Maximization of wind turbine energy generation constrained to noise reduction using Reinforcement Learning M. de Frutos∗,1, D. Huergo1, O. Mari ̃no1 and E. Ferrer1,2 1 ETSIAE-UPM-School of Aeronautics, Universidad Polit ́ecnica de Madrid, Plaza Cardenal Cisneros 3, E-28040 Madrid, Spain 2 Center for Computational Simulation, Universidad Polit ́ecnica de Madrid, Campus de Montegancedo, Boadilla del Monte, 28660 Madrid, Spain (∗) e-mail : m.defrutos@upm.es Keywords: Reinforcement Learning, Wind Energy, Wind Turbine Control, Aeroacous- tics, Renewable energy, Reduced order model (ROM). The growing need for sustainable energy sources has led to a focus on improving wind turbine performance. As wind farms increase in size and number, concerns about their environmental impact have also risen. This study explores the use of reinforcement learning techniques (RL) to enhance wind turbine energy generation while tackling the crucial issue of reducing noise. We propose a control method that optimizes power production within specified noise limits. Previously, our research group developed a deep reinforcement learning methodology for optimizing the control parameters of a wind turbine, enabling it to dynamically adjust to varying wind conditions, [1]. However, this work involves revisiting classic RL algorithms [2], with the aim of accelerating the training process and ensuring algorithm convergence. We advocate for a control methodology that harmonizes power production while adhering to predefined noise generation limits. To achieve this objective, we employ a Multi-Objective Reinforcement Learning framework. This approach allows us to identify a Pareto optimal policy for controlling the wind turbine, striking a balance between maximizing power output and minimizing noise generation. Additionally, we delve into the analysis of the impact of various control parameters on the noise mechanisms of wind turbines [3], which helps us to define the state-reward structure of the reinforcement learning agent. REFERENCES [1] D. Soler, D.Huergo, O.Mariño, M. de Frutos, E. Ferrer, Reinforcement learning to maximise wind turbine energy generation, under review at Expert Systems with Ap- plications. [2] Richard S. Sutton and Andrew G. Barto, Reinforcement Learning: An Introduction, MIT Press, 2018. [3] Siegfried Wagner, Rainer Bareiß, Gianfranco Guidati, Noise mechanisms of wind turbines. Wind Turbine Noise, 1st Edition , Springer, 1996.