Hello,

on arbitrary unstructured grids the computation on cell face values can be complicated, expecially when using local grid refinement. CFD-Wiki page "Gradient Computation" describes two approaches, but the first ("cell-based methods") is only first order or arbitrary grids, and for the second one, it is not clear how to estimate accuracy. Another possibility is mentioned in Ferziger and Peric, and it involves auxiliary nodes and gradients at cell centers. But in order to calculate the gradient at cell centers, cell face values are needed (at least when using Gauss theorem: see again page "Gradient Computation"), which sends us in a logical loop. Anyone knows of methods for cell face values computation, which are second order accurate on arbitrary grids?

Thanks, Sergio

PS thanks to the authors of CFD-Wiki, it is a very useful resource.

hi,

u can check up on "A Second-Order Time-Accurate Finite Volume Method for Unsteady Incompressible Flow on Hybrid Unstructured Grids" by Dongjoo Kim and Haecheon Choi. jcp 162 p411-428.

hope it helps. if u managed to find similar papers, pls post them here. tks

this seems good too:

Improved linear interpolation practice for finite-volume schemes on complex grids

Int. J. Numer. Meth. Fluids 2002; 38:625â€"645