Homec4science
Diffusion µSpectre (feat/merge_elasto_plasticity) Browse

muSpectre/plastic_comparator.py
e5adbcb3328bfeat/merge_elasto_plast…

Last Change
Raw File

plastic_comparator.py
View Options

import numpy as np
import scipy.sparse.linalg as sp
import itertools
import sys
sys.path.append("build-clang-5.0-dbg/language_bindings/python")
import muSpectre as msp
import matplotlib.pyplot as plt
# turn of warning for zero division
# (which occurs in the linearization of the logarithmic strain)
np.seterr(divide='ignore', invalid='ignore')
# ----------------------------------- GRID ------------------------------------
# read phase indicator from micrograph: 0=soft, 1=hard
phase = np.load('odd_image.npz')['phase']
phase = phase[:3, :3]
# grid dimensions
Nx = phase.shape[0] # number of pixels in x-direction
Ny = phase.shape[1] # number of pixels in y-direction
shape = [Nx,Ny] # number of voxels in all directions
# ----------------------------- TENSOR OPERATIONS -----------------------------
def deserialise_t4(t4, dim=2):
turnaroud = np.arange(dim**2).reshape(dim,dim).T.reshape(-1)
retval = np.zeros([dim*dim, dim*dim])
for i,j in itertools.product(range(dim**2), repeat=2):
retval[i,j] = t4[:dim, :dim, :dim, :dim, 0,0].reshape(dim**2, dim**2)[turnaroud[i], turnaroud[j]]
pass
return retval
def t2_to_goose(t2_msp, dim):
t2_goose = np.zeros((dim,dim,Nx, Ny))
for i in range(Nx):
for j in range(Ny):
t2_goose[:,:,i,j] = t2_msp[:, i + Nx*j].reshape(dim, dim).T
pass
pass
return t2_goose
def t2_vec_to_goose(t2_msp_vec, dim):
return t2_to_goose(t2_msp_vec.reshape(dim*dim, Nx*Ny)).reshape(-1)
def t2_from_goose(t2_goose, dim):
nb_pix = Nx*Ny
t2_msp = np.zeros((dim**2, nb_pix), order='F')
for i in range(Nx):
for j in range(Ny):
view = t2_msp[:, i + Nx*j].reshape(dim, dim).T
view = t2goose[:,:,i,j].T
pass
pass
return t2_msp
# tensor operations / products: np.einsum enables index notation, avoiding loops
# e.g. ddot42 performs $C_ij = A_ijkl B_lk$ for the entire grid
trans2 = lambda A2 : np.einsum('ijxy ->jixy ',A2 )
ddot22 = lambda A2,B2: np.einsum('ijxy ,jixy ->xy ',A2,B2)
ddot42 = lambda A4,B2: np.einsum('ijklxy,lkxy ->ijxy ',A4,B2)
ddot44 = lambda A4,B4: np.einsum('ijklxy,lkmnxy->ijmnxy',A4,B4)
dot11 = lambda A1,B1: np.einsum('ixy ,ixy ->xy ',A1,B1)
dot22 = lambda A2,B2: np.einsum('ijxy ,jkxy ->ikxy ',A2,B2)
dot24 = lambda A2,B4: np.einsum('ijxy ,jkmnxy->ikmnxy',A2,B4)
dot42 = lambda A4,B2: np.einsum('ijklxy,lmxy ->ijkmxy',A4,B2)
dyad22 = lambda A2,B2: np.einsum('ijxy ,klxy ->ijklxy',A2,B2)
dyad11 = lambda A1,B1: np.einsum('ixy ,jxy ->ijxy ',A1,B1)
# eigenvalue decomposition of 2nd-order tensor: return in convention i,j,x,y
# NB requires to swap default order of NumPy (in in/output)
def eig2(A2):
swap1i = lambda A1: np.einsum('xyi ->ixy ',A1)
swap2 = lambda A2: np.einsum('ijxy->xyij',A2)
swap2i = lambda A2: np.einsum('xyij->ijxy',A2)
vals,vecs = np.linalg.eig(swap2(A2))
vals = swap1i(vals)
vecs = swap2i(vecs)
return vals,vecs
# logarithm of grid of 2nd-order tensors
def ln2(A2):
vals,vecs = eig2(A2)
return sum([np.log(vals[i])*dyad11(vecs[:,i],vecs[:,i]) for i in range(3)])
# exponent of grid of 2nd-order tensors
def exp2(A2):
vals,vecs = eig2(A2)
return sum([np.exp(vals[i])*dyad11(vecs[:,i],vecs[:,i]) for i in range(3)])
# determinant of grid of 2nd-order tensors
def det2(A2):
