Anomaly Detection, Classification and Identification Tool (ADCIT)

The Anomaly Detection, Classification and Identification Tool (ADCIT) is an open source Matlab and Python code used for detection, classification and identification of anomalies in power system state estimation. Outputs of weighted least squares (WLS) and extended Kalman filter (EKF) state estimators, developed in Matlab, are used as inputs for machine learning algorithms developed in Python. The ADCIT can address hard anomaly cases; for example, it can detect and classify the case when load is abruptly changed at multiple nodes simultaneously, or when false data injection attack targets multiple states at the same time. Additionally, the ADCIT does not require retraining of the machine learning algorithm in the presence of network topology changes. Application of the ADCIT within power grid energy management system can help system operator to design proper countermeasures in case of an anomaly occurrence.