[lund8619]

I am working on a simulation that I would like to use viscoelastic properties for. I have simulation data from the model before applying the viscoelastic properties and also for after. After applying, the model is failing due to a negative element. The part that is confusing me is that the negative element is being seen long before my pressure wave reaches the viscoelastic material. I have attached images of the two pressure plots, which show a pressure reading that to me, is unexplained.

The only difference between the two simulation is the addition of a Prony Shear Relaxation property.

Values for the table are as follows:

Temperature = 27 [C]
Relative Moduli(i) = 5.5556
Relaxation Time(i) = 2.5e-5 [s]

These values were chosen based on the information found on page 6 of the following document

http://ansys.net/ansys/papers/nonlin...ng_viscoel.pdf

I have also linked this page to a previous issue that is still unsolved with this model. I would like to resolve both issues, but achieving a successful viscoelastic simulation is my priority
http://www.cfd-online.com/Forums/cfx...tructural.html

[ghorrocks]

Have you checked your convergence when it crashes? I bet the solver is going bezerk.

[lund8619]

Glenn,

I do not know how to check the convergence when the simulation crashes. I tried running two separate simulations using low and high convergence criteria and the results are different from one another, but surprisingly, the one with worse convergence criteria ran longer. I have included images of these two runs.

Is there any reason as to why the higher convergence model would crash earlier?

[ghorrocks]

This sounds like your model is highly numerically unstable so anything can happen. I would not read too much into how quickly things crash other than you need to improve numerical stability.

You can do that by:
* Smaller time steps
* Double precision numerics
* Better mesh quality (this one is the usual problem, you can always do a better mesh)
* better initial condition