A Survey of AI in System of Systems With a Focus on Power Electronic Systems—Part II: Maintenance and Forecasting

Artificial intelligence (AI) is increasingly pivotal in modernizing the lifecycle of power electronic systems (PESs). Extending the discussion on the design and control phases presented in Part-I of the survey, Part II focuses on the transformative role of AI in the maintenance and forecasting phases of PES. Specifically, we analyze how AI techniques are applied to power electronic components, power electronic converters (PECs), and their emerging applications. This survey analyzes how AI techniques monitor PES’s health by detecting faults/anomalies, and estimating the remaining useful life (RUL), and in turn enable fault prevention before complete failure. Distinct from prior surveys, it also discusses soft-fault detection in complex wired networks such as aircraft, representing a system-of-systems application. In the context of forecasting, AI enables accurate prediction of PES parameters, such as power efficiency, thermal behavior, components’ lifetime, load demand, and energy generation. This survey identifies open challenges and provides forward-looking insight to guide researchers and engineers in developing robust, efficient, and autonomous maintenance and forecasting solutions for PES.