Coefficients of Launder-Sharma k-epsilon model
 

Can the coefficients used in LS k-epsilon model be varied?

For example, if I modify the Ce2 coefficient from 1.92 to 1.88, is it acceptable?

What I'm suffering now is that while I try to simulate a supersonic flat plate using ARSM based on LS k-epsilon model, the results agree well to experimental results if I set up the coefficient Ce2 as 1.88 rather than 1.92 which is originally proposed.

When I set up the Ce2 as 1.92, it over-predicts the friction velocity and production term is too large that provides higher turbulent kinetic energy level.

Is there anyone who suffers from similar problem?