Fluids

Fluids

Multiphase Euler-Euler divergence

    • michelletramper
      Subscriber
      Hi all!nI have a problem with running transient multiphase simulations in patient-specific aortas. I have developped a model using the three-phase euler-euler model, to model blood flow in aortas. I have applied this model on several geometries, and some of them convergence properly, however I also have 2 geometries where divergence occurs. The perculiar thing is that I run the exact same settings for a single phase models (same mesh, boundary conditions etc.) and then everything works fine (no divergence at all).Does this problem occur more often and where should I try to find the troubles?.Kind regards,nnMichellen
    • Rob
      Ansys Employee
      Assuming the mesh is good have you got any regions with higher velocities? The common problem with Eulerian multiphase models tend to be in the phase interaction, that can often be transient. Less common is that as the solver is different it can be stiffer with only one phase set. n
    • michelletramper
      Subscriber
      Hi,nThanks for your quick answers. Yes the mesh is definetely good. It's applicable for the single phase models and for other geometries. Good that you mention that, because one of them indeed has a coarctation where higher velocities are present. Do you have a suggestion for a workaround? Right now, I'm not using any additional interaction forces except drag force interactions. nWhat I'm trying right now is running for 3 cardiac cycles with smaller timestep and more iterations per timestep before data extraction.Thanks.nnMichellen
    • Rob
      Ansys Employee
      That sounds good, also check the system residence time to make sure material has also had time to flow through the domain. Check the flow monitors are identical between cycle 1-2 and 2-3 before taking data for analysis: you may need to run a few more cycles to reach the repeatable cycle. n
    • michelletramper
      Subscriber
      Hi,nnYes thank you! I hope this works and otherwise I'll come back to this thread n
    • Rob
      Ansys Employee
      I might be hiding.... n
Viewing 5 reply threads
  • You must be logged in to reply to this topic.