Fault Diagnosis of Power Transformer Based on Multi-La yer SVM Classifier
Lü Gan-yun, CHENG Hao-zhong, DONG Li-xin, ZHAI Hai-bao (Department of Electrical Engineering,Shanghai Jiaotong University, Shanghai 200 030, China)
Support Vector Machine (SVMs) is a novel machine learning method based on statis tical learning theory (SLT). SVM is powerful for the problem with small sample, nonlinear and high dimension. A multi-layer SVM classifier is applied here to f ault diagnosis of power transformer. Through a special data dealing process, con tents of five characteristic gases obtained by DGA are transformed, and 6 charac teristic components for fault diagnosis are distilled for SVMs. The multi-layer SVM classifier, trained with the sampling data from the above dealing process, identifies out the four types of transformer states. The test results show that the proposed classifier has an excellent performance on training speed and corre ct ratio.