I'm trying to model a load (a long mountain with vertical sides) placed on an originally unstressed elastic half space. I know what the solution should be based on an analytical solution. The solution I'm getting with COMSOL doesn't match that. Instead of peaking slightly below the mountain, the 2nd invariant of the stress just keeps increasing deeper in the halfspace. One issue I've thought of might be that the "half space" part in the COMSOL model isn't a semi-infinite halfspace just a very large domain with respect to the moutain, but trying to make the space larger leads to increasingly high stresses throughout. I'm also not certain what the boundary conditions for this finite space should be; I've tried having only the top surface free with the others fixed (as in the attached image) as well as having top and sides free or top and bottom free and none give a stress distribution matching the analytical solution. I'd appreciate any insight into why the COMSOL solution might be different than expected.
↧