!-------------------------------------------------------------------------------------------------- !> @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 use IO use config use math use material use constitutive use FEsolving use discretization use thermal_isothermal use thermal_conduction use damage_none use damage_nonlocal use results implicit none private logical, public :: & terminallyIll = .false. !< at least one material point is terminally ill !-------------------------------------------------------------------------------------------------- ! General variables for the homogenization at a material point real(pReal), dimension(:,:,:), allocatable, public :: & homogenization_F0, & !< def grad of IP at start of FE increment homogenization_F !< def grad of IP to be reached at end of FE increment real(pReal), dimension(:,:,:), allocatable, public :: & !, protected :: & Issue with ifort homogenization_P !< first P--K stress of IP real(pReal), dimension(:,:,:,:,:), allocatable, public :: & !, protected :: & homogenization_dPdF !< tangent of first P--K stress at IP !-------------------------------------------------------------------------------------------------- type :: tNumerics integer :: & nMPstate !< materialpoint state loop limit real(pReal) :: & subStepMinHomog, & !< minimum (relative) size of sub-step allowed during cutback in homogenization subStepSizeHomog, & !< size of first substep when cutback in homogenization stepIncreaseHomog !< increase of next substep size when previous substep converged in homogenization end type tNumerics type(tNumerics) :: num type :: tDebugOptions logical :: & basic, & extensive, & selective integer :: & element, & ip, & grain end type tDebugOptions type(tDebugOptions) :: debugHomog !-------------------------------------------------------------------------------------------------- interface module subroutine mech_init(num_homog) class(tNode), pointer, intent(in) :: & num_homog !< pointer to mechanical homogenization numerics data end subroutine mech_init module subroutine mech_partition(subF,ip,el) real(pReal), intent(in), dimension(3,3) :: & subF integer, intent(in) :: & ip, & !< integration point el !< element number end subroutine mech_partition module subroutine mech_homogenize(ip,el) integer, intent(in) :: & ip, & !< integration point el !< element number end subroutine mech_homogenize module subroutine mech_results(group_base,h) character(len=*), intent(in) :: group_base integer, intent(in) :: h end subroutine mech_results ! -------- ToDo --------------------------------------------------------- module function mech_RGC_updateState(P,F,F0,avgF,dt,dPdF,ip,el) logical, dimension(2) :: mech_RGC_updateState real(pReal), dimension(:,:,:), intent(in) :: & P,& !< partitioned stresses F,& !< partitioned deformation gradients F0 !< partitioned initial deformation gradients real(pReal), dimension(:,:,:,:,:), intent(in) :: dPdF !< partitioned stiffnesses real(pReal), dimension(3,3), intent(in) :: avgF !< average F real(pReal), intent(in) :: dt !< time increment integer, intent(in) :: & ip, & !< integration point number el !< element number end function mech_RGC_updateState end interface ! ----------------------------------------------------------------------- public :: & homogenization_init, & materialpoint_stressAndItsTangent, & homogenization_forward, & homogenization_results contains !-------------------------------------------------------------------------------------------------- !> @brief module initialization !