[KimBest]

Hi,

I'm trying to simulate the effect of varying inlet velocity on a single 2D sinusoidal hill in an neutral atmospheric boundary layer. Everything runs without errors, however when I vary the inlet velocity the flow field is the same for all simulations. By this I mean, the velocities do scale as they should, but there is no non-linearity due to the presence of the hill. I am using the turbineSiting tutorial as a template with all the boundary conditions unchanged (I intent to also simulate the presence of actuator disks, but the problem is the same with or without them) and the same k-epsilon turbulence model.
The mesh is scaled down so the hill has a height of 0.04 m (my reference height for u*). If I run the simulations for very low speeds (<12 m/s), there is no separation across the hill. Higher than this means the number of timesteps diverges, but then the case converges again when the inlet velocity is around 25 m/s and above. In both cases where I have convergence, the flow field looks exactly the same when I scale to each ones data range in paraView.

I have added a layer to the mesh so that the first cell centre height (=0.84mm) is greater than Ks (=0.08mm, where z0 is 4e-6m). I think that the problem may lie here however I'm pretty sure I have done this right. Any help would be great as I'm runnning out of time on this project and have tried everything!
I have attached all the relevant files for running my case in how it differs from the turbineSiting tutorial.

Thanks,


Kim