Decentralized multi-objective nonlinear droop control-based strategy for power sharing in low-voltage agricultural microgrids

Abstract

The increasing demand for efficient and sustainable energy solutions in the agri-food sector, particularly in low-voltage microgrids (MG) deployed in rural areas, presents specific challenges related to power quality, resource availability, and operational resilience. These systems, typically composed of heterogeneous distributed generation (DG) units, require control strategies capable of addressing multiple and often conflicting objectives, such as minimizing operational costs and emissions, improving system efficiency through reduced distribution losses, and extending the lifespan of energy storage systems, all while ensuring system stability and reliability. In this context, this thesis proposes a decentralized control strategy based on nonlinear droop characteristics and user-guided multi-criteria operation adjustment, tailored for the autonomous operation of agricultural microgrids with grid-forming (GFM) generators. The proposed approach enables proportional power-sharing based on user-defined multi-objective criteria, sequential dispatch of DG units, and local voltage and frequency regulation, even in the absence of high-bandwidth communication infrastructure. Given the unavailability of detailed historical power demand data in rural settings, the strategy was designed to operate independently of such datasets, offering a robust initial solution for the development and deployment of agricultural microgrid projects. The proposed methodology was experimentally validated through real-time tests conducted on a low-voltage microgrid prototype under a Hardware-in-the-Loop (HIL) framework at the Microgrid Research Laboratory (MGLab) of Aalborg University, Denmark. The preliminary experimental results demonstrated the effectiveness of the strategy in enabling decentralized coordination of DG units, maintaining system stability, and significantly reducing the considered operational objectives, approximately between 10% to 72% in operational costs and 20% to 68% in emissions depending on the selected online power reserve, when compared to the conventional resistive droop control benchmark.