机械制造与自动化2026，Vol.55Issue(3)：100-106,7.DOI:10.19344/j.cnki.issn1671-5276.2026.03.020
多维信息分析增强张量奇异谱分解及多源信息融合诊断应用
Multidimensional Information Analysis on Enhanced Tensor Singular Spectrum Decomposition and Its Application for Multi-Source Information Fusion Diagnosis
摘要
Abstract
A tensor is a generalization of vectors and matrices to arbitrary dimensions,capable of efficiently presenting and manipulating multidimensional data.It is widely used in fields such as image processing,signal processing,and high-dimensional data mining.Based on the theory of Flexible Tensor Singular Value Decomposition(FTSVD),this paper proposes a novel tensor singular spectrum decomposition algorithm by integrating generalized multidimensional information analysis of power spectrum.Apart from overcoming issues in traditional tensor decomposition methods,such as low sparsity of core tensors and non-uniqueness of tensor ranks,the proposed algorithm achieves adaptive iterative selection of the optimal tensor-embedded dimension,making it widely applicable to the fusion processing of multi-source high-order data.Comparative analysis with the analysis result of processing multi-sensor fault signals of rolling bearings by existing methods demonstrates that the proposed method has significant advantages in signal denoising and feature extraction performance.关键词
张量分解/奇异值分解/故障诊断/信号处理/滚动轴承Key words
tensor decomposition/singular value decomposition/fault Diagnosis/signal processing/rolling bearing分类
信息技术与安全科学引用本文复制引用
邢得富,卢佳庆,杨山,胡文超,宫元,尧伟和,张飞斌..多维信息分析增强张量奇异谱分解及多源信息融合诊断应用[J].机械制造与自动化,2026,55(3):100-106,7.基金项目
国家自然科学基金项目(52105109) （52105109）
