广东石油化工学院学报2025，Vol.35Issue(6)：68-73,6.DOI:10.26962/j.cnki.1991.2025.0054
基于TTAO-VMD与CWT融合的轴承早期故障诊断方法
Bearing Early Fault Diagnosis Method Based on the Fusion of TTAO-VMD and CWT
摘要
Abstract
Aiming at the challenges of weak early fault signals in rolling bearings and susceptibility to environmental noise interference,this paper proposes a fault diagnosis method based on a dual-channel feature fusion approach combining Multi-Scale Convolutional Neural Network(MSCNN)and Long Short-Term Memory(LSTM).First,the Triangular Topology Aggregation Optimization(TTAO)algorithm is employed to optimize the penalty factor and mode components of Variational Mode Decomposition(VMD).The TTAO algorithm,known for its fast convergence and ability to avoid local optima,enables adaptive decomposition of bearing fault signals.Effective intrinsic mode functions are selected using the Spearman correlation coefficient,and the signal is reconstructed.The reconstructed one-dimensional signal and the two-dimensional time-frequency images generated by Continuous Wavelet Transform(CWT)from the original vibration signal are fed into a dual-channel MSCNN.This process fully integrates time-domain and time-frequency domain features,enhancing the extraction capability for weak fault information.At last,an LSTM network is utilized to learn temporal features and perform fault classification.Experimental results demonstrate that the proposed method achieves 100%accuracy in early fault diagnosis on the CWRU bearing dataset,significantly outperforming traditional approaches,thus verifying its effectiveness and superiority.关键词
滚动轴承/早期故障诊断/多尺度卷积神经网络/长短期记忆网络/特征融合/三角拓扑聚合优化算法Key words
rolling bearing/early fault diagnosis/Multi-Scale Convolutional Neural Network/Long Short-Term Memory/feature fusion/Triangular Topology Aggregation Optimization分类
机械制造引用本文复制引用
邓志超,张清华..基于TTAO-VMD与CWT融合的轴承早期故障诊断方法[J].广东石油化工学院学报,2025,35(6):68-73,6.基金项目
国家自然科学基金重点项目(61933013) （61933013）
广东省自然科学基金面上项目(2022A1515010599) （2022A1515010599）
广东石油化工学院博士启动项目(2020bs006) （2020bs006）
