Research of Early Warning of Failure with Load Tendency Based on Non-intrusive Load Monitoring in Microgrid

摘要

As the prospective study of non-intrusive load monitoring (NILM) in load decomposition, the fault detection and classification is promising. After describing the structure of offshore oil platform power system connected with offshore wind farm, this paper presented a framework, for using NILM for fault detection and electricity behavior in offshore oil platform microgrid. The data acquisition from smart power meter was adopted to train the designed algorithm and strategy with GPU, the moving average convergence divergence strategy and differential value prediction line for the early warning of failure with installation load tendency was approached to solve the problem: “Early stopping–But when?”