ANFIS-based diagnosis and location of stator interturn faults in PM brushless DC motors

An automatic scheme for fault diagnosis and location of stator-winding interturns in permanent-magnet brushless dc motors is presented. System performances under healthy and faulty operation are obtained via a discrete-time model. Waveform of the electromagnetic torque is monitored and processed using discrete Fourier transform and short-time Fourier transform to derive proper diagnostic indices. Two adaptive neuro-fuzzy inference systems (ANFIS) are developed to automate the fault diagnosis process. Test results show an acceptable performance for ANFIS in detecting the fault.