Interference fits, radial stress and plotting along a path

General Questions
1

Hello,
I’m a stress/structures engineer and I regularly use ansys and msc tools for mechanical systems and interactions of those systems with housings and structures. We use a lot of contact and interference fits in our models!

Just recently we’ve been base lining contact management and performance in the packages available to us, using the same test - a simple shaft in a socket with interference fit. I decided I would try to solve this using prepomax/calculix over the weekend. The software solution is compared against a classical analytical approach away from stress concentrations/risers.

The setup and running was fine, once I understood the steps and generally the solve time was pretty good using 4 cores. The only stumbling block is that gmsh wouldn’t mesh the socket if I split the upper and lower surfaces using the vertexes of the axial edges. I wanted to do this to create a history plot along an edge.

After a little searching I see that you can add a keyword to transform the orthogonal axis system to a cylindrical one, and it would be great if that could be defined in the GUI at some point. I need to add that keyword and re-run.

We are assessing the contact pressures and the radial stress in the socket, away from the socket edges. In ansys it’s straightforward to create a construction line and plot stresses along that line. How would the forum suggest doing similar with prepomax?

I tried to import a line element or axis in the step file (made using freecad) but that doesn’t seem to be permitted. I tried splitting a surface to get a radial edge, but that creates problems with meshing.

I can get a history plot along an edge (say the socket axial surface edge) so maybe I could split up the socket in to separate bodies in freecad, import them and make compound meshes from the individual body cells?

And maybe I haven’t found the option yet, but it would also be very useful to set fringe plot limits (for example in the attached image, show values between say 50-100MPa

All help always appreciated, and maybe I can contribute back in to the group.

(btw, my thanks go to the developers, I really like using this software - it’s simple and straightforward, mostly intuitive).

The contact pressure results are in line with both analytical and industry standard software results…

image
image1008×767 259 KB

2

Yes, that could work. PrePoMax can create compound parts. Or you could partition only the surface, avoiding the need for compound creation.

Plotting along a line is not currently supported but you could use ParaView for this purpose. I showed it in this tutorial: https://www.youtube.com/watch?v=wRcTD1zSbLU

I also showed there how to convert the results to cylindrical CSYS in ParaView.

This can be set by double-clicking on the legend and switching Min/max limit type to Manual.

1 Like
3

If I remember well you must convert the .frd CalculiX results to some Paraview compatible format such as vtk/vtu/rmed, and then use Paraview to made such kind of plot along a line.

For converting the results I use an external utility called Calmed8, I found it years ago in the old Salome forum and is the best by the moment to me after testing several ones.

If you need this utility send me a message, because I don´t believe that is available now as it was very old post. I know that sharing directly an executable and libraries is not very GNU/GPL or even secure, but is the best that I can offer :slight_smile:

1 Like
4

Thanks for the quick responses.

I figured paraview might come in to it. OK, I will look at the tutorial, do the conversion and report back when I’m done.

5

In the tutorial, I used FreeCAD for conversion to VTK but you can find more options here: Converting results to ParaView format

I mostly use the ccx2paraview tool.

2 Likes
6

I’ll try the ccx2para tool first then

1 Like
7

Ok, understand the fringe plot limits now…

image
image1015×746 253 KB

8

Hi, @stresser ! Can you give more details about this model? It’s always nice to check the performance of CalculiX and PrePoMax as pre/post processor with analytical results :smiley:

9

https://roymech.org/Useful_Tables/Mechanics/Cylinders.html

10

Hey, once I’m done with the paraview side of things I’ll come back and give some more details. I can’t really release some of what I’m doing, but I can give an idea of how well prepomax performs. And, as ANYS posted an analytical methodology we could do the comparison with hand calcs here.

11

Thanks Stresser. At the end of the link there is the tribology calculator to make it easier.

https://www.tribology-abc.com/calculators/e6_2.htm

12

I covered interference fits in my tutorials once: https://www.youtube.com/watch?v=sos-6ilPIZc

I can recommend Shigley’s Mechanical Engineering Design for hand calcs.

1 Like