Coordinates transmation/solving in phyical space
 

Hello,

I'm now trying to modify my cartesian FVM Navier Stokes code to solve flow around an airfoil. I'm aware that I can either solve in the physical space ie generate structured grids and solve using interpolation or I can do coordinate transformation on my existing equation.

Anyone care to comment which is a more preferred mtd, in terms of speed and ease of programming? Or it's simply a matter of choice?

Moreover, there is the possibility that the code will be extending to 3D. Btw, I'm using a fractional step mtd to solve the NS eqns.

Thanks!

Re: Coordinates transmation/solving in phyical spa
 

Solving in physical space is preferred since it is easier to deal with grid control for viscous computations which is particularly important not only for turbulent flow but also for high reynolds number laminar flow. Coordinate transformation requires grid metrics which not only makes coding that much more difficult but you also need to ensure that the transformed equations preserve uniform flow by satisfying the so-called Geometric Conservation Law (GCL).