_pInt,ns
l = l + 1_pInt
iRhoB(s,t,instance) = l
enddo
enddo
do c = 1_pInt,2_pInt
do s = 1_pInt,ns
l = l + 1_pInt
iRhoD(s,c,instance) = l
enddo
enddo
do s = 1_pInt,ns
l = l + 1_pInt
iGamma(s,instance) = l
enddo
do s = 1_pInt,ns
l = l + 1_pInt
iRhoF(s,instance) = l
enddo
do s = 1_pInt,ns
l = l + 1_pInt
iTauF(s,instance) = l
enddo
do s = 1_pInt,ns
l = l + 1_pInt
iTauB(s,instance) = l
enddo
do t = 1_pInt,4_pInt
do s = 1_pInt,ns
l = l + 1_pInt
iV(s,t,instance) = l
enddo
enddo
do c = 1_pInt,2_pInt
do s = 1_pInt,ns
l = l + 1_pInt
iD(s,c,instance) = l
enddo
enddo
if (iD(ns,2,instance) /= constitutive_nonlocal_sizeState(instance)) & ! check if last index is equal to size of state
call IO_error(0_pInt, ext_msg = 'state indices not properly set ('//PLASTICITY_NONLOCAL_label//')')
!*** determine size of postResults array
outputsLoop: do o = 1_pInt,Noutput(instance)
select case(constitutive_nonlocal_outputID(o,instance))
case( rho_ID, &
delta_ID, &
rho_edge_ID, &
rho_screw_ID, &
rho_sgl_ID, &
delta_sgl_ID, &
rho_sgl_edge_ID, &
rho_sgl_edge_pos_ID, &
rho_sgl_edge_neg_ID, &
rho_sgl_screw_ID, &
rho_sgl_screw_pos_ID, &
rho_sgl_screw_neg_ID, &
rho_sgl_mobile_ID, &
rho_sgl_edge_mobile_ID, &
rho_sgl_edge_pos_mobile_ID, &
rho_sgl_edge_neg_mobile_ID, &
rho_sgl_screw_mobile_ID, &
rho_sgl_screw_pos_mobile_ID, &
rho_sgl_screw_neg_mobile_ID, &
rho_sgl_immobile_ID, &
rho_sgl_edge_immobile_ID, &
rho_sgl_edge_pos_immobile_ID, &
rho_sgl_edge_neg_immobile_ID, &
rho_sgl_screw_immobile_ID, &
rho_sgl_screw_pos_immobile_ID, &
rho_sgl_screw_neg_immobile_ID, &
rho_dip_ID, &
delta_dip_ID, &
rho_dip_edge_ID, &
rho_dip_screw_ID, &
excess_rho_ID, &
excess_rho_edge_ID, &
excess_rho_screw_ID, &
rho_forest_ID, &
shearrate_ID, &
resolvedstress_ID, &
resolvedstress_external_ID, &
resolvedstress_back_ID, &
resistance_ID, &
rho_dot_ID, &
rho_dot_sgl_ID, &
rho_dot_sgl_mobile_ID, &
rho_dot_dip_ID, &
rho_dot_gen_ID, &
rho_dot_gen_edge_ID, &
rho_dot_gen_screw_ID, &
rho_dot_sgl2dip_ID, &
rho_dot_sgl2dip_edge_ID, &
rho_dot_sgl2dip_screw_ID, &
rho_dot_ann_ath_ID, &
rho_dot_ann_the_ID, &
rho_dot_ann_the_edge_ID, &
rho_dot_ann_the_screw_ID, &
rho_dot_edgejogs_ID, &
rho_dot_flux_ID, &
rho_dot_flux_mobile_ID, &
rho_dot_flux_edge_ID, &
rho_dot_flux_screw_ID, &
velocity_edge_pos_ID, &
velocity_edge_neg_ID, &
velocity_screw_pos_ID, &
velocity_screw_neg_ID, &
slipdirectionx_ID, &
slipdirectiony_ID, &
slipdirectionz_ID, &
slipnormalx_ID, &
slipnormaly_ID, &
slipnormalz_ID, &
fluxdensity_edge_posx_ID, &
fluxdensity_edge_posy_ID, &
fluxdensity_edge_posz_ID, &
fluxdensity_edge_negx_ID, &
fluxdensity_edge_negy_ID, &
fluxdensity_edge_negz_ID, &
fluxdensity_screw_posx_ID, &
fluxdensity_screw_posy_ID, &
fluxdensity_screw_posz_ID, &
fluxdensity_screw_negx_ID, &
fluxdensity_screw_negy_ID, &
fluxdensity_screw_negz_ID, &
maximumdipoleheight_edge_ID, &
maximumdipoleheight_screw_ID, &
accumulatedshear_ID )
mySize = totalNslip(instance)
case(dislocationstress_ID)
mySize = 6_pInt
case default
end select
if (mySize > 0_pInt) then ! any meaningful output found
constitutive_nonlocal_sizePostResult(o,instance) = mySize
constitutive_nonlocal_sizePostResults(instance) = constitutive_nonlocal_sizePostResults(instance) + mySize
endif
enddo outputsLoop
do s1 = 1_pInt,ns
f = slipFamily(s1,instance)
!*** burgers vector, mean free path prefactor and minimum dipole distance for each slip system
burgers(s1,instance) = burgersPerSlipFamily(f,instance)
lambda0(s1,instance) = lambda0PerSlipFamily(f,instance)
minDipoleHeight(s1,1:2,instance) = minDipoleHeightPerSlipFamily(f,1:2,instance)
peierlsStress(s1,1:2,instance) = peierlsStressPerSlipFamily(f,1:2,instance)
do s2 = 1_pInt,ns
!*** calculation of forest projections for edge and screw dislocations. s2 acts as forest for s1
forestProjectionEdge(s1,s2,instance) &
= abs(math_mul3x3(lattice_sn(1:3,slipSystemLattice(s1,instance),phase), &
lattice_st(1:3,slipSystemLattice(s2,instance),phase))) ! forest projection of edge dislocations is the projection of (t = b x n) onto the slip normal of the respective slip plane
forestProjectionScrew(s1,s2,instance) &
= abs(math_mul3x3(lattice_sn(1:3,slipSystemLattice(s1,instance),phase), &
lattice_sd(1:3,slipSystemLattice(s2,instance),phase))) ! forest projection of screw dislocations is the projection of b onto the slip normal of the respective splip plane
!*** calculation of interaction matrices
interactionMatrixSlipSlip(s1,s2,instance) &
= interactionSlipSlip(lattice_interactionSlipSlip(slipSystemLattice(s1,instance), &
slipSystemLattice(s2,instance), &
phase), instance)
!*** colinear slip system (only makes sense for fcc like it is defined here)
if (lattice_interactionSlipSlip(slipSystemLattice(s1,instance), &
slipSystemLattice(s2,instance), &
phase) == 3_pInt) then
colinearSystem(s1,instance) = s2
endif
enddo
!*** rotation matrix from lattice configuration to slip system
lattice2slip(1:3,1:3,s1,instance) &
= math_transpose33( reshape([ lattice_sd(1:3, slipSystemLattice(s1,instance), phase), &
-lattice_st(1:3, slipSystemLattice(s1,instance), phase), &
lattice_sn(1:3, slipSystemLattice(s1,instance), phase)], [3,3]))
enddo
!*** combined projection of Schmid and non-Schmid contributions to the resolved shear stress (only for screws)
!* four types t:
!* 1) positive screw at positive resolved stress
!* 2) positive screw at negative resolved stress
!* 3) negative screw at positive resolved stress
!* 4) negative screw at negative resolved stress
do s = 1_pInt,ns
do l = 1_pInt,lattice_NnonSchmid(phase)
nonSchmidProjection(1:3,1:3,1,s,instance) = nonSchmidProjection(1:3,1:3,1,s,instance) &
+ nonSchmidCoeff(l,instance) * lattice_Sslip(1:3,1:3,2*l,slipSystemLattice(s,instance),phase)
nonSchmidProjection(1:3,1:3,2,s,instance) = nonSchmidProjection(1:3,1:3,2,s,instance) &
+ nonSchmidCoeff(l,instance) * lattice_Sslip(1:3,1:3,2*l+1,slipSystemLattice(s,instance),phase)
enddo
nonSchmidProjection(1:3,1:3,3,s,instance) = -nonSchmidProjection(1:3,1:3,2,s,phase)
nonSchmidProjection(1:3,1:3,4,s,instance) = -nonSchmidProjection(1:3,1:3,1,s,phase)
forall (t = 1:4) &
nonSchmidProjection(1:3,1:3,t,s,instance) = nonSchmidProjection(1:3,1:3,t,s,phase) &
+ lattice_Sslip(1:3,1:3,1,slipSystemLattice(s,instance),phase)
enddo
endif
enddo initializeInstances
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
end subroutine constitutive_nonlocal_init
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief sets the initial microstructural state for a given instance of this plasticity
!--------------------------------------------------------------------------------------------------
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)
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, &
introduced named constants for homogenization_type, phase_plasticitiy, and phase_elasticity to replace expensive string comparison. Labels for different schemes need to be stored in material.f90 because they are needed when parsing the config file
2013-11-27 13:34:05 +05:30
phase_plasticity ,&
PLASTICITY_NONLOCAL_ID
* 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, &
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, &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
instance, &
renamed "maxNmatIDs" accordingly to "maxNinstances"
2014-03-04 19:17:04 +05:30
maxNinstances
* 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
renamed "maxNmatIDs" accordingly to "maxNinstances"
2014-03-04 19:17:04 +05:30
maxNinstances = int(count(phase_plasticity == PLASTICITY_NONLOCAL_ID),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)))
introduced named constants for homogenization_type, phase_plasticitiy, and phase_elasticity to replace expensive string comparison. Labels for different schemes need to be stored in material.f90 because they are needed when parsing the config file
2013-11-27 13:34:05 +05:30
if (PLASTICITY_NONLOCAL_ID == phase_plasticity(material_phase(1,i,e))) &
fixed bug in nonlocal state init that was introduced in rev 2438 and used to reset states from other constitutions
2013-05-24 19:13:44 +05:30
state(1,i,e)%p = 0.0_pReal
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
enddo
enddo
renamed "maxNmatIDs" accordingly to "maxNinstances"
2014-03-04 19:17:04 +05:30
do instance = 1_pInt,maxNinstances
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
ns = totalNslip(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
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
if (rhoSglRandom(instance) > 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)))
introduced named constants for homogenization_type, phase_plasticitiy, and phase_elasticity to replace expensive string comparison. Labels for different schemes need to be stored in material.f90 because they are needed when parsing the config file
2013-11-27 13:34:05 +05:30
if (PLASTICITY_NONLOCAL_ID == phase_plasticity(material_phase(1,i,e)) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
.and. instance == phase_plasticityInstance(material_phase(1,i,e))) then
use named state variable indices in "stateInit"
2013-05-24 02:40:31 +05:30
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
densityBinning = rhoSglRandomBinning(instance) / 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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
do while(meanDensity < rhoSglRandom(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
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)
introduced named constants for homogenization_type, phase_plasticitiy, and phase_elasticity to replace expensive string comparison. Labels for different schemes need to be stored in material.f90 because they are needed when parsing the config file
2013-11-27 13:34:05 +05:30
if (PLASTICITY_NONLOCAL_ID == phase_plasticity(material_phase(1,ip,el)) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
.and. instance == phase_plasticityInstance(material_phase(1,ip,el))) 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
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
state(1,ip,el)%p(iRhoU(s,t,instance)) = state(1,ip,el)%p(iRhoU(s,t,instance)) + 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)))
introduced named constants for homogenization_type, phase_plasticitiy, and phase_elasticity to replace expensive string comparison. Labels for different schemes need to be stored in material.f90 because they are needed when parsing the config file
2013-11-27 13:34:05 +05:30
if (PLASTICITY_NONLOCAL_ID == phase_plasticity(material_phase(1,i,e)) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
.and. instance == 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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
from = 1_pInt + sum(Nslip(1:f-1_pInt,instance))
upto = sum(Nslip(1:f,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
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
noise(j) = math_sampleGaussVar(0.0_pReal, rhoSglScatter(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
enddo
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
state(1,i,e)%p(iRhoU(s,1,instance)) = rhoSglEdgePos0(f,instance) + noise(1)
state(1,i,e)%p(iRhoU(s,2,instance)) = rhoSglEdgeNeg0(f,instance) + noise(1)
state(1,i,e)%p(iRhoU(s,3,instance)) = rhoSglScrewPos0(f,instance) + noise(2)
state(1,i,e)%p(iRhoU(s,4,instance)) = rhoSglScrewNeg0(f,instance) + 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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
state(1,i,e)%p(iRhoD(from:upto,1,instance)) = rhoDipEdge0(f,instance)
state(1,i,e)%p(iRhoD(from:upto,2,instance)) = rhoDipScrew0(f,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
enddo
endif
enddo
enddo
endif
enddo
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
end subroutine constitutive_nonlocal_stateInit
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief sets the relevant state values for a given instance of this plasticity
!--------------------------------------------------------------------------------------------------
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
pure function constitutive_nonlocal_aTolState(instance)
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
integer(pInt), intent(in) :: instance ! 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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
real(pReal), dimension(constitutive_nonlocal_sizeState(instance)) :: &
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
combined some state indices to an array with a more generic name
2013-05-24 14:32:30 +05:30
integer(pInt) :: ns, t, c
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
ns = totalNslip(instance)
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
combined some state indices to an array with a more generic name
2013-05-24 14:32:30 +05:30
forall (t = 1_pInt:4_pInt)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
constitutive_nonlocal_aTolState(iRhoU(1:ns,t,instance)) = aTolRho(instance)
constitutive_nonlocal_aTolState(iRhoB(1:ns,t,instance)) = aTolRho(instance)
combined some state indices to an array with a more generic name
2013-05-24 14:32:30 +05:30
endforall
forall (c = 1_pInt:2_pInt) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
constitutive_nonlocal_aTolState(iRhoD(1:ns,c,instance)) = aTolRho(instance)
constitutive_nonlocal_aTolState(iGamma(1:ns,instance)) = aTolShear(instance)
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
end function constitutive_nonlocal_aTolState
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief calculates quantities characterizing the microstructure
!--------------------------------------------------------------------------------------------------
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
subroutine constitutive_nonlocal_microstructure(state, Fe, Fp, gr, ip, el)
polishing
2013-05-23 17:55:56 +05:30
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, &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
lattice_st, &
lattice_mu, &
lattice_nu, &
lattice_structure, &
LATTICE_bcc_ID, &
LATTICE_fcc_ID
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
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +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
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
