【作者】 李志强；

【导师】 夏长亮；

【作者基本信息】 天津大学 ， 电机与电器， 2009， 博士

【摘要】 无刷直流电机具有调速范围宽、功率密度高、电磁污染小、运行可靠等显著优点,在伺服控制、电动车辆、机器人技术及家用电器等工业和民用领域得到了越来越广泛的应用。为进一步提高其性能,本文主要从无位置传感器控制、新型逆变器拓扑结构控制及智能控制等方面进行了深入研究。为提高无刷直流电机无位置传感器控制的可靠性,本文分析了线反电势与换相时刻对应关系,对传统反电势法进行了改进,研究了无需移相π/3电角度的无刷直流电机转子位置检测方法。通过总结线反电势变化规律、提出新的线反电势检测方法及对滤波误差相移补偿,实现了基于线反电势的无刷直流电机无位置传感器控制,并进行了仿真和实验验证。该策略硬件电路结构简单,算法易于实现,避免了传统反电势法在移相π/3电角度过程中带来的误差,提高了位置检测的快速性与准确度。随着无刷直流电机的广泛应用,越来越多的学者关注于维持电机高性能的同时降低驱动系统成本。针对无刷直流电机的四开关逆变器拓扑结构,本文分析了电机C相电流与电机工作模式的关系,通过重新设计控制策略,改进了对四开关逆变器控制需要检测两相电流的方法,实现了基于C相电流的四开关逆变器无刷直流电机单电流传感器控制。该策略在四开关逆变器基础上,进一步节省了成本,同时保证了电机的性能,为无刷直流电机控制技术的发展提供了新的思路。传统的PID控制算法在控制非线性系统时随动性差,不易满足高精度伺服系统的控制要求。本文研究了基于小脑模型神经网络与PID复合控制的智能控制系统,通过小脑模型神经网络强大的映射能力,来实现参数的在线调节,使无刷直流电机在不同的运行环境下都能快速、准确的跟踪给定指令。仿真与实验结果表明,控制系统兼顾了常规控制器快速性及智能控制器精确性的优点,转速超调量小,调节时间短,有效的抑制了负载扰动的影响,具有较好的静态和动态性能。电梯门机是无刷直流电机驱动系统的一个应用方向,为提高电梯门机的安全性与舒适性,本文设计了无位置传感器控制的无刷直流电机电梯门机控制系统。该系统以TMS320F2812 DSP数字信号处理器为控制核心,主要包括电源电路、驱动电路、逆变器、信号检测与保护电路、光耦隔离电路等部分。系统实现了电梯门机的门宽自学习、平滑变速门机运行曲线及遇阻安全保护等功能,节省了成本,减小了开关门噪音,提高了运行可靠性。更多还原

【Abstract】 Brushless DC (BLDC) motor has found widespread applications in the fields of servo systems, electrical vehicle, robot control and household appliance etc, due to its extensive speeding range, high power density, light electromagnetic pollution and high reliability. To further improve performance of BLDC motor drive, the thesis makes deep investigations on the following aspects, including intelligent control, sensorless control and new topological structure of inverter.To improve the reliability of sensorless BLDC motor drive, relationship between line back EMF and commutation instant is analyzed to improve ordinary back-EMF method, and a novel rotor position detection method withoutπ/3 phase angle shift is proposed. Position detecting circuit is improved, commutation rules are re-arranged and error from filtering is compensated. It can be drawn from simulation and experimental results that the method avoids the error introduced by phase shifting of measured signals, and improve precision and rapidity of position detection. .Recently, cost reduction of BLDC motor drive has drawn more and more attention of researchers. Four-switch inverter can reduce cost of drivers by replacing power devices with capacitors. This work designs four-switch inverter for three phase BLDC motor. The relationship between current of phase C and working mode of motor is analyzed. Control strategy based on two phase current measurement is improved. A novel control strategy based on single current sensor is proposed and implemented for four-switch three phase BLDC motor drive. The proposed method achieves good performance with reduced cost, which proves it a new valuable research direction.Despite its simplicity and good steady performance, traditional PID controller could hardly meet the requirement of precision and fast responding when used in controlling nonlinear system. In this work, a novel approach of compound control for BLDC motor using Cerebella Model Neural Network controller and PID strategy is proposed and researched. Based on strong mapping capability of neural network, the parameters of controller can be adjusted online so that the motor speed can track reference value very fast and exactly under different working conditions. Simulation and experiment results prove that the proposed intelligent controller has excellent dynamic and steady performance with very short setting time, extremely reduced overshot, and dramatically restrained effect of load disturbance. Elevator door drive is one of important applications of BLDC motor. A sensorless BLDC motor control system is designed for it in this work. The system uses TMS320F2812 DSP as core control unit, and includes power circuits, drive circuits, inverter, signal measurement circuits, protection circuits, optically coupled isolation circuits, and so on. The major functions, such as self-learning of door width, smooth track of speed reference curve, and protective action when blocked, are totally implemented with reduced cost, low noise, and high operational reliability.更多还原