A review of recent AI applications in next-generation power electronics
Ashkan Safari, Arman Oshnoei, Frede Blaabjerg
Abstract
Power electronics (PELS) is an important part of modern technology by enabling energy conversion, control, and management, which is essential for powering renewable energy systems (RES), electric vehicles (EVs), industrial automation, and advanced consumer electronics, thereby driving sustainability and innovation. Consequently, the efficiency and performance of PELS components are always in priority. To this end, the potential applications of generative and non-metaheuristic AI-driven algorithms in the control, maintenance, and design of PELS systems are reviewed in this paper. Covering a spectrum of models, including generative adversarial networks (GANs), neural networks (NNs), quantum neural networks (QNNs), fuzzy interfaces, and reinforcement learning (RL), this paper investigates their applications in control, maintenance, and design of PELS. The review also provides the implications of big data, generative, and non-metaheuristic AI-driven algorithms in efficiency enhancement, reliability, and sustainability of PELS-integrated energy infrastructure. Also, some of the recently funded projects in the U.S. and Europe, as well as the associated challenges and future research directions, are investigated in this emerging sector.