Fluids

Fluids

Problems reaching convergence

    • simonhenn
      Subscriber

      Hello,

      I am simulating an airflow over a flate plate. While I have solved earlier problems where I often got a floating point exception the solution still doesn't really converge. These are the oscillating residuals:

    • Rob
      Ansys Employee
      The residuals suggest some transients in the flow. The pressure plot implies it's not converged, so have a look at images of pressure, velocity etc to see what's still changing: possibly a separation zone is moving around which may mean there's a low pressure region that's not very stable. nHow does the mesh look when you plot the flow results onto it? Have good cells doesn't necessarily mean you have a good mesh. n
    • simonhenn
      Subscriber
      Thanks for your answer! What exactly do you mean by plotting the flow results onto the mesh?
    • Rob
      Ansys Employee
      Plot contours with the mesh on that isosurface/plane on and node values off. It lets us see the cell data and we can judge if the mesh is resolving the flow gradients. n
    • simonhenn
      Subscriber
      Here are some pictures of the complete Mesh and possibly critical regions. Let me know if you need some more info. n
    • Rob
      Ansys Employee
      Top contour shows some of the issue: you've got a fairly big jump in cell size and a very stretched cell right where the flow separation is trying to happen. It's also a fairly complex shape for the flow so could take a lot of iterations to sort itself out if the mass flow splits become part of the solution. n
    • simonhenn
      Subscriber
      Ok, so I'm gonna make the transition between the cell sizes smoother. Then are the high aspect ratio boundary layer cells no problem anymore?n
    • Rob
      Ansys Employee
      You need to be wary if the solution is changing down the length of those cells. Inflation layers are very good for attached flow to give you a cheap mesh with low y+ however they're not good if you have flow separation or jet impingement. n
Viewing 7 reply threads
  • You must be logged in to reply to this topic.