Hello all,
I am new to spectral element solvers, and I am currently working on NEKO to analyse the capabilities of the code and exploit the GPU computation capabilities as well. As a part of this I am trying to analyse a thick airfoil with sharp trailing edge and at a Re 75000. I have the following questions below:
1. What are the mesh quality metrics that we to look for in the spectral element code? What is the aspect ratio that I need to be targeting near the walls and for the whole field? If we aim for a high resolution near the walls it is difficult to maintain the aspect ratio for the far-field region.
2. How to decide on the timestep for the simulation? Is it estimated based on the CFL number to be maintained and the velocity of the flow, polynomial order? If that is the case is CFL around 0.5 good enough?
3. I have constraints on the number of cells as I have only 4GPU's I can utilise and as there are no turbulence models available, I need to know the size of the domain that I can use especially in the z direction. Since its a turbulent case can I use a single cell in Z and apply periodic conditions or since turbulence is 3D I will have to resolve enough in Z direction as well?
I designed a rectangular configuration in gmsh as cells in C grid configuration are not yet defined in NEKO. Do let me know if you need any other info.

Thanks
Pranay