I'm using CFX5.5.1. I have a problem concerned with simulating a flow between flat plates (Poiseuille flow) that has an exact solution. The flow is driven by pressure difference and the velocity profile is supposed to be parabolic.

First, I guessed the pressure difference between the inlet and outlet by using the velocity at the center line between the flat plates (10 m/s) and the 1 atm, 20 C air for calculation. So that, the pressure difference was 0.143149 Pa.

Then, I created a domain and specified the boundary condition as follows, at the inlet ---> Inlet / static pressure 0.143149 Pa (Ref Pressure 101325 Pa) at the out let---> Outlet / static reference pressure 0.0 Pa at the top wall ----> Wall / no-slip stationary at the center, back and front planes ---> Symmetrical planes

After running, I found that the results are not as what I've expected.

From the manual in the Advice on Flow Modelling section, it illustrates that this type of Boundary Specifying is not reliable. (Static pressure at the inlet and Static pressure at the outlet)

So, I've tried to change methods of specifying the BC in so many ways, such as,

1. Mass flow rate at the inlet and static pressure at the out let (The most robustness) 2. Total pressure at the inlet and static pressure at the outlet 3. Total pressure at the inlet and mass flow rate at the outlet

By applying these BC methods, such a parabolic velocity profile hasn't been found.

Could anybody explain what went wrong please?

Jackie

OK. Thank you everybody that come in this topic. I already found my fault. It's just wrong setting in some parameter.

Jackie