【作者】 黄招彬；

【导师】 游林儒；

【作者基本信息】 华南理工大学 ， 检测技术与自动化装置， 2014， 博士

【摘要】 作为垂直升降的交通工具，电梯在高层建筑和公共场所已经成为重要的建筑设备而不可或缺。电梯曳引系统的控制性能集中体现在起动溜车和运行抖动的抑制性能。在起动阶段一般采用零位置伺服控制，在电梯制动器打开瞬间，突加的负载扰动会造成电梯溜车；而起动零位置伺服阶段的突加负载扰动抑制性能，主要受转子磁极初始位置辨识精度、起动过程的转子位置与速度检测精度和控制算法影响。在运行阶段采用S曲线给定的速度伺服控制，低速阶段的速度跟踪性能主要受转子位置与转速检测精度影响，而高速阶段的速度跟踪性能还受控制环输出电压饱和影响。为保证电梯运行的控制性能，首先必须应对转子磁极初始位置辨识精度、低速阶段速度检测精度和高速阶段电压饱和问题。因此，本文的研究内容主要包括以下三个方面：(1)研究永磁同步电机转子磁极初始位置辨识精度的提高方法；(2)研究低速阶段转子位置与速度的检测优化方法；(3)研究提高母线电压利用率的脉宽调制优化算法，增加控制环输出电压调节范围。为了提高转子初始位置的辨识精度，本文介绍了三种经典的静止型转子初始位置辨识方法，包括脉冲电压注入法、高频旋转电压注入法和高频脉振电压注入法，分析了三种方法辨识过程中基于摩擦模型的运动情况，给出了转子静止的边界条件，提出了超出边界条件的转子运动抑制方法。通过抑制辨识过程中转子的运动，提高转子初始位置的辨识精度。同时，针对无凸极或者凸极率较低的永磁同步电机，提出了基于动态磁阻差异的静止型转子初始位置辨识方法，通过注入高频电压信号、提取二次谐波电流分量中转子位置与极性信息来实现转子初始位置辨识，仿真与实验结果验证了该辨识方法的可行性。为了提高低速阶段的转子位置与转速检测精度，本文通过研究基于纯统计规律的多项式外插法和基于纯系统模型的转子位置龙伯格观测器，提出了基于多项式外插法的矢量跟踪观测器方法。该方法通过对编码器信号进行多项式外插处理，得到基于龙伯格拓扑的矢量跟踪观测器参考输入，然后由观测器估计转子位置与速度。该方法兼具多项式外插法的线性预估性能和观测器对系统变化（给定随动与负载扰动）的动态响应性能，对于编码器检测精度不高的系统，其低速检测性能有显著的提高。针对转子磁极位置偏差或者母线电压偏低造成的高速运行电压饱和问题，本文提出了一种新颖的考虑死区与最小脉宽限制的综合式脉宽调制（Integrated PWM, IPWM）方法。该方法综合了SVPWM、DPWM1、三段式PWM和一段式PWM，通过参考电压线性修正、调制算法优化和IPWM新死区插入，实现了在提高输出电压基波幅值线性度和抑制总谐波畸变的基础上提高母线电压的利用率，从而拓宽了控制环输出电压的调节范围，增加了电梯曳引系统运行的稳定性和鲁棒性，仿真与实验结果验证了该IPWM方法的优越性能。更多还原

【Abstract】 As vertical transportation in high-rise buildings and public places, the elevator hasbecome important and indispensable construction equipment. The control performance of theelevator traction system is mainly manifested in the suppression performance of the startingslide and the running vibrating. In the starting stage, the system usually adopts zero positionservo control, while the sudden load disturbance can cause the elevator slide at the momentthe elevator brake opens. The load disturbance suppression performance at the starting stagemainly depends on the initial position precision, the rotor position and speed detectionprecision and control algorithm. In the running stage, the system employs the S curve givenspeed servo control. The velocity tracking performance at low speed is mainly influenced bythe rotor position and speed detection precision, while the performance at high speed isaffected by the output voltage saturation as well.To ensure the control performance of elevator traction system, the key problems come tothe initial position precision, the rotor position and velocity precision at low speed and theoutput voltage saturation at high speed. Therefore, this paper focuses on the following threeaspects:(a) the methods to improve the identification precision of the initial position forpermanent magnet synchronous motor rotor,(b) the rotor position and speed detectionoptimization methods at low speed,(c) the pulse-width modulation (PWM) optimizationalgorithm to increase the utilization ratio of bus voltage, which expands the output voltagerange of the control loop.In order to improve the identification precision of the rotor initial position, this paperintroduces three kinds of classic rotor initial position identification methods at standstill,including pulse voltage injection method, high frequency rotating voltage injection methodand high frequency pulsating voltage injection method, analyses the rotor movement duringidentification process based on friction model, presents the rotor standstill boundaryconditions and proposes the rotor motion suppression methods beyond the boundary condition.By suppressing the rotor movement during identification process, the identification precisionof the rotor initial position is improved. Simultaneously, for non-salient or low salientpermanent magnet synchronous motor, another rotor initial position identification methodbased on dynamic magnetic resistance difference at standstill is proposed. This novel methodinjects high frequency voltage signal to the stator coils and extracts the rotor position andpolarity information from the second harmonic current component to achieve the rotor initialposition identification, and simulation and experimental results verify its feasibility. Aiming to improve the rotor position and speed detection precision at low speed, afterstudying the polynomial extrapolation method based on pure statistical law and the rotorposition Luenberger observer based on pure system model, this paper proposed a vectortracking observer method based on polynomial extrapolation. The proposed method puts thethe polynomial extrapolation result on the basis of the encoder signal as the reference input ofthe vector tracking observer based on the Luenberger topology, and then estimates the rotorposition and speed by the observer. This method combines the linear prediction performanceof polynomial extrapolation method and the excellent dynamic performance of theLuenberger observer including refference follow-up and load disturbance. For the systemwithout high-precision encoder, the proposed method significantly improves the rotor positionand speed detection performance at low speed.Aim at voltage saturation problem at high speed caused by rotor position deviation or lowbus voltage, this paper proposes a novel IPWM (Integrated PWM) method with dead time andminimum pulse-width limitation considered. The method combines SVPWM (space vectorPWM), DPWM1(discontinuous PWM), three-segment PWM and one-segment PWM, by thereference voltage linear correction, modulation algorithm optimization and IPWM newdead-band insertion. With the output voltage amplitude linearity improved and the totalharmonic distortion suppressed at the same time, the method achieves the highest utilizationrate of bus voltage, which broadens the voltage regulating range of the control loop andincreases the stability and robustness of the elevator traction system operation. The simulationand experiment results demonstrate the superior performance of the IPWM method.更多还原