Work

Metaheuristic Optimization of Multi-PV Placement in Distribution Networks with Dynamic Load Profiles

Year: 2026
Type: article
Abstract: The strategic integration of photovoltaic (PV) systems in distribution networks is essential for reducing power losses and enhancing voltage stability. However, traditional planning approaches often neglect the dynamic nature of daily load variations, leading to suboptimal PV deployments. This study addresses this gap by proposing a time-varying, load-sensitive optimization framework for multi-PV placement and sizing in the IEEE 33-bus radial distribution system. The objective is to minimize 24-hour active power losses across industrial, residential, and commercial load scenarios. A voltage-dependent load model and empirical PV generation data are integrated into the load flow analysis. The Coyote Optimization Algorithm (COA) is employed to solve a mixed-integer nonlinear programming (MINLP) problem, which involves both discrete variables (PV bus locations) and continuous variables (PV sizes), governed by nonlinear power flow constraints. The performance of COA is benchmarked against Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). COA achieved the lowest active power loss in all cases. Under commercial loading, three PV units reduced active power loss by 38.48% relative to the base case and improved the voltage profile. In this scenario, Wilcoxon signed-rank tests found significant differences between COA and both PSO and GA over 30 runs. These findings underscore the importance of realistic load modelling and advanced optimization in planning resilient, high-performance PV deployments. The proposed methodology provides a realistic and statistically validated framework for PV planning studies involving distributed renewable generation.
Language: English
Open Access status: diamond
OpenAlex ID: W7236589961
FWCI:
Cites: 37
Cited by:
Related to: 10