Stator fault estimation in induction motors using particle swarm optimization

The use of induction motors is extensive in industry. The working conditions of these motors make them subject to many faults. These faults must be detected in an early stage before they lead to catastrophic failures. This paper presents a scheme for detecting inter-turn faults in the stator windings of induction motors and estimating the fault severity. Detection of incipient inter-turn faults prevents further insulation failure. The proposed algorithm monitors the spectral content of stator currents to detect the fault. After the fault is detected and identified, a particle swarm approach is used to estimate the fault severity. The swarm estimator update is based on the error between the measured data and a complete model of the faulty motor. An experimental setup is used to validate the developed scheme and to implement an online fault detector.