Fluids

Fluids

Floating point exception

    • Chad Gouws
      Subscriber

      I keep getting this error:


       


      "Error: floating point exception


       


      Error Object: #f"


       


      I am simulating a porous-jump turbine with a casing attached to determine the drag and power coefficient using k-epsilon model. If I set the casing to be "interior", the simulation runs "correctly" but then the incompressible gas flows through the casing wall, thereby negating its presence. But when I try to set the casing to be a wall boundary, the error occurs and all but one tiny section of the domain is zero velocity.


       


      I can't seem to understand why the wall function causes errors when the mesh doesn't change but the boundary condition does. 

    • Rob
      Ansys Employee

      Please can you post some images of the geometry and mesh? Indicate on those the flow speeds etc.

    • seeta gunti
      Ansys Employee

      Helo Chad,


      What is the maximum velocity of the domain? Can you please post some pics for more understanding of the issue.

    • Chad Gouws
      Subscriber

      The inlet velocity is 8 m/s, the maximum shown in the simulation is in the order of hundreds of metres per second.


      Velocity inlet, pressure outlet, symmetry and porous jump boundaries were used.


      The first image is the porous jump bare turbine, the simulation ran correctly

    • Rob
      Ansys Employee

      Please check your mesh quality, I think that's causing the problems.

    • Chad Gouws
      Subscriber

      Would making it much finer around the casing do the job?

    • Rob
      Ansys Employee

      Possibly, you need to resolve the geometry with enough cells to calculate the flow and also avoid skew cells, big jumps in cell size etc.  Have a look in the Meshing tutorials attached to the R19.x help for some pointers.

    • seeta gunti
      Ansys Employee

      Hello Chad,


      Your mesh is really bad. If you check the mesh on top surface marked 1, you have only two cells and the region I marked 2 is also not recommended. You might be having very high skewed cells in that region. Can you do the mesh check in Fluent? will the mesh pass the mesh check?


    • Chad Gouws
      Subscriber


      Thank you so much rwoolhou and seeta, I really appreciate the help.


      Would this be seen as a better quality mesh? It has been refined right at the casing and the given improvements were made. I did a mesh quality check and ran the mesh improve function. The solution continuity in residuals is now diverging. 

    • Rob
      Ansys Employee

      You're welcome.


      The mesh is better, but I'd check the cells on the yellow surface of the second image. There looks to be a small radius that's skewed the mesh and the area where the 4 surfaces come together may also need some investigation.  You may also need to reduce the maximum cell size.


       

    • seeta gunti
      Ansys Employee

      In the second picture of your mesh, I could see the skewed cells. What is the maximum skewness of your mesh? fluent will report orthogonality. Is it with in the limit?



      At the end of the edge, I could see again two cells. You need to refine further and check for skewness at the marked region. 

Viewing 10 reply threads
  • You must be logged in to reply to this topic.