1 School of Mechatronics Engineering, Anhui University of Science & Technology, Huainan 232001, China
2 Joint Research Center for Occupational Medicine and Health of IHM, Anhui University of Science & Technology, Huainan 232001, China
3 School of Computer Science and Engineering, Anhui University of Science & Technology, Huainan 232001, China
Show less
文章历史+
Received
Published
Revised
2025-06-09
2026-05-18
2025-11-19
Online Date
2026-08-06
PDF (14027K)
摘要
为了解决跨模态故障数据的近邻结构失真问题,提出了一种基于跨模态近邻互补空间学习(cross-modal neighbor complementary space learning,CNCSL)的轴承故障诊断方法,实现了不同模态间的近邻结构互补,有效改善了轴承故障诊断的准确性。针对不同模态间的数据维数不一致导致的互补困境,沿着空间学习的实现方式反向思考,设计了原始数据空间与跨模态近邻互补空间的空间转化策略,构建所有模态的跨模态近邻互补簇心,从而得到原始数据更真实的近邻结构且不受跨模态数据维数不一致的影响。然后在该簇心的基础上构建跨模态近邻互补散布,从而将更真实的近邻结构嵌入跨模态近邻互补空间,增强空间的近邻聚集性,最后通过最小化跨模态近邻互补散布同时最大化所有模态的成对相关性,构建CNCSL模型的目标优化函数,并在理论上推导出该模型的解析解,进而利用空间投影直接获得跨模态近邻互补特征,该特征不仅能够尽可能保持更真实的近邻结构,而且拥有良好的鉴别力。在Paderborn轴承数据集和实验室实验平台上进行的实验结果表明了CNCSL方法的有效性。
Abstract
To address the issue of neighbor-structure distortion in cross-modal fault data, a bearing fault diagnosis method based on CNCSL (cross-modal neighbor complementary space learning) was proposed. By this method, neighbor-structure complementarity among different modalities was achieved, and the accuracy of bearing fault diagnosis was effectively improved. Aiming at the complementary dilemma caused by inconsistent dimensions of data among different modalities, the implementation methods of space learning were considered in a reverse manner. A space-transformation strategy between the original data space and the cross-modal neighbor complementary space was designed, and cross-modal neighbor complementary cluster centers for all modalities were constructed. Thus a more authentic neighbor structure of the original data was obtained in the cluster centers, and the cluster centers was not affected by the inconsistent dimensions of the cross-modal data. Then, on the basis of the cluster centers, cross-modal neighbor complementary scatters were constructed to embed a more authentic neighbor structure into the cross-modal neighbor complementary space, which can enhance the neighbor-aggregation property of the space. Finally, by minimizing the cross-modal neighbor complementary scatters while maximizing the pairwise correlations of all modalities, the objective optimization function of the CNCSL model was constructed. The analytical solutions of this model were theoretically derived, and then cross-modal neighbor complementary features were directly obtained through space projection. These features not only preserve a more authentic neighbor structure as much as possible, but also exhibit good discriminability. The effectiveness of the CNCSL method is demonstrated by experimental results on the Paderborn bearing dataset and laboratory experimental platforms.
针对这些问题,现提出一种基于跨模态近邻互补空间学习(cross-modal neighbor complementary space learning,CNCSL)的轴承故障诊断方法。为了解决跨模态故障数据的近邻结构失真问题,现设计一种原始数据空间与跨模态近邻互补空间的空间转化策略,同时在跨模态近邻互补空间中构建近邻互补簇心,接着借助理论推导获得原始数据空间中所有模态的近邻互补簇心的显式描述。然后以近邻互补簇心为基础,进一步构建跨模态近邻互补散布,该散布明确考虑不同模态间的近邻互补性。最后通过最小化跨模态近邻互补散布同时最大化所有模态的成对相关性,构建CNCSL模型的目标函数并对其进行优化求解;为跨模态轴承故障诊断中近邻结构失真问题提供一种新颖且有效的解决方案,为复杂工业场景下多模态故障数据的特征学习与诊断优化提供新的思路和理论支撑。
1 基于跨模态近邻互补空间学习的轴承故障诊断方法
为了解决跨模态故障数据的近邻结构失真问题,提出了一种基于跨模态近邻互补空间学习(cross-modal neighbor complementary space learning,CNCSL)的轴承故障诊断方法。针对不同模态间的数据维数不一致导致的互补困境,沿着空间学习的实现方式反向思考,设计了原始数据空间与跨模态近邻互补空间的空间转化策略,构建所有模态的跨模态近邻互补簇心,然后在该簇心的基础上构建跨模态近邻互补散布,最后通过最小化跨模态近邻互补散布同时最大化所有模态的成对相关性,构建CNCSL模型的目标优化函数,并在理论上推导出该模型的解析解。
WangXuechun, LiXiang, YangSuixian. Bearing fault diagnosis based on MCNN-MSA-BiGRU[J]. Science Technology and Engineering, 2025, 25(11): 4534-4542.
