文献类型：期刊文献

英文题名：A novel statistical filtering framework for extracting fault-induced characteristics from bearing noisy vibration signal

作者：Liao, Zhiqiang[1];Cai, Renchao[1];Jia, Baozhu[1];Chen, Peng[2];Song, Xuewei[1]

机构：[1]Guangdong Ocean Univ, Naval Architecture & Shipping Coll, Zhanjiang 524088, Peoples R China;[2]Mie Univ, Grad Sch Environm Sci & Technol, Tsu 5148507, Japan

年份：2026

卷号：184

外文期刊名：DIGITAL SIGNAL PROCESSING

收录：SCI-EXPANDED(收录号：WOS:001835151900001)、、EI(收录号：20263021167714)、Scopus(收录号：2-s2.0-105045464792)、WOS

基金：This research was supported by the National Natural Science Foun-dation of China (Grant Nos. 52401418 52201355) and the Guangdong Province Overseas Famous Teacher Project under MS202500036. The authors would acknowledge many colleagues who provided construc-tive comments on improving the research.

语种：英文

外文关键词：Statistical filtering; Feature enhancement; Noise suppression; Bearing fault diagnosis

外文摘要：Bearing fault vibration signal features are often submerged by strong background noise, resulting in inaccurate feature extraction and poor fault diagnosis effectiveness. To address this issue, we propose a curvature-sensitive Otsu (CSO) statistical filtering (CSO-SF) method to enhance the fault signal and bearing fault diagnosis. The method first constructs a multi-dimensional frequency feature matrix to represent the characteristics of fault signals and reference signals, then extracts the principal component from the matrix to evaluate the difference between fault signals and reference signals. The difference as an index used for statistical filtering can significantly highlight the fault frequency. The designed adaptive thresholding criterion serving as an index can precisely match the optimal geometric separation point to choose the fault signal frequency band. The CSO-SF filtered signal can effectively enhance the fault signal feature, and its envelope spectrum can detect the bearing fault characteristic frequency. The feasibility and the superiority of the presented method are verified by the simulation data, public dataset, experimental platform data, and comparison experiments. The results show that the proposed method can effectively enhance the fault signal and diagnose bearing faults.