Dear Stefano, I will just add that if you intend to use a Discontinuous Galerkin method, there is an example of definition of such a method in the framework of the generic assembly language in the test programs of the interface interface/test/python/demo_laplacian_DG.py interface/test/matlab/demo_laplacian_DG.m
Yves. Le 11/01/2017 à 10:19, Andriy Andreykiv a écrit : > Dear Stefano, > > From your problem description I see you're using discontinuous fem > method for your main variable. > I don't think you can use it directly (unless you're in Discontinuous > Galerkin, but then I don't know much about it.) > Please instead of > FEM_TYPE= 'FEM_PK_DISCONTINUOUS(3,1)'; > try to use > FEM_TYPE= 'FEM_PK(3,1)'; > > Best regards, > Andriy > > On 10 January 2017 at 16:09, Stefano Brambilla > <[email protected] <mailto:[email protected]>> wrote: > > Hello, > > I'm using Getfem 5.0 on Scientific Linux. > > I'm trying to solve a coupled diffusion-transport problem in a > mesh 3D-1D; i have some issues assemblying the stiffness matrix > for the laplace operator (for diffusion term). > I tried building it both with the already-made function > getfem::asm_stiffness_matrix_for_laplacian and with the > generic_assembly class, that means: > getfem::generic_assembly > assem("M$1(#1,#1) += sym(comp(vGrad(#1).vGrad(#1)) (:,i,k, : > ,i,k) )"); > assem.push_mi(mim); > assem.push_mf(mf_u); > assem.push_mat(AM); > assem.assembly(); > > both seems to generate a singular matrix. > I guess that my errors are into the boundary conditions or in the > choice of fem methods and integration method; which is probably > confirmed, sinced i have the same problems with the test code > tests/laplacian.cc > > i have, as output using an iterative method as GMRES: > > Level 2 Warning in /opt/getfem/5.0/include/gmm/gmm_precond_ilut.h, > line 172: pivot 1 too small. try with ilutp ? > Level 2 Warning in /opt/getfem/5.0/include/gmm/gmm_precond_ilut.h, > line 172: pivot 2 too small. try with ilutp ? > ... > Level 2 Warning in /opt/getfem/5.0/include/gmm/gmm_precond_ilut.h, > line 172: pivot 21078 too small. try with ilutp ? > Time to compute preconditionner : 0.07 seconds > iter 0 residual 1 > iter 50 residual 0.117633 > iter 100 residual 0.125247 > iter 150 residual 0.137942 > iter 200 residual 0.159473 > iter 250 residual 0.226573 > iter 300 residual 0.147597 > iter 350 residual 0.150442 > iter 400 residual 0.151829 > iter 450 residual 0.161334 > iter 500 residual 0.177289 > iter 550 residual 0.248701 > Gmres is blocked, exiting > Total time to solve : 2.59 seconds > ============================================ > | An error has been detected !!! | > ============================================ > Error in main.cpp, line 339 : > Solve procedure has failed > > > and, with SuperLU: > > Level 1 Warning in ../../src/getfem_superlu.cc, line 217: SuperLU > solve failed: info =4 > Floating point exception > > > i am using this .param file: > %=================================== > % MESH > %=================================== > NX=10; > MESH_NOISED = 0; > LX=1; > LY=2; > LZ=1; > INCLINE =0; > FT = 1; > RESIDUAL=1E-10; > ROOTFILENAME = '/home/pacs_student/Desktop/pacs_shared/mesh' > %=================================== > % GETFEM DESCRIPTORS > %=================================== > % GetFem type for the 3D mesh > MESH_TYPE = 'GT_PK(3,1)'; > % GetFem type for the Finite Element Method > FEM_TYPE= 'FEM_PK_DISCONTINUOUS(3,1)'; > % GetFem type for the Integration Method > INTEGRATION = 'IM_TETRAHEDRON(6)'; > DATA_FEM_TYPE = 'FEM_PK_DISCONTINUOUS(3,1)'; > GENERIC_DIRICHLET=0; > > > > Thank you all for your help!! > Best regards > Stefano Brambilla > > > > > _______________________________________________ > Getfem-users mailing list > [email protected] <mailto:[email protected]> > https://mail.gna.org/listinfo/getfem-users > <https://mail.gna.org/listinfo/getfem-users> > > > > > _______________________________________________ > Getfem-users mailing list > [email protected] > https://mail.gna.org/listinfo/getfem-users -- Yves Renard ([email protected]) tel : (33) 04.72.43.87.08 Pole de Mathematiques, INSA-Lyon fax : (33) 04.72.43.85.29 20, rue Albert Einstein 69621 Villeurbanne Cedex, FRANCE http://math.univ-lyon1.fr/~renard ---------
_______________________________________________ Getfem-users mailing list [email protected] https://mail.gna.org/listinfo/getfem-users
