多约束鲁棒非负矩阵分解的冲击特征频带自适应分解方法

梁霖; 崔旭军; 胡文昊; 代普淼

doi:10.7652/xjtuxb202504016

您当前的位置：

首页 >

文章列表页 >

多约束鲁棒非负矩阵分解的冲击特征频带自适应分解方法

—————————————————————— | 更新时间：2026-01-06

- 多约束鲁棒非负矩阵分解的冲击特征频带自适应分解方法
- Adaptive Decomposition Method for Impact Characteristic Frequency Bands of Multi-Constrained Robust Non-Negative Matrix Factorization
- 西安交通大学学报 2025年59卷第4期页码：171-179
- 作者机构：
  
  1.西安交通大学仪器科学与技术学院， 710049，西安
  2.西安交通大学机械工程学院， 710049，西安
- 作者简介：
  
  梁霖（1973—），男，教授，博士生导师。
- 基金信息：
  
  国家自然科学基金资助项目(52075431)
- DOI：10.7652/xjtuxb202504016
  中图分类号： TH133.33
- 收稿：2024-08-03，
  
  网络首发：2024-12-18，
  
  纸质出版：2025-04-10
- 稿件说明：
移动端阅览
梁霖, 崔旭军, 胡文昊, 等. 多约束鲁棒非负矩阵分解的冲击特征频带自适应分解方法[J]. 西安交通大学学报, 2025,59(4):171-179.

LIANG Lin, CUI Xujun, HU Wenhao, et al. Adaptive Decomposition Method for Impact Characteristic Frequency Bands of Multi-Constrained Robust Non-Negative Matrix Factorization[J]. Journal of Xi’an Jiaotong University, 2025, 59(4): 171-179.
梁霖, 崔旭军, 胡文昊, 等. 多约束鲁棒非负矩阵分解的冲击特征频带自适应分解方法[J]. 西安交通大学学报, 2025,59(4):171-179. DOI： 10.7652/xjtuxb202504016.

LIANG Lin, CUI Xujun, HU Wenhao, et al. Adaptive Decomposition Method for Impact Characteristic Frequency Bands of Multi-Constrained Robust Non-Negative Matrix Factorization[J]. Journal of Xi’an Jiaotong University, 2025, 59(4): 171-179. DOI： 10.7652/xjtuxb202504016.

摘要

针对非负矩阵分解（NMF）在轴承故障的冲击振动频带分解中存在能量导向引起的频带混叠问题，提出了一种面向冲击故障频带自适应分解的多约束鲁棒非负矩阵分解方法。首先，采用

散度自适应加权的误差函数来避免信号未知分布引起的风险，以信号频带数量作为分解秩的选择参考，通过对基矩阵施加正交约束实现自适应频带划分；其次，结合周期冲击响应时频谱的光滑频域和稀疏时域的特性，引入具有良好物理意义的光滑、稀疏约束，构建了面向振动时频谱的多约束鲁棒非负矩阵分解模型；最后，借助正则化技术和Stiefel流形优化方法设计了求解算法。仿真和实验结果表明，与多种NMF方法和典型频带选择方法相比，在面对缺陷引起的微弱冲击时，低频区间中往往存在着多种干扰源影响，所提分解模型能准确提取出高频区间的冲击响应频带，避免了能量导向的传统频带分解方式不足，约束项的引入则有效地提升了NMF的求解结果，增强了NMF方法在冲击特征频带微弱时的辨识能力。

