Multi-fidelity data fusion through parameter space reduction with applications to automotive engineering

TL;DR

This paper introduces a multi-fidelity data fusion method that reduces parameter space dimensionality using active subspaces and nonlinear transformations, improving response surface accuracy in high-dimensional, data-scarce engineering problems.

Contribution

It proposes a novel multi-fidelity approach combining active subspaces and nonlinear level-set learning for high-dimensional function approximation.

Findings

Enhanced accuracy of Gaussian process response surfaces with low intrinsic dimensionality.

Effective reduction of parameter space improves multi-fidelity modeling in automotive engineering.

Validated on benchmark functions and a car aerodynamics problem.

Abstract

Multi-fidelity models are of great importance due to their capability of fusing information coming from different numerical simulations, surrogates, and sensors. We focus on the approximation of high-dimensional scalar functions with low intrinsic dimensionality. By introducing a low dimensional bias we can fight the curse of dimensionality affecting these quantities of interest, especially for many-query applications. We seek a gradient-based reduction of the parameter space through linear active subspaces or a nonlinear transformation of the input space. Then we build a low-fidelity response surface based on such reduction, thus enabling nonlinear autoregressive multi-fidelity Gaussian process regression without the need of running new simulations with simplified physical models. This has a great potential in the data scarcity regime affecting many engineering applications. In this work we present a new multi-fidelity approach that involves active subspaces and the nonlinear level-set learning method, starting from the preliminary analysis previously conducted in Romor et al. 2020. The proposed framework is tested on two high-dimensional benchmark functions, and on a more complex car aerodynamics problem. We show how a low intrinsic dimensionality bias can increase the accuracy of Gaussian process response surfaces.