return (A2[0,0]*A2[1,1]*A2[2,2]+A2[0,1]*A2[1,2]*A2[2,0]+A2[0,2]*A2[1,0]*A2[2,1])-\
(A2[0,2]*A2[1,1]*A2[2,0]+A2[0,1]*A2[1,0]*A2[2,2]+A2[0,0]*A2[1,2]*A2[2,1])
# inverse of grid of 2nd-order tensors
def inv2(A2):
A2det = det2(A2)
A2inv = np.empty([3,3,Nx,Ny])
A2inv[0,0] = (A2[1,1]*A2[2,2]-A2[1,2]*A2[2,1])/A2det
A2inv[0,1] = (A2[0,2]*A2[2,1]-A2[0,1]*A2[2,2])/A2det
A2inv[0,2] = (A2[0,1]*A2[1,2]-A2[0,2]*A2[1,1])/A2det
A2inv[1,0] = (A2[1,2]*A2[2,0]-A2[1,0]*A2[2,2])/A2det
A2inv[1,1] = (A2[0,0]*A2[2,2]-A2[0,2]*A2[2,0])/A2det
A2inv[1,2] = (A2[0,2]*A2[1,0]-A2[0,0]*A2[1,2])/A2det
A2inv[2,0] = (A2[1,0]*A2[2,1]-A2[1,1]*A2[2,0])/A2det
A2inv[2,1] = (A2[0,1]*A2[2,0]-A2[0,0]*A2[2,1])/A2det
A2inv[2,2] = (A2[0,0]*A2[1,1]-A2[0,1]*A2[1,0])/A2det
return A2inv
# ------------------------ INITIATE (IDENTITY) TENSORS ------------------------
# identity tensor (single tensor)
i = np.eye(3)
# identity tensors (grid)
I = np.einsum('ij,xy' , i ,np.ones([Nx,Ny]))
I4 = np.einsum('ijkl,xy->ijklxy',np.einsum('il,jk',i,i),np.ones([Nx,Ny]))
I4rt = np.einsum('ijkl,xy->ijklxy',np.einsum('ik,jl',i,i),np.ones([Nx,Ny]))
I4s = (I4+I4rt)/2.
II = dyad22(I,I)
#---------------------------- µSpectre init -----------------------------------
resolution = list(phase.shape); resolution.append(1)
dim = len(resolution)
center = np.array([r//2 for r in resolution])
incl = resolution[0]//5
## Domain dimensions
lengths = [float(r) for r in resolution]
## formulation (small_strain or finite_strain)
formulation = msp.Formulation.finite_strain
## build a computational domain
rve = msp.Cell(resolution, lengths, formulation)
def get_E_nu(bulk, shear):
Young = 9*bulk*shear/(3*bulk + shear)
Poisson = Young/(2*shear) - 1
return Young, Poisson
E, nu = get_E_nu(.833, .386)
H = (2000.0e6/200.0e9,2.*2000.0e6/200.0e9)
tauy0 = ( 600.0e6/200.0e9,2.* 600.0e6/200.0e9)
## define the material properties of the matrix and inclusion
mat = msp.material.MaterialHyperElastoPlastic1_3d if dim == 3 else msp.material.MaterialHyperElastoPlastic1_2d
hard = mat.make(
rve, "hard", E, nu, tauy0[1], H[1])
soft = mat.make(
rve, "soft", E, nu, tauy0[0], H[0])
## assign each pixel to exactly one material
for i, pixel in enumerate(rve):
print(pixel)
if (phase[pixel[0], pixel[1]] == 1).all():
hard.add_pixel(pixel)
else:
soft.add_pixel(pixel)
## define the convergence tolerance for the Newton-Raphson increment
tol = 1e-5
## tolerance for the solver of the linear cell
cg_tol = 1e-8
eq_tol = 1e-8
#-------------------------End µSpertre init -----------------------------------
# ------------------------------------ FFT ------------------------------------
# 2-D projection operator (only for non-zero frequency, associated with the mean)
# - allocate / support function
Ghat4_2 = np.zeros([2,2,2,2,Nx,Ny]) # projection operator
x_2 = np.zeros([2 ,Nx,Ny],dtype='int64') # position vectors
q_2 = np.zeros([2 ,Nx,Ny],dtype='int64') # frequency vectors
delta = lambda i,j: np.float(i==j) # Dirac delta function
# - set "x_2" as position vector of all grid-points
x_2[0],x_2[1] = np.mgrid[:Nx,:Ny]
# - convert positions "x_2" for frequencies "q_2"
for i in range(2):
freq = np.arange(-(shape[i]-1)/2,+(shape[i]+1)/2,dtype='int64')
q_2[i] = freq[x_2[i]]
# - compute "Q = ||q_2||", and "norm = 1/Q" being zero for the mean (Q==0)
# NB: avoid zero division
q_2 = q_2.astype(np.float)
Q = dot11(q_2,q_2)
Z = Q==0
Q[Z] = 1.
