Print Email Facebook Twitter Sailing the Wind: Evaluating the Impact of COMA on Multi-Agent Active Wake Control in Wind Farms Title Sailing the Wind: Evaluating the Impact of COMA on Multi-Agent Active Wake Control in Wind Farms: What is the effect of COMA on the problem of AWC compared to single-agent RL algorithms? Author Filimon, Mihai (TU Delft Electrical Engineering, Mathematics and Computer Science) Contributor de Weerdt, M.M. (mentor) Neustroev, G. (mentor) Pawełczak, Przemysław (graduation committee) Degree granting institution Delft University of Technology Programme Computer Science and Engineering Project CSE3000 Research Project Date 2023-06-30 Abstract The close proximity of wind turbines to one another in a wind farm can lead to inefficiency in terms of power production due to wake effects. One technique to mitigate the losses is to veer from their individual optimal direction. As such, the wakes can be steered away from downstream turbines in order to increase the overall power output. Multi-Agent Reinforcement Learning (MARL) models the interactions between wind turbines and determines an optimal control strategy through agents that learn the collective consequences of their actions. Toanalyse the benefit of multi-agent cooperation and centralised critic evaluation, I investigated the effect of Counterfactual Multi-Agent Policy Gradients (COMA) on Active Wake Control. Ultimately, experiments on wind farms of three and sixteen turbines indicate that the algorithm performs moderately, yet worse than single-agent Reinforcement Learning. In addition, high computation costs hinder its application on real-life environments. Subject Active Wake ControlWind EnergyMulti Agent Reinforcement Learning To reference this document use: http://resolver.tudelft.nl/uuid:5f0c0e5f-59aa-4fd6-8d94-deeefa5b60a4 Part of collection Student theses Document type bachelor thesis Rights © 2023 Mihai Filimon Files PDF CSE3000_Final_Paper_Mihai ... ilimon.pdf 399.4 KB Close viewer /islandora/object/uuid:5f0c0e5f-59aa-4fd6-8d94-deeefa5b60a4/datastream/OBJ/view