[stephan]

Dear avrachan,

as far as I know, there is no issue with HLBM and the BGK dynamics are validated.
The problems you are facing are very likely due to the choice of parameters.

1) I would suggest to paste your code here, if you like.
2) Please try to reproduce only one configuration from one reference at a time.
3) Then follow the book of Krüger et al. to choose your parameters. I would propose to choose a real physical characteristic length (the side length of your problem domain for example). Then choose the grid spacing and the viscosity and the real physical flow speed (or the Reynolds number alternatively). Then set the lattice velocity to be smaller then 0.1. This will give you the relaxation time and the time step size delta t.
4) Double-check the settings and the results with respect to this one particular physical setting until the solver converges.
5) Please print the converter of this particular setting again.

BR
Stephan

[stephan]

Dear avrachan,

can we fix a single set of physical parameters you are trying to simulate?
Like, a set of fixed Re, volume fraction, L.
Then choose some stable set of discrete parameters (for example with the help of Krueger et al. book, there’s a section on choosing parameters).
Then simulate that for a range of resolutions in diffusive scaling (Ma->0, where dt scales as dx^2).

Could you please try that? Does it work then?

If that approximates a single configuration in your target plot, you can move on to the next one.

BR
Stephan

[stephan]

Dear avrachan,

thank you for your post.

– Could you please explain how the quantities in the linked plots are defined exactly?
– Which range of porosity / permeability would you like to model?
– What are the mesh sizes you consider?
– How is the set up of the domain? How large is it? How many spheres are in it? Is there an inflow and an outflow region? Which boundary conditions are you using? What is Reynolds, Mach, and relaxation time exactly?

In this dissertation (https://dx.doi.org/10.5445/IR/1000161726), Section “4.5 Homogenized Navier–Stokes equations” describes a test case based on laminar flow through an array of spheres.

BR
Stephan

BR
Stephan

[stephan]

Dear Bobbie,

thanks for your post.
We will release the User Guide 1.8 before the end of August – hopefully earlier.

BR
Stephan

[stephan]

Dear aaron,

thank you for your post.

I am not sure, if I understand this correctly.

From my point of view, there is no change in viscosity whatsoever due a specific forcing method.
This happens only, if you formulate the force to actually do this.
But that’s a matter of the force itself, and not of the method you use for forcing.
Please have a look in the Krüger et al. 2017 book.
There is a specific chapter that compares all the forcing schemes and explains the effects they have.

BR
Stephan

[stephan]

Dear sx lee,

thank you for your understanding.
We are working on it.
The user guide for 1.8 will be released during the year, not sooner than June.
Please note, that the earlier version for 1.7 does contain valuable information also.

BR
Stephan

• This reply was modified 11 months, 3 weeks ago by stephan.

[stephan]

Dear eulerboiler,

that is correct, we have a cumulant dynamics in OpenLB 1.8.
It is however limited to a specific lattice in three dimensions.
You could try yourself and implement a 2D version (if you have a reference with the specific equations to be coded).

If you need support for these kind of things, please consider visiting our yearly spring school (https://www.openlb.net/spring-school/).
Otherwise, you are welcome to share your progress in this forum.

BR
Stephan

[stephan]

Dear hhan,

sorry, it has been a while.
However, your question is interesting.

So, in our code, we make use of the model described here:
[1] https://doi.org/10.1017/jfm.2012.155
and here:
[2] https://doi.org/10.1016/j.jcp.2005.03.022

In the latter reference [2], Equation (15) has an additional sqrt(2) which should be deleted.
Nevertheless, in the same notation, let’s assume that bar(Q) := sqrt(2 sum Q_ij Q_ij).
Then we replace the Q_ij with PiNeq_ij computed in OpenLB and insert this bar(PiNeq) into Equation (15).
The sqrt(2) in our code hence resembles the prefactor of 2 when computing the magnitude of the strain rate tensor ||S|| = sqrt(2 sum S_ij S_ij) (which is a common thing to do in CFD, see [1], page 516).
All of this, also leads to same expressions as in the paper you referenced above (https://doi.org/https://doi.org/10.1016/j.camwa.2018.08.018).

Let me know if you disagree here.
Please also consider testing the tgv3d example with SmagorinskyBGKdynamics (once including the sqrt(2) and once without it in the EffectiveOmega computation) and post the results here in the forum.

Thanks again for posting and BR,
Stephan

[stephan]

Dear liu,

thank you for your post.
The best example to start would be examples/turbulence/tgv3d/.
I suggest to change the 3D functors to 2D, and remove all the lines you do not require.
The intialization should also be simplified, such that only defineU and defineRho is left.
The velocity field for u_0 is constructed as functor and then passed to defineU.
The pressure for p_0 (given for example in the book of Krüger et al. (2017)) is constructed as a functor, then transformed to density and passed to defineRho.

If you want to have specific support on creating your own applications with OpenLB, please consider taking part in our upcoming Spring School: https://www.openlb.net/spring-school-2025/

BR
Stephan

[stephan]

Dear sfraniatte,

we are currently working on this.
Similar features should be included in an upcoming release.
It is straight-forward to add a layer of boundary cells, where the relaxation time is increased.
Although this is the least effective approach, it should work well as a first attempt.

If you like to have more information or guided support for your specific implementation, please consider taking part in our next Spring School in May: https://www.openlb.net/spring-school-2025/

BR
Stephan

[stephan]

Dear Dan,

as you describe it, your application would have to at least use a two-component (e.g. Shan-Chen or Free energy) or free surface model coupled to a particle model (e.g. HLBM).

Please consider joining the Consortium (https://www.openlb.net/consortium/) or visiting our Spring School (https://www.openlb.net/spring-school-2025/) to get extended support for this.

BR
Stephan

[stephan]

Dear atanuchaudhury,

thank you for posting.
Please see my other replies to your questions (you have asked this question in another post before).

BR
Stephan

[stephan]

Dear atanuchaudhury,

thank you for posting.
To increase the Reynolds number and still get stable results, please consider the standard literature on LBM, e.g. the book by Krüger et al. (https://doi.org/10.1007/978-3-319-44649-3).
Basic techniques for this are choosing suitable discretization parameters with experimental stability bounds or using turbulence models as well as advanced collision operators.

We cannot offer full support on debugging your complete application.
You might want to consider taking part in our upcoming spring school for this.
More information is given here: https://www.openlb.net/spring-school-2025/

BR
Stephan

[stephan]

Dear atanuchaudhury,

thank you for posting.
Please have a look at our doxygen and user guide documentations to learn more about the correct usage of functors.

Unfortunately, we cannot offer full support on debugging your complete application.

You might want to consider taking part in our upcoming spring school for this.
More information is given here: https://www.openlb.net/spring-school-2025/

BR
Stephan

[stephan]

Dear Dirk,

thank you for posting.
Please also adapt everything else in your modified example to the 2-component case.
Then, the boundary conditions have to made consistent with this reduction.
I rather expect that you have to modify several other sections of the .cpp file to make this work properly. However, you can do this (as you already have started) with looking at the other examples with 2 components.

BR
Stephan

[Author]

Posts