Scholarlink

This work investigates the application of input state discretization techniques in a Q-learning-based controller used for the collective pitch control of a wind turbine, to reduce convergence time and improve steady-state error. The proposed methods are evaluated using simulations, demonstrating their potential to enhance the performance of Q-learning control techniques for power output control in wind turbines, when compared to a uniform distribution.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Exponential and Quadratic State Discretizations for Wind Turbine Collective Pitch Control Based on Q-Learning

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • A. Gil Maciá,
  • J. Enrique Sierra-García,
  • Matilde Santos

摘要

This work investigates the application of input state discretization techniques in a Q-learning-based controller used for the collective pitch control of a wind turbine, to reduce convergence time and improve steady-state error. The proposed methods are evaluated using simulations, demonstrating their potential to enhance the performance of Q-learning control techniques for power output control in wind turbines, when compared to a uniform distribution.