Scaled Conjugate Gradient Backpropagation-Based Fault Analysis System for Induction Motor

A fault analysis system for induction motor based on scaled conjugate gradient backpropagation is proposed in this paper. An induction motor is simulated in MATLAB with the specifications. Faults are created in the simulated induction motor as a part of data generation process for the artificial neural network. The fault data is used for the supervised learning which is the input for the training. As mentioned SCGB neural network is used as a learning algorithm. The data is divided into training, testing and validation data. The training performance is presented in the paper. The faults considered are symmetrical as well as unsymmetrical faults. The detected results for line to ground, line to line, double line, double line to ground, triple line, triple line to ground faults are analysed and presented in this paper.