A novel MRAS-based adaptive observer for sensorless induction motor drive

Sensorless induction motor (IM) drive systems are widely used in industrial applications for their reliability, robustness and flexibility particularly in harsh work environments. Previously developed speed estimators using observer-based techniques had steady state speed disturbances and poor performance at very low speed range of IM drives. A novel model reference adaptive system (MRAS) is proposed for sensorless induction motor drives to achieve high performance at very low speeds as well as zero-speed. In this proposed scheme, an adaptive pseudoreduced-order flux observer (APFO) is used instead of the adaptive full-order flux observer (AFFO). The stator currents are estimated using a closed-loop observer with a [2 x 2] feedback gain matrix, whereas the rotor fluxes are estimated by an open-loop observer. In comparison with AFFO, the proposed APFO method requires less computational burden and estimation time and provides better speed estimations especially at very low and zero speeds. In addition, the stator resistance under a wide range of variations is estimated using the same observer. The convergence speed for the stator resistance estimator is higher than of the conventional AFFO based methods.; In addition a pole-placement technique is also studied and designed, such that the observer poles are fully independent of motor running speed, thus extending the operational speed range of the sensorless induction motor drive from very low to very high speeds.; The MRAS based APFO has been evaluated first by computer simulations using the induction motor d-q models in both stationary and rotational reference frames. The proposed APFO has also been fully validated by laboratory experimental studies. The APFO system is implemented on the PC interfaced with a 'LAB Windows/CVI' I/O acquisition board. A microcontroller based three-phase IGBT inverter is used to drive a 10 hp direct field-oriented induction motor. Both the experimental and simulation results for the sensorless field-oriented induction motor drive have demonstrated a close agreement between the estimated and the actual speeds at both very low speeds as well as zero speed.; This thesis fully investigates the principles, models, simulation, and experimental laboratory results of the developed MRAS based pseudoreduced-order flux observer for sensorless field-oriented induction motor drive.