DAMASK_EICMD/code/homogenization.f90

880 lines
46 KiB
Fortran

!--------------------------------------------------------------------------------------------------
! $Id$
!--------------------------------------------------------------------------------------------------
!> @author Franz Roters, Max-Planck-Institut für Eisenforschung GmbH
!> @author Philip Eisenlohr, Max-Planck-Institut für Eisenforschung GmbH
!> @author Denny Tjahjanto, Max-Planck-Institut für Eisenforschung GmbH
!> @brief homogenization manager, organizing deformation partitioning and stress homogenization
!--------------------------------------------------------------------------------------------------
module homogenization
use prec, only: &
pInt, &
pReal, &
p_vec
!--------------------------------------------------------------------------------------------------
! General variables for the homogenization at a material point
implicit none
private
type(p_vec), dimension(:,:), allocatable, public :: &
homogenization_state0 !< pointer array to homogenization state at start of FE increment
real(pReal), dimension(:,:,:,:), allocatable, public :: &
materialpoint_F0, & !< def grad of IP at start of FE increment
materialpoint_F, & !< def grad of IP to be reached at end of FE increment
materialpoint_P !< first P--K stress of IP
real(pReal), dimension(:,:,:,:,:,:), allocatable, public :: &
materialpoint_dPdF !< tangent of first P--K stress at IP
real(pReal), dimension(:,:,:), allocatable, public :: &
materialpoint_results !< results array of material point
type(p_vec), dimension(:,:), allocatable, public, protected :: &
homogenization_state !< pointer array to current homogenization state (end of converged time step)
integer(pInt), dimension(:,:), allocatable, public, protected :: &
homogenization_sizeState !< size of state array per grain
integer(pInt), public, protected :: &
materialpoint_sizeResults, &
homogenization_maxSizePostResults
real(pReal), dimension(:,:), allocatable, public, protected :: &
materialpoint_heat
type(p_vec), dimension(:,:), allocatable, private :: &
homogenization_subState0 !< pointer array to homogenization state at start of homogenization increment
real(pReal), dimension(:,:,:,:), allocatable, private :: &
materialpoint_subF0, & !< def grad of IP at beginning of homogenization increment
materialpoint_subF !< def grad of IP to be reached at end of homog inc
real(pReal), dimension(:,:), allocatable, private :: &
materialpoint_subFrac, &
materialpoint_subStep, &
materialpoint_subdt
integer(pInt), dimension(:,:), allocatable, private :: &
homogenization_sizePostResults !< size of postResults array per material point
integer(pInt), private :: &
homogenization_maxSizeState
logical, dimension(:,:), allocatable, private :: &
materialpoint_requested, &
materialpoint_converged
logical, dimension(:,:,:), allocatable, private :: &
materialpoint_doneAndHappy
public :: &
homogenization_init, &
materialpoint_stressAndItsTangent, &
materialpoint_postResults
private :: &
homogenization_partitionDeformation, &
homogenization_updateState, &
homogenization_averageStressAndItsTangent, &
homogenization_averageHeat, &
homogenization_postResults
contains
!--------------------------------------------------------------------------------------------------
!> @brief module initialization
!--------------------------------------------------------------------------------------------------
subroutine homogenization_init()
#ifdef HDF
use hdf5, only: &
HID_T
use IO, only : &
HDF5_mappingHomogenization
#endif
use, intrinsic :: iso_fortran_env ! to get compiler_version and compiler_options (at least for gfortran 4.6 at the moment)
use math, only: &
math_I3
use debug, only: &
debug_level, &
debug_homogenization, &
debug_levelBasic, &
debug_e, &
debug_g
use IO, only: &
IO_error, &
IO_open_file, &
IO_open_jobFile_stat, &
IO_write_jobFile, &
IO_write_jobIntFile, &
IO_timeStamp
use mesh, only: &
mesh_maxNips, &
mesh_NcpElems, &
mesh_element, &
FE_Nips, &
FE_geomtype
use constitutive, only: &
constitutive_maxSizePostResults
use constitutive_damage, only: &
constitutive_damage_maxSizePostResults
use constitutive_thermal, only: &
constitutive_thermal_maxSizePostResults
use crystallite, only: &
crystallite_maxSizePostResults
use material
use homogenization_none
use homogenization_isostrain
use homogenization_RGC
implicit none
integer(pInt), parameter :: FILEUNIT = 200_pInt
integer(pInt) :: e,i,p,myInstance
integer(pInt), dimension(:,:), pointer :: thisSize
character(len=64), dimension(:,:), pointer :: thisOutput
character(len=32) :: outputName !< name of output, intermediate fix until HDF5 output is ready
logical :: knownHomogenization
#ifdef HDF
integer(pInt), dimension(:,:), allocatable :: mapping
integer(pInt), dimension(:), allocatable :: InstancePosition
allocate(mapping(mesh_ncpelems,4),source=0_pInt)
allocate(InstancePosition(material_Nhomogenization),source=0_pInt)
#endif
!--------------------------------------------------------------------------------------------------
! parse homogenization from config file
if (.not. IO_open_jobFile_stat(FILEUNIT,material_localFileExt)) & ! no local material configuration present...
