Gradient-Informed Machine Learning in Electromagnetics

TL;DR

This paper introduces a gradient-informed surrogate modeling approach combining Isogeometric Analysis, proper orthogonal decomposition, and Gaussian process regression to efficiently simulate nonlinear electromagnetic devices.

Contribution

It presents a novel non-intrusive surrogate modeling framework that leverages the differentiability of IGA for efficient sensitivity analysis in electromagnetic simulations.

Findings

Effective reduction in computational cost for nonlinear models

Accurate surrogate models for parametric electromagnetic devices

Enhanced sensitivity analysis capabilities

Abstract

Simulation techniques such as the finite element method are essential for designing electrical devices, but their computational cost can be prohibitive for repeated or real-time computations. Projection-based model order reduction techniques mitigate this by reducing the model size and complexity, yet face challenges when extended to nonlinear or non-affine parametric models. In this work, Isogeometric Analysis (IGA) is combined with proper orthogonal decomposition and Gaussian process regression to construct a non-intrusive surrogate model of a parametric nonlinear model of a permanent magnet synchronous machine. The differentiable nature of IGA allows for computationally efficient extraction of parametric sensitivities, which are leveraged for gradient-enhanced surrogate modeling.