高速轻型并联机器人集成优化设计与控制

发布时间：2019-06-14 20:56

【摘要】：并联机器人以其大负载自重比、高刚度及高精度等优点被广泛用。随着现代工业中对生产效率要求的不断提高,并联机器人已被越来越多地应用于高速磨削、搬运及装配中,同时为降低机器人的生产成本及减少能源消耗,考虑轻量化的现代设计方法已逐渐应用于机器人的研制中。在高速运行工况下,轻量化的结构通常会引入弹性变形与振动,当采用传统针对刚性机器人的分析及控制方法时,将无法保证机器人的跟踪精度和动态性能。为解决并联机器人在高速运行时存在的共性问题,本文以具有典型结构的3RRR并联机器人为对象,开展了面向控制系统设计的并联机器人刚柔耦合动力建模;综合机构优化、驱动与传统部件参数匹配及控制系统参数调整于一体的集成优化设计;基于模型的轨迹优化及控制研究,并进行了性能测试与实验研究。首先对3RRR并联机器人进行运动学与刚体动力学建模,以此为基础,从欧拉梁的基本假设出发,考虑杆件端部的刚性特征,采用曲率有限元(CFE,Curvature based finite element)对杆件进行离散化建模,并推导并联机器人杆件中常用的悬臂梁及简支梁模型,从而考虑杆件端部刚性提出改进的曲率有限元(ICFE,improved curvature based finite element)方法。根据推导的杆件模型及小变形假设,提出杆件弹性位移与机器人刚体耦合运动的建模方法。根据上述模型并结合Kane方程建立3RRR并联机构的刚柔耦合动力学模型。为对推导的动力学模型进行验证,分别对模态与加速度响应进行分析,并将推导模型的计算值与ABAQUS模仿真值进行对比,同时根据模态分析对模型进行简化。结果显示,随着对杆件端部特征的考虑,模态分析与加速度响应分析结果与ABAQUS仿真结果吻合度得到显著提高,由于对模型的合理简化且求解时避免了对刚柔耦合运动约束方程的迭代,本文推导的模型保证了较高的计算精度与计算效率。为实现高速并联机器人良好的综合性能,提出集机构优化、驱动传动部件参数匹配及控制系统参数调整于一体的集成优化设计方法。为适应高速运行要求,在机构优化时考虑全局条件数、速度性能、加速能力及基频等运动学与动力学指标;为保证驱动及传动部件的经济性及参数匹配,建立其约束模型及参数库,并选择成本为优化目标;为取得高精度的控制性能,设计了动力学前馈加PD的控制算法,并选择系统跟踪误差为优化目标。从而建立包含机构参数、驱动与传动部件参数及控制系统参数的优化模型,并采非支配遗传算法NSGAII对模型求解,最终完成对高速并联机器人的集成优化设计。优化结果表明,集成优化设计有效提高了高速并联机器人的综合性能。为对高速并联机器人进行弹性位移(FD,flexible displacement)限制与残余振动(RV,Residue vibration)抑制,本文从轨迹规划的角度出发,考虑杆件柔性,对给定轨迹与点到点快速定位两种情况下的残余振动抑制与弹性位移限制问题进行研究。针对给定轨迹时的残余振动抑制问题,考虑到并联机器人各阶频率随位置改变的特点,将多模态输入整形(IS,Input shaping)与粒子群优化(PSO,Particle swarm optimization)及控制相结合,建立以残余振动为优化目标,以多模态输入整形器参数为优化变量的优化模型,并利用PSO对控制模型进行离线优化。仿真结果表明,优化后的整形器可对残余振动进行显著的抑制,且随着整形器数目的增加,残余振动进一步减小,同时避免了现有方法对整形器参数实时更新带来的计算量大的问题。对于快速点到点运动,将时间最优规划与多模态输入整形相结合,提出两步优化方法,首先采用分段伪谱法(GPM,Gauss pseudospectral method)对时间最优问题进行求解,之后采用多模态输入整形对得到的轨迹进行残余振动抑制,仿真结果表明,两步优化法实现了时间最优时的残余振动抑制与弹性位移限制。为解决含柔性环节高速轻型(HSLW,high-speed and light-weight)并联机器人的轨迹跟踪控制问题,提出了基于积分流形与高增益观测器的复合控制算法。首先根据刚度矩阵引入小参数,并基于积分流形将刚柔耦合动力学模型降阶为快速与慢速两个子系统,然后分别对两个子系统进行控制算法设计,对慢速子系统采用反演控制,同时考虑杆件弹性位移对末端轨迹的影响,设计校正力矩,实现了杆件弹性位移对机器人末端运动的弹性补偿。对快速子系统采用滑模变结构控制,从而保证流形成立。为解决曲率变化率无法直接测量的问题,引入高增益观测器,并对稳定性进行证明。对于上述复合控制算法,证明了整体系统的稳定性进行,并给出小参数的选取范围。最后,为验证复合控制算法的有效性,将设计的算法与奇异摄动及基于刚体动力学的反演控制算法进行对比,仿真结果表明,本文提出的基于积分流形与观测器的复合控制方法在振动抑制与轨迹跟踪方面均具有明显的优势。为对上述理论研究进行验证,参考优化设计结果对3RRR并联机器人进行了本体设计及驱动与传动部件选型,并采用工控机、实时操作系统及高速通讯总线的控制架构设计了控制系统;采用激光跟踪仪对机器人重复精度测试进行测试;为对建模精度进行验证,采用LMS振动测试仪对系统进行了模态测试与加速度响应测试;本文最后对基于多模态输入整形与PSO的规划方法及基于分段伪谱法与多模态输入整形的轨迹规划方法进行了实验验证。
[Abstract]:The parallel robot has the advantages of large load self-weight ratio, high rigidity and high precision. With the increasing demand of production efficiency in the modern industry, the parallel robot has been applied to the high-speed grinding, handling and assembly, in order to reduce the production cost of the robot and reduce the energy consumption. The modern design method considering the light weight has been applied to the development of the robot. In the high-speed operation, the light-weight structure usually introduces the elastic deformation and vibration, and when the traditional analysis and control method for the rigid robot is adopted, the tracking accuracy and the dynamic performance of the robot cannot be ensured. In order to solve the common problems of parallel robot in high-speed operation, a three-RRR parallel robot with a typical structure is used as the object, and the rigid-flexible coupling power modeling of parallel robot designed for the control system is carried out, and the comprehensive mechanism is optimized. The integration and optimization design of the parameter matching and control system parameters of the drive and the traditional part are adjusted, and the performance test and the experimental research are carried out based on the model-based trajectory optimization and control