A parallel asynchronous decomposition for on-line contingency planning

Abstract

Traditional formulations of security-constrained-optimal-power-flows represent contingencies by hard constraints. The disadvantages are four-fold. First, the conflicts among contingencies must be arbitrated apriori, before their effects are known. Second, the feasible region shrinks with an increase in the number of contingencies. Third, computational time increases with the number of contingencies. Fourth, hard constraints provide poor models of fuzzy quantities such as equipment ratings and operating guidelines. This paper develops a modeling framework without these disadvantages. Specifically, it allows for soft constraints and always has feasible solutions. The effects of conflicts among contingencies are displayed so system operators can arbitrate them in an informed manner. Moreover, each contingency can be handled asynchronously and in parallel. In other words, the computational time, for handling an arbitrarily large number of contingencies, remains the same as for performing an optimal power flow without any contingencies (provided that a computer is dedicated to each contingency).