Fluids

Fluids

The question of discrete phase volume in (D)DPM modelling

Tagged: , ,

    • Sardar
      Subscriber

      As far as I know, the (D)DPM modelling does not take into account the volume of the discrete phase - only its mass and density are defined (which mathematically does result in some volume but the volume of the powder is not "added" to the slurry during injection.)

      Now please imagine the scenario below:

      a 100 litre container is to be filled up with a slurry. The slurry contains 80% VOLUMETRIC water and 20% powder. The powder is to be injected into the domain later.

      So, how much of the container should I initially (i.e. before injection) fill with water in my simulation? Should I fill ONLY 80% of the container with water, and add the discrete phase? If so, then this leaves us with only 80% of the container occupied with water and powder, while in reality the whole container is to be filled. Or should I FULLY fill the container with water initially regarless of the fact that, again, in reallity, only 80% of the container is occuppied with water?

    • Karthik R
      Administrator
      Hello It depends on what you are modeling. Technically, if 80% of your vessel is filled with water, 20% is air. If you are ignoring this, then you might wish to fill the entire container with water. Does that make sense?
      Karthik
    • Sardar
      Subscriber
      Thank you for your reply . If I want to rephrase my question, I can put it this way in two possible scenarios:
      What is wrong with filling 80% of the vessel with water, and then inject powders to fill the remaining 20%? In this scenario, the problem would be that, since injection does not add to the volume of the slurry, only 80% of the vessel will be involved in the simulation, while in reality the vessel is full.
      What is the problem with filling up (i.e. 100%) the vessel with water and then inject powders? I would say this does make better sense because at least we will have the entire volume of the vessel involved, just like that in real world. However, it is not in line with reality either, because in reality 80% is water, not 100% as we initially assumed in this second scenario.

      Thank you
    • Karthik R
      Administrator
      Hello In these scenarios, with you say 80% water, is it 80% by volume of the mixture or 80% of the volume of the vessel? Please make sure that you understand the initial conditions. This will be the key to answering your question.
      Karthik
    • Sardar
      Subscriber
      Since in real world, the vessel is full with the mixture, 80% of the vessel is water and again 80% of the mixture is water. If I am not missing a particular point here, I still cannot see your point how theses two can be different in this particular case.

    • Rob
      Ansys Employee
      But DPM has no volume, and at 20% loading the model isn't really applicable. I'd switch to the Eulerian Granular approach which makes the whole problem moot.
    • DrAmine
      Ansys Employee
      DDPM takes care about the particle volume displacement. DPM does not.
    • Sardar
      Subscriber
      Thank you. For 20% (which is larger than 10-12%) I too am using granular as you suggested and turn on the DDPM.
      Can you please say why this is a trivial problem for a volume fraction as large as 20%?
    • Sardar
      Subscriber
      Amine, with DDPM turned on, does the simulation add to the initial water volume, as your comment suggests?
      Thank you
    • DrAmine
      Ansys Employee
      It does consider the volume of particles in the domain. I substract it from the full /cell volume keeping the rest available for the carrier phase. Have a look into the Theory Guide :).
    • Sardar
      Subscriber
      sorry for calling you back, but can you please tell if I leave some volume for my ddpm particles, again, how should I specify the tank top boundary? Because, I will thus have to fill the empty volume of the tank top with air on the one hand, and on the other, injecting particles will force the air out of the tank domain through the top roof. So, if I choose a wall (symmetry or zero-slip) for top, that will not be a proper assumption. And if I make the top a pressure outlet, I will have to specify the dreaded backflow turbulence values, about which I have zero data. I am gradually being convinced that...
Viewing 10 reply threads
  • You must be logged in to reply to this topic.