Fluids

Fluids

PBM changes the multiphase flow

    • Philxyz
      Subscriber

      Hi,

      I'm currently simulating a multiphase flow in 2 phases with pbm. phase 1 is gasoline and phase 2 is water. The goal is to simulate an emulsification process with water as the dispersed phase.

      As my multiphase model I'm using the Euler and Mixture model to compare them. (see attached screenshots)

      In the beginning I was simulating the flow without any PBM activated, just So I know that it is running stable. The Problem is that as soon as I activate the PBM my flow pattern changes completely, while it is still running stable and converging.

      So the question is if anybody knows why the flow pattern changes or how to fix it.

      Thank you

    • DrAmine
      Ansys Employee
      First of all you need to pay attention the kernels you want to use: almost none of them have been calibrated for your flow system. Check the references in the Guide and do a literature review. Second: now you have a polydisperse flow with changing diameter-> affects Interfacial area, drag force and non-drag forces. I won't use surface tension force for this kind of subjects. Cheers!
    • DrAmine
      Ansys Employee
      Bit looks more weird than just my explanation: please check the boundary condition if something went wrong. Also check without any breakage and coalescence and just use a dummy polydisperse by just initializing and using a 1 for certain bin size at inlet. Debuging will help to understand!
    • DrAmine
      Ansys Employee
      And Report back!
    • Rob
      Ansys Employee
      Another question, which way is up in your model?
    • Philxyz
      Subscriber
      thanks for the advice. I'll run it as suggested and see what happens.
      I attached another screenshot with the coordinate system. up is in the Y (+) direction. So Gasoline comes in from the left side and water from the top.

    • Rob
      Ansys Employee
      Hmm, if you plot phase fraction on the walls what do you see.
    • Philxyz
      Subscriber
      You can see how the water is flowing directly onto the wall and further downwards.
      In comparison without PBM the water is also flowing onto the wall but is getting pushed towards the outlet immediately

    • Rob
      Ansys Employee
      Perfect. Now look at the droplet sizes in the two cases.
    • DrAmine
      Ansys Employee
      Go through what we have shared yesterday with the debugging and check the diameter predicted with your current model. Also share residuals and convergence behavior of both runs.
    • DrAmine
      Ansys Employee
      Check the size Check the Size Check the Size :) :) small droplets are swept away.. Large droplets have more resistance and they accumulate...
    • Philxyz
      Subscriber
      I don't have any droplet size in the model without PBM except the default value (1e-05), but I think I'm getting what you mean. The flow of the individual droplets looks suspiciously close to the flow of water without PBM. Meaning that fluent is tracking the flow of the individual droplet and depicting it as the actual phase (water). I hope I make sense right now :)
      Disabling the surface tension force seems to do the trick. The flow looks to the one without the PBM (the simulation isn't completely done yet). Also I'm letting in just 1 BIN size at the moment.



    • DrAmine
      Ansys Employee
      Good progress.
    • Philxyz
      Subscriber
      Little Update: The last simulation is run through without any problems and without any change in the flow pattern.
      What I got from different simulations is as stated before: depending on the BIN size I set up at the inlet my flow pattern changes.
      The changing diameter, or rather a different diameter than before, changes drag and non-drag forces and the interfacial area of the droplets as DrAmine already pointed out.
      So basically if I have a flow in which I want to track droplets via PBM I can't compare flow patterns of simulations with different inlet droplet sizes.

    • DrAmine
      Ansys Employee
      Because as mentioned from the beginning: you need to "review" the kernels used for breakup and coalescence which have not been calibrated for this kind of system. So actually either you set a limit for the bin sizes (Top), adjust the kernels or you stick to a monodisperse formulation. There is another model which try's to enforce a special treatment where the secondary phase accumulates. Is is the GENTOP approach, relies , however, on the PBE and so on the kernels you are using.
    • Philxyz
      Subscriber
      Thank you for all the advice
    • DrAmine
      Ansys Employee
      Welcome!
Viewing 16 reply threads
  • You must be logged in to reply to this topic.