Optimum clustering of active distribution networks using back tracking search algorithm

Abstract

A microgrid has become the main building block of future distribution networks, demanding a systematic procedure for its optimal construction. Large distribution systems can be divided into clusters of distributed energy resources serving a group of distributed loads, known as microgrids to facilitate powerful control and operation. This paper compares the clustering of smart distribution systems into a set of microgrids based on two different objective functions. The probabilistic nature of intermittent distributed generators and loads is considered by performing a probabilistic Backward-Forward sweep power flow. The probabilistic approach aims to determine the optimal virtual cut set lines that split the system into self sufficient microgrids. The Back Tracking Search optimization algorithm is used to cluster a 69-bus distribution system with optimally allocated distributed generation units. The design concept, problem formulation, solution algorithms, probabilistic model and other graph related theories are presented in this paper.