The usage of AI-Based Anomaly Detection in SCADA Systems

Supervisory Control and Data Acquisition (SCADA) Systems are critical for industrial infrastructures, especially to ensure the ability to seamlessly monitor and control assets remotely. However, the significant threats of cybersecurity breaches and operational faults stand in the way of complete reliance on SCADA systems. This review paper investigates the recent advancements in artificial intelligence (AI) and its ability to serve as an effective tool to detect anomalies which traditional rule-based systems often miss. AI-Driven anomaly detection is examined from two perspectives: operational anomalies-including process deviations, abnormal readings and equipment malfunctions-and cybersecurity anomalies such as intrusions and network attacks. Various modern AI methodologies are evaluated, alongside potential challenges like real-time processing demands and explainability.