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
integer(pInt) neighbor_el, & ! element number of neighboring material point
neighbor_ip, & ! integration point of neighboring material point
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
instance, & ! my instance of this plasticity
neighbor_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
phase, &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
neighbor_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
ns, & ! total number of active slip systems at my material point
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
neighbor_ns, & ! total number of active slip systems at 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
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, &
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, &
correction of hardening coefficients also enabled for bcc for fcc: now applied to all hardening coefficients, not only hirth, lomer, and glissile
2013-08-14 17:14:30 +05:30
nRealNeighbors ! number of really existing neighbors
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
integer(pInt), dimension(2) :: neighbors
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) :: &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
neighbor_rhoExcess, & ! excess density at neighboring material point
neighbor_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)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
instance = phase_plasticityInstance(phase)
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
combined some state indices to an array with a more generic name
2013-05-24 14:32:30 +05:30
forall (s = 1_pInt:ns, t = 1_pInt:4_pInt)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoSgl(s,t) = max(state(gr,ip,el)%p(iRhoU(s,t,instance)), 0.0_pReal) ! ensure positive single mobile densities
rhoSgl(s,t+4_pInt) = state(gr,ip,el)%p(iRhoB(s,t,instance))
use named state variable indices in "microstructure"
2013-05-24 02:58:45 +05:30
endforall
combined some state indices to an array with a more generic name
2013-05-24 14:32:30 +05:30
forall (s = 1_pInt:ns, c = 1_pInt:2_pInt) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoDip(s,c) = max(state(gr,ip,el)%p(iRhoD(s,c,instance)), 0.0_pReal) ! ensure positive dipole densities
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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)), &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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)), &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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
correction of hardening coefficients also enabled for bcc for fcc: now applied to all hardening coefficients, not only hirth, lomer, and glissile
2013-08-14 17:14:30 +05:30
!*** 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)
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
make correction of hardening coefficients thread safe
2012-10-05 21:35:51 +05:30
myInteractionMatrix = 0.0_pReal
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
myInteractionMatrix(1:ns,1:ns) = interactionMatrixSlipSlip(1:ns,1:ns,instance)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
if (lattice_structure(phase) == LATTICE_bcc_ID .or. lattice_structure(phase) == LATTICE_fcc_ID) then ! only fcc and bcc
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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
correction of hardening coefficients also enabled for bcc for fcc: now applied to all hardening coefficients, not only hirth, lomer, and glissile
2013-08-14 17:14:30 +05:30
myInteractionMatrix(s,1:ns) = correction * myInteractionMatrix(s,1:ns)
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
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) &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
tauThreshold(s) = lattice_mu(phase) * 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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
neighbor_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))))
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
neighbor_el = mesh_ipNeighborhood(1,n,ip,el)
neighbor_ip = mesh_ipNeighborhood(2,n,ip,el)
if (neighbor_el > 0 .and. neighbor_ip > 0) then
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
neighbor_phase = material_phase(gr,neighbor_ip,neighbor_el)
neighbor_instance = phase_plasticityInstance(neighbor_phase)
neighbor_ns = totalNslip(neighbor_instance)
if (.not. phase_localPlasticity(neighbor_phase) &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
.and. neighbor_instance == instance) then ! same instance should be same structure
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
if (neighbor_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
combined some state indices to an array with a more generic name
2013-05-24 14:32:30 +05:30
forall (s = 1_pInt:ns, c = 1_pInt:2_pInt)
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
neighbor_rhoExcess(c,s,n) = &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
max(state(gr,neighbor_ip,neighbor_el)%p(iRhoU(s,2*c-1,neighbor_instance)), 0.0_pReal) &! positive mobiles
- max(state(gr,neighbor_ip,neighbor_el)%p(iRhoU(s,2*c,neighbor_instance)), 0.0_pReal) ! negative mobiles
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
neighbor_rhoTotal(c,s,n) = &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
max(state(gr,neighbor_ip,neighbor_el)%p(iRhoU(s,2*c-1,neighbor_instance)), 0.0_pReal) &! positive mobiles
+ max(state(gr,neighbor_ip,neighbor_el)%p(iRhoU(s,2*c,neighbor_instance)), 0.0_pReal) & ! negative mobiles
+ abs(state(gr,neighbor_ip,neighbor_el)%p(iRhoB(s,2*c-1,neighbor_instance))) & ! positive deads
+ abs(state(gr,neighbor_ip,neighbor_el)%p(iRhoB(s,2*c,neighbor_instance))) & ! negative deads
+ max(state(gr,neighbor_ip,neighbor_el)%p(iRhoD(s,c,neighbor_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) = &
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
math_mul33x3(invFe, mesh_ipCoordinates(1:3,neighbor_ip,neighbor_el) &
fixed bug in backstress calculation for elements with periodic neighbors
2013-05-23 18:06:48 +05:30
- mesh_ipCoordinates(1:3,ip,el))
normal_latticeConf = math_mul33x3(math_transpose33(invFp), mesh_ipAreaNormal(1:3,n,ip,el))
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
if (math_mul3x3(normal_latticeConf,connection_latticeConf(1:3,n)) < 0.0_pReal) then ! neighboring connection points in opposite direction to face normal: must be periodic image
fixed bug in backstress calculation for elements with periodic neighbors
2013-05-23 18:06:48 +05:30
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
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
neighbor_rhoExcess(1:2,1:ns,n) = rhoExcess
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
! free surface -> use central values instead
connection_latticeConf(1:3,n) = 0.0_pReal
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
neighbor_rhoExcess(1:2,1:ns,n) = rhoExcess
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
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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
m(1:3,1:ns,1) = lattice_sd(1:3,slipSystemLattice(1:ns,instance),phase)
m(1:3,1:ns,2) = -lattice_st(1:3,slipSystemLattice(1:ns,instance),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
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
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
neighbors(1) = 2_pInt * dir - 1_pInt
neighbors(2) = 2_pInt * dir
connections(dir,1:3) = connection_latticeConf(1:3,neighbors(1)) &
- connection_latticeConf(1:3,neighbors(2))
rhoExcessDifferences(dir) = neighbor_rhoExcess(c,s,neighbors(1)) &
- neighbor_rhoExcess(c,s,neighbors(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) &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
rhoTotal(c) = (sum(abs(rhoSgl(s,[2*c-1,2*c,2*c+3,2*c+4]))) + rhoDip(s,c) + sum(neighbor_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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
tauBack(s) = - lattice_mu(phase) * burgers(s,instance) / (2.0_pReal * pi) &
* (rhoExcessGradient_over_rho(1) / (1.0_pReal - lattice_nu(phase)) + 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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
state(gr,ip,el)%p(iRhoF(1:ns,instance)) = rhoForest
state(gr,ip,el)%p(iTauF(1:ns,instance)) = tauThreshold
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
end subroutine constitutive_nonlocal_microstructure
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief calculates kinetics
!--------------------------------------------------------------------------------------------------
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
subroutine constitutive_nonlocal_kinetics(v, dv_dtau, dv_dtauNS, tau, tauNS, &
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
tauThreshold, c, Temperature, ipc, 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
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
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
integer(pInt), intent(in) :: ipc, & !< current grain number
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
ip, & !< current integration point
el, & !< current element number
c !< dislocation character (1:edge, 2:screw)
real(pReal), intent(in) :: Temperature !< temperature
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,ip,el)))), &
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
intent(in) :: tau, & !< resolved external shear stress (without non Schmid effects)
tauNS, & !< resolved external shear stress (including non Schmid effects)
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
tauThreshold !< threshold shear stress
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: - 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 temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,ip,el)))), &
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
intent(out) :: v, & !< velocity
dv_dtau, & !< velocity derivative with respect to resolved shear stress (without non Schmid contributions)
dv_dtauNS !< velocity derivative with respect to resolved shear stress (including non Schmid contributions)
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
integer(pInt) :: instance, & !< current instance of this plasticity
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
ns, & !< short notation for the total number of active slip systems
s !< index of my current slip system
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, &
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
tauEff, & !< effective shear stress
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
tPeierls, & !< waiting time in front of a peierls barriers
tSolidSolution, & !< waiting time in front of a solid solution obstacle
vViscous, & !< viscous glide velocity
dtPeierls_dtau, & !< derivative with respect to resolved shear stress
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
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
instance = phase_plasticityInstance(material_phase(ipc,ip,el))
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
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
"dv_dtau" argument in "constitutive_nonlocal_kinetics" not optional anymore renamed "screwStressProjection" to "nonSchmidProjection"
2013-08-21 12:53:02 +05:30
dv_dtau = 0.0_pReal
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
dv_dtauNS = 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
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
if (abs(tau(s)) > tauThreshold(s)) then
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
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
!* Effective stress includes non Schmid constributions
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
!* 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
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
tauEff = max(0.0_pReal, abs(tauNS(s)) - tauThreshold(s)) ! ensure that the effective stress is positive
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
tauRel_P = min(1.0_pReal, tauEff / criticalStress_P) ! ensure that the activation probability cannot become greater than one
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
tPeierls = 1.0_pReal / fattack(instance) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
* exp(activationEnergy_P / (KB * Temperature) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
* (1.0_pReal - tauRel_P**pParam(instance))**qParam(instance))
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
if (tauEff < criticalStress_P) then
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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)
"dv_dtau" argument in "constitutive_nonlocal_kinetics" not optional anymore renamed "screwStressProjection" to "nonSchmidProjection"
2013-08-21 12:53:02 +05:30
else
dtPeierls_dtau = 0.0_pReal
* kinetics now according to Mohles; enables solid solution hardening * changed example in material.config accordingly
2011-01-26 15:47:42 +05:30
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
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
tauEff = abs(tau(s)) - tauThreshold(s)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
tauRel_S = min(1.0_pReal, tauEff / criticalStress_S) ! ensure that the activation probability cannot become greater than one
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
tSolidSolution = 1.0_pReal / fattack(instance) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
* exp(activationEnergy_S / (KB * Temperature) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
* (1.0_pReal - tauRel_S**pParam(instance))**qParam(instance))
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
if (tauEff < criticalStress_S) then
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
dtSolidSolution_dtau = tSolidSolution * pParam(instance) * qParam(instance) &
"dv_dtau" argument in "constitutive_nonlocal_kinetics" not optional anymore renamed "screwStressProjection" to "nonSchmidProjection"
2013-08-21 12:53:02 +05:30
* activationVolume_S / (KB * Temperature) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
* (1.0_pReal - tauRel_S**pParam(instance))**(qParam(instance)-1.0_pReal) &
