Extremely High Stress Issue and Interpretation

[DanielFET]

We are designing a flow check valve rated for 7500 psi and using Solidworks Simulation to analyze the check valve poppet stress distribution. Please see image-01 for the section cut of the valve assembly.

After running the static FEA with non-penetration contact set between the poppet head and the casing shoulder, I'm getting really high stresses around the two contact points shown on image-02.

We did the hand calculation based on formula: Contact Stress = Contact Force / Contact Area, and we got 64.4 KSI contact stress which is about 1/2 of the maximum number the FEA gave us.

We don't know if the high stresses are singularities or not. We have put a rounded corners around the high stress location but no luck, the stress is still very high.

My question is how to interoperate the stress result?

Does the material yield or not yield?

How to make the conclusion persuasive? I don't want to say "the red area is singularity and next to it is the green, hence ok. .."

Should I trust the result from the FEA analysis?

Are there anything else we can do to get a meaningful result?

Image -01 click this Link
Image -02 click this Link

 

[MartinShane]

I would suspect that this is a singularity caused by the mesh at the intersection of the two bodies. Rounded corner or not, the contact point between the two would appear to be sharp. (see sketch)

What you probably need to do is a Convergence check - comparing stress results obtained with progressively finer meshing. If there is a singularity the stress result will increase to infinity, if not the results will eventually start to level out. Simulation has some capability to do this automatically in static tests with solids - Look at the h-adaptive and p-adaptive solutions and convergence graphs - otherwise you can manually refine the mesh and plot your resulting max stress to check.

 

[corus]

It appears to contact at two positions which suggests that the contact is behaving like the stress concentrations around shrink fit cylinders. There are ways round this, such as using a shoulder fillet, or notches, to reduce the stress concentration by effectively reducing the local stiffness there. The assessment of these stresses is against fatigue damage, but if yours is yield stress then you could assess what the effect this would have if a small area was left with some plastic defomration and how this would affect subsequent loads. Perhaps it may yield (no pun) rounded edges or the notch you want?