Fault diagnosis of DC motor based on parameter estimation and Fuzzy ARTMAP

In order to improve the motor quality and realize system monitor, a new approach to diagnose motor faults is presented to estimate the electromechanical parameters of motor based on the block-pulse function series and classify motor faults based on Fuzzy ARTMAP neural network. The electromechanical parameters of motors can be acquired on line without stopping and loading motor. Thus the running state of motor is obtained. Because of the strong pattern recognition ability of Fuzzy ARTMAP neural network, it is employed to diagnose the type and magnitude of faults. The effectiveness of the proposed method is verified by the results of simulations and experiments.