Research On False Data Injection Attack Of Smart Grid Based On Bi-Level Programming

As an important energy industry,electricity is the lifeblood of national economic development.With the rapid development of economy,the power demand of our country continues to grow strongly,and the construction of electric power industry has entered the period of overall smart grid construction.Smart grid is a new type of power grid based on physical grid and highly integrated with modern communication,computer,automation and other information technologies.The security of smart grid is related to the healthy and stable development of national economy,and its safe and reliable operation must be ensured.But in recent years,with the development of power grid technology and information technology,smart grid is very vulnerable to false data injection attacks,so it is of great theoretical significance and application value to study the pattern and principle of false data injection attacks.This dissertation first reviews the background of false data injection attack and related research results,and introduces the attack principle and protection mechanism of false data injection attack,then,taking the false data injection attacks of smart grid as the research object,combining the Stackelberg game theory,the bi-level programming theory and the random programming theory,the results of numerical simulation are used to investigate the influence of the degree of attack on the smart grid under the condition of attacker’s tendency or uncertainty.Considering that the direct solution of multi-objective bi-level programming model is very complex,this dissertation first transforms the bi-level programming model into a single-level multiobjective optimization model by using the Karush-Kuhn-Tucker condition of the lower level problem,then,the linear weighted approximation method is used to transform the problem into a single objective nonlinear programming problem,which is solved by software programming,and the following conclusions are obtained:(1)It is almost impossible to maximize the operating costs when the number of measurements being attacked is small,and more measurements must be attacked when the attacker is only seeking the maximum operating costs,the operating costs of the power system have also increased.Therefore,power system operators can reduce the number of measurements that may be successfully attacked by effectively protecting some measurement devices,which can significantly reduce the loss caused by the attack.(2)With less protection from grid operators,the number of attacks from power grid attackers against load and line measurements will gradually increase;the experimental results also show that with the same degree of attack,the attack quantity injected by the line measurements is obviously less than that injected by the load measurements.Therefore,power system operators can focus on the effective protection of load measurements to reduce the number of measurements that may be successfully attacked,and significantly reduce the economic losses of the power system.In this dissertation,a multi-objective bi-level programming model considering the attacker’s attack tendency and a stochastic bi-level programming model considering the attacker’s attack degree uncertainty are established respectively,to make the multiobjective bi-level programming model more realistic,and to explore the feasible defense strategy of power grid operators under the consideration of the possible influence of attackers,it is of great significance to the establishment of a safe and economic new type of smart grid to ensure the steady development of national economy.