Adaptive Dynamic Droop Control for Photovoltaic-Rich Distribution Networks Considering Photovoltaic Uncertainties

Abstract

ABSTRACT The integration of photovoltaic (PV) units into distribution networks is becoming crucial as renewable energy sources gain more prominence. However, the intermittent nature of solar energy causes voltage fluctuations and power losses that need to be addressed. To address these challenges associated with integrating PV units into distribution networks, a novel approach is proposed that provides reactive power ancillary support using a decentralized local dynamic droop control (DDC) strategy. The proposed approach, combined with a meta-heuristic reptile search algorithm (RSA) technique that identifies the appropriate sizing and placement of multiple PV arrays in distribution networks, will substantially improve the voltage profile and full utilization of the PV inverter’s capacity to provide reactive power support over a 24-hour period. The PV systems are modeled using the Lambert-W function, which simplifies the estimation of PV at its maximum power point (MPP) without resorting to computationally intensive MPP tracking techniques. The study leverages 24 hours of realistic data, including ambient temperature and solar irradiance, together with the inverter’s terminal voltage, to estimate the amount of reactive power an inverter can handle. The proposed method substantially improves the voltage profile with full utilization of the PV inverter’s capacity to provide reactive power support over a 24-hour period. The study demonstrates that this approach prioritizes a decentralized network and PV inverters over a centralized management system, making it a promising option for future PV systems that require improved grid compatibility and integration without the installation of additional reactive power devices. In addition, a customized performance index based on the integral square of the voltages at various buses where PVs are located in IEEE-33 and IEEE-69 test bus systems, summed over a 24-hour time window, was calculated for comparative analysis. The proposed method results in an improvement of 93.9% and 95.6% over the base case test systems, whereas conventional droop control results in an improvement of 33.47% and 30.9% only over the base case, respectively, for the test systems.