节点文献
直驱精密平面并联运动平台的动力学建模与轮廓控制
Dynamic Analysis and Contouring Control of A2-DoF Planar Parallel Manipulator Driven by Linear Motors
【作者】 张刚;
【导师】 丁汉;
【作者基本信息】 上海交通大学 , 机械电子工程, 2014, 博士
【摘要】 精密运动平台是现代数控机床和电子制造装备中的关键部件,决定了数控机床和电子制造装备的精度和生产效率。精密运动平台的结构形式多种多样,按机构型式来分,有串联运动平台和并联运动平台;按驱动形式来分,有间接驱动和直接驱动运动平台。直接驱动并联运动平台采用直驱电机直接驱动,其直接驱动机构可以提高机械系统的传动刚度和传动精度,其多支链闭环结构可以消除串联运动平台的误差积累效应,减小运动部件的质量,提高运动平台的运动速度和负载能力,因此,直驱并联运动平台是潜在的高速度、高精度运动平台。开发直驱并联运动平台可进一步满足现代生产制造对精密运动平台高速度、高精度的需求,具有重要的理论意义和应用价值。本论文以直线电机驱动的精密平面并联运动平台为研究对象,从操作空间动力学分析、直驱并联运动平台的尺寸综合与动态优化设计、基于操作空间动力学模型的运动控制器设计、全局任务空间的交叉耦合轮廓控制等多个方面,对直驱精密平面并联运动平台在高速度、高精度轨迹跟踪中的若干理论方法及关键技术进行了深入的理论分析和实验研究。本论文首先分析并设计了平面并联运动平台的构型,建立了显式的平面并联运动平台正/逆向位置、速度和加速度方程。基于拉格朗日第一类方程,建立了平面并联运动平台的操作空间动力学模型。基于平面并联运动平台的运动学方程和操作空间动力学,全面地分析了其各项运动学性能指标(包括:工作空间、条件数、刚度、灵巧性等)和动态性能指标(包括:动态灵巧性、动态可操作性等)。在此基础上,提出平面并联运动平台的全局综合运动学性能指标(Global and Comprehensive Performance Index, GCPI)和全局综合动态性能指标(Global and Comprehensive Dynamic Performance Index,GCDPI),平面并联运动平台的动态优化设计则表示为以全局综合动态性能指标GCDPI为目标函数的非线性有约束优化问题。基于动态优化设计的计算结果,设计完成并联运动平台的机械结构,搭建了直驱精密平面并联运动平台的原型样机。为了实现平面并联运动平台关节空间的精密轨迹跟踪控制,本文基于平面并联运动平台操作空间动力学模型,提出了基于模型的前馈控制器(Model-based feedforwardcontrol, MFC)。基于模型的前馈控制器与级联PID/PI控制器和速度前馈控制器(Velocityfeedforward control, VFC)相结合,构成复合PID/PI+VFC/MFC轨迹跟踪控制器,实现平面并联运动平台关节空间高速度、高精度的轨迹跟踪控制。仿真和实验结果表明,复合PID/PI+VFC/MFC轨迹跟踪控制器可以补偿动平台运动过程中负载变化、哥氏力、离心力等非线性因素对控制系统的影响,提高关节空间的轨迹跟踪精度。在动平台以角速度=10rad/s,最大速度vm ax=250mm/s,最大加速度a=2510mm/s2m ax跟踪一直径为50mm的圆弧轨迹时,两直线电机位置跟踪误差的均方值分别为0.0118mm和0.0141mm,最大跟踪误差分别为0.0366和0.0424mm。与基于加速度前馈(Acceleration feedforwardcontrol, AFC)的复合PID/PI+VFC/AFC轨迹跟踪控制器相比,关节空间直线电机位置跟踪误差均方值和最大值都降低了15%以上。为了实现直驱精密运动平台操作空间的轮廓控制,本文在复合PID/PI+VFC/MFC轨迹跟踪控制器的基础上,进一步提出基于全局任务空间的交叉耦合轮廓控制器(Cross-coupled Controller, CCC)。基于平面并联运动平台运动学分析结果和密切圆轮廓误差模型,在平面并联运动平台的全局任务空间内计算轮廓误差,采用PI型交叉耦合控制器对直线电机的速度输出信号进行补偿,构成基于全局任务空间的闭环交叉耦合轮廓控制系统。通过广义轮廓误差传递函数(Generalized controur error transfer function,GCETF)建立非耦合轮廓控制系统与交叉耦合轮廓控制系统轮廓误差之间的函数关系,交叉耦合轮廓控制器的参数整定与稳定性分析问题则转化为基于GCETF的控制器设计与稳定性分析问题。轮廓控制实验结果表明,平面并联运动平台的交叉耦合轮廓控制器可实现高速度、大曲率的精密轮廓控制。在动平台以角速度=10rad/s,最大速度vm ax=201.06mm/s,最大加速度am ax=2008.6mm/s2跟踪一长轴为50mm,短轴为40mm的椭圆轮廓时,轮廓误差的均方值为0.0047mm,最大轮廓误差为0.0255mm;在动平台以角速度=5rad/s,最大速度vm ax=175.93mm/s,最大加速度am ax=1768.6mm/s2跟踪定线段长度为60mm的伯努利双纽线轮廓时,轮廓误差的均方值为0.0037mm,最大轮廓误差为0.0326mm。与各轴独立的复合PID/PI+VFC/MFC轨迹跟踪控制器相比,交叉耦合轮廓控制器在平面并联运动平台的高速椭圆、伯努利双纽线轮廓控制中,轮廓误差均方值和最大值降低了24.29%和17.95%以上。本论文的研究成果对于开发新型的精密并联运动平台、对于直驱精密并联运动平台的机构分析与综合、高速高精度运动控制算法的理论分析与实验研究具有参考价值。
【Abstract】 Precision motion platform are key components of modern manufacturing equipments,which determine the accuracy and production efficiency of precision CNC machine toolsand electronic manufacturing equipments. From the viewpoint of mechanism, precisionmotion platform can be devided into serial platforms and parallel platforms (which is alsocalled parallel manipulator). From the viewpoint of drive type, they can be devided intodirect-drive motion platforms and indirect-drive motion platforms. The motion platformdirect driven by linear motors can improve the rigidity and reliability of transmission system,resulting in better control performances, higher transmission and positioning accuracy of themotion platform. At the same time, the multi-loop kinematic structure of parallel mechanismcan compensate for the error accumulation effects and achieve better tracking performance,which results a potential solutions to high-speed and high-precision motion applications. Thedevelopment of precision motion platform driven by linear motors can meet therequirements of modern manufacturing equipments and is of great theoretical and industrialsignificance.The subject investigated in this dissertation is a2-DoF high-precision planar parallelmanipulator direct driven by linear motors. Several fundamental problems, including thedimensional synthesis, inverse dynamic analysis and dynamic optimization of parallelmechanism, the task space dynamic model-based trajectory tracking control and thecontouring control of the2-DoF planar parallel manipulator are fully investigated andintensively studied. Experimental results are presented to validate the proposed trajectorytracking control and contouring control stragegies.Based