Analysis of Automatic Contingency Selection Algorithms

This paper presents a method for evaluating the effectiveness of the presently used Automatic Contingency Selection (ACS) algorithms in capturing contingencies which give out-of-limit conditions. The evaluation is given in terms of both capture and false- alarms rates, which are computed a priori together with a stopping criterion. Furthermore a method is given for enhancing the ACS algorithm´s effectiveness in terms of maximizing the capture rate and minimizing the false-alarm rate. The theory is given for selecting the set of weights in the scalar performance index (PI) for both real power and voltage problems in order to circumvent some of the misranking problems, and more importantly, how to select the threshold value of the PI in order to capture all the critical contingencies while minimizing the capture of the non- critical ones. Since the PI used in the paper is similar to that used in the present ACS algorithms, the results give a systematic method for tuning these algorithms in order to maximize their effectiveness for the particular power system under study.