Mailman 3 April 2010 - SfePy

updating entities
by freevryheid 20 Apr '10

20 Apr '10

I noticed in the examples that some of the entities in the problem description file are updated i.e. materials, ecbs and solvers. The updates do not not seem to change the entity parameters but rather add to them. Two questions: 1. Can updates be used to change the entity parameters? e.g. change material parameters. 2. Can all entities be updated? e.g change equations when specific failure criterion is met. Thanks -- You received this message because you are subscribed to the Google Groups "sfepy-devel" group. To post to this group, send email to sfepy...(a)googlegroups.com. To unsubscribe from this group, send email to sfepy-devel...(a)googlegroups.com. For more options, visit this group at http://groups.google.com/group/sfepy-devel?hl=en.

3 3

probe during post_process
by freevryheid 20 Apr '10

20 Apr '10

I presume it's possible to probe cell data in a function call or during post_processing. I'm interested in extrapolating stresses averaged in the elements to the nodes. Could I apply var.evaluate_at(...)? In [49]: from sfepy.fem.evaluate import eval_term_op In [52]: stress = eval_term_op(vec, 'de_cauchy_stress.i1.Omega(solid.D,u)', pb) In [54]: var Out[54]: FieldVariable:s In [59]: stress Out[59]: array([[[[ 2.38035802], [-19.45504858], [ 2.5780901 ]]], ... In [68]: var.data_from_data(stress) In [71]: var.evaluate_at(nm.array[0.0,0.0]) --------------------------------------------------------------------------- TypeError Traceback (most recent call last) /home/grassy/sfepy_dev/sfepy/sfepy/interactive/__init__.pyc in <module>() ----> 1 2 3 4 5 TypeError: 'builtin_function_or_method' object is unsubscriptable In [77]: var.data Out[77]: deque([array([ 2.38035802, -19.45504858, 2.5780901 , ..., -1.92049253, -1.47623867, 1.56398786])]) -- You received this message because you are subscribed to the Google Groups "sfepy-devel" group. To post to this group, send email to sfepy...(a)googlegroups.com. To unsubscribe from this group, send email to sfepy-devel...(a)googlegroups.com. For more options, visit this group at http://groups.google.com/group/sfepy-devel?hl=en.

2 2

p-Laplacian
by Dominique 16 Apr '10

16 Apr '10

Is it possible to model the weak form of equations involving the p- Laplacian in SfePy? I am no expert but I believe the relevant term might be integral( |grad(u)|^p grad(u) . grad(w) ) where p > -1 is some real number. I am not seeing it in the terms table, but it might be there in a form unknown to me. Thanks, Dominique -- You received this message because you are subscribed to the Google Groups "sfepy-devel" group. To post to this group, send email to sfepy...(a)googlegroups.com. To unsubscribe from this group, send email to sfepy-devel...(a)googlegroups.com. For more options, visit this group at http://groups.google.com/group/sfepy-devel?hl=en.

2 3

regions
by freevryheid 15 Apr '10

15 Apr '10

I notice that one cannot define a region by elements i.e. element <id> Is this difficult to implement?

3 8

cache_override
by freevryheid 15 Apr '10

15 Apr '10

I noticed the setting for cache_override in time step problems: fe = { 'chunk_size' : 1000, 'cache_override' : False, } What is its purpose?

