Fluids

Fluids

Extract a Surface heat transfer coefficient field in Fluent

    • LordMacharius
      Subscriber

      Hello everyone !


      I'm new here so maybe my post is in the wrong category or have wrong tags, I apologize in advance.


      So, as I said in the title, I would like to extract a surface heat transfer coefficient field from FLUENT.


      Here is my study :


      I'm working on a circular air-oil heat exchanger. Due to the fact of my complex geometry (i have a total of 306 pipe, 153 per half circle), i would need almost 581M (roughly extrapolation/estimation), I need to decompose my case in two parts : Air flow simulation and oil flow simulation.


      On my Air flow simulation, I modelized a quarter of my tube (for the rest, i will use symmetry). I have an inlet temperature for the air and a fixed temperature on the "wall" of my pipes. I would like to simulate the air flow around the pipe, and extract the Surface heat transfert coefficient all-along the pipe (not an average, I suppose that it will decrease with the distance from the start of the pipe) and use this surface heat transfert coefficent field as a boundary condition on the pipes in my oil flow simulation.


      So my question is, how to extract that coefficient as a field around the pipe (as a matrix for exemple) and apply it as boundary condition ?


      I hope I have been clear enough


      Best Regards


      Lord


      PS : I work with ANSYS Academic 17.0


      PPS : If needed, I can provide more informations/pictures

    • DrAmine
      Ansys Employee

      Provide more informations/pictures and try to highlight the differences between pipe and tube.

    • Rob
      Ansys Employee

      The external HTC will alter within the system, but I think you can write a profile out. You will need to make sure the HTC you add into Fluent is correct for the flow temperature and reference. Additionally the profile will contain x, y & z position so you need the inner pipe model to be in the same location. 


      Is there a reason for not calculating the external HTC from theory or modelling the whole unit in one go? You do know that you can use nonconformal mesh between the two parts to give a non-matching mesh at the pipe surface & model the pipe as a thin wall? 

    • LordMacharius
      Subscriber

      For my calculations, and from the recommendations of my professor, I'm using the k-omga SST turbulence model. From there, I have determined that I should have a Wall Yplus in the range of 0,02 to 1,6. With this in mind, and the caracteristics of my flow, I need a first cell's height of 0,02 mm. This leads to a mesh of 581 millions elements for a full-sized unit and almost 80 millions elements for the geometry posted earlier, wich only have the air flow. So I can't compute on the whole unit in one go.


      I am currently trying different settings with non-conformal mesh to decrease the size of my mesh.


      I will take the external HTC proposition in serious consideration. But I will be still stuck on how to write a profile out of my study, with the x,y,z position.

    • DrAmine
      Ansys Employee

      You can then decouple the heat exchanger analysis say you make a fluent fluent coupling by sending the temperature and HTC back and forth between both analyses. You will need to find the optimal synchronization point and coupling frequency. 


       


      Regarding all in one analysis make a coarse run with high Re mesh to get a reference result. With a smart mesh method u can perhaps reduce the overall count for Low Re run. 

    • Rob
      Ansys Employee

      Once you have a converged solution go back to the Setting up Physics tab and look for Zones -> Profiles. You may be able to write out the heat transfer coefficient for the wetted wall. Read this into the model with the oil flow and set the external HTC to that profile: I've not tested this so you may find it's not that simple. 


      Alternatively run a 2d-axisymmetric model and take the surface average HTC from each layer/row and apply that manually to the tubeside model. 

Viewing 5 reply threads
  • You must be logged in to reply to this topic.