µSpectre

Copyright © 2018 Till Junge

µSpectre 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, or (at your option) any later version.

µSpectre 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 GNU Emacs; see the file COPYING. If not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.

Building µSpectre

µSpectre is a CMake project that uses C++14. It depends on the Boost unit test framework for testing and uses uses python3 as secondary API. You will need a modern C++ compiler (µSpectre was tested with gcc-6 , gcc-7, clang-4 and clang5) and CMake version 3.7.0 or higher.

Compilation

$ git clone https://c4science.ch/source/muSpectre.git
$ cd build
$ cmake -DCMAKE_BUILD_TYPE=Release ..
$ make

µSpectre makes use of expression templates, as a result, production code must be compiled with the CMAKE_BUILD_TYPE flag set to Release in order to get performance (under gcc, non-optimised code is about 50 times slower than the release). Be careful with parallel compilation: compiling µSpectre is quite memory-hungry because of the use of expression templates, a parallel make -j requires currently about 10 GB of RAM under GCC.

Simple usage example

The following is a simple example for using µSpectre through its convenient Python interface

#!/usr/bin/env python3

import numpy as np
import pyMuSpectre as µ

# setting the geometry
resolution = [51, 51]
center = np.array([r//2 for r in resolution])
incl = resolution[0]//5

lengths = [7., 5.]
formulation = µ.Formulation.small_strain

# creating the periodic cell
rve = µ.SystemFactory(resolution,
                      lengths,
                      formulation)
hard = µ.material.MaterialHooke2d.make(
    rve, "hard", 10e9, .33)
soft = µ.material.MaterialHooke2d.make(
    rve, "soft",  70e9, .33)


# assign a material to each pixel
for i, pixel in enumerate(rve):
    if np.linalg.norm(center - np.array(pixel),2)<incl:
        hard.add_pixel(pixel)
    else:
        soft.add_pixel(pixel)

tol = 1e-5
cg_tol = 1e-8

# set macroscopic strain
Del0 = np.array([[.0, .0],
                 [0,  .03]])
if formulation == µ.Formulation.small_strain:
    Del0 = .5*(Del0 + Del0.T)
maxiter = 401
verbose = 2

solver = SolverCG(rve, cg_tol, maxiter, verbose=False)
r = µ.solvers.newton_cg(rve, Del0, solver, tol, verbose)
print("nb of {} iterations: {}".format(solver.name(), r.nb_fev))

You can find more examples using both the python and the c++ interface in the bin/ and tests folder.