Reliability assessment of distribution networks through graph theory, topology similarity and statistical analysis

This study aims to investigate the topology characteristic indices which have a close relationship with reliability of distribution networks via mathematical tools. First, a practical network is extracted into a topology and presented by weighted adjacency matrix associated with a failure degree matrix, which contains more significant information than that of original adjacency matrix. Second, singular value sequence and vertex/edge sharing are proposed as two similarity indices. Finally, similarity functions are applied on two cases to establish the relationship with reliability by curve fitting and clustering. optimisation strategy based on modified contingency-load-loss index is elaborated to improve the reliability and guide to distribution network planning.