I am working on a 2D flow inside a diverging tube using Simpler algorithm. My problem is that my velocity converges very well, but the mass conservation is not very good. It seems to me that the velocity equations converge much faster than the pressure correction equation. Does anybody have any suggestions on how to solve this problem?

Thanks a lot,

yonghong

What do you mean by not very good? Do you mean it converges too slowly or it converges to a wrong solution such that the total mass is not conserved?

If it's converging too slowly. It depends on how you solve the pressure equation. If you solve it by simple Jacobian iteration, it could be rather slow.

One way to accelerate it is use so called red-black ordering. It's basically doing updating the unknowns first by odd grids and then by even grids. So that all the new values will be reused.

http://www.jics.utk.edu/PCUE/MOD9_IT/sld042.htm

or

http://www.jics.utk.edu/PCUE/MOD9_IT/index.htm

You should check if the mass balance is satisfied and the total mass is conserved. Maybe there is something wrong with the spefification of your boundary conditions. You should especially take care of the outflow boundaries.

I meant that the mass is not conserved in each grid point, though the overall mass balance between the inlet and outlet of the flow is OK. I tried straight duct flow with the same code, it worked perfectly. Any suggestions?

Thanks a lot

yonghong

Is the flow compressible or incompressible?

Is the flow viscous or invicid?

Is the solution symmetric if the tube is symmetric? If not, problem will be at boundaries. Put a reverse direction flow, is it the same as before?

Is the flow steady? If yes, can check the energy by Bernoulli Eqn. As you said velocity is okay, then pressure is okay, the mass conservation should be okay.

good luck