基于经验模态分解算法的直驱XY平台迭代学习控制

发布时间：2018-07-23 08:48

【摘要】：工业加工的技术水平是国家生产力和工业实力的重要指标,由于数控机床具有快速精确的加工能力,近年来一直备受业界的瞩目。永磁直线同步电动机兼有永磁同步电机体积小、运行效率高、动态性能好和直线电机结构简单、精度高、速度快的双重优点,已成为高速超高速、高档数控装备中的主流驱动方式。此外,由其驱动的用以实现平面定位和平面进给的XY双轴工作台,在数控镗、铣、钻床以及各种加工中心中应用广泛。运动控制的精度取决于伺服控制结构和机械结构,由于数控装备的机械结构已趋于成熟,所以为提高加工精度,有必要设计性能良好的控制器。首先,本文介绍了永磁直线同步电机和直驱XY平台的结构、工作原理及发展应用,分析了系统受到的扰动,并建立了永磁直线同步电机和直驱XY平台的数学模型。定义跟踪误差和和轮廓误差,推导直线指令、圆弧指令和任意轮廓指令下的轮廓误差表达式。然后,针对传统控制方法对存在不确定性、时变性和非线性因素的实际系统的局限性,本文为实现永磁直线同步电机伺服系统的精确跟踪控制,设计了迭代学习位置控制器和PI速度控制器。针对传统交叉耦合控制器与单轴反馈控制器相互独立设计、缺乏系统化的问题,本文为实现直驱XY平台的精确协调控制,同时建立单轴反馈控制系统和双轴交叉耦合迭代学习控制结构,并利用瞬态响应指直接给出各控制器参数的表达式。由于迭代学习控制利用误差和上一次控制输入计算当前控制输入,因此误差的累加会直接影响系统的收敛速度和稳定性,甚至导致系统发散。针对上述问题,本文利用经验模态分解算法分解迭代过程中的跟踪误差和轮廓误差,剔除其中的发散分量,改善迭代学习控制系统和交叉耦合迭代学习控制系统的收敛性和跟踪精度与轮廓精度。最后,在Matlab/Simulink下搭建永磁直线同步电机和直驱XY平台的控制系统仿真模型,编写程序实现迭代学习控制过程,验证设计的迭代学习控制器、交叉耦合迭代学习控制器的有效性。编写程序实现误差信号的经验模态分解,剔除发散分量后重新运行系统,并对两次仿真结果进行比较分析,验证提出新方法的有效性。
[Abstract]:The technical level of industrial processing is an important indicator of national productivity and industrial strength. Due to the rapid and accurate machining ability of NC machine tools, it has been attracting the attention of the industry in recent years. The permanent magnet linear synchronous motor has the advantages of small size, high running efficiency, good dynamic performance and the dual advantages of simple structure, high precision and high speed. High-grade CNC equipment in the mainstream drive mode. In addition, the XY double axis worktable driven by it is widely used in NC boring, milling, drilling machine and various machining centers. The precision of motion control depends on the servo control structure and mechanical structure. Since the mechanical structure of CNC equipment has become mature, it is necessary to design a controller with good performance for improving machining accuracy. Firstly, this paper introduces the structure, working principle and development and application of permanent magnet linear synchronous motor and direct-drive XY platform, analyzes the disturbance of the system, and establishes the mathematical model of permanent magnet linear synchronous motor and direct-drive XY platform. The tracking error and contour error are defined, and the contour error expressions under straight line instruction, arc instruction and arbitrary contour instruction are derived. Then, aiming at the limitation of the traditional control method to the actual system with uncertain, time-varying and nonlinear factors, this paper aims to realize the precise tracking control of the permanent magnet linear synchronous motor servo system. Iterative learning position controller and Pi speed controller are designed. To solve the problem that the traditional cross coupling controller and uniaxial feedback controller are independent of each other and lack of systematization, this paper aims to realize the precise coordination control of the direct drive XY platform. At the same time, the uniaxial feedback control system and the biaxial cross-coupling iterative learning control structure are established, and the expressions of the controller parameters are obtained directly by using the transient response index. Because the iterative learning control uses the error and the last control input to calculate the current control input, the accumulation of errors will directly affect the convergence speed and stability of the system, and even lead to the divergence of the system. In this paper, the tracking error and contour error in the iterative process are decomposed by the empirical mode decomposition algorithm, and the divergent components are eliminated. The convergence, tracking accuracy and contour accuracy of iterative learning control systems and cross-coupled iterative learning control systems are improved. Finally, the control system simulation model of permanent magnet linear synchronous motor and direct-drive XY platform is built under Matlab / Simulink. The iterative learning control process is realized by programming, and the effectiveness of the designed iterative learning controller and cross-coupled iterative learning controller is verified. A program is written to decompose the error signal, remove the divergent component and rerun the system. The results of two simulations are compared and analyzed to verify the effectiveness of the proposed method.
【学位授予单位】：沈阳工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TG659

【相似文献】