机械制造与自动化2025，Vol.54Issue(2)：37-41,5.DOI:10.19344/j.cnki.issn1671-5276.2025.02.007
基于累积模糊熵与Bi-LSTM的滚动轴承退化趋势预测
Prediction of Rolling Bearing Degradation Trend Based on Cumulative Fuzzy Entropy and Bi-LSTM
摘要
Abstract
In order to compensate for the deficiency that the fuzzy entropy feature has difficulty in reflecting the monotony deterioration trend of rolling bearing in the process of degradation,and predict the bearing degradation trend more accurately,a prediction method combining the cumulative fuzzy entropy feature index and Bi-LSTM is proposed.The feature sequence of the vibration signal is extracted according to the fuzzy entropy algorithm,and the cumulative fuzzy entropy feature index is obtained through the cumulative transformation.The degradation trend of the bearing is predicted by the Bi-LSTM based on features of historical state degradation.The experimental results show that the cumulative fuzzy entropy characteristic index can not only reflect the early degradation of the bearing,but also has good monotonicity.And for the same cumulative fuzzy entropy feature index,the Bi-LSTM prediction model,compared with LSTM and GRU prediction models,has higher prediction accuracy.关键词
滚动轴承/退化趋势预测/累积模糊熵/双向长短期记忆神经网络Key words
rolling bearing/degradation trend prediction/cumulative fuzzy entropy/bidirectional long-short term memory neural network分类
机械制造引用本文复制引用
唐斌,池茂儒,赵明花..基于累积模糊熵与Bi-LSTM的滚动轴承退化趋势预测[J].机械制造与自动化,2025,54(2):37-41,5.基金项目
国家自然科学基金区域创新发展联合基金项目(U21A20168) （U21A20168）
