What happens if I use RANS model(eg KEpsilon) on a laminar artery flow
 

What happens if I use RANS model(eg KEpsilon) on a laminar artery flow which is slightly transition to turbulent fow OR just stay laminar?

The artery geometry has an irregular shape where there is possibility of transitioning from laminar to turbulent for most cases in the stenoid(narrow) region , but for some of arteries flow remains laminar. What happens if I use KEpsilon for modeling all my 50 patients?

Please see the figure attached in order to see the geometry.


Thanks,
Farzad

You should see that k (turbulent kinetic energy) and epsilon (turbulent dissipation rate) remain small.

Dear Domenico Lahaye
for most cases k and epsilon are small and negligible. can I accept results from Kepsilon as my final results or what should I do?


Thanks,
Farzad

What happens if I use RANS model(eg KEpsilon) on a laminar artery flow
 

k-epsilon model could be problematic when k goes to zero, in that case k-omega model has better behavior and almost reduces to laminar flow.
Laminar to turbulent transition modeling is unsteady in its nature, therefore quite complicated to solve within RANS context, but (with some limitations) there are models that consider transition and eventually could be suitable for your problem setup.
https://en.wikipedia.org/wiki/Gamma-Re_Transition_Model

Dear Baker
I just used Gamma-Re Transition model and velocity was higher compare to the kEpsilon model(Please see attached figure). I just have one concern, what happens if fluid remain laminar and do not transition to turbulence? can I rely on Gamma-Re results?

Also, I am not interested in transition from laminar to turbulent. I just want to use correct model for the flow regime which I do not know if it is turbulent or laminar or in between. Also, I just want to use it as a steady state(which in this case I do not think it can capture transition at all).

Thanks,
Farzad

What happens if I use RANS model(eg KEpsilon) on a laminar artery flow
 

As you've already seen, choosing appropriate turbulence model is essential as the results vary considerably.
But the problem that you are addressing is not that trivial, namely already the averaging in the RANS context has the limitations considering the physical nature of laminar-turbulent transition.
The Gamma-Re model is based on the incorporated empirical results with the intention to cover the cases with laminar and turbulent flow and the transition between. If you do the research on the topic, you'll find enough information sources regarding validation and the model accuracy and it's limitations.

Thanks for your help. Actually, I cant do LES(transient) simulation due to high number of patients and it computational costs, but for one case, I can compare averaged LES(Transient) results with kEpsilon, kOmega and kOmegaSSTLM to see which one of them has closer results to LES.


Thanks,
Farzad

Dear Farzad,

Just wondering whether you managed to find out which solver is best for your application. and which of them gave you closer match with LES.

Thanks
Azi