Fluids

Fluids

Meshing at interface, floating point exception issue!

    • Anadi_Mondal
      Subscriber

      Hello,

      I have no problem to run the simulation but I am getting 'floating point exception' at a certain flow time(say 0.5s).I am using coupled solver. I tried reducing URF ,Courant number, step size, and increasing number of iteration per time step, k limit .I also check the boundary condition. Still I have the 'floating point exception' issue.

      At the middle of simulation ,it's showing below message and getting crushed.

      1)Divergence detected in amg solver: pressure coupled

      2)Turbulent viscosity limited to viscosity ratio of 1.000000e+05 in 5 cells. And, number of cells are increasing until crushed.

      I checked the orthogonal quality (average around 0.98) and skewness (average close to 0.00932, max ,0.5), but I have high aspect ratio(min 35.174, max 108.21, average 60.977 as mesh is long in axial direction.

      Here I have attached the mesh at interface and at cross section. Would you tell me if I have any problem with the mesh at interface?

      Or Will you give me any idea to skip the 'floating point exception' issue?


      Regards,

      Anadi

    • Rob
      Ansys Employee
      If you're adding DPM to the face with stretched cells it could cause an issue, and could lead to divergence. How much mass (and volume fraction) of DPM are you adding and are the particles inert?
    • RK
      Ansys Employee
      Hello,
      Is there mesh motion involved, for example, is dynamic mesh enabled? Are you using a UDF? What solution methods and boundary conditions are you using?
    • Anadi_Mondal
      Subscriber

      I am using UDF as I have defined evaporation ,entrainment source term, droplet injection ,liquid and vapor mass and momentum term . I did not enable dynamic mesh. If I need to enable then which mesh methods and options I should select?
      I am using coupled solver .Here is the details of boundary conditions:
      I have sliced a pipe into 3 sections (i,e slice at two position to make 3 different area). I am using the first surface area(injection _wall) to apply liquid mass flux , 2nd surface area(stabilization_wall) to stabilize the liquid film and finally 3rd surface area(annular wall). I have a certain vapor velocity at inlet and two surface injection (one at the interior surface (see the image above with first post) and another from annular wall). I have calculated k and omega value and a certain operating pressure. Here is the pipe with zones and boundary:
      If you need any more info, please let me know.
      Regards Anadi



    • Anadi_Mondal
      Subscriber

      I have attach an image of the face/surface that I am using for surface droplet injections. This annular region is the upper portion of a BWR .So , obviously the volume fraction of DPM is between 0-20%. Yes, the particles are inert.
      I have attached another image describing the boundaries ,zones, internal surface .Please see the image and kindly let me know if you think I am wrong anywhere or need to change /modify anything. As I have divided full pipe into two zones(pre annular and annular zone) ,at the beginning of simulation I select cell zone materials as water-vapor . I have added mass and momentum source term through UDF with annular zone.
      This is annular flow ,so vapor and droplet should be at the core and liquid film on the wall of pipe.
      Regards Anadi




    • Anadi_Mondal
      Subscriber
      & If I use Coupled solver I cannot continue simulation more than flow time of 0.5s, but for PISO solver I have no problem to continue for longer time (say 2s). But for PISO case, no wall film is forming on pipe wall after a certain distance. Here I have attached an image of film thickness on pipe wall for Coupled and PISO solver. Is there any particular reason of this difference.? I am using same settings for both cases except solver. After 0.5s I am getting divergence and floating point exception if I use Coupled solver.
      Regards Anadi


    • Anadi_Mondal
      Subscriber
      I mistakenly upload the same image for 0.6s and 2s for PISO solver. Here is the film thickness for 2s using PISO solver!


    • Rob
      Ansys Employee
      Which multiphase or film model are you using?
    • Anadi_Mondal
      Subscriber

      I am using DPM to simulate the core (vapor and droplets) and EWF to simulate liquid film on the pipe wall. Is there any problem with environment?
      Regards Anadi

    • Rob
      Ansys Employee
      You'll need to reduce the boundary facet aspect ratio, and ensure the EWF courant number stays below about 0.03
    • Anadi_Mondal
      Subscriber

      I have attached the aspect ratio of mesh and Courant number. So I need to reduce the courant number to 0.03 and would you tell the range for aspect ratio? I am using time step size between 0.0001 and 0.0008, and I have a velocity at axial direction of 12 m/s.
      Regards Anadi


    • Rob
      Ansys Employee
      Film courant number, it'll be reported in the TUI as you run the model.
Viewing 11 reply threads
  • You must be logged in to reply to this topic.