so you are using some coupling similar to the immersed boundary method, correct?
What kind of external boundary conditions do you have? Is your system closed? Do you have periodic boundary condition? Pressure or velocity boundary conditions?
If you have a mass increase of 1% after only 50000 time steps and the reason is not the external boundary, then something is wrong.
I assume that the particle forces acting on the fluid enter the computations in the collision step?
When you know the force vector at a given fluid lattice node, you compute the lattice force (one for each population) for the collision step after some specific rule. In the famous Guo paper from 2002 you have a rich variety of such expressions. All of them obey mass conservation up to machine precision. I recommend that you take this paper and check your lattice force definition.
the LB algorithm is not interested where an external force comes from. I am also coupling particles to the fluid and I have never observed mass violation. So please check the list of possible error sources posted above.
Yes, the coupling method uses the concept of the immersed boundary.
The system is closed. In my study case, particles are being injected into a closed square column filled with fluid. All the walls are described by bounce-back (no-slip) boundary condition except the top wall which is described by the free-slip boundary condition.
Your assumption is correct. The particle forces are incorporated in the collision step. Additionally, I would like to note that I am using the LBM scheme of Eggels & Somers[sup]1[/sup].
I have downloaded the paper you suggested already. I will study it in detail and check the source of error. Thank you for your kind suggestion.
Regards,
Radompon
(1) Eggels & Somers. Numerical Simulation of Free Convective Flow using the lattice-Boltzmann Scheme, Int. J. Heat and Fluid Flow;16:1995
Hi everyone,
I have simulated flow past a circular cylinder at Reynolds<200 using a D2Q9 lattice(LBM).
Also, I want to do it Immersed boundary method-LBM,
I read article (Vortex Structures and Behavior of a Flow Past Two Rotating Circular Cylinders
Arranged Side-by-Side) but I have many problems, such as:
What is the r in Eq. (14)?
How could I calculate delta-function?
How should I transfer Eq.(11) from the Eulerian to the Lagrangian frame?
you should read a more fundamental article about the IBM first. At the beginning, it is a bit hard to understand everything, so you have to take some time for that. May you should start with the Peskin article which you find here.