Journal International Association on Electricity Generation, Transmission and Distribution
Year : 2010, Volume : 22and23, Issue : 1
First page : ( 33) Last page : ( 37)
Print ISSN : 2229-4449.

Alternator fault diagnosis using artificial neural network

Thoke A.S.¹, H.OD., Jain Anamika², Assistant Engineer

¹Department of Electrical Engineering, NI.T Raipur, C.G. India, (e-mail: asthoke@yahoo.co.in)

²Chhattisgarh State Electricity Board, Dangania, Raipur C.G, India (e-mail: anamikajugnu@vahoo.com).

Abstract

Artificial Intelligence (AI) techniques are finding increasing applications in the field of power system operation, control and protection. The paper reports application of Aliificial Neural Network (ANN) for fault diagnosis/detection of large number of faults and abnormal operating conditions of an Alternator (stator/rotor side). Feedforward Multilayer Neural Network architecture with Backpropagation of error has been used in the study. MATLAB and its NEURAL NET toolbox have been used. Effects of num bel' of hidden layers, learning rates and training algoritluns are reported. The number of neurons in input, hidden and output layers is 10, 4, and 10 respectively. Ten types of faults/abnormal operating conditions were considered and were classified correctly. The results are based on offline studies.

Top