First of all, I would like to thank the team around the OpenLB-Projekt for their interesting work.

Here is my Problem:
I would like to simulate flows in microfluidic channels. The idea is to investigate the quality of mixing elements in laminar flow. I set up a 3D-Tmixer with the two inlets having the same velocity.
… ___________________________
.-> l…I<- velocity inlet (uniform)
.-> I___________…__________I<-
… I .I .I.I
pressure or velocity outlet

The setup is supposed to be symmetrical.
I used ConstRhoBGK and BGK-Algorithm for a single Component with bounce-back boundaries.

At low Reynoldsnumbers the incompressible flow is symmetric. At a certain point the fluid swaps to the other side und engulfment flow is prevailing (N. Kockmann, Transport Phenomena in Micro Process Engineering).
I am just getting no engulfment flow! Is it maybe in the nature of the method!
I tried the very everything thing. Changing Re, inlet flow ratio, geometry itself. And I am fairly sure about the units und boundary parameters.

Second question: How reliable is OpenLB on predicting pressure loss (BGK)?

Can anybody help me with this?



It is difficult to tell from your description where the problem could be. Could you be more specific about what is supposed to happen when the fluid “swaps”? What is an engulfment flow? Why is it expected to happen?

Two quick remarks though:

  • If the simulation setup is fully symmetric, the flow will always be symmetric, no matter what Reynolds number you use. In other numerical methods, an instability may occur, because the underlying grid or the numerical method is not fully symmetric. But in lattice Boltzmann, you need to manually break the symmetry by either adding a weak noise or building a slightly non-symmetric geometry.
  • Just to see what happens, I would suggest that you try BGKdynamics instead of ConstRhoBGKdynamics.

Thanks for the hints, Mr. Latt!

For the setup there is a good article from Lindken, Hussong, Esh, Westerweel:

“Experimental and numerical investigation of the symmetry and flow bifurcation in a T-shaped micromixer”

I changed my initial velocity field to be unsymmetric. The results I still have to analyse, but things look better.