摘要
由于工业机器人的灵活性,被广泛应用于机器人焊缝磨削任务中。但由于机器人的弱刚性,在焊缝磨削过程中系统容易发生颤振,因此对加工过程中的颤振监测是保证加工质量的基础。针对在加工振动信号处理过程中的模态混叠现象,提出了一种基于排列熵算法改进的经验模态分解方法,通过排列熵算法检测振动信号中的异常信号并剔除。通过相关系数法提取相关性最大的固有模态函数的能量熵作为特征值,同时提取方差、峰峰值、均方根和峭度4种时域特征。利用遗传算法优化BP神经网络(back propagation neural network,BPNN)建立颤振辨识模型,最后将提取的5种特征参数作为特征向量代入辨识模型中对加工状态进行监测。试验结果显示,提出的改进经验模态分解算法结合遗传算法优化的BPNN模型能够有效地对机器人焊缝磨削中的颤振进行监测。
Industrial robots are widely used in robot welding seam grinding tasks due to their flexibility.However,due to their weak rigidity,the system is prone to chattering in welding seam grinding process.Therefore,monitoring chattering in machining process is the basis to ensure machining quality.Here,aiming at modal aliasing phenomena in processing vibration signals,an improved empirical mode decomposition method based on permutation entropy algorithm was proposed.The permutation entropy algorithm was used to detect and eliminate abnormal signals in vibration signals.The correlation coefficient method was used to extract the energy entropy of the intrinsic mode function with the largest correlation as eigenvalue.Meanwhile,4 time-domain features of variance,peak to peak value,root mean square and kurtosis were extracted.The genetic algorithm was used to optimize BP neural network and establish a chattering recognition model.Finally,the extracted 5 feature parameters were taken as feature vectors to input into the recognition model for monitoring machining state.Test results showed that the proposed improved empirical mode decomposition algorithm combined with BP neural network model optimized with genetic algorithm can effectively monitor chatter in robot weld seam grinding.
作者
刘伟
刘旺
曹大虎
葛吉民
万林林
陈加
LIU Wei;LIU Wang;CAO Dahu;GE Jimin;WAN Linlin;CHEN Jia(Hunan Provincial Key Laboratory of High Efficiency and Precision Machining of Difficult-to-cut Material,School of Mechanical Engineering,Hunan University of Science and Technology,Xiangtan 411201,China)
出处
《振动与冲击》
EI
CSCD
北大核心
2024年第9期131-138,174,共9页
Journal of Vibration and Shock
基金
湖南省自然科学基金省市联合基金(2021JJ50116)。
关键词
机器人磨削
颤振监测
改进经验模态分解
遗传算法
BP神经网络
robot grinding
chatter monitoring
improved empirical mode decomposition(EMD)
genetic algorithm
BP neural network
作者简介
第一作者:刘伟,男,博士,副教授,1986年生。
