Hi I’m new to FEA/PrePoMax. This might be a silly question but I was wondering if there was a way to split surfaces without breaking all of the dependent features/contacts/constraints. I want to apply a remote load on a subsection of a surface. I saw there was a tool called Split Part Mesh Using Surfaces, but I would probably have to import my model geometry again with the surfaces to cut and then that would require me to re-select all of the required nodes/surfaces for the dependent features. I was wondering if there was a way to do this without having to reinsert the geometry with surfaces to cut and then regenerate the mesh since that would require me to reselect all the points/surfaces for the below features. If there’s also a better way to do this besides Split Part Mesh Using Surfaces that would also be helpful information!
As some context, this is a motion table that tilts and I was looking for the reaction force from the bolts (blue) that are holding the flanged base to a larger plate. I used a remote force applied to the entire surface of the flanged base plate, but I want the remote force to be applied only to the area of the pole that sticks out of the base plate. The circle is meant to represent the center of mass of the object that is being tilted. Thank you!
There’s also a tool to split surfaces using two points. But usually partitioning is done in CAD software before importing geometry to PrePoMax. For example, I can tell you how to do it in FreeCAD. You could try doing this in CAD and then use the Regenerate Using Other Files option in PrePoMax to swap the models. PrePoMax tries to reassign the features properly. Of course, it may not always work depending on the degree of changes.
One important setting when selecting locations for analysis feature is the choice between “Selection by location” and "Selection by ID”. From the manual:
Two selection methods are available - by location and by ID. The first one is based on the location of the user selection (point or area). Location data is saved for when there is a need to recreate the selection (remeshing, regenerating). If the selection is recreated the location data is used to create the selection again. The second one uses the location of the user selection to get the geometry IDs of the underlying geometry and saves only the IDs. When the selection needs to be recreated the geometry IDs are directly used again. In standard usage, like remeshing and regenerating, those two methods are equivalent. However, if the geometry changes before remeshing or regenerating (Regenerate Using Other Files or Swap Part Geometries), the results of the selection may change. If the topology (numbering of vertices, edges, and faces) does not change and only the geometry (dimensions) changes, the ID-based selection will always select the same vertex, edge, face of part while the location-based method might fail to do so.
I see thank you for responding so quickly! Sorry it took so long to respond. I wasn’t able to return to work on this project and test out your tips until now. I think I was able to split the surface so I can choose only a part of the surface to apply a force to.
There’s no problem with the shell model calculated in the linear range and the nonlinear solid. When I change the model to solid and apply compression support, I can’t get the results to converge.
Thank you all for the helpful advice! However, the case I have is not a flanged square tube on concrete. I have a flanged square tube that is connected to a flat metal plate with nuts and bolts so I do want to capture the stress and deflection on that plate as well. This is a picture for more context. In the middle is where the bottom plate connects to a flanged shaft collar which then connects to a shaft. The whole plate rotates about this shaft. I was planning on doing the stress analysis for the bolts that are holding the flanged shaft collar the plate in a different analysis file.
There are a few ways to model the bolts. In most FEA software, it’s usually best to use 1D (e.g. beam) elements in such cases, but not really in CalculiX. So you can either replace them with constraints (even rigid body) or model them as solid cylinders (IMO the best way in CalculiX).
Are you suggesting that I should model the bolts as solid cylinders where they would attach the flanged square tubing to the base plate? Would you suggest applying a similar method as in this video?
Not necessarily, you could also try some simplified approaches with springs/beams or even just rigid body or coupling constraints. However, in the end, it’s usually easiest to just use solid cylinders to represent bolts. You can find many examples of that on the forum, sometimes even with just one cylinder for the shank (without modeling the head/nut): Reaction forces in history output. Baseplate with bolt connection
