Reducción del espacio de búsqueda aplicada a la planificación óptima de sistemas transmisión de energía eléctrica de gran escala

Abstract

Transmission Network Expansion Planning (TNEP) is an essential activity to ensure that the electric power system can meet future demand at the lowest cost while meeting technical and environmental constraints. The current challenge for implementing any TNEP methodology is the computational complexity derived from the size of transmission systems and the high number of candidate lines, which results in large search spaces and therefore unmanageable computation times. For that reason, in this work, we propose a methodology based on Search Space Reduction (SSR) using clustering algorithms to produce a subset or group of candidate lines in which the optimal solution is ideally found. The performance of unsupervised algorithms such as Kmeans, PAM, and CLARA is compared, considering the overload, cost benefit, and least effort criteria as clustering methods. To validate that the search subspace contains the optimal solution, the Improved version of DE-PBILc hybrid metaheuristic (IDE PBILc) is used as an optimization technique. Finally, the methodology is applied to four electrical test systems of different scales: the 6-bus Garver system, the 24-bus IEEE system, the modified 118-bus IEEE system, and the modified 300-bus IEEE system. This results in significant reductions in processing time, 12.05%, 98.85%, 78.81%, and 95.76%, respectively, due to the reduction in computational complexity. It is also verified that the optimal solution is part of the subspace obtained for each of the test systems.