Discussion Forum

I have now attached the .mph file that I have been working on.

Please can someone explain how I may not have the zero flux boundary condition on the hexagon's 6 edges?

Thanks,
Kartiek

I don't have access to COMSOL right now, but I think it would help if you defined a material "air" (with the acoustic properties of air) and then filled the holes with that.

Hi Brian,

Thank you for your suggestion but unfortunately it doesn't help me. I tried changing the geometry to defining the a hexagon with a hole with air and the rest with the material. Unfortunately, now I cannot even implement the zero flux boundary condition on the hole.

In any case, I should be able to find a solution of a pde on the given shape without thinking of any physics at all and applying the boundary conditions that I require.

I find using the coefficient form the simplest because I can accordingly reduce my 3D problem to a 2D problem knowing the specifics of the problem and then defining the equations I need accordingly in the coefficient form.


The Periodic Boundary Conditions, if they are to work, should be able to override the zero flux boundary condition but this isn't happening.

Any more inputs would be greatly appreciated. I really need to get this simulation to work!

Thanks,
Kartiek.

In fact, the problem is completely general!

As per a model example given in the help in comsol 4.1, we make a plane rectangle and solve the simplest of pde equations on it after applying periodic boundary conditions on the rectangle's two opposing edges.

If you follow this through, it STILL shows that the zero flux boundary condition is applied on the 4 edges.

So there are 2 possibilities -

1) the GUI interface shows it wrong, but the internal workings of COMSOL assume that the zero flux boundary condition is overriden by the periodic boundary condition (not shown in the GUI which still shows that zero flux BC is implemented).

2) COMSOL cannot implement Periodic Boundary Conditions in Coefficient Form PDE interface.

Please inform me about which is the the true case?

Kartiek.

Update : It seems that it's more of a case of the former - case 1.

COMSOL does seem to implement the boundary conditions properly and the zero flux condition is overrided as it should be.

The reason I believe this to be true is because of the following quick test I did -

I wrote down the basic wave equation using coefficient form PDE interface. I took a rectangle, and applied periodic BC to opposing edges. Then I did an eigenfrequency study.

Then, I made one small change to the equation - I put the gamma term in the equation to a non zero value in the x direction. I did the eigenfrequency study again.

The results were the same. The reason to try a non zero gamma term is the following - It is a constant, and doesn't affect the wave equation. However, it affects the zero flux boundary condition. Since no changes were registered in the two studies, the zero flux boundary condition does seem to be overrided, but this is not confirmed by the GUI and is a source of confusion.

Can one of the forum moderators please confirm or deny if what I'm doing is correct? It would be a great clarification.
A file is again attached.

Regards,
Kartiek.

Update : It seems that it's more of a case of the former - case 1.

COMSOL does seem to implement the boundary conditions properly and the zero flux condition is overrided as it should be.

The reason I believe this to be true is because of the following quick test I did -

I wrote down the basic wave equation using coefficient form PDE interface. I took a rectangle, and applied periodic BC to opposing edges. Then I did an eigenfrequency study.

Then, I made one small change to the equation - I put the gamma term in the equation to a non zero value in the x direction. I did the eigenfrequency study again.

The results were the same. The reason to try a non zero gamma term is the following - It is a constant, and doesn't affect the wave equation. However, it affects the zero flux boundary condition. Since no changes were registered in the two studies, the zero flux boundary condition does seem to be overrided, but this is not confirmed by the GUI and is a source of confusion.

Can one of the forum moderators please confirm or deny if what I'm doing is correct? It would be a great clarification.
A file is again attached.

Regards,
Kartiek.

Hi Kartiek,

I also confronted the same problem as you did. When using the periodic condition in the Weak form PDE Module, I found the zero flux setting is defined by default, and cannot be cancelled when setting the periodic boundary condition.

I want to ask that whether you have solved such a problem.

Many thanks for your reply.

Best regards,
Jia

PS: I am using COMSOL 4.3.

Hi
I am going to solve Richard's Equation using COMSOl which is so similar to Heat Equation

I have a problem with Coefficient "c" because my coefficient is not constant is a function of dependent parameter what should I do with that?

u_t = K* u_xx
K is a function of u

How can I define it for COMSOL

Thanks a lot

Did you solve it? I am facing the same limitation
Thanks

Hi
Unfortunately, no the COMSOL can not define this constant.
try Wolfram Mathmatica it is so better