噪声与振动控制2025，Vol.45Issue(4)：130-135,230,7.DOI:10.3969/j.issn.1006-1355.2025.04.021
基于多特征提取与蜣螂算法优化的轴承故障诊断
Bearing Fault Diagnosis Based on Multi-feature Extraction and Dung Beetle Algorithm Optimization
摘要
Abstract
To address the issue of bearing vibration signals being susceptible of noise interference and the low accuracy of individual feature quantities,a method based on Wavelet Packet Decomposition(WPD),effective time-domain features(absolute mean value,waveform index),frequency-domain features(root-mean-square frequency),and Multi-scale Symbolic Dynamic Entropy(MSDE)is proposed for bearing fault diagnosis.Firstly,the wavelet packet decomposition for the bearing vibration signal is performed to extract multi-frequency band features,and the optimal components are selected based on cor-relation coefficients for signal reconstruction.Secondly,sensitive features are extracted from both time and frequency do-mains,and the MSDE values of the reconstructed signals are calculated to form multi-feature vectors.Lastly,these vectors are fed into a Support Vector Machine(SVM)and optimized by the Dung Beetle Optimizer(DBO)to recognize different types of bearing faults.Results demonstrate that this method can extract fault features from multiple perspectives with higher accuracy and faster recognition than single-feature approaches.关键词
故障诊断/多特征/多尺度符号动力学熵/蜣螂算法Key words
fault diagnosis/multi-feature/multi-scale symbolic dynamic entropy/dung beetle algorithm分类
机械制造引用本文复制引用
谢锋云,樊秋阳,孙恩广,王阳,宋成杰,朱海燕..基于多特征提取与蜣螂算法优化的轴承故障诊断[J].噪声与振动控制,2025,45(4):130-135,230,7.基金项目
国家自然科学基金资助项目(52265068,52162045) （52265068,52162045）
江西省自然科学基金资助项目(20224BAB204050,20232ACB204022) （20224BAB204050,20232ACB204022）
载运工具与装备教育部重点实验室资助项目(KLCEZ2022-02) （KLCEZ2022-02）
