Results from a software for CFD that uses FEM
 

Hello,

I have developed a software using FEM (finite element method) for solving flows for incompressible fluids.
For validating the software I am trying to compare the results that I obtain with this software (for a lid driven cavity flow), with the reference (Ghia), that is the reference in this kind of domain (lid driven cavity).

I am enclosing the comparison for different reynolds numbers.

The meaning of the pictures is the following:
- vel30x30Re400.JPG. Simulation for a 30 x 30 grid and 400 as reynolds number (Horizontal velocity component in the vertical axis of simetry of the cavity).
- vel30x30uRe1000.jpg. Simulation for a 30 x 30 grid and 1000 as reynolds number (Horizontal velocity component in the vertical axis of simetry of the cavity).
- vel30x30vRe1000.jpg. Simulation for a 30 x 30 grid and 1000 as reynolds number (Vertical velocity component in the horizontal axis of simetry of the cavity).
- vel30x30uRe2500.jpg. Simulation for a 30 x 30 grid and 2500 as reynolds number (Horizontal velocity component in the vertical axis of simetry of the cavity).
- vel30x30vRe2500.jpg. Simulation for a 30 x 30 grid and 2500 as reynolds number (Vertical velocity component in the horizontal axis of simetry of the cavity).

I would like to know if the results I am enclosing could be regarded as valid.
I don't have analysed the impact of mesh size in the results. Maybe for more accuracy I should use a finer grid. I am thinking of using a 50x50 grid instead of a 30x30 grid.

Has anyone simulate this kind of domain with a such finner grid?

Thanks in advance for any value information / feedback you can give me.

Best regards,
Hector.