Stochastic framework for balancing smart grid's performance enhancement and resilience to cyber threat

The reliability of a power grid is heavily dependent on the interdependence between power grid components and its associated communication and control networks. Moreover, the system operators' expertise in dealing with cascading failures can play a pivotal role during contingencies. In this dissertation, a stochastic Markov-chain based model,namely, Interdependent Stochastic Abstract State Space Evolution (I-SASE), is developed, which captures the dynamics of cascading failures in the power grid and the benefits and risk of information through the communication network. A previously developed Markov-chain based model is refined to capture the trade-off between the benefits of having a robust communication infrastructure and its vulnerability to cyber-attacks. A State-space reduction of the complex interactions of the power grid components is achieved by judiciously choosing the state variables of the Markov chain. The impact of systems operators' probability of error during a mitigation action was incorporated into the model as a function of the state variables of the Markov chain. A point of diminishing return is observed beyond which the effect of cyber threat and human errors outweighs the benefits of having more information. An optimal level of interdependence minimizing the expected value of a cascade is achieved between the power grid and the communication network. The benefits of the I-SASE model are threefold. First, the model captures the interdependence and dynamic interactions between the layers of the power grid, which is more realistic compared to the current literature that does not include the effect of cyber threat in an analytical model. Second, the model incorporates the benefits of having inter-connectivity with communication network through effective implementation of load shedding. On the other hand, the model captures the harm of having excessive information through cyber threat and system operator error. This leads to finding an optimal level of inter-connectivity that maximizes the benefits of information for a given level of cyber threat and operator error. Third, the model produces the distribution of the size of a blackout analytically considering the potential harm from cyber threat and operator error, which can be used by the utilities to minimize cascading failures in their design process using statistics such as keeping the expected values of transmission line failures under a threshold. The model is analytic, scalable, and tractable.