* tauRel_S**(pParam(instance)-1.0_pReal)
"dv_dtau" argument in "constitutive_nonlocal_kinetics" not optional anymore renamed "screwStressProjection" to "nonSchmidProjection"
2013-08-21 12:53:02 +05:30
else
dtSolidSolution_dtau = 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
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
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
tauEff = abs(tau(s)) - tauThreshold(s)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
mobility = burgers(s,instance) / viscosity(instance)
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
vViscous = mobility * tauEff
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)
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
dv_dtau(s) = v(s) * v(s) * (dtSolidSolution_dtau / meanfreepath_S &
+ mobility / (vViscous * vViscous))
dv_dtauNS(s) = v(s) * v(s) * dtPeierls_dtau / meanfreepath_P
* 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 &
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == ipc)&
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,*)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
write(6,'(a,i8,1x,i2,1x,i1)') '<< CONST >> nonlocal_kinetics at el ip ipc',el,ip,ipc
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,*)
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
write(6,'(a,/,12x,12(f12.5,1x))') '<< CONST >> tauThreshold / MPa', tauThreshold / 1e6_pReal
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
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
write(6,'(a,/,12x,12(f12.5,1x))') '<< CONST >> tauNS / MPa', tauNS / 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
"dv_dtau" argument in "constitutive_nonlocal_kinetics" not optional anymore renamed "screwStressProjection" to "nonSchmidProjection"
2013-08-21 12:53:02 +05:30
write(6,'(a,/,12x,12(e12.5,1x))') '<< CONST >> dv_dtau', dv_dtau
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
write(6,'(a,/,12x,12(e12.5,1x))') '<< CONST >> dv_dtauNS', dv_dtauNS
* 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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
end subroutine constitutive_nonlocal_kinetics
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief calculates plastic velocity gradient and its tangent
!--------------------------------------------------------------------------------------------------
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
subroutine constitutive_nonlocal_LpAndItsTangent(Lp, dLp_dTstar99, Tstar_v, Temperature, state, ipc, 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, &
nonSchmid behavior now ready to use for bcc with phenopowerlaw and nonlocal constitutive model. nonSchmid tensors according to Koester,Ma,Hartmaier,2012. extended lattice_Sslip with non-Schmid tensors (the full non-symmetric tensors are required for the tangent and cannot be retrieved from the symmetrized Mandel notation)
2013-08-05 14:53:21 +05:30
math_mul33xx33, &
added non-schmid structure
2013-01-22 05:20:28 +05:30
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
removed a bunch of dummy functions (delta state for all constitutive_XXX except nonlocal, microstructure for none, j2 and phenopowerla), additional simplifications for none, averageBurgers is only relevant for RGC
2013-10-14 16:24:45 +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
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, &
nonSchmid behavior now ready to use for bcc with phenopowerlaw and nonlocal constitutive model. nonSchmid tensors according to Koester,Ma,Hartmaier,2012. extended lattice_Sslip with non-Schmid tensors (the full non-symmetric tensors are required for the tangent and cannot be retrieved from the symmetrized Mandel notation)
2013-08-05 14:53:21 +05:30
lattice_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
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
integer(pInt), intent(in) :: ipc, & !< current grain number
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
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
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
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
type(p_vec), intent(inout) :: 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
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
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
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)
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
integer(pInt) instance, & !< current instance of this plasticity
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
phase, & !< phase
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
ns, & !< short notation for the 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
i, &
j, &
k, &
l, &
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
t, & !< dislocation type
s, & !< index of my current slip system
sLattice !< index of my current slip system according to lattice order
real(pReal), dimension(3,3,3,3) :: dLp_dTstar3333 !< derivative of Lp with respect to Tstar (3x3x3x3 matrix)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,ip,el))),8) :: &
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
rhoSgl !< single dislocation densities (including blocked)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,ip,el))),4) :: &
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
v, & !< velocity
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
tauNS, & !< resolved shear stress including non Schmid and backstress terms
dv_dtau, & !< velocity derivative with respect to the shear stress
dv_dtauNS !< velocity derivative with respect to the shear stress
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,ip,el)))) :: &
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
tau, & !< resolved shear stress including backstress terms
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
gdotTotal, & !< shear rate
tauBack, & !< back stress from dislocation gradients on same slip system
removed obsolete "dead zone" scaling
2013-08-21 14:09:43 +05:30
tauThreshold !< threshold shear stress
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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
phase = material_phase(ipc,ip,el)
instance = phase_plasticityInstance(phase)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
ns = totalNslip(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
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
combined some state indices to an array with a more generic name
2013-05-24 14:32:30 +05:30
forall (s = 1_pInt:ns, t = 1_pInt:4_pInt)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoSgl(s,t) = max(state%p(iRhoU(s,t,instance)), 0.0_pReal) ! ensure positive single mobile densities
rhoSgl(s,t+4_pInt) = state%p(iRhoB(s,t,instance))
use named state variable indices in "kinetics" and "microstructure"
2013-05-24 03:10:00 +05:30
endforall
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
tauBack = state%p(iTauB(1:ns,instance))
tauThreshold = state%p(iTauF(1:ns,instance))
use named state variable indices in "kinetics" and "microstructure"
2013-05-24 03:10:00 +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
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
!*** get resolved shear stress
nonSchmid behavior now ready to use for bcc with phenopowerlaw and nonlocal constitutive model. nonSchmid tensors according to Koester,Ma,Hartmaier,2012. extended lattice_Sslip with non-Schmid tensors (the full non-symmetric tensors are required for the tangent and cannot be retrieved from the symmetrized Mandel notation)
2013-08-05 14:53:21 +05:30
!*** for screws possible non-schmid contributions are also taken into account
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
sLattice = slipSystemLattice(s,instance)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
tau(s) = math_mul6x6(Tstar_v, lattice_Sslip_v(1:6,1,sLattice,phase))
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
tauNS(s,1) = tau(s)
tauNS(s,2) = tau(s)
if (tau(s) > 0.0_pReal) then
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
tauNS(s,3) = math_mul33xx33(math_Mandel6to33(Tstar_v), nonSchmidProjection(1:3,1:3,1,s,instance))
tauNS(s,4) = math_mul33xx33(math_Mandel6to33(Tstar_v), nonSchmidProjection(1:3,1:3,3,s,instance))
nonSchmid behavior now ready to use for bcc with phenopowerlaw and nonlocal constitutive model. nonSchmid tensors according to Koester,Ma,Hartmaier,2012. extended lattice_Sslip with non-Schmid tensors (the full non-symmetric tensors are required for the tangent and cannot be retrieved from the symmetrized Mandel notation)
2013-08-05 14:53:21 +05:30
else
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
tauNS(s,3) = math_mul33xx33(math_Mandel6to33(Tstar_v), nonSchmidProjection(1:3,1:3,2,s,instance))
tauNS(s,4) = math_mul33xx33(math_Mandel6to33(Tstar_v), nonSchmidProjection(1:3,1:3,4,s,instance))
nonSchmid behavior now ready to use for bcc with phenopowerlaw and nonlocal constitutive model. nonSchmid tensors according to Koester,Ma,Hartmaier,2012. extended lattice_Sslip with non-Schmid tensors (the full non-symmetric tensors are required for the tangent and cannot be retrieved from the symmetrized Mandel notation)
2013-08-05 14:53:21 +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
enddo
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
forall (t = 1_pInt:4_pInt) &
tauNS(1:ns,t) = tauNS(1:ns,t) + tauBack ! add backstress
tau = tau + tauBack ! add backstress
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 dislocation velocity and its tangent and store the velocity in the state array
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
! edges
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
call constitutive_nonlocal_kinetics(v(1:ns,1), dv_dtau(1:ns,1), dv_dtauNS(1:ns,1), &
tau(1:ns), tauNS(1:ns,1), tauThreshold(1:ns), &
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
1_pInt, Temperature, ipc, ip, el)
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
v(1:ns,2) = v(1:ns,1)
dv_dtau(1:ns,2) = dv_dtau(1:ns,1)
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
dv_dtauNS(1:ns,2) = dv_dtauNS(1:ns,1)
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
!screws
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
if (lattice_NnonSchmid(phase) == 0_pInt) then ! no non-Schmid contributions
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
forall(t = 3_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)
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
dv_dtauNS(1:ns,t) = dv_dtauNS(1:ns,1)
rearranged arguments of "constitutive_nonlocal_kinetics", got "tauTreshold" as argument, not complete state don't call "constitutive_nonlocal_kinetics" twice for edges in case of nonSchmid behavior, but just call once and copy results from positive to negative edges
2013-08-21 13:25:34 +05:30
endforall
else ! take non-Schmid contributions into account
do t = 3_pInt,4_pInt
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
call constitutive_nonlocal_kinetics(v(1:ns,t), dv_dtau(1:ns,t), dv_dtauNS(1:ns,t), &
tau(1:ns), tauNS(1:ns,t), tauThreshold(1:ns), &
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
2_pInt , Temperature, ipc, ip, el)
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
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
dislocation velocity was not correctly stored to state in last version
2013-08-21 19:24:36 +05:30
!*** store velocity in state
forall (t = 1_pInt:4_pInt) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
state%p(iV(1:ns,t,instance)) = v(1:ns,t)
dislocation velocity was not correctly stored to state in last version
2013-08-21 19:24:36 +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
!*** 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 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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
gdotTotal = sum(rhoSgl(1:ns,1:4) * v, 2) * burgers(1:ns,instance)
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +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 s = 1_pInt,ns
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
sLattice = slipSystemLattice(s,instance)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
Lp = Lp + gdotTotal(s) * lattice_Sslip(1:3,1:3,1,sLattice,phase)
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
! Schmid contributions to tangent
forall (i=1_pInt:3_pInt,j=1_pInt:3_pInt,k=1_pInt:3_pInt,l=1_pInt:3_pInt) &
dLp_dTstar3333(i,j,k,l) = dLp_dTstar3333(i,j,k,l) &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
+ lattice_Sslip(i,j,1,sLattice,phase) * lattice_Sslip(k,l,1,sLattice,phase) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
* sum(rhoSgl(s,1:4) * dv_dtau(s,1:4)) * burgers(s,instance)
in kinetics: non schmid stresses only influence peierls mechanism, but not solid solution hardening; as a result the derivative of the velocity with respect to the resolved stress has to be split into a Schmid and a non-Schmid part
2013-08-21 17:51:52 +05:30
! non Schmid contributions to tangent
if (tau(s) > 0.0_pReal) then
nonSchmid behavior now ready to use for bcc with phenopowerlaw and nonlocal constitutive model. nonSchmid tensors according to Koester,Ma,Hartmaier,2012. extended lattice_Sslip with non-Schmid tensors (the full non-symmetric tensors are required for the tangent and cannot be retrieved from the symmetrized Mandel notation)
2013-08-05 14:53:21 +05:30
forall (i=1_pInt:3_pInt,j=1_pInt:3_pInt,k=1_pInt:3_pInt,l=1_pInt:3_pInt) &
dLp_dTstar3333(i,j,k,l) = dLp_dTstar3333(i,j,k,l) &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
+ lattice_Sslip(i,j,1,sLattice,phase) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
* ( nonSchmidProjection(k,l,1,s,instance) * rhoSgl(s,3) * dv_dtauNS(s,3) &
+ nonSchmidProjection(k,l,3,s,instance) * rhoSgl(s,4) * dv_dtauNS(s,4) ) &
* burgers(s,instance)
nonSchmid behavior now ready to use for bcc with phenopowerlaw and nonlocal constitutive model. nonSchmid tensors according to Koester,Ma,Hartmaier,2012. extended lattice_Sslip with non-Schmid tensors (the full non-symmetric tensors are required for the tangent and cannot be retrieved from the symmetrized Mandel notation)
2013-08-05 14:53:21 +05:30
else
