Power Electronics process 80% of electricity generated and utilized today, being a part of the critical infrastructure of green transition. Since the 1990s, academic research in power electronics has started focusing on AI-based methods. There is still a lack of practical implementation. AI-Power will bridge the gap to industry applications by developing tailored physics-informed AI tools for energy efficiency boost and life-cycle-cost reduction of power electronic systems. It will consolidate the competitive edge of the energy sector and bring new opportunities in CO2 reduction.

Key Drivers

• Energy efficiency
• Life cycle cost
• Time-to-Market

Key Trends

• Product+service
• Data+physics
• Digitalization

The evolving electrical energy systems require new principles and operational practices. It, in turn, means we must develop and implement new and updated system tools. AI-Power project strives to address three significant challenges. It will develop tailored physics-informed AI tools for the design, operation, and maintenance of power electronic systems.

• How to obtain and understand the most relevant database?
• How to develop computational-efficient and affordable AI solutions?
• How to quantify uncertainty and interpret the results from AI?

AI-Power will bridge the gap to industry applications by developing tailored physics-informed AI tools for energy efficiency boost and life-cycle-cost reduction of power electronic systems. It will consolidate the competitive edge of the Danish energy sector and bring new opportunities in CO2 reduction. It involves four industry partners: Xtel (IoT), Neurospace (machine learning), Danfoss (power electronics), and Nordic Ultra Capacitor Engineering Research Center (e-mobility) and two universities. It estimates that the outcomes will create 200 million DKK business opportunities and 100 jobs within 3-5 years after the project, increasing to two billion DKK with 1,000 jobs in the long term.