Nonlinear coupling calculation method based on Kolmogorov-Arnold attention dynamic spatiotemporal graph for remaining useful life prediction
- 9 January 2026
- journal article
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (4)
Abstract
Remaining useful life (RUL) prediction is a key aspect of prognostics and health management. It is essential for ensuring the reliability and operational safety of mechanical equipment. Graph Attention Networks (GATs) are an emerging approach for RUL prediction through the innovative integration of graph structures and attention mechanisms. However, existing graph structures fail to establish the correlations between different sensors at different timestamps. Furthermore, the attention mechanism of traditional GATs exhibits insufficient capability in capturing complex nonlinear dependencies between nodes. To overcome these limitations, this paper proposes a nonlinear coupling spatiotemporal graph attention network (NCST-GAT) for RUL prediction. First, a spatiotemporal graph structure construction method is designed to integrate sensor nodes across timestamps. Then, a multiplicative Kolmogorov-Arnold attention mechanism is proposed to replace the standard multilayer perceptron in attention scoring functions, significantly enhancing the modeling of nonlinear interactions among nodes and yielding more expressive feature representations to predict RUL. Experimental results on the CMAPSS and PHM2010 datasets demonstrate that the proposed NCST-GAT achieves an average improvement of 6.1% and 15.6% for RMSE and prognostic Score metrics respectively over state-of-the-art methods, exhibiting strong predictive performance and generalization capability.Keywords
Funding Information
- Science and Technology Project of Hebei Education Department (QN2024107)
- National Key Research and Development Program of China (2022YFB3303803)
- National Key Research and Development Program of China (2024YFB3310900)
- National Natural Science Foundation of China (52405100)
- National Natural Science Foundation of China (62073118)
This publication has 41 references indexed in Scilit:
- GLC: A dual-perspective approach for identifying influential nodes in complex networksExpert Systems with Applications, 2024
- KAN-ODEs: Kolmogorov–Arnold network ordinary differential equations for learning dynamical systems and hidden physicsComputer Methods in Applied Mechanics and Engineering, 2024
- Research on gas turbine health assessment method based on physical prior knowledge and spatial-temporal graph neural networkApplied Energy, 2024
- An attention-based temporal convolutional network method for predicting remaining useful life of aero-engineEngineering Applications of Artificial Intelligence, 2023
- Remaining useful life prediction of bearings with attention-awared graph convolutional networkAdvanced Engineering Informatics, 2023
- Machine learning-based fatigue life prediction of metal materials: Perspectives of physics-informed and data-driven hybrid methodsEngineering Fracture Mechanics, 2023
- A dual‐stage attention‐based Conv‐LSTM network for spatio‐temporal correlation and multivariate time series predictionInternational Journal of Intelligent Systems, 2021
- Spatio-temporal graph convolutional neural network for remaining useful life estimation of aircraft enginesAerospace Systems, 2020
- Remaining useful life prediction using multi-scale deep convolutional neural networkApplied Soft Computing, 2020
- On a Constructive Proof of Kolmogorov’s Superposition TheoremConstructive Approximation, 2009