Offline RL algorithm boosts wind farm power by 10% in wind tunnel tests

Researchers have developed a novel offline reinforcement learning algorithm, MTD3-BC, to optimize wind farm power generation by controlling yaw systems. This algorithm, validated through wind tunnel experiments, effectively reduces wake effects and increases farm-level power by approximately 10% compared to a greedy strategy. MTD3-BC achieves this without requiring a wake model or extensive interaction with simulators, significantly lowering computational costs and training time. AI

IMPACT This research demonstrates a more computationally efficient method for optimizing renewable energy generation using AI.

RANK_REASON Academic paper detailing a new algorithm and experimental validation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• greedy strategy
• Modified Twin Delayed Deep Deterministic Policy Gradient with Behavior Cloning
• MTD3-BC
• reinforcement learning
• Wake effects on the energy loss as a function of the scattering angle for thin-film-transmitted H2 + ions
• Wake model for horizontal-axis wind and hydrokinetic turbines in yawed conditions
• wind farm
• wind tunnel
• Yaw control systems for tailsitting biplane aircraft

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →