A Review of Reactive Power Optimization in Distribution Networks with High Penetration of Renewable Energy

Abstract

The global energy mix is undergoing an accelerated transition towards cleaner, lower-carbon sources. As a result, the large-scale integration of a high proportion of renewable energy into distribution networks is an irreversible trend. However, since the output of renewable energy sources is highly random, intermittent, and fluctuating, and given that the high r/x parameter characteristics of distribution grids result in a pronounced active-reactive power coupling, the traditional unidirectional voltage regulation mode in distribution grids is unable to effectively address issues such as bidirectional power flows and frequent voltage overshoots. Consequently, reactive power and voltage control face significant challenges. To this end, this paper provides a systematic and logically coherent review of the problem of reactive power optimization in distribution networks with a high proportion of renewable energy. First, we clarify the mechanisms by which the grid integration of renewable energy sources affects distribution network voltages. This study then systematically summarizes methods for active power regulation, reactive power compensation, tap-changer voltage regulation, and multi-device coordinated control. Subsequently, this study proposes an approach to multi-timescale, hierarchical, and zoned modeling that coordinates active and reactive power. This study then conducted a comparative analysis of uncertainty management techniques, including stochastic optimization, robust optimization, and fuzzy theory. Finally, this study summarizes the application characteristics of traditional mathematical methods, intelligent algorithms, and data-driven methods. This naturally and appropriately leads to the conclusion that the coordination of generation, grid, load and storage, combined with multi-timescale optimization, represents the current mainstream control framework, and that data-driven methods hold immense potential for application. However, it also objectively and cautiously highlights various issues in existing research, including uncertainties that are difficult to address, multi-scale collaborative complexity, communication and privacy constraints, and a lack of interpretability in AI models. Consequently, the future direction of development should focus on collaboration, intelligence and decentralization, whilst the refinement of relevant technologies will directly and effectively underpin the safe and efficient operation of distribution networks with a high proportion of renewable energy.