Robust factor information fusion for industrial safety through TSFD-Based phase space reconstruction and noise reduction

•The research elucidates the fusion mechanism, proposes an SFEP analysis framework.•Explore the characteristics of SFEP and TSFBD, the research is conducted based on TSFBD.•Employ optimal phase space and KOLPP to address dimensionality and noise reduction. Industrial system safety is a core goal of...

Full description

Saved in:

Bibliographic Details
Published in:	Information fusion Vol. 128; p. 103917
Main Authors:	Li, Shasha, Cui, Tiejun
Format:	Journal Article
Language:	English
Published:	Elsevier B.V 01.04.2026
Subjects:	Data noise reduction Factor classification Industrial system safety Robust factor information fusion System fault evolution Time series fault data (TSFD) Factor classification System fault evolution Data noise reduction Time series fault data (TSFD) Industrial system safety Robust factor information fusion
ISSN:	1566-2535
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	•The research elucidates the fusion mechanism, proposes an SFEP analysis framework.•Explore the characteristics of SFEP and TSFBD, the research is conducted based on TSFBD.•Employ optimal phase space and KOLPP to address dimensionality and noise reduction. Industrial system safety is a core goal of industrial production, and robust factor information fusion is vital for data-driven industrial system fault prediction, early warning, and analysis. To explore system fault evolution characteristics, this paper constructs time series fault data (TSFD) by integrating multi- source industrial fault-impacting factor measurements, and proposes methods for optimal phase space construction, dimension reduction, noise reduction, and factor classification. It first analyzes system fault evolution process and TSFD characteristics, then presents the algorithm flow and validates via case study. For TSFD processing, optimal time delay and embedding dimension are determined to build the optimal phase space matrix; the KOLPP method retains local and global features to obtain key matrices, realizing TSFD optimization. Based on the robust factor information fusion method, the factor classification was obtained, including core factors, supplementary factors and irrelevant factors. Finally, algorithm challenges are discussed, and its characteristics summarized. This study enhances industrial fault management to ensure production reliability.
ISSN:	1566-2535
DOI:	10.1016/j.inffus.2025.103917