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OA

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.