Great software, just the tutorials and documentation are bit on the slim side. One thing I'm struggling with is adding a point to an geometry. For example I have a cube (solid->primitives) , add a load to on rib and calculate the stresses/displacements when constrained on four corners: works fine. Now I want to constrain it on four points arbitrarily placed on the surface or internally in the body, I can draw the points but the solver doesn't like it, throws fatal errors. I guess it doesn't understand my points are part of this body, how to explain? I tried remeshing, no luck.
Seccondly I would be interested in more background information on how the solver works internally. I need a solution for a pretty extended case and want to verify the solutions are actually physically correct. Specifically I'm looking for ppb accuracy on a macroscoping but very thin plate (shell) where stress and displacement (static) are resolved through the thickness of the shell. I know shells are faforred, but are they still if wer want ppb accuracy and resolution through the thickness? Or should I maybe consider full 3D solving? Any pointers to background information on the numerical methods used would be very helpfull.
Dear Bastiaan, 1.- You can prescribe loads/constraints to geometry, but remember that mesh & geometry should be associated in any way inorder that FEMAP to be successful to exporting the data to the solver. Said that, the practical method I will use is to split the cube by planes passing through the points you draw, then remesh your model, and for sure you will have a node at every point, then you will be able to apply loads & constrainst at your points, OK?. You can use commands like GEOMETRY > SOLID > SLICE to trim the solid by planes. It is important to understant the philosophy that is behind FEMAP, after that everything is more easy!. 2.- I don't understabd what means PPB, or words like "faforred", sorry!. Shell CQUAD4 elements are perfect if the structure works like a Shell, it means stress varying linearly in the thickness, and element size times bigger than the thickness. You are free to mesh with solid element (in this case I strongly suggest to use solid CHEXA elements, please forgot at all TET meshing) but please note that you need minimum three element in the thickness in order to capture correctly the stress gradients, OK?. Then make your numbers, depending the model dimensions you will have millions of nodes!!. Take care and good luck!. Best regards, Blas.