AI-powered physics model that can detect hidden magnetic patterns responsible for wasting energy inside electric motors. The study focuses on “magnetic hysteresis loss,” a process where energy is lost as heat when magnetic fields inside motors repeatedly change direction. This problem is especially important for electric vehicles, where improving motor efficiency could significantly reduce energy consumption and extend battery performance.
The research examined tiny magnetic structures known as “maze domains,” which appear inside soft magnetic materials used in motor cores. These complex, labyrinth-like patterns become unstable at high temperatures and can increase energy loss. Scientists have struggled for years to fully understand how these domains behave because they are influenced by multiple interacting factors, including heat, entropy, and microscopic material structures.
To solve this challenge, the research team created an explainable AI framework called the entropy-feature-eXtended Ginzburg-Landau (eX-GL) model. The system combines advanced mathematics, machine learning, and topological data analysis to study microscopic images of magnetic domains. By mapping these structures into a “free-energy landscape,” the AI was able to identify hidden energy barriers and explain how magnetic reversal processes occur inside materials as temperatures change.
The findings could have major implications for the future of electric vehicles and energy-efficient technologies. By understanding exactly how magnetic energy is lost, engineers may be able to design motors that generate less heat and operate more efficiently. Researchers also believe the AI framework could be adapted for studying other complex physical systems beyond magnetism, opening new possibilities for material science, electronics, and next-generation energy technologies.