research. Firstly, the kinematics and rigid body dynamics of the 3RRR parallel robot are modeled. Based on this, the rigid character of the end of the rod is considered from the basic assumption of the Euler beam, and the finite element of curvature (CFE, Curvature based final element) is used to model the rod. In this paper, the conventional cantilever beam and simply-supported beam model in the rod of the parallel robot are derived, and the improved curvature finite element (ICFE) method is proposed in consideration of the rigidity of the end of the rod. According to the derived rod model and the small deformation hypothesis, the modeling method of the elastic displacement of the rod and the coupling motion of the rigid body of the robot is proposed. Based on the above model and the Kane equation, the rigid-flexible coupling dynamics model of the 3RRR parallel mechanism is established. In order to validate the derived dynamic model, the modal and acceleration response are analyzed, and the calculated value of the derived model is compared with the ABAQUS imitation true value, and the model is simplified according to the modal analysis. The results show that with the consideration of the characteristic of the end of the rod, the result of modal analysis and acceleration response analysis is significantly improved with the results of the ABAQUS simulation, and the iteration of the rigid-flexible coupling motion equation is avoided because of the reasonable simplification of the model and the solution. The model of this paper guarantees higher calculation accuracy and calculation efficiency. In order to realize the good comprehensive performance of the high-speed parallel robot, the integrated optimization design method of the integration of the mechanism optimization, the parameter matching of the drive transmission components and the parameter adjustment of the control system is put forward. in ord to meet that requirement of high-speed operation, the kinematic and dynamic indexes of global condition number, speed performance, acceleration capability and fundamental frequency are taken into account in the optimization of the mechanism, and the constraint model and the parameter library are established to ensure the economic and the parameter matching of the drive and the transmission component, and the selection cost is the optimization target; In order to obtain high-precision control performance, the control algorithm of dynamic feedforward and PD is designed, and the tracking error of the system is selected as the optimization target. So as to establish an optimization model containing the parameters of the mechanism, the parameters of the drive and the transmission components and the parameters of the control system, and the non-dominant genetic algorithm NSGAII is used for solving the model, and finally, the integrated optimization design of the high-speed parallel robot is finished. The optimization results show that the integrated optimization design effectively improves the comprehensive performance of the high-speed parallel robot. In order to restrain the elastic displacement (FD, flexible displacement) of the high-speed parallel robot and the residual vibration (RV), this paper, from the point of the trajectory planning, considers the flexibility of the rod. The problem of residual vibration suppression and elastic displacement limitation for a given track and point-to-point rapid positioning is studied. Aiming at the problem of residual vibration suppression in a given track, considering the characteristics of the change of the