A Discrete-Event System Approach for Modeling and Mitigating Power System Cascading Failures

A power system cascading failure can propagate through sequential tripping of components in the network. As a result, a complete or partial shutdown may occur. In this article, we develop a new systematic approach to identify and prevent cascading failures in power systems using supervisory control of discrete-event systems (DESs). We build the DES model for a power system in a modular fashion by first modeling its components as (small) automata and then combining these automata using parallel composition. To overcome state explosion, we use online lookahead control that can significantly reduce the number of states to be considered. Since some events such as line trips cannot be disabled but can be preempted by forcing some forcible events such as load shedding, we extend supervisory control of DES to include forcible events. This extension allows us to control power systems effectively. The proposed control is implemented in an implementation platform that we build in MATLAB. The platform uses MATPOWER to simulate a power system and then control it using the proposed DES controller. Simulation studies are carried out for IEEE 6-, 30-, and 118-bus systems. The results verify the effectiveness of the proposed approach.