WFCRL
收藏资源简介:
WFCRL是由法国国家信息与自动化研究所和巴黎高等师范学院联合创建的首个开源风电场控制强化学习环境套件。该数据集包含10个风电场布局,其中5个基于真实风电场,涵盖了从7到91个风力涡轮机的不同规模。数据集提供了两种风电场模拟器接口:静态模拟器FLORIS和动态模拟器FAST.Farm,支持从静态到动态模拟器的迁移学习策略。WFCRL旨在通过多智能体强化学习优化风电场的功率输出,同时减少风力涡轮机的结构损伤,适用于风电场控制、功率最大化及疲劳负载管理等研究领域。
WFCRL is the first open-source reinforcement learning environment suite for wind farm control, jointly created by the Institut National de Recherche en Informatique et en Automatique (INRIA) and École Normale Supérieure de Paris. This dataset includes 10 wind farm layouts, 5 of which are based on real-world wind farms, covering various scales ranging from 7 to 91 wind turbines. The dataset provides interfaces for two wind farm simulators: the static simulator FLORIS and the dynamic simulator FAST.Farm, supporting transfer learning strategies between static and dynamic simulators. WFCRL aims to optimize the power output of wind farms via multi-agent reinforcement learning while reducing structural damage to wind turbines, and is suitable for research fields including wind farm control, power maximization, and fatigue load management.




