Fault detection with discrete-time measurements: An application for the cyber security of power networks

Abstract

This article concerns an application of a model-based fault detection and isolation (FDI) method for the cyber security of power systems in a realistic framework, where the system dynamics are expressed in continuous-time, whilst system measurements are applied to an FDI filter in discrete-time samples. Towards the development of a tractable approach for high dimensional nonlinear systems, an existing optimization-based technique for residual generator design is reviewed. However, this requires that both system dynamics and measurements are in the same time scale, i.e., either continuous or discrete. To this end, we investigate different variants of discrete-time modeling approaches for state-space systems, specifically tailored to meet the needs of the existing FDI filter design methodology. Finally,the efficiency and limitations of the presented scheme are illustrated through simulation results for a two-area power system network, in which the objective is the diagnosis of a cyber attack at the Automatic Generation Control signal.