Fluids

Fluids

How to model free convection warming of liquid in a plastic bag

    • dodonn4
      Subscriber

      Hi everyone,

      I am working on a simulation in which a blood bag is removed from a cooler and placed in ambient air. The goal is to determine the temperature on the outside of the blood bag when the center of the blood reaches a certain temperature.

      I am trying to model free convection in Fluent, but keep struggling with the boundary conditions. Other examples I have found online involved a hot solid object cooling off in air, but my system involves a cold liquid warming up in air through free convection and conduction through the plastic bag wall.

      I have tried a few different models that essentially all have a large enclosure around the blood bag to represent the warmer ambient air, but had no luck. The boundary conditions applied included pressure inlets on every side of the enclosure except the top, which is a pressure outlet, and then wall boundary conditions with shell conduction applied to the plastic bag wall.

      Could anyone recommend a different approach to this problem, or possibly share some helpful references/documentation to guide me?

      Thanks in advance!

    • michael.hanchak
      Subscriber
      Couple of ideas:
      1) use "outflow" boundary condition for the top and pressure-inlt for the sides (and bottom). You could also try "wall" for the sides.
      2) if the blood is cold and thick, model that as a solid with the right density, k, and Cp. In other words, first ignore convection in the blood.
      I always try to simplify the problem as much as possible. Ignoring blood convection and using standard HTC for convection, what does the hand calculation predict?
      Hope this helps, Mike
    • dodonn4
      Subscriber
      Thank you! I tried your first idea and it seemed to work fairly well. However, there is no convection happening inside the bag, and I cannot figure out how to solve that. After running the simulation the temperature gradient of the blood inside the bag is a constant value, When reading the documentation, it states that one of the thermal boundary conditions I can apply at the wall is a convective boundary, but I cannot seem to find that anywhere when setting my boundary conditions. In the thermal tab of my wall boundary condition, the only options for "thermal conditions" are: Heat Flux, Temperature, and Coupled.
      The closest way I could find for setting convection is by making the blood volume a radiator type boundary condition and inputting the heat transfer coefficient there, but I do not think the radiator condition is accurate.
      Any advice for be greatly appreciated!
    • michael.hanchak
      Subscriber
      Glad it helped! A couple of ideas:
      Make sure gravity is turned on.
      The interfaces between air-to-bag, and bag-to-blood should be of type "coupled." The convection will be determined by the flow solver (conduction plus advection)
      Do you have temperature dependent density for the blood? If not, at least the Boussinesq approximation value? Without density changes or Boussinesq there will be no thermal convection in the blood.
    • dodonn4
      Subscriber
      I think my problem is with the interfaces. When making the named selections, I create an interface between the faces for air-to-bag interface, then I make another for the faces for bag-to-blood interface and they end up being the same exact interface. I get an error stating there are overlapping named selections, and when I highlight each individual named selection they are the same thing. For example, if I hide the geometry for the blood inside the bag and then highlight the named selection for the bag-to-blood interface, it shows the faces from the air-to-bag interface. Essentially, it seems to be acting like the bag is not there.
      Could this be an error in my geometry setup? I created the bag by copy-and-pasting the faces of the blood volume and giving those surfaces a shell thickness during meshing. I can highlight and see there are different surfaces for the bag and the blood, but when I try to apply a named selection it seems to mess things up.
    • Rob
      Ansys Employee
      If you did the share topology step the surface between the blood and air should be a wall and wall:shadow pair. Heat will pass through the surface (you can set a thickness too for heat transfer) and you can inflate both sides to get a good near wall mesh.
    • dodonn4
      Subscriber
      Thanks for the suggestion, Rob!
      Yes, I did do the share topology step, and there is the wall and wall:shadow pair for that surface. I set the thickness and turned on shell conduction.
      However, I am still stuck figuring out the convection. My previous fluent model corrupted so I had to start over, but now I can't seem to get the blood or ambient air volume to load into fluent. All of my other named selections loaded fine, and there are internal boundary conditions named "interior-blood_domain" and "interior-air_domain" present, but when I select them an error appears: "Surface for zone interior-blood_domain doesn't exist. It must have been deleted. It can be created using "Zone Surface panel".
      I did not have this issue is my previous (corrupted) model, and I used the same named selection choices/methods, so I am confused on how this happened, and how to resolve it.
      Just to see, I ran the transient simulation anyway, and within a few iterations the internal temperature of the blood was equivalent to the ambient temperature of the air. From my time step settings, this temperature change occurred over a few minutes, so I know something is incorrect.

    • Rob
      Ansys Employee
      You set the fluids on the cell zones. Face zones (boundary conditions) are for walls etc.
Viewing 7 reply threads
  • You must be logged in to reply to this topic.