摘要
针对目前的风力发电功率预测方法并不能很好地捕捉局部信息,预测精度相对较低的问题,提出了一种基于改进Transformer模型的风力发电功率预测模型。通过模糊信息粒化和时间卷积网络对数据局部特征进行提取,并将其输入Transformer模型中对发电功率进行预测。结果表明,该模型的RMSE、MAE和R~2分别为0.170、0.121、0.826,且其区间覆盖率指标数值为0.944。该模型能准确有效预测风力发电功率。
In response to the current issue that wind power generation power prediction methods cannot effectively capture local information and have relatively low prediction accuracy,an improved Transformer-based wind power generation power prediction model is proposed.By using fuzzy information granulation and temporal convolution networks to extract local features from the data,these features are then input into the Transformer model for power generation prediction.The results show that the RMSE,MAE,and R2 of the model are 0.170,0.121,and 0.826,respectively.Moreover,the interval coverage index value is 0.944,indicating that the model can accurately and effectively predict wind power generation power.
作者
张东
ZHANG Dong(Shandong Guohua Times Investment and Development Co.Ltd.,Ji’nan 250001,China)
出处
《工程建设与设计》
2025年第17期164-167,共4页
Construction & Design for Engineering
