【作者】 宫玉琳；

【导师】 曲永印；

【作者基本信息】 长春理工大学 ， 机械电子工程， 2013， 博士

【摘要】 永磁同步电动机伺服系统广泛应用于工业、农业、国防、航空航天、社会生活等各个领域。随着科技的不断进步,对于伺服控制系统的要求也越来越高。因此,研究先进的永磁同步电动机伺服系统控制策略、开发高性能永磁同步电动机伺服系统,对提高我国工业发展水平、促进国防现代化建设,具有极其重要的现实意义和实用价值。永磁同步电动机伺服系统作为一个非线性、多变量、强耦合的时变系统,在运行过程中,参数变化、外界扰动和噪声等非线性因素将使系统性能变差。因此,要提高永磁同步电动机伺服系统性能,必须采用先进的控制策略以克服参数变化和外界扰动对系统产生的不利影响。本文针对提高永磁同步电动机伺服系统性能进行了深入分析,结合信号处理的方法,提出了基于复合自适应逆控制策略的永磁同步电动机伺服控制系统,采用位置环、速度环、电流环控制结构。其中,速度环和电流环采用基于矢量控制的id=0控制策略,通过id=0控制策略可以使电动机获得较宽的调速范围和较好的转矩性能。永磁同步电动机伺服系统的关键性能指标是位置控制精度,为了提高位置控制精度,系统位置环采用复合自适应逆控制策略。当控制器为对象的逆模型时,位置输出跟随位置输入,同时对系统存在的扰动进行消除。在自适应逆控制系统中,被控对象的输入同时驱动对象和对象模型,对象输出和对象模型的输出之差即为对象的噪声,利用该噪声去驱动对象的逆模型并在对象输入中予以消除。自适应逆控制策略从根本上消除了噪声对位置输出的影响,提高了伺服系统的控制性能。为了进一步提高自适应逆控制的性能,本文对变步长最小均方(Least mean square, LMS)算法进行了分析和改进。提出了基于相关误差的非线性变步长LMS算法,提高了LMS算法的收敛速度并有效地克服了参数扰动对算法的影响。非线性滤波器作为非线性自适应逆控制的重要组成部分,直接影响着自适应逆控制策略的控制效果。为了更好的实现非线性自适应逆控制,本文将动态径向基函数(Radial Basis Function,RBF)神经网络和FIR滤波器相结合构成非线性滤波器,并采用混沌多群体粒子群优化(Chaos Multi-population Particle Swarm Optimization, CMPSO)算法对非线性滤波器的权值进行离线优化,进一步提高了自适应滤波器的收敛速度和精度,有效地实现了系统建模、逆建模以及扰动消除器的设计,进而提高了自适应逆控制策略的控制效果。最后,本文将基于动态RBF神经网络和FIR的非线性滤波器与基于相关误差的非线性变步长LMS算法相结合,实现了永磁同步电动机伺服系统自适应逆控制。进一步将PI控制策略与自适应逆控制策略相结合,提出了一种复合自适应逆控制策略。仿真和实验结果表明,基于本文提出的复合自适应逆控制策略的永磁同步电动机伺服系统具有较好的动态响应、较高的稳态精度和较强的抗干扰能力,证明了本文所提出的控制策略的有效性和先进性。更多还原

【Abstract】 The permanent magnet synchronous motor control system has been widely used in industry, agriculture, aerospace and various fields of social life. With the continuous progress of science and technology, servo system requirements have become increasingly stringent. So study the high-performance permanent magnet synchronous motor servo system control strategy and development high-performance permanent magnet synchronous motor servo system products has extremely important practical significance and practical value to improve industrial levels and promote the development of defense industry.The permanent magnet synchronous motor servo system as a nonlinear, multivariable, strong coupling varying systems is difficult to describe in precise mathematical model. In the process, the external disturbance and other nonlinear factors will lead to poor system performance. Therefore, to improve the performance of the control system of permanent magnet synchronous motor servo system, advanced control strategies must be used to overcome parameter variations, external disturbances and make the system has good tracking performance and strong robustness.This paper proposed composite adaptive inverse control strategy for permanent magnet synchronous motor servo control system based on nonlinear signal processing method. The system adopts the position loop, speed loop and current loop control structure. The speed loop and current loop based on id=0control method. The id=0control method can make permanent magnet synchronous motor achieves wide speed range and good torque performance.The key performance of permanent magnet synchronous motor servo system is position accuracy. In order to improve it, the adaptive inverse control strategy is used in the position loop. The adaptive inverse control strategy combined with the idea of inverse control and adaptive control. When the controller is the inverse of the plant, the output of position can track the input precisely. Meanwhile the whole disturbance between the plant and model derive both the plant and model, and this disturbance was subtracted from the input of plant by driving the inverse of the model, and the noise and disturbance of system can be eliminated ultimately. This strategy solves the problem of high performance control of AC servo system preferably.To further enhance the performance of adaptive inverse control, an improved variable step size LMS algorithm based on dependent errors was proposed. This variable step size LMS algorithm improves the convergence speed and overcome the impact of parameter disturbances.Nonlinear filter as a key component of the nonlinear adaptive inverse control directly impact the performance of adaptive inverse control strategy. In order to better achieve the nonlinear adaptive inverse control, an improved nonlinear adaptive filter based on dynamic RBF neural network and FIR filter is proposed. Meanwhile, the chaos multi-population particle swarm optimization(CMPSO) algorithm is used to training of the nonlinear filter weights offline, and the training results as the initial weights of the nonlinear filter. This method improves the control performance of adaptive inverse control strategy.Combined with the proposed nonlinear filter and variable step size LMS algorithm, the permanent magnet synchronous motor servo system based on adaptive inverse control is realized. Further, a composite adaptive inverse control strategy is proposed based on the conventional PI control strategy and adaptive inverse control strategy. Simulation and testing results show that the composite adaptive inverse control strategy for permanent magnet synchronous motor servo system has good dynamic response, steady-state precision and disturbance-rejection ability. The validity and advantage of the proposed control strategy is proved.更多还原