LBM Algorithm

Lattice Boltzmann Method LBM: Algorithms
#1

Hi
Dear all,
i just start c++ coding of poiseuille flow and i have problem about the algorithm. i want to know which one come first in loop
calculating collision, macroscopic properties, calculating equilibrium distribution, BC or streaming. and where we should apply boundary condition. and if we have bounceback BC is this require we apply collision too for those node that have bounceback.

best regards

#2

Hello,

in my code I do it in this order:

  1. equilibrium
  2. collision
  3. propagation
  4. macroscopic properties
  5. back to 1.
    If you have obstacles (bounce-back), their treatment should be combined together with propagation, i.e., steps 1, 2, and 4 are usually not modified by boundary conditions.
    Since I use half-way bounce-back (physical walls are between fluid and wall nodes), there is no collision on the wall nodes. In other boundary condition implementation (wet boundaries), collision is also applied on the wall, at least at the wall nodes in contact with the fluid.
    Maybe you should read also this article.

Good luck,
Timm

#3

Hi
Dear Timm
thank you very much.
It was very helpful.
Good luck
sajad

#4

Hi, guys,
I want use the interpolation, but I can’t treat it correctly.
I applied second order interpolation according to paper of Lallemand and Li-Shi Luo in 2000.
Do everybody have any sample code or good reference to remark interpolation algorithm?
I wait for your suggestions.

#5

Dear Timm, Sajad,

Can you please update the link in your post?

Best
Alekhine

#6

Hi!

I’ve noticed there’s no activity in this topic for long time but still I’ll try before starting a new topic. It’ll be helpful to know about the source posted by Tim. Additionally, if someone reads this, do you know how to scale the velocity in LB to the real physical units?

Kind regards.
Thank you.

Timm Wrote:

Hello,

in my code I do it in this order:

  1. equilibrium
  2. collision
  3. propagation
  4. macroscopic properties
  5. back to 1.
    If you have obstacles (bounce-back), their
    treatment should be combined together with
    propagation, i.e., steps 1, 2, and 4 are usually
    not modified by boundary conditions.
    Since I use half-way bounce-back (physical walls
    are between fluid and wall nodes), there is no
    collision on the wall nodes. In other boundary
    condition implementation (wet boundaries),
    collision is also applied on the wall, at least at
    the wall nodes in contact with the fluid.
    Maybe you should read also
    th
    is article    .

Good luck,
Timm

#7

The link posted by Tim was moved here:
http://wiki.palabos.org/literature:latt_08

For LBM unit conversion, I’d advise reading the following:
http://lbmworkshop.com/wp-content/uploads/2011/08/2011-08-22_Edmonton_scaling.pdf

Nic