Regarding this recent Post I think there is something not fully right.
I found another related issue that made recall this post.
Hydrostatic “Pressure” is a direction dependent load.
Prepomax discretize the Pressure (*DLOAD) into a Traction (*CLOAD) which is Local Coordinate system dependent. In the Silo problem for example, One would expect hydrostatic pressure to act always perpendicular to the surface no matter the coordinates systems around. If one applies a local load in Cylindrical on the silo surface, the hydrostatic pressure will follow the local coordinate system.
Hydrostatic pressure as implemented is not a “Pressure (DLOAD)” anymore.
I have found the issue when tring to apply the Patch load to the Silo.
True. The problem with *dload is that it can use only one pressure value per element. If the elements are large, using *cload in nodes produces smoother pressure field.
I see only one solution, to add a model check that will find the nodes with defined items using different coordinate systems and warn the user about it.
I commented this with Victor as Mecway has the same aproach.
He also recognizes the problem but has made an interesting observation. Even if the element has a single value , ccx will distribute the pressure and will end up averaging over the contiguous nodes so the difference should not be that much. what do you think?
I think it would be of great benefit to keep hydrostatic Pressure independent of the coordinate systems arround. In reservoirs in general it is very common to have local loads due to support clips , nozzles, reinforcing pads,… that would affect the hydrostatic pressure in the current implementation. It could also be affected by boundary conditions applied with a local coordinate system.
I did not test the *load implementation, but I can try it. If the mesh is fine enough, this should not matter. It would only be noticeable for coarse meshes.
The current implementation interpolates the hydrostatic load given at the nodal locations over the element face using the interpolation functions. Then, it integrates the interpolated field over the element face to obtain the actual nodal equivalent forces.
