#Gmsh mesh.random factor code
That’s why, i don’t really understand why the code is doing this, i mean i feel like when plotting e on my gmsh mesh the values do not correspond to the vertex in an order way or something like that) ? I cannot understand that in any way and i can not explain it. So I think i didn’t clear enough my problem, so i am stuck now because if I want to reproduce the same algorithm but just using a gmsh mesh ( I need to use it because i have relatively complex geometry to test on) as the code in my previous comment ( I mean really changing just the mesh) in this case i don’t see the field anymore, even when I use the same geometry(It means BoxMesh but reproduced with gmsh it means an msh file) (this is what i want to see on the same mesh using gmsh but i don’t see it )
So my question is : i don’t Know if i am representing well e on my mesh? I don’t understand where is the trick and how can i correct it ? import osįile_mean_kalman << and Thank you very much for your response.īelow is the code with the built in mesh and it works perfectly fine, i am using beta_x > beta_y=beta_z, so as you can see if you run the code that the field is correlated in direction x but on both direction y and z the gradient of the field is discontinuous. The thing is when i plot e on the mesh, i can see that my field is not ok, since i have not more correlation on direction x because i made beta_x>beta_y>beta_z So, the thing is I cannot reduce the code this is it and it’s not too hard just calculate stiffness matrix in all directions and the mass matrix and then calculate e. I have just modified the title of the problem so it can be more clear, i did that because i couldn’t modify the other one. M_sp=_matrix((val,col,row))įactor=cholesky(M_sp,ordering_method=“natural”)Į=spsolve(M_sp+Kx_sp+Ky_sp+Kz_sp,alpha*y)įile_mean_kalman = File(‘results/realisation.pvd’) Mesh = Mesh(“calcul_little_specimen.xml”)
#Gmsh mesh.random factor how to
I have read the same problem founded by another person but still cannot understand how to fix that.
#Gmsh mesh.random factor download
I am sure that he does this, because i have tried my code on a box generated by BoxMesh of fenics and my program worked and if i check mesh.cells() and my XML file i can see that the ordering had been changed, So that’s why i am not having my good matricesĪnd to be honest, i don’t know how to fix this thing? Does dolfin destroys the ordering of any mesh and build another ordering (in an ascending way) ? how do you guys get the good stiffness matrix and mass one when you download a mesh of your type? I think that dolfin when downloaded the XML file which has already convert it from MSH, he changes the ordering of the numbering of the elements in ascending way. The thing is when i assemble the stiffness matrix and the mass matrix and try to generate an exemple (MonteCarlo) using these matrices, i have figured out that they are not the good ones. I am trying to read a gmsh file (.msh) using dolfin convert to XML and then download it with dolfin.
I hope that one of you guys can help me because i have been stuck here for a week.
0 kommentar(er)
