基于数字孪生与BP神经网络的RV减速器故障诊断  

RV Reducer fault diagnosis based on digital twin and BP neural network

作　　者：孙笑笑 江本赤 陈智君 李公文 SUN Xiaoxiao;JIANG Benchi;CHEN Zhijun;LI Gongwen(School of Mechanical and Automotive Engineering,Anhui Polytechnic University,Wuhu 241000,China;Yangtze River Delta HIT Robot Technology Research Institute,Wuhu 241007,China;Wuhu Anpu Robot Technology Research Institute Co.,Ltd.,Wuhu 241060,China)

机构地区：[1]安徽工程大学机械与汽车工程学院,安徽芜湖241000 [2]长三角哈特机器人产业技术研究院,安徽芜湖241007 [3]芜湖安普机器人产业技术研究院有限公司,安徽芜湖241060

出　　处：《安徽科技学院学报》2025年第2期104-111,共8页Journal of Anhui Science and Technology University

基　　金：机器视觉检测安徽省重点实验室开放基金项目(KLMVI-2023-HIT-04);芜湖市科技项目(2023pt08);安徽未来技术研究院企业合作项目(2023qyhz02)。

摘　　要：本文为解决RV减速器的故障诊断及分类问题,提出一种基于数字孪生和神经网络的故障诊断方法。首先,构建RV减速器的数字孪生模型,并建立数据通信完成虚实交互;其次,基于BP神经网络构建减速器故障诊断模型,并引入Adam算法对其优化;最后,结合仿真数据库和传感器实时数据对减速器孪生模型和诊断模型进行修正更新。训练集的诊断准确率达到96.5%,测试集的准确率达到96%,验证了该方法的可行性。本文通过融合数字孪生技术与BP神经网络,可以有效提升故障诊断的准确率。To propose a fault diagnosis method based on digital twin and neural network to solve the problem of RV reducer fault diagnosis and classification.Firstly,the digital twin model of RV reducer is constructed,and the data communication is established to complete the virtual-real interaction.Secondly,the fault diagnosis model of the reducer is constructed based on BP neural network,and the Adam algorithm is introduced to optimize it.Finally,the twin model and diagnostic model of the reducer are modified and updated by combining the simulation database and real-time data of the sensor.The experiment shows that the diagnostic accuracy of training set is 96.5%,and the accuracy of test set is 96%,which verifies the feasibility of this method.By integrating digital twin technology and BP neural network,the accuracy of fault diagnosis can be effectively improved.

关 键 词：数字孪生 神经网络 RV减速器 故障诊断 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]