TY - GEN
T1 - Graph Attention-Based Interpretable Deviation Network for Circulating Water Pump Situational Awareness
AU - Zhang, Le
AU - Cheng, Wei
AU - Zhang, Shuo
AU - Nie, Zelin
N1 - Publisher Copyright:
© 2025 IEEE.
PY - 2025
Y1 - 2025
N2 - The increasing complexity and interconnectivity of modern process industry systems, particularly in critical infrastructure such as circulating water pump systems in nuclear and chemical plants, pose significant challenges for ensuring operational security and system-level situational awareness. Conventional deep learning models, while effective in temporal pattern recognition, exhibit limited capability in capturing spatial dependencies across heterogeneous sensor networks, thereby restricting their ability to represent dynamic system behavior. To address these limitations, this paper presents a novel method titled graph attention-based deviation network (GADN4SA) for circulating water system situational awareness. The proposed method dynamically constructs task-aware sensor graphs based on signal correlations, integrates both structural and semantic representations, and leverages attention-guided mechanisms to model long-range spatial-temporal dependencies. In addition, a structured deviation scoring module is incorporated to enable interpretable anomaly quantification and attribution. Extensive experiments on both simulated and real-world datasets demonstrate that the proposed method consistently outperforms state-of-the-art baselines in terms of robustness, accuracy, and generalization, highlighting its effectiveness for deployment in safety-critical industrial environments.
AB - The increasing complexity and interconnectivity of modern process industry systems, particularly in critical infrastructure such as circulating water pump systems in nuclear and chemical plants, pose significant challenges for ensuring operational security and system-level situational awareness. Conventional deep learning models, while effective in temporal pattern recognition, exhibit limited capability in capturing spatial dependencies across heterogeneous sensor networks, thereby restricting their ability to represent dynamic system behavior. To address these limitations, this paper presents a novel method titled graph attention-based deviation network (GADN4SA) for circulating water system situational awareness. The proposed method dynamically constructs task-aware sensor graphs based on signal correlations, integrates both structural and semantic representations, and leverages attention-guided mechanisms to model long-range spatial-temporal dependencies. In addition, a structured deviation scoring module is incorporated to enable interpretable anomaly quantification and attribution. Extensive experiments on both simulated and real-world datasets demonstrate that the proposed method consistently outperforms state-of-the-art baselines in terms of robustness, accuracy, and generalization, highlighting its effectiveness for deployment in safety-critical industrial environments.
KW - attention mechanism
KW - circulating water pump
KW - graph neural network
KW - situational awareness
UR - https://www.scopus.com/pages/publications/105032664527
U2 - 10.1109/INDIN64977.2025.11279721
DO - 10.1109/INDIN64977.2025.11279721
M3 - 会议稿件
AN - SCOPUS:105032664527
T3 - IEEE International Conference on Industrial Informatics (INDIN)
BT - 2025 IEEE 23rd International Conference on Industrial Informatics, INDIN 2025
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 23rd International Conference on Industrial Informatics, INDIN 2025
Y2 - 12 July 2025 through 15 July 2025
ER -