基于ERNIE的数控机床故障分类研究  

Research on fault classification of CNC machine tools based on ERNIE

作　　者：张嘉琛 褚燕华[1] 王丽颖[1] ZHANG Jiachen;CHU Yanhua;WANG Liying(School of Digital and Intelligent Industry,Inner Mongolia University of Science&Technology,Baotou 014010,China)

机构地区：[1]内蒙古科技大学数智产业学院,内蒙古包头014010

出　　处：《内蒙古科技大学学报》2024年第4期371-377,共7页Journal of Inner Mongolia University of Science and Technology

基　　金：内蒙古自治区直属高校基本科研业务费项目(2024XKJX018)。

摘　　要：数控机床故障诊断中存在效率低、准确率不高的问题,为提高数控机床的维护效率,准确快速地识别和分类故障文本信息,研究提出基于深度学习的ERNIE-CoCBi-Att模型,该模型利用ERNIE进行词向量嵌入,采用Concat操作并引入注意力机制,提高分类的准确性。实验结果表明,ERNIE预训练模型能够捕捉到更为丰富的语义信息,BiGRU能够有效捕捉上下文信息,CNN则在有效提取局部重要特征表现出色,这些特性的共同作用,使得ERNIE-CoCBi-Att模型在处理故障文本分类任务时优势显著。与对应改进前模型相比,正确率(Accuracy)、精确率(Precision)、召回率(Recall)和F1值(F1-Score)4项指标均提升0.5%以上。The problems of low efficiency and low accuracy exist in the fault diagnosis of CNC machine tools.In order to improve the maintenance efficiency of CNC machine tools,identify and classify fault text information accurately and quickly,this paper proposes the ERNIE-CoCBi-Att model based on deep learning,which uses ERNIE to carry out word vector embedding,Meanwhile,concat operation and attention mechanism are adopted to improve the accuracy of classification.The experimental results show that ERNIE pretrained model can capture richer semantic information,BiGRU performs well in capturing context information and CNN can effectively extract local important features.Together,these features make ERNIE-CoCBi-Att model have significant advantages in dealing with fault text classification tasks.Compared with the model before corresponding improvement,the four indexes of Accuracy,Precision,Recall and F1-Score were all increased by more than 0.5%.

关 键 词：数控机床 深度学习 文本分类 注意力机制 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]