Order of numerical scheme
I did a 2nd order calculation of a laminar flow in CFX which followed a GCI study (refinement factor r=2, 3 grids, asymptotic range (1.001113)). I observed a rate of convergence of p=1.15.
I expected a value a bit less than 2 due to numerical effects but can a value close to 1 still be correct?
Thank you very much in advance!
Yes. All the different bits of the equations can converge at different rates - advection could be second order, bit diffusion could be first, etc etc. So if I find I am in the asymptotic range p which is close then I am happy.
|All times are GMT -4. The time now is 19:48.|