POWER GRID ANOMALY LOCATING METHOD BASED ON MAXIMUM EIGENVECTOR

A power grid anomaly locating method based on the maximum eigenvector. The method comprises the steps of: selecting sample nodes to construct a data source matrix; collecting state data at the moment when an anomaly is detected, and state data at TW-1 moments after the moment when the anomaly is detected, so as to construct a window matrix; performing data processing on the window matrix, so as to obtain a normalized non-Hermitian matrix; calculating a sample covariance matrix S of the non-Hermitian matrix, and obtaining the maximum eigenvector of the matrix S; calculating an anomaly locating index, and performing feature scaling-up processing on elements in the anomaly locating index, so as to obtain an improved anomaly locating index lmax; and performing comparative analysis on element values of lmax, and then outputting the serial number of an abnormal node. By means of the method, element distribution of the maximum eigenvector is more stable, such that the maximum eigenvector is not prone to being affect