Suggestions for improving grid quality while creating hybrid grid
 

Dear all,

I have been trying to create Vertical Axis Wind Turbine (VWAT) 3D mesh using a hybrid meshing technique. I created structured mesh around the blade boundary layer with pretty much good quality. Then I tried to fill the remaining of the rotating domain with unstructured cells. In doing so, I created a block with help of outer unstructured rotating domain and structured domain around the blade. After creating the block, I tried the following to achieve a good quality mesh.
1. In size field, I altered decaying to 0.9 and observed the result. Also, I tried altering the background spacing. I applied grid influencer, with T-rex BC changed to 'Adapt'. Still the skewness in transition region between structured and unstructured grid didn't improved.
2. I tried creating T-Rex layer around the structured block then fill the other region with isotropic layer. Also, altered T-rex governing criteria like 'Growth rate, max layer, skewness criteria, isotropic seed layers'. That didn't work.
3. I tried altering aspect ratio of the pyramids formed on structured domain to 0.1 and 0.9. Both didn't work.


I also used connector to make a connection between inner block and outer domain and manually created domain and then block for different section changing the spacing of connector. Still, the skewness didn't improved.

I am not sure if understood your request here but I suggest to start running your CFD simulation with a course mesh and then after you got your result verified and validated, think which part of mesh needs to be improved in order to obtain better results. Physics to physics, geometry to geometry and solver to solver this might differ.

While creating a unstructured block on a structured domain ,I found that the structured domain need to have smooth transition of cells. Domain having high aspect ratio surface will create highly skewed pyramid thus generating low quality unstructured cells.
Thank you for your suggestions.

I made a structured mesh on blockMesh, OF. The blades are circular arcs and the mesh is parametrized: https://github.com/EdgarAMO/VAWT-structured-mesh