Scholarlink

The precision of wind power forecasting significantly influences the operation of power systems. Traditional models often fall short in delivering adequate accuracy for wind power predictions. To address this limitation, this paper presents an innovative short-term wind power forecasting model based on NGO-optimized CNN-BiGRU-Attention. The Pearson correlation coefficient is first applied to identify wind power parameters with strong correlations. Then, CNN is employed to extract feature vectors, and the BiGRU network models the temporal dynamics of these features across past and future time steps. An Attention mechanism is incorporated to assign adaptive importance weights to the hidden states of the BiGRU, enhancing the model's focus on key information. Additionally, the Northern Goshawk Optimization (NGO) algorithm is utilized to fine-tune the parameters of the integrated framework. Comparative simulations demonstrate the proposed model’s effectiveness and superior performance, validating its practical applicability and robustness.

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

Short-Term Wind Power Prediction Method Based on NGO Optimization and Composed of CNN-BiGRU-Attention

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Duanchao Li,
  • Zhilong Huang,
  • Wei Wang,
  • Li Zhang,
  • Feixiang Peng,
  • Jun Tao

摘要

The precision of wind power forecasting significantly influences the operation of power systems. Traditional models often fall short in delivering adequate accuracy for wind power predictions. To address this limitation, this paper presents an innovative short-term wind power forecasting model based on NGO-optimized CNN-BiGRU-Attention. The Pearson correlation coefficient is first applied to identify wind power parameters with strong correlations. Then, CNN is employed to extract feature vectors, and the BiGRU network models the temporal dynamics of these features across past and future time steps. An Attention mechanism is incorporated to assign adaptive importance weights to the hidden states of the BiGRU, enhancing the model's focus on key information. Additionally, the Northern Goshawk Optimization (NGO) algorithm is utilized to fine-tune the parameters of the integrated framework. Comparative simulations demonstrate the proposed model’s effectiveness and superior performance, validating its practical applicability and robustness.