Figures from this paperAbstractFigures from this paperAbstractFigures from this paperAbstractFigures from this paper

Abstract

Transformer undertakes the heavy responsibility of power transmission and transformation in the power system, and once a failure occurs, it will seriously affect the safe and reliable operation of the power system. Based on the data of characteristic gas dissolved in transformer oil, this paper uses kernel principal component analysis method to detect transformer fault. Firstly, the gas data is normalized and a Gaussian kernel function is selected to map the normalized data to a high-dimensional feature space, and then a fault detection model is built using the off-line data for transformer fault detection. The simulation results show that the detection accuracy of T2 statistic and SPE statistic is 100 percent and 98.442 percent respectively, which is higher than that of traditional principal component analysis.