A study of reliability assessment and topology similarity: Methodology, relationship, and applications

Abstract

Summary This paper attempts to investigate the relationship between topology and reliability of a distribution network associated with a singular value sequence (SVS) to accurately represent the crucial features of the distribution network topology, in which a novel algorithm called topology similarity analysis (TSA) is proposed for the first time and adopted to calculate the reliability indices of the distribution networks. Under such framework, SVS is treated as a similarity index that constructs a standard database of distribution network topology. Moreover, TSA outperforms other typical algorithms used in distribution network reliability assessment with a lower calculation costs and higher efficiency. Two case studies, eg, standard IEEE 33‐bus system and large‐scale practical 148‐bus system, are carried out at first to evaluate the effectiveness of the proposed approach. Then, TSA incorporated with splitting method is applied in the 148‐bus case. Simulation results demonstrate that TSA is 2 to 2000 times faster than that of other typical algorithms. Besides, the overall calculation time of TSA just grows insignificantly as the scale of the distribution network increases, together with a guaranteed less than 0.005% precision error.

Unsupervised machine learning techniques applied to composite reliability assessment of power systems

“…In order to speed up the process of assessing the reliability of electrical systems (generation, transmission, and distribution), several works in the related literature have proposed innovative strategies. Highlight can be given to the use of variance reduction techniques, 8,12,21,24 fuzzy representation methods, 7,14 fast sorting algorithm, 17 multiparametric linear programming, 28 topology similarity analysis, 30 state merging technique, 26 parallel and distributed computing, 4,11,19,20 smart population‐based optimization methods, 16,18,23‐25 and pattern recognition techniques 9,10,13,15,20,21,27,29 . In general, many of these strategies are hybrid tools that combine the use of different techniques with Monte Carlo simulation (MCS) methods 6 …”