call IO_open_file(FILEUNIT,material_configFile) ! ... open material.config file
if (any(homogenization_type == HOMOGENIZATION_NONE_ID)) &
call homogenization_none_init()
if (any(homogenization_type == HOMOGENIZATION_ISOSTRAIN_ID)) &
call homogenization_isostrain_init(FILEUNIT)
if (any(homogenization_type == HOMOGENIZATION_RGC_ID)) &
call homogenization_RGC_init(FILEUNIT)
close(FILEUNIT)
!--------------------------------------------------------------------------------------------------
! write description file for homogenization output
call IO_write_jobFile(FILEUNIT,'outputHomogenization')
do p = 1,material_Nhomogenization
i = homogenization_typeInstance(p) ! which instance of this homogenization type
knownHomogenization = .true. ! assume valid
select case(homogenization_type(p)) ! split per homogenization type
case (HOMOGENIZATION_NONE_ID)
outputName = HOMOGENIZATION_NONE_label
thisOutput => null()
thisSize => null()
case (HOMOGENIZATION_ISOSTRAIN_ID)
outputName = HOMOGENIZATION_ISOSTRAIN_label
thisOutput => homogenization_isostrain_output
thisSize => homogenization_isostrain_sizePostResult
case (HOMOGENIZATION_RGC_ID)
outputName = HOMOGENIZATION_RGC_label
thisOutput => homogenization_RGC_output
thisSize => homogenization_RGC_sizePostResult
case default
knownHomogenization = .false.
end select
write(FILEUNIT,'(/,a,/)') '['//trim(homogenization_name(p))//']'
if (knownHomogenization) then
write(FILEUNIT,'(a)') '(type)'//char(9)//trim(outputName)
write(FILEUNIT,'(a,i4)') '(ngrains)'//char(9),homogenization_Ngrains(p)
do e = 1,homogenization_Noutput(p)
write(FILEUNIT,'(a,i4)') trim(thisOutput(e,i))//char(9),thisSize(e,i)
enddo
endif
enddo
close(FILEUNIT)
!--------------------------------------------------------------------------------------------------
! allocate and initialize global variables
allocate(homogenization_state0(mesh_maxNips,mesh_NcpElems))
allocate(homogenization_subState0(mesh_maxNips,mesh_NcpElems))
allocate(homogenization_state(mesh_maxNips,mesh_NcpElems))
allocate(homogenization_sizeState(mesh_maxNips,mesh_NcpElems), source=0_pInt)
allocate(homogenization_sizePostResults(mesh_maxNips,mesh_NcpElems), source=0_pInt)
allocate(materialpoint_heat(mesh_maxNips,mesh_NcpElems), source=0.0_pReal)
allocate(materialpoint_dPdF(3,3,3,3,mesh_maxNips,mesh_NcpElems), source=0.0_pReal)
allocate(materialpoint_F0(3,3,mesh_maxNips,mesh_NcpElems), source=0.0_pReal)
materialpoint_F0 = spread(spread(math_I3,3,mesh_maxNips),4,mesh_NcpElems) ! initialize to identity
allocate(materialpoint_F(3,3,mesh_maxNips,mesh_NcpElems), source=0.0_pReal)
materialpoint_F = materialpoint_F0 ! initialize to identity
allocate(materialpoint_subF0(3,3,mesh_maxNips,mesh_NcpElems), source=0.0_pReal)
allocate(materialpoint_subF(3,3,mesh_maxNips,mesh_NcpElems), source=0.0_pReal)
allocate(materialpoint_P(3,3,mesh_maxNips,mesh_NcpElems), source=0.0_pReal)
allocate(materialpoint_subFrac(mesh_maxNips,mesh_NcpElems), source=0.0_pReal)
allocate(materialpoint_subStep(mesh_maxNips,mesh_NcpElems), source=0.0_pReal)
allocate(materialpoint_subdt(mesh_maxNips,mesh_NcpElems), source=0.0_pReal)
allocate(materialpoint_requested(mesh_maxNips,mesh_NcpElems), source=.false.)
allocate(materialpoint_converged(mesh_maxNips,mesh_NcpElems), source=.true.)
allocate(materialpoint_doneAndHappy(2,mesh_maxNips,mesh_NcpElems), source=.true.)
!--------------------------------------------------------------------------------------------------
! allocate and initialize global state and postresutls variables
elementLooping: do e = 1,mesh_NcpElems
myInstance = homogenization_typeInstance(mesh_element(3,e))
IpLooping: do i = 1,FE_Nips(FE_geomtype(mesh_element(2,e)))
#ifdef HDF
InstancePosition(myInstance) = InstancePosition(myInstance)+1_pInt
mapping(e,1:4) = [instancePosition(myinstance),myinstance,e,i]
#endif
select case(homogenization_type(mesh_element(3,e)))
case (HOMOGENIZATION_none_ID)
homogenization_sizePostResults(i,e) = 0_pInt
case (HOMOGENIZATION_ISOSTRAIN_ID)
homogenization_sizePostResults(i,e) = homogenization_isostrain_sizePostResults(myInstance)
case (HOMOGENIZATION_RGC_ID)
if (homogenization_RGC_sizeState(myInstance) > 0_pInt) then
allocate(homogenization_state0(i,e)%p(homogenization_RGC_sizeState(myInstance)))
allocate(homogenization_subState0(i,e)%p(homogenization_RGC_sizeState(myInstance)))
allocate(homogenization_state(i,e)%p(homogenization_RGC_sizeState(myInstance)))
homogenization_state0(i,e)%p = 0.0_pReal
homogenization_sizeState(i,e) = homogenization_RGC_sizeState(myInstance)
endif
homogenization_sizePostResults(i,e) = homogenization_RGC_sizePostResults(myInstance)
end select
enddo IpLooping
enddo elementLooping
#ifdef HDF
call HDF5_mappingHomogenization(mapping)
#endif
!--------------------------------------------------------------------------------------------------
! write state size file out
call IO_write_jobIntFile(777,'sizeStateHomog',size(homogenization_sizeState))
write (777,rec=1) homogenization_sizeState
close(777)
homogenization_maxSizeState = maxval(homogenization_sizeState)
homogenization_maxSizePostResults = maxval(homogenization_sizePostResults)
materialpoint_sizeResults = 1 & ! grain count
