Nusselt number and Heat Transfer Coefficient in Forced Convection

lordofthethingslordofthethings The ShireMember


The setup: Air is flowing at 160°C inside a rubber tube 0.5 m long and 0.054 m inner diameter, at a flow rate of 0.2 kg/s. The environment temperature is 23 °C. 

The goal : What does the temperature distribution across the wall thickness look like?

A test-setup was created for the above conditions, and using a thermal imaging camera, the temperature across the wall thickness was obtained. Different rubber compounds of the same constructions were tested, and for the same input, different temperature distributions were obtained.

The simulation:

The flow is in the turbulent regime, and so I used the k-Epsilon, realizable model with enhanced wall treatment in fluent. For the above conditions, I get the following results for a steady-state analysis:

T_inner wall = 152 °C

T_outer wall = 95 °C

Surface Heat Transfer Coefficient (Area averaged) = around 200 W/m²K

Analytical Nusselt Number (Dittus-Boelter Equation) obtained was almost equal to the one calculated using the above HTC.

The problem: the simulated temperatures do not match reality. In reality the inner wall temperature in contact with the fluid is almost always much lesser than the fluid temperature itself. I tried to manually give a film coefficient for the inner wall to get the desired inner wall temperature in Ansys Mechanical. It is way smaller than the analytical HTC obtained from the Nusselt number. 

How can I change me analysis procedure in order to realize the actual wall temperatures obtained in tests?



  • RobRob UKForum Coordinator
    edited February 21

    Read the documentation to understand what you're setting. The HTC you set is on the outside of the domain: ie the boundary between the mesh and "outside". Did you model the solid bit in Fluent? Have you resolved the near wall flow? 

    Chances are the CFD results are about right (assuming you're more-or-less there with the models & mesh) for the boundary conditions you set. 

  • lordofthethingslordofthethings The ShireMember
    edited February 21

    I modelled the solid bit in Fluent. I have shared the topology between the flow volume and the solid part in Spaceclaim. I did not apply an HTC manually on the outside of the fluid because I felt it would be overconstraining the model. 

    The HTC you set is on the outside of the domain: ie the boundary between the mesh and "outside". 

    Do you mean the region between the solid and the fluid? Due to the shared topology, interfaces are automatically coupled in my model. Do you suggest taking out the shared topology and applying HTC directly in the "outside" part of the fluid?

    I'll check the documentation. Appreciate your initial response!

Sign In or Register to comment.