Hybrid physics-informed artificial intelligence for high-fidelity modeling and optimization of electrical systems

The M.Tech project investigates the potential of hybrid physics-informed machine learning (PIML) approaches for the modeling and optimization of electrical machines and drives. The research aims to capture the potential of machine learning integrated with domain-specific physics to overcome problems like the scarcity of data, improve the interpretability of models, and ensure compliance with the fundamental laws of physics. The focus is on developing models that are both efficient and accurate for advanced electrical systems. The project will analyze different types of hybrid physics-informed neural networks (PINNs) such as Deep Operator Networks (DeepONets), Fourier Neural Operators, and Extreme Learning Machines (ELM) augmented with PINNs. These kinds of models are pivotal

in enabling real-time diagnostics, creating digital twins, fault detection and control and optimization of processes with high degrees of fidelity. The study intends to show the emerging shift in paradigm from traditional opaque AI methods to transparent and physics-informed approaches to offer reliable and scalable solutions for Industry 4.0 prospects of electrical drives and machines.

Message: Hi FE, I need help with “Hybrid physics-informed artificial intelligence for high-fidelity modeling and optimization of electrical systems” in “Electrical Engineering”