Blog posts

Conference visit at GPPS Beijing 19

We presented our previous work of turbulence model UQ at GPPS Beijing 19. Based on a metamodel forward propagation UQ framework, we conclude the following:

  1. SA model has limitations in predicting flow separations, where the discrepancy is most evident and the uncertainty is the largest.
  2. Shared sensitive coefficients to Reynolds stress, shock and separation include σ, κ, cb1 and cw2, all of which are source term related.
  3. Re-calibration of these coefficients to match the measured data of shock and separation leads to contradictory results, implying the limitation of the original SA model form.

Full text of the GPPS paper is available on ResearchGate.

Good experience to meet friends and fellow researchers. The weather of September in Beijing is a gift.

(credited to Yekai, 2019.09.16)

A friendly kitty in Tsinghua.

New poster to ASME Turbo Expo 19

We submitted a draft poster, entitled “Towards Improved Prediction of Compressor Flow by Uncertainty Quantification of Spalart-Allmaras Turbulence Model” (GT2019-92244),  to ASME Turbo Expo 19.

The poster is about our recent progress on sensitivity analysis of Spalart-Allmaras turbulence model coefficients on predicting compressor-like flow features. After that, we will conduct similar sensitivities studies on predicting airfoil stall and compressor stall boundary. Those results could hopefully provide some hint for future model modifications.


Updated on June 21, 2019:

We have won the People’s Choice Best Poster Award! In this year, 46 posters were presented by students from all around the world, and the award goes to the poster with the highest number of vote. Thanks to all the audiences who stood by and voted!