forall (i=1_pInt:3_pInt,j=1_pInt:3_pInt,k=1_pInt:3_pInt,l=1_pInt:3_pInt) &
dLp_dTstar3333(i,j,k,l) = dLp_dTstar3333(i,j,k,l) &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
+ lattice_Sslip(i,j,1,sLattice,phase) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
* ( nonSchmidProjection(k,l,2,s,instance) * rhoSgl(s,3) * dv_dtauNS(s,3) &
+ nonSchmidProjection(k,l,4,s,instance) * rhoSgl(s,4) * dv_dtauNS(s,4) ) &
* burgers(s,instance)
nonSchmid behavior now ready to use for bcc with phenopowerlaw and nonlocal constitutive model. nonSchmid tensors according to Koester,Ma,Hartmaier,2012. extended lattice_Sslip with non-Schmid tensors (the full non-symmetric tensors are required for the tangent and cannot be retrieved from the symmetrized Mandel notation)
2013-08-05 14:53:21 +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
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 &
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == ipc)&
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,*)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
write(6,'(a,i8,1x,i2,1x,i1)') '<< CONST >> nonlocal_LpandItsTangent at el ip ipc ',el,ip,ipc
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
end subroutine constitutive_nonlocal_LpAndItsTangent
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief (instantaneous) incremental change of microstructure
!--------------------------------------------------------------------------------------------------
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
subroutine constitutive_nonlocal_deltaState(deltaState, state, Tstar_v, ipc,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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
use lattice, only: lattice_Sslip_v ,&
lattice_mu, &
lattice_nu
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
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
integer(pInt), intent(in) :: ipc, & ! current grain number
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
ip, & ! current integration point
el ! current element number
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 reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
integer(pInt) phase, &
instance, & ! current instance of this plasticity
moved remobilization of dead dislocations from dotState to deltaState
2012-05-18 19:05:44 +05:30
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 temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,ip,el))),10) :: &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
deltaRho, & ! density increment
deltaRhoRemobilization, & ! density increment by remobilization
deltaRhoDipole2SingleStress ! density increment by dipole dissociation (by stress change)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,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 temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,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 temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,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 temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,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 &
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == ipc)&
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,*)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
write(6,'(a,i8,1x,i2,1x,i1)') '<< CONST >> nonlocal_deltaState at el ip ipc ',el,ip,ipc
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 reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
phase = material_phase(ipc,ip,el)
instance = phase_plasticityInstance(phase)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
ns = totalNslip(instance)
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
combined some state indices to an array with a more generic name
2013-05-24 14:32:30 +05:30
forall (s = 1_pInt:ns, t = 1_pInt:4_pInt)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
rhoSgl(s,t) = max(state(ipc,ip,el)%p(iRhoU(s,t,instance)), 0.0_pReal) ! ensure positive single mobile densities
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoSgl(s,t+4_pInt) = state(ipc,ip,el)%p(iRhoB(s,t,instance))
v(s,t) = state(ipc,ip,el)%p(iV(s,t,instance))
use named state variable indices in "deltaState"
2013-05-24 03:16:15 +05:30
endforall
combined some state indices to an array with a more generic name
2013-05-24 14:32:30 +05:30
forall (s = 1_pInt:ns, c = 1_pInt:2_pInt)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
rhoDip(s,c) = max(state(ipc,ip,el)%p(iRhoD(s,c,instance)), 0.0_pReal) ! ensure positive dipole densities
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
dUpperOld(s,c) = state(ipc,ip,el)%p(iD(s,c,instance))
use named state variable indices in "deltaState"
2013-05-24 03:16:15 +05:30
endforall
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
tauBack = state(ipc,ip,el)%p(iTauB(1:ns,instance))
combined some state indices to an array with a more generic name
2013-05-24 14:32:30 +05:30
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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
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
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
sLattice = slipSystemLattice(s,instance)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
tau(s) = math_mul6x6(Tstar_v, lattice_Sslip_v(1:6,1,sLattice,phase)) + 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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
dLower = minDipoleHeight(1:ns,1:2,instance)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
dUpper(1:ns,1) = lattice_mu(phase) * burgers(1:ns,instance) &
/ (8.0_pReal * pi * (1.0_pReal - lattice_nu(phase)) * abs(tau))
dUpper(1:ns,2) = lattice_mu(phase) * burgers(1:ns,instance) / (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) &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
deltaRhoDipole2SingleStress(s,8_pInt+c) = rhoDip(s,c) * deltaDUpper(s,c) / (dUpperOld(s,c) - dLower(s,c))
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 (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
combined some state indices to an array with a more generic name
2013-05-24 14:32:30 +05:30
forall (s = 1_pInt:ns, c = 1_pInt:2_pInt) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
state(ipc,ip,el)%p(iD(s,c,instance)) = dUpper(s,c)
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
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
fixed bug in deltaState; used to lead to unpredictable values in accumulated shear, since deltaState was not defined at this state position
2013-06-12 02:33:23 +05:30
deltaState%p = 0.0_pReal
forall (s = 1:ns, t = 1_pInt:4_pInt)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
deltaState%p(iRhoU(s,t,instance)) = deltaRho(s,t)
deltaState%p(iRhoB(s,t,instance)) = deltaRho(s,t+4_pInt)
fixed bug in deltaState; used to lead to unpredictable values in accumulated shear, since deltaState was not defined at this state position
2013-06-12 02:33:23 +05:30
endforall
forall (s = 1:ns, c = 1_pInt:2_pInt) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
deltaState%p(iRhoD(s,c,instance)) = deltaRho(s,c+8_pInt)
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 &
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == ipc)&
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 reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
write(6,'(a,/,8(12x,12(e12.5,1x),/))') '<< CONST >> dislocation remobilization', deltaRhoRemobilization(1:ns,1:8)
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
end subroutine constitutive_nonlocal_deltaState
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief calculates the rate of change of microstructure
!--------------------------------------------------------------------------------------------------
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
function constitutive_nonlocal_dotState(Tstar_v, Fe, Fp, Temperature, state, state0, timestep, subfrac, ipc,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, &
introduced named constants for homogenization_type, phase_plasticitiy, and phase_elasticity to replace expensive string comparison. Labels for different schemes need to be stored in material.f90 because they are needed when parsing the config file
2013-11-27 13:34:05 +05:30
phase_plasticity ,&
PLASTICITY_NONLOCAL_ID
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, &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
lattice_st ,&
lattice_mu, &
lattice_nu, &
lattice_structure, &
LATTICE_bcc_ID, &
LATTICE_fcc_ID
* 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
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
integer(pInt), intent(in) :: ipc, & !< current grain number
line break in lines > 132
2013-01-22 16:36:39 +05:30
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
!*** output variables
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(constitutive_nonlocal_sizeDotState(phase_plasticityInstance(material_phase(ipc,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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
integer(pInt) :: phase, &
instance, & !< current instance of this plasticity
neighbor_instance, & !< instance of my neighbor's plasticity
line break in lines > 132
2013-01-22 16:36:39 +05:30
ns, & !< short notation for the total number of active slip systems
c, & !< character of dislocation
n, & !< index of my current neighbor
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
neighbor_el, & !< element number of my neighbor
neighbor_ip, & !< integration point of my neighbor
neighbor_n, & !< neighbor index pointing to me when looking from my neighbor
line break in lines > 132
2013-01-22 16:36:39 +05:30
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
blocked density does no longer contribute to dislocation flux; used to generate source term for mobile density if burgers vector had any non-zero component perpendicular to blocking boundary.
2014-01-08 22:01:50 +05:30
sLattice !< index of my current slip system according to lattice order
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,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 temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,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, &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
neighbor_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)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
my_rhoSgl !< single dislocation densities of central ip (positive/negative screw and edge without dipoles)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
my_v, & !< dislocation glide velocity of central ip
neighbor_v, & !< dislocation glide velocity of enighboring ip
line break in lines > 132
2013-01-22 16:36:39 +05:30
gdot !< shear rates
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,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 temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,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 temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(3,totalNslip(phase_plasticityInstance(material_phase(ipc,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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
neighbor_F, & !< total deformation gradient of my neighbor
line break in lines > 132
2013-01-22 16:36:39 +05:30
my_Fe, & !< my elastic deformation gradient
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
neighbor_Fe, & !< elastic deformation gradient of my neighbor
line break in lines > 132
2013-01-22 16:36:39 +05:30
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 &
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == ipc)&
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,*)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
write(6,'(a,i8,1x,i2,1x,i1)') '<< CONST >> nonlocal_dotState at el ip ipc ',el,ip,ipc
- 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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
phase = material_phase(ipc,ip,el)
instance = phase_plasticityInstance(phase)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
ns = totalNslip(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
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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
forall (s = 1_pInt:ns, t = 1_pInt:4_pInt)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
rhoSgl(s,t) = max(state(ipc,ip,el)%p(iRhoU(s,t,instance)), 0.0_pReal) ! ensure positive single mobile densities
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoSgl(s,t+4_pInt) = state(ipc,ip,el)%p(iRhoB(s,t,instance))
v(s,t) = state(ipc,ip,el)%p(iV(s,t,instance))
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
endforall
forall (s = 1_pInt:ns, c = 1_pInt:2_pInt)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
rhoDip(s,c) = max(state(ipc,ip,el)%p(iRhoD(s,c,instance)), 0.0_pReal) ! ensure positive dipole densities
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
endforall
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoForest = state(ipc,ip,el)%p(iRhoF(1:ns,instance))
tauThreshold = state(ipc,ip,el)%p(iTauF(1:ns,instance))
tauBack = state(ipc,ip,el)%p(iTauB(1:ns,instance))
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
if (numerics_timeSyncing) then
fixed bug in deltaState; used to lead to unpredictable values in accumulated shear, since deltaState was not defined at this state position
2013-06-12 02:33:23 +05:30
forall (s = 1_pInt:ns, t = 1_pInt:4_pInt)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoSgl0(s,t) = max(state0(ipc,ip,el)%p(iRhoU(s,t,instance)), 0.0_pReal)
rhoSgl0(s,t+4_pInt) = state0(ipc,ip,el)%p(iRhoB(s,t,instance))
v0(s,t) = state0(ipc,ip,el)%p(iV(s,t,instance))
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
endforall
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
where (abs(rhoSgl0) * mesh_ipVolume(ip,el) ** 0.667_pReal < significantN(instance) &
.or. abs(rhoSgl0) < significantRho(instance)) &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
rhoSgl0 = 0.0_pReal
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
- 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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
if (timestep <= 0.0_pReal) then ! if illegal timestep...
