Asian Basic and Applied Research Journal

Optimal power flow (OPF) is a nonlinear constrained optimisation problem in which generator dispatch and voltage-control variables are selected to minimise an operating objective while satisfying the AC power-flow equations and engineering limits. This study develops and evaluates a particle swarm optimisation (PSO) framework for the standard IEEE 14-bus test system. The formulation minimises quadratic generation cost and explicitly checks active-power limits, reactive-power limits and bus-voltage limits through a penalty-assisted feasibility mechanism. A swarm of 30 particles was evolved for 100 iterations using acceleration coefficients c1 = c2 = 2.0 and a linearly decreasing inertia weight from 0.9 to 0.4. The algorithm was assessed over 20 independent runs. In the best run, total generation cost decreased from the reference power-flow cost of $8,172.00/h to $8,081.535/h, a reduction of 1.107%. Real-power loss decreased from 13.393 MW to 9.285 MW, corresponding to a 30.67% reduction. The optimised bus-voltage range was 1.0145 to 1.0600 p.u., satisfying the adopted 0.94 to 1.06 p.u. limits. The best PSO objective was only 0.000115% above the approximately $8,081.526/h deterministic MATPOWER benchmark. Across 20 runs, the mean cost was $8,085.218/h with a standard deviation of $9.014/h, while the mean execution time was 1.669 s. The results indicate that a carefully constrained PSO implementation can provide near-benchmark OPF solutions while retaining the flexibility of population-based search.