how to define a moving wall boundary condition in MRT #367

ztdepztdep · 2021-08-31T02:20:48Z

ztdepztdep
Aug 31, 2021

I want to compute the lid-driven cavity problem, but could you please tell me how to define a moving wall boundary condition.

llaniewski · 2021-08-31T03:22:33Z

llaniewski
Aug 31, 2021
Maintainer

Quick search for "MovingWall" provides a nice example in one of the multiphase models:

TCLB/models/multiphase/d2q9_pf_velocity/Dynamics.c.Rt

Lines 1365 to 1378 in 9e1ceb1

    
           CudaDeviceFunction void MovingWall_S_Dirichlet(){ 
        
           	// The bounce-back formula for a Dirichlet boundary condition with a prescribed wall velocity 
        
           	// eq 5.26 p180 from The Lattice Boltzmann Method: Principles and Practice 
        
           	// T. Krüger, H. Kusumaatmaja, A. Kuzmin, O. Shardt, G. Silva, E.M. Viggen 
        
           	// TODO: IMPLEMENT ACTUAL BOUNDARY CONDITION 
        
           	g[2] = g[4]; 
        
           	g[5] = g[7] + VelocityX/6.0; 
        
           	g[6] = g[8] - VelocityX/6.0; 
        
           	h[2] = h[4]; 
        
           	h[5] = h[7]; 
        
           	h[6] = h[8]; 
        
           }

In this case g[...] are standard LBM densities, and the S in the function name stands for "South", so bottom wall. From this you should be able to construct the moving wall condition you need.

3 replies

ztdepztdep Aug 31, 2021
Author

hello sir:
I have tried, it always gives inf value of density and velocity. I can't figure it out where is the problem. I am wondering why this simple boundary condiction not be included in the standard d2q9 solver.

ztdepztdep Sep 1, 2021
Author

`CudaDeviceFunction void MovingWall_N_Dirichlet(){

f[4] = f[2];
f[7] = f[5] - WallX/6.0;
f[8] = f[6] + WallX/6.0;

}

I added this piece of code into the code. But it cann't work. hope you can help me out of this.
`

llaniewski Sep 1, 2021
Maintainer

@ztdepztdep I recommend doing the tutorial on model development (https://docs.tclb.io/tutorials/model-development/1.-finite-difference-wave-equation/) to understand how the note-types work. You'll learn how to add your new node types to Run, etc.

ztdepztdep · 2021-08-31T04:09:04Z

ztdepztdep
Aug 31, 2021
Author

`

   <?xml version="1.0"?>
   <CLBConfig version="2.0" output="output/" permissive="true">

<Geometry nx="120" ny="120">
	<MRT>
		<Box/>
	</MRT>
	<MovingWall_N mask="ALL">
		<Box dy="-1"/>
	</MovingWall>

	<Wall mask="ALL">
  		<Box nx="1"/>
		<Box dx="-1"/>
		<Box ny="1"/>
    </Wall>
    
</Geometry>

<Model>
	<Param name="VelocityX" value="1.0"/>
	<Param name="Viscosity" value="0.12"/>
</Model>
<VTK Iterations="5000" what="U,Rho"/>
<Log Iterations="5000" name="log.txt"/>
<Solve Iterations="100000"/>

` I added movingwall NodeType, and compiled it sucessfully. but when i run the xml file, it feed back the error about : error ! Error while parsing drivencavity-2D.xml: Start-end tags mismatch

5 replies

TravisMitchell Aug 31, 2021

This is likely because you have "Start-end tags mismatch" for the MovingWall_N element in your xml.

ztdepztdep Aug 31, 2021
Author

<MovingWall_N mask="ALL">	
<Box dy="-1"/>	
</MovingWall>

the north element is the last row of box ,as i have defined in the xml file.
i cann't figure it out what is the problem

TravisMitchell Aug 31, 2021

You have <MovingWall_N> closed with </ MovingWall> ... These tags do not match.

ztdepztdep Aug 31, 2021
Author

Great! Thank you. It works now. I will upload the computational results once I have completed it.

ztdepztdep Aug 31, 2021
Author

no, I have tried many parameters combinations, it cann't give any meaningful results. full of "inf".

how to define a moving wall boundary condition in MRT #367

Uh oh!

ztdepztdep Aug 31, 2021

Replies: 2 comments · 8 replies

Uh oh!

llaniewski Aug 31, 2021 Maintainer

Uh oh!

ztdepztdep Aug 31, 2021 Author

Uh oh!

ztdepztdep Sep 1, 2021 Author

Uh oh!

llaniewski Sep 1, 2021 Maintainer

Uh oh!

Uh oh!

ztdepztdep Aug 31, 2021 Author

Uh oh!

TravisMitchell Aug 31, 2021

Uh oh!

Uh oh!

ztdepztdep Aug 31, 2021 Author

Uh oh!

Uh oh!

TravisMitchell Aug 31, 2021

Uh oh!

ztdepztdep Aug 31, 2021 Author

Uh oh!

ztdepztdep Aug 31, 2021 Author

ztdepztdep
Aug 31, 2021

Replies: 2 comments 8 replies

llaniewski
Aug 31, 2021
Maintainer

ztdepztdep Aug 31, 2021
Author

ztdepztdep Sep 1, 2021
Author

llaniewski Sep 1, 2021
Maintainer

ztdepztdep
Aug 31, 2021
Author

ztdepztdep Aug 31, 2021
Author

ztdepztdep Aug 31, 2021
Author

ztdepztdep Aug 31, 2021
Author