Uncertainty quantification in autoencoders predictions: Applications in aerodynamics

A data-driven model is compared to classical equation-driven approaches to investigate its ability to predict quantity of interest and their uncertainty when studying airfoil aerodynamics. The focus is on autoencoders and the effect of uncertainties due to the architecture, the hyperparamaters and t...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Journal of computational physics Ročník 506; s. 112951
Hlavní autori: Saetta, Ettore, Tognaccini, Renato, Iaccarino, Gianluca
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Elsevier Inc 01.06.2024
Predmet:
Airfoil aerodynamics
Autoencoder
Machine learning
Uncertainty quantification
Airfoil aerodynamics
Uncertainty quantification
Autoencoder
Machine learning
ISSN:0021-9991, 1090-2716
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:A data-driven model is compared to classical equation-driven approaches to investigate its ability to predict quantity of interest and their uncertainty when studying airfoil aerodynamics. The focus is on autoencoders and the effect of uncertainties due to the architecture, the hyperparamaters and the choice of the training data (internal or model-form uncertainties). Comparisons with a Gaussian Process regression approach clearly illustrate the autoencoder advantage in extracting useful information on the prediction confidence even in the absence of ground truth data. Simulations accounting for internal uncertainties are also compared to the impact of the variability induced by uncertain operating conditions (external uncertainties) showing the importance of accounting for the total uncertainty when establishing prediction confidence. •Proposed autoencoder ensemble distinguishes between different sources of uncertainty.•The approach correctly identifies regions of low prediction confidence.•The standard deviation of the ensemble is a reasonable proxy for the actual error.
ISSN:0021-9991
1090-2716
DOI:10.1016/j.jcp.2024.112951