Abstract

In response to the issue of frequency band mixing caused by energy orientation in the decomposition of impact vibration frequency bands related to bearing faults using non-negative matrix factorization (NMF)

a multi-constrained robust non-negative matrix factorization method for adaptive decomposition of impact fault bands is proposed. Firstly

an error function with adaptive weighted

-divergence is utilized to mitigate risks arising from unknown signal distributions

with the number of signal frequency bands serving as a reference for selecting the decomposition rank. Adaptive band division is achieved by imposing orthogonal constraints on the basis matrix. Secondly

considering the smooth frequency domain of the periodic impact response spectrogram and the sparse time domain characteristics

physically meaningful smooth and sparse constraints are introduced to establish a multi-constrained robust non-negative matrix factorization model tailored to vibration time-frequency

spectra. Finally

a solution algorithm is designed leveraging regularization techniques and Stiefel manifold optimization methods. Simulation and experimental results demonstrate that

compared to various NMF methods and typical band selection methods

when facing weak impacts caused by defects

the low-frequency range often harbors multiple sources of interference. The proposed decomposition model accurately extracts the impact response bands in the high-frequency range

circumventing the shortcomings of traditional energy-oriented band decomposition methods. The introduction of constraints effectively enhances the solution results of NMF

reinforcing its discriminative capability when dealing with weak impact characteristic frequency bands.

关键词

Keywords

references

SMITH W A , BORGHESANI P , NI Qing , et al . Optimal demodulation-band selection for envelope-based diagnostics: a comparative study of traditional and novel tools [J ] . Mechanical Systems and Signal Processing , 2019 , 134 : 106303 .

ANTONI J . Fast computation of the kurtogram for the detection of transient faults [J ] . Mechanical Systems and Signal Processing , 2007 , 21 ( 1 ): 108 - 124 .

MOSHREFZADEH A , FASANA A . The autogram: an effective approach for selecting the optimal demodulation band in rolling element bearings diagnosis [J ] . Mechanical Systems and Signal Processing , 2018 , 105 : 294 - 318 .

MAURICIO A , SMITH W A , RANDALL R B , et al . Improved envelope spectrum via feature optimisation-gram (IESFOgram): a novel tool for rolling element bearing diagnostics under non-stationary operating conditions [J ] . Mechanical Systems and Signal Processing , 2020 , 144 : 106891 .

SUN Jing , WANG Zhihui , SUN Fuming , et al . Sparse dual graph-regularized NMF for image co-clustering [J ] . Neurocomputing , 2018 , 316 : 156 - 165 .

LEPLAT V , GILLIS N , ANG A M S . Blind audio source separation with minimum-volume beta-divergence NMF [J ] . IEEE Transactions on Signal Processing , 2020 , 68 : 3400 - 3410 .

WODECKI J , KRUCZEK P , BARTKOWIAK A , et al . Novel method of informative frequency band selection for vibration signal using nonnegative matrix factorization of spectrogram matrix [J ] . Mechanical Systems and Signal Processing , 2019 , 130 : 585 - 596 .

王华庆 , 王梦阳 , 宋浏阳 , 等 . 双约束非负矩阵分解的复合故障信号分离方法 [J ] . 振动工程学报 , 2020 , 33 ( 3 ): 590 - 596 .

WANG Huaqing , WANG Mengyang , SONG Liu yang , et al . Method of compound fault signal separation using double constraints non-negative matrix factorization [J ] . Journal of Vibration Engineering , 2020 , 33 ( 3 ): 590 - 596 .

薛红涛 , 丁殿勇 , 李汭铖 , 等 . 基于分量加权重构和稀疏NMF的轮毂电机轴承复合故障特征提取方法 [J ] . 机械工程学报 , 2023 , 59 ( 9 ): 146 - 156 .

XUE Hongtao , DING Dianyong , LI Ruicheng , et al . Feature extraction method based on component weighted reconstruction and sparse NMF for bearing compound faults of in-wheel motor [J ] . Journal of Mechanical Engineering , 2023 , 59 ( 9 ): 146 - 156 .

LIANG Lin , SHAN Lei , LIU Fei , et al . Impulse feature extraction of bearing faults based on convolutive nonnegative matrix factorization [J ] . IEEE Access , 2020 , 8 : 88617 - 88632 .

LIANG Lin , DING Xingyun , LIU Fei , et al . Feature extraction using sparse kernel non-negative matrix factorization for rolling element bearing diagnosis [J ] . Sensors , 2021 , 21 ( 11 ): 3680 .

