机器人砂带磨削复杂型面零件技术研究

发布时间：2018-07-12 17:36

本文选题：机器人 + 砂带磨削　；参考：《哈尔滨工业大学》2017年硕士论文

【摘要】：本文的主要研究内容是使用机器人抓持水龙头对其工件表面进行磨削加工。基于人机和美学的考虑,水龙头表面多为复杂曲面,在以往的水龙头表面磨削加工过程中多采用人工磨削,但是人工磨削加工尺寸一致性差,并且磨削过程中产生的金属粉尘对工人的健康危害极大。采用专用磨床加工,解决了人工磨削的诸多弊端,但是其可拓展性差,不适合水龙头这类造型频繁变化的零件。采用机器人加工尽管在精度方面与专用磨床相比有所不及,但是其加工柔性高,非常适合加工水龙头这类造型变化频繁,批量不大,对尺寸精度与形状精度要求并不高的零件。因此本课题采用机器人砂带磨削的方式对水龙头表面进行磨削加工。针对于水龙头造型频繁变化,新造型水龙头磨削调整时间长的问题,利用BP神经网络建立磨削过程模型,实现磨削参数的离线设定,缩短新造型水龙头磨削所需要的试磨时间。基于以上设想,选择对磨削表面粗糙度影响最大的三个因素:砂带目数,砂带速度,磨削正压力,进行了大量的基础工艺试验,探究以上三个因素对磨削表面粗糙度的影响。对这些实验数据进行归一化处理,将处理后的数据作为训练样本,对神经网络进行训练,利用训练好的神经网络对磨削参数进行预测,并利用MATLAB GUI工具箱编写了磨削参数预测界面,实现离线磨削参数设定。并将离线设定的磨削参数运用到实际的水龙头磨削加工中,达到缩短新工件调整时间的目的。针对水龙头曲面造型复杂,难以编写磨削轨迹的问题,本文中采用离线编程软件与示教器混合编程的方式。在离线编程软件RobotStudio中实现对加工轨迹的精确编程,但是对于非加工轨迹以及相关外围信号配置则通过示教器完成,结合两者的优势,缩短编程时间。针对加工区域轨迹的编程,本文中根据水龙头表面不同部分的特征,划分不同的磨削区域,选择对应的磨削方法。使用RobotStudio中的Machining PowerPac轨迹编程插件完成磨削轨迹的编程。在RobotStudio软件中对编制好的磨削轨迹进行仿真验证,检查有无干涉碰撞以及机器人关节超限的问题,仿真检查无误后,将磨削轨迹程序输入到机器人控制柜中,进行慢速磨削轨迹验证,检查无误后进行实际加工。观察机器人砂带磨削后的水龙头零件,表面磨削一致性好,曲面光滑平顺,达到了预想设定的磨削效果。
[Abstract]:The main research content of this paper is grinding the workpiece surface with robot holding faucet. Based on human-computer and aesthetic considerations, the surface of faucet is mostly complex surface, and manual grinding is used in the previous grinding process of faucet surface, but the dimension consistency of manual grinding is poor. And the metal dust produced during grinding is harmful to workers' health. The malpractice of manual grinding is solved by using special grinding machine, but its expansibility is poor, so it is not suitable for parts with frequently changing shape such as faucet. Although the robot is inferior to the special grinding machine in precision, it has high flexibility, which is very suitable for machining the parts with such kinds of shapes as faucet, such as frequent change, small batch, and low requirement of dimensional precision and shape precision. Therefore, the robot belt grinding method is used to grinding the faucet surface. In order to solve the problem of frequent change of tap shape and long grinding adjustment time of new modeling faucet, the grinding process model is established by using BP neural network to realize off-line setting of grinding parameters and to shorten the test grinding time required for new modeling faucet grinding. Based on the above assumption, three factors that have the greatest influence on the grinding surface roughness are selected: the number of belt mesh, the speed of the belt, and the grinding positive pressure. A large number of basic technological tests are carried out to explore the effect of the above three factors on the grinding surface roughness. These experimental data are normalized, the processed data are taken as training samples, the neural network is trained, and the grinding parameters are predicted by the trained neural network. The prediction interface of grinding parameters is compiled by MATLAB GUI toolbox to realize the parameter setting of off-line grinding. The off-line grinding parameters are applied to the actual grinding process of the faucet to shorten the adjustment time of the new workpiece. In view of the complex surface modeling of the faucet, it is difficult to write the grinding track. In this paper, the mixed programming mode of off-line programming software and teaching device is adopted. In the off-line programming software Robot Studio, the precise programming of the machining trajectory is realized, but the non-machining trajectory and the related peripheral signal configuration are accomplished by the teacher, combining the advantages of the two, the programming time is shortened. According to the programming of machining area, according to the characteristics of different parts of faucet surface, different grinding areas are divided and the corresponding grinding methods are selected in this paper. Using Machining PowerPac trajectory programming plug-in Robot Studio to complete grinding trajectory programming. In the software of Robot Studio, the grinding track is simulated and verified to check whether there are interference collisions and the problems of the robot joint exceeding the limit. After the simulation check, the grinding track program is input into the robot control cabinet. To verify the slow grinding track, check the accuracy of the actual processing. After grinding the faucet parts with robot abrasive belt, the surface grinding consistency is good and the surface is smooth and smooth, which achieves the desired grinding effect.
【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;TG580.619.2

【参考文献】