Pressure distribution (traction) applied to model with all tetra respect the total load (sum). Same pressure distribution applied to a model with C3D8 (also C3D20 tried) doesn't reach the theoretical load. I tried to reduce mesh size but not negligible different remains
I reproduce 3 different simple example
Comments
But it's odd that hex20 is worse than tet. Would you mind sharing the files, especially the coarse hex mesh?
I think the main source of error is that due to the faceted shape, the x and z values in the formulas will be consistently inside the cylinder. I changed the formula to replace 5 with sqrt(x^2+z^2) to represent the actual radius, and that improved the reaction force to -9.97 kN on the original hex8 mesh. The remaining error is probably because the total area is smaller than the ideal cylinder's.
Curiously, this seems to be the same problem with a solution described by @kennethfugate recently https://mecway.com/forum/discussion/comment/7514