On-site surrogates for large-scale calibration

TL;DR

This paper introduces a Bayesian on-site surrogate modeling approach for high-dimensional, unstable computer simulations, improving calibration accuracy and computational efficiency in complex engineering problems.

Contribution

It develops a novel Bayesian methodology with on-site surrogates tailored for high-dimensional, unstable simulators, extending calibration frameworks to handle real-world industrial data.

Findings

Effective on toy data and honeycomb seal example

Balances data fidelity with computational tractability

Handles numerical instabilities and missing data

Abstract

Motivated by a computer model calibration problem from the oil and gas industry, involving the design of a honeycomb seal, we develop a new Bayesian methodology to cope with limitations in the canonical apparatus stemming from several factors. We propose a new strategy of on-site design and surrogate modeling for a computer simulator acting on a high-dimensional input space that, although relatively speedy, is prone to numerical instabilities, missing data, and nonstationary dynamics. Our aim is to strike a balance between data-faithful modeling and computational tractability in a calibration framework--tailoring the computer model to a limited field experiment. Situating our on-site surrogates within the canonical calibration apparatus requires updates to that framework. We describe a novel yet intuitive Bayesian setup that carefully decomposes otherwise prohibitively large matrices by exploiting the sparse blockwise structure. Empirical illustrations demonstrate that this approach performs well on toy data and our motivating honeycomb example.