Calculating dx+ and/or dz+
 

Good afternoon,

I am running various scale-adaptive simulations as part of a university research project. Naturally, such simulations entail a good deal of post/pre processing. I have found most of the germane quantities (such as y+, integral length scale, and SGS viscosity) relatively straightforward to post/pre process, either directly or with CFFs. With this said, I have yet to find a way to calculate dx+/dz+ in Fluent. Does anyone know of a way these can be calculated (using CFFs or otherwise)?

While I am fairly certain my grid spacing is adequate, I would like to show some contours of the grid in terms of viscous units.

I found only one other thread with a similar topic (https://www.cfd-online.com/Forums/ma...866-dx-dz.html). However, I think there was some confusion.

Any help is greatly appreciated!

dx+ and dz+
 

It will be straightforward to calculate and provided the mesh is structured. Otherwise, you need to find a way to define streamwise and transverse dimensions of the cells. This is not easy for unstructured meshes.

dx+ and dz+
 

Thanks for your reply, vinerm.

Unfortunately my geometry is not conducive to structured meshing. I assumed it would require a good bit of work to calculate, given my meshes are unstructured tetra. I just wanted to make sure I was not overlooking something in Fluent.

As a side note for anyone else interested, in ANSYS "Quick Guide to Setting up LES-type Simulations" the authors claim dx+/dz+ can be approximated using the CFF "sqrt(face-area-magnitude)*y-plus/cellwall-distance." I suspect this is only accurate to within an order of magnitude though.

dx+ and dz+
 

Thought I would follow-up by saying this: the CFF I posted above provides a fairly good approximation. I wrote the necessary code to calculate the viscous grid (for my particular case) and found the above CFF to be within 25% of the calculated values. The approximation provided by the CFF degrades as skewness increases.