-------------------------------------------------------------------------------------------------- subroutine homogenization_init class (tNode) , pointer :: & num_homog, & num_homogGeneric, & debug_homogenization print'(/,a)', ' <<<+- homogenization init -+>>>'; flush(IO_STDOUT) debug_homogenization => config_debug%get('homogenization', defaultVal=emptyList) debugHomog%basic = debug_homogenization%contains('basic') debugHomog%extensive = debug_homogenization%contains('extensive') debugHomog%selective = debug_homogenization%contains('selective') debugHomog%element = config_debug%get_asInt('element',defaultVal = 1) debugHomog%ip = config_debug%get_asInt('integrationpoint',defaultVal = 1) debugHomog%grain = config_debug%get_asInt('grain',defaultVal = 1) if (debugHomog%grain < 1 & .or. debugHomog%grain > homogenization_Nconstituents(material_homogenizationAt(debugHomog%element))) & call IO_error(602,ext_msg='constituent', el=debugHomog%element, g=debugHomog%grain) num_homog => config_numerics%get('homogenization',defaultVal=emptyDict) num_homogGeneric => num_homog%get('generic',defaultVal=emptyDict) num%nMPstate = num_homogGeneric%get_asInt ('nMPstate', defaultVal=10) num%subStepMinHomog = num_homogGeneric%get_asFloat('subStepMin', defaultVal=1.0e-3_pReal) num%subStepSizeHomog = num_homogGeneric%get_asFloat('subStepSize', defaultVal=0.25_pReal) num%stepIncreaseHomog = num_homogGeneric%get_asFloat('stepIncrease', defaultVal=1.5_pReal) if (num%nMPstate < 1) call IO_error(301,ext_msg='nMPstate') if (num%subStepMinHomog <= 0.0_pReal) call IO_error(301,ext_msg='subStepMinHomog') if (num%subStepSizeHomog <= 0.0_pReal) call IO_error(301,ext_msg='subStepSizeHomog') if (num%stepIncreaseHomog <= 0.0_pReal) call IO_error(301,ext_msg='stepIncreaseHomog') call mech_init(num_homog) if (any(thermal_type == THERMAL_isothermal_ID)) call thermal_isothermal_init if (any(thermal_type == THERMAL_conduction_ID)) call thermal_conduction_init if (any(damage_type == DAMAGE_none_ID)) call damage_none_init if (any(damage_type == DAMAGE_nonlocal_ID)) call damage_nonlocal_init end subroutine homogenization_init !-------------------------------------------------------------------------------------------------- !> @brief parallelized calculation of stress and corresponding tangent at material points !-------------------------------------------------------------------------------------------------- subroutine materialpoint_stressAndItsTangent(dt) real(pReal), intent(in) :: dt !< time increment integer :: & NiterationHomog, & NiterationMPstate, & i, & !< integration point number e, & !< element number myNgrains real(pReal), dimension(discretization_nIPs,discretization_Nelems) :: & subFrac, & subStep logical, dimension(discretization_nIPs,discretization_Nelems) :: & requested, & converged logical, dimension(2,discretization_nIPs,discretization_Nelems) :: & doneAndHappy integer :: m !-------------------------------------------------------------------------------------------------- ! initialize restoration points do e = FEsolving_execElem(1),FEsolving_execElem(2) do i = FEsolving_execIP(1),FEsolving_execIP(2); call constitutive_initializeRestorationPoints(i,e) subFrac(i,e) = 0.0_pReal converged(i,e) = .false. ! pretend failed step ... subStep(i,e) = 1.0_pReal/num%subStepSizeHomog ! ... larger then the requested calculation requested(i,e) = .true. ! everybody requires calculation if (homogState(material_homogenizationAt(e))%sizeState > 0) & homogState(material_homogenizationAt(e))%subState0(:,material_homogenizationMemberAt(i,e)) = & homogState(material_homogenizationAt(e))%State0( :,material_homogenizationMemberAt(i,e)) if (damageState(material_homogenizationAt(e))%sizeState > 0) & damageState(material_homogenizationAt(e))%subState0(:,material_homogenizationMemberAt(i,e)) = & damageState(material_homogenizationAt(e))%State0( :,material_homogenizationMemberAt(i,e)) enddo enddo NiterationHomog = 0 cutBackLooping: do while (.not. terminallyIll .and. & any(subStep(FEsolving_execIP(1):FEsolving_execIP(2),& FEsolving_execElem(1):FEsolving_execElem(2)) > num%subStepMinHomog)) !