Scholarlink

<p>Stabilizing the output voltage of permanent magnet synchronous generator (PMSG)-based wind turbines is a critical challenge due to the inherent nonlinearities and time-varying nature of wind patterns. Traditional linear controllers, such as PID, often fail to adapt to these dynamic conditions, while the performance of conventional model predictive control (MPC) is limited by the accuracy of the underlying system model. To address these issues, this paper proposes a novel hybrid control strategy that integrates a nonlinear artificial neural network (ANN) for precise system identification with an MPC scheme. A feedforward backpropagation ANN is employed to accurately model the PMSG dynamics (R ≈ 0.999), serving as the predictive foundation for the MPC optimizer. Comparative simulations reveal that the proposed ANN-MPC strategy significantly outperforms traditional PID controllers, reducing voltage settling times from approximately 3.5&#xa0;s to under 0.1&#xa0;s during wind speed transients. Experimental validation further confirms the robustness of the approach, demonstrating a settling time of 0.7&#xa0;s and limiting peak voltage overshoot to approximately 3.5% under variable wind conditions. These findings indicate that the ANN-MPC framework offers superior disturbance rejection and voltage stability, presenting a viable solution for enhancing the power quality and reliability of modern wind energy conversion systems (WECSs).</p>

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

Output voltage stabilization of a PMSG wind turbine by nonlinear ANN identification-based MPC

  • 期刊论文

  • 发表日期：

    2026年7月9日

查看全文
  • Kerim Karabacak

摘要

Stabilizing the output voltage of permanent magnet synchronous generator (PMSG)-based wind turbines is a critical challenge due to the inherent nonlinearities and time-varying nature of wind patterns. Traditional linear controllers, such as PID, often fail to adapt to these dynamic conditions, while the performance of conventional model predictive control (MPC) is limited by the accuracy of the underlying system model. To address these issues, this paper proposes a novel hybrid control strategy that integrates a nonlinear artificial neural network (ANN) for precise system identification with an MPC scheme. A feedforward backpropagation ANN is employed to accurately model the PMSG dynamics (R ≈ 0.999), serving as the predictive foundation for the MPC optimizer. Comparative simulations reveal that the proposed ANN-MPC strategy significantly outperforms traditional PID controllers, reducing voltage settling times from approximately 3.5 s to under 0.1 s during wind speed transients. Experimental validation further confirms the robustness of the approach, demonstrating a settling time of 0.7 s and limiting peak voltage overshoot to approximately 3.5% under variable wind conditions. These findings indicate that the ANN-MPC framework offers superior disturbance rejection and voltage stability, presenting a viable solution for enhancing the power quality and reliability of modern wind energy conversion systems (WECSs).