Research on power grid anomaly detection based on high-dimensional random matrix theory

Abstract

Power grid anomaly detection is a basic task of the power system, and the evaluation results are of great significance for the power grid to find out the potential dangers of operation as soon as possible and take timely measures to prevent them from happening. In recent years, with the rapid development of green energy and high technology, the power grid is developing towards the smart grid. A large number of measurement devices are deployed in the power grid, and the power grid operation data is collected, transmitted and analyzed in real time. The data-driven power big data analysis approach emerges as the times require. In this paper, a data-driven approach is proposed. It constructs a high-dimensional random matrix based on power system operating data, uses the sliding window method and the random matrix single-ring theorem to continuously analyze the high-dimensional matrix, and introduces the mean spectral radius (MSR) as a specific evaluation index to realize the power grid operation situation assessment. The simulation data test results of IEEE 39 bus system verifies the feasibility and correctness of the approach for evaluating the operation situation of the power grid. This developed approach can identify abnormal conditions at the early stage of system failure, and has great guiding significance for fault inspection of power grids.