How to Deal With FEA Singularities in ANSYS Workbench

[@sdakshin]

thanks for this video. i have 2 questions in mind.1. in the stress vs. elements graph, what can we deduce about the mesh size where both graph's lines meet? will it be correct to call it an optimum mesh? or is there another way of looking at it?2. i am curious about the usage of parameters and how it increases solving time. i'd like to try it. say just for varying mesh sizes.

[@АлександрСоклаков-в7э]

one element away? it's bad idea. it is not mesh-independent result. And You can't compare two results: fillet and non-fillet. It's pointless.

[@JavierD]

I think he is showing the same CAD but one with the fillet. The one element away I agree, I have a singularity but the values increase gradually to the max point/element is not an asymptote

[@dennisjbonillalamboy646]

What would you suggest be done in the cases where getting rid of the sharp edge is not really an ideal option?

[@jevgenij879]

You are right. I have checked sharp corner with not fillet, with different mesh size(from 2mm to 0.25mm). It is really very mesh-dependent result, much larger stress one node away with smaller mesh size.

[@ibrahimkeremkoyuncuoglu6022]

I would like to ask a questions. If we don't have filled edge on our design, and local stress - as your sharp edge model does - un-realisticly calculated like over the material yield stress. We can estimate what is the approx. real value with using probe. But, when we use Lineriazed Equiblirum Stress (LES) for stress range at this specified high stress point than we gonna caltulate really huge Stress Range. And this is affect our fatigue life calculation according to ASME Sec.VIII Div.2 and EN 13445 (pressure equipment design codes) calculations. So how can we prevent to have this locally high stress (came from singularity) ? As i indicate before, our design has no filled edge, only sharp edge. Thank you for you support.

[@SimuTechGroup]

Hi Ibrahim, feel free to drop us a note at [e-mail] if you're interested in our consulting services and we'd be more than happy to help! Otherwise there is the Ansys student community to find posts/articles: studentcommunity.ansys.com/

[@frankrusso5569]

Should the bottom x axis of your mesh convergence plot be # of nodes?

[@viniciusvazfoz]

Good good

[@Prabh_Dhillon1993]

Niceeeeeee

[@Prabh_Dhillon1993]

It’s very descriptive and educational, this video