Fluids

Fluids

For DPM model,grid size at boundary layer need to be larger than particle size?

    • 2019023289
      Subscriber

      /* EDIT by DrAmine */

      Hello,

      Could you please help me?

      The DPM model needs to meet two conditions: 1. The particle size must be far smaller than the mesh size, enough to be regarded as a point masses
      2. the particle volume fraction should be less than 10vol%-20vol%.

      DPM model and vicous  model have contradictory requirements on mesh size. Viscous model requires mesh size to be small enough, while DPM model requires mesh size to be larger than particle size.

      But for boundary layer grids, the size needs to be very small, even smaller than the particle size, in order to calculate accurately.

      Is it ok if the mesh size at the boundary layer is smaller than the particle size? Is the error in this condition acceptable?

      Really Appreciate!

    • Rob
      Ansys Employee

      Both statements are true: and this results in a mesh that's either too fine at the wall for the particles, or too coarse for the flow.  The solver will work if particles are too big, but wall collision is only calculated in the near wall cell. 

    • 2019023289
      Subscriber

      Thank you,sir.

      Sorry,i am not mean using MPM model.I mean  is it ok to use DPM model  if the mesh size at the boundary layer is smaller than the particle size? And I don't understand 'The solver will work if particles are too big, but wall collision is only calculated in the near wall cell. '

      Thank you,again.

    • DrAmine
      Ansys Employee

      I need to edit your post by removing the screenshot from the documentation.

    • DrAmine
      Ansys Employee

      MPM is not like DPM: there you need to resolve the paritcles to capture the forces properly. 10 to 20 cells per particle are required!

    • 2019023289
      Subscriber

      Thank you, Mr,Ali

      So, is it ok to use DPM model  if the mesh size at the boundary layer is smaller than the particle size? And I don't want use MPM model.

    • DrAmine
      Ansys Employee

      Are you now trying to use DPM or MPM?

       

      I can write chapters on this... But for now:

      For DPM it is recommended to keep particle size smaller than the mesh size: because one can control that in an easy way.

      For sure that cannot be always fulfilled espcially when one wants to resolve boundary layers properly. Just do the right mesh for the flow and accept this paradoxon and try to increate particle streams to make statistics better and flow rate per stream / parcel small to get things well distributed.

       

      So start running and then do a sensitivity

    • 2019023289
      Subscriber

      Thank you,sir.

      I try to use DPM.

      You mean it's OK to  use DPM model  if the mesh size at the boundary layer is smaller than the particle size  if I increate particle streams.

      And increating particle streams can decrease this error?

    • DrAmine
      Ansys Employee

      Yes. Just take all that into consideration when evaluating the results.

       

    • Rob
      Ansys Employee

      Increasing particle streams will reduce the parcel mass: that tends to give a more stable solution as the DPM source is more spread out. 

      Particles larger than the near wall cell may cause stability issues (hence more streams) and can mean the particle is closer to the wall than it's radius would permit. Otherwise, it'll work, but you need to be aware of the conflicts when analysing the results. 

Viewing 9 reply threads
  • You must be logged in to reply to this topic.