Rhie - Chow Interpolation ?

Javier Goicochea

I have been writting a 1D CFD code for solving the Navier Stokes Eqs. I have found that the Rhie - Chow interpolation takes to many iteration to get a good convergence (i.e. Mass error < 1.0E-12). A have modified the relaxation factors for velocity and pressure, within the range of 0.8-0.5 for velocity and 0.5-0.1 for pressure, having almost the same results. I am using a SIMPLE algorithm, with a direct solver (1D - TDMA).

I actually test the code with a very simple problem (viscosity = 1.0E-3, density = 1000.0, dx = 100 * 50 CV, Area = 1.0E-4, every thing in IU, Re = 1000). Although, I have a good rate of convergence at begin of the iterations (e.g. Mass error = 1.0E-8, for the first 300 iterations), it takes a lot to get a Mass error < 1.0E-12. (Fortran 90 - Double Precision for any single real variable).

So I wonder if there is any kind of speed up factor that could be used with the Rhie - Chow interpolation, for this purpose ?

Thanks in advance, Goicox (goicox@hotmail.com)