+ 1 + homogenization_maxSizePostResults & ! homogSize & homogResult
+ homogenization_maxNgrains * (1 + crystallite_maxSizePostResults & ! crystallite size & crystallite results
+ 1 + constitutive_damage_maxSizePostResults &
+ 1 + constitutive_thermal_maxSizePostResults &
+ 1 + constitutive_maxSizePostResults) ! constitutive size & constitutive results
allocate(materialpoint_results(materialpoint_sizeResults,mesh_maxNips,mesh_NcpElems))
write(6,'(/,a)') ' <<<+- homogenization init -+>>>'
write(6,'(a)') ' $Id$'
write(6,'(a15,a)') ' Current time: ',IO_timeStamp()
#include "compilation_info.f90"
if (iand(debug_level(debug_homogenization), debug_levelBasic) /= 0_pInt) then
write(6,'(a32,1x,7(i8,1x))') 'homogenization_state0: ', shape(homogenization_state0)
write(6,'(a32,1x,7(i8,1x))') 'homogenization_subState0: ', shape(homogenization_subState0)
write(6,'(a32,1x,7(i8,1x))') 'homogenization_state: ', shape(homogenization_state)
write(6,'(a32,1x,7(i8,1x))') 'homogenization_sizeState: ', shape(homogenization_sizeState)
write(6,'(a32,1x,7(i8,1x),/)') 'homogenization_sizePostResults: ', shape(homogenization_sizePostResults)
write(6,'(a32,1x,7(i8,1x))') 'materialpoint_dPdF: ', shape(materialpoint_dPdF)
write(6,'(a32,1x,7(i8,1x))') 'materialpoint_F0: ', shape(materialpoint_F0)
write(6,'(a32,1x,7(i8,1x))') 'materialpoint_F: ', shape(materialpoint_F)
write(6,'(a32,1x,7(i8,1x))') 'materialpoint_subF0: ', shape(materialpoint_subF0)
write(6,'(a32,1x,7(i8,1x))') 'materialpoint_subF: ', shape(materialpoint_subF)
write(6,'(a32,1x,7(i8,1x))') 'materialpoint_P: ', shape(materialpoint_P)
write(6,'(a32,1x,7(i8,1x))') 'materialpoint_heat: ', shape(materialpoint_heat)
write(6,'(a32,1x,7(i8,1x))') 'materialpoint_subFrac: ', shape(materialpoint_subFrac)
write(6,'(a32,1x,7(i8,1x))') 'materialpoint_subStep: ', shape(materialpoint_subStep)
write(6,'(a32,1x,7(i8,1x))') 'materialpoint_subdt: ', shape(materialpoint_subdt)
write(6,'(a32,1x,7(i8,1x))') 'materialpoint_requested: ', shape(materialpoint_requested)
write(6,'(a32,1x,7(i8,1x))') 'materialpoint_converged: ', shape(materialpoint_converged)
write(6,'(a32,1x,7(i8,1x),/)') 'materialpoint_doneAndHappy: ', shape(materialpoint_doneAndHappy)
write(6,'(a32,1x,7(i8,1x),/)') 'materialpoint_results: ', shape(materialpoint_results)
write(6,'(a32,1x,7(i8,1x))') 'maxSizeState: ', homogenization_maxSizeState
write(6,'(a32,1x,7(i8,1x))') 'maxSizePostResults: ', homogenization_maxSizePostResults
endif
flush(6)
if (debug_g < 1 .or. debug_g > homogenization_Ngrains(mesh_element(3,debug_e))) &
call IO_error(602_pInt,ext_msg='component (grain)')
end subroutine homogenization_init
!--------------------------------------------------------------------------------------------------
!> @brief parallelized calculation of stress and corresponding tangent at material points
!--------------------------------------------------------------------------------------------------
subroutine materialpoint_stressAndItsTangent(updateJaco,dt)
use numerics, only: &
subStepMinHomog, &
subStepSizeHomog, &
stepIncreaseHomog, &
nHomog, &
nMPstate
use math, only: &
math_transpose33
use FEsolving, only: &
FEsolving_execElem, &
FEsolving_execIP, &
terminallyIll
use mesh, only: &
mesh_element, &
mesh_NcpElems, &
mesh_maxNips
use material, only: &
plasticState, &
damageState, &
thermalState, &
mappingConstitutive, &
homogenization_Ngrains
use crystallite, only: &
crystallite_heat, &
crystallite_F0, &
crystallite_Fp0, &
crystallite_Fp, &
crystallite_Lp0, &
crystallite_Lp, &
crystallite_dPdF, &
crystallite_dPdF0, &
crystallite_Tstar0_v, &
crystallite_Tstar_v, &
crystallite_partionedF0, &
crystallite_partionedF, &
crystallite_partionedFp0, &
crystallite_partionedLp0, &
crystallite_partioneddPdF0, &
crystallite_partionedTstar0_v, &
crystallite_dt, &
crystallite_requested, &
crystallite_converged, &
crystallite_stressAndItsTangent, &
crystallite_orientations
use debug, only: &
debug_level, &
debug_homogenization, &
debug_levelBasic, &
debug_levelSelective, &
debug_e, &
debug_i, &
debug_MaterialpointLoopDistribution, &
debug_MaterialpointStateLoopDistribution
use math, only: &
math_pDecomposition
implicit none
real(pReal), intent(in) :: dt !< time increment
logical, intent(in) :: updateJaco !< initiating Jacobian update
logical :: rate_sensitivity
integer(pInt) :: &
NiterationHomog, &
NiterationMPstate, &
g, & !< grain number
i, & !< integration point number
e, & !< element number
myNgrains
!--------------------------------------------------------------------------------------------------
! initialize to starting condition
if (iand(debug_level(debug_homogenization), debug_levelBasic) /= 0_pInt) then
!$OMP CRITICAL (write2out)
write(6,'(/a,i5,1x,i2)') '<< HOMOG >> Material Point start at el ip ', debug_e, debug_i
write(6,'(a,/,3(12x,3(f14.9,1x)/))') '<< HOMOG >> F0', &
math_transpose33(materialpoint_F0(1:3,1:3,debug_i,debug_e))
write(6,'(a,/,3(12x,3(f14.9,1x)/))') '<< HOMOG >> F', &
math_transpose33(materialpoint_F(1:3,1:3,debug_i,debug_e))
!$OMP END CRITICAL (write2out)
endif
!--------------------------------------------------------------------------------------------------
! initialize restoration points of ...
