论文信息 - Research on Optimization of Feature Extracting Based on PD Fingerprints in Pattern Recognition

Research on Optimization of Feature Extracting Based on PD Fingerprints in Pattern Recognition

In order to investigate the effect of feature extraction on pattern classification for partial discharge (PD) signals appearing potential insulating failures in high voltage apparatus while operation, the PD-fingerprints acquired by Hipotronics DDX-7000 digital PD detector are taken as database and method on the feature extracting from the database is carried out, and its practicability is demonstrated by the mathematics and experiment, respectively. The BP neural network is made of three layers and the transfer function of hidden layer and output layer are tansig, then the influence of the structure of neural network on recognition results is studied at the same time. As a result, the optimal characteristic vector with obvious separability and the number of hidden layer are obtained, and achievements of research show that the network convergence is not only quickly, but also the recognition rate very high so much as up to 100%.

[1] E. Gulski. Discharge pattern recognition in high voltage equipment , 1995 .

[2] Kai Gao,et al. PD pattern recognition for stator bar models with six kinds of characteristic vectors using BP network , 2002 .

[3] E. Gulski,et al. Classification of partial discharges , 1993 .

[4] Sergios Theodoridis,et al. Pattern Recognition, Fourth Edition , 2008 .

[5] E. Gulski,et al. Computer-aided recognition of discharge sources , 1992 .

[6] Birsen Yazici,et al. Statistical pattern analysis of partial discharge measurements for quality assessment of insulation systems in high voltage electrical machinery , 2003, 4th IEEE International Symposium on Diagnostics for Electric Machines, Power Electronics and Drives, 2003. SDEMPED 2003..