on the structural analysis of the parallel manipulator, explicit expression ofinverse and direct kinematics of2-DoF planar parallel manipulator are derived. TheLagrangian formulation is applied to derive explicit expressions of task space dynamicmodel of the2-DoF planar parallel manipulator. The kinematic and dynamic performanceindices, including the workspace, the conditioning index, the dynamic manipulability and dynamic dexterity are analyzed. A global and comprehensive performance index (GCPI) anda global and comprehensive dynamic performance index (GCDPI) are proposed for thedimendional synthesis and dynamic optimization. Simulation results show that the2-DoFplanar parallel manipulator with optimized kinematic and inertial variables has larger, moreisotropic and more uniform dynamic manipulability in the prescribed task space of theparallel manipulator.In order to achieve high precision trajectory tracking control of the2-DoF parallelmanioulator, a task space dynamic model-based feedforward controller (MFC) is proposedin this dissertation. The model-based feedforward controller is combined with a cascadePID/PI controller and a velocity feedforward controller (VFC) to construct a hybridPID/PI+VFC/MFC trajectory tracking controller. The task space dynamic model and currentloop controller of linear motor are used to derive the transfer function of model-basedfeedforward controller. Experimental results show that, when the moving platform iscommanded to track a circle with a radius of50mm at an angular velocity of=10rad/s, avelocity ofvm ax=250mm/s and an acceleration ofam ax=2510mm/s2, the Root Mean Square(RMS) values of axial tracking errors are0.0118mm and0.0141mm, respectively, themaximum absolute values of axial tracking errors are0.0366mm and0.0424mm,respectively. Compared with conventional hybrid PID/PI+VFC/AFC trajectory trackingcontroller, the hybrid PID/PI+VFC/MFC trajectory tracking controller has superior trackingperformance, which reduces axial tracking errors by at least15%in the high-feedratecircular tracking experiments.In order to achieve high precision contouring control of the2-DoF parallel manipulator,a global task space based cross-coupled contouring controller (CCC) is proposed in thisdissertation. The contouring controller is based on a PI type cross-coupled controller andcircular approximation of contour error in the global task space of the parallel manipulator.The generalized contour error transfer function (GCETF), which establishes the relationshipbetween the contour error of uncoupled controuring control system and cross-coupledcontrouring control system, is used to derive the control parameters and stability analysis ofthe cross-coupled controller. The task space based contouring controller is combined withthe hybrid PID/PI+VFC/MFC trajectory tracking controller to achieve high-precisioncontouring control of the2-DoF planar parallel manipulator. Experimental results show that,when the moving platform is commanded to tracking an ellipse with long-axis of50mm andshort-axis of40mm at an angular speed of=10rad/s, a velocity ofvm ax=201.06mm/s and an acceleration ofam ax=2008.6mm/s2, the Root Mean Square (RMS) and maximum absolutevalues of the contour errors are0.0047mm and0.0255mm, respectively. When the movingplatform is commanded to tracking a lemniscate of Bernoulli with fixed length of60mm inthe X-axis at an angular speed of=5rad/s, a velocity ofvm ax=175.95mm/s and anacceleration ofa max=1768.6mm/s2, the RMS and maximum absolute values of the contourerrors are0.0037mm and0.0326mm, respectively. Compared with uncoupled contouringcontroller, the cross-coupled controller has superior contouring tracking performance, whichreduces contouring errors by at least24.29%and17.95%in the high-feedrate elliptical andlemniscate contouring experiments.The research results of this dissertation is useful for the development and application ofnew-type precision motion platform. It is of great reference value to the structural synthesisof parallel mechanism and to the theoretical and experimental studies of high-speed andhigh-precision motion control algorithms.