$OMP PARALLEL DO PRIVATE(m) elementLooping1: do e = FEsolving_execElem(1),FEsolving_execElem(2) myNgrains = homogenization_Nconstituents(material_homogenizationAt(e)) IpLooping1: do i = FEsolving_execIP(1),FEsolving_execIP(2) if (converged(i,e)) then subFrac(i,e) = subFrac(i,e) + subStep(i,e) subStep(i,e) = min(1.0_pReal-subFrac(i,e),num%stepIncreaseHomog*subStep(i,e)) ! introduce flexibility for step increase/acceleration steppingNeeded: if (subStep(i,e) > num%subStepMinHomog) then ! wind forward grain starting point call constitutive_windForward(i,e) if(homogState(material_homogenizationAt(e))%sizeState > 0) & homogState(material_homogenizationAt(e))%subState0(:,material_homogenizationMemberAt(i,e)) = & homogState(material_homogenizationAt(e))%State (:,material_homogenizationMemberAt(i,e)) if(damageState(material_homogenizationAt(e))%sizeState > 0) & damageState(material_homogenizationAt(e))%subState0(:,material_homogenizationMemberAt(i,e)) = & damageState(material_homogenizationAt(e))%State (:,material_homogenizationMemberAt(i,e)) endif steppingNeeded else if ( (myNgrains == 1 .and. subStep(i,e) <= 1.0 ) .or. & ! single grain already tried internal subStepping in crystallite num%subStepSizeHomog * subStep(i,e) <= num%subStepMinHomog ) then ! would require too small subStep ! cutback makes no sense if (.not. terminallyIll) then ! so first signals terminally ill... print*, ' Integration point ', i,' at element ', e, ' terminally ill' endif terminallyIll = .true. ! ...and kills all others else ! cutback makes sense subStep(i,e) = num%subStepSizeHomog * subStep(i,e) ! crystallite had severe trouble, so do a significant cutback call crystallite_restore(i,e,subStep(i,e) < 1.0_pReal) call constitutive_restore(i,e) if(homogState(material_homogenizationAt(e))%sizeState > 0) & homogState(material_homogenizationAt(e))%State( :,material_homogenizationMemberAt(i,e)) = & homogState(material_homogenizationAt(e))%subState0(:,material_homogenizationMemberAt(i,e)) if(damageState(material_homogenizationAt(e))%sizeState > 0) & damageState(material_homogenizationAt(e))%State( :,material_homogenizationMemberAt(i,e)) = & damageState(material_homogenizationAt(e))%subState0(:,material_homogenizationMemberAt(i,e)) endif endif if (subStep(i,e) > num%subStepMinHomog) then requested(i,e) = .true. doneAndHappy(1:2,i,e) = [.false.,.true.] endif enddo IpLooping1 enddo elementLooping1 !$OMP END PARALLEL DO NiterationMPstate = 0 convergenceLooping: do while (.not. terminallyIll .and. & any( requested(:,FEsolving_execELem(1):FEsolving_execElem(2)) & .and. .not. doneAndHappy(1,:,FEsolving_execELem(1):FEsolving_execElem(2)) & ) .and. & NiterationMPstate < num%nMPstate) NiterationMPstate = NiterationMPstate + 1 !-------------------------------------------------------------------------------------------------- ! deformation partitioning !$OMP PARALLEL DO PRIVATE(myNgrains,m) elementLooping2: do e = FEsolving_execElem(1),FEsolving_execElem(2) myNgrains = homogenization_Nconstituents(material_homogenizationAt(e)) IpLooping2: do i = FEsolving_execIP(1),FEsolving_execIP(2) if(requested(i,e) .and. .not. doneAndHappy(1,i,e)) then ! requested but not yet done m = (e-1)*discretization_nIPs + i call mech_partition(homogenization_F0(1:3,1:3,m) & + (homogenization_F(1:3,1:3,m)-homogenization_F0(1:3,1:3,m))& *(subStep(i,e)+subFrac(i,e)), & i,e) crystallite_dt(1:myNgrains,i,e) = dt*subStep(i,e) ! propagate materialpoint dt to grains crystallite_requested(1:myNgrains,i,e) = .true. ! request calculation for constituents else crystallite_requested(1:myNgrains,i,e) = .false. ! calculation for constituents not required anymore endif enddo IpLooping2 enddo elementLooping2 !$OMP END PARALLEL DO !-------------------------------------------------------------------------------------------------- ! crystallite integration converged = crystallite_stress() !ToDo: MD not sure if that is the best logic !-------------------------------------------------------------------------------------------------- ! state update !