【作者】 王相兵；

【导师】 童水光；

【作者基本信息】 浙江大学 ， 流体机械及工程， 2014， 博士

【摘要】 随着计算机技术和产品设计技术的飞速发展,高速化、高精度、抗冲击、低噪声、轻量化和长寿命成为机械产品性能的发展趋势,其动态特性研究越来越成为机械产品设计的关键,正成为国内外专家研究的热点。工程机械臂是各种工程机械设备的关键、重要部件,是一种多功能高效机械部件,其动力学性能决定着工程机械整机的性能发挥。工程机械的作业工况通常比较的恶劣,工程机械臂所受的载荷情况非常复杂,为保证其工作安全可靠,提高施工质量和作业效率,对其进行准确的动力学分析非常必要。工程机械臂通常是冗余自由度、强非线性、刚柔耦合的多体系统,其动力学及特性研究涉及到多体动力学以及控制、结构动态优化、运动学逆问题等方面的问题。目前技术条件下对工程机械臂动力学的研究在产品建模、仿真模拟、数值计算等关键技术均处于探索阶段,而用户对工程机械产品的性能要求日益增长。因此,对于工程机械臂系统动力学及特性研究具有重要的理论价值和实际的工程意义。本文以工程机械臂系统为研究对象,主要研究多体动力学尤其是柔性多体动力学理论在建模、仿真方面的应用,重点是进行理论性的分析。对工程机械臂的动力学建模、仿真、数值计算以及运动控制、结构动态优化等方面作了细致的研究工作,并结合液压挖掘机及大型船用挖掘机机械臂实例进行研究,主要研究内容归纳如下：1)研究了工程机械臂的多刚体动力学方程及柔性多体动力学方程的建模方法,主要讨论了等效有限元方法和拉格朗日定理,重点研究了柔性梁的多体动力学理论基础。2)采用基于集中质量和惯量的等效有限元法建立液压挖掘机机械臂的动力学微分方程,运用龙格-库塔(Runge-Kutta)数值求解法对运动微分方程进行推导和数值求解,对各臂杆的动力学特性进行了分析,并运用ADAMS软件进行仿真分析。根据柔性多体动力学理论,采用模态函数描述臂架的弹性变形,采用LAGRANGE定理和虚功原理建立液压挖掘机臂架系统刚柔耦合的非线性动力学方程,并通过数值求解、动力学仿真的方法对建立的动力学方程进行分析。对已建立的动力学方程利用MATLAB进行数值求解,运用仿真软件ADAMS及NASTRAN建立液压挖掘机机械臂刚柔耦合模型并进行仿真分析,并对比二者结果。运用数值求解的方法进行模态计算和动力学响应分析,求解相关几何参数的一阶固有频率灵敏度,分析影响机械臂动力学特性的主要模态参数。3)考虑大型船用挖掘机机械臂的轴向变形,采用拉格朗日定理建立其柔性机械臂动力学方程,采用数值求解的方法,利用复模态分析中的状态矢量法对机构动力学控制方程进行了解耦,得到机械臂在标称运动附近振动固有频率和大范围刚性转动角速度的关系。应用UG、NASTRAN、ADAMS等软件建立了柔性臂架、柔性变幅绳、柔性抓斗提升钢丝绳、刚性支架、刚性回转平台的船用挖掘机虚拟样机模型。根据船用挖掘机机械臂系统结构及工况特点,施加载荷并进行动力学分析,通过后处理模块,得到系统位移、速度、加速度等动态性能,以及连接点的铰接力和动臂的动态应力等特性。4)研究液压挖掘机机械臂的轨迹规划及运动控制。分析了工程机械臂控制策略理论基础,根据控制原理及机械臂铲斗运动轨迹跟踪控制表达式,建立工作装置机械臂PID控制模型及基于RBF神经网络的PID控制模型,运用Matlab及相关软件对控制模型进行仿真计算,分析了系统的动态响应和稳态误差,并得出基于RBF神经网络的PID控制优于常规控制的结论,能够实现挖掘机工装轨迹的智能控制。5)利用模态分析、谐响应分析确定对系统动态性能影响最大的模态频率,以此为基础构建动态优化的目标函数,再利用灵敏度分析缩减设计变量以提高动态优化的效率。利用有限元分析软件中的优化模块对液压挖掘机工作装置机械臂结构系统进行了动态优化设计。采用拉格朗日乘子法将液压挖掘机工作装置机械臂结构约束动态优化问题转化为无约束的动态优化问题,得到此拉格朗日函数的极小值点并且此拉格朗日函数极值点将逐步逼近原目标函数的约束最优点,由收敛准则可得近似最优解。以有限元分析为基础,将正交试验法与BP神经网络结合,建立了工作装置机械臂结构设计变量与动态特性之间的非线性映射关系的神经网络模型,用遗传算法优化神经网络模型得到最优解。更多还原

