Research on the Disturbance Detection Method Based on Random Matrix Eigenvalue

Traditional anomaly detection has some shortcomings, such as time delay, low sensitivity and lack of overall situation. The wide area measurement system (WAMS) and synchronous phase measurement unit (PMU) that can be widely applied can improve the anomaly detection level of distribution network. The maximum and minimum eigenvalue method of random matrix is first applied in the field of cognitive radio to detect weak signals in radio networks. An algorithm for anomaly detection of distribution network based on the maximum and minimum eigenvalue method is proposed in the paper. The method adopts the global sampling data of PMU to detect anomaly of distribution network in real time, so as to improve the detection sensitivity. The algorithm and its threshold are deduced through theoretical analysis. The effectiveness and feasibility of the method are verified by simulation of short-circuit anomaly and harmonic anomaly. Case analysis shows that this method can detect disturbance signal quickly, sensitively and accurately, and has good robustness.