一种短期风电功率组合预测方法  

A Combined Short-term Wind Power Prediction Method

作　　者：曹权 李岩[1] 李双明 Cao Quan;Li Yan;Li Shuangming(College of Energy and Power Engineering,Nanjing University of Science and Technology,Nanjing Jiangsu 210094,China)

机构地区：[1]南京理工大学能源与动力工程学院,江苏南京210094

出　　处：《电气自动化》2020年第3期28-30,67,共4页Electrical Automation

摘　　要：为提高风电功率的短期预测精度,对组合预测进行了研究。选用差分自回归移动平均模型(ARIMA)与埃尔曼神经网络(Elman)模型,建立新的组合预测模型。首先,用单一的ARIMA预测模型和Elman预测模型对风电功率进行预测;然后,在单独模型预测的结果上,再次利用Elman神经网络进行预测;最后,将组合预测的结果与两单一模型的预测结果进行分析比较。结果表明,组合预测模型比各自单一预测模型有更高的预测精度。Combined forecasting was studied to improve the short-term wind power prediction accuracy.A new combined prediction model was established by using the autoregressive integrated moving average model(ARIMA)and the Elman neural network model.Firstly,wind power was predicted through a single ARIMA prediction model and Elman neural network prediction model.Secondly,on the basis of individual model prediction results,the Elman neural network was used again for prediction.Finally,the results of the combined prediction were compared with those of two individual models.The results showed that the combined prediction model had higher forecasting accuracy than the single prediction model.

关 键 词：风电功率短期预测 差分自回归移动平均 埃尔曼神经网络法 组合预测模型 预测精度 

分 类 号：TM614[电气工程—电力系统及自动化]