Adaptive procedure for masking effect compensation in contingency selection algorithms

An adaptive procedure developed for the compensation of the masking effect in automatic performance index (PI) contingency ranking algorithms is described. A vector-norm-based PI formulation is introduced. It allows a quantitative description of the masking effect. The advantage of the adaptive procedure is that the security analysis is carried out cyclicly. The verification of the ranking after each exact outage simulation allows updating factors to be estimated, taking into account a sequence of former system states. These factors are assumed to be valid at the time of the next security check in order to update the second-order PI ranking. Thus, the adaptive procedure combines the computational speed of the standard PI/sub 2/-methods with the accuracy of the masking-free PI/sub 20/-formation. Extensive tests for realistic European power systems considering actual load and topology changes indicate a substantial reduction of the masking effect and therefore a more reasonable contingency ranking.<>