Heat transfer in thermal store

Hello,

I am attempting to simulate the transfer of heat from a thermal store to an arbitrary external system - the thermal store is isolated in ANSYS for simplification. I have already defined the energy contained within the store using the energy source term, but I am unsure as to how to define the removal of heat from the store using Fluent. I.e. I need to know how to define a removal of 800W from a store of 3kWh capacity. Could anyone offer some guidance?

Many thanks in advance.

Comments

  • Can you have this as wall flux -800W ?

  • @YasserSelima Hi,

    I had considered this, but in truth, I need to run the removal of 800W, for example, for a large period of time, and then repeat this for other values xW, yW, and so on over and over again. Hence, I hoped someone might have an alternative method of simply removing the total energy - that's in Joules, not Watts - as opposed to removing the heat at intervals as this will take a very long time. My aim is to view the temperature degradation after the desired removal of energy from the initial store. Can anyone advise further?

  • Use a user defined function, DEFINE_SOURCE and you can attach it to your zone

  • RobRob UKForum Coordinator

    You need to be more specific on what you're trying to model. Source terms will remove heat from the volume, wall fluxes from the surface. Both are valid approaches but one may be better suited to your model. Is the thermal store flowing (ie convection) as Mechanical is faster if it's a solid region that's cooling due to the underlying solver.

  • tomk88tomk88 Member
    edited January 14

    Hi @Rob,

    The store itself is static, i.e. it is simply a stationary volume of hot water. There is a coil containing fluid which runs through the store and extracts heat from it. I am simply trying to model what the temperature degradation in the store will be at different points, bearing in mind stratification will take place as it cools, after a certain amount of heat is extracted from the store due to heat transfer between the store and coil. I then want to use this final state as the IC for a following model where I will extract more heat, at a different rate.

  • RobRob UKForum Coordinator

    OK, so you need to account for buoyancy in the water and forced flow in the coil. Modelling all of this in one go will be computationally expensive as the time step will be limited by the speed in the coil, but the overall duration by the change in temperature in the tank. Breaking into sub models will also be complex as the coil fluid temperature will be a function of the tank and also the flow in the tank.

    We're not permitted to give detailed advice on here, so:

    • Work through all of the heat transfer mechanisms and steps with your supervisor to find all of the unknowns (heat transfer coefficients etc)
    • Sketch these out on paper to see what's missing, and how CFD will answer these questions
    • Post the results of that on here: we can say if you're on the right track, and other Community members can give more detailed advice
  • Adding to @Rob advice, you need also to check the feasibility of using the temperature distribution in the tank at no load condition, as boundary condition for another case where the tank is loaded.

  • tomk88tomk88 Member
    edited January 20

    @Rob,


    I've simplified the model so the coil is now approximated by a length of pipe of identical surface area to the coil. I also have a single heat flux from the coupled wall of the outer water to the inner water. Now I have heat transfer from the water store to the water inside the pipe. I can clearly see an increase in temperature of the water inside the pipe from inlet to outlet by a reasonable amount. However, the temperature of the store seems to remain the same as it was initially, despite running the simulation for a long time. Heat cannot escape from anywhere else as the outside faces of the store are assumed as adiabatic. The store temperature has decreased only by 0.001K, which is a lot less than I would expect. The ratio of store volume to pipe volume is 111:1. Could you advise why the store temp is not changing? Is this reasonable, and perhaps I have oversized the store?

  • RobRob UKForum Coordinator

    If you check Q=m cp dT for the store against the heat transfer into the coil what do you get?

    However, I'm not clear what you mean by "single heat flux from the coupled wall of the outer water to the inner water". How are you extracting heat from the store, and where is it going?

  • tomk88tomk88 Member
    edited January 20

    @Rob , I've added a heat flux out of the bottom of the store of equal value (taking into account the different surface areas of bottom of store and pipe surface) and the heat distribution looks perfect now in both the store and the heat pipe. Is this an incorrect boundary condition or is this valid? Bearing in mind I am simulating stratification in the tank, so the colder water will settle at the bottom of the tank over large periods of time.

  • tomk88tomk88 Member
    edited January 20

    @Rob ,

    To reply to your question: I have applied a heat flux between the water store domain and the water in the pipe domain. This, and the temperature difference between the store and water in the pipe, is causing heat to be transferred between the two.

  • RobRob UKForum Coordinator

    As the pipe is connected to the vessel and pipe core you should just have a coupled wall with no heat transfer rates set. If you are setting any heat flux values inside the domain you've not connected faces and are unlikely to balance energy.

  • I am just going to massively simplify my model and use this as my heat transfer mechanism. I have ran the simulations and I receive reasonable temperature distributions. Thank you for your help! @Rob , @YasserSelima.

  • You are welcome, @tomk88 ! Good luck!

  • tomk88tomk88 Member
    edited January 21


    @Rob , I am curious as to why this did not work. When I neglected any heat fluxes and simply ran the cold water (298K) through the pipe core in the thermal store (343K), the temperature after 1 hour on the pipe surface was physically impossible - it ranged from -6K to 4000K in different regions which is absurd. This was using the coupled wall method you suggested, with the interfaces of cold water domain against inner pipe surface, and also the interface of hot water store against outer pipe surface were the coupled walls. Clearly I have defined something incorrectly. Could you estimate what this might be?

  • @Rob, I am curious as to why this did not work. Using the coupled wall method and neglecting any heat flux, I received temperature distributions along the pipe wall which were absurd and physically impossible. Could you advise as to why this might be the case?

  • RobRob UKForum Coordinator

    That's convergence issues. One problem is having to stabilise two flow volumes, the other is the mesh and timestep required to capture the flow features.

    Try starting by removing the water in the pipe and just looking at the heat transfer & flow in the vessel.

Sign In or Register to comment.