Key points are not available for this paper at this time.
Abstract Supervisory control and data acquisition (SCADA) systems are vital in monitoring and controlling industrial processes through the web. However, while such systems result in lower costs, greater utilisation efficiency, and improved reliability, they are vulnerable to cyberattacks, with consequences ranging from the inconvenience and minor disruption to severe physical damage and even loss of life. The authors evaluate the security of the Dam system in the form of Common Criteria, develop safety goals to improve this safety, and focus on threats and risks to the dam SCADA system. Finally proposes an anomaly‐based machine‐learning framework for detecting malicious network attacks in the SCADA system of a dam. Three unsupervised classification algorithms are considered: hierarchical clustering, local outlier factor, and isolation forest. It is shown that the hierarchical clustering algorithm achieves the highest precision and F‐score of the three algorithms. Overall, the results confirm the effectiveness of anomaly‐based detection algorithms in enhancing the robustness of SCADA systems toward malicious attacks. At the same time, it complies with the security objectives of Common Criteria, achieving the safety and protection of the dam.
Lu et al. (Thu,) studied this question.