Intelligent wind farm control via deep reinforcement learning and high-fidelity simulations

Data-Driven Wind Farm Control via Multiplayer Deep Reinforcement Learning

“…A DRL-based yaw control method that aimed to handle stepwise-varying inflow conditions was designed in [13], but it was built upon a steady-state wind farm model. Another two wind farm control approaches via DRL were proposed in [14], [15]. They are model-free and can adapt to time-varying wind speeds.…”

“…These essential features render our method to have enhanced performance. • Distinct from the most recent DRL-based wind farm control approaches [12], [13], [14], [15], a special distractor network is employed in our method. This design not only mitigates the non-Markovian feature induced by the stochastic nature of wind conditions but also enhances our method's robustness.…”

“…It is noteworthy that the method proposed in this paper is partially built upon [14], [15], [4], and substantial new contributions have been made. Specifically, though Refs.…”

Intelligent Wind Farm Control via Grouping-Based Reinforcement Learning

“…It has attracted worldwide research interest and been applied to many important fields [15], [16], [17]. Notably, several recent studies [18], [19], [20], [21] successfully applied DRL to address wind farm control tasks. They verified the feasibility of employing DRL to maximize wind farms' economic profitability.…”

Data-driven fluid mechanics of wind farms: A review

“…An overview of the objectives of control and diagnosis and also wind-speed forecasting can be found in Refs. [14][15][16][17][18][19][20][21][22][23] . Wind-farm flow modeling with a data-driven approach is covered in this review article.…”