Mechanical Condition Monitoring Of On-load Tap-changer Based On Vibration Analysis

On-load Tap-Changer?OLTC?is the key component and the only movable part of the power transformer for voltage regulation,and its operation status directly affects the safe and reliable operation of the transformer and the power system.According to statistics,transformer faults caused by abnormal state of OLTC account for more than 20%of total faults,and the mechanical faults of OLTC account for up to 90%.Therefore,it is necessary to research mechanical performance monitoring of OLTC,this is of great significance to ensure the safe and reliable operation of OLTC and power transformer.Since the vibration signal generated by the impact of static and dynamic contact during the switching of OLTC gear is closely related to the mechanical performance of the switch body,this paper firstly builds a vibration characteristics simulation test platform based on a CM-type OLTC.By studying the simulation method of typical mechanical faults of OLTC,the vibration signals of OLTC under normal and typical mechanical faults are tested and analyzed,which provides important data support for subsequent research.Aiming at the time-domain waveform characteristics of the vibration signal during OLTC switching,this paper uses the combination of adaptive morphological combinatorial filtering and Morlet wavelet transform to extract the characteristic index K and carry out statistical analysis.Meanwhile,introduces the operating state coefficient E_co to quantitatively describe the statistical results.The results show that the consistency of the vibration signal during OLTC switching is good,E_co is about 9.4.When OLTC has mechanical faults,E_coo decreases obviously,indicating that the vibration characteristics of OLTC are significantly different between normal condition and mechanical fault condition.Based on this,the abnormal state of OLTC can be accurately identified,but E_co is limited in distinguishing different mechanical fault.Due to the low-dimensional chaotic characteristics of the vibration signal during OLTC switching,a Density-Based Spatial Clustering of Application with noise?DBSCAN?algorithm is proposed to study the dynamic characteristics of OLTC vibration signals in high-dimensional space.DBSCAN clustering is performed on the vibration signal of OLTC by reasonably setting the threshold and distance threshold.The discrete points of the signal are defined according to the clustering result to analyze the vibration characteristics of OLTC under different operating conditions.The results show that the discrete points of OLTC vibration signals in different operating states have significant differences,which indicates that the proposed OLTC vibration signal clustering algorithm and the proposed feature index are correct,according to which the different operating states of OLTC can be accurately identified.In addition,with the gradual popularization of vacuum OLTC,in order to study the monitoring technology of vacuum OLTC operation status,based on the vibration characteristics of vacuum OLTC switching,this paper uses DBSCAN algorithm to calculate and analyze the vibration signals of OLTC switching,and compares the differences of the discrete points and action points between vacuum OLTC and oil arc extinguishing OLTC after clustering vibration signals.The results show that the vibration signal of vacuum OLTC switching has the characteristics of multi-pulse distribution and long pulse duration,and the number of discrete points and total action points in phase space is more than that of oil arc extinguishing OLTC.The conclusions can provide reference for the research of mechanical condition monitoring technology of vacuum OLTC.