3 Replies Latest reply on Nov 6, 2016 2:52 PM by Bill McEachern

    2-D Laminar steady state simulation of backward facing step fails to provide accurate results

    Enrique Angola



      I am trying to perform a well known simulation of a 2-D Laminar steady state Re = 389 flow over a backward-facing-step. I tried the simulation in another software using the following boundary conditions:


      1) Inlet vel = 0.544 m/s

          pressure gradient = 0

      2) Outlet pressure = Atmospheric pressure

          Vel. gradient = 0

      3) no slip walls


      Using a standard quadrilateral mesh of 50,000 total cells with mesh refinements around the step I got very accurate results when comparing with experimental data.


      Then I tried performing the same simulation on Solidworks FlowSim and results were off. On Solidworks I am using the same model for the channel, and I generated a very similar mesh. Boundary conditions on the 2-D simulation were the following:


      1) Inlet velocity = 0.544 m/s
         check box: fully developed flow

      2) Pressure outlet: set static pressure to atmospheric pressure (101300 Pa in this case)

      3) Ideal walls


      On my model I have a very long channel before the step, so I expect to see the usual parabolic 2-D channel velocity profile, but I don't. I see  a parabolic profile at the inlet which becomes completely linear around the middle of the channel on the way to the step. When I plot the velocity profile right before the step, and right after the step, the results are far from experimental results and/or my other simulation.

      Is there anything I am not setting up properly on my simulation on FlowSim? I even tried doing mesh refinements on the walls, but results stay the same.


      Thank you