基于小波变换和时间序列的风功率超短期预测模型研究  

Application time series based on wavelet transform on wind power ultra-short term prediction

作　　者：苏展 张志霞[1] 朴在林[1] 孙卓 赵丽华 王强 Su Zhan Zhang Zhixia Piao Zailin Sun Zhuo Zhao Lihua Wang Qiang(College of Information and Electrical Engineering, Shenyang Agriculture University, Shenyang 110866, China State Grid Corporation of Liaoyang, Liaoyang 111000, China State Grid Corporation of Chaoyang, Chaoyang 122000, China State Grid Corporation of Jinzhou, Jinzhou 121000, China)

机构地区：[1]沈阳农业大学信息与电气工程学院,辽宁沈阳110866 [2]国网辽阳供电公司,辽宁辽阳111000 [3]国网朝阳供电公司,辽宁朝阳122000 [4]国网锦州供电公司,辽宁锦州121000

出　　处：《可再生能源》2017年第9期1381-1386,共6页Renewable Energy Resources

基　　金：国家"十二五"科技支撑项目(2012BAJ26B00)

摘　　要：为提高风功率超短期预测模型的精确度,利用小波变换将原始风功率时间序列进行分解和重构,得到相应的高频序列和低频序列。对不同序列建立相应的自回归移动平均模型,并且进行拉格朗日乘子检验,验证是否具有拉格朗日乘子效应,从而建立相应的自回归条件异方差模型或广义自回归条件异方差模型,将所得的预测结果进行线性叠加组合得出最终结果。通过算例分析及与其他几种预测模型预测结果的对比,结果表明小波变换和时间序列结合的风功率超短期预测模型可以有效提高风功率超短期预测精度。In order to improve the accuracy of the ultra short term forecasting model of wind power,original wind power time series are decomposed and reconstructed by wavelet transform to get corresponding high frequency sequences and low-frequency sequences.The corresponding autoregressive moving average model is established according to different sequences,as well as the Lagrangian multiplier test is mainly used to verify whether there is a Lagrangian multiplier effect,so as to establish the corresponding autoregressive conditional heteroskedasticity model or generalized autoregressive conditional heteroscedasticity model.The final forecasting results are obtained through the linear superposition and combination.Through the case study and the comparison with the forecasting results of several other forecasting models,the consequences show that the combination model of wavelet transform and time series can effectively improve the accuracy of the ultra short term forecasting model of wind power.

关 键 词：小波变换 自回归移动平均模型 时间序列 风功率预测 超短期 

分 类 号：TK81[动力工程及工程热物理—流体机械及工程] TM711[电气工程—电力系统及自动化]