Hello everyone,
My aim is capture the stress fields around the crack-tip for building up the Paris curve. For this, I model cracks of discrete lengths (as measured in the experiments) for every simulation.
The experimental sample is cooled down and it is observed that the cooling process is crucial to determine the point of crack initiation. The residual stresses from the cooling process needs to be accounted before the mechanical loading step (of experiments). So I model the cooling step (without any crack) and transfer the stress fields to a model with crack before loading this model.
From the experiments, let's say the crack propagates from 'a1' mm to 'a2' mm in 'N1' cycles. Now my question is, should I model the cooling everytime (with a virtual closed crack of desired length) before doing a mechanical loading (where the crack can open) for every discrete crack length that I have to work on??
Is there a better way of performing this?
I think I am missing something here. I believe similar things would have been done by a lot of people. Looking forward to some ideas and suggestions.
Thanks for your time. Please contact me for any further information.
Regards,
Shree
P.S. I am working in Ansys on a bi-material interface crack if that is relevant.
Hi Mr. Bakir,
Thanks for the response firstly. I understand your point of view. I am not performing a coupled thermo-mechanical analysis for two reasons.
1. I can do a thermal cooling only once and can map stress results to many mechanical analyses without repeating thermal every time.
2. It is done this way, as in the future we would want to tweak the material properties of a downstream analysis alone without disturbing the thermal analysis.
I posed this question more from a physical perspective.
Nevertheless, thank you once again.
- Badges
- Science topic