Deep Learning-Based Anomaly Detection for Time-Series Data in Industrial Control Systems

Industrial control systems(ICS) are widely used in key areas such as industry, energy, transportation, and water conservancy that can control the operation of production equipment. With the widespread use of sensors and actuators in ICS, these systems generate large amounts of multivariate time series data. Currently, there remain significant challenges in anomaly detection in ICS due to the lack of anomaly labels, high data volatility, and the demands of ultra-low inference times. The paper proposes a deep learning-based multivariate anomaly detection method that combines Transform and GAN to detect anomalous behaviours in ICS. The method was tested on the SWaT and WADI datasets collected from industrial control systems, and experimental results show that the model proposed in this paper can outperform state-of-the-art baseline methods in anomaly diagnosis performance and timeliness.