I've written my own code program simulating 2D, incompressible and viscous flow about an airfoil. I'm using for that structured C-grid generation (far-field boundary is 15xcord). The problem is that my code program diverges after about 5000 itterations. Mister Ganesh who is an active and helpfull on this Site suggested to use non-reflecting BC at far-field instead of free stream BC. My question is: Are that type of BC can be applied to incompressible flow, and how can i implement non reflecting BC.

Most of the non reflecting BC are derived from the characteristics of the Euler equation and they usually try to absorb correctly the outgoing acoustic,vorticity and entropy wave and are derived for systems where the speed of sound is finite.

You can try implementing an absorbing buffer zone.A simple way of implementing it is as follows:

The equation would be

df/dt=RHS+sigma* f

where the factor sigma=sigma0*((x-xb)/D)^beta.sigma0 is a constant based on the grid ( I use sigma0=2/dx for most cases, beta=3 ).D is the thickness of the sponge layer.xb is the location where the sponge layer begins.A sponge layer of 10-20 points works for me.

--------------------------------------|------------|

xb<---D----->xn

I have not tested the buffer zone for incompressible systems and am not sure if they might help. =)