Abstract Shock wave–boundary layer interaction (SWBLI) is one of the most prevalent challenges in the field of fluid mechanics. Excessive interaction may lead to strong flow separation, which is an important factor causing the deterioration of aircraft performance. Reynolds averaged Navier–Stokes (RANS) models remain the primary methods for predicting flow separation for engineering applications. RANS simulation results typically have uncertainty owing to the drawbacks of the model structure and the variability of the parameters. Therefore, evaluation of a turbulence model based on nominal predictions may be insufficient for reflecting its actual simulation capabilities. In this study, the Bayesian method is employed to re-evaluate the simulation capabilities of three k–ω models for hypersonic SWBLI flows: the original k–ω model and two k–ω models with compressibility corrections (Wilcox's and Zeman's corrections). This is achieved by perturbing the model parameters and combining with experimental data. Bayesian model evaluation based on different datasets is performed using the embedded model error method, and the three turbulence models are compared from three aspects: posterior mean of the quantities of interests, model error, and Bayesian evidence. The results show that the Bayesian calibration can effectively improve the predictive abilities of the three models for wall pressure and wall heat flux. Different from the large differences in the nominal results, the performance of the two calibrated k–ω models with compressibility corrections are very close and superior to that of the original model. Moreover, the evaluation results of this study are also proven to be applicable to other similar flows.

    Highlights The Bayesian method is employed to re-evaluate three k–ω models. The embedded model error method is used to estimate the model error. Turbulence models are compared from three aspects: posterior mean, model error, and Bayesian evidence. The applicability of evaluation results is investigated.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Bayesian model evaluation of three k–ω turbulence models for hypersonic shock wave–boundary layer interaction flows


    Beteiligte:
    Li, Jin-ping (Autor:in) / Zeng, Fan-zhi (Autor:in) / Chen, Shu-sheng (Autor:in) / Zhang, Kai-ling (Autor:in) / Yan, Chao (Autor:in)

    Erschienen in:

    Acta Astronautica ; 189 ; 143-157


    Erscheinungsdatum :

    2021-08-27


    Format / Umfang :

    15 pages




    Medientyp :

    Aufsatz (Zeitschrift)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch