Self-tuned NFC based speed ripple minimization of a faulty induction motor

This paper presents a self-tuned neuro-fuzzy controller (NFC) based speed ripple minimization of a vector controlled faulty induction motor (FIM) with broken rotor bars. First, the performance of the FIM is investigated in terms of speed ripple under the open-loop condition. Then, a new mechanical model of induction motor is developed incorporating the speed ripple. Based on this model a new NFC is proposed to tolerant the effect of the fault under an indirect field oriented control scheme. The proposed NFC compensates the faulty condition by minimizing the supply frequency related speed ripples instead of directly working on the low frequency signature speed ripples which a FIM exhibits. Based on the knowledge of motor control and intelligent algorithms an unsupervised self-tuning method is developed to adjust weights of the proposed NFC. The convergence of the weights is also discussed. The complete drive is experimentally implemented using a digital signal processor board DS-1104 for a laboratory 250 W faulty IM. The effectiveness of the proposed NFC is tested both in simulation and experiment.