Thermal matching is a key stage of the development process for a gas turbine engine where component models are verified to ensure the correct metal temperature distribution has been used in life calculations. The thermal match involves adjusting parameters of a thermal model in order to match an experimental temperature distribution, usually obtained from a thermal paint test. Current methodologies involve manually adjusting parameters, which is both time consuming and leads to variation in the matches achieved. This paper presents a new method to conduct thermal matching, where Gaussian process regression is utilised to obtain a surrogate model from which optimal parameters for matching are obtained. This standardised procedure removes subjectivity from the match and gives faster and more consistent matches. The method is introduced and demonstrated for a number of cases involving a leading edge impingement system that has been isolated from a high pressure turbine blade.


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Thermal matching using Gaussian process regression


    Beteiligte:
    Pearce, Robert (Autor:in) / Ireland, Peter (Autor:in) / Romero, Eduardo (Autor:in)


    Erscheinungsdatum :

    2020-05-01


    Format / Umfang :

    9 pages




    Medientyp :

    Aufsatz (Zeitschrift)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch




    Sunspot cycle prediction using Warped Gaussian process regression

    Gonçalves, Ítalo G. / Echer, Ezequiel / Frigo, Everton | Elsevier | 2019


    A confidence-based roadmap using Gaussian process regression

    Okadome, Y. | British Library Online Contents | 2017



    HETEROSCEDASTIC GAUSSIAN PROCESS REGRESSION USING NEAREST NEIGHBOR POINT ESTIMATES

    Robani, Muhammad D. / Palar, Pramudita S. / Zuhal, Lavi Rizki | TIBKAT | 2021


    Learning Uncertainty using Clustering and Local Gaussian Process Regression

    Zhang, Yiming / Ghosh, Sayan / Asher, Isaac et al. | AIAA | 2019