do e = FEsolving_execElem(1),FEsolving_execElem(2)
myNgrains = homogenization_Ngrains(mesh_element(3,e))
forall(i = FEsolving_execIP(1,e):FEsolving_execIP(2,e), g = 1:myNgrains)
plasticState(mappingConstitutive(2,g,i,e))%partionedState0(:,mappingConstitutive(1,g,i,e)) = &
plasticState(mappingConstitutive(2,g,i,e))%state0(:,mappingConstitutive(1,g,i,e))
damageState(mappingConstitutive(2,g,i,e))%partionedState0(:,mappingConstitutive(1,g,i,e)) = &
damageState(mappingConstitutive(2,g,i,e))%state0(:,mappingConstitutive(1,g,i,e))
thermalState(mappingConstitutive(2,g,i,e))%partionedState0(:,mappingConstitutive(1,g,i,e)) = &
thermalState(mappingConstitutive(2,g,i,e))%state0(:,mappingConstitutive(1,g,i,e))
crystallite_partionedFp0(1:3,1:3,g,i,e) = crystallite_Fp0(1:3,1:3,g,i,e) ! ...plastic def grads
crystallite_partionedLp0(1:3,1:3,g,i,e) = crystallite_Lp0(1:3,1:3,g,i,e) ! ...plastic velocity grads
crystallite_partioneddPdF0(1:3,1:3,1:3,1:3,g,i,e) = crystallite_dPdF0(1:3,1:3,1:3,1:3,g,i,e) ! ...stiffness
crystallite_partionedF0(1:3,1:3,g,i,e) = crystallite_F0(1:3,1:3,g,i,e) ! ...def grads
crystallite_partionedTstar0_v(1:6,g,i,e) = crystallite_Tstar0_v(1:6,g,i,e) ! ...2nd PK stress
endforall
forall(i = FEsolving_execIP(1,e):FEsolving_execIP(2,e))
materialpoint_subF0(1:3,1:3,i,e) = materialpoint_F0(1:3,1:3,i,e) ! ...def grad
materialpoint_subFrac(i,e) = 0.0_pReal
materialpoint_subStep(i,e) = 1.0_pReal/subStepSizeHomog ! <<added to adopt flexibility in cutback size>>
materialpoint_converged(i,e) = .false. ! pretend failed step of twice the required size
materialpoint_requested(i,e) = .true. ! everybody requires calculation
endforall
forall(i = FEsolving_execIP(1,e):FEsolving_execIP(2,e), homogenization_sizeState(i,e) > 0_pInt) &
homogenization_subState0(i,e)%p = homogenization_state0(i,e)%p ! ...internal homogenization state
enddo
NiterationHomog = 0_pInt
cutBackLooping: do while (.not. terminallyIll .and. &
any(materialpoint_subStep(:,FEsolving_execELem(1):FEsolving_execElem(2)) > subStepMinHomog))
!$OMP PARALLEL DO PRIVATE(myNgrains)
elementLooping1: do e = FEsolving_execElem(1),FEsolving_execElem(2)
myNgrains = homogenization_Ngrains(mesh_element(3,e))
IpLooping1: do i = FEsolving_execIP(1,e),FEsolving_execIP(2,e)
converged: if ( materialpoint_converged(i,e) ) then
#ifndef _OPENMP
if (iand(debug_level(debug_homogenization), debug_levelBasic) /= 0_pInt &
.and. ((e == debug_e .and. i == debug_i) &
.or. .not. iand(debug_level(debug_homogenization),debug_levelSelective) /= 0_pInt)) then
write(6,'(a,1x,f12.8,1x,a,1x,f12.8,1x,a,i8,1x,i2/)') '<< HOMOG >> winding forward from', &
materialpoint_subFrac(i,e), 'to current materialpoint_subFrac', &
materialpoint_subFrac(i,e)+materialpoint_subStep(i,e),'in materialpoint_stressAndItsTangent at el ip',e,i
endif
#endif
!--------------------------------------------------------------------------------------------------
! calculate new subStep and new subFrac
materialpoint_subFrac(i,e) = materialpoint_subFrac(i,e) + materialpoint_subStep(i,e)
!$OMP FLUSH(materialpoint_subFrac)
materialpoint_subStep(i,e) = min(1.0_pReal-materialpoint_subFrac(i,e), &
stepIncreaseHomog*materialpoint_subStep(i,e)) ! introduce flexibility for step increase/acceleration
!$OMP FLUSH(materialpoint_subStep)
steppingNeeded: if (materialpoint_subStep(i,e) > subStepMinHomog) then
! wind forward grain starting point of...
