Scholarlink

With the rapid development of wind power generation, it is important to reduce the maintenance expenses of wind farms by alleviating fatigue damage in wind turbines. This paper proposes an optimization strategy to mitigate fatigue damage in wind turbines. Firstly, an estimation model for thrust and torque is proposed based on aerodynamic principles, which is then refined by the application of Physics-Informed Neural Network (PINN). Subsequently, an enhanced linear cumulative damage theory is utilized to estimate the fatigue damage of wind turbines. An optimization model is then established based on the fatigue damage model. Additionally, to suppress the noise and latency in the Automatic Generation Control System (AGCS), the model’s resilience to interference is improved by the integration of Gaussian noise reduction and robust optimization model. Simulations are constructed on a dataset encompassing the operational states of 100 wind turbines. The results indicate that the proposed model exhibits superiority in suppressing fatigue damage and is beneficial to resist the interference.

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

Active Power Optimization Considering Fatigue Damage in Wind Farms Using Physics-Informed Neural Networks

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Zijun Chen,
  • Rongxiang Zhang,
  • Bo Li

摘要

With the rapid development of wind power generation, it is important to reduce the maintenance expenses of wind farms by alleviating fatigue damage in wind turbines. This paper proposes an optimization strategy to mitigate fatigue damage in wind turbines. Firstly, an estimation model for thrust and torque is proposed based on aerodynamic principles, which is then refined by the application of Physics-Informed Neural Network (PINN). Subsequently, an enhanced linear cumulative damage theory is utilized to estimate the fatigue damage of wind turbines. An optimization model is then established based on the fatigue damage model. Additionally, to suppress the noise and latency in the Automatic Generation Control System (AGCS), the model’s resilience to interference is improved by the integration of Gaussian noise reduction and robust optimization model. Simulations are constructed on a dataset encompassing the operational states of 100 wind turbines. The results indicate that the proposed model exhibits superiority in suppressing fatigue damage and is beneficial to resist the interference.