Hello everyone, I’m interrested in using FEM in narrow regions and use BEM for the radiation part to compute the pressure at very far field away from the narrow region. I would like to know if the FEM computed the pressure and the velocity at the near filed boundary, can I use them as input to the BEM integral equation (BIE) to compute a very far field pressure?
Does it means that the Boundary Integral Equation served only as post-processing trough the BEM equation?
If I suppose that BEM is just a post-processing; do I have any restriction regarding the high frequency calculation?
Best regards, Kossivi
Dear Kossivi Gokpi
I suppose you have to implement an iterative procedure to couple the two methods. In principle, you use a solution of FEM as an input of BEM.
Maybe one iteration is sufficient, but you can check.
Regards
If the boundary conditions of the FE problem in the narrow region are completely known, you can first solve the problem in the narrow region. After that, you can solve the problem in the infinite domain using the BEM. It is incorrect to consider the BEM as a postprocessing tool. Similar to FEM, the BEM is an analysis tool. The BEM is very suitable for linear problems with infinite domains.
If the boundary conditions of the narrow region are not completely known, you should solve a global system of equations including FEM equations, BEM equations, and interface equations. This is not straightforward.
It is perfectly possible to use different techniques that apply to different regions of a solution domain and obtain a converged solution. For example, if you have an area of the domain where you have code that tackles the problem well, and another area where the code requirements are different (either because the math or physics are different), you can connect the two domains in an iteration and have the solution converge. I did this many years ago at NASA with good results.
You can pose the problem split into adjacent or overlapping domains, but the communication across boundaries must be dealt with carefully. This a powerful idea for solving complex problems.
Hello M. M. R. Hematiyan,
Many thanks for your reply.
What I'm trying to deal with is an acoustic fluid-structure problem with FEM. But do to the fact that at far field the domain with be too large thus very time consuming, I thought to solve the FEM problem with my FEM code to for example 10 meters far from the structure and then I know the pressure and the normal velocity at the infinite boundary at 10m. So at a second level I used the Boundary Integral Equation (BIE) to fill in the pressure and the normal velocity on the boundary to compute the pressure at 25meters or 100 m for exemple.
Is that way is correct?
With regards.
Kossivi
Your approach is correct; however, please note that you should solve another problem using the BEM in the other domain (the infinite domain outside 10 m).
Hello M. R. Hematiyan
Many Thanks for your reply.
So it means that the pressure and velocity become the boundary condition at the surface of the 10 m radius?
Thanks.
Dear Kossivi, I suggest you to contact my colleague Claudio Testa from CNR-INM who is an expert in this kind of problems. His email address is the following [email protected].
Best regards,
Luca
Hello Luca Greco
Many thanks for your reply. I will contact him.
With regards,
Kossivi
- Badges
- Science topic
- Similar topics
- Engineering
- Computer Engineering
- Computer