While solving compressible high speed flow over an airplane, after firing the job i am getting negative densities found in more than 100 cells after few iterations then run stopped. i have meshed using Proam(trimmed cell mesh) and prostar(cartesian mesh),Boundry conditons are input-inlet flow,output-pressure. i also have tried with reducing relaxation factor.

how to solve this problem......

Reduce the relaxation factor of pressure upto 0.05 and reduce other terms considerably and restart the analysis. Let the flow field to develop and later on you can keep on increase the U.R.Factors (pressure up to 0.7). Apply the boundary conditions as per the guidence given in user guide capter 6,7 (useful points may help you in some extent like extrude one layer of cells at the boundary when you are applying pressure B.C on the trimmed cells)for compressible flows.

Actually I would recommend to use another cfd-solver. Star-CD is vulnerable of negative densities. I had also enormous problems with neg. den. and the combination of inlet and pressure boundary seems to be accountable for it. A colleage of mine works with Fluent, and so we decided to generate a benchmark with pressure as inlet and a inlet boundary as outlet.

The same model (mesh) with same boundaries. Fluent without a problem , Star crashes after 30 iterations. And by way I am by far not a Star beginner, I know a lot of possible tricks to get a solution converged, but that's not the way it should be. It means always additional work and TIME.

To be honest I prefer Star and the Star environment but if you don't have time and you need a converged solution use Fluent. The Fluent solver is more robust.

Piet

Are you using densit.f (the density user subroutine)?

I have experienced getting negative densities, but managed to solve that by using densit.f to provide the density.

Have you tried turning on the environmental pressure and mean value options for the outlet pressure boundary? That might help.

set real constant 89 and 90 maybe useful.

What do constants 89 and 90 do?

RCONS 89 X: specify mininum temperature limit X (default 54.) RCONS 90 X: specify maximum temperature limit X (default 20000.)

for normal air the mininum temperature limit should be 293 I think. I have solved the negative problem by set RCONS 89=293 RCONS 90=4000

Thanks, I will try these out.

This is a common problem and has been dicussed in previous forums; try the search button above. Further, please read this one:

www.cfd-online.com/Forum/starcd_archive_2005.cgi/read/8106

It is always advisable to search; however, during my search I did not find the advice which Fengwei provided about constants 89 and 90. I don't think it hurts to ask again in case someone else has found a new solution to share.

Regards, Ant.