[rg12g09]

Hi
I am reasonably new to Star and trying to simulate flow over an isolated ridge.
I have Tke and Tke disp as boundary coniditions at my inlet and initally the velocity profile produced is accurate to wind tunnnel experimental results I am trying to validate.
After 300 or so iterations the velocity at the crest no longer shows the expected velocity increase near to the surface.

When the TKE and TKE disp values are plotted I have a huge spike of Tke very close to the wall which gradually does decrease but not completly. This the causes a large change in Tke disp
I understand this is whats causing the incorrect velocity profile but am not sure how to prevent it occuring. Is this a failure of the wall functions ?
I have tried runnning the model using various turbulence models (currently using low y+ Ke realised) and many different y+ values varying from 0.019 to 2.

Thanks