采用HHT与CNN的刀具磨损状态监测  

Tool Wear Monitoring Method Based on Hilbert Huang Transform and Convolutional Neural Network

作　　者：周粤 段现银 ZHOU Yue;DUAN Xianyin(School of Mechanical Automation,Wuhan University of Science and Technology,Wuhan 430081,China;Hubei Key Laboratory of Mechanical Transmission and Manufacturing Engineering,Wuhan University of Science and Technology,Wuhan 430081,China)

机构地区：[1]武汉科技大学机械自动化学院,湖北430081 [2]武汉科技大学机械传动与制造工程湖北省重点实验室,湖北430081

出　　处：《组合机床与自动化加工技术》2023年第10期169-173,178,共6页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：湖北省重点研发计划项目(2022BAA059)。

摘　　要：为实现刀具磨损状态监测的智能化及高精度,提出一种将希尔伯特黄变换改进算法与卷积神经网络相结合的监测方法。首先,采用敏感固有模态函数筛选方法改进希尔伯特黄变换,提取出刀具信号的时频特征;其次,运用MATLAB微调卷积神经网络生成刀具磨损监测迁移模型;最后,运用典型的卷积神经网络迁移模型进行实验验证。结果表明,与传统时频变换相比,希尔伯特黄变换提取的时频图更加精细,能有效防止频谱泄露的问题,刀具磨损识别平均准确率达到94.09%,提升近15%;与希尔伯特黄变换相比,改进后的希尔伯特黄变换能避免虚假固有模态分量问题,监测效果进一步提升,达到96.8%,证明了所提监测方法的有效性。In order to realize the automation and high precision of tool wear condition monitoring,a tool wear monitoring method based on improved Hilbert Huang transform algorithm and convolution neural network is proposed.First,the Hilbert Huang transform is improved by using the sensitive natural mode function screening method to extract the time-frequency characteristics of the tool vibration signal,and then the finely tuned convolutional neural network migration model is used to predict the tool wear state.The experimental results on three typical convolutional neural network models show that,compared with traditional time-frequency transform,Hilbert Huang transform can effectively prevent the problem of spectrum leakage,the extracted time-frequency map is more refined,and the average accuracy of tool wear identification reaches 94.09%,increasing by nearly 15%;Compared with the Hilbert Huang transform,the improved Hilbert Huang transform can avoid false natural mode components,and the prediction effect is further improved,reaching 96.8%,which verifies the effectiveness of this method for tool wear monitoring.

关 键 词：希尔伯特黄变换 卷积神经网络 敏感固有模态函数 刀具磨损状态监测 

分 类 号：TH16[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]