Analysis of Flexible Distribution Network Structures Incorporating Load Diversity and Distributed Generation

Abstract

Effective reduction of power losses in radial distribution systems can be achieved through strategic network reconfiguration—disconnecting existing branches and connecting alternative ones—and through optimal siting of distributed generation (DG) units, in which load characteristics play a decisive role. Since electrical loads vary with voltage levels and differ by consumer category, they can be represented by a combination of fixed-impedance, power, and current components. Identifying the proportion of these elements for each load category enables the formulation of more adaptive and realistic models for feeder reconfiguration and DG allocation. In this context, the current work introduces a mathematical framework that explicitly links load composition with consumer categories and reformulates a nonlinear polynomial load model into a tractable quadratic form suitable for linear solvers. The developed framework is evaluated on standard 16- and 118-bus distribution systems. For the 16-bus network, the flexible formulation achieved power losses of 66.4–67.08 kW, which closely match the exact nonlinear model results (66.9–67.04 kW) while reducing computation time from about 2000–2230 s to nearly 1.2 s. For the 118-bus system, losses of 410.05–470.03 kW were obtained, compared with 409.38–471.10 kW for the exact model, with processing time reduced from approximately 58 000–157 000 s to 26–68 s. According to the results, the suggested framework preserves high solution accuracy while providing orders-of-magnitude improvements in computational efficiency. The developed approach, therefore, offers a practical and flexible tool for real-time distribution network reconfiguration and DG planning under voltage-dependent and category-based load modeling.