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) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
gdot(1_pInt:ns,t) = rhoSgl(1_pInt:ns,t) * burgers(1:ns,instance) * 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 &
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == ipc)&
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
sLattice = slipSystemLattice(s,instance)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
tau(s) = math_mul6x6(Tstar_v, lattice_Sslip_v(1:6,1,sLattice,phase)) + 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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
dLower = minDipoleHeight(1:ns,1:2,instance)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
dUpper(1:ns,1) = lattice_mu(phase) * burgers(1:ns,instance) &
/ (8.0_pReal * pi * (1.0_pReal - lattice_nu(phase)) * abs(tau))
dUpper(1:ns,2) = lattice_mu(phase) * burgers(1:ns,instance) &
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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
if (lattice_structure(phase) == LATTICE_bcc_ID) then ! BCC
multiplication now sensitive of crystal structure, yet multiplication in bcc identical to fcc except for the contribution of edges which is completely dropped for bcc
2013-08-05 14:56:37 +05:30
forall (s = 1:ns, sum(abs(v(s,1:4))) > 0.0_pReal)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
rhoDotMultiplication(s,1:2) = sum(abs(gdot(s,3:4))) / burgers(s,instance) & ! assuming double-cross-slip of screws to be decisive for multiplication
* sqrt(rhoForest(s)) / lambda0(s,instance) ! & ! mean free path
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
! * 2.0_pReal * sum(abs(v(s,3:4))) / sum(abs(v(s,1:4))) ! ratio of screw to overall velocity determines edge generation
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
rhoDotMultiplication(s,3:4) = sum(abs(gdot(s,3:4))) / burgers(s,instance) & ! assuming double-cross-slip of screws to be decisive for multiplication
* sqrt(rhoForest(s)) / lambda0(s,instance) ! & ! mean free path
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
! * 2.0_pReal * sum(abs(v(s,1:2))) / sum(abs(v(s,1:4))) ! ratio of edge to overall velocity determines screw generation
multiplication now sensitive of crystal structure, yet multiplication in bcc identical to fcc except for the contribution of edges which is completely dropped for bcc
2013-08-05 14:56:37 +05:30
endforall
else ! ALL OTHER STRUCTURES
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
if (probabilisticMultiplication(instance)) then
multiplication now sensitive of crystal structure, yet multiplication in bcc identical to fcc except for the contribution of edges which is completely dropped for bcc
2013-08-05 14:56:37 +05:30
meshlength = mesh_ipVolume(ip,el)**0.333_pReal
where(sum(rhoSgl(1:ns,1:4),2) > 0.0_pReal)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
nSources = (sum(rhoSgl(1:ns,1:2),2) * fEdgeMultiplication(instance) + sum(rhoSgl(1:ns,3:4),2)) &
/ sum(rhoSgl(1:ns,1:4),2) * meshlength / lambda0(1:ns,instance) * sqrt(rhoForest(1:ns))
multiplication now sensitive of crystal structure, yet multiplication in bcc identical to fcc except for the contribution of edges which is completely dropped for bcc
2013-08-05 14:56:37 +05:30
elsewhere
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
nSources = meshlength / lambda0(1:ns,instance) * sqrt(rhoForest(1:ns))
multiplication now sensitive of crystal structure, yet multiplication in bcc identical to fcc except for the contribution of edges which is completely dropped for bcc
2013-08-05 14:56:37 +05:30
endwhere
do s = 1_pInt,ns
if (nSources(s) < 1.0_pReal) then
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
if (sourceProbability(s,ipc,ip,el) > 1.0_pReal) then
multiplication now sensitive of crystal structure, yet multiplication in bcc identical to fcc except for the contribution of edges which is completely dropped for bcc
2013-08-05 14:56:37 +05:30
call random_number(rnd)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
sourceProbability(s,ipc,ip,el) = rnd
multiplication now sensitive of crystal structure, yet multiplication in bcc identical to fcc except for the contribution of edges which is completely dropped for bcc
2013-08-05 14:56:37 +05:30
!$OMP FLUSH(sourceProbability)
endif
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
if (sourceProbability(s,ipc,ip,el) > 1.0_pReal - nSources(s)) then
multiplication now sensitive of crystal structure, yet multiplication in bcc identical to fcc except for the contribution of edges which is completely dropped for bcc
2013-08-05 14:56:37 +05:30
rhoDotMultiplication(s,1:4) = sum(rhoSglOriginal(s,1:4) * abs(v(s,1:4))) / meshlength
endif
else
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
sourceProbability(s,ipc,ip,el) = 2.0_pReal
multiplication now sensitive of crystal structure, yet multiplication in bcc identical to fcc except for the contribution of edges which is completely dropped for bcc
2013-08-05 14:56:37 +05:30
rhoDotMultiplication(s,1:4) = &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
(sum(abs(gdot(s,1:2))) * fEdgeMultiplication(instance) + sum(abs(gdot(s,3:4)))) &
/ burgers(s,instance) * sqrt(rhoForest(s)) / lambda0(s,instance)
made random multiplication switchable via material.config
2012-12-03 18:29:38 +05:30
endif
multiplication now sensitive of crystal structure, yet multiplication in bcc identical to fcc except for the contribution of edges which is completely dropped for bcc
2013-08-05 14:56:37 +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
multiplication now sensitive of crystal structure, yet multiplication in bcc identical to fcc except for the contribution of edges which is completely dropped for bcc
2013-08-05 14:56:37 +05:30
if (iand(debug_level(debug_constitutive),debug_levelExtensive) /= 0_pInt &
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == ipc)&
multiplication now sensitive of crystal structure, yet multiplication in bcc identical to fcc except for the contribution of edges which is completely dropped for bcc
2013-08-05 14:56:37 +05:30
.or. .not. iand(debug_level(debug_constitutive),debug_levelSelective) /= 0_pInt )) then
write(6,'(a,/,4(12x,12(f12.5,1x),/))') '<< CONST >> sources', nSources
write(6,*)
endif
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
#endif
multiplication now sensitive of crystal structure, yet multiplication in bcc identical to fcc except for the contribution of edges which is completely dropped for bcc
2013-08-05 14:56:37 +05:30
else
rhoDotMultiplication(1:ns,1:4) = spread( &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
(sum(abs(gdot(1:ns,1:2)),2) * fEdgeMultiplication(instance) + sum(abs(gdot(1:ns,3:4)),2)) &
* sqrt(rhoForest(1:ns)) / lambda0(1:ns,instance) / burgers(1:ns,instance), 2, 4)
multiplication now sensitive of crystal structure, yet multiplication in bcc identical to fcc except for the contribution of edges which is completely dropped for bcc
2013-08-05 14:56:37 +05:30
endif
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
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
if (.not. phase_localPlasticity(material_phase(ipc,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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
if (any( abs(gdot) > 0.0_pReal & ! any active slip system ...
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
.and. CFLfactor(instance) * abs(v) * timestep &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +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)
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
#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 ', &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
maxval(abs(v), abs(gdot) > 0.0_pReal &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
.and. CFLfactor(instance) * abs(v) * timestep &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
constitutive_nonlocal_dotState = DAMASK_NaN ! -> return NaN and, hence, enforce cutback
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
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
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 !!!
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
m(1:3,1:ns,1) = lattice_sd(1:3, slipSystemLattice(1:ns,instance), phase)
m(1:3,1:ns,2) = -lattice_sd(1:3, slipSystemLattice(1:ns,instance), phase)
m(1:3,1:ns,3) = -lattice_st(1:3, slipSystemLattice(1:ns,instance), phase)
m(1:3,1:ns,4) = lattice_st(1:3, slipSystemLattice(1:ns,instance), phase)
* had to correct nonlocal slip system compatibility and flux calculation. last update gave wrong results.
2010-10-15 18:49:26 +05:30
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
my_Fe = Fe(1:3,1:3,ipc,ip,el)
my_F = math_mul33x33(my_Fe, Fp(1:3,1:3,ipc,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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
do n = 1_pInt,FE_NipNeighbors(FE_celltype(FE_geomtype(mesh_element(2,el)))) ! loop through my neighbors
neighbor_el = mesh_ipNeighborhood(1,n,ip,el)
neighbor_ip = mesh_ipNeighborhood(2,n,ip,el)
neighbor_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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
if (neighbor_n > 0_pInt) then ! if neighbor exists, average deformation gradient
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
neighbor_instance = phase_plasticityInstance(material_phase(ipc,neighbor_ip,neighbor_el))
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
neighbor_Fe = Fe(1:3,1:3,ipc,neighbor_ip,neighbor_el)
neighbor_F = math_mul33x33(neighbor_Fe, Fp(1:3,1:3,ipc,neighbor_ip,neighbor_el))
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
Favg = 0.5_pReal * (my_F + neighbor_F)
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.
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
neighbor_v = 0.0_pReal ! needed for check of sign change in flux density below
neighbor_rhoSgl = 0.0_pReal
if (neighbor_n > 0_pInt) then
introduced named constants for homogenization_type, phase_plasticitiy, and phase_elasticity to replace expensive string comparison. Labels for different schemes need to be stored in material.f90 because they are needed when parsing the config file
2013-11-27 13:34:05 +05:30
if (phase_plasticity(material_phase(1,neighbor_ip,neighbor_el)) == PLASTICITY_NONLOCAL_ID &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
.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
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
if(numerics_timeSyncing .and. (subfrac(ipc,neighbor_ip,neighbor_el) /= subfrac(ipc,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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
forall (s = 1:ns, t = 1_pInt:4_pInt)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
neighbor_v(s,t) = state0(ipc,neighbor_ip,neighbor_el)%p(iV(s,t,neighbor_instance))
neighbor_rhoSgl(s,t) = max(state0(ipc,neighbor_ip,neighbor_el)%p(iRhoU(s,t,neighbor_instance)), 0.0_pReal)
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
endforall
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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
forall (s = 1:ns, t = 1_pInt:4_pInt)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
neighbor_v(s,t) = state(ipc,neighbor_ip,neighbor_el)%p(iV(s,t,neighbor_instance))
neighbor_rhoSgl(s,t) = max(state(ipc,neighbor_ip,neighbor_el)%p(iRhoU(s,t,neighbor_instance)), 0.0_pReal)
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
endforall
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
where (neighbor_rhoSgl * mesh_ipVolume(neighbor_ip,neighbor_el) ** 0.667_pReal < significantN(instance) &
.or. neighbor_rhoSgl < significantRho(instance)) &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
neighbor_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)), &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
mesh_ipAreaNormal(1:3,neighbor_n,neighbor_ip,neighbor_el)) ! calculate the normal of the interface in (average) deformed configuration (now pointing from my neighbor to me!!!)
normal_neighbor2me = math_mul33x3(transpose(neighbor_Fe), normal_neighbor2me_defConf) &
/ math_det33(neighbor_Fe) ! interface normal in the lattice configuration of my neighbor
area = mesh_ipArea(neighbor_n,neighbor_ip,neighbor_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)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
if (neighbor_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
allow flux between neighboring material points if one of both got a velocity of zero
2014-01-08 22:05:10 +05:30
.and. v(s,t) * neighbor_v(s,t) >= 0.0_pReal ) then ! ... only if no sign change in flux density
blocked density does no longer contribute to dislocation flux; used to generate source term for mobile density if burgers vector had any non-zero component perpendicular to blocking boundary.
