基于风速时空信息的BP神经网络超短期风速预测研究

摘要

基于风速历史数据统计法和基于地理信息与数值预报的物理方法都不能经济、有效、准确地对超短期风速做出预测。为了满足超短期风速预测的时效性和准确性，提出了基于风速历史数据和周边风速数据的风速时空信息BP神经网络超短期风速预测的思想，并研究了基于风速时空信息BP神经网络风速预测模型。建立基于MATLAB平台的BP神经网络预测程序，并实例验证了基于风速时空信息BP神经网络风速预测方法具有更高的精确度、时效性和经济性。%The ultra-short term wind speed forecast method based on historical wind speed statistics and geographical infor-mation physical and numerical prediction is not economical, effective and accurate. In order to meet requirements of time-liness and accuracy of the ultra-short-term wind forecasting, the paper proposes a new idea of BP neural network ultra-short term wind speed forecast based on historical and spatial wind speed data, and studies the BP neural network wind speed prediction model. The BP neural network prediction program is established on MATLAB platform,and it is verified through examples that this prediction method is of higher accuracy,timeliness and economy.