Hello Everyone
I am going to calculate the spurious velocity and physical velocity in the Shan-Chen model. So I simulated a static droplet in the vapor domain. The convergence is ok and I could develop the Palabos using other schemes like EDM. But when I want to calcuate spurious velocity or physical velocity through the computevelocity the spurious magnitude is very large. We expect it is in order of 0.01 or even smaller but the max of velocity calculated through the computevelocity() is about 0.3 which is 10 times greater than what we expected.
Do you know why the difference is so large? And I know the max velocity is LBM cannot exceed 0.1.
I also calculated force to determine uphy=u+F/2rho but it has also the same order as 0.3.
Hi,
You know the spurious velocity magnitude depends on the density ratio. I guess you run a high density ratio problem. Then, 0.3 for the Shan-Chen model is a normal spurious velocity magnitude, isn’t it?!
Did you see Peng Yuan and Laura Schaefer’s paper [Phys. Fluids 18, 042101 (2006)]? They have reported Us_max=0.08 for T/Tc~0.6… Then, I expect Us_max=0.3 for lower T/Tc (high density ratios)!
Cheers
I remember having the same problem, e.g. using psi = 1-exp(-rho) with moderate density ratios and ending up with astronomically large spurious velocities (also around 0.2 / 0.3). In my case the problem was that I miscalculated the velocity. When you simply calculate the velocity using the first order moment, you will end up with these large magnitudes. The computeVelocity-Function in Palabos, uses the function that is provided in the dynamics class. I dont know which dynamics class you are using, but it probably determines the velocity with said first order moment.
What is puzzling me though, when you calculate the force corrected velocity, i.e. u_phys = u + F/(rho*2), the velocities should be smaller. At least in my case this fixed the problem and I could reproduce the results from the Yuan & Schaefer Publication stated in the previous post. Could there be a problem with the way you calculate u_phys?
Thanks for your responses. I simulate a droplet in a vapor phase with the density ratio about 1000 and As you said for the Yuan and Shaefer method is not unusual. I had this spurious magnitude when I was using Yuan and Shaefer method but now I am using the EDM method proposed by Kupershtokh which is expected to have by far smaller spurious velocity magnitude but my results do not show a noticeable difference in their magnitude. Is there any body who has applied EDM method?
For kk answer, as you know the physical velocity does not have direct influence on the method. Generally, I do not have problem with the Yuan and Shaefer model because in their case the spurious velocity is large compared with the EDM method. And I am trying to utilize EDM to reduce the spurious velocity.