2014-01-08 22:01:50 +05:30
lineLength = neighbor_rhoSgl(s,t) * neighbor_v(s,t) &
* math_mul3x3(m(1:3,s,t), normal_neighbor2me) * area ! positive line length that wants to enter through this interface
where (compatibility(c,1_pInt:ns,s,n,ip,el) > 0.0_pReal) & ! positive compatibility...
rhoDotFlux(1_pInt:ns,t) = rhoDotFlux(1_pInt:ns,t) &
+ lineLength / mesh_ipVolume(ip,el) & ! ... transferring to equally signed mobile dislocation type
* 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...
rhoDotFlux(1_pInt:ns,topp) = rhoDotFlux(1_pInt:ns,topp) &
+ lineLength / mesh_ipVolume(ip,el) & ! ... transferring to opposite signed mobile dislocation type
* compatibility(c,1_pInt:ns,s,n,ip,el) ** 2.0_pReal
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
introduced named constants for homogenization_type, phase_plasticitiy, and phase_elasticity to replace expensive string comparison. Labels for different schemes need to be stored in material.f90 because they are needed when parsing the config file
2013-11-27 13:34:05 +05:30
if (phase_plasticity(material_phase(1,opposite_ip,opposite_el)) /= PLASTICITY_NONLOCAL_ID) &
* 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.
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
my_rhoSgl = rhoSgl
my_v = v
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) then
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
if (subfrac(ipc,ip,el) == 0.0_pReal) then
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
my_rhoSgl = rhoSgl0
my_v = v0
elseif (neighbor_n > 0_pInt) then
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
if (subfrac(ipc,neighbor_ip,neighbor_el) == 0.0_pReal) then
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
my_rhoSgl = rhoSgl0
my_v = 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
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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +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!!!)
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)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
if (my_v(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)
allow flux between neighboring material points if one of both got a velocity of zero
2014-01-08 22:05:10 +05:30
if (my_v(s,t) * neighbor_v(s,t) >= 0.0_pReal) then ! no sign change in flux density
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +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
else ! sign change in flux density means sign change in stress which does not allow for dislocations to arive at the 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
transmissivity = 0.0_pReal
endif
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
lineLength = my_rhoSgl(s,t) * my_v(s,t) &
* math_mul3x3(m(1:3,s,t), normal_me2neighbor) * area ! positive line length of mobiles that wants to leave through this interface
rhoDotFlux(s,t) = rhoDotFlux(s,t) - lineLength / mesh_ipVolume(ip,el) ! subtract dislocation flux from current type
rhoDotFlux(s,t+4_pInt) = rhoDotFlux(s,t+4_pInt) &
+ lineLength / mesh_ipVolume(ip,el) * (1.0_pReal - transmissivity) &
* sign(1.0_pReal, my_v(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
* 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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoDotSingle2DipoleGlide(1:ns,2*c-1) = -2.0_pReal * dUpper(1:ns,c) / burgers(1:ns,instance) &
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoDotSingle2DipoleGlide(1:ns,2*c) = -2.0_pReal * dUpper(1:ns,c) / burgers(1:ns,instance) &
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoDotSingle2DipoleGlide(1:ns,2*c+3) = -2.0_pReal * dUpper(1:ns,c) / burgers(1:ns,instance) &
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoDotSingle2DipoleGlide(1:ns,2*c+4) = -2.0_pReal * dUpper(1:ns,c) / burgers(1:ns,instance) &
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
renamed "maxNmatIDs" accordingly to "maxNinstances"
2014-03-04 19:17:04 +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) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoDotAthermalAnnihilation(1:ns,c+8_pInt) = -2.0_pReal * dLower(1:ns,c) / burgers(1:ns,instance) &
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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
if (lattice_structure(phase) == LATTICE_fcc_ID) then ! only fcc
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
forall (s = 1:ns, colinearSystem(s,instance) > 0_pInt) &
rhoDotAthermalAnnihilation(colinearSystem(s,instance),1:2) = - rhoDotAthermalAnnihilation(s,10) &
* 0.25_pReal * sqrt(rhoForest(s)) * (dUpper(s,2) + dLower(s,2)) * edgeJogFactor(instance)
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
selfDiffusion = Dsd0(instance) * exp(-selfDiffusionEnergy(instance) / (KB * Temperature))
vClimb = atomicVolume(instance) * selfDiffusion / ( KB * Temperature ) &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
* lattice_mu(phase) / ( 2.0_pReal * PI * (1.0_pReal-lattice_nu(phase)) ) &
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)), &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
- 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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
if (numerics_integrationMode == 1_pInt) then ! save rates for output if in central integration mode
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
rhoDotFluxOutput(1:ns,1:8,ipc,ip,el) = rhoDotFlux(1:ns,1:8)
rhoDotMultiplicationOutput(1:ns,1:2,ipc,ip,el) = rhoDotMultiplication(1:ns,[1,3])
rhoDotSingle2DipoleGlideOutput(1:ns,1:2,ipc,ip,el) = rhoDotSingle2DipoleGlide(1:ns,9:10)
rhoDotAthermalAnnihilationOutput(1:ns,1:2,ipc,ip,el) = rhoDotAthermalAnnihilation(1:ns,9:10)
rhoDotThermalAnnihilationOutput(1:ns,1:2,ipc,ip,el) = rhoDotThermalAnnihilation(1:ns,9:10)
rhoDotEdgeJogsOutput(1:ns,ipc,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 &
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
.and. ((debug_e == el .and. debug_i == ip .and. debug_g == ipc)&
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 reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
if ( any(rhoSglOriginal(1:ns,1:4) + rhoDot(1:ns,1:4) * timestep < -aTolRho(instance)) &
.or. any(rhoDipOriginal(1:ns,1:2) + rhoDot(1:ns,9:10) * timestep < -aTolRho(instance))) 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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
forall (s = 1:ns, t = 1_pInt:4_pInt)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
constitutive_nonlocal_dotState(iRhoU(s,t,instance)) = rhoDot(s,t)
constitutive_nonlocal_dotState(iRhoB(s,t,instance)) = rhoDot(s,t+4_pInt)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
endforall
forall (s = 1:ns, c = 1_pInt:2_pInt) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
constitutive_nonlocal_dotState(iRhoD(s,c,instance)) = rhoDot(s,c+8_pInt)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
forall (s = 1:ns) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
constitutive_nonlocal_dotState(iGamma(s,instance)) = sum(gdot(s,1:4))
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
end function constitutive_nonlocal_dotState
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)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
use math, only: math_mul3x3, &
moved quaternion disorientation to lattice because it requires the knowledge of the lattice structure.
2014-02-28 18:58:27 +05:30
math_qRot
* 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 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, &
moved quaternion disorientation to lattice because it requires the knowledge of the lattice structure.
2014-02-28 18:58:27 +05:30
lattice_sd, &
lattice_qDisorientation
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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
neighbor_e, & ! element index of my neighbor
neighbor_i, & ! integration point index of my neighbor
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
phase, &
neighbor_phase, &
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
textureID, &
neighbor_textureID, &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
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))))) :: &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
my_compatibility ! my_compatibility for current element and ip
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(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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
real(pReal) my_compatibilitySum, &
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))))
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
phase = material_phase(1,i,e)
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
textureID = material_texture(1,i,e)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
instance = phase_plasticityInstance(phase)
ns = totalNslip(instance)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
slipNormal(1:3,1:ns) = lattice_sn(1:3, slipSystemLattice(1:ns,instance), phase)
slipDirection(1:3,1:ns) = lattice_sd(1:3, slipSystemLattice(1:ns,instance), phase)
restructured constitutive_nonlocal_compatibility and incorporated the "surfaceTransmissivity" in the compatibility calculation
2011-02-16 22:08:18 +05:30
!*** start out fully compatible
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
my_compatibility = 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) &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
my_compatibility(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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
!*** Loop thrugh neighbors and check whether there is any my_compatibility.
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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
neighbor_e = mesh_ipNeighborhood(1,n,i,e)
neighbor_i = mesh_ipNeighborhood(2,n,i,e)
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
!* FREE SURFACE
!* Set surface transmissivity to the value specified in the material.config
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
if (neighbor_e <= 0_pInt .or. neighbor_i <= 0_pInt) 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) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
my_compatibility(1:2,s1,s1,n) = sqrt(surfaceTransmissivity(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.
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
neighbor_phase = material_phase(1,neighbor_i,neighbor_e)
if (neighbor_phase /= phase) then
if (.not. phase_localPlasticity(neighbor_phase) .and. .not. phase_localPlasticity(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) &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
my_compatibility(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)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
if (grainboundaryTransmissivity(instance) >= 0.0_pReal) then
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
neighbor_textureID = material_texture(1,neighbor_i,neighbor_e)
if (neighbor_textureID /= textureID) then
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
if (.not. phase_localPlasticity(neighbor_phase)) 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
forall(s1 = 1_pInt:ns) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
my_compatibility(1:2,s1,s1,n) = sqrt(grainboundaryTransmissivity(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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +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
!* 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:
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
!* The my_compatibility 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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
!* the number of compatible slip systems is minimized with the sum of the original my_compatibility values exceeding one.
!* Finally the smallest my_compatibility 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
moved quaternion disorientation to lattice because it requires the knowledge of the lattice structure.
2014-02-28 18:58:27 +05:30
absoluteMisorientation = lattice_qDisorientation(orientation(1:4,1,i,e), &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
orientation(1:4,1,neighbor_i,neighbor_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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
my_compatibility(1,s2,s1,n) = math_mul3x3(slipNormal(1:3,s1), math_qRot(absoluteMisorientation, slipNormal(1:3,s2))) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
* abs(math_mul3x3(slipDirection(1:3,s1), math_qRot(absoluteMisorientation, slipDirection(1:3,s2))))
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
my_compatibility(2,s2,s1,n) = abs(math_mul3x3(slipNormal(1:3,s1), math_qRot(absoluteMisorientation, slipNormal(1:3,s2)))) &
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
* 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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
my_compatibilitySum = 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.
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
do while (my_compatibilitySum < 1.0_pReal .and. any(belowThreshold(1:ns)))
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
thresholdValue = maxval(my_compatibility(2,1:ns,s1,n), belowThreshold(1:ns)) ! screws always positive
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
nThresholdValues = real(count(my_compatibility(2,1:ns,s1,n) == thresholdValue),pReal)
where (my_compatibility(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.
