Working with convection near a no-slip boundary
I am looking to use the convection only to calculate some near wall terms. However, it appears to be rather unreliable.
For instance, I've set up a user defined scalar in steady flow on a well resolved grid with a good solution. My UDS value on the wall is zero (no-slip) and If I have no diffusion and set a source term S to be:
S = density(VxdVx/dx + Vy*dVx/dy)
I should return phi = Vx. In the bulk flow this is true however close to the wall I experience all sorts of gradients that will ruin my near wall analysis should I desire changing the source term.
Are there rules of UDS' I can use to severely reduce this error? Is something in particular occuring? Can I use different variables than those I've written in my expression for the source term?
Is there a specific way to mitigate this?
