4 Replies Latest reply on Dec 29, 2015 12:05 PM by Glenn Whyte

    Pressure drop: Simple analysis not converging

    Shelon Marini

      Hello people,


      Frequently I have to analyse the pressure drop through a piping, and many times I face the same problem: the goals do not converge rightly.


      Here is a simple example: a linear pipe with diameter 250mm, and lenght 15m.


      Boundary conditions:

      Outlet velocity = 30 m/s

      Static pressure at the inlet = 101325 Pa

      Real wall at pipe faces

      The fluid is air



      Static pressure at the outlet


      So the pressure drop will be (static pressure at the inlet - static pressure at the outlet)


      I've set the goals convergence as the unic stop criteria, and changed the delta to 10Pa.


      Here are the results:





      As you can see, the static pressure at the outlet is varing more than 5000Pa even in the last interactions, what is a hugeness.


      Does someone know why is this happening, and how to solve the problem?


      It is interesting to note that if I reduce the pipe's lenght, to 5m for example, the goal converge well.


      Thanks for your attention!

        • Re: Pressure drop: Simple analysis not converging
          Amit Katz

          The fact that changing your domain size alleviates this problem leads me to believe it may have something to do with your mesh. How many cells do you have along the diameter of the pipe in the problematic calculation?

          One other thing I would try is setting an outlet mass flow rate or volumetric flow rate. By setting a constant velocity you are forcing a linear profile at the outlet which may be messing with your simulation.

            • Re: Pressure drop: Simple analysis not converging
              Shelon Marini

              Hello Amit!


              I have tried to use a volume flow rate instead of velocity, but no changes in the results.


              Regarding the mesh, I confess that I'm not familiar with all parameters, but I set the Level of initial mesh to 5, The minimum gap size to 0.05m, and the Minimum wall thickness to 0.00095m (which is the thickness of the pipe). With these configurations, the mesh in a section of the pipe has 16x16 cells. But also the same problem occur, without any apparent improvement.


              Any other idea of what can be wrong?


              Thank you very much!

                • Re: Pressure drop: Simple analysis not converging
                  Amit Katz

                  Your pipe diameter is 250mm, so I would set that as the minimum gap. You don't need to worry about the wall thickness in this case, since unless I'm mistaken this is a simple internal flow situation with no ribs or small features inside the pipe, and there is no heat transfer involved. Try performing the simulation with a level 3 mesh, 250mm minimum gap, make sure that "narrow channel refinement" is enabled and let's see what results you get back.

                  Also if you upload your project I could take a look at it sometime when I have a free moment.

              • Re: Pressure drop: Simple analysis not converging
                Glenn Whyte

                Hi Shelon,


                A couple of other things I'd recommend taking a look at, in addition to what Amit mentioned:


                • You mention defining a "real wall" on the pipe walls. What conditions do you have attached to this? By default, the wall will be recognized as a no-slip wall, with the roughness specified in the general settings (default - 0 roughness), generally you don't need a secondary condition unless you want to control the roughness of specific faces.
                • If you monitor the average static pressure at the outlet goal, rather than the bulk average pressure, do you still see the same oscillations?


                As Amit said, the project file will help if you need a deeper review, but I would recommend taking a look at the Flow Simulation tutorials - the very first chapter should replicate exactly what you need to do, so if you reproduce that process and its settings on your model, I'd be interested to see if the issue still persists.