数控机床电动主轴WPD-TSNE-SVM模型故障诊断
Fault Diagnosis of CNC Machine Tool Electric Spindle with WPD-TSNE-SVM Model
李坤宏 1江桂云 2朱代兵3
作者信息
- 1. 重庆工业职业技术学院机械工程与自动化学院,重庆 401120
- 2. 重庆大学机械与运载工程学院,重庆 400030
- 3. 重庆红江机械有限责任公司,重庆 402162
- 折叠
摘要
为了提高数控机床电动主轴故障诊断效率,设计了一种WPD-TSNE-SVM组合模型.利用小波包方法分解主轴振动信号,并完成样本集TSNE降维的过程,利用SVM完成重构特征的故障分类.构建数控机床主轴信号混合特征空间向量,并进行故障诊断分析.研究结果表明:TSNE方法训练样数据形成规律分布特点,采用非线性SVM多故障分类器实现小波包混合特征的故障准确分类.根据径向基核函数建立的非线性SVM诊断方法获得更高准确率.该方法诊断轴承运行故障,获得更高维护效率,确保数控机床主轴运行稳定性.
Abstract
In order to improve the fault diagnosis efficiency of motorized spindle of NC machine tool,a WPD-TSNE-SVM combined model was designed.The main shaft vibration signal is decomposed by using the wavelet packet method,and the dimensionality reduction process of sample set TSNE is completed,and the fault classification of reconstructed features is completed via SVM.The mixed feature space vector of NC machine tool spindle signal was constructed,and the fault diagnosis was analyzed.The results show that the training sample data of TSNE method form regular distribution characteristics,and nonlinear SVM multi-fault classifier is used to achieve the accurate fault classification of wavelet packet mixed features.The nonlinear SVM diagnosis method based on the radial basis kernel function can achieve the higher accuracy.This method can diagnose the running faults of bearings,obtain higher maintenance efficiency,and ensure the running stability of CNC machine tool spindle.
关键词
数控机床/电动主轴/故障诊断/小波包分解Key words
CNC machine tool/electric spindle/fault diagnosis/wavelet packet decomposition引用本文复制引用
基金项目
重庆市科技计划应用开发重大项目(cstc2015yykfC40001)
出版年
2024
国家科技期刊平台
NSTL
万方数据