【Abstract】 With the rapid development of computer technology and digital design technology, high speed, high precision, lightweight and impact resistant, low noise,and long life have become the development trend of mechanical products performance. The dynamic characteristic research more and more becomes the key of the mechanical products design, is becoming hot topics in the study of domestic and foreign experts. Engineering mechnical arm is the critical and important parts of various engineering machinery and equipment, it is a kind of multi-functional efficient mechanical parts.the performance play of machine is determined by the engineering arm. Working condition of engineering mechanical tending to be bad,and load of engineering mechanical arm being very complicated, it is necessary to conduct kinetic analysis, to ensure the safety and reliability of its work,and to improve construction quality and operational efficiency.The engineering mechanical arm generally is a mufti-body system with redundant freedom, strong nonlinear, coupled with rigid and flexible characters.The system is related to multi-body dynamics, control theory, structural dynamics and kinematic reverse problem, etc.Under the current condition, the Research on the engineering arm dynamics is in the exploration stage such as in the product modeling, simulation and numerical calculation, etc,but the user’s requirements for the performance of the engineering mechanical product is increasing. So it is of greattheoretic value and real engineering significance to study the engineering mechanical arm dynamics.In this article, the engineering arm system is considered as the study object. The application of mufti-body dynamics, especially the flexible mufti-body dynamics theory in the modeling and simulation is mainly studied, and focus on the theoretic analysis. Much research has been done on the dynamics modeling,simulation,numerical calculation, and motion control, structures dynamic optimization,etc.and the study is conducted Combined with the examples of hydraulic excavators and large marine excavator, and the main content is summarized as follows:1).The modeling method of multi-body dynamics equations and flexible multi-body dynamics equations of the engineering mechanical arm was researched,the equivalent finite element method and Lagrange theorems was mainly discussed,and the relevant theory has been studied, especially focusing on the multi-body dynamics.theoretical basis of the flexible beam.2).The modal functions are adopted to describe elastic deformation of the mechanical arm,Lagrange theorem and the principle of virtual work is used to establish rigid-flexible coupling nonlinear dynamic equations of arm frame system.And the established dynamic equation are analyzed through the numerical solution, dynamic simulation method.The dynamic equations are numerically solved byMATLAB, the rigid-flexible coupling model of hydraulic excavator’s mechanical arm is established and simulated by the simulation software ADAMS and NASTRAN, it is showed that the modeling method of dynamic equations adopted in this paper is correct by contrasting the both results. The modal and dynamic response is calculated by applying the numerical solution method,the sensitivity of the first natural frequency for the Geometric parameters related is solved, the main modal parameters are analyzed which influence dynamic characteristics of mechanical arm.3).Considering axial deformation of large marine excavator engineering arm, its flexible engineering mechanical arm dynamic equations by using Lagrange theorems, the method of numerical solution is adopting, institutional dynamics equations were decoupled by using complex state vector method in the complex modal analysis, and the relationship between the vibration natural frequencies in the vicinity of nominal movement and a wide range of rotation angular velocity is gotten.The virtual prototype model of marine excavator including a flexible boom, flexible grab hoisting rope,flexible luffing rope,rigid frame,rigid rotating platform is established by using soft such as UG,NASTRAN,ADAMS,etc. the load is applied and dynamic analysis is conducted according to structure and work conditions of marine excavator mechanical arm system,the dynamic performance of the system is gotten by post-processing module such as the displacement, velocity, acceleration, etc.4).The manipulator trajectory planning and motion control of hydraulic excavator’s mechanical arm are researched.the theoretical foundation of control strategy for the engineering mechanical arm is analyzed, based on control theory and manipulator bucket trajectory tracking control expression,the PID control model and PID control model based on RBF neural network of working device’s mechanical arm,the controlling model simulation is conducted by using matlab and related software, dynamic response and steady state error of system are analyzed, the conclusion which PID control based on RBF neural network is superior to conventional PID control is drawn,intelligent control of the hydraulic excavator’s working device trace Can be achieved.5).Modal analysis and harmonic response analysis is used to determine the modal frequency of the biggest influence on the system’s dynamic performance,based on this dynamic optimization objective function is built.and the sensitivity analysis is used to reduce design variables in order to improve the efficiency of dynamic optimization.The structure system dynamic structure optimization design for hydraulic excavator working device’s mechanical arm was conducted using finite element analysis software optimization module,Constrained dynamic optimization problem of hydraulic excavator working device structure is changed into unconstrained dynamic optimization problem by using the Lagrange multiplier method.Get the minimum point of Lagrange function and the Lagrange function extremum points gradually close to the constraint optimal point of the original objective function, the optimal solution can be approximated based on the convergence criterion.Based on finite element analysis,the orthogonal test method combined with BP neural network, the neural network model of the nonlinear mapping relationship between working device design variables and the dynamic characteristics is set up. The optimal solution is gotten by using genetic algorithm to optimize neural network model.更多还原