Improved contingency measures for operation and planning applications

Abstract

Contingency selection methods should be capable of: (a) separating critical (contingencies that have thermal and voltage limit violations) and noncritical (contingencies that have no thermal or voltage limit violations) contingencies; (b) having no false alarms where a contingency is considered to be critical based on the magnitude of the contingency measure but has no thermal or voltage limit violations; (c) having no misclassifications where a contingency that is critical and has thermal or voltage limit violations has a small contingency measure and is not considered to be critical; (d) ranking contingencies based on the magnitude of the largest thermal or voltage limit violation; Present contingency measures [1-5] do not perform well in meeting objectives (a-d) on heavily loaded systems since these contingency measures sum contributions for all branches (or all buses for voltage limit problems) for thermal limit problems. These contributions are relatively large for a heavily loaded system and since so many branch contributions for thermal problems (bus voltage contributions for voltage problems) are added for large data bases, the contingency measures lose their accuracy and validity for selecting critical contingencies (contingencies with branch contributions greater than some threshold). A set of three fundamentally new contingency measures are proposed that clearly separate critical and noncritical contingencies; have no misclassification or false alarm problems, and rank contingencies based on their worst thermal limit violation. The contingency measures are weighted by a probability of occurrence, which is extremely important in on-line operation based on [6] and