Automatic diagnosis and location of open-switch fault in brushless DC motor drives using wavelets and neuro-fuzzy systems

The faulty performance of permanent-magnet (PM) brushless dc motor drives is studied under open-switch conditions. The wavelet transform is used to extract diagnostic indices from the current waveform of the motor dc link. An intelligent agent based on adaptive neuro-fuzzy inference systems (ANFIS) is developed to automate the fault identification and location process. ANFIS is trained offline using simulation results under various healthy and faulty conditions obtained from a lumped-parameter, network model. ANFIS testing shows that the system could not only detect the open-switch fault, but also identify the faulty switch. Good agreement between simulation results and measured waveforms confirms the effectiveness of the proposed methodology.