Learning Forum

[shaheen wahab]

Hello Everyone,

I have been trying to simulate the flow past a cylindrical pipe in a rectangular channel (3D). I could generate the desired inflation layers in the side walls and also on the pipe. But it seems there is some problem with the mesh as I am not able to get the actual result. I get some results (contours) which are visible wrong. What could be done to improve the mesh quality?

The following are my mesh settings:

1. Max face size = 0.25 m

2. Target skewness = 0.7 (I had to do this as the quality was bad )

3. Inflation layers -> Sidewalls and in the boundary of the pipe. (Both the inflation layers from the pipe and the sidewalls merge and looks somewhat like an 'L' shaped)

4. Face sizing (Sphere of influence) = 4 m radius

Thank you so much 

With regards

Shaheen  

This what the two Inflation layers look like (L shaped)

[Rob]

I'd probably look at solving this in 2d as there won't be much (any) flow in the z direction unless you're really interested in the end effects.  

If you do want 3d then read up on swept mesh and inflation: it's not set in the same way as an unstructured mesh. I'd also consider splitting the domain so you can concentrate mesh in the cylinder & wake region. Bodies of influence can cause some pretty big jumps in cell size so need to be used with some care. 

[shaheen wahab]

Thank you so much for your reply. I will try doing it in 2D and check for the things that you mentioned

With regards

Shaheen 

[shaheen wahab]

Hi,

Here is the solution.

1. I used Sphere of influence around the pile (0.1 m element size), inflation layers at the channel and the pile (which looked like an L, previous description)

2. Used a growth factor of 1.05 and max face size of 5 m. 

3. Target skewness = 0.7 (I had to do this as the quality was bad )

The mesh quality was found to be good. 

The simulation went well as well. 

Thank you 

With regards

Shaheen