crystallite_partionedF0(1:3,1:3,1:myNgrains,i,e) = crystallite_partionedF(1:3,1:3,1:myNgrains,i,e) ! ...def grads
crystallite_partionedFp0(1:3,1:3,1:myNgrains,i,e) = crystallite_Fp(1:3,1:3,1:myNgrains,i,e) ! ...plastic def grads
crystallite_partionedLp0(1:3,1:3,1:myNgrains,i,e) = crystallite_Lp(1:3,1:3,1:myNgrains,i,e) ! ...plastic velocity grads
crystallite_partioneddPdF0(1:3,1:3,1:3,1:3,1:myNgrains,i,e) = crystallite_dPdF(1:3,1:3,1:3,1:3,1:myNgrains,i,e)! ...stiffness
crystallite_partionedTstar0_v(1:6,1:myNgrains,i,e) = crystallite_Tstar_v(1:6,1:myNgrains,i,e) ! ...2nd PK stress
forall (g = 1:myNgrains)
plasticState(mappingConstitutive(2,g,i,e))%partionedState0(:,mappingConstitutive(1,g,i,e)) = &
plasticState(mappingConstitutive(2,g,i,e))%state(:,mappingConstitutive(1,g,i,e))
damageState(mappingConstitutive(2,g,i,e))%partionedState0(:,mappingConstitutive(1,g,i,e)) = &
damageState(mappingConstitutive(2,g,i,e))%state(:,mappingConstitutive(1,g,i,e))
thermalState(mappingConstitutive(2,g,i,e))%partionedState0(:,mappingConstitutive(1,g,i,e)) = &
thermalState(mappingConstitutive(2,g,i,e))%state(:,mappingConstitutive(1,g,i,e))
end forall
if (homogenization_sizeState(i,e) > 0_pInt) &
homogenization_subState0(i,e)%p = homogenization_state(i,e)%p ! ...internal state of homog scheme
materialpoint_subF0(1:3,1:3,i,e) = materialpoint_subF(1:3,1:3,i,e) ! ...def grad
!$OMP FLUSH(materialpoint_subF0)
elseif (materialpoint_requested(i,e)) then steppingNeeded ! already at final time (??)
if (iand(debug_level(debug_homogenization), debug_levelBasic) /= 0_pInt) then
!$OMP CRITICAL (distributionHomog)
debug_MaterialpointLoopDistribution(min(nHomog+1,NiterationHomog)) = &
debug_MaterialpointLoopDistribution(min(nHomog+1,NiterationHomog)) + 1
!$OMP END CRITICAL (distributionHomog)
endif
endif steppingNeeded
else converged
if ( (myNgrains == 1_pInt .and. materialpoint_subStep(i,e) <= 1.0 ) .or. & ! single grain already tried internal subStepping in crystallite
subStepSizeHomog * materialpoint_subStep(i,e) <= subStepMinHomog ) then ! would require too small subStep
! cutback makes no sense
!$OMP FLUSH(terminallyIll)
if (.not. terminallyIll) then ! so first signals terminally ill...
!$OMP CRITICAL (write2out)
write(6,*) 'Integration point ', i,' at element ', e, ' terminally ill'
!$OMP END CRITICAL (write2out)
endif
!$OMP CRITICAL (setTerminallyIll)
terminallyIll = .true. ! ...and kills all others
!$OMP END CRITICAL (setTerminallyIll)
else ! cutback makes sense
materialpoint_subStep(i,e) = subStepSizeHomog * materialpoint_subStep(i,e) ! crystallite had severe trouble, so do a significant cutback
!$OMP FLUSH(materialpoint_subStep)
#ifndef _OPENMP
if (iand(debug_level(debug_homogenization), debug_levelBasic) /= 0_pInt &
.and. ((e == debug_e .and. i == debug_i) &
.or. .not. iand(debug_level(debug_homogenization), debug_levelSelective) /= 0_pInt)) then
write(6,'(a,1x,f12.8,a,i8,1x,i2/)') &
'<< HOMOG >> cutback step in materialpoint_stressAndItsTangent with new materialpoint_subStep:',&
materialpoint_subStep(i,e),' at el ip',e,i
endif
#endif
!--------------------------------------------------------------------------------------------------
! restore...