norm = 1./Q
norm[Z] = 0.
# - set projection operator [grid of tensors]
for i, j, l, m in itertools.product(range(2), repeat=4):
Ghat4_2[i,j,l,m] = norm*delta(i,m)*q_2[j]*q_2[l]
# 2-D (inverse) Fourier transform (for each tensor component in each direction)
fft = lambda x: np.fft.fftshift(np.fft.fftn (np.fft.ifftshift(x),[Nx,Ny]))
ifft = lambda x: np.fft.fftshift(np.fft.ifftn(np.fft.ifftshift(x),[Nx,Ny]))
# functions for the 2-D projection 'G', and the product 'G : K^LT : (delta F)^T'
G = lambda A2_2 : np.real( ifft( ddot42(Ghat4_2,fft(A2_2)) ) ).reshape(-1)
K_dF = lambda dFm_2: trans2(ddot42(K4_2,trans2(dFm_2.reshape(2,2,Nx,Ny))))
G_K_dF = lambda dFm_2: G(K_dF(dFm_2))
# --------------------------- CONSTITUTIVE RESPONSE ---------------------------
# constitutive response to a certain loading and history
# NB: completely uncoupled from the FFT-solver, but implemented as a regular
# grid of quadrature points, to have an efficient code;
# each point is completely independent, just evaluated at the same time
def constitutive(F,F_t,be_t,ep_t):
# function to compute linearization of the logarithmic Finger tensor
def dln2_d2(A2):
vals,vecs = eig2(A2)
K4 = np.zeros([3,3,3,3,Nx,Ny])
for m, n in itertools.product(range(3),repeat=2):
gc = (np.log(vals[n])-np.log(vals[m]))/(vals[n]-vals[m])
gc[vals[n]==vals[m]] = (1.0/vals[m])[vals[n]==vals[m]]
K4 += gc*dyad22(dyad11(vecs[:,m],vecs[:,n]),dyad11(vecs[:,m],vecs[:,n]))
return K4
# elastic stiffness tensor
C4e = K*II+2.*mu*(I4s-1./3.*II)
# trial state
Fdelta = dot22(F,inv2(F_t))
be_s = dot22(Fdelta,dot22(be_t,trans2(Fdelta)))
lnbe_s = ln2(be_s)
tau_s = ddot42(C4e,lnbe_s)/2.
taum_s = ddot22(tau_s,I)/3.
