Data-Integrity Aware Stochastic Model for Cascading Failures in Power Grids

The reliable operation of power grids during cascading failures is heavily dependent on the interdependencies between the power grid components and the supporting communications and control networks. Moreover, the system operators' expertise in dealing with cascading failures can play a pivotal role during contingencies. In this paper, a dynamical probabilistic model is developed based on Markov-chains, which captures the dynamics of cascading failures in the power grid. Specifically, a previously developed Markov-chain based model is extended to capture the trade-off between the benefits of having a robust communication infrastructure and its vulnerability from data integrity (e.g., cyber-attacks). State-space reduction of the complex interactions between power grids, communication networks and system operators is achieved by judiciously specifying the state variables of the Markov chain. The impact of system operators' probability of error during a cascade-mitigation action is incorporated into the model as a function of the state variables of the Markov chain. A point of diminishing returns is observed beyond which the effect of information infidelity outweighs the benefits of having more information. For a given level of cyber threat, an optimal size of a communication network is observed that minimizes the expected number of transmission-line failures before the cascade stops.