I simulated 2-D Laminar pipe flow, and used periodic boundary conditions at outlet and inlet of pipe. I tied two method respectively to define the periodic boundary conditions: pressure gradient and mass flow rate(per 2*Pi). However, all the simulation results didn't agree with the analytical resutls. I cannot figure out the reasons.

Any suggetions would be preferred. Thanks.

1.You are using the axissymmetric model..? if you are,just specify one side of your 2-D geometry as axis. and pay attention to specify the correct pressure gradient at periodic boundary according to the laminar theory.. 2.Maybe you can try another method,i.e,writting a velocity inlet UDF and using velocity inlet and pressure oullet instead of periodic boundary... 3.By the way ,are you doing 3-D circular tube numerical simulation..? and you just simplified it into a 2-D axissymetric model?

Hi, Majestyezh. Thanks for your kind help.

I am trying to simulate 2-D fully developed laminar pipe flow as the first step of my research. As you suggested I specified the centerline of tube as axis(X axis) and used steady axisymmetric model(the reason I adoped steady model, I thought, was that the pipe length could be not necessarily long enough to obtain fully develped flow).

The B.C of my future research topic is periodic, therefore, in this 2-D-fully-developed-laminar-pipe flow simulation, I used translationally periodic B.C.. I input pressure gradient(small enough to gurantte laminar flow)for this periodic B.C., and I surely got probolic velocity profile of laminar flow, however, the scale of velocity is smaller than that of it should be(I compared the simulation result with analytical result). I also tried Mass Flowrate input,and got similar problem.

While, I tried Velocity profile inlet and outflow outlet(maybe I could later try pressure outlet as your suggestion 2), simulation results agree with analytical results.

So, I cannot figure out why Periodic B.C with steady model doesn't work for the case of 2-D-F.D.-Laminar-Pipe flow?

I figured out the question: not enough nodes on calculation domain.