Mitigating the Risk of Cascading Blackouts: A Data Inference Based Maintenance Method

Abstract

The risk of cascading blackouts (RCB) is of great significance in practice because cascading outages can have catastrophic consequences. As there is a positive relationship between the probability of cascading blackouts and that of component failures, an effective way to mitigate the RCB is to perform maintenance. However, this approach is of limited value when considering extremely complicated cascading outages, such as those in particularly large systems. In this paper, we propose a methodology to efficiently identify the most influential component(s) for mitigating the RCB in a large-power system based on inference from the simulation data. First, we establish a data-based analytic relationship between the adopted maintenance strategies and the estimated RCB. Then, we formulate the component maintenance decision-making problem as a nonlinear 0–1 programming problem. We then quantify the credibility of the estimated RCB and develop an adaptive method to determine the minimum required number of simulations, which is a crucial parameter in the optimization model. Finally, we devise two heuristic algorithms to efficiently identify approximately optimal solutions. The proposed method is then validated by way of numerical experiments based on IEEE standard systems and an actual provincial system.