hi friends i am trying to write CFD programme for heat exchanger with porosity approch,for momentum equation i am using SIMPLE method . when i use density of fluid 1.0 there is no problem for convergence .When i change density of fluid to 1000 for water my pressure correction equation is not converging . i would like to know what should be the convergence criteria for pressure correction . thank you

look van and raithby 1984 paper, they explained about convergence criteria. or else give some 50 iterations for each global iteration, this may help you

the SIMPLE alorithm solve the pressure volocity problem in a iteration way . in each iteration, you need not to have a fully convergent results, you only need to iterate for several times, for example 5 iteration. The convergence criteria for the outer iteration is that the velocity fields must satisfy the divergence conditino. Regards

Hi,

It's right that you don't need a fully convergence for every (outer) iteration, but it can enhance behaviour dramatically if you solve pressure correction more accuratly than the momentum and other scalar equation. For example run (inner) iteration for about reducing residual to one order of magnitude, pressure correction for about two orders. You can get trouble when using bad iterative solvers for the pressure correction, choose cgstab-like algorithms...

Hope this helps...

thanks all. i appreciate your help. i would like to know one small thing if somebody has observed it how the convergence time vary with change of density ?

the pressure correction equation is now a convection diffusion equation instead of a pure poisson equation in the case of incompressible flow, so it should converge better within inner iterations... see the paper: a collocated finite volume method for predicting flow at all speed, lilek and peric, journal of numerical methods in fluids...

Hi,

if you've only changed the density and not the velocity, grid, ... you have a much higher reynolds-number. This may cause your convergence problem. E.g. if you do a steady calculation and the flow gets unsteady due to higher Re, convergence problems may arise...

Regards Michael

Please inform me about the following: When we use the Simple algorithm for flow calculations, do we enter the inlet boundary conditions as velocity, static pressure, or total pressure?

Sincerely, Kamal