Scholarlink

<p>Wind speed forecasting is vital for managing wind power generation, a renewable energy source. To improve prediction accuracy, this study proposes a hybrid model combining WT-VMD and ANFIS-DELM. Firstly, a two-stage data processing method is introduced, integrating wavelet transform (WT) and variational mode decomposition (VMD) to effectively capture intrinsic characteristics by decomposing trend components and residuals of the sequence. Additionally, the sparrow search algorithm (SSA) optimizes hyperparameters of the adaptive neuro-fuzzy inference system (ANFIS) and weights of the prediction sequence in the deep extreme learning machine (DELM) to achieve precise predictions. The experimental results confirm that the mean absolute percentage error (MAPE) of the proposed wind speed prediction model on the dataset covering the four seasons-spring, summer, autumn, and winter-is superior to that of other benchmark models, with values of 0.4412%, 2.0187%, 1.2146% and 5.1015%, respectively.</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

A hybrid wind speed forecasting model with two-stage data processing based on adaptive neuro-fuzzy inference systems and deep learning algorithms

  • 期刊论文

  • 发表日期：

    2025年1月3日

查看全文
  • Zhongda Tian,
  • Donglai Wei

摘要

Wind speed forecasting is vital for managing wind power generation, a renewable energy source. To improve prediction accuracy, this study proposes a hybrid model combining WT-VMD and ANFIS-DELM. Firstly, a two-stage data processing method is introduced, integrating wavelet transform (WT) and variational mode decomposition (VMD) to effectively capture intrinsic characteristics by decomposing trend components and residuals of the sequence. Additionally, the sparrow search algorithm (SSA) optimizes hyperparameters of the adaptive neuro-fuzzy inference system (ANFIS) and weights of the prediction sequence in the deep extreme learning machine (DELM) to achieve precise predictions. The experimental results confirm that the mean absolute percentage error (MAPE) of the proposed wind speed prediction model on the dataset covering the four seasons-spring, summer, autumn, and winter-is superior to that of other benchmark models, with values of 0.4412%, 2.0187%, 1.2146% and 5.1015%, respectively.