Dear Sir,

Thank you for the reply. Would you be kind enough to share usage of structural hot spot method for standard FEA case instead for weld simulation?

Thanks in Advance.

Dear Sir,

I also tried using small fillets in the sharp corners, as you said it also does not solve the issue. It only leta say stabilize the maximum stress when element size is reduced.

Secondly I've tried using elasto-plastic material, by using bi-linear kinematic hardening with tangent modulus of zero, but even then the maximum stress did not converge when element size is reduced. Also stress concentration in the corner was still visible.

My Intention is to do fatigue simulation, so acquiring a maximum stress is critical to me to my next step. So only option is to think that the part fails in static stage, Or to avoid the stress concentration and interpolate the stress similar to stress hot spot method and for this I don't have solid evidence to support my argument,

I'm really stuck on this stage because I do not have a solid reason to move forward.

Any advice would be highly appreciated!

Thnaks.

Best Regards,

Rashiga

Hi Rashiga,

How did you induce interference fit to your model?

My two cents worth is to apply prescribed displacement of 0.01 mm to the contact geometry. And examines the load flow from your results to identify if forces / stresses are being transferred from one component to the next as expected since your analysis is dealing with two distinct materials.

All the best.

piong

Dear Sir,

I only have used ANSYS workbench up to now, so I still learning how to integrate APDL snipets to ANSYS mechanical.

I understand what you meant by there's no point in refining the mesh since the stress concentration will reach up to yield stress and stop from further increasing.

Up to now I have used ncode to obtain fatigue results. I will try to see if there's any equations available for notch sensitivity for similar cases.

Also I've noticed another weird phenomena when the mesh size is reduced near the contact surface. To get a better idea on fatigue life I have also used to analyse the "signed von mises stress"(elemental) of the components. These results were processed using "Hyperview" software.(please check the image attached)

At the point of the stress concentration highly negative stress is visible which is obvious since it is due to the compressive stress from the interference fit. But right next to the highly negative stress there's highly positive stress. And there's no visible transition from the highly positive stress to highly negative stress.

This is also one phenomena I cannot explain.

Dear Madam,

The prescribed interference is applied to the geometry directly to the contact surface will simulate it effect. I checked the contact pressure from the contact surfaces but the contact pressure was highly concentrated to the corner. It was similar to hertz contact stress.

Dear SIr,

Would like to apologize for my ignorance. I was able to run the the APDL snipet you have provided.

With the elasto plastic material there were no stress concentration but most of the area of the component were under yield stress.

Regarding Signed Von Mises stress,

I wasn't able to find the what exact method is used signing (+/-) . I found the following description from a research paper.(Please find the attached image) https://ac.els-cdn.com/S2452321617301373/1-s2.0-S2452321617301373-main.pdf?_tid=4c5cbf41-758c-464b-a324-1890c4fb780d&acdnat=1532877933_01f9d827a2cc4a959ebef644939538a7

Can't we use the sign of von mises stress to clarify a stress as tensile or compressive stress? Since the sign will determine by the direction which principle stresses occur. So if we use the same reasoning can there be highly tensile and compressive stress in same surface without any gradual transition?

Finally I would like to thank you for taking your time for answering my questions, Sir!

Dear Sir,

Thank you for the detailed explanation.

In Hyperview the sign is obtained by the sign of the largest absolute principle stress.

I perfectly understand in a mathematical stand point it possible to have stress jumps. What bothered me was how it can happen in a practical stand point.

Regarding the stress concentration simulation. I've used the APDL snippet with elasto plastic material properties. But for one component the maximum stress was higher than yield stress. I'm trying to figure out what went wrong there.

Best Regards,

Rashiga

Dear Claudio Pedrazzi

Thank you for taking your time and effort to answer the questions I had. Now I understand how to tackle this sort of issue in FEA.

Thank you.

Best Regards,

Rashiga

Dear Sir,

Thanks again for the explanation.

Thank you.

Best Regards,

Rashiga