saturated tank using evaporation-condensation model

Hello everyone,

I am trying to simulate a 3D liquid receiver tank filled with R134a when the entering flow is 5K subcooled assuming that the tank is at the saturated state initially. the geometry and initial volume fractions are attached. The goal is to see the state of the fluid at the tank outlet in steady-state conditions and also to see the equilibrium level, i.e if the tank is filled and the exit reaches the same inlet temperature or the tank maintain an equilibrium level with a temperature between the inlet and the saturated temperatures.


I am using Eulerian model to use two-resistant heat transfer models suggested by User guide to get rid of the tuning coefficients of the Lee model. But the pressure in the tank increases to unphysical values at very first iterations and I don't know why it happens. Can anyone give me a clue about it? Besides, the outlet mass flow is an order of magnitude higher than the inlet mass flow rate.

The primary phase is liquid and one evaporation-condensation mechanism is activated.

Bcs are mass flow inlet (in the top) and pressure outlet (in the bottom).

I highly appreciate the experts comments on this simple geometry but apparently complex physics. 😎

Comments

  • RobRob UKForum Coordinator

    Depending on settings you may need to account for the static head of liquid, so the outlet pressure may be too low. What temperature is the system relative to the boiling point? Again, if boiling occurs too quickly the gas has to go somewhere. What time step are you using?

  • Thanks for your comment,

    static head is accounted using operating density set to zero.

    the initial temperature of the gas is at saturated temperature and the inlet liquid is 5K subcooled.

    In fact, I used polynomial density of gas via temperature, and apparently as the liquid comes in, the density is constant and so pressure goes up since gas was like incompressible liquid!

    I am using Peng-Robinson density for gas now, and mass flow outlet BC instead of pressure outlet. I had a strange pressure changes at the very small times steps at the beginning but it becomes physical now using larger times steps (0.01 s).

    The simulation is running to see if I will have a physical volume fraction and temperature or not. I will update it here the possible problems of the model. Like now I can see that the liquid temperature is developing a strange colder column from outlet into the fluid that seems strange.

    I highly appreciate if the topic is followed after the update.

  • Dear @Rob

    I tried both pressure outlet and mass flow outlet BCs.

    Mass flow outlet: the pressure in the tank goes up to unphysically high values

    Pressure outlet: the outlet mass flow rate becomes positive and almost 10 times more than the inlet flow. Also the form of the VF contour seems strange and unphysical. please find it below after some time steps (TS = 0.01 s)

    P.S: I set one mass trasnfer mechanism (evaporation-condensation) to model both evaporation and condensation in the tank to find the equilibrium state. I found in the User guide that "From" must be the liquid and "To" must be the gas phase so I assume one mass transfer mechanism can model both evaporation and condensation at the interface.

    Please let me know for any tips!

    Thank you in advance, best regards.


  • RobRob UKForum Coordinator

    You're correct re the mass transfer. I'd use Ideal Gas over the Real Gas options as it tends to be more stable. If the first time steps are odd, what temperature did you patch the gas at? If it's already sub cooled the first thing it'll try and do is contract & condense.

  • Thanks for the comment

    I can try ideal gas too.

    both gas and liquid are patched at a temperature almost equal to the saturation temperature. gas a bit higher and liquid a bit lower to avoid spontaneous condensaton/evaporation. So the initial state of the tank is saturated and the inlet liquid is 5K subcooled.

    What is the contract & condense that I should do?

  • RobRob UKForum Coordinator

    What pressure did you patch at? With a compressible gas if you don't get the initial conditions right it'll try and expand/contract to suit the conditions. If you run your model on does the system stabilise and start to drain?

  • Thank you very much Rob,

    It is patched at 5 bar and the initial temperature of the tank is equal to the saturation temperature at this pressure (288.8 K). The inlet fluid is at 283.8 K accordingly (5k sub-cooled)

    I have tried to outlet BC and different results, both wrong: 😓

    pressure outlet: pressure in the tank is decreased a bit that seems normal due to pressure of the liquid column, so in the tank must be a bit less that the outlet pressure (5 bar). BUT the outlet mass flow is positive and much higher than the inlet flow! 😟

    Mass flow outlet: set equal to inlet mass flow, but the pressure goes up to very high values. 😔

    Maybe VOF should be tried using default tuning coefficients. I tried it before and at least the volume fraction contour was more physical than Eulerian model. but the same high pressure/outlet mass flow was observed with VOF as well! 😕

    There is no other BC that can be tried I dont know why the models are failed like this.

  • RobRob UKForum Coordinator
    edited October 16

    Did you calculate the properties at 5 bar gauge or 5 bar absolute? What's the operating pressure in your model?

  • Thanks Rob,

    I calculated them at 5 bar absolute. operating pressure is 1 bar as default. could it be the problem?! 🧐

  • RobRob UKForum Coordinator

    Probably, you'll be 1 bar out on the saturation temperature: you're running the model at 6 bar absolute if you've not changed the operating pressure.

  • Thank you Rob,

    To conclude, the problem was as you correctly said the operating pressure. I set it to zero and the simulation becomes physical now. 😊

This discussion has been closed.