面向小型金属铸件的机器人自动抛磨关键技术研究

发布时间：2018-08-13 14:12

【摘要】：目前,小型金属铸件的抛磨以人工作业为主,抛磨环境非常恶劣,并且人工效率比较低,工件均一性不佳,抛磨质量难以得到保证。随着人工成本的增加以及对工人健康防护的重视,人工抛磨势必会被机器抛磨甚至自动化抛磨所替代。本文结合国家863计划项目子课题“面向精密抛磨的机器人力位混合控制技术研究与工程实现”,对面向小型金属铸件的机器人自动抛磨关键技术,包括机器人自动抛磨工艺分析和系统设计、抛磨质量建模以及机器人力控制策略进行研究。首先,结合人工抛磨经验,进行了小型金属铸件的抛磨工艺分析和离线仿真分析。根据力控制精度要求,设计了被动力控砂带机结构和主动力控机器人末端工具,提出了基于力传感器的机器人主动力控制和基于气缸缓冲的砂带机被动力控制相结合的机器人自动抛磨系统,并搭建了主被动力控制机器人自动抛磨控制系统,奠定了本课题的硬件基础。然后,通过剖析抛磨质量的各个影响因素,提出了以砂带速度、机器人速度、抛磨力和工件表面曲率为模型自变量,表面粗糙度为模型因变量的基于支持向量回归抛磨质量模型。采用遍历、遗传和粒子群三种算法对抛磨质量模型参数分别进行寻优,遴选速度更快、均方误差更小的算法。之后利用寻优参数进行了抛磨质量建模和实验验证。基于所研制的机器人自动抛磨系统,提出了相应的机器人抛磨主被动力控制策略。通过实验比较中值滤波、均值滤波和卡尔曼滤波三种滤波算法,选出抗干扰能力强且预测性能好的算法,并确定其关键参数。通过推导工具重力补偿公式,解决其位姿变化对传感器的影响。提出笛卡尔空间机器人抛磨阻抗算法,并探讨其阻抗参数对控制系统的影响。针对抛磨作业环境多变问题,提出面向机器人抛磨的模糊阻抗算法,并通过实验验证其鲁棒性。最后,根据方案和算法设计,搭建了机器人自动抛磨系统标定实验平台,并设计了机器人抛磨人机交互软件,包括系统通讯、运动控制、力监控和力控制等模块。利用激光跟踪仪进行机器人自动抛磨系统标定实验,开展机器人抛磨重力补偿实验、恒力控制实验和机器人自动抛磨实验验证。实验结果表明,该机器人自动抛磨系统的力控制精度在±5N以内,恒力抛磨的工件光洁度更高、表面粗糙度更低、抛磨表面连续均匀性更好。
[Abstract]:At present, the grinding of small metal castings is dominated by manual work. The polishing environment is very bad, and the manual efficiency is low, the uniformity of workpiece is not good, and the quality of polishing is difficult to be guaranteed. With the increase of labor cost and the emphasis on health protection of workers, manual grinding will be replaced by machine polishing or even automatic grinding. In this paper, the key technology of robot automatic grinding for small metal castings is discussed, which is a subproject of the national 863 project, "Research and engineering realization of robot force position hybrid control technology for precision grinding". It includes automatic polishing process analysis and system design, grinding quality modeling and robot force control strategy. Firstly, the grinding process and off-line simulation of small metal castings are analyzed based on manual grinding experience. According to the requirement of force control precision, the structure of dynamic sand belt machine and the end tool of main power control robot are designed. A robot automatic polishing and grinding system based on force sensor and pneumatic cylinder buffer is proposed, and the automatic grinding control system is built. Laid the hardware foundation of this subject. Then, by analyzing the factors that affect the grinding quality, the paper puts forward that the sand belt velocity, robot speed, polishing force and surface curvature of workpiece are the independent variables of the model. Support vector regression based polishing quality model with surface roughness as a dependent variable. The ergodic, genetic and particle swarm optimization algorithms are used to optimize the parameters of the polishing quality model respectively. The selection speed is faster and the mean square error is smaller. After that, the grinding quality modeling and experimental verification are carried out by optimizing parameters. Based on the developed automatic polishing system for robot, a dynamic control strategy for the main body of robot polishing is proposed. Three filtering algorithms, median filter, mean filter and Kalman filter, are compared experimentally. The algorithms with strong anti-interference ability and good prediction performance are selected, and the key parameters are determined. By deducing the formula of gravity compensation, the influence of the position and pose change on the sensor is solved. A grinding impedance algorithm for Cartesian space robot is proposed, and the influence of impedance parameters on the control system is discussed. A fuzzy impedance algorithm for robot polishing is proposed and its robustness is verified by experiments. Finally, according to the scheme and algorithm design, the calibration experiment platform of robot automatic polishing system is built, and the man-machine interactive software of robot polishing is designed, including system communication, motion control, force monitoring and force control, and so on. The laser tracker is used to calibrate the robot automatic polishing system, to carry out the robot polishing gravity compensation experiment, the constant force control experiment and the robot automatic polishing experiment verification. The experimental results show that the force control accuracy of the automatic polishing system is less than 卤5N, the finish of the workpiece with constant force polishing is higher, the surface roughness is lower, and the continuous uniformity of the polishing surface is better.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】