ECPE White Paper 'Artificial Intelligence in Power Electronics'

This is the authors' abstract. We don't add key points for this paper.

The White Paper was prepared by an ECPE Task Force comprising of almost 40 renowned international experts from industry and academia and is intended to promote dialogue within the power electronics community. AI is changing how we live and work. The ECPE network therefore asked what impact AI already has on power electronics and where its future potential lies. The White Paper aims to raise awareness, provide a platform for discussion within the community and strengthen collaboration between industry and academia. AI's key potential lies in speed and, with it, cost: shorter development and testing times for faster time-to-market, optimized production with fewer failures and rejects, and intelligent condition and service-life monitoring in the field. The vision is not a fully automated engineering process but an AI-assisted engineer with outstanding work efficiency.The White Paper concentrates on the use of AI methods and tools in power electronics, not on the role of power electronics in AI data centres. AI is looked at through three lenses: pattern recognition with machine learning and deep neural networks, large language models as design assistants, and reinforcement learning for exploratory converter development.