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
if (my_compatibilitySum + thresholdValue * nThresholdValues > 1.0_pReal) &
where (abs(my_compatibility(1:2,1:ns,s1,n)) == thresholdValue) &
my_compatibility(1:2,1:ns,s1,n) = sign((1.0_pReal - my_compatibilitySum) &
/ nThresholdValues, my_compatibility(1:2,1:ns,s1,n))
my_compatibilitySum = my_compatibilitySum + 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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
where (belowThreshold(1:ns)) my_compatibility(1,1:ns,s1,n) = 0.0_pReal
where (belowThreshold(1:ns)) my_compatibility(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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
compatibility(1:2,1:ns,1:ns,1:Nneighbors,i,e) = my_compatibility
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
end subroutine constitutive_nonlocal_updateCompatibility
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
!*********************************************************************
!* calculates quantities characterizing the microstructure *
!*********************************************************************
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
pure function constitutive_nonlocal_dislocationstress(state, Fe, ipc, 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
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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
use lattice, only: lattice_mu, &
lattice_nu
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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
integer(pInt), intent(in) :: ipc, & ! current grain ID
ip, & ! current integration point
el ! current element
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) :: &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
Fe ! elastic deformation gradient
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
type(p_vec), dimension(homogenization_maxNgrains,mesh_maxNips,mesh_NcpElems), intent(in) :: &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
state ! microstructural state
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
!*** input/output variables
!*** output variables
real(pReal), dimension(3,3) :: constitutive_nonlocal_dislocationstress
!*** local variables
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
integer(pInt) neighbor_el, & ! element number of neighbor material point
neighbor_ip, & ! integration point of neighbor material point
instance, & ! my instance of this plasticity
neighbor_instance, & ! instance of this plasticity of neighbor 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
phase, &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
neighbor_phase, &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
ns, & ! total number of active slip systems at my material point
neighbor_ns, & ! total number of active slip systems at neighbor 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-)
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
dir, &
deltaX, deltaY, deltaZ, &
side, &
j
integer(pInt), dimension(2,3) :: periodicImages
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
real(pReal) x, y, z, & ! coordinates of connection vector in neighbor lattice frame
xsquare, ysquare, zsquare, & ! squares of respective coordinates
distance, & ! length of connection vector
segmentLength, & ! segment length of dislocations
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, &
R, Rsquare, Rcube, &
denominator, &
flipSign, &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
neighbor_ipVolumeSideLength, &
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
detFe
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
real(pReal), dimension(3) :: connection, & ! connection vector between me and my neighbor in the deformed configuration
connection_neighborLattice, & ! connection vector between me and my neighbor in the lattice configuration of my neighbor
connection_neighborSlip, & ! connection vector between me and my neighbor in the slip system frame of my neighbor
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, minCoord, &
meshSize, &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
coords, & ! x,y,z coordinates of cell center of ip volume
neighbor_coords ! x,y,z coordinates of cell center of neighbor ip volume
real(pReal), dimension(3,3) :: sigma, & ! dislocation stress for one slip system in neighbor material point's slip system frame
Tdislo_neighborLattice, & ! dislocation stress as 2nd Piola-Kirchhoff stress at neighbor material point
invFe, & ! inverse of my elastic deformation gradient
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
neighbor_invFe, &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
neighborLattice2myLattice ! mapping from neighbor 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)) :: &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
neighbor_rhoExcess ! excess density at neighbor 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 temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,ip,el))),8) :: &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
rhoSgl ! single dislocation density (edge+, edge-, screw+, screw-, used edge+, used edge-, used screw+, used screw-)
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
logical inversionError
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
phase = material_phase(ipc,ip,el)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
instance = phase_plasticityInstance(phase)
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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
forall (s = 1_pInt:ns, t = 1_pInt:4_pInt)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
rhoSgl(s,t) = max(state(ipc,ip,el)%p(iRhoU(s,t,instance)), 0.0_pReal) ! ensure positive single mobile densities
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoSgl(s,t+4_pInt) = state(ipc,ip,el)%p(iRhoB(s,t,instance))
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
endforall
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
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
call math_invert33(Fe(1:3,1:3,ipc,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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
do neighbor_el = 1_pInt,mesh_NcpElems
ipLoop: do neighbor_ip = 1_pInt,FE_Nips(FE_geomtype(mesh_element(2,neighbor_el)))
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
neighbor_phase = material_phase(ipc,neighbor_ip,neighbor_el)
if (phase_localPlasticity(neighbor_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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
neighbor_instance = phase_plasticityInstance(neighbor_phase)
neighbor_ns = totalNslip(neighbor_instance)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
call math_invert33(Fe(1:3,1:3,1,neighbor_ip,neighbor_el), neighbor_invFe, detFe, inversionError)
neighbor_ipVolumeSideLength = mesh_ipVolume(neighbor_ip,neighbor_el) ** (1.0_pReal/3.0_pReal) ! reference volume used here
forall (s = 1_pInt:neighbor_ns, c = 1_pInt:2_pInt)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
neighbor_rhoExcess(c,1,s) = state(ipc,neighbor_ip,neighbor_el)%p(iRhoU(s,2*c-1,neighbor_instance)) & ! positive mobiles
- state(ipc,neighbor_ip,neighbor_el)%p(iRhoU(s,2*c,neighbor_instance)) ! negative mobiles
neighbor_rhoExcess(c,2,s) = abs(state(ipc,neighbor_ip,neighbor_el)%p(iRhoB(s,2*c-1,neighbor_instance))) & ! positive deads
- abs(state(ipc,neighbor_ip,neighbor_el)%p(iRhoB(s,2*c,neighbor_instance))) ! negative deads
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
endforall
Tdislo_neighborLattice = 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
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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
if (neighbor_el /= el .or. neighbor_ip /= ip &
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
.or. deltaX /= 0_pInt .or. deltaY /= 0_pInt .or. deltaZ /= 0_pInt) then
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
neighbor_coords = mesh_cellCenterCoordinates(neighbor_ip,neighbor_el) &
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
+ [real(deltaX,pReal), real(deltaY,pReal), real(deltaZ,pReal)] * meshSize
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
connection = neighbor_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))
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
!* this ensures, that the central MP never sits on a neighbor dislocation segment
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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
connection_neighborLattice = math_mul33x3(neighbor_invFe, connection)
segmentLength = min(neighbor_ipVolumeSideLength, distance)
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 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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
!* loop through all slip systems of the neighbor 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
!* and add up the stress contributions from egde and screw excess on these slip systems (if significant)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
do s = 1_pInt,neighbor_ns
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
if (all(abs(neighbor_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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
connection_neighborSlip = math_mul33x3(lattice2slip(1:3,1:3,s,neighbor_instance), &
renaming of "instance" and thelike to be consistent with naming scheme introduced in rev2662
2013-11-21 19:05:43 +05:30
connection_neighborLattice)
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
!* edge contribution to stress
sigma = 0.0_pReal
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
x = connection_neighborSlip(1)
y = connection_neighborSlip(2)
z = connection_neighborSlip(3)
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 = 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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
if (abs(neighbor_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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
x = connection_neighborSlip(1) + sign(0.5_pReal * segmentLength, &
state(ipc,neighbor_ip,neighbor_el)%p(iRhoB(s,1,neighbor_instance)) &
- state(ipc,neighbor_ip,neighbor_el)%p(iRhoB(s,2,neighbor_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
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) &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
* neighbor_rhoExcess(1,j,s)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sigma(2,2) = sigma(2,2) - real(side,pReal) &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
* (flipSign * 2.0_pReal * lattice_nu(phase) * z / denominator + z * lambda / Rcube) &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
* neighbor_rhoExcess(1,j,s)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sigma(3,3) = sigma(3,3) + real(side,pReal) &
* flipSign * z / denominator &
* (1.0_pReal - zsquare / Rsquare - zsquare / denominator) &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
* neighbor_rhoExcess(1,j,s)
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) &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
* x * z / Rcube * neighbor_rhoExcess(1,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 * x / denominator &
* (1.0_pReal - zsquare / Rsquare - zsquare / denominator) &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
* neighbor_rhoExcess(1,j,s)
removed prefix "constitutive_nonlocal_" from all private variables
2013-05-24 01:26:36 +05:30
sigma(2,3) = sigma(2,3) - real(side,pReal) &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
* (lattice_nu(phase) / R - zsquare / Rcube) * neighbor_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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
x = connection_neighborSlip(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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
if (abs(neighbor_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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
y = connection_neighborSlip(2) + sign(0.5_pReal * segmentLength, &
state(ipc,neighbor_ip,neighbor_el)%p(iRhoB(s,3,neighbor_instance)) &
- state(ipc,neighbor_ip,neighbor_el)%p(iRhoB(s,4,neighbor_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
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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
sigma(1,2) = sigma(1,2) - real(side,pReal) * flipSign * z * (1.0_pReal - lattice_nu(phase)) / denominator &
* neighbor_rhoExcess(2,j,s)
sigma(1,3) = sigma(1,3) + real(side,pReal) * flipSign * y * (1.0_pReal - lattice_nu(phase)) / denominator &
* neighbor_rhoExcess(2,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
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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
!* scale stresses and map them into the neighbor material point's lattice configuration
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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
sigma = sigma * lattice_mu(neighbor_phase) * burgers(s,neighbor_instance) &
/ (4.0_pReal * pi * (1.0_pReal - lattice_nu(neighbor_phase))) &
* mesh_ipVolume(neighbor_ip,neighbor_el) / segmentLength ! reference volume is used here (according to the segment length calculation)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
Tdislo_neighborLattice = Tdislo_neighborLattice &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
+ math_mul33x33(math_transpose33(lattice2slip(1:3,1:3,s,neighbor_instance)), &
math_mul33x33(sigma, lattice2slip(1:3,1:3,s,neighbor_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) &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
rhoExcessDead(c,s) = state(ipc,ip,el)%p(iRhoB(s,2*c-1,instance)) & ! 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(ipc,ip,el)%p(iRhoB(s,2*c,instance)) ! 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)
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
sigma(1,3) = - (rhoExcessDead(1,s) + rhoExcessDead(2,s) * (1.0_pReal - lattice_nu(phase))) &
* neighbor_ipVolumeSideLength * lattice_mu(phase) * burgers(s,instance) &
/ (sqrt(2.0_pReal) * pi * (1.0_pReal - lattice_nu(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
sigma(3,1) = sigma(1,3)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
Tdislo_neighborLattice = Tdislo_neighborLattice &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
!* map the stress from the neighbor MP's lattice configuration into the deformed configuration
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 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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
neighborLattice2myLattice = math_mul33x33(invFe, Fe(1:3,1:3,1,neighbor_ip,neighbor_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
constitutive_nonlocal_dislocationstress = constitutive_nonlocal_dislocationstress &
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
+ math_mul33x33(neighborLattice2myLattice, &
math_mul33x33(Tdislo_neighborLattice, &
math_transpose33(neighborLattice2myLattice)))
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
remove dotTemperature (returned always 0 anyway) to make it easier to replace it with a heating rate funtction remove potentially dangerous short circuiting in homogenization files added doxygen comment for nonlocal (giving the functions a name) fixed wrong definition of maxnchungs introcuded into dislotwin with last commit, causes trouble with hex
2013-10-09 11:42:16 +05:30
end function 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
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief return array of constitutive results
!--------------------------------------------------------------------------------------------------
pure function constitutive_nonlocal_postResults(Tstar_v,Fe,state,dotState,ipc,ip,el)
use math, only: &
math_mul6x6, &
math_mul33x3, &
math_mul33x33, &
pi
use mesh, only: &
mesh_NcpElems, &
mesh_maxNips
use material, only: &
homogenization_maxNgrains, &
material_phase, &
phase_plasticityInstance, &
phase_Noutput
use lattice, only: &
lattice_Sslip_v, &
lattice_sd, &
lattice_st, &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
lattice_sn, &
lattice_mu, &
lattice_nu
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
implicit none
real(pReal), dimension(6), intent(in) :: &
Tstar_v !< 2nd Piola Kirchhoff stress tensor in Mandel notation
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 !< microstructure state
type(p_vec), intent(in) :: dotState ! evolution rate of microstructural state
integer(pInt), intent(in) :: &
ipc, & !< component-ID of integration point
ip, & !< integration point
el !< element
real(pReal), dimension(constitutive_nonlocal_sizePostResults(&
phase_plasticityInstance(material_phase(ipc,ip,el)))) :: &
constitutive_nonlocal_postResults
integer(pInt) :: &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
phase, &
renamed "maxNmatIDs" accordingly to "maxNinstances"
2014-03-04 19:17:04 +05:30
instance, & !< current instance of this plasticity
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
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
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,ip,el))),8) :: &
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)
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,ip,el))),4) :: &
gdot, & !< shear rates
v !< velocities
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,ip,el)))) :: &
