Advancing Electrical Machine Modeling with Physics-Informed AI

Dr. Nyangon has authored a new scholarly paper titled “Hybrid Physics-Informed Artificial Intelligence for High-Fidelity Modeling and Optimization of Electrical Machines,” recently submitted for publication in the Journal of Frontiers in Artificial Intelligence.

The paper presents a comprehensive review of how physics-informed machine learning (PIML) is transforming the design, monitoring, and control of electrical machines and drives. By integrating domain-specific physical laws with advanced machine learning techniques, PIML addresses critical challenges such as data scarcity, parameter sensitivity, and model robustness—while significantly improving interpretability and computational efficiency.

A central focus is placed on hybrid physics-informed neural network (PINN) architectures that enable high-fidelity modeling, real-time diagnostics, digital twins, fault detection, and optimization. The study highlights cutting-edge approaches including Deep Operator Networks, Fourier Neural Operators, Extreme Learning Machine–enhanced PINNs, Graph-Based PINNs, and domain decomposition methods.

Through state-of-the-art case studies and simulations, the paper underscores a paradigm shift from traditional black-box models toward transparent, physics-informed, and scalable AI solutions—positioning PIML as a key enabler for next-generation Industry 4.0 applications.