Fluids

Fluids

Super Strange behaviour for multiphase modelling with combustion

    • HorstvanGrass
      Subscriber
      Hi,nI'm trying to simulate a combustion of H2+O2->H2O with later condensation of water. nHere is my setup:n-real gas soave redlich kwong, critical O2 and H2n-Eddy dissipation model for chemical interactionn-turbulence k omega SST due to its dampening propertiesnBecause of convergence issues I did the following steps:n1)First calculate without multiphase modellingn2)Interpolate all data in the multiphase modellingn3)Deactivate the Volume fraction equation for a first iteration to have the same results as in (1)n4a)Startet with Mixture(with still deactivated Volume fraction) multiphase model, always divergencen4b)Startet with Eulerian (with still deactivated Volume fraction), not great but no divergence Now my question: Why the more complex Eulerian model is converging? I thought the mixture multiphase modell is not as accurate but much more stable.nWhich equations does the Volume fraction explicitly activate, I don't get it completely from the Theory Guide?.Thx for any help!n
    • Surya Deb
      Ansys Employee
      nI am trying to understand your setup here a bit. When you interpolate the data, do you interpolate Velocity , Pressure and Temperature? Instead of doing this, why don't you start with the Multiphase but just have 1 phase inside your domain. This will be essentially similar to a model without multiphase. nDo you see convergence issue for this kind of setup as well? The interpolation for this will make more sense. Although it is difficult to remark without understanding the model and steps in more details.nThe Mixture multiphase averages the properties in cells based on the volume fractions of the phases present. I am not sure if the initial interpolation without any volume fraction evolution causes this.nThe Euler-Euler model on the other hand has a separate set of governing equations for each phase. Due to the initially interpolated data, it might be that is does not fail immediately. But have you checked if the results make sense or not?.Regards,nSuryan
Viewing 1 reply thread
  • You must be logged in to reply to this topic.