Symmetry Constraints Question

[edwardWR65A]

Hello there,

I am doing an analysis on an assembly. I have 2 models, 1 is a ¼ scale assembly model and 1 is a ½ scale assembly model. Why is it that I get a high stress concentration (red circled) on a particular part (in this case: front beam pressing) in the ¼ model (Figure 1) but not on the ½ model (Figure 2)? Any help would be much appreciated.

Figure 1: ¼ model shows 1 part having stress concentration at a weld location and along an edge. I have approximated the stress at the ‘389.5’ MPa location using whats described in the Why do stresses keep going up when the mesh is refined in Nastran | Inventor Nastran | Autodesk Know..... Autodesk knowledge article. It brings it down to 200 MPa.

Figure 2: 1/2 model shows 1 part having significantly less stress in the middle near the fold(where the probe is) 46 MPa.

Originally I was expecting in theory the ¼ model to have a quicker simulation time and be just as accurate hence why I did that first. Are the differences in stresses due to a combination of ‘harsh’ symmetry constraints I have applied in the ¼ model and in general a sharp edge producing high stress? Or is there something that I have done wrong? 

I am running version of Inventor Nastran 2023. 23.0.0.23

I have tried to attach my assembly model via 'pack and go' or compressed folder to help show you what I mean but unfortunately its still greater than the 71 MB limit. Any other ideas on how to attach this? Otherwise below are my loads and constraints etc:

Full model looks like:

1/4 model looks like:

1/2 model looks like:

Kind regards,

Ed

[John_Holtz]

Hi Ed,

Most of the file size is caused by the mesh in the Nastran environment. To reduce the size, you can delete the mesh from the analysis (and all the analyses in the model if there is more than one). Hopefully that will reduce the size below the 71 MB limit. As long as someone uses the same version of Nastran that you are using, regenerating the mesh should give the same number of nodes and elements. In the Nastran environment, right-click on "Mesh Model" in the model tree and choose "Delete Mesh".

The answer to your question is the results of the 1/4 symmetry model will be the same as the 1/2 symmetry model if everything in the analysis is really the same (or approximately the same). The difference may be that the mesh "through the thickness" is smaller in the 1/4 symmetry model because there is an extra face that determines the mesh "through the thickness". In other words,

• Imagine the part is 10 mm thick and the mesh size is 5 mm. For the 1/2 symmetry model, the element size on the top and bottom faces is 5 mm, but you have no control over how long the solid elements are extending into the volume. They could be 5 mm (so giving 2 elements through the thickness), or they could be 10 mm (giving 1 element through the thickness).
• For the 1/4 symmetry model, the 5 mm mesh size on the symmetry plane gives 2 elements through the thickness at least for the solid elements close to the symmetry face. (Further from the symmetry face, the element size can become larger.)

However, something does look wrong in the 1/4 symmetry results. The unexpected high stress is not confined to the circled area but extends all along the "top" edge of the plate and extending to the right in Figure 1. The 1/2 symmetry model does not show any hints of that type of stress. Hopefully you will be able to attached the model after deleting the mesh.

John

[edwardWR65A]

Hello John,

Thank you for your helpful response. I will / have attached 2 assemblies. Yes, I thought the ¼ assembly model did not look correct.

In regard to your bullet point explanations, do you mean this in Figure 1? From memory, all parts were 10 mm thick with 5 mm element meshes on parts of interest. Parts not of interest had a general mesh size of 15 mm. Welds were 3mm mesh size.

Figure 1: ¼ assembly model. Multiple single elements identified.

Kind regards,

Ed