I have a hemispherical-shaped part in abaqus, and i want to define acoustic infinite elements on the round surface. To accomplish this i have taken the following steps:
- define a reference point at the center of the part
- define a skin on the round surface of the part
- create an acoustic infinite section (category: other; type: acoustic infinite). This section references a material with the density and bulk modulus properties defined.
- assign the section to the skin of the part.
- assign ACIN3D3 (infinite) elements to the skin in mesh module. The rest of the part was meshed with acoustic finite elements (solid, homogeneous section).
What makes me doubt that i am using the elements correctly is the fact that when i run a datacheck for the job, i get the warning that some of the elements are distorted, which i believe may cause the results to be inaccurate. I don't know how to make these elements undistorted. I attach an image: on the left, on white is the skin that i defined, and the green region was meshed with the finite acoustic elements; on the right the distorted infinite elements are shown. Do you know if i'm using the infinite elements correctly? why are the elements distorted and how can i make them undistorted?. I thank you very much for any help in advance.
Hi David
I don't think that the warning about distorted elements is related to the infinite elements. It looks like the distorted elements are within the solid part, not on the boundary, and you should be able get mesh statistics like Jacobian ratio and skewness in Abaqus (I use ANSYS myself, but it's not too different). You can try to improve element quality by refining the mesh with a smaller element size, but normally an acoustic mesh isn't too sensitive to mesh quality - as long as you have enough nodes pr wavelength you should be ok.
As far as I can tell, you are using the infinite elements correct, you can verify this by solving with and without the infinite elements - if the infinite elements work, the sound pressure in the model should be much lower with infinite elements than without.
Hope it helps,
Martin
Also, animate the sound pressure (not the level) and watch if the waves are travelling outwards.
I second that it your warning about distortion is related to the solid acoustic elements instead. Removing the infinite elements is also a good idea, but a very direct way to assess them is to run the model in a dynamic implicit or explicit step where you input some type of disturbance into the center of the domain. A propagating wave will approach the boundary of your domain, and if it reflects off the boundary then something is not ideal. It should not be reflected at all at the boundary, by definition of the infinite assumption.
Hi
It seems you apply the elements correctly.
I concur with the above.You can check your model also for things like consistent element normals on the boundary and on the inside if you have surface elements there, loose nodes etc.
If still in doubt - run something supersimple, e.g. a rectangular box. The infinite elements should not work as well for this geometry but should be able to do a decent job and make sure that your deck is ok.If the deck is ok, then the model is amiss.
/Claes
- Badges
- Science topic