Monitoring robot machine tool sate via neural ODE and BP-GA

IF 2.7 3区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

Measurement Science and Technology Pub Date : 2023-12-17 DOI:10.1088/1361-6501/ad166d

Guangyi Zhu, Xi Zeng, Zheng Gong, Zhuohan Gao, Renquan Ji, Yisen Zeng, Pei Wang, Congda Lu

{"title":"Monitoring robot machine tool sate via neural ODE and BP-GA","authors":"Guangyi Zhu, Xi Zeng, Zheng Gong, Zhuohan Gao, Renquan Ji, Yisen Zeng, Pei Wang, Congda Lu","doi":"10.1088/1361-6501/ad166d","DOIUrl":null,"url":null,"abstract":"Tool wear during robotic polishing affects material removal rates and surface roughness, leading to erratic and inconsistent polishing quality. Therefore, a method that can predict the tool state is needed to replace the robot end tool in time. In this paper, based on the cutting-edge neural ordinary differential equations (Neural ODE) and BP neural network optimization based on genetic algorithm (BP-GA), we propose a method to identify the tool state during robotic machining: firstly, a new training method of Neural ODE is proposed to avoid the model from falling into poor stationary points, and then on this basis, Neural ODE is utilized to predict the changes of vibration signals during robot machining; secondly, the predicted vibration signals of the tool are processed using variable modal decomposition method to extract the eigen kurtosis index and envelope entropy of the modal function as the vibration signal eigenvectors, and compare them with the traditional vibration signal eigenvectors. Finally, the predicted tool states were identified using BP-GA, and numerical experiments yielded an F1 score of 91.76% and an accuracy of 96.55% for model identification.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":"4 12","pages":""},"PeriodicalIF":2.7000,"publicationDate":"2023-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Measurement Science and Technology","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1088/1361-6501/ad166d","RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, MULTIDISCIPLINARY","Score":null,"Total":0}

引用次数: 0

Abstract

Tool wear during robotic polishing affects material removal rates and surface roughness, leading to erratic and inconsistent polishing quality. Therefore, a method that can predict the tool state is needed to replace the robot end tool in time. In this paper, based on the cutting-edge neural ordinary differential equations (Neural ODE) and BP neural network optimization based on genetic algorithm (BP-GA), we propose a method to identify the tool state during robotic machining: firstly, a new training method of Neural ODE is proposed to avoid the model from falling into poor stationary points, and then on this basis, Neural ODE is utilized to predict the changes of vibration signals during robot machining; secondly, the predicted vibration signals of the tool are processed using variable modal decomposition method to extract the eigen kurtosis index and envelope entropy of the modal function as the vibration signal eigenvectors, and compare them with the traditional vibration signal eigenvectors. Finally, the predicted tool states were identified using BP-GA, and numerical experiments yielded an F1 score of 91.76% and an accuracy of 96.55% for model identification.

查看原文本刊更多论文

通过神经 ODE 和 BP-GA 监控机器人机床状态

机器人抛光过程中的工具磨损会影响材料去除率和表面粗糙度，导致抛光质量不稳定和不一致。因此，需要一种能预测工具状态的方法来及时更换机器人终端工具。本文基于前沿的神经常微分方程（Neural ODE）和基于遗传算法的 BP 神经网络优化（BP-GA），提出了一种识别机器人加工过程中刀具状态的方法：首先，提出了一种新的神经 ODE 训练方法，以避免模型陷入较差的静止点，然后在此基础上利用神经 ODE 预测机器人加工过程中振动信号的变化；其次，利用变模态分解方法对预测的刀具振动信号进行处理，提取模态函数的特征峰度指数和包络熵作为振动信号特征向量，并与传统的振动信号特征向量进行比较。最后，利用 BP-GA 对预测的工具状态进行识别，数值实验结果表明，模型识别的 F1 得分为 91.76%，准确率为 96.55%。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Measurement Science and Technology 工程技术-工程：综合

CiteScore

4.30

自引率

16.70%

发文量

656

审稿时长

4.9 months

期刊介绍： Measurement Science and Technology publishes articles on new measurement techniques and associated instrumentation. Papers that describe experiments must represent an advance in measurement science or measurement technique rather than the application of established experimental technique. Bearing in mind the multidisciplinary nature of the journal, authors must provide an introduction to their work that makes clear the novelty, significance, broader relevance of their work in a measurement context and relevance to the readership of Measurement Science and Technology. All submitted articles should contain consideration of the uncertainty, precision and/or accuracy of the measurements presented. Subject coverage includes the theory, practice and application of measurement in physics, chemistry, engineering and the environmental and life sciences from inception to commercial exploitation. Publications in the journal should emphasize the novelty of reported methods, characterize them and demonstrate their performance using examples or applications.