Fluids

Fluids

Heat transfer and fluid flow in Pipe

    • jhjang
      Subscriber

      Hello everyone,


         I am trying to simulate 2D laminar developing flow and heat transfer in grooved enhanced pipe. The details are 


      Diameter 40mm, Length of pipe 400mm


      meshmesh detailmesh detail smooth


      Grooved meshgrooved mesh detailgrooved mesh inflation


       


      Inlet velocity 0.0125375m/s


      pressure outlet


      heat flux  50w/m^2



       


      Results


      When I calculate Nusselt number in this way…


      Surface integral, Area-weighted average, wall fluxes, Nusselt number & Skin friction factor, I feel these results are not in the right pattern...


      Area-Weighted Average


      Skin Friction Coefficient





      wall 0.048432428    (Smooth)             0.37105531  (grooved)


       


      Area-Weighted Average


      Surface Nusselt Number





       


      wall 15.567972          (smooth)             14.906946   (grooved)


       


      here the trend of results for both Nu and skin friction is in opposite order.


       


       Please help me to calculate local and average Nusselt number and friction factor


       


       

    • DrAmine
      Ansys Employee

      Please stop creating several threads for the same issue especially as we have started answering in the other thread. Are you now doing a 2D or 3D? 

    • jhjang
      Subscriber

      Thankyou Abenjadj sir,


      I am doing on 3D but since i was getting wrong results so i tried 2D also.

    • DrAmine
      Ansys Employee

      Just focus on what I have already asked: Have you adjusted the reference values to match your case. The default values are not valid for every case. 


      Moreover what do you mean with "here the trend of results for both Nu and skin friction is in opposite order." Do you have exp. results or references?


      Just to remind: Simple laminar flow you will have Cf to be proportional to the inverse of the square root of Reynolds number at external flat surface. Whereas the Nu number is proportional to the square root of the Reynolds number.


       


      For simple basic geometries we can refer to some well know correlations.

    • seeta gunti
      Ansys Employee

      Hi,


       


      Following video tutorial will help you on how to model the heat transfer in pipes from geometry to post processing. Please go through the following links.





      Regards,


      Seeta

    • jhjang
      Subscriber

      Hi, Seeta Gunti


        Thank you for this video. I have seen this video and know some more things but still, I have the problem how to get friction factor and Nu over the curve.


       

    • jhjang
      Subscriber

      Hello Amin sir,


       I do not have experimental or reference results. here what I have the opposite trend means friction factor and Nusselt number is less for the enhanced surface than the smooth surface. when I change the reference value of length as the diameter of the pipe then nusselt number start from around 10 but when I keep the reference length as default then nusselt number start around 800-900.


      I just extract the results directly using area weighted average, wall heat flux, nusselt number, and skin friction coefficient.

    • DrAmine
      Ansys Employee

      Ensure for well resolved boundary layer for your laminar case. You need to carry out a mesh sensitivity analysis to see if the trend is mesh dependent or not. 


      How big is the device Reynolds number?

    • jhjang
      Subscriber

      I have checked for Re 500 and 1000.


      Sir, will you please elaborate me on how to get Friction factor and Nu number directly from the software for any case.

    • atulsingh92
      Subscriber

      Hello @Abenhadj,


      A question about reference values, here. (I am doing a similar case of enhanced wall tubes with grooves, in 3D and turbulent along with periodic bc)


      I wanted to know exactly "how" can I tune this reference value to my case. Changing the reference value pretty much changes the Nusselt. So essentially I could end up having any nusselt nunmber, for any reference temperature.


      Now when you mention "adjusted to your case", does this mean, I should look at a similar experimental setup, for a particular reynolds, know what Nusselt does the experiments show, and then reverse calculate the neccessary reference temperature to input for my case?
      Or is there some other method to tune this value?

Viewing 9 reply threads
  • You must be logged in to reply to this topic.