Interval state estimation based defense mechanism against cyber attack on power systems

State estimation is critical for the normal operation of power system. In conventional state estimator, bad data detector (BDD) is used to prevent power systems from being misled by erroneous measurements. However, many studies have pointed out that hackers can exploit security vulnerabilities of BDD to launch cyber-attack on power systems. In order to detect cyber-attack effectively, a novel defense mechanism based on interval state estimation is proposed. In the defense mechanism, the lower- and upper bounds of state variables are modeled as a dual bi-level optimization problem, whose objective is to maximize and minimize the boundary values of the state variables. In addition, the uncertainty of load forecasting is modeled as a parametric Gaussian distribution. Finally, the effectiveness of the proposed method is verified on standard IEEE 30-bus system.