New Method for Transformer Fault Diagnosis Using Probabilistic Neural Network Based on Principal Component Analysis

Power transformers has an important role in electrical power transmission and its interruption has financial losses, thus its condition monitoring is essential andperformance of this equipment is effective for power system reliability. In this paper, proposed method has advantages of both principal component analysis andprobabilistic neural network. Firstly, main feature is extracted from primary and secondary three phase currents and searchcoils differential voltage by wavelet transform. Then, after decomposition of signals, redundant data is reduced by applying PCA. Finally, this information isused as probabilistic neural network inputs. Diagonal elements of confusion matrix proves appropriate performance oftransformer fault diagnosis usingprobabilistic neural network based on principal component analysis . Classification results show after six level decomposition ofsignals by mother wavelet ‘db5’ , 99.11% of samples are correctly classified.

کلیدواژه ها:

fault diagnostic ، transformer ، internal fault