Hi Arvind, I can't think of a simple way. You can build a mesh from the slice, and then build a mesh_fem on that mesh, but mapping each slice node to the corresponding mesh_fem dof will be painful as you will need to find the correspondance using dof coordinates
-- Julien Quoting Arvind Ajoy <[EMAIL PROTECTED]>: > Hi > > Sorry to trouble you again. I have a question regarding meshes and slices. > > Before I ask my question, I wish to explain the geometry of the problem. I > have a > cylinder meshed using prisms that have been extruded from triangular > elements. > The axis of the cylinder is the x-axis. > > I am able to solve a Laplacian system to give me a quantity U(x,y,z). I need > to find the > quantity V(y,z) = U(x0,y,z) + U(x1,y,z) + .... + U(xn,y,z). I do this by > first taking slices at > x = x1, x2 .... xn; then generating meshes from each of these silces; and > finally computing > U(xi,y,z) on each mesh (using gf_compute()). I find that the gf_compute() > takes roughly > 1sec at each xi. > On the other hand, if I do not create a mesh from each slice, but > interpolate directly on the > slice, I find that I can perform this operation in 0.04sec per xi. I > understand that this increase in > speed is because interpolating using a slice does not interpolate the > meshfem basis. > > Is there a way for me to convert the values on a slice to that on a mesh? > > i.e. I will calculate V(y,z) using interpolation on slices, and then finally > convert this to values > on a mesh. > > I look forward to your suggestions. > > Thank you for your help. > > Regards > Arvind > _______________________________________________ Getfem-users mailing list [email protected] https://mail.gna.org/listinfo/getfem-users