crystallite_Fp(1:3,1:3,1:myNgrains,i,e) = crystallite_partionedFp0(1:3,1:3,1:myNgrains,i,e) ! ...plastic def grads
crystallite_Lp(1:3,1:3,1:myNgrains,i,e) = crystallite_partionedLp0(1:3,1:3,1:myNgrains,i,e) ! ...plastic velocity grads
crystallite_dPdF(1:3,1:3,1:3,1:3,1:myNgrains,i,e) = crystallite_partioneddPdF0(1:3,1:3,1:3,1:3,1:myNgrains,i,e) ! ...stiffness
crystallite_Tstar_v(1:6,1:myNgrains,i,e) = crystallite_partionedTstar0_v(1:6,1:myNgrains,i,e) ! ...2nd PK stress
forall (g = 1:myNgrains)
plasticState(mappingConstitutive(2,g,i,e))%state(:,mappingConstitutive(1,g,i,e)) = &
plasticState(mappingConstitutive(2,g,i,e))%partionedState0(:,mappingConstitutive(1,g,i,e))
damageState(mappingConstitutive(2,g,i,e))%state(:,mappingConstitutive(1,g,i,e)) = &
damageState(mappingConstitutive(2,g,i,e))%partionedState0(:,mappingConstitutive(1,g,i,e))
thermalState(mappingConstitutive(2,g,i,e))%state(:,mappingConstitutive(1,g,i,e)) = &
thermalState(mappingConstitutive(2,g,i,e))%partionedState0(:,mappingConstitutive(1,g,i,e))
end forall
if (homogenization_sizeState(i,e) > 0_pInt) &
homogenization_state(i,e)%p = homogenization_subState0(i,e)%p ! ...internal state of homog scheme
endif
endif converged
if (materialpoint_subStep(i,e) > subStepMinHomog) then
materialpoint_requested(i,e) = .true.
materialpoint_subF(1:3,1:3,i,e) = materialpoint_subF0(1:3,1:3,i,e) + &
materialpoint_subStep(i,e) * (materialpoint_F(1:3,1:3,i,e) - materialpoint_F0(1:3,1:3,i,e))
materialpoint_subdt(i,e) = materialpoint_subStep(i,e) * dt
materialpoint_doneAndHappy(1:2,i,e) = [.false.,.true.]
endif
enddo IpLooping1
enddo elementLooping1
!$OMP END PARALLEL DO
NiterationMPstate = 0_pInt
convergenceLooping: do while (.not. terminallyIll .and. &
any( materialpoint_requested(:,FEsolving_execELem(1):FEsolving_execElem(2)) &
.and. .not. materialpoint_doneAndHappy(1,:,FEsolving_execELem(1):FEsolving_execElem(2)) &
) .and. &
NiterationMPstate < nMPstate)
NiterationMPstate = NiterationMPstate + 1
!--------------------------------------------------------------------------------------------------
! deformation partitioning
! based on materialpoint_subF0,.._subF,crystallite_partionedF0, and homogenization_state,
! results in crystallite_partionedF
!$OMP PARALLEL DO PRIVATE(myNgrains)
elementLooping2: do e = FEsolving_execElem(1),FEsolving_execElem(2)
myNgrains = homogenization_Ngrains(mesh_element(3,e))
IpLooping2: do i = FEsolving_execIP(1,e),FEsolving_execIP(2,e)
if ( materialpoint_requested(i,e) .and. & ! process requested but...
.not. materialpoint_doneAndHappy(1,i,e)) then ! ...not yet done material points
call homogenization_partitionDeformation(i,e) ! partition deformation onto constituents
crystallite_dt(1:myNgrains,i,e) = materialpoint_subdt(i,e) ! propagate materialpoint dt to grains
crystallite_requested(1:myNgrains,i,e) = .true. ! request calculation for constituents
else
crystallite_requested(1:myNgrains,i,e) = .false. ! calculation for constituents not required anymore
endif
enddo IpLooping2
enddo elementLooping2
!$OMP END PARALLEL DO
!--------------------------------------------------------------------------------------------------
! crystallite integration
! based on crystallite_partionedF0,.._partionedF
! incrementing by crystallite_dt
rate_sensitivity = .false. ! request rate sensitive contribution to dPdF
call crystallite_stressAndItsTangent(updateJaco,rate_sensitivity) ! request stress and tangent calculation for constituent grains
!--------------------------------------------------------------------------------------------------
! state update
!$OMP PARALLEL DO
elementLooping3: do e = FEsolving_execElem(1),FEsolving_execElem(2)
IpLooping3: do i = FEsolving_execIP(1,e),FEsolving_execIP(2,e)
if ( materialpoint_requested(i,e) .and. &
.not. materialpoint_doneAndHappy(1,i,e)) then
if (.not. all(crystallite_converged(:,i,e))) then
materialpoint_doneAndHappy(1:2,i,e) = [.true.,.false.]
materialpoint_converged(i,e) = .false.
