Efficiency Improvement in MV Distribution System through Feeder Reconfiguration (Case Study: Adwa Distribution System)

Abstract

The electric power industry is increasingly challenged by rising load demand, aging infrastructure, and the complex spatial distribution of electrical loads, all of which undermine distribution system efficiency. These are especially pronounced in radial networks, where high technical losses and voltage deviations compromise power quality and reliability. Adwa feeder, a 15 kV radial distribution system in Ethiopia, embodies these concerns, with technical losses reaching 940.84 kW, voltage levels dropping to 0.83 per unit, and line loading peaking at 148.14 percent. This study explores feeder reconfiguration using Particle Swarm Optimization (PSO) to improve energy efficiency, voltage stability, and line overloading under varied load conditions. Load flow analysis was conducted using the Backward/Forward Sweep method in MATLAB under peak, medium, and minimum loading conditions, with total demands of 7,839 kW, 4,710 kW, and 2,160 kW, respectively. PSO identified an optimal switching scheme by opening sectionalizing switches 52 and 71, keeping tie switches 79 and 82 open, and closing 80 and 81. The optimized configuration reduced peak losses by 35.7%, lowering them to 604.87 kW and improving minimum bus voltage to 0.9161 per unit. For medium and minimum loads, losses dropped from 570 kW to 370 kW and from 140 kW to 90 kW, respectively, improvements of 35% and 36% with corresponding voltage profile enhancements. To validate robustness, Monte Carlo simulations (1,000 iterations, ±10% load variation, ±0.05 PF deviation) confirmed the optimized topology sustained losses near 605 kW and voltages above 0.9 per unit under uncertainty. Furthermore, upgrading the most overloaded segment (Line 1–2) to an AAC 150 conductor further improves losses to 524.32 kW. While this upgrade alone provided an 8.6% gain beyond reconfiguration, the combined effect achieved a total loss reduction of 44.3% from the original case. These results demonstrate that intelligent feeder reconfiguration, enhanced by PSO and supported by probabilistic analysis, provides a scalable, cost-effective solution for improving performance in radial distribution networks like Adwa's.