rhoForest, & !< forest dislocation density
tauThreshold, & !< threshold shear stress
tau, & !< current resolved shear stress
tauBack !< back stress from pileups on same slip system
real(pReal), dimension(totalNslip(phase_plasticityInstance(material_phase(ipc,ip,el))),2) :: &
rhoDip, & !< current dipole dislocation densities (screw and edge dipoles)
rhoDotDip, & !< evolution rate of 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
real(pReal), dimension(3,totalNslip(phase_plasticityInstance(material_phase(ipc,ip,el))),2) :: &
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
real(pReal), dimension(3,totalNslip(phase_plasticityInstance(material_phase(ipc,ip,el)))) :: &
n_currentconf !< slip system normal (unit vector) in current configuration
real(pReal), dimension(3,3) :: &
sigma
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
phase = material_phase(ipc,ip,el)
instance = phase_plasticityInstance(phase)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
ns = totalNslip(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
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
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
forall (s = 1_pInt:ns, t = 1_pInt:4_pInt)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoSgl(s,t) = state(ipc,ip,el)%p(iRhoU(s,t,instance))
rhoSgl(s,t+4_pInt) = state(ipc,ip,el)%p(iRhoB(s,t,instance))
v(s,t) = state(ipc,ip,el)%p(iV(s,t,instance))
rhoDotSgl(s,t) = dotState%p(iRhoU(s,t,instance))
rhoDotSgl(s,t+4_pInt) = dotState%p(iRhoB(s,t,instance))
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
endforall
forall (s = 1_pInt:ns, c = 1_pInt:2_pInt)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoDip(s,c) = state(ipc,ip,el)%p(iRhoD(s,c,instance))
rhoDotDip(s,c) = dotState%p(iRhoD(s,c,instance))
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
endforall
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
rhoForest = state(ipc,ip,el)%p(iRhoF(1:ns,instance))
tauThreshold = state(ipc,ip,el)%p(iTauF(1:ns,instance))
tauBack = state(ipc,ip,el)%p(iTauB(1:ns,instance))
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) &
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
gdot(1:ns,t) = rhoSgl(1:ns,t) * burgers(1:ns,instance) * 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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
sLattice = slipSystemLattice(s,instance)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
tau(s) = math_mul6x6(Tstar_v, lattice_Sslip_v(1:6,1,sLattice,phase)) + 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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
dLower = minDipoleHeight(1:ns,1:2,instance)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
dUpper(1:ns,1) = lattice_mu(phase) * burgers(1:ns,instance) &
/ (8.0_pReal * pi * (1.0_pReal - lattice_nu(phase)) * abs(tau))
dUpper(1:ns,2) = lattice_mu(phase) * burgers(1:ns,instance) &
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
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
m(1:3,1:ns,1) = lattice_sd(1:3,slipSystemLattice(1:ns,instance),phase)
m(1:3,1:ns,2) = -lattice_st(1:3,slipSystemLattice(1:ns,instance),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
forall (c = 1_pInt:2_pInt, s = 1_pInt:ns) &
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
m_currentconf(1:3,s,c) = math_mul33x3(Fe(1:3,1:3,ipc,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) &
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
n_currentconf(1:3,s) = math_mul33x3(Fe(1:3,1:3,ipc,ip,el), &
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
lattice_sn(1:3,slipSystemLattice(s,instance),phase))
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
outputsLoop: do o = 1_pInt,phase_Noutput(material_phase(ipc,ip,el))
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
select case(constitutive_nonlocal_outputID(o,instance))
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_mobile_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_immobile_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dip_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_edge_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_edge_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_edge_mobile_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_edge_immobile_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_edge_pos_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_edge_pos_mobile_ID)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_edge_pos_immobile_ID)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = 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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_edge_neg_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_edge_neg_mobile_ID)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_edge_neg_immobile_ID)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = 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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dip_edge_ID)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDip(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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_screw_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_screw_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_screw_mobile_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_screw_immobile_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_screw_pos_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_screw_pos_mobile_ID)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(1:ns,3)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_screw_pos_immobile_ID)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(1:ns,7)
- 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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_screw_neg_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_screw_neg_mobile_ID)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoSgl(1:ns,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
cs = cs + ns
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_sgl_screw_neg_immobile_ID)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = 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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dip_screw_ID)
replaced all remaining occurrences of state indices
2013-05-24 17:18:34 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDip(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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (excess_rho_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (excess_rho_edge_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (excess_rho_screw_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_forest_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (delta_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (delta_sgl_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (delta_dip_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (shearrate_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (resolvedstress_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (resolvedstress_back_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (resolvedstress_external_ID)
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
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
sLattice = slipSystemLattice(s,instance)
moved reading in of lattice type and elastic constants to lattice module removed structure type for hex, fcc, bcc, now defining slip/twin systems for each phase found in material.config constitutive modules will only be initialized if needed homogenizedC function is only needed for models incorporating twinning in a physical way (titanmod and dislotwin)
2014-03-09 02:20:31 +05:30
constitutive_nonlocal_postResults(cs+s) = math_mul6x6(Tstar_v, lattice_Sslip_v(1:6,1,sLattice,phase))
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (resistance_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_ID)
added two new outputs: "rho_dot_flux_mobile" and "rho_dot_sgl_mobile" corrected outputs of dislocation density rates: take into account that blocked dislocation density is a signed quantity
2014-01-07 21:25:28 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoDotSgl(1:ns,1:4),2) &
+ sum(rhoDotSgl(1:ns,5:8)*sign(1.0_pReal,rhoSgl(1:ns,5:8)),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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_sgl_ID)
added two new outputs: "rho_dot_flux_mobile" and "rho_dot_sgl_mobile" corrected outputs of dislocation density rates: take into account that blocked dislocation density is a signed quantity
2014-01-07 21:25:28 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoDotSgl(1:ns,1:4),2) &
+ sum(rhoDotSgl(1:ns,5:8)*sign(1.0_pReal,rhoSgl(1:ns,5:8)),2)
cs = cs + ns
case (rho_dot_sgl_mobile_ID)
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoDotSgl(1:ns,1:4),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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_dip_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_gen_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotMultiplicationOutput(1:ns,1,ipc,ip,el) &
+ rhoDotMultiplicationOutput(1:ns,2,ipc,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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_gen_edge_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotMultiplicationOutput(1:ns,1,ipc,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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_gen_screw_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotMultiplicationOutput(1:ns,2,ipc,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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_sgl2dip_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotSingle2DipoleGlideOutput(1:ns,1,ipc,ip,el) &
+ rhoDotSingle2DipoleGlideOutput(1:ns,2,ipc,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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_sgl2dip_edge_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotSingle2DipoleGlideOutput(1:ns,1,ipc,ip,el)
added separate output of dipole formation rate for edge and screw
2012-08-27 21:27:31 +05:30
cs = cs + ns
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_sgl2dip_screw_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotSingle2DipoleGlideOutput(1:ns,2,ipc,ip,el)
added separate output of dipole formation rate for edge and screw
2012-08-27 21:27:31 +05:30
cs = cs + ns
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_ann_ath_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotAthermalAnnihilationOutput(1:ns,1,ipc,ip,el) &
+ rhoDotAthermalAnnihilationOutput(1:ns,2,ipc,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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_ann_the_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotThermalAnnihilationOutput(1:ns,1,ipc,ip,el) &
+ rhoDotThermalAnnihilationOutput(1:ns,2,ipc,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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_ann_the_edge_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotThermalAnnihilationOutput(1:ns,1,ipc,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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_ann_the_screw_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotThermalAnnihilationOutput(1:ns,2,ipc,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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_edgejogs_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = rhoDotEdgeJogsOutput(1:ns,ipc,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 two new outputs: "rho_dot_flux_mobile" and "rho_dot_sgl_mobile" corrected outputs of dislocation density rates: take into account that blocked dislocation density is a signed quantity
2014-01-07 21:25:28 +05:30
case (rho_dot_flux_mobile_ID)
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoDotFluxOutput(1:ns,1:4,ipc,ip,el),2)
cs = cs + ns
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_flux_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoDotFluxOutput(1:ns,1:4,ipc,ip,el),2) &
added two new outputs: "rho_dot_flux_mobile" and "rho_dot_sgl_mobile" corrected outputs of dislocation density rates: take into account that blocked dislocation density is a signed quantity
2014-01-07 21:25:28 +05:30
+ sum(rhoDotFluxOutput(1:ns,5:8,ipc,ip,el)*sign(1.0_pReal,rhoSgl(1:ns,5:8)),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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_flux_edge_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoDotFluxOutput(1:ns,1:2,ipc,ip,el),2) &
added two new outputs: "rho_dot_flux_mobile" and "rho_dot_sgl_mobile" corrected outputs of dislocation density rates: take into account that blocked dislocation density is a signed quantity
2014-01-07 21:25:28 +05:30
+ sum(rhoDotFluxOutput(1:ns,5:6,ipc,ip,el)*sign(1.0_pReal,rhoSgl(1:ns,5:6)),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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (rho_dot_flux_screw_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = sum(rhoDotFluxOutput(1:ns,3:4,ipc,ip,el),2) &
added two new outputs: "rho_dot_flux_mobile" and "rho_dot_sgl_mobile" corrected outputs of dislocation density rates: take into account that blocked dislocation density is a signed quantity
2014-01-07 21:25:28 +05:30
+ sum(rhoDotFluxOutput(1:ns,7:8,ipc,ip,el)*sign(1.0_pReal,rhoSgl(1:ns,7:8)),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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (velocity_edge_pos_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (velocity_edge_neg_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (velocity_screw_pos_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (velocity_screw_neg_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (slipdirectionx_ID)
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = m_currentconf(1,1:ns,1)
cs = cs + ns
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (slipdirectiony_ID)
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = m_currentconf(2,1:ns,1)
cs = cs + ns
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (slipdirectionz_ID)
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = m_currentconf(3,1:ns,1)
cs = cs + ns
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (slipnormalx_ID)
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = n_currentconf(1,1:ns)
cs = cs + ns
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (slipnormaly_ID)
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = n_currentconf(2,1:ns)
cs = cs + ns
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (slipnormalz_ID)
added output of slip direction and slip normal changed name of fluxdensity outputs
2013-05-23 13:49:36 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = n_currentconf(3,1:ns)
cs = cs + ns
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (fluxdensity_edge_posx_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (fluxdensity_edge_posy_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (fluxdensity_edge_posz_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (fluxdensity_edge_negx_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (fluxdensity_edge_negy_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (fluxdensity_edge_negz_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (fluxdensity_screw_posx_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (fluxdensity_screw_posy_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (fluxdensity_screw_posz_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (fluxdensity_screw_negx_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (fluxdensity_screw_negy_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (fluxdensity_screw_negz_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (maximumdipoleheight_edge_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case (maximumdipoleheight_screw_ID)
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case(dislocationstress_ID)
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
sigma = constitutive_nonlocal_dislocationstress(state, Fe, ipc, 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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
case(accumulatedshear_ID)
also rename matID/i to instance like in the other constitutive models
2014-02-28 18:33:21 +05:30
constitutive_nonlocal_postResults(cs+1_pInt:cs+ns) = state(ipc,ip,el)%p(iGamma(1:ns,instance))
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
enumerators for case switch of outputs as already used in phenopowerlaw removed obsolete output 'boundarylayer'
2013-11-29 17:59:47 +05:30
end select
enddo outputsLoop
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 temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
end function constitutive_nonlocal_postResults
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
removed temperature integration and corresponding data structures and debugging options temperature is stored in crystallite, but homogeneous on one IP (not an component (grain) quantity and an input value parsed in by the BVP solver. introduced heat, a component (grain) quantity which is homogenized before returned to the heat transfer solver. went ahead with removal of dummy functions in homogenization and constitutive, this time mainly reduced function signatures to reflect actually needed quantities.
2013-10-16 18:34:59 +05:30
end module constitutive_nonlocal