I am using a code that is a reynolds independent flow solver. As such it assumes laminar flow. However, in the real world separation does occur. I can not use a transient simulation. But I want to know areas in which separation is likely to occur.

I have a theory / method: by altering the boundary conditions (flow direction) by small amounts and seeing what resulting changes there are to the flow, if there are large changes in the solution in a particular area with a small change in boundary conditions, then this would represent an area with potential flow separation.

Has anyone got any thoughts on this, or tried similar tolerance analysis.

Cheers

Have a look at the book "Asymptotic theory of separated flow" by Sychev et al. (Cambridge University Press).

Strictly speaking you need viscosity for separation to occur (no matter how small) since it corresponds to a weak, rather than strong, solution to the inviscid equations of motion.