Hi everyone,

I've written a NS solver based on the projection mtd/colocated FVM. I've run the cavity test for Re=100 and 400 and they give good results (max 5% error) compared with Ghia. My grid size is 128X128 uniform, same as Ghia. However at Re=1000, my u,v values increases t. 1.3 and beyond as time step increases. I know it's impossible for u,v to go beyond 1.0.

I'm trying to figure out what I should change in order to get the correct ans as Re increases.

One is to use non-uniform grid, with more concentrations near the wall. Is that correct?

Is it necessary to increase the grid size as Re increases? However Ghia uses only 128X128 uniform grids and he can do up to Re=10000

If grid size is the same, is the necessary to decrease the time step size if Re increases?

Thanks in advance!

Yes, you need to decrease time step as Re increases by CFL condition.

also to be precise, using the same CFL number while you increase your time step

Dear sir,

it is very recommended to use stretched grid for cavity problem for Re>=400. but u may try CFL number=1 have a good luck