MO-MODBO-MCS-Based Stochastic Multi-Objective Planning of Distributed Generators and EV Charging Stations in Radial Distribution Networks Under Voltage-Dependent Loads

Abstract

The incorporation of clean distributed generators (DGs) to fast-charging EV hubs presents significant operational challenges to modern distribution grids. The intermittent nature of solar and wind generation, combined with the uncertainty of EV charging demand, makes the optimal planning of these resources particularly challenging. Inappropriate placement can increase distribution line losses, cause unacceptable voltage deviations, and lead to thermal overloading. To address these challenges, we developed a multi-objective planning framework that combines a modified dung beetle optimizer with Monte Carlo simulation (MO-MODBO-MCS) under voltage-dependent load uncertainty. Our formulation solves a 15-dimensional joint decision vector to determine the optimal bus locations, megawatt capacities, and technology types for three DGs and two fast-charging plazas across four static ZIP load profiles (CNL, IDL, RSL, and CML). Testing on the IEEE 38-bus and IEEE 69-bus radial distribution systems demonstrates that the proposed method simultaneously reduces expected active power losses, total voltage deviation, and proxy capital investment costs. Across 100 independent trial runs under commercial loading, MO-MODBO-MCS exhibited the lowest run-to-run variability, achieving a standard deviation of σ=0.0543 kW (0.232% relative variance) and outperforming the NSGA-II-MCS and GA-MCS benchmarks without crowding distance collapse. The framework incorporates a lower-truncated Gaussian distribution for vehicle fleets to prevent non-physical negative charging power under extreme load variations. The proposed approach provides a comprehensive multi-objective planning architecture for distribution network operators (DNOs) to coordinate DG and EVCS deployment under voltage-dependent load models and stochastic operational regimes.