[syler3321]

Hi,
I am simulating a flow over 2d hills. Curvature pretty much comes into the equation here. There will be an extra strain rate dv/dx. I am using the k epsilon model to run my simulation and playing with the curvature correction term does not exactly give accurate results. The maximum value allowed is 1.25. I am thinking about incorporating the richardson number into the k epsilon model. but since the richardson number is a function of dv/dx and du/dy, i would need to have a Cmu that is not invariant. I understand that CFX does not have the realizable k epsilon model in its list of models so this is a bugger. i have tried inputting a CCL function into Cmu but it returns an error message saying that i cannot input variables when defining Cmu. Any suggestions how i can tackle the effects of curvature?

Cheers
Ralph