Developing the Next Generation Power Electronics for Emerging Energy Systems

Power electronics is the backbone of future energy systems, including data centers, EVs, and grid-scale energy storage. These high-impact applications demand increased efficiency, density, and dynamic performance from power electronics systems. However, traditional power converters favor simple topologies with large, lumped switching cells and discrete magnetics, relying heavily on bulky passive components for energy storage. These designs fail to leverage the rapid advancements in semiconductor materials and are approaching fundamental performance limits. As wide-bandgap devices emerge and new opportunities arise in evolving energy systems, granular power electronics with integrated magnetics are becoming increasingly compelling.

In pursuit of this vision, this talk will first introduce a systematic matrix coupling approach to merging multiple magnetic components into one. Two distinct architectures are then developed to tackle two critical challenges for emerging energy systems: (1) how to deliver a massive amount of current to a tiny area (e.g., AI computing processors) with extreme power density (>1A/mm2); and (2) how to deliver power to a massive number of modular loads (e.g., data center servers, batteries, solar cells) with extreme energy efficiency (above 99%). The matrix coupling theory and the two proposed architectures are enlightening the path to extreme-performance granular power electronics systems that will benefit a broader range of applications. Besides, machine learning methods for MIMO power flow control and magnetics core loss modeling will be explored, fostering a synergistic cycle of “Power for AI and AI for Power.”