Performance evaluation and accuracy enhancement of a day-ahead wind power forecasting system in China

详细信息    查看全文

• 作者：Pan Zhao^a ; ^{pzhao0204@gmail.com} ; Jiangfeng Wang^a ; Junrong Xia^a ; Yiping Dai^a ; ^{ypdai@mail.xjtu.edu.cn} ; Yingxin Sheng^b ; Jie Yue^b
• 关键词：Artificial neural networks ; Kalman filter ; Numerical weather prediction ; Wind power forecasting
• 刊名：Renewable Energy
• 出版年：2012
• 出版时间：July 2012
• 年：2012
• 卷：43
• 期：Complete
• 页码：234-241
• 全文大小：843 K

文摘

Wind power forecasting system is useful to increase the wind energy penetration level. Latest statistics show that China has been the biggest wind energy market throughout the world. However, few studies have been published to introduce the wind energy forecasting technologies in China. This paper presents the performance evaluation and accuracy enhancement of a novel day-ahead wind power forecasting system in China. This system consists of a numerical weather prediction (NWP) model and artificial neural networks (ANNs). The NWP model is established by coupling the Global Forecasting system (GFS) with the Weather Research and Forecasting (WRF) system together to predict meteorological parameters. In addition, Kalman filter has been integrated in this system to reduce the systematic errors in wind speed from WRF and enhance the forecasting accuracy. The numerical results from a real world case are proven the effectiveness of this forecasting system in terms of the raw wind speed correction and wind power forecasting accuracy. The Normalized Root Mean Square Error (NRMSE) has a month average value of 16.47 % , which is an acceptable error margin for allowing the use of the forecasted values in electric market operations. This forecasting system is profitable for increasing the wind energy penetration level in China.