智能系统学报2025，Vol.20Issue(2)：363-375,13.DOI:10.11992/tis.202401013
改进滑动粗粒化和集成波动色散熵的故障诊断方法
Fault diagnosis using improved sliding coarsening and integrated fluctuation-based dispersion entropy
摘要
Abstract
In multiscale fluctuation-based dispersion entropy,multiscale coarse graining loses the information between adjacent points in the reconstructed subsequences.Additionally,the length decreases as the scale factor increases,and the features extracted through this coarse-grained method are not conducive to fault classification.To address this prob-lem,this paper proposes a method of n steps sliding.This method ensures that the information between points is pre-served under the given scale factor,maintaining the length of the reconstructed sequence to be consistent with the origin-al sequence.Aiming at the problem that the mapping technology in the fluctuation dispersion entropy is too simple,in-tegrated dispersion entropy is used to extract features from the reconstructed sequence,enhancing the accuracy of en-tropy calculations.The algorithm is verified using bearing datasets from Case Western Reserve University and other in-stitutions,the proposed method notably improves fault diagnosis accuracy.关键词
滑动粗粒化/序列重构/故障诊断/故障分类/集成波动色散熵/滚动轴承/振动信号/特征提取Key words
sliding coarsening/sequence reconstruction/fault diagnosis/fault classification/integrated fluctuation-based dispersion entropy/rolling bearing/vibration signal/feature extraction分类
信息技术与安全科学引用本文复制引用
穆凌霞,田璐,冯楠,汪红鑫,张建,吴世海,刘丁..改进滑动粗粒化和集成波动色散熵的故障诊断方法[J].智能系统学报,2025,20(2):363-375,13.基金项目
国家自然科学基金项目(62373299,62127809) （62373299,62127809）
陕西省重点研发计划项目(2024GX-YBXM-093) （2024GX-YBXM-093）
中国博士后科学基金项目(2022MD723834) （2022MD723834）
陕西省科协青年人才托举计划项目(20210114). （20210114）
