电网技术2012，Vol.36Issue(9)：221-225,5.
基于Adaboost的BP神经网络改进算法在短期风速预测中的应用
Application of Adaboost-Based BP Neural Network for Short-Term Wind Speed Forecast
摘要
Abstract
It is significant for economic dispatching of power grids containing large-scale wind farms to forecast wind speed more accurately. In allusion to the defect of insufficient accuracy in current short-term wind speed forecasting by neural network, auto-regressive moving-average (ARMA) time series analysis, Kalman filtering and so on, the Adaboost algorithm was led in to improve back propagation (BP) neural network algorithm, and an Adaboost-based BP neural network method was proposed and applied to short-term wind speed forecasting. Results of analyzing calculation example showed that using the proposed Adaboost-based BP neural network the accuracy of one or two hour-ahead wind speed forecasting was superior to respective forecasting accuracy by neural network and ARMA time series analysis, and the mean absolute percentage error of wind speed forecasting by the proposed algorithm was lower than 7.5% in high wind speed period (higher than10m/s). Thus the proposed method is applicable in engineering application.关键词
风速预测/Adaboost/BP神经网络Key words
wind speed forecast/ Adaboost/ BP neural network分类
信息技术与安全科学引用本文复制引用
吴俊利,张步涵,王魁..基于Adaboost的BP神经网络改进算法在短期风速预测中的应用[J].电网技术,2012,36(9):221-225,5.基金项目
国家重点基础研究发展计划项目(973项目)(2010CB227206) （973项目）
国家863高技术基金项目(2011AA05A101). （2011AA05A101）
