Optimal Sizing of PV Systems in a Utility Distribution Feeder Using OpenDSS

This paper presents a computational approach for optimal sizing of photovoltaic (PV) systems within a main distribution feeder of a real-world distribution network with thousands of electrical nodes. The analysis was performed through an OpenDSS-MATLAB Quasi-Static Time Series (QSTS) environment and initially involves modal analysis to identify the designated PV system locations. A conventional Particle Swarm Optimization (PSO) algorithm determines the optimal size for individual PV units, considering constraints such as Reverse Power Flow (RPF) and voltage limits. Results show the impact of increasing PV capacity on RPF violations and total system power loss. The approach was initially validated on IEEE 34 and IEEE 123 test systems and then applied to a real-world utility main distribution feeder. The results indicated an increase in PV capacity of up to 11.8 %. Specifically, the PV capacities at sites 1, 2, and 3 of the real-world distribution feeder were increased to 5.27MW, 2.49MW, and 1.6MW, respectively. Therefore, such a heuristic approach implemented in MATLAB and coupled with the utility distribution network in OpenDSS can facilitate the PV sizing studies for real-world distribution networks.