Learning large scale industrial physics simulations

TL;DR

This paper discusses the development of fast surrogate models for industrial physics simulations, including reduced-order modeling and machine learning approaches for complex geometries, to improve design processes.

Contribution

It introduces novel reduced-order modeling techniques and machine learning methods tailored for large-scale industrial physics simulations, with open-source contributions.

Findings

Effective reduced-order models for non-linear structural and thermal analysis

Machine learning approaches for non-parameterized geometrical variability

Open-source tools and data for the community

Abstract

In an industrial group like Safran, numerical simulations of physical phenomena are integral to most design processes. At Safran's corporate research center, we enhance these processes by developing fast and reliable surrogate models for various physics. We focus here on two technologies developed in recent years. The first is a physical reduced-order modeling method for non-linear structural mechanics and thermal analysis, used for calculating the lifespan of high-pressure turbine blades and performing heat analysis of high-pressure compressors. The second technology involves learning physics simulations with non-parameterized geometrical variability using classical machine learning tools, such as Gaussian process regression. Finally, we present our contributions to the open-source and open-data community.