Is it possible that modelling water flow between 2 parallel smooth plates with a laminar solution produces a non-linear correlation between the discharge and the pressure gradient. The problem is that unexpectedly my parallel plates model cannot be described by the cubic law, as there are quite high non linear pressure losses. i tried to refine the mesh but the results don't change. Can anybody help me, thanks a lot.

I think this could be helpful. I noticed that the longer the parallel plates the more the system approaches the cubic law. In fact, gradually increasing the length of my model from 32cm to 2m, gave a final linear relationship between the gradient and the discharge (with a very small error). Do you think this behaviour is caused by the presence of no flow boundaries on the upper lower and lateral surfaces? Thanks a lot

Boundary layers need time/distance to develop unless set correctly at the inlet BC.