A New Vibration Analysis Approach for Monitoring the Working Condition of a High-Voltage Shunt Reactor

With the rapid development of power grids, the voltage level and capacity of high-voltage shunt reactors (HVSRs) are increasing year by year, and HVSR faults are increasing, especially HVSRs core winding faults. The HVRSs core winding faults seriously threaten the safe and stable operation of power grids. To solve this problem, we propose a feature extraction method for HVSR core winding faults identification to improve the accuracy of faults identification. This method relies on the combined application of qualitative analysis of phase space reconstruction (PSR) and quantitative calculation of an improved K-means clustering method. First, we employ PSR to qualitatively analyze the HVSR vibration signal and extract the phase trajectory feature quantity of the vibration signal. Then, this paper uses an improved K-means clustering optimized by an improved grasshopper optimization algorithm (GOA) to cluster the phase trajectory features, so as to obtain the cluster center coordinates. Furthermore, we employ the distance from the cluster center displacement vector to the origin and the angle change to perform quantitative calculations to realize fault identification. Finally, the fault simulation experimental data sets of 10kV and 20kV HVSRs verify the effectiveness of the proposed method. The experiment results show that the proposed method has better accuracy and can truly reflect the state characteristics of the HVSR core winding. The proposed high-accuracy method helps to improve the efficiency of on-site HVSRs condition assessment and maintenance.