高速直线伺服系统的学习前馈控制算法及谐振抑制研究

发布时间：2018-03-30 03:05

本文选题：直线伺服系统　切入点：前馈控制　出处：《浙江理工大学》2017年硕士论文

【摘要】：直线伺服系统是一种直接驱动给进的伺服系统,不存在中间任何的的传动环节,直接输出直线运动,具有结构简单、适合高速度和高加速度运动、易于调节和控制等特点,因而被广泛的应用在了高速高精数控系统、半导体制造和精密仪器等领域中。为了发挥其高速高精的特性,需要不断研究控制算法以及设计运动控制器。本文首先介绍了直线伺服系统的二自由度控制策略,针对反馈控制只能保持系统的鲁棒性,而无法提高伺服系统的轨迹跟踪性能的问题。提出一种线性前馈控制器,并对前馈控制器进行参数化。该前馈控制器可以根据伺服系统的建模精度以及输入轨迹规划的阶数调节自身的阶数,到达不同的伺服控制精度。为解决直线伺服固有时滞特性对轨迹跟踪性能的影响,以前馈加反馈复合控制器为基础,分析直线伺服时滞特性对其轨迹跟踪性能的影响。在此基础上设计了前馈环节上的延时控制器,针对延时控制器位于前馈通道时作为一个超前环节控制上无法实现的问题进行了延时控制器重新设计,将延时参数分为两部分,分别调节理想轨迹指令按照整数倍伺服周期延时及前馈控制信号滤波延时以达到补偿时滞的目的。其次随着不断提高的直线伺服系统带宽,直线伺服系统本身的柔性特性的影响愈发明显,导致机械谐振,本文引入“弹簧-阻尼”双惯量模型,对直线伺服系统的谐振特性进行了分析。选用被动式谐振抑制方案,在控制器中加入陷波滤波器,通过matlab仿真验证陷波滤波器对于机械谐振抑制的效果。针对直线伺服系统可以重复运动的特性,对上述前馈控制器、延时控制器、陷波滤波器采用迭代学习控制的方法,可以自动优化控制器参数,提升伺服系统性能。分别介绍了梯度下降法、牛顿法、割线法,对这些迭代方法进行了理论推导,并带入控制器参数,推导可以进行仿真及实验的迭代公式。最后,介绍了实验室的直线伺服系统的硬件和软件平台,并在此之上选择适合的迭代方法,对三种控制器的参数进行迭代优化仿真及实验,验证控制器的作用和迭代优化方法的效果。
[Abstract]:Linear servo system is a kind of servo system with direct drive and feed, there is no transmission link in the middle, and the direct output linear motion has the characteristics of simple structure, suitable for high speed and high acceleration motion, easy to adjust and control, etc. Therefore, it has been widely used in the fields of high speed and high precision numerical control system, semiconductor manufacturing and precision instrument, etc. In order to give play to its characteristics of high speed and high precision, It is necessary to study the control algorithm and design the motion controller. Firstly, this paper introduces the two-degree-of-freedom control strategy of the linear servo system, aiming at the feedback control can only maintain the robustness of the system. This paper presents a linear feedforward controller, which can not improve the track tracking performance of servo system. The feedforward controller can adjust its order according to the modeling accuracy of servo system and the order of input trajectory planning. In order to solve the influence of inherent delay characteristics of linear servo on trajectory tracking performance, the feedforward and feedback composite controller is used to solve the problem. The effect of linear servo delay on trajectory tracking performance is analyzed. Based on this, a delay controller based on feedforward link is designed. The delay controller is redesigned to solve the problem that the delay controller can not be realized as a leading link when the delay controller is located in the feed-forward channel. The delay parameters are divided into two parts. The ideal trajectory instruction is adjusted according to integer multiple servo period delay and feedforward control signal filtering delay to compensate the delay. Secondly, with the increasing bandwidth of linear servo system, The influence of the flexibility of linear servo system becomes more and more obvious, which leads to mechanical resonance. In this paper, the "spring-damping" dual inertia model is introduced to analyze the resonance characteristics of linear servo system. The notch filter is added to the controller, and the effect of notch filter on mechanical resonance suppression is verified by matlab simulation. The notch filter adopts iterative learning control method, which can automatically optimize the controller parameters and improve the performance of servo system. The gradient descent method, Newton method and Secant method are introduced, and the theoretical derivation of these iterative methods is given. Finally, the hardware and software platform of the linear servo system in the laboratory is introduced, and the suitable iterative method is selected on the basis of introducing the parameters of the controller, and deducing the iterative formula that can be used in simulation and experiment. The parameters of the three kinds of controllers are simulated and experimented to verify the function of the controller and the effectiveness of the iterative optimization method.
【学位授予单位】：浙江理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP273;TM921.541

【相似文献】