[wimblo]

Hi,

I am currently simulating the airflow in a small room with two identical inlets and one outlet. I'm using a std k-e with enhanced wall function model. No other models are enabled for now.

In order to have a fully developped flow at the inlet, I did two simulations.

1) I ran a simulation with a long duct on which I imposed a mass-flow-inlet boundary condition of 0.2388kg/s. I then used the define>profile option to export the velocity profile (x, y and z velocity) at the outlet of that duct.

2) I ran a second simulation with the room i'm testing. I used the profile from the first simulation for my inlets. I used a velocity-inlet boundary condition with a cartesian coordinate system (with the profile from simulation 1 as x, y and z velocity)

Problem is, when I monitor the mass flow rate at the inlets, I get the following results:
inlet1 = 0.1177kg/s
inlet2 = 0.23118 kg/s

I am using the Least Squares profile interpolation method. The results are even worst if I use the other interpolation methods.

I've tried refining the mesh in both simulations and making the boundary mesh identical with no success.

I know I could just modelize the actual duct in the second simulation, but using the profiles would make my live easier in future simulations.

Does anybody know how I could make it work properly?

Thanks.

[LuckyTran]

I think you are getting the right result, meaning they should have different mass-flow rates.

But have you considered instead using a mass flow inlet and specifying the velocity profile? The actual value of velocity will float, but its shape is constrained.

Btw if you are going to simulate a long pipe anyway, then might as well include it in your model. For getting a fully developed profile, I would use periodic BC's rather than a long pipe. The problem is you don't know what mass-flow rate to do this simulation at.

[wimblo]

Thanks for the quick reply.

I will include the pipe in my model and compare the results.