基于频率校正的机器人动作周期提取研究
发布时间：2018-08-22 11:43
【摘要】：随着工业自动化程度的不断扩大,关节型工业机器人广泛应用于自动化工厂。为了防止工业机器人机械系统发生故障,运用声发射、振动等技术对其进行状态检测是非常重要的。关节型机器人的机械臂的运动通常是向前和前后摆动的,因此,所测信号的幅值是随着它的旋转动作而变化的。当旋转速度加大时,信号强度也随之加强;当转速减小或者变向时,信号强度随之变小。在对信号做进一步研究分析之前,将信号按照整周期截取出来将是很必要的,它的关键技术在于确定周期性信号的频率和信号动作的起始点。论文针对该问题,首先利用相位差频谱校正法修正实测信号的包络谱得出机器人摆动动作的精确频率,进而得到信号的周期时间长度。而对机器人动作周期信号周期起始位置的确定,本文提出了两种方法:第一种是通过定位一个周期内的最低点来得到机器人动作的起始位置。这一方法中,使用低通滤波器对包络信号进行滤波处理,以便得到光滑的包络曲线。滤波器的截止频率的选取,是根据其能否保留包络信号中大部分能量来决定的;第二种是利用互相关法提取信号的初始相位,然后将其转化到时间上,得到信号的周期起始位置。最终,机器人摆动动作的整个周期就能够被划分出来。论文对提出的方法进行了仿真和实验信号的有效验证。
[Abstract]:With the continuous expansion of industrial automation, joint industrial robots are widely used in automation plants. In order to prevent the mechanical system of industrial robot from malfunction, it is very important to use acoustic emission and vibration technology to detect its state. The motion of the manipulator is usually forward and forward swing, so the amplitude of the measured signal varies with the rotation of the robot. When the rotation speed increases, the signal intensity also increases, and when the rotational speed decreases or changes direction, the signal intensity becomes smaller. Before further study and analysis of the signal, it is necessary to intercept the signal according to the whole period. Its key technology is to determine the frequency of the periodic signal and the starting point of the signal action. In order to solve this problem, the phase difference spectrum correction method is first used to correct the envelope spectrum of the measured signal to obtain the precise frequency of the robot's oscillating action, and then the period time length of the signal is obtained. In this paper, two methods are proposed to determine the starting position of the cycle signal. The first method is to get the starting position of the robot by locating the lowest point in a period. In this method, the low pass filter is used to filter the envelope signal to obtain a smooth envelope curve. The cutoff frequency of the filter is determined by whether it can retain most of the energy in the envelope signal. The second is to extract the initial phase of the signal by using cross-correlation method, and then convert it to time. Get the starting position of the signal cycle. Eventually, the whole cycle of the robot's oscillating motion can be divided. The method is validated by simulation and experiment.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242
本文编号：2196998
[Abstract]:With the continuous expansion of industrial automation, joint industrial robots are widely used in automation plants. In order to prevent the mechanical system of industrial robot from malfunction, it is very important to use acoustic emission and vibration technology to detect its state. The motion of the manipulator is usually forward and forward swing, so the amplitude of the measured signal varies with the rotation of the robot. When the rotation speed increases, the signal intensity also increases, and when the rotational speed decreases or changes direction, the signal intensity becomes smaller. Before further study and analysis of the signal, it is necessary to intercept the signal according to the whole period. Its key technology is to determine the frequency of the periodic signal and the starting point of the signal action. In order to solve this problem, the phase difference spectrum correction method is first used to correct the envelope spectrum of the measured signal to obtain the precise frequency of the robot's oscillating action, and then the period time length of the signal is obtained. In this paper, two methods are proposed to determine the starting position of the cycle signal. The first method is to get the starting position of the robot by locating the lowest point in a period. In this method, the low pass filter is used to filter the envelope signal to obtain a smooth envelope curve. The cutoff frequency of the filter is determined by whether it can retain most of the energy in the envelope signal. The second is to extract the initial phase of the signal by using cross-correlation method, and then convert it to time. Get the starting position of the signal cycle. Eventually, the whole cycle of the robot's oscillating motion can be divided. The method is validated by simulation and experiment.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242
【参考文献】
相关期刊论文 前10条
1 陈亮;;IFR发布最新全球服务机器人统计报告[J];机器人产业;2016年03期
2 范学明;;机械设备故障诊断与监测的常用方法及其发展趋势[J];化工管理;2014年23期
3 李明;廖明夫;王四季;;风力机现场动平衡的信号预处理方法研究[J];噪声与振动控制;2014年03期
4 王田苗;陶永;;我国工业机器人技术现状与产业化发展战略[J];机械工程学报;2014年09期
5 郭俊华;伍星;柳小勤;闫天延;;转子动平衡中振动信号幅值相位的提取方法研究[J];机械与电子;2011年10期
6 李建民;赵鹏;侯文;郑宾;郑浩鑫;;基于相关理论的相位差算法的误差研究[J];中北大学学报(自然科学版);2009年06期
7 韩长生;;基于LabVIEW的动不平衡信号提取技术研究[J];机床与液压;2009年07期
8 熊建国;;工业机器人的应用和发展趋势[J];才智;2009年01期
9 胡劲松;杨世锡;郭荣;;基于模式匹配的振动信号整周期采样方法[J];振动、测试与诊断;2008年02期
10 曹雏清;陈国聪;;刚性转子现场动平衡仪的设计与开发[J];机械制造;2007年10期
相关硕士学位论文 前2条
1 杨平;声发射技术在轴承故障诊断中的应用[D];沈阳工业大学;2009年
2 杨黎明;声发射技术用于段修货车轴承故障诊断研究[D];西南交通大学;2003年
，本文编号：2196998
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2196998.html
