A Model Predictive Approach to Preventing Cascading Failures of Power Systems

Abstract: Power system blackouts are usually triggered by the initial contingency and then deteriorate as the branch outage spreads quickly. Thus, it is crucial to eliminate the propagation of cascading outages in its infancy. In this paper, a model predictive approach is proposed to protect power grids against cascading blackout by timely shedding load on buses. The cascading dynamics of power grids is described by a cascading outage model of transmission lines coupled with the DC power flow equation. In addition, a nonlinear convex optimization formulation is established to characterize the optimal load shedding for the mitigation of cascading failures. As a result, two protection schemes are designed on the basis of the optimization formulation. One scheme carries out the remedial action once for all, while the other focuses on the consecutive protection measures. Saddle point dynamics is employed to provide a numerical solution to the proposed optimization problem, and its global convergence is guaranteed in theory. Finally, numerical simulations on IEEE 57 Bus Systems have been implemented to validate the proposed approach in terms of preventing the degradation of power grids.