Abstract
The reliability of a power system depends, among other things, on the operating states of the system. For reliability analysis for long-term planning purposes, it may be necessary to consider a large set of representative operating states, and clustering techniques can be applied to reduce this set to make the analysis computationally tractable. However, the values of reliability indices for the system may be dominated by the contributions from a relatively small number of high-impact operating states that are not easily captured in the analysis. The objective of this paper is to identify high-impact operating states in the context of power system reliability analysis based on clustering techniques. A secondary objective is to characterize features of these operating states to better understand how they could be recognized, prepared for, and if possible avoided. An approach and an algorithm are proposed, and they are illustrated using a reliability analysis case study for a region of the Norwegian transmission grid with realistic operating states for the Nordic power system
