Research on a Position-Sensorless IPM Motor Drive System Using Position Estimation Based on Magnetic Saliency

点击翻译按钮获取中文摘要

This research aims to build up a position-sensorless IPM motor drive system (3.7 kW) applying a position estimation proposed by the head investigator, and to study characteristic of the drive system.First of all, we have design and build up an experimental drive using a 3.7-kW interior permanent magnet synchronous motor (IPM motor). A harmonic current detection circuit, a PWM signal generator, and an interface circuit connecting a digital signal processor (DSP) have been developed. The circuits are characterized by a sophisticated configuration which can automatically acquire ripple currents generated by a voltage-source PWM inverter to estimate the rotor position of the IPM motor. The acquisition system is implemented in programmable logic devices (PLDs). Moreover, it is shown theoretically and experimentally that introduction of weighting coefficient into the PWM method proposed by the head investigator can enlarge the output voltage of the inverter by 15%.On the other hand, based on … More simulations using MATLAB/Simulink, it has been investigated that the control performance of the drive system is different whether the control system has current minor loops or not. As a result, it is shown that the d-axis motor current can not be adjusted to zero even if the d-axis voltage reference is set to zero, and that addition of the current minor loops to the control system improves not only current response but also stability of the position and speed control systems. In addition, a simulation method, which can simulate the drive system accurately and efficiently, is developed using the MATLAB/Simulink. The simulation method is characterized by assumption that a digital-controlled PWM inverter corresponds to discrete time systems with variable sampling time. Simulations in five conditions have confirmed validity and effectiveness of the method. The simulation method makes it possible to eliminate current and voltage errors, and execution time in the method is about 100 times faster than that of the conventional method. Less