frequency of each step of the parallel robot along with the position, the multi-modal input shaping (IS) and the particle swarm optimization (PSO) and the control are combined, and the residual vibration is established as an optimization target. The parameter of the multi-modal input shaper is the optimization model of the optimization variable, and the control model is optimized off-line by using the PSO. The simulation results show that the optimized shaper can obviously restrain the residual vibration, and the residual vibration is further reduced with the increase of the number of the shaper, and the problem of large calculation amount brought by the prior method to the real-time updating of the parameter of the shaper is avoided. for the fast point-to-point motion, the time optimal planning and the multi-mode input shaping are combined, a two-step optimization method is proposed, and the time optimal problem is firstly solved by adopting a section pseudo-spectrum method (GPM, Gauss pseudo-spectrum method), The results show that the two-step optimization method achieves the limit of the residual vibration and the elastic displacement when the time is optimal. In order to solve the problem of tracking control of high-speed light (HSLW, high-speed and light-weight) parallel robot with flexible link, a compound control algorithm based on integral manifold and high gain observer is proposed. firstly, a small parameter is introduced according to a stiffness matrix, a rigid-flexible coupling dynamic model is reduced to a fast and slow two subsystems based on an integral manifold, At the same time, considering the influence of the elastic displacement of the rod on the end track, the correction moment is designed, and the elastic compensation of the elastic displacement of the rod on the end movement of the robot is realized. The rapid sub-system is controlled by a sliding mode variable structure, so that the manifold is established. In order to solve the problem that the rate of change of curvature cannot be measured directly, a high-gain observer is introduced and the stability is proved. For the above-mentioned compound control algorithm, the stability of the whole system is proved, and the selection range of the small parameters is given. Finally, in order to verify the effectiveness of the compound control algorithm, the design algorithm is compared with the singular perturbation and the inversion control algorithm based on the rigid body dynamics. The simulation results show that, In this paper, the compound control method based on the integral manifold and the observer has obvious advantages in the aspects of vibration suppression and track tracking. In order to verify the above-mentioned theoretical research, the design of the 3RRR parallel robot and the selection of the drive and transmission components are carried out with reference to the optimized design results, and the control system is designed by the control architecture of the industrial personal computer, the real-time operating system and the high-speed communication bus. In this paper, a laser tracker is used to test the repeated precision of the robot, and the model test and the acceleration response test are carried out to the system by using the LMS vibration tester to verify the modeling accuracy. In this paper, the planning method based on multi-mode input shaping and PSO and the trajectory planning method based on the segmentation pseudo-spectrum method and the multi-mode input shaping are verified.
【学位授予单位】：哈尔滨工业大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP242

【相似文献】