Artificial neural network approaches for fault classification: comparison and performance

作者：Nagpal Tapsi^*; Brar Yadwinder Singh

来源：Neural Computing & Applications, 2014, 25(7-8): 1863-1870.

DOI：10.1007/s00521-014-1677-y

• Power transformer
• Probabilistic Neural Network
• Backpropagation Neural Network
• Fault Diagnosis

This manuscript focuses the implementation of artificial neural network-based algorithms to classify different types of faults in a power transformer, meant particularly for NonDestructive Test for transformer fault classification. The performance analysis of Probabilistic Neural Network (PNN) and Backpropagation Network classifiers has been carried out using the database of dissolved gases collected from Punjab State Electricity Board, Patiala, India. Features from the preprocessed data have been extracted using dimensionality reduction technique, i.e., principal component analysis. The selected features were used as inputs to the Backpropagation Network and PNN classifiers. A comparative study of the two intelligent classifiers has been carried out, which reveals that PNN classifier outperforms the Backpropagation Network classifier.

• 出版日期2014-12