Identifying three-phase induction motor faults using artificial neural networks

Artificial neural network based fault identification scheme implementation for a three-phase induction motor

“…The main kinds of faults include overloading, single phasing, unbalanced voltage, locked rotor, ground fault, undervoltage, and overvoltage. A detailed discussion of these faults is given in Kolla and Varatharasa [10]. From this information it can be noted that the motor voltages and currents have distinct characteristics during these faults.…”

“…The three-phase voltage and current information is used in identifying faults using ANNs in this paper. While the applicability of the ANN method for identifying various simulated faults is demonstrated in [10], the present work implements the ANN scheme and tests it in real time for those faults that do not damage the motor.…”

“…The motor current and speed were used as inputs. Kolla and Varatharasa used ANNs for fault identification in a three-phase induction motor [10]. They used software simulation to produce data to train an ANN to determine some of the main faults in an induction motor.…”

“…A 1/3 HP squirrel-cage, three-phase induction motor is used for this purpose. Section 2 reviews concepts of induction motor faults and artificial neural networks, and previous methods of ANN use for fault identification [10]. Details of the proposed hardware system are described in Section 3.…”

Motor Current Demodulation Analysis applied with Neural Networks and Genetic Algorithms for Rotor Bar Faults Diagnosis

“…Many researchers [12][13][14][15][16][17] dealing with MCSA techniques for motor diagnosis have been established in basis of using Fast Fourier Transformation (FFT) because it requires a simple and inexpensive calculation tools, this spectrum technique is mainly used to detect fault signature components; however, this technique have shown no accurate results with non-loaded motors or with frequency disturbance caused by motor vibrations that affects greatly fault components in current spectrum. This is way many research [18][19][20] have been conducted to overcome this weakness by Motor Current Demodulation Analysis technique.…”