Adaptive online control of cascading blackouts

We study online control algorithms to be deployed in the event of a cascading power system failure. The control mechanism is computed after the initial event that sets off the cascade and is applied as the cascade unfolds, with the goal of reaching a stable state while shedding a minimum amount of load. We focus on robust controls, using models of line outages that explicitly account for noise. Computational experience is presented using simulated cascading failures of the U.S. Eastern Interconnect.