MRT stability issues

Hello Everyone,

I’m using a D3Q19 MRT implementation on basic validation cases. In my understanding, MRT is more stable as smaller lb-viscosities, i.e. higher Re numbers. On the other hand, applying a very moderate Re number of about 200 with a high lb-velocity (e.g. v_lb~0.1) I observe stability issues, even though tau is about 0.53… Thus, I’m wondering if I misunderstood something, or if there is something wrong. I’m aware of the fact, that v_lb~0.1 is pretty close to the upper limit, but thought that MRT is more stable in such situations!? However, using half the lb-velocity, my simulation runs stable and shows pretty good results.

Please, could someone provide me with some more information or (even better) some theoretical discussion (references are welcome!),

Kind regards,

Hi Francois,

Can I ask which MRT model you implemented? Also, what validation cases you are looking at? And depending on your test case, what boundary conditions did you implement?
In my experience, MRT is tricky to set up because of all the free parameters, but in theory it can give you better results than BGK when set up properly. Have you observed better stability?