基于传感器数据的CNC机床设备故障行为分析与特征提取方法研究  

Research on Fault Behavior Analysis and Feature Extraction Methods of CNC Machine Tools Based on Sensor Data

作　　者：郭钊 申景凤 GUO Zhao;SHEN Jingfeng(CRRC Nanjing Puzhen Co.,Ltd.,Nanjing Jiangsu 210031)

机构地区：[1]中车南京浦镇车辆有限公司,江苏南京210031

出　　处：《中国科技纵横》2024年第20期108-110,共3页China Science & Technology Overview

摘　　要：本文深入研究基于传感器数据的CNC机床设备故障行为分析和特征提取方法。首先,对CNC机床进行概述,包括其定义、故障数据来源及预处理方式。接着,详细探讨故障数据的聚类分析以及多种故障行为诊断方法,如时域、频域、时频、时变和多尺度特征分析等。在特征选择部分,本文阐述特征选择的过程、方法及结果,成功识别出一系列具有代表性的故障特征值。这些研究成果为CNC机床设备的故障预测与诊断提供了坚实依据,有助于提高设备运行效率和稳定性,降低维护成本。未来研究应继续探索更多有效的故障检测与诊断方法,推动现代工业制造领域发展。This paper deeply studies the fault behavior analysis and feature extraction methods of CNC machine tools based on sensor data.First,an overview of the CNC machine is given,including its definition,fault data sources and pre-processing.Then,the cluster analysis of fault data and a variety of fault behavior diagnosis methods,such as timedomain,frequency-domain,time-frequency,time-varying and multi-scale feature analysis,are discussed in detail.In the feature selection part,this paper expounds the process,method and results of feature selection,and successfully identifies a series of representative fault characteristic values.These research results provide a solid basis for the fault prediction and diagnosis of CNC machine tool equipment,which helps to improve the operation efficiency and stability of the equipment and reduce the maintenance cost.In the future,more effective fault detection and diagnosis methods should be explored to promote the development of modern industrial manufacturing.

关 键 词：传感器 CNC 特征提取 

分 类 号：TG659[金属学及工艺—金属切削加工及机床]