Mid‐term vulnerability analysis of power systems under intentional attacks

Abstract

This study addresses the vulnerability analysis of power systems over a time horizon. The authors introduce a model which could be used by system operators to assess where and when, over a specific time horizon, their power systems are most vulnerable to intentional attacks. This new time-phased vulnerability analysis is modelled as a bilevel programming problem in which in the upper level, an attacker determines the best attack plan, including the best locations and the best times over the time horizon, for launching attacks. In the lower level, the system operator minimises the system operation and load shedding costs. Using duality theory, the bilevel optimisation problem is converted into a mathematical programming with equilibrium constraints which is subsequently converted to a single-level mixed integer linear programming problem by means of linearisation techniques. This model is tested on modified Garver 6-bus test system, modified IEEE 24-bus reliability test system, and IEEE 300-bus test system. The results appreciate the capability of the proposed model and show that it is indispensable to consider the time in analysing the vulnerability of power systems.