Disturbance Signal Recognition Method of Power System Based on Constrained Fuzzy Clustering

Traditional methods for identifying disturbance signals in power systems can only solve the single disturbance problem, and can not identify multiple disturbance signals efficiently. To avoid the disadvantages of traditional methods, a method for identifying disturbance signals based on constrained fuzzy clustering is proposed. By calculating the envelope average value of the fitting signal, the difference between the fitting signal and the envelope average value is obtained. The difference is used as a new fitting signal to obtain the minimum eigenvalue component. The signal can be decomposed into several different eigenvalue functions by mode decomposition. The signal after mode decomposition meets the routing information protocol standard. Based on the new sparse vector, the aliasing disturbance signal is extracted. Fourier transform is used to describe the basic change of disturbance signal, and the order derivative of disturbance signal in power system is calculated. The transformation parameters in transmission process are obtained. The spatial coordinate system of acquisition points is obtained. Based on the coordinate system, the optimal window criterion is selected, and the disturbance signals are clustered to the central position by the constrained fuzzy clustering method. By establishing weight coefficient matrix, setting iteration times, and error compensation, new clustering centers are obtained to effectively identify the noise disturbance signals. The experimental results show that the maximum recognition accuracy of this method can reach 98%, which provides support for the normal operation of power system.