Data-Driven Anomaly Detection in Modern Power Systems

With the development of information and communication techniques, big data in smart grid makes it possible to provide new solutions to efficient power system operations. However, these technologies are vulnerable to malicious cyber attacks targeting on electricity markets or physical systems, which may lead to severe reliability and economic issues. In this chapter, two state-of-the-art data-driven anomaly detection methods are introduced. To demonstrate their applications in power systems, two groups of forecasting models, i.e., deterministic and probabilistic models, are developed to provide very-short-term electricity price forecasting. Therefore, the anomaly behavior in the electricity market could be detected in both deterministic and probabilistic manner. Case studies showed that the developed deterministic and probabilistic anomaly methodologies outperformed benchmarks.