PCA and Kriging for the efficient exploration of consistency regions in Uncertainty Quantification

Gianmarco Aversano (1) , John Camilo Parra-Alvarez (2) , Benjamin J. Isaac (2) , Sean T. Smith (2) , Axel Coussement (1) , Olivier Gicquel (3) , Alessandro Parente (1)

1 Ecole Polytechnique de Bruxelles
2 University of Utah
3 EM2C - Laboratoire d'Énergétique Moléculaire et Macroscopique, Combustion

For stationary power sources such as utility boilers, it is useful to dispose of parametric models able to describe their behavior in a wide range of operating conditions, to predict some Quantities of Interest (QOIs) that need to be consistent with experimental observations. The development of predictive simulation tools for large scale systems cannot rely on full-order models, as the latter would lead to prohibitive costs when coupled to sampling techniques in the model parameter space. An alternative approach consists of using a Surrogate Model (SM). As the number of QOIs is often high and many SMs need to be trained, Principal Component Analysis (PCA) can be used to encode the set of QOIs in a much smaller set of scalars, called PCA scores. A SM is then built for each PCA score rather than for each QOI. The advantage of reducing the number of variables is twofold computational costs are reduced (less SMs need to be trained) and information is preserved (correlation among the original variables). The strategy is applied to a CFD model simulating the Alstom 15 MWth oxy-pilot Boiler Simulation Facility (BSF). In practice, experiments cannot provide full coverage of the pulverized-coal utility boiler due to both practicality and costs. Values of the model's parameters which guarantee consistency with the experimental data of this test facility for 121 QOIs are found, by training a SM based on the combination of Kriging and PCA, using only 5 latent variables. (C) 2018 The Authors. Published by Elsevier Inc. on behalf of The Combustion Institute.

Mots clés: PCA; Bound-to-Bound Data Collaboration; Uncertainty Quantification; Surrogate models

Domaines: Génie chimique; Milieux fluides et réactifs; Applications [stat.AP]

Gianmarco Aversano, John Camilo Parra-Alvarez, Benjamin J. Isaac, Sean T. Smith, Axel Coussement, et al.. PCA and Kriging for the efficient exploration of consistency regions in Uncertainty Quantification. Proceedings of the Combustion Institute, 2019, 37 (4), pp.4461-4469. ⟨10.1016/j.proci.2018.07.040⟩. ⟨hal-02398468⟩