[kennyboy]

I am currently doing a 2D simulation of a rotary positive displacement pump which uses the wankel mechanism (used by the wankel engine).

Some questions I have are

1. for dynamic mesh, why does it only work for triangular meshes? I read somewhere it is because of stability issues but I dont understand why though.

2. how do i know what mesh size is suitable? I was also told that mesh study is unfeasible for dynamic mesh cases. so how else can i do this?

3. how do i know what is the suitable time step for the dynamic mesh? for me i start off with a random time step which results in a negative cell volume generated during dynamic mesh movement. so i decided to keep lowering the time step until i stopped seeing negative volume.

4. how do i know which dynamic mesh type is suitable for me? smoothing, layering, remeshing. currently i am using smoothing and remeshing (according to a recommendaiton by my mentor but i am not sure why.)

[ramkr]

Hello! I can feel how this reply would be very less of importance or value to you now. Yet, seeing hardly any answers/reply in MOST of the dynamic mesh threads of the forums, I decided to still give information reg your doubts.

1. for dynamic mesh, why does it only work for triangular meshes? I read somewhere it is because of stability issues but I dont understand why though.

ANS. Generally, triangular meshes in 2D (tetrahedral meshes in 3D) are more quickly generated in the ANSYS meshing tool. And the deformation-stretching of cells too, which is a part of moving mesh type analysis, occurs best with unstructured cells, although there would be a higher trade off with computational time.

2. how do i know what mesh size is suitable? I was also told that mesh study is unfeasible for dynamic mesh cases. so how else can i do this?

ANS. Mesh size, type depends solely on type of problem you wish to solve. Eg in dynamic meshes there are cases where quad cells would suit (Piston-valve in 2D-opening, closing). Mesh study becomes unfeasible on due of computational time most often. Mostly it's best to refer to research articles on FSI/ moving mesh carried out using commercial CFD soft wares or you may go through you tube links demonstrating dynamic mesh using fluent.

3. how do i know what is the suitable time step for the dynamic mesh? for me i start off with a random time step which results in a negative cell volume generated during dynamic mesh movement. so i decided to keep lowering the time step until i stopped seeing negative volume.

ANS. time step size-determined by CFL condition, grid minimum cell size order, etc. For a complex problem of object moving large distances in the domain ,2D size could be upto 1e-04 seconds while 3D would not go beyond 1e-05 sec.

Yes, you are going in the right direction by checking convergence and mesh motion update. Additionally, ensure that the remeshing size interval is kept as 1 instead of default frequency of 5 in fluent dynamic mesh to avoid this type of error in 2D


4. how do i know which dynamic mesh type is suitable for me? smoothing, layering, remeshing. currently i am using smoothing and remeshing (according to a recommendaiton by my mentor but i am not sure why.)

ANS. Pls read through these sites>> https://wenku.baidu.com/view/81e3223...de2bd89ba.html
https://www.sharcnet.ca/Software/Ans...sh_update.html

Thank you
Hope it is helpful