Learning Forum

[joshuahutton]

Hello All,

I am having some issues with my large transient structural model that have to deal with singularities. I simplified the geometry significantly and now have lots of sharp edges. I could go back and add small radii, but then my day long solve times will just go up more and that's already a problem. A coworker suggested just filtering out these results by post processing and getting rid of the top say 0.1% of the stress values. I can manually export the results and do a histogram type of analysis in excel to find out what the 99.9% stress value is, but is there a way in Ansys to get this value?

I understand that this could lead to accuracy issues and there might be other ways to deal with the singularity, but for me, the time lost for the extra solve time is a greater issue and the real stresses I am concerned about are well away from these singularities.

Thanks

Joshua

[peteroznewman]

A better way is to create an output scoped to elements that exclude a layer of elements at the few singularity locations that coincide with the location of peak stress.
Many locations where there is a singularity may never have stress values that rise to a level of concern, so those don't need special treatment.
It could happen that the peak stress in the model occurs at a location that has no singularity, so filtering stress at that location would discard valid data.

[joshuahutton]

Thanks for the response. I was planning on comparing the maximum stress value with the "filtered" value and if they are drastically different then it would indicate a singularity. Yes, I know this is subjective and could lead to inaccuracies, but like I said, I have looked at the results and the real stress values are located well away from the singularities.
I am interested in your better way solution, but am unsure how to implement it. Can you give me some more detail or an example?

[peteroznewman]

If you have looked at your results and the peak stress is located well away from singularities, then there is no need for filtering.
If you had a model where the peak stress is at a singularity, then you need to create an output scoped to nodes instead of geometry. There are a sequence of steps you use in the worksheet for a named selection.

[peteroznewman]

Even cleaner looking result using Elements instead of Nodes.

[joshuahutton]

Thanks I'm having trouble duplicating your example. Could you send me your example?

[peteroznewman]

[joshuahutton]

Thanks