else
materialpoint_doneAndHappy(1:2,i,e) = homogenization_updateState(i,e)
materialpoint_converged(i,e) = all(homogenization_updateState(i,e)) ! converged if done and happy
endif
!$OMP FLUSH(materialpoint_converged)
if (materialpoint_converged(i,e)) then
if (iand(debug_level(debug_homogenization), debug_levelBasic) /= 0_pInt) then
!$OMP CRITICAL (distributionMPState)
debug_MaterialpointStateLoopdistribution(NiterationMPstate) = &
debug_MaterialpointStateLoopdistribution(NiterationMPstate) + 1_pInt
!$OMP END CRITICAL (distributionMPState)
endif
endif
endif
enddo IpLooping3
enddo elementLooping3
!$OMP END PARALLEL DO
enddo convergenceLooping
NiterationHomog = NiterationHomog + 1_pInt
enddo cutBackLooping
if (.not. terminallyIll ) then
call crystallite_orientations() ! calculate crystal orientations
!$OMP PARALLEL DO
elementLooping4: do e = FEsolving_execElem(1),FEsolving_execElem(2)
IpLooping4: do i = FEsolving_execIP(1,e),FEsolving_execIP(2,e)
call homogenization_averageStressAndItsTangent(i,e)
materialpoint_heat(i,e) = homogenization_averageHeat(i,e)
enddo IpLooping4
enddo elementLooping4
!$OMP END PARALLEL DO
else
!$OMP CRITICAL (write2out)
write(6,'(/,a,/)') '<< HOMOG >> Material Point terminally ill'
!$OMP END CRITICAL (write2out)
endif
end subroutine materialpoint_stressAndItsTangent
!--------------------------------------------------------------------------------------------------
!> @brief parallelized calculation of result array at material points
!--------------------------------------------------------------------------------------------------
subroutine materialpoint_postResults
use FEsolving, only: &
FEsolving_execElem, &
FEsolving_execIP
use mesh, only: &
mesh_element
use material, only: &
plasticState, &
damageState, &
thermalState, &
material_phase, &
homogenization_Ngrains, &
microstructure_crystallite
use constitutive, only: &
constitutive_postResults
use crystallite, only: &
crystallite_sizePostResults, &
crystallite_postResults
implicit none
integer(pInt) :: &
thePos, &
theSize, &
myNgrains, &
myCrystallite, &
g, & !< grain number
i, & !< integration point number
e !< element number
!$OMP PARALLEL DO PRIVATE(myNgrains,myCrystallite,thePos,theSize)
elementLooping: do e = FEsolving_execElem(1),FEsolving_execElem(2)
myNgrains = homogenization_Ngrains(mesh_element(3,e))
myCrystallite = microstructure_crystallite(mesh_element(4,e))
IpLooping: do i = FEsolving_execIP(1,e),FEsolving_execIP(2,e)
thePos = 0_pInt
theSize = homogenization_sizePostResults(i,e)
materialpoint_results(thePos+1,i,e) = real(theSize,pReal) ! tell size of homogenization results
thePos = thePos + 1_pInt
if (theSize > 0_pInt) then ! any homogenization results to mention?
materialpoint_results(thePos+1:thePos+theSize,i,e) = homogenization_postResults(i,e) ! tell homogenization results
thePos = thePos + theSize
endif
materialpoint_results(thePos+1,i,e) = real(myNgrains,pReal) ! tell number of grains at materialpoint
thePos = thePos + 1_pInt
grainLooping :do g = 1,myNgrains
theSize = (1 + crystallite_sizePostResults(myCrystallite)) + &
(1 + plasticState(material_phase(g,i,e))%sizePostResults) + &
(1 + damageState(material_phase(g,i,e))%sizePostResults) + &
(1 + thermalState(material_phase(g,i,e))%sizePostResults)
materialpoint_results(thePos+1:thePos+theSize,i,e) = crystallite_postResults(g,i,e) ! tell crystallite results
thePos = thePos + theSize
enddo grainLooping
enddo IpLooping
enddo elementLooping
!$OMP END PARALLEL DO
end subroutine materialpoint_postResults
!--------------------------------------------------------------------------------------------------
!> @brief partition material point def grad onto constituents
!--------------------------------------------------------------------------------------------------
subroutine homogenization_partitionDeformation(ip,el)
use mesh, only: &
mesh_element
use material, only: &
homogenization_type, &
homogenization_maxNgrains, &
HOMOGENIZATION_NONE_ID, &
HOMOGENIZATION_ISOSTRAIN_ID, &
HOMOGENIZATION_RGC_ID
use crystallite, only: &
crystallite_partionedF0, &
crystallite_partionedF
use homogenization_isostrain, only: &
homogenization_isostrain_partitionDeformation
use homogenization_RGC, only: &
homogenization_RGC_partitionDeformation
implicit none
integer(pInt), intent(in) :: &
ip, & !< integration point
el !< element number
chosenHomogenization: select case(homogenization_type(mesh_element(3,el)))
case (HOMOGENIZATION_NONE_ID) chosenHomogenization
crystallite_partionedF(1:3,1:3,1:homogenization_maxNgrains,ip,el) = 0.0_pReal
crystallite_partionedF(1:3,1:3,1:1,ip,el) = &
spread(materialpoint_subF(1:3,1:3,ip,el),3,1)
case (HOMOGENIZATION_ISOSTRAIN_ID) chosenHomogenization
call homogenization_isostrain_partitionDeformation(&
crystallite_partionedF(1:3,1:3,1:homogenization_maxNgrains,ip,el), &
materialpoint_subF(1:3,1:3,ip,el),&
el)
case (HOMOGENIZATION_RGC_ID) chosenHomogenization
call homogenization_RGC_partitionDeformation(&
crystallite_partionedF(1:3,1:3,1:homogenization_maxNgrains,ip,el), &
materialpoint_subF(1:3,1:3,ip,el),&
homogenization_state(ip,el), &
ip, &
el)
end select chosenHomogenization
end subroutine homogenization_partitionDeformation
!--------------------------------------------------------------------------------------------------
!> @brief update the internal state of the homogenization scheme and tell whether "done" and
!> "happy" with result
!--------------------------------------------------------------------------------------------------
function homogenization_updateState(ip,el)
use mesh, only: &
mesh_element
use material, only: &
homogenization_type, &
homogenization_maxNgrains, &
HOMOGENIZATION_RGC_ID
use crystallite, only: &
crystallite_P, &
crystallite_dPdF, &
crystallite_partionedF,&
crystallite_partionedF0
use homogenization_RGC, only: &
homogenization_RGC_updateState
implicit none
integer(pInt), intent(in) :: &
ip, & !< integration point
el !< element number
logical, dimension(2) :: homogenization_updateState
chosenHomogenization: select case(homogenization_type(mesh_element(3,el)))
case (HOMOGENIZATION_RGC_ID) chosenHomogenization
homogenization_updateState = &
homogenization_RGC_updateState( homogenization_state(ip,el), &
homogenization_subState0(ip,el), &
crystallite_P(1:3,1:3,1:homogenization_maxNgrains,ip,el), &
crystallite_partionedF(1:3,1:3,1:homogenization_maxNgrains,ip,el), &
crystallite_partionedF0(1:3,1:3,1:homogenization_maxNgrains,ip,el),&
materialpoint_subF(1:3,1:3,ip,el),&
materialpoint_subdt(ip,el), &
crystallite_dPdF(1:3,1:3,1:3,1:3,1:homogenization_maxNgrains,ip,el), &
ip, &
el)
case default chosenHomogenization
homogenization_updateState = .true.
