可再生能源2025，Vol.43Issue(9)：1213-1220,8.
基于历史发电及气象数据的双阶段风力发电量预测方法
The forecasting method for two-stage wind power generation based on historical power gener-ation and meteorological data
摘要
Abstract
In order to predict wind power generation accurately and reliably,and further ensure the safe and stable operation of power system,a two-stage wind power generation forecasting method based on historical generation data and meteorological data is proposed,aiming at the defects of forecasting models that only use historical generation data or meteorological data.First,the historical generation data and ARIMA model are used to achieve the preliminary prediction of generation,and the residual sequence is obtained.Then,combining meteorological data and residual series,the residual series is predicted based on BiLSTM model.At the same time,PSO hyperparameter optimization algorithm is used to realize the order optimization of ARIMA model and the hyperparameter optimization of BiLSTM model.Secondly,a complete algorithm flow including training,testing and optimization is constructed.Finally,the experiment shows that the prediction accuracy of the two-stage wind power generation prediction model is higher than that of the single model,and the prediction accuracy is increased by 5.3%compared with the optimal model.关键词
风力发电/发电量预测/ARIMA模型/PSO算法/BiLSTM算法Key words
wind power generation/power generation forecast/ARIMA model/PSO algorithm/BiLSTM algorithm分类
能源科技引用本文复制引用
顾凌龙,何非,龙苏岩,岳紫玉..基于历史发电及气象数据的双阶段风力发电量预测方法[J].可再生能源,2025,43(9):1213-1220,8.基金项目
国家电网公司科技项目(5108-202355437A-3-2-ZN). （5108-202355437A-3-2-ZN）