taud_s = tau_s-taum_s*I
taueq_s = np.sqrt(3./2.*ddot22(taud_s,taud_s))
N_s = 3./2.*taud_s/taueq_s
phi_s = taueq_s-(tauy0+H*ep_t)
phi_s = 1./2.*(phi_s+np.abs(phi_s))
# return map
dgamma = phi_s/(H+3.*mu)
ep = ep_t + dgamma
tau = tau_s -2.*dgamma*N_s*mu
lnbe = lnbe_s-2.*dgamma*N_s
be = exp2(lnbe)
P = dot22(tau,trans2(inv2(F)))
# consistent tangent operator
a0 = dgamma*mu/taueq_s
a1 = mu/(H+3.*mu)
C4ep = ((K-2./3.*mu)/2.+a0*mu)*II+(1.-3.*a0)*mu*I4s+2.*mu*(a0-a1)*dyad22(N_s,N_s)
dlnbe4_s = dln2_d2(be_s)
dbe4_s = 2.*dot42(I4s,be_s)
K4 = (C4e/2.)*(phi_s<=0.).astype(np.float)+C4ep*(phi_s>0.).astype(np.float)
K4 = ddot44(K4,ddot44(dlnbe4_s,dbe4_s))
K4 = dot42(-I4rt,tau)+K4
K4 = dot42(dot24(inv2(F),K4),trans2(inv2(F)))
return P,P[:2,:2],K4[:2,:2,:2,:2],be,ep
# function to convert material parameters to grid of scalars
param = lambda soft,hard: soft*np.ones([Nx,Ny])*(1.-phase)+\
hard*np.ones([Nx,Ny])* phase
# material parameters
K = param( 0.833 , 0.833 ) # bulk modulus
mu = param( 0.386 , 0.386 ) # shear modulus
H = param(2000.0e6/200.0e9,2.*2000.0e6/200.0e9) # hardening modulus
tauy0 = param( 600.0e6/200.0e9,2.* 600.0e6/200.0e9) # initial yield stress
# ---------------------------------- LOADING ----------------------------------
# stress, deformation gradient, plastic strain, elastic Finger tensor
# NB "_t" signifies that it concerns the value at the previous increment
ep_t = np.zeros([ Nx,Ny])
P = np.zeros([3,3,Nx,Ny])
F = np.array(I,copy=True)
F_t = np.array(I,copy=True)
be_t = np.array(I,copy=True)
# initialize macroscopic incremental loading
ninc = 250
epsbar = np.linspace(0.0,0.2,ninc+1)[1:]
stretch = np.exp(np.sqrt(3.0)/2.0*epsbar)
barF = np.array(I,copy=True)
barF_t = np.array(I,copy=True)
rve.set_uniform_strain(np.eye(dim))
mspF = rve.get_strain()
mspbarF = mspF.copy()
mspbarF_t = mspF.copy()
mspP, mspK = rve.evaluate_stress_tangent(mspF)
mspG_K_dF = lambda x: rve.directional_stiffness(x.reshape(F.shape)).reshape(-1)
mspG = lambda x: rve.project(x).reshape(-1)
# initial tangent operator: the elastic tangent
K4_2 = (K*II+2.*mu*(I4s-1./3.*II))[:2,:2,:2,:2]
# incremental deformation
for inc,lam in zip(range(1,ninc+1),stretch[:24]):
print('=============================')
print('inc: {0:d}'.format(inc))
# set macroscopic deformation gradient (pure-shear)
barF = np.array(I,copy=True)
barF[0,0] = (lam)
barF[1,1] = 1./(lam)
print("Step {}:\n{}".format(inc, barF[:,:,0,0]))
# first iteration residual: distribute "barF" over grid using "K4"
b = -G_K_dF((barF-barF_t)[:2,:2])
F += barF-barF_t
# parameters for Newton iterations: normalization and iteration counter
Fn = np.linalg.norm(F)
iiter = 0
# iterate as long as the iterative update does not vanish
while True:
# solve linear system using the Conjugate Gradient iterative solver
dFm,i = sp.cg(tol=1.e-8,
A = sp.LinearOperator(shape=(2*2*Nx*Ny,2*2*Nx*Ny),matvec=G_K_dF,dtype='float'),
b = b,
maxiter = 500,
)
if i: raise IOError('CG-solver failed, i = {}'.format(i))
# apply iterative update to (3-D) DOFs
F[:2,:2] += dFm.reshape(2,2,Nx,Ny)
# compute residual stress and tangent, convert to residual
P,P_2,K4_2,be,ep = constitutive(F,F_t,be_t,ep_t)
b = -G(P_2)
# check for convergence, print convergence info to screen
print('norm: {0:10.9e}, nb_newon = {1}'.format(np.linalg.norm(dFm)/Fn, iiter))
if np.linalg.norm(dFm)/Fn<1.e-5 and iiter>0:
print("F:\n{}".format(F[:,:,0,0]))
print("P:\n{}".format(P[:,:,0,0]))
break
if iiter>9: raise IOError('Maximum number of Newton iterations have been exceeded')
iiter += 1
# end-of-increment: update history
barF_t = np.array(barF,copy=True)
F_t = np.array(F ,copy=True)
be_t = np.array(be ,copy=True)
ep_t = np.array(ep ,copy=True)
np.savez('odd_ouput_inc%03d.npz'%inc,ep=ep)
c4science · Help