基于KPCA-PSO-LSSVM的轴承寿命预测研究
Research on Bearing Remaining Useful Life Prediction Based on KPCA-PSO-LSSVM
丁国荣 1王文波 1赵姣姣1
作者信息
- 1. 武汉科技大学理学院 武汉 430065
- 折叠
摘要
为了预测不同工况下对于滚动轴承的最大剩余使用寿命(RUL),提出了一种基于核主成分分析(KPCA)结合粒子群优化最小二乘支持向量机(PSO-LSSVM)的滚动轴承RUL预测框架.该方法首先从时域、频域以及小波包域进行轴承故障特征提取,得到一系列退化特征;其次,在尽可能多保留退化特征的前提下,运用KPCA方法进行特征约简;最后采用PSO-LSSVM构建结合的模型来预测滚动轴承的RUL.通过美国智能维护中心(IMS)提供的多组轴承衰退振动信号对模型进行验证,实验结果表明,相比较于PSO-LSSVM和KPCA-LSSVM模型,论文提出的KPCA-PSO-LSSVM的轴承剩余寿命预测方法具有更低的预测误差,可以比较准确出拟合滚动轴承的退化情况.
Abstract
In order to predict the remaining useful life(RUL)of rolling bearings under different working conditions,a rolling bearing RUL prediction framework based on kernel principal component analysis(KPCA)combined with particle swarm optimiza-tion least squares support vector machine(PSO-LSSVM)is proposed.Firstly,fault features from the time domain,frequency do-main and wavelet packet domain are extracted to obtain a series of degraded features.Secondly,on the premise of retaining as many degraded features as possible,KPCA method is used to reduce features.Finally,PSO-LSSVM is used to build a combined model to predict the RUL of rolling bearings.The model is verified by multiple sets of bearing recession vibration signals provided by the intel-ligent maintenance system(IMS).The experimental results show that compared with the PSO-LSSVM and KPCA-LSSVM models,the remaining life forecasting approach of the bearing of the KPCA-PSO-LSSVM proposed in this paper has a lower forecasting er-ror,which can accurately determine the degradation of the fitting rolling bearing.
关键词
剩余寿命预测/KPCA-PSO-LSSVM/退化特征提取Key words
remaining useful life prediction/KPCA-PSO-LSSVM/degradation feature extraction引用本文复制引用
基金项目
国家自然科学基金面上项目(61671338)
国家自然科学基金面上项目(51877161)
出版年
2024
国家科技期刊平台
NSTL
万方数据