Static Reconfiguration of Unbalanced Distribution Systems with Variable Power using Selective Bat Algorithm

Abstract

Distribution network reconfiguration (DNR) is one of the most studied and important optimization problems in Electrical Engineering field, being well established throughout the years, as seen in the specialized literature. The majority of studies treats the problem in a conservative way, i.e., considering the systems as balanced with fixed demand. However, the real characteristics of distribution systems may take into consideration the unbalanced nature of loads, as well as the variable demand characteristic according to the type of consumer connected. This comes closer to reflect the real operation of such systems, and thus, studies considering DNR with these aspects are becoming more important in recent years. In this context, the paper here presented proposes modeling unbalanced systems with variable demand using a static reconfiguration (SR). As SR considers the whole studied period to define one single set of switches, it is opposed to a dynamic reconfiguration (DR), which is performed in short periods. Considering the load profiles presented, a comparison is made to discuss the advantages and disadvantages of SR in the expanse of DR in unbalanced systems. To solve the DNR problem, a selective bat algorithm (SBAT) is used, as it presents characteristics from particle swarm optimization (PSO) and harmony search (HS). Tests for SR and DR with variable demand are performed in one balanced system (33 buses) and three unbalanced systems (19, 25 and 123 buses). Results for SR using SBAT are compared with selective PSO and selective HS.