A Multi-Objective Intelligent Control Method for Active Distribution Power Network Using Proximal Policy Optimization Agents

To effectively mitigate the operational risks of active distribution power grid with growing stochastics and uncertainties caused by renewable generation, this paper proposes a novel multi-objective intelligent control method based on the Proximal Policy Optimization (PPO) algorithm for regulating various types of controllers of active distribution network with a high ratio of photovoltaic and power electronics-based devices. The OpenDSS tool is used to model power equipment in distribution networks by incorporating the time-series load patterns, weather index, PV generation profiles for representing realistic operational scenarios. The intelligent control method successfully achieves the objectives of voltage regulation, power loss reduction and minimized control costs. Case studies demonstrate that this method significantly enhances power grid performance, exhibiting adaptability and robustness to changing environmental conditions.