8 Replies Latest reply on Apr 16, 2011 8:11 AM by Rich Bayless

    Flow Anomaly Near Boundaries

    Joel Turco
    Joel Turco Apr 10, 2011 10:42 PM

      G'day,

       

      We are running a number of external flow scenarios and continue having issues with the flow results near the boundaries of the computational domain.

      Our system is fairly simple, we are trying to analyse subsea flow over a simplified seabed and around block shaped objects.

       

      We can't understand why the velocity is speeding up near the domain edges (see attached picture).

      Our input velocity is in the X-Direction and no matter what velocity we have the same problem.

       

      Any help on this matter would be much appreciated.

      Thank you!

        • Re: Flow Anomaly Near Boundaries
          Bill McEachern
          Bill McEachern Apr 11, 2011 8:51 AM (in response to Joel Turco)

          What exactly are you doing? You are obviously running a transient analysis. at the time given on the plot is 13.5 seconds. what is the time varying parameter (BC?)  and what is the objective of the analysis? I take it, it is an external analysis. If so what is the free stream velocity set as the BC?

            • Re: Flow Anomaly Near Boundaries
              Joel Turco
              Joel Turco Apr 11, 2011 9:02 PM (in response to Bill McEachern)

              I am trying to analyse the flow profile in and around this block-like arrangement underwater.  Also, I will look to determined the pressures that the block faces are exposed to under various conditions.  I have modelled the seabed as a planar surface with a specified roughness and the blocks sit on top of it.  The botom of the computational domain is flush against the bottom of the planar surface, such that the entire upper face of this planar surface (the 'seabed') is contained within the computational domain.

               

              Initially, I was running steady state analyses and encountered the same issue with flow near the walls of the computational domain.  However, I thought I would try a transient analysis to a) see if it made any difference, and b) allow for fluctuating inlet velocities down the track.  In this scenario, the free stream velocity is constant at approx 5.3 m/s in the x-direction.

               

              Thanks again for your help.

            • Re: Flow Anomaly Near Boundaries
              Joe Galliera
              Joe Galliera Employee Apr 11, 2011 5:26 PM (in response to Joel Turco)

              Could it be perhaps that you are modeling as a solid the outer boundary, like a wind tunnel?  For an external flow, you do not model the boundaries but the external envelope is determined by the computational domain.  By simply sharing your files with you VAR, they would be able to help you to clear up this issue.

                • Re: Flow Anomaly Near Boundaries
                  Joel Turco
                  Joel Turco Apr 11, 2011 9:09 PM (in response to Joe Galliera)

                  I thought I may have inadvertently done that at first but I don't think that is the case.  I used the external flow option and simply extended the computational domain to an appropriate size.

                   

                  I will contact my VAR, as you said.  Thanks!

                  • Re: Flow Anomaly Near Boundaries
                    Joel Turco
                    Joel Turco Apr 14, 2011 10:59 PM (in response to Joe Galliera)

                    Thanks for your help so far.  We have experimented with increasing and decreasing the size of the computational domain.  The best results we have obtained so far are achieved when the bottom of the computational domain is set below the seabed surface.  Therefore, the surface of the seabed is within the computational domain.

                     

                    The unusual results that appeared at the long sides of the domain are resolved by making them symmetry boundaries.  However, making the downstream boundary a symmetry condition does not fix the problem and gives very unrealistic results.

                     

                    We are still experiencing stagnation points at the bottom of the inlet boundary and a lot of stagnation at the downstream boundary, as shown in the images attached.

                     

                    Once again, any advice would be much appreciated.

                    We have contacted Solidworks and they are having similar issues.

                      • Re: Flow Anomaly Near Boundaries
                        Joe Galliera
                        Joe Galliera Employee Apr 14, 2011 11:15 PM (in response to Joel Turco)

                        Weird.  I could possibly remodel what I see in your pics, but I'm sure I'll miss some detail from your model.  Can you share with me?  My email address is in my user profile.

                          • Re: Flow Anomaly Near Boundaries
                            Joe Galliera
                            Joe Galliera Employee Apr 16, 2011 12:32 AM (in response to Joe Galliera)

                            We solved their problem offline... here is the resolution.

                             

                            Joel,

                            I think that the manual pressure dependency (151 kPa to 101 kPa from 0 m at sea floor to 4.5 m at top of sea) for ambient that you defined makes the difference.  I set it as a pressure potential and put the pressure that is at the origin (see picture); it figures out the pressure at the top for you after running a few iterations.  Now it only gets to 110 kPa (instead of 101 kPa) but that could possibly be caused by your own calcs for pressure at the sea floor accounts for sea water with salinity, thus here you might need to define a new water fluid with different density to account for the difference.