Data Integrity attack detection in energy control systems- Normal, Fault and contingency dataset

The IEEE 24 bus system is modelled in Digsilent’s Power factory, and normal operating data points are generated under quasi dynamic simulation and N-2 contingencies. The data is then cleaned, and a Gaussian noise component is added. An optimization problem to generate false data injection attack vectors are then modelled using MATLAB. This data contains normal, contingency and attack data points for the necessary variables in a power network. The attack data has 3 parts, with each part generated by a different level of attack. A machine learning algorithm can be trained to detect and classify false data injection attacks in a power network using this data.