corrected stress integration scheme: now use norm(Lpresiduum) as a target function for Armijo's rule instead of whole tensor Lp; also corrected the guess for the improvement in Lp
This commit is contained in:
parent
ca3d21a3b6
commit
9d1bc584d0
|
@ -2592,6 +2592,7 @@ use debug, only: debug_verbosity, &
|
|||
use constitutive, only: constitutive_homogenizedC, &
|
||||
constitutive_LpAndItsTangent
|
||||
use math, only: math_mul33x33, &
|
||||
math_mul33xx33, &
|
||||
math_mul66x6, &
|
||||
math_mul99x99, &
|
||||
math_transpose3x3, &
|
||||
|
@ -2629,8 +2630,8 @@ real(pReal), dimension(3,3):: Fg_new, & ! deformation
|
|||
Lp_constitutive, & ! plastic velocity gradient resulting from constitutive law
|
||||
residuum, & ! current residuum of plastic velocity gradient
|
||||
residuum_old, & ! last residuum of plastic velocity gradient
|
||||
deltaLp, &
|
||||
expectedImprovement, &
|
||||
deltaLp, & ! direction of next guess
|
||||
gradientR, & ! derivative of the residuum norm
|
||||
A, &
|
||||
B, &
|
||||
BT, &
|
||||
|
@ -2645,6 +2646,7 @@ real(pReal), dimension(3,3,3,3):: C ! 4th rank ela
|
|||
real(pReal), dimension(6,6):: C_66 ! simplified 2nd rank elasticity tensor
|
||||
real(pReal) p_hydro, & ! volumetric part of 2nd Piola-Kirchhoff Stress
|
||||
det, & ! determinant
|
||||
expectedImprovement, &
|
||||
steplength0, &
|
||||
steplength, &
|
||||
steplength_max, &
|
||||
|
@ -2689,7 +2691,6 @@ endif
|
|||
!* feed local variables
|
||||
|
||||
Fp_current = crystallite_subFp0(1:3,1:3,g,i,e)
|
||||
Tstar_v = crystallite_Tstar_v(1:6,g,i,e)
|
||||
Lpguess_old = crystallite_Lp(1:3,1:3,g,i,e) ! consider present Lp good (i.e. worth remembering) ...
|
||||
Lpguess = crystallite_Lp(1:3,1:3,g,i,e) ! ... and take it as first guess
|
||||
|
||||
|
@ -2815,7 +2816,7 @@ LpLoop: do
|
|||
if (steplength <= steplength0) then
|
||||
|
||||
!* Armijo rule is fullfilled (norm of residuum was reduced by at least 0.01 of what was expected)
|
||||
if (math_norm33(residuum) <= math_norm33(residuum_old + 0.01_pReal * steplength * expectedImprovement)) then
|
||||
if (math_norm33(residuum) <= math_norm33(residuum_old) + 0.01_pReal * steplength * expectedImprovement) then
|
||||
|
||||
!* reduced step length -> return to normal step length
|
||||
if (steplength < steplength0) then
|
||||
|
@ -2838,7 +2839,7 @@ LpLoop: do
|
|||
|
||||
!* no NaN, Armijo rule fullfilled, and sign of residuum did not change -> accelerate if not yet at maximum speed
|
||||
if (.not. any(residuum /= residuum) &
|
||||
.and. math_norm33(residuum) <= math_norm33(residuum_old + 0.01_pReal * steplength * expectedImprovement) &
|
||||
.and. math_norm33(residuum) <= math_norm33(residuum_old) + 0.01_pReal * steplength * expectedImprovement &
|
||||
.and. sum(residuum * residuum_old) >= 0.0_pReal) then
|
||||
if (steplength + 1_pInt <= steplength_max) then
|
||||
steplength = steplength + 1.0_pReal
|
||||
|
@ -2906,10 +2907,12 @@ LpLoop: do
|
|||
deltaLp(k,l) = deltaLp(k,l) - inv_dR_dLp(3*(k-1)+l,3*(m-1)+n) * residuum(m,n)
|
||||
enddo; enddo; enddo; enddo
|
||||
|
||||
expectedImprovement = 0.0_pReal
|
||||
gradientR = 0.0_pReal
|
||||
do k=1,3; do l=1,3; do m=1,3; do n=1,3
|
||||
expectedImprovement(k,l) = expectedImprovement(k,l) - dR_dLp(3*(k-1)+l,3*(m-1)+n) * deltaLp(m,n)
|
||||
gradientR(k,l) = gradientR(k,l) + dR_dLp(3*(k-1)+l,3*(m-1)+n) * residuum(m,n)
|
||||
enddo; enddo; enddo; enddo
|
||||
gradientR = gradientR / math_norm33(gradientR)
|
||||
expectedImprovement = math_mul33xx33(deltaLp, gradientR)
|
||||
endif
|
||||
|
||||
jacoCounter = jacoCounter + 1_pInt ! increase counter for jaco update
|
||||
|
|
Loading…
Reference in New Issue