Abstract

This study presents an optimization strategy to improve the operation of electrical distribution networks by altering their topology. The main goals were to reduce real power losses and improve service reliability. These objectives were combined into a single objective using a weighted-sum approach. The study used two metaheuristic methods, namely, the genetic algorithm (GA) and particle swarm optimization (PSO), to identify the best switching setups. To maintain the network as radial and within operational limits, graph-theoretic connectivity rules, such as loop, common, and prohibited branch vector rules, were applied. For each possible network layout, a sweep-based power-flow algorithm computed the nodal voltages, branch currents, and total active losses. Reliability was measured using the minimal cut set technique, and indices, such as SAIFI, SAIDI, and ENS, were calculated. The method was first tested on the IEEE 33-bus network and then used on a real 43-bus, 11 kV feeder. The results show that PSO is more efficient than GA, making it a practical choice for optimizing distribution systems.