I use the qurey tool to evaluate the forces that are transmitted between the components for predimensioning of the bolts. I definetly want to avoid contact / nonlinearities in the simulation.
These forces are in good correlation with the reaction forces I get from the history output at the boundary conditions.
I cannot model the bolts as rigid, because the output of the forces for the reference points seem to be wrong.
Now that I modeled the bolts as solids, I ask myself what would be a suitable elastic modulus for theses solids?
The stiffness of the bolts has a significant impact on the forces in each bolt and on the loads transmitted in each of the boundary conditions.
Does anyone have reference values for suitable material stiffness of the bolts when not considering pretension and contact?
If pretension and contact were considered, the loads would be transmitted through a larger cross section (compression area) and the connection would therefore be stiffer than just the bolt. So I would assume, that the bolt should have quite a high elastic modulus.
What do you think?
You can find some information regarding bolt stiffness estimation for simplified models in “Practical Finite Element Analysis for Mechanical Engineers” by D. Madier. However, a typical approach (also discussed there) is to use beam elements to represent bolts (connection to parts being fastened is realized in a similar way as in your file). PrePoMax currently doesn’t support beam elements but you could add them manually to the exported input file.
By the way, how did you model those partitions for structured mesh around the holes ? Manually in CAD software or with the help of some automated tool ?
For the mesh I prepared a paramatric sketch in cad where I can change the legth of the squares edge and the hole diameter. I copy/paste that on each hole an “print” breps on the surface.
Then I can define a coarse mesh with only one element per edge /curvature and the holes will still be meshed properly. I find that to be quicker in the end compared to fiddeling with the settings of the mesh or doing it in gmesh.
Well, it seems to be a limitation of Netgen’s quad-dominated meshing algorithm. It often generates triangles even for simple rectangular plates and cylindrical shells. In such cases, it can be avoided by changing the element size but for more complicated geometry it’s unavoidable.
PrePoMax currently uses Netgen as an outside mesher. As far as I understand the meshing in Netgen, first a triangular mesh is generated and then neighbouring triangles are merged in quads. That is why sometimes some tringles are left in the quad meshes.
