电路引用Principal component and hierarchical cluster analysis as applied to transformer partial discharge data with particular reference to transformer condition monitoring.pdfVIP

2008 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 23, NO. 4, OCTOBER 2008 Principal Component and Hierarchical Cluster Analyses as Applied to Transformer Partial Discharge Data With Particular Reference to Transformer Condition Monitoring Tadeja Babnik, Member, IEEE, Raj K. Aggarwal, Senior Member, IEEE, and Philip J. Moore, Senior Member, IEEE Abstract— This paper analyses partial discharges obtained by re- PD activity can provide an early indication of an incipient fault mote radiometric measurements from a power transformer with condition. a known internal defect. Since fingerprints of remote radiometric PD detection methods generally fall into two categories: elec- measurements are not available, the formation of clusters with sim- ilar features obtained from captured partial discharge data is cru- trical methods [1]–[5] and acoustic methods [6]–[8]. Depending cial. Hierarchical cluster analysis technique is used as a method for on the sensors being employed, both methods can be further grouping different signals. Investigation based on Euclidian and classified into two sub-categories: non-intrusive monitoring and Mahalanobis distance measures and Ward and Average linkage al- gorithms were performed on partial discharge data pre-processed intrusive monitoring. Non-intrusive monitoring uses sensors by principal component analysis. As a result of the analysis, a clear installed outside the transformers, such as coupling capacitors, separation of partial discharges emanating from the transformer high frequency current transformers, piezoelectric type ac- and discharges emanating from its surrounding is achieved; thi