Question about Fluent post-processing regarding calculation of Wall Heat Flux

Hello,

I am simulating a rotating disk in a free flow. I am using an axi-symmetric boundary condition and swirl is put on. The domain is a square, where the bottom boundary is the axi-symmetry boundary, the right boundary is the rotating wall (rotating around the symmetry axis) and the top and left boundary are ambient air with a fixed temperature. Furthermore there is a fixed temperature boundary condition on the plate.

Furthermore I am using a k-omega SST turbulence model. My mesh at the wall also has a yplus lower than 1.

I am calculating the heat flux from the start of the plate (the point at the wall and the symmetry axis at the bottom right) to half of the plate. However I have a question about how this wall heat flux is computed. In the theory guide it states that this is done by using: q=k_f * (dT/dn), however when I calculate the temperature gradient with the post processing and multiply it with the k_f (thermal conductivity of fluid) I am not obtaining the wall heat flux also calculated by the post processing. Could somebody maybe help me on what I am doing wrong in computing the wall heat flux, the same way as the post processing?

Thank you in advance!

Kind regards,

Cars

Best Answer

  • CarsvoCarsvo Member
    Accepted Answer

    Hello!

    I have checked the results with a simple case of which I knew the boundary condition at the wall. It seems that the gradient calculated in the post processing is not always the same as the one used in the solver. Therefore there was a slight error. It is now solved. Thank you!

    Kind regards,

    Cars

Answers

  • RobRob UKForum Coordinator

    There are still wall effects, which are covered in the Fluent Theory Guide. The most common cause for error in wall heat transfer relates to the temperatures used.

  • CarsvoCarsvo Member
    Accepted Answer

    Hello!

    I have checked the results with a simple case of which I knew the boundary condition at the wall. It seems that the gradient calculated in the post processing is not always the same as the one used in the solver. Therefore there was a slight error. It is now solved. Thank you!

    Kind regards,

    Cars

  • RobRob UKForum Coordinator

    I always do post processing in Fluent as it means I'm working with the solver and data rather than transferred data for this reason. Differences tend to be negligible but can be noticeable if the gradients are complicated or the mesh isn't sufficiently resolved.

This discussion has been closed.