$OMP PARALLEL DO PRIVATE(m) elementLooping3: do e = FEsolving_execElem(1),FEsolving_execElem(2) IpLooping3: do i = FEsolving_execIP(1),FEsolving_execIP(2) if (requested(i,e) .and. .not. doneAndHappy(1,i,e)) then if (.not. converged(i,e)) then doneAndHappy(1:2,i,e) = [.true.,.false.] else m = (e-1)*discretization_nIPs + i doneAndHappy(1:2,i,e) = updateState(dt*subStep(i,e), & homogenization_F0(1:3,1:3,m) & + (homogenization_F(1:3,1:3,m)-homogenization_F0(1:3,1:3,m)) & *(subStep(i,e)+subFrac(i,e)), & i,e) converged(i,e) = all(doneAndHappy(1:2,i,e)) ! converged if done and happy endif endif enddo IpLooping3 enddo elementLooping3 !$OMP END PARALLEL DO enddo convergenceLooping NiterationHomog = NiterationHomog + 1 enddo cutBackLooping if (.not. terminallyIll ) then call crystallite_orientations() ! calculate crystal orientations !$OMP PARALLEL DO elementLooping4: do e = FEsolving_execElem(1),FEsolving_execElem(2) IpLooping4: do i = FEsolving_execIP(1),FEsolving_execIP(2) call mech_homogenize(i,e) enddo IpLooping4 enddo elementLooping4 !$OMP END PARALLEL DO else print'(/,a,/)', ' << HOMOG >> Material Point terminally ill' endif end subroutine materialpoint_stressAndItsTangent !-------------------------------------------------------------------------------------------------- !> @brief update the internal state of the homogenization scheme and tell whether "done" and !> "happy" with result !-------------------------------------------------------------------------------------------------- function updateState(subdt,subF,ip,el) real(pReal), intent(in) :: & subdt !< current time step real(pReal), intent(in), dimension(3,3) :: & subF integer, intent(in) :: & ip, & !< integration point el !< element number integer :: c logical, dimension(2) :: updateState real(pReal) :: dPdFs(3,3,3,3,homogenization_Nconstituents(material_homogenizationAt(el))) updateState = .true. chosenHomogenization: select case(homogenization_type(material_homogenizationAt(el))) case (HOMOGENIZATION_RGC_ID) chosenHomogenization do c=1,homogenization_Nconstituents(material_homogenizationAt(el)) dPdFs(:,:,:,:,c) = crystallite_stressTangent(c,ip,el) enddo updateState = & updateState .and. & mech_RGC_updateState(crystallite_P(1:3,1:3,1:homogenization_Nconstituents(material_homogenizationAt(el)),ip,el), & crystallite_partitionedF(1:3,1:3,1:homogenization_Nconstituents(material_homogenizationAt(el)),ip,el), & crystallite_partitionedF0(1:3,1:3,1:homogenization_Nconstituents(material_homogenizationAt(el)),ip,el),& subF,& subdt, & dPdFs, & ip, & el) end select chosenHomogenization end function updateState !-------------------------------------------------------------------------------------------------- !> @brief writes homogenization results to HDF5 output file !-------------------------------------------------------------------------------------------------- subroutine homogenization_results use material, only: & material_homogenization_type => homogenization_type integer :: p character(len=:), allocatable :: group_base,group call results_closeGroup(results_addGroup('current/homogenization/')) do p=1,size(material_name_homogenization) group_base = 'current/homogenization/'//trim(material_name_homogenization(p)) call results_closeGroup(results_addGroup(group_base)) call mech_results(group_base,p) group = trim(group_base)//'/damage' call results_closeGroup(results_addGroup(group)) select case(damage_type(p)) case(DAMAGE_NONLOCAL_ID) call damage_nonlocal_results(p,group) end select group = trim(group_base)//'/thermal' call results_closeGroup(results_addGroup(group)) select case(thermal_type(p)) case(THERMAL_CONDUCTION_ID) call thermal_conduction_results(p,group) end select enddo end subroutine homogenization_results !-------------------------------------------------------------------------------------------------- !> @brief Forward data after successful increment. ! ToDo: Any guessing for the current states possible? !-------------------------------------------------------------------------------------------------- subroutine homogenization_forward integer :: ho do ho = 1, size(material_name_homogenization) homogState (ho)%state0 = homogState (ho)%state damageState(ho)%state0 = damageState(ho)%state enddo end subroutine homogenization_forward end module homogenization