A sensorless induction motor drive using a least mean square speed estimator and the matrix converter

摘要

This paper presents a novel least mean square (LMS) estimator for a sensor-less drive of a three phase induction motor. Also, the proposed system includes the use of the matrix converter instead of the two-level inverter to improve the estimator performance. The system studied consists of a three phase induction motor driven by a matrix converter, a hysteris current controller, and an indirect field oriented controller. Also, a proportional plus integral speed controller and the proposed speed estimator are used. The LMS estimator includes a step size factor (SSF). The value of the SSF affects the dynamic performance of the speed estimator. Different simulation results of the overall system are conducted to depict the dynamic performance of the induction motor including the effect of the SSF, used in the LMS estimator. The simulation results show that the low value of the SSF gives high estimation accuracy when the actual speed is nearly constant. Meanwhile, the estimated motor speed follows the actual value with a higher time lag during the speed change. On the other hand, the high step size value reduces the time lag during the speed change but reduces the estimation accuracy during the steady state. A variable LMS SSF is introduced to achieve both advantages of the low and high SSF. The simulation results show improvement in the dynamic performance of the estimator regarding the low time lag during the speed change and the high estimation accuracy during the steady state. The simulation results show excellent of the proposed estimator using LMS with matrix converter driven by variable SSF in reference speed tracking. The main advantage of this proposed estimator is reducing the mathematical calculation time while maintaining high estimation accuracy. This leads to using a slower processor and smaller memory which reduces the drive cost.