Multi-Level Graph Attention Network-Based Anomaly Detection in Industrial Control System

Industrial control systems (ICSs) are vital to critical infrastructure in energy, manufacturing, and other industries. As ICSs become increasingly interconnected, their complexity grows, making them more vulnerable to cyber attacks and system failures. This growing complexity underscores the critica...

Full description

Saved in:
Bibliographic Details
Main Authors: Longxin Lin, Anyang Gu, Feiyan Min, Shan Zhou
Format: Article
Language:English
Published: MDPI AG 2025-04-01
Series:Actuators
Subjects:
Online Access:https://www.mdpi.com/2076-0825/14/5/210
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850127746156462080
author Longxin Lin
Anyang Gu
Feiyan Min
Shan Zhou
author_facet Longxin Lin
Anyang Gu
Feiyan Min
Shan Zhou
author_sort Longxin Lin
collection DOAJ
description Industrial control systems (ICSs) are vital to critical infrastructure in energy, manufacturing, and other industries. As ICSs become increasingly interconnected, their complexity grows, making them more vulnerable to cyber attacks and system failures. This growing complexity underscores the critical need for advanced anomaly detection techniques to ensure the safe and reliable operation of ICSs. To address this need, we propose a novel method, the physical process and controller graph attention network (PCGAT), which constructs multi-level graphs based on physical process and controller information. Experiments on two real-world ICS datasets demonstrate that PCGAT achieves superior performance and enables the localization of anomalies within specific physical processes. Moreover, by leveraging graph attention networks (GATs), PCGAT enhances interpretability in anomaly detection.
format Article
id doaj-art-51d9e1452c8e4c76bcfa163786eb159e
institution OA Journals
issn 2076-0825
language English
publishDate 2025-04-01
publisher MDPI AG
record_format Article
series Actuators
spelling doaj-art-51d9e1452c8e4c76bcfa163786eb159e2025-08-20T02:33:35ZengMDPI AGActuators2076-08252025-04-0114521010.3390/act14050210Multi-Level Graph Attention Network-Based Anomaly Detection in Industrial Control SystemLongxin Lin0Anyang Gu1Feiyan Min2Shan Zhou3College of Information Science and Technology, Jinan University, Guangzhou 510632, ChinaSchool of Intelligent Systems Science and Engineering, Jinan University, Zhuhai 519070, ChinaCollege of Information Science and Technology, Jinan University, Guangzhou 510632, ChinaCollege of Information Science and Technology, Jinan University, Guangzhou 510632, ChinaIndustrial control systems (ICSs) are vital to critical infrastructure in energy, manufacturing, and other industries. As ICSs become increasingly interconnected, their complexity grows, making them more vulnerable to cyber attacks and system failures. This growing complexity underscores the critical need for advanced anomaly detection techniques to ensure the safe and reliable operation of ICSs. To address this need, we propose a novel method, the physical process and controller graph attention network (PCGAT), which constructs multi-level graphs based on physical process and controller information. Experiments on two real-world ICS datasets demonstrate that PCGAT achieves superior performance and enables the localization of anomalies within specific physical processes. Moreover, by leveraging graph attention networks (GATs), PCGAT enhances interpretability in anomaly detection.https://www.mdpi.com/2076-0825/14/5/210industrial control systemsanomaly detectiongraph neural networksensors and actuators
spellingShingle Longxin Lin
Anyang Gu
Feiyan Min
Shan Zhou
Multi-Level Graph Attention Network-Based Anomaly Detection in Industrial Control System
Actuators
industrial control systems
anomaly detection
graph neural network
sensors and actuators
title Multi-Level Graph Attention Network-Based Anomaly Detection in Industrial Control System
title_full Multi-Level Graph Attention Network-Based Anomaly Detection in Industrial Control System
title_fullStr Multi-Level Graph Attention Network-Based Anomaly Detection in Industrial Control System
title_full_unstemmed Multi-Level Graph Attention Network-Based Anomaly Detection in Industrial Control System
title_short Multi-Level Graph Attention Network-Based Anomaly Detection in Industrial Control System
title_sort multi level graph attention network based anomaly detection in industrial control system
topic industrial control systems
anomaly detection
graph neural network
sensors and actuators
url https://www.mdpi.com/2076-0825/14/5/210
work_keys_str_mv AT longxinlin multilevelgraphattentionnetworkbasedanomalydetectioninindustrialcontrolsystem
AT anyanggu multilevelgraphattentionnetworkbasedanomalydetectioninindustrialcontrolsystem
AT feiyanmin multilevelgraphattentionnetworkbasedanomalydetectioninindustrialcontrolsystem
AT shanzhou multilevelgraphattentionnetworkbasedanomalydetectioninindustrialcontrolsystem