Modelo optimización robusta para la planificación de la topología de un sistema de distribución a gran escala

Abstract

The integration of technologies such as distributed energy resources, microgrids, and smart grids has significantly transformed the design and operation of distribution networks, requiring new planning models that ensure efficiency and continuity of supply. The lack of access to detailed data from real distribution networks motivates this project, which proposes synthetic networks that replicate real-world characteristics without exposing sensitive information. To achieve this goal, geographic information systems, graph theory, and network statistics are employed to design both secondary and primary distribution systems. In this context, the algorithms define service drops, locate and size distribution transformers, secondary network layouts, structure primary feeders, and position support elements. Within the secondary system, length and proximity constraints are enforced using a KD-Tree to optimize spatial organization, alongside a DFS-P algorithm that merges depth-first search with tree pruning. This methodology places transformers under spatial constraints and avoids unrealistic configurations. For the primary distribution system, minimum spanning trees and optimal routes are devised based on distances among substations, load nodes, and distribution transformers. The distribution system design incorporates street and load maps, processed to correct geometric and attribute inconsistencies. Finally, the algorithms were implemented in an area served by Empresa Eléctrica Regional Centro Sur C.A., demonstrating that this methodology yields realistic and robust networks with statistics comparable to realistic systems.