Cyber-security analysis in smart grid SCADA systems: A game theoretic approach

Power grid systems deliver electricity from suppliers that generate the power to be transmitted through grids of interconnected networks and distributed to customers. Smart grids refer to power grid systems that rely on information and communication technology to perform their functions. Smart grid Supervisory Control and Data Acquisition (SCADA) systems monitor, control and manage most of the critical processes of the smart grid. Such processes may include complex interactions among various hardware and software components. Damage of any form to the SCADA system may have a deleterious effect on the functionality of the smart grid leading to economic losses, electrical blackouts or even human fatalities in worst cases. Recent documented incidents, aimed at disrupting SCADA systems have mostly been from the cyber world (cyber-attacks). The crucial nature and instrumentality of SCADA systems make them an appealing target to cyber-attacks. Substantial research has been done in applying game theory to address security issues of critical infrastructures including SCADA controls in smart grid systems. However, most existing approaches lack details in modeling the attack actions and rely on payoff values that tend to be excessively subjective making it hard to repeat the analysis in a systematic manner. This thesis presents an analytical game theoretic approach to analyzing security of SCADA smart grids. The proposed game model is a two-player, non-zero sum, sequential game between the SCADA attacker and SCADA administrator. The payoffs for both the attacker’s and the defender’s actions are formulated into utility functions to minimize the reliance on subjective data and mimic actual attack behaviors. The methodology used for equilibrium analysis in this thesis is backward maximin induction whose solution is compared with results obtained by Nash equilibrium to evaluate the method for analysis of the proposed approach. Further evaluation of the approach is provided by an intuitive reasoning on two considerations of a real-world scenario.