YANG Yongsheng , MING Anbo , ZHANG Youyun , et al . Discriminative non-negative matrix factorization (DNMF) and its application to the fault diagnosis of diesel engine [J ] . Mechanical Systems and Signal Processing , 2017 , 95 : 158 - 171 .

GABOR M , ZDUNEK R , ZIMROZ R , et al . Non-negative tensor factorization for vibration-based local damage detection [J ] . Mechanical Systems and Signal Processing , 2023 , 198 : 110430 .

LIANG Lin , WEN Haobin , LIU Fei , et al . Feature extraction of impulse faults for vibration signals based on sparse non-negative tensor factorization [J ] . Applied Sciences , 2019 , 9 ( 18 ): 3642 .

LEE J H . Enhancement of decomposed spectral coherence using sparse nonnegative matrix factorization [J ] . Mechanical Systems and Signal Processing , 2021 , 157 : 107747 .

LIANG Lin , DING Xingyun , WEN Haobin , et al . Impulsive components separation using minimum-determinant KL-divergence NMF of bi-variable map for bearing diagnosis [J ] . Mechanical Systems and Signal Processing , 2022 , 175 : 109129 .

DING C , LI Tao , PENG Wei , et al . Orthogonal nonnegativematrix t-factorizations for clustering [C ] // Proceedings of the 12th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining . New York, NY, USA : ACM , 2006 : 126 - 135 .

CHEN Z , CICHOCKI A . Nonnegative matrix factorization with temporal smoothness and/or spatial decorrelation constraints [J ] . Signal Processing , 2005 , 68 : 1 - 6 .

SALEHANI Y E , GAZOR S . Smooth and sparse regularization for NMF hyperspectral unmixing [J ] . IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2017 , 10 ( 8 ): 3677 - 3692 .

ESSID S , FÉVOTTE C . Smooth nonnegative matrix factorization for unsupervised audiovisual document structuring [J ] . IEEE Transactions on Multimedia , 2013 , 15 ( 2 ): 415 - 425 .

VIRTANEN T . Monaural sound source separation by nonnegative matrix factorization with temporal continuity and sparseness criteria [J ] . IEEE Transactions on Audio Speech and Language Processing , 2007 , 15 ( 3 ): 1066 - 1074 .

FABREGAT R , PUSTELNIK N , GONÇALVES P , et al . Solving NMF with smoothness and sparsity constraints using PALM [EB/OL ] . ( 2021-03-18 ) [ 2024-10-10 ] . https://arxiv.org/abs/1910.14576 https://arxiv.org/abs/1910.14576 .

GILLIS N , HIEN L T K , LEPLAT V , et al . Distributionally robust and MULTI-OBJECTIVE nonnegative matrix factorization [J ] . IEEE Transactions on Pattern Analysis and Machine Intelligence , 2022 , 44 ( 8 ): 4052 - 4064 .

QIAN Yuntao , JIA Sen , ZHOU Jun , et al . Hyperspectral unmixing via L 1/2 sparsity-constrained nonnegative matrix factorization [J ] . IEEE Transactions on Geoscience and Remote Sensing , 2011 , 49 ( 11 ): 4282 - 4297 .

YOO J H , CHOI S J . Nonnegative matrix factorization with orthogonality constraints [J ] . Journal of Computing Science and Engineering , 2010 , 4 ( 2 ): 97 - 109 .

DAGA A P , FASANA A , MARCHESIELLO S , et al . The Politecnico di Torino rolling bearing test rig: description and analysis of open access data [J ] . Mechanical Systems and Signal Processing , 2019 , 120 : 252 - 273 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

面向高海拔空冷发电机防晕性能演化规律的电热耦合研究

隧道内燃料电池汽车氢气泄漏扩散特性及影响因素

结构-语义-残差协同的图像压缩框架：面向隐私保护与高保真重建

冷库水平空气幕热湿阻隔性能数值模拟研究

扇贝仿生密封腔室内超临界二氧化碳泄漏流湍动能耗散机理研究