[Eslam Reda]

Hello every body, I am facing a trouble with the pressure correction method in a simple flow in pipe problem. When I use very low values for density (1 kg/m3), the code converges. While when using heavy densities (1000 kg/m3) the program diverges. Please I need your help in this annoying trouble.

[RenardP]

Hello,

What kind of help do you need ? Do you have a specific question ? What solver do you use ?

RenardP

[Eslam Reda]

Hello, let's start:

1. I have a simple laminar water flow in a pipe problem. The inlet and outlet pressures are known (0.035 Pa and 0 Pa respectively), the inlet radial velocity is set to zero while the outlet one is floating and also the inlet and outlet axial velocities are floating.
2. I use an ordinary SIMPLE algorithm, Fortran program. I tried it as explicit and implicit giving the same problem.
3. The trouble is the divergence of the reults.
4. The strange thing is when I set the density to unity, rather than 1000 kg/m3 for water, the program converges and gives the correct results.

I need your help in this problem as it is considered as my start to the more difficult practical problems.

Thanks so much.
Eslam Reda.

[RenardP]

OK,

the SIMPLE algorithm is a classical one so I think you have implemented a correct version.

In your place I would check the grid and the implementation of the boundary conditions. Usually a divergent solution is caused by bad implementation of the boundary conditions.

Also if you can find an analytical solution for a simplified version of your tube problem you can test various parts of your implementation (e.g a Poiseuille flow in a duct).

Paul

[Eslam Reda]

Hello again,
Surely I checked the code for a simple Poiseuielle flow problem. The program diverges at high densities but gives the expected analytical results for low densities. I wish to get the data, Dx, Dy, under-relaxation factor and Dt, for a solved water flow in a pipe problem.