end select chosenHomogenization
end function homogenization_updateState
!--------------------------------------------------------------------------------------------------
!> @brief derive average stress and stiffness from constituent quantities
!--------------------------------------------------------------------------------------------------
subroutine homogenization_averageStressAndItsTangent(ip,el)
use mesh, only: &
mesh_element
use material, only: &
homogenization_type, &
homogenization_maxNgrains, &
HOMOGENIZATION_NONE_ID, &
HOMOGENIZATION_ISOSTRAIN_ID, &
HOMOGENIZATION_RGC_ID
use crystallite, only: &
crystallite_P,crystallite_dPdF
use homogenization_isostrain, only: &
homogenization_isostrain_averageStressAndItsTangent
use homogenization_RGC, only: &
homogenization_RGC_averageStressAndItsTangent
implicit none
integer(pInt), intent(in) :: &
ip, & !< integration point
el !< element number
chosenHomogenization: select case(homogenization_type(mesh_element(3,el)))
case (HOMOGENIZATION_NONE_ID) chosenHomogenization
materialpoint_P(1:3,1:3,ip,el) = sum(crystallite_P(1:3,1:3,1:1,ip,el),3)
materialpoint_dPdF(1:3,1:3,1:3,1:3,ip,el) &
= sum(crystallite_dPdF(1:3,1:3,1:3,1:3,1:1,ip,el),5)
case (HOMOGENIZATION_ISOSTRAIN_ID) chosenHomogenization
call homogenization_isostrain_averageStressAndItsTangent(&
materialpoint_P(1:3,1:3,ip,el), &
materialpoint_dPdF(1:3,1:3,1:3,1:3,ip,el),&
crystallite_P(1:3,1:3,1:homogenization_maxNgrains,ip,el), &
crystallite_dPdF(1:3,1:3,1:3,1:3,1:homogenization_maxNgrains,ip,el), &
el)
case (HOMOGENIZATION_RGC_ID) chosenHomogenization
call homogenization_RGC_averageStressAndItsTangent(&
materialpoint_P(1:3,1:3,ip,el), &
materialpoint_dPdF(1:3,1:3,1:3,1:3,ip,el),&
crystallite_P(1:3,1:3,1:homogenization_maxNgrains,ip,el), &
crystallite_dPdF(1:3,1:3,1:3,1:3,1:homogenization_maxNgrains,ip,el), &
el)
end select chosenHomogenization
end subroutine homogenization_averageStressAndItsTangent
!--------------------------------------------------------------------------------------------------
!> @brief derive average heat from constituent quantities (does not depend on choosen
!! homogenization scheme)
!--------------------------------------------------------------------------------------------------
real(pReal) function homogenization_averageHeat(ip,el)
use mesh, only: &
mesh_element
use material, only: &
homogenization_Ngrains
use crystallite, only: &
crystallite_heat
implicit none
integer(pInt), intent(in) :: &
ip, & !< integration point number
el !< element number
integer(pInt) :: &
Ngrains
!--------------------------------------------------------------------------------------------------
! computing the average heat
Ngrains = homogenization_Ngrains(mesh_element(3,el))
homogenization_averageHeat= sum(crystallite_heat(1:Ngrains,ip,el))/real(Ngrains,pReal)
end function homogenization_averageHeat
!--------------------------------------------------------------------------------------------------
!> @brief return array of homogenization results for post file inclusion. call only,
!> if homogenization_sizePostResults(i,e) > 0 !!
!--------------------------------------------------------------------------------------------------
function homogenization_postResults(ip,el)
use mesh, only: &
mesh_element
use material, only: &
homogenization_type, &
HOMOGENIZATION_NONE_ID, &
HOMOGENIZATION_ISOSTRAIN_ID, &
HOMOGENIZATION_RGC_ID
use homogenization_isostrain, only: &
homogenization_isostrain_postResults
use homogenization_RGC, only: &
homogenization_RGC_postResults
implicit none
integer(pInt), intent(in) :: &
ip, & !< integration point
el !< element number
real(pReal), dimension(homogenization_sizePostResults(ip,el)) :: homogenization_postResults
homogenization_postResults = 0.0_pReal
chosenHomogenization: select case (homogenization_type(mesh_element(3,el)))
case (HOMOGENIZATION_NONE_ID) chosenHomogenization
case (HOMOGENIZATION_ISOSTRAIN_ID) chosenHomogenization
homogenization_postResults = homogenization_isostrain_postResults(&
ip, &
el, &
materialpoint_P(1:3,1:3,ip,el), &
materialpoint_F(1:3,1:3,ip,el))
case (HOMOGENIZATION_RGC_ID) chosenHomogenization
homogenization_postResults = homogenization_RGC_postResults(&
homogenization_state(ip,el),&
ip, &
el, &
materialpoint_P(1:3,1:3,ip,el), &
materialpoint_F(1:3,1:3,ip,el))
end select chosenHomogenization
end function homogenization_postResults
end module homogenization