[lollipopcorn]

Thanks for your prompt reply.
Just an update. I removed “march=native” flag and now it works on M1.

[lollipopcorn]

Dear Max,

Thanks for your reply. When I say condensation I was actually meaning the precipitation of one fluid in the other fluid. The setup is that fluid 1 is being injected from the left end, displacing fluid 2. (rectangle geometry, constant pressure at right end, bounceback at top and bottom, BouncbackRho1=0.7, BouncebackRho2=0.3). When I increase G from 3 to 4, after the start of the simulation, though most of fluid 1 is still in the left region, I see some little bubbles of fluid 1 precipitation in fluid 2 at up and bottom walls in the middle and right region. it feels like part of fluid one dissolves in fluid 2 and then precipitates at the top and bottom boundary walls.

Why does this happen? especially, if its due to the partial miscibility of two fluids, then, why does it happen when G is 4 but doesn’t when G is 3. since G is a measurement of repulsive forces between fluids, shouldn’t the fluids be “more immiscible” when G is 4?

Thanks,
John

[lollipopcorn]

Thanks a lot for your reply. your reply solves my problem, except that it seems the first value pushed back is for the second lattce, i.e.

std::vector<T> rho0;
rho0.push_back( 1 ); //this rho is for second lattice
rho0.push_back( 1 ); //first lattice
…

I noticed that I can implement MRT, just wondering does openLB support MRT with multicomponents? if so, do you have any examples that I can study?

Thanks a lot for sharing this excellent library.

John

[lollipopcorn]

I found the problem: velocity can only be added if only the adjacent lattices have the material 1. If you want to simulate a situation where there are 2 flows, you need to switch the flow material number such that the material number adjacent to the INLET is 1.

[Author]

Posts