Transition SST vs LES

Hi! I am doing a study of a rear wing of a SAE formula. The preliminary study was the 2D study of the single profile at various AOAs, the study of two profiles (main and flap) and finally the study of three profiles (slat, main and flap). A triangular hybrid mesh with inflation (y_plus = 1) and the k-omega SST model was used. As we had problems with post profile flow separation, we decided to do the positioning through an optimization to maximize lift and the cl to cd ratio.

At this point, I wanted to go and research LSB (laminar separation bubble). The tests were performed at a Re=1*10^6 (v = 20m / s). Are there any turbulent laminar transition models? Can they be activated with the k-omega SST (Transition SST)? Or is it better to switch to a 3D studio and use LES?

Best Answer

  • RobRob UKForum Coordinator
    Accepted Answer

    If you refine the mesh until the solution stops changing (use monitors) you need to decide the criterion. Note that as the solution may be transient it might be more difficult than for a pure steady case.

Answers

  • RobRob UKForum Coordinator

    There are transition models in Fluent, and they're covered in the manual(s). LES will be computationally expensive as you'll need to resolve the eddies in space (mesh) and time (small time steps) but may give a more accurate solution if done properly.

  • Thanks for the reply.

    I am trying to run a study via Transition SST and have set up the mesh following the instructions in the manual for desired y + and grow rate. Having no experimental data for the three-way profile, I wanted to validate the study (mesh, setup) by studying a NACA 2415 profile and comparing the results to the experimental ones. If I get an error of less than 3%, can I consider my mesh & setup also valid for the case of the three profile? My problem is that I don't have experimental data for my three-profile geometry and I'm looking for a way to validate the study. I am using the SST transition because I am in 2D.

  • RobRob UKForum Coordinator
    Accepted Answer

    If you refine the mesh until the solution stops changing (use monitors) you need to decide the criterion. Note that as the solution may be transient it might be more difficult than for a pure steady case.

  • I had read that during the preparation of the mesh it is good practice to have a discretization around the wall of delta (y) = 20 * delta (x) = 40 * delta (z).

    In the case of the 2D study, I have interpreted this relationship in such a way that if I have a first layer thick value of 1.8e-5, the airfoil edge sizing is set with an edge sizing of 20 * 1.8e -5. Is it correct up to this point?

    In the case of Transition SST it is recommended to have 75/100 streamwise. What correlation is there between streamwise and edge sizing on foil?

    Because if it is recommended to have "75-100 mesh nodes in the streamwise direction", the edge sizing set as above becomes excessive

  • RobRob UKForum Coordinator

    That cell thickness is possibly not unreasonable. The usual approach is to calculate the cell height for a y+ of about 1 and then check y+ once the solution is complete. Remeshing (including adaption) is then used as needed.

    In terms of the streamwise mesh, you're wanting to keep the cell aspect ratio at a reasonable value, and capture any changes in the flow pattern along the wing chord. You are likely to need more cells at the leading edge and potentially wherever separation occurs.

Sign In or Register to comment.