I have a model that is composed of multiple bodies. There is a conformal mesh at the boundary between bodies, so the bodies share nodes at the boundary. I noticed though that when I export the nodal strain results to Matlab, these shared nodes have multiple results for strain. How do I know which value to use? How do I better understand how to interpret these nodal results? Can I sum these results or average them somehow?
(Here is what the documentation says about this situation if it's any help: "For result contours that are scoped to more than one body, and that share nodes by more than one body, the export file will contain multiple result listings for each shared node. Furthermore, if the result type is a degree of freedom result, such as temperature and displacement results, then the result values for a given shared node are identical (and redundant). In addition, if the result type is an element nodal result, like stress or strain, then the result values for a given shared node can be different because the solver can compute different element nodal result values for a node shared by different bodies.")
Dear Joseph,
In my opinion the problem you are describing is called discontinuity in numerical analysis. If nodes from both sides of the contacts are merged together, the difference between the results should be very small. In your case, I suggest averaging the result.
Hamidreza Ahmadi Moghaddam Thank you. One extra detail: the nodes are shared since I opted to share coincident topology between the bodies when I was creating the model in SpaceClaim. So, there is certainly a discontinuity in the material properties. Would you still expect this same type of discontinuity that you are describing even with the bodies sharing nodes?
It is very difficult to suggest since you did not provide any meshed geometry, however, based on what I understood from your explanation, Yes I do.
- Badges
- Science topic