文献类型：期刊文献

英文题名：Rotating Machinery Structural Faults Feature Enhancement and Diagnosis Based on Multi-Sensor Information Fusion

作者：Jia, Baozhu[1,2,3];Liang, Guanlong[1,2,3];Huang, Zhende[1,2,3];Song, Xuewei[1,2,3];Liao, Zhiqiang[1,2,3]

机构：[1]Guangdong Ocean Univ, Naval Architecture & Shipping Coll, Zhanjiang 524088, Peoples R China;[2]Tech Res Ctr Ship Intelligence & Safety Engn Guang, Zhanjiang 524088, Peoples R China;[3]Guangdong Prov Key Lab Intelligent Equipment South, Zhanjiang 524088, Peoples R China

年份：2025

卷号：13

期号：7

外文期刊名：MACHINES

收录：SCI-EXPANDED(收录号：WOS:001535529600001)、、Scopus(收录号：2-s2.0-105011618396)、WOS

基金：This research was funded by the National Natural Science Foundation of China (52201355, 52071090), the Program for Scientific Research Start-Up Funds of Guangdong Ocean University (060302132304, 060302132101), the Zhanjiang Non-funded Science and Technology Tesearch Project (2022B01049, 2023B01046), and the Postgraduate Education Innovation Project of Guangdong Ocean University (202545).

语种：英文

外文关键词：rotating machinery structural faults; faults feature enhancement and diagnosis; multi-sensor information fusion; symmetric dot pattern

外文摘要：To address the challenges posed by the difficulty of extracting fault features from rotating machinery with weak fault features, this paper proposes a rotating machinery structural faults feature enhancement and diagnosis method based on multi-sensor information fusion. Firstly, Savitzky-Golay filtering suppresses noise and enhances fault features. Secondly, the designed multi-sensor symmetric dot pattern (SDP) transformation method fuses multi-source information of the rotating machinery structural faults, providing more comprehensive and richer fault feature information for diagnosis. Finally, the ResNet18 model performs fault diagnosis. To validate the feasibility and effectiveness of the proposed method, two datasets verify its performance. The accuracy of the experimental results was 99.16% and 100%, respectively, demonstrating the feasibility and effectiveness of the proposed method. To further validate the superiority of the proposed method, it was compared with different 2D signal transformation methods. The comparison results indicate that the proposed method achieves the best fault diagnosis accuracy compared to other methods.