[Felix_Sp]

Hello again!

A question in regard to this thread:

Coupling different domains with different governing equations

I wondered how the workflow for such an approach is. First, one does an FNPF simulation with a specific domain length etc. Then one copies the .state files into another folder with a CFD module in it. Enabling H 10 4 in DiveMesh and B 92 61 in Reef3D seems to not be enough to actually simulate the problem.

So my questions are:

• Is the workflow correct?
• Is another print option needed to generate the input for coupling (like P 230 or P 240)?
• Are relaxations zones allowed in the FNPF file?
• Does the start of the grid of the CFD simulation have to correspond to the end of the grid of the FNPF simulation?
• Is the inflow taken from setting H 21 H 22 and H 23 in DiveMesh?
• Are I 230, 231, 232 and 241 needed for coupling?

Thanks for the help in advance!

[kamath]

The workflow is correct and should work.
You dont need to copy folders away. Just replace the input files in the same location, the output folders have unique names (_FNPF_ and _CFD_), so you won't loose any data.

After an FNPF run is completed (with P 40 2, P 41 1 for state file output every time step), run DM for your control.txt for the CFD run with H 10 4 in it. You can change xstart/end, no. of procs etc.
A new folder named REEF3D_CFD_HDC_Input should be created.
If you want to start at a particular time (generally, you want to) include I 241 xx (xx=time you want).
Optional use I 30 1 to start with the CFD domain initialised with the FNPF result at that chosen start time.

The CFD domain should begin somewhere before the end of the FNPF domain, as the end of the FNPF domain will be a beach isn't it?

H 21.... I 241.... are used to further optimise printout and start points if you'd want.

From your description, you have all the steps and I don't see where you are missing something, but hopefully my description here helps you figure that out or refine the question down the problem?

[Felix_Sp]

Thanks for your response!

So I tried the exact workflow again, but it seems like my CFD-boundary is getting no input.

Maybe it helps, if I further describe the situation:

I have got a long two-dimensional channel for the FNPF module and a short two-dimensional channel for the CFD-module (see attachment; these are just coarse examples). The two meshes overlap at a certain point (before dampening beach), where the input should be taken for the CFD-simulation.

Just applying H 10 4 in DiveMesh And B 92 61 in Reef3D, without further changes creates the HDC_Input, put no input seems to arrive at the boundary (the basic .ctrl files are also in the attachment). I tried the copy and paste method and applying it in the same folder.

Thanks again for your help in advance.

Edit: I found my mistake. Somehow I thought, that i do not have to declare the inlet as a Dirichlet boundary condition (maybe I thought is implicitly done with B 92 61). It works now!

[Hao666]

Hi Arun
I don't know why my account can't post at present, so I can only ask questions in this way. At present, my problem is that when I use the B270 function, my simulationis calculated in a strange way, and there is an error as shown in the following images, and the calculated force is null. But when I do not use the B270 function, the calculation can be carried out normally, there is no error, and the force is normal. I would like to ask how this problem can be solved. All simulation tests are carried out by REEF3D 21.09.

[Hao666]

Originally Posted by kamath

The workflow is correct and should work.
You dont need to copy folders away. Just replace the input files in the same location, the output folders have unique names (_FNPF_ and _CFD_), so you won't loose any data.

After an FNPF run is completed (with P 40 2, P 41 1 for state file output every time step), run DM for your control.txt for the CFD run with H 10 4 in it. You can change xstart/end, no. of procs etc.
A new folder named REEF3D_CFD_HDC_Input should be created.
If you want to start at a particular time (generally, you want to) include I 241 xx (xx=time you want).
Optional use I 30 1 to start with the CFD domain initialised with the FNPF result at that chosen start time.

The CFD domain should begin somewhere before the end of the FNPF domain, as the end of the FNPF domain will be a beach isn't it?

H 21.... I 241.... are used to further optimise printout and start points if you'd want.

From your description, you have all the steps and I don't see where you are missing something, but hopefully my description here helps you figure that out or refine the question down the problem?

Hi Arun
I don't know why my account can't post at present, so I can only ask questions in this way. At present, my problem is that when I use the B270 function, my simulationis calculated in a strange way, and there is an error as shown in the following images, and the calculated force is null. But when I do not use the B270 function, the calculation can be carried out normally, there is no error, and the force is normal. I would like to ask how this problem can be solved. All simulation tests are carried out by REEF3D 21.09 and I use S 1 to import stl. As if I can not use S1 and B270 at the same time.

[kamath]

Hi Felix,

I only just realised what you meant!

Yes, B 98 is the boundary condition for wave generation, which you will have to define. It is technically also possible to use B 98 2 (but I have not explored the consequences)


B 92 61 for coupling, as you can see is defined as a wave type. So therefore yes, you will need to use a B 98 call after the B 92 call, irrespective of the type of B 92 call.


@hao: I will start a new thread for you