Research on Non-Intrusive Load Identification Method Based on Edge Computing

This study proposes a non-intrusive load identification method based on edge computing. By constructing a repository of load steady-state characteristics and employing temporal and spectral feature analysis techniques, precise identification of operational states of various electrical appliances in household or industrial environments is achieved. The study introduces a load monitoring technology roadmap tailored for edge computing frameworks and develops a platform for load equipment analysis and testing. Furthermore, the paper explores load identification methods based on statistical and machine learning algorithms (such as XGBoost and enhanced AlexNet network), validating their effectiveness and reliability through experiments in practical applications. This research not only offers novel insights into load management for smart grids but also advances the application of edge computing technology in the energy sector.