New adaptive vector control for induction motors and its potential application for motor fault detection

This paper presents a new scheme of adaptive vector control for inverter-fed induction motors, which can identify crucial motor parameters in a real-time manner. The total system has a novel configuration consisting of two subsystems, a vector control system in the synchronous frame and an adaptive identification system in the stationary frame. The proposed adaptive identification system is new and can adaptively and simultaneously identify two motor parameters varying dynamically with temperature, magnetic saturation or motor faults, which are related directly with control performance and possibly with real-time motor fault detection and diagnosis. High performance of the proposed scheme is verified through practical evaluation tests of the adaptive vector control system. Variation of torque generated by the system to motor frame temperature variation of about 50°C is only 0.5%. Linearity of torque generation up to 300% of torque rating is ensured. Deviation from ideal response is almost constant up to 300% of torque rating, whose magnitude is only about 1% of torque rating. Inverse of rotor time constant, which is one of the identified parameters for high control performance, seems to present also straight and valuable information on rotor-bar breaks where rotor resistance increases, inductance decreases and consequently the change of inverse of rotor time constant as their ratio is much enhanced