本文選題：機(jī)器人 + 砂帶磨削��； 參考：《哈爾濱工業(yè)大學(xué)》2017年碩士論文

【摘要】：本文的主要研究內(nèi)容是使用機(jī)器人抓持水龍頭對其工件表面進(jìn)行磨削加工。基于人機(jī)和美學(xué)的考慮,水龍頭表面多為復(fù)雜曲面,在以往的水龍頭表面磨削加工過程中多采用人工磨削,但是人工磨削加工尺寸一致性差,并且磨削過程中產(chǎn)生的金屬粉塵對工人的健康危害極大。采用專用磨床加工,解決了人工磨削的諸多弊端,但是其可拓展性差,不適合水龍頭這類造型頻繁變化的零件。采用機(jī)器人加工盡管在精度方面與專用磨床相比有所不及,但是其加工柔性高,非常適合加工水龍頭這類造型變化頻繁,批量不大,對尺寸精度與形狀精度要求并不高的零件。因此本課題采用機(jī)器人砂帶磨削的方式對水龍頭表面進(jìn)行磨削加工。針對于水龍頭造型頻繁變化,新造型水龍頭磨削調(diào)整時(shí)間長的問題,利用BP神經(jīng)網(wǎng)絡(luò)建立磨削過程模型,實(shí)現(xiàn)磨削參數(shù)的離線設(shè)定,縮短新造型水龍頭磨削所需要的試磨時(shí)間。基于以上設(shè)想,選擇對磨削表面粗糙度影響最大的三個(gè)因素:砂帶目數(shù),砂帶速度,磨削正壓力,進(jìn)行了大量的基礎(chǔ)工藝試驗(yàn),探究以上三個(gè)因素對磨削表面粗糙度的影響。對這些實(shí)驗(yàn)數(shù)據(jù)進(jìn)行歸一化處理,將處理后的數(shù)據(jù)作為訓(xùn)練樣本,對神經(jīng)網(wǎng)絡(luò)進(jìn)行訓(xùn)練,利用訓(xùn)練好的神經(jīng)網(wǎng)絡(luò)對磨削參數(shù)進(jìn)行預(yù)測,并利用MATLAB GUI工具箱編寫了磨削參數(shù)預(yù)測界面,實(shí)現(xiàn)離線磨削參數(shù)設(shè)定。并將離線設(shè)定的磨削參數(shù)運(yùn)用到實(shí)際的水龍頭磨削加工中,達(dá)到縮短新工件調(diào)整時(shí)間的目的。針對水龍頭曲面造型復(fù)雜,難以編寫磨削軌跡的問題,本文中采用離線編程軟件與示教器混合編程的方式。在離線編程軟件RobotStudio中實(shí)現(xiàn)對加工軌跡的精確編程,但是對于非加工軌跡以及相關(guān)外圍信號配置則通過示教器完成,結(jié)合兩者的優(yōu)勢,縮短編程時(shí)間。針對加工區(qū)域軌跡的編程,本文中根據(jù)水龍頭表面不同部分的特征,劃分不同的磨削區(qū)域,選擇對應(yīng)的磨削方法。使用RobotStudio中的Machining PowerPac軌跡編程插件完成磨削軌跡的編程。在RobotStudio軟件中對編制好的磨削軌跡進(jìn)行仿真驗(yàn)證,檢查有無干涉碰撞以及機(jī)器人關(guān)節(jié)超限的問題,仿真檢查無誤后,將磨削軌跡程序輸入到機(jī)器人控制柜中,進(jìn)行慢速磨削軌跡驗(yàn)證,檢查無誤后進(jìn)行實(shí)際加工。觀察機(jī)器人砂帶磨削后的水龍頭零件,表面磨削一致性好,曲面光滑平順,達(dá)到了預(yù)想設(shè)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.
【學(xué)位授予單位】：哈爾濱工業(yè)大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TP242;TG580.619.2

【參考文獻(xiàn)】

相關(guān)期刊論文 前8條

1 齊俊德;張定華;李山;;考慮去除深度的砂帶磨削路徑規(guī)劃方法[J];機(jī)械科學(xué)與技術(shù);2016年09期

2 張海洋;楊文玉;張家軍;陳巍;;葉片機(jī)器人砂帶磨拋的軌跡規(guī)劃研究[J];機(jī)電工程;2014年05期

3 陳廉清;郭建亮;楊勛;遲軍;趙霞;;基于進(jìn)化神經(jīng)網(wǎng)絡(luò)的磨削粗糙度預(yù)測模型[J];計(jì)算機(jī)集成制造系統(tǒng);2013年11期

4 齊立哲;甘中學(xué);孫云權(quán);湯青;，

本文編號：2117965