2 1

Nonlinear source term
by Dominique 15 Apr '10

15 Apr '10

Hi, I'm trying to modify an existing example, complicating it bit by bit. I started from sinbc.py which is essentially the Laplace equation with Dirichlet boundary conditions specified by functions. I know how to add a constant source term: material_1 = { 'name' : 'coef', 'region' : 'Omega', 'values' : {'val' : 1.0}, } material_2 = { 'name' : 'source_term', 'region' : 'Omega', 'mode' : 'here', 'values' : {'val' : 1.0}, } integral_1 = { 'name' : 'i1', 'kind' : 'v', 'quadrature' : 'gauss_o2_d2', # The elements are 2_3_P1 } equations = { 'Temperature' : """dw_laplace.i1.Omega( coef.val, w, u ) = dw_volume_lvf.i1.Omega( source_term.val, w )""" } That works fine. But how do I add a nonlinear source term specified by a function? I tried the following: material_2 = { 'name' : 'source_term', 'region' : 'Omega', 'mode' : 'function', 'function' : 'source_term_func', } def source_term_func(ts, coor, mode=None, region=None, ig=None, extra_arg=None): r = -25 * nm.sin(3*coor[:,0]) * nm.cos(4*coor[:,1]) r.shape = (r.shape[0], 1, 1) out = {} out['val'] = r return out functions = { 'ebc_sin' : (ebc_sin,), 'ebc_sin2' : (ebc_sin2,), 'source_term_func' : (source_term_func,), } But all I'm getting is an overflow error: sfepy: left over: ['__builtins__', 'ebc_sin', '__file__', '__name__', 'ebc_sin2', 'source_term_func', 'data_dir', 'nm', '_filename', '__package__', 'amplitude', '__doc__'] sfepy: reading mesh (/Users/dpo/local/src/sfepy/meshes/2d/ square_unit_tri.mesh)... sfepy: ...done in 0.00 s sfepy: setting up domain edges... sfepy: ...done in 0.00 s sfepy: creating regions... sfepy: Omega sfepy: Gamma_Top sfepy: Gamma_Bottom sfepy: ...done in 0.01 s sfepy: equation "Temperature": sfepy: dw_laplace.i1.Omega( coef.val, w, u ) = dw_volume_lvf.i1.Omega( source_term.val, w ) sfepy: setting up dof connectivities... sfepy: ...done in 0.00 s sfepy: describing geometries... sfepy: ...done in 0.00 s sfepy: using solvers: nls: newton ls: ls sfepy: matrix shape: (112, 112) sfepy: assembling matrix graph... sfepy: ...done in 0.00 s sfepy: matrix structural nonzeros: 702 (5.60e-02% fill) sfepy: updating materials... sfepy: coef sfepy: source_term sfepy: ...done in 0.00 s sfepy: updating variables... sfepy: ...done Warning: overflow encountered in multiply sfepy: nls: iter: 0, residual: inf (rel: nan) sfepy: rezidual: 0.00 [s] sfepy: solve: 0.00 [s] sfepy: matrix: 0.00 [s] Warning: overflow encountered in multiply sfepy: linear system not solved! (err = inf) Warning: overflow encountered in multiply sfepy: linesearch: iter 1, (inf < inf) (new ls: 1.000000e-01) Warning: overflow encountered in multiply sfepy: linesearch: iter 1, (inf < inf) (new ls: 1.000000e-02) Warning: overflow encountered in multiply sfepy: linesearch: iter 1, (inf < inf) (new ls: 1.000000e-03) Warning: overflow encountered in multiply sfepy: linesearch: iter 1, (inf < inf) (new ls: 1.000000e-04) Warning: overflow encountered in multiply sfepy: linesearch: iter 1, (inf < inf) (new ls: 1.000000e-05) Warning: overflow encountered in multiply sfepy: linesearch: iter 1, (inf < inf) (new ls: 1.000000e-06) Warning: overflow encountered in multiply sfepy: linesearch: iter 1, (inf < inf) (new ls: 1.000000e-07) sfepy: linesearch failed, continuing anyway sfepy: nls: iter: 1, residual: inf (rel: nan) I'm obviously doing something wrong... Thanks! Dominique

2 4

nodal forces
by freevryheid 14 Apr '10

14 Apr '10

In linear elastic analyses, how do I get access to the global stiffness matrix - I want to calculate nodal forces.

3 7

conversion formulas of elastic constants
by Robert Cimrman 14 Apr '10

14 Apr '10

Hi, you can test ([0]) a new ElasticConstants class [1] that implements all possible conversion formulas for: Young’s modulus Poisson’s ratio bulk modulus Lamé’s first parameter shear modulus = Lamé’s second parameter P-wave modulus = longitudinal wave modulus Given any two constants, it can determine the remaining, see [2], [3]. Basically, at [2] there are all the formulas without the P-wave modulus, while at [3] there are all the constants, but not all the formulas (60 for six constants). It uses sympy to some extent for the moment, but later we could just use the resulting relations converted to Python functions. Let me know if it works for you or not. thanks, r. [0] http://github.com/rc/sfepy [1] http://docs.sfepy.org/doc-devel/src/sfepy/mechanics/matcoefs.html [2] http://www.efunda.com/formulae/solid_mechanics/mat_mechanics/calc_elastic_c… [3] http://en.wikipedia.org/wiki/Lam%C3%A9_parameters

2 2

closing issues
by Robert Cimrman 13 Apr '10

13 Apr '10

Hi (Andre), shall we close issues 110, 111? The changes are already in the master repo. What about 116 - are you satisfied with the answer? r.

2 4

Probing data
by freevryheid 12 Apr '10

12 Apr '10

I see some probing functions in the code although no examples or documentation. Is this still under development? I'm looking for a way to extract info from the results file.

3 30