Hosting Capacity Discovery and Multi-Objective DG Allocation in Real Radial Distribution Systems via PSO
Abstract
The rapid integration of Distributed Generation (DG) units is transforming traditional passive radial distribution networks into active grids with bidirectional power flow. Determining the optimal allocation and physical hosting capacity of DG units is essential for maintaining grid stability while maximizing technical gains. This study presents a multi-objective optimization framework using empirical peak-load field data from an active medium-voltage distribution network in Türkiye. A custom Particle Swarm Optimization (PSO) algorithm was developed in MATLAB to simultaneously minimize active and reactive power losses while improving the Voltage Deviation Index (VDI). The baseline network model and load flow calculations established in DIgSILENT PowerFactory were rigorously cross-validated with MATPOWER, demonstrating complete mathematical agreement with a negligible deviation (<0.001%). Simulation results under both capacity-constrained (10 MW limit) and unconstrained Hosting Capacity Discovery scenarios demonstrate that single-objective optimization causes negative trade-offs across other operational metrics. Conversely, multi-objective hybrid approaches provide a balanced, globally common optimum operating point best suited to the grid’s natural electrical structure, consistently identifying a reliable ultimate hosting capacity threshold of 21.78 MW at Busbar 4. These findings offer a concrete, empirical decision-support framework for active distribution network planning and modernization.