Representative contingency identification using data mining

This paper studies an approach to identify representative operating and contingency scenarios for the design of underfrequency load-shedding (UFLS) schemes. In small isolated power systems, contingency scenarios are outages of generating units. Not only N-1 outages are considered, but simultaneous outages of several units must be considered. Cluster analysis and classification trees are used to group scenarios in terms of system frequency and power-system parameters and to identify representative scenarios. The approach has been applied to the design of the UFLS scheme of one of the Spanish isolated power systems. Clustering techniques yielded to satisfactory results, i.e. representative operating and contingency scenarios can be identified. In addition, classification trees are able to classify new operating and contingency scenarios according to the clusters obtained.