[3]
FuZ, LiuZ. TRA-ACGAN: a motor bearing fault diagnosis model based on an auxiliary classifier generative adversarial network and transformer network[J]. ISA Transactions, 2024, 149: 381-393.
HongYang, WuQinmu. Motor bearing fault diagnosis based on adaptive local collaboration federated learning[J]. Science Technology and Engineering, 2025, 25(8): 3217-3225.
[6]
WangZ, GaoY, YuJ, et al. Data-driven fault diagnosis of PEMFC water management with segmented cell and deep learning technologies[J]. International Journal of Hydrogen Energy, 2024, 67: 715-727.
[7]
WangY, ShenJ, YangS, et al. Knowledge and data dual-driven fault diagnosis in industrial scenarios: a survey[J]. IEEE Internet of Things Journal, 2024, 11(11): 19256-19277.
[8]
ElshenawyL M, GafarA A, AwadH A, et al. Fault detection of wind turbine system based on data-driven methods: a comparative study[J]. Neural Computing and Applications, 2024, 36(17): 10279-10296.
[9]
LuoZ D, LiH X. Dynamic partial-least-squares-based fault detection for nonlinear distributed parameter systems[J]. IEEE Transactions on Instrumentation and Measurement, 2024, 73: 1-9.
[10]
LuoL, WangW, BaoS, et al. Robust and sparse canonical correlation analysis for fault detection and diagnosis using training data with outliers[J]. Expert Systems with Applications, 2024, 236. DOI: 10.1016/j.eswa.2023.121434.
[11]
ZhangY, DingJ, LiY, et al. Multi-modal data cross-domain fusion network for gearbox fault diagnosis under variable operating conditions[J]. Engineering Applications of Artificial Intelligence, 2024, 133. DOI: 10.1016/j.engappai.2024.108236.
[12]
ChenZ, DingS X, ZhangK, et al. Canonical correlation analysis-based fault detection methods with application to alumina evaporation process[J]. Control Engineering Practice, 2016, 46: 51-58.
MaoXiangde, DongHaiying, LiangJin. Fault diagnosis of traction rectifier based on t-distribution stochastic neighbor embedding fused with optimal multi-frequency band box dimension[J]. Transactions of China Electrotechnical Society, 2025, 40(2): 531-543.
[15]
WangS, JuY, FuC, et al. A reversible residual network-aided canonical correlation analysis to fault detection and diagnosis in electrical drive systems[J]. IEEE Transactions on Instrumentation and Measurement, 2024, 73: 1-10.
[16]
ZhangK, PengK, DingS X, et al. A correlation-based distributed fault detection method and its application to a hot tandem rolling mill process[J]. IEEE Transactions on Industrial Electronics, 2019, 67(3): 2380-2390.
[17]
WangW, WangY, et al. Distributed monitoring of nonlinear plant-wide processes based on GA-regularized kernel canonical correlation analysis[J]. Reliability Engineering & System Safety, 2024, 252. DOI: 10.1016/j.ress.2024.110421.
[18]
WangY, ChangD, FuZ, et al. Partially view-aligned representation learning via cross-view graph contrastive network[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2024, 34(8): 7272-7283.
[19]
WangY, BaoD, LiS. Dynamic graph embedding PCA to extract spatio-temporal information for fault detection[J]. IEEE Transactions on Industrial Informatics, 2024, 21(2): 1714-1723.
[20]
PengC, ShengY, GuiW, et al. A rolling bearing fault diagnosis method based on multimodal knowledge graph[J]. IEEE Transactions on Industrial Informatics, 2024, 20(11): 13047-13057.
[21]
ZengY, ZhangR, ZhangH, et al. GCCNet: a novel network leveraging gated cross-correlation for multi-view classification[J]. IEEE Transactions on Multimedia, 2024, 27: 1086-1099.
[22]
WuF, ChenK, QiuG, et al. Robust open circuit fault diagnosis method for converter using automatic feature extraction and random forests considering nonstationary influence[J]. IEEE Transactions on Industrial Electronics, 2024, 71(10): 13263-13273.
[23]
FuS, PengQ, HeY, et al. Unsupervised multiplex graph diffusion networks with multi-level canonical correlation analysis for multiplex graph representation learning[J]. Science China Information Sciences, 2025, 68(3). DOI: 10.1007/s11432-025-4492-y.
[24]
LiS, FengK, XuY, et al. Cross-modal zero-sample diagnosis framework utilizing non-contact sensing data fusion[J]. Information Fusion, 2024, 110. DOI: 10.1016/j.inffus.2024.102453.
[25]
PuH, ZhangK, AnY. Restricted sparse networks for rolling bearing fault diagnosis[J]. IEEE Transactions on Industrial Informatics, 2023, 19(11): 11139-11149.
LiangHongtao, KongLingchao, LiuGuozhu, et al. ASL-CatBoost method for wind turbine fault detection integrated with digital twin[J]. Journal of System Simulation, 2024, 36(4): 873-887.