摘要
准确的风速预测对风电扩大并网规模具有积极的推动作用。针对风速的波动性和随机性特征,提出了一种基于EMD、GPR和ISTA的短期风速预测模型。通过EMD对原始风速序列进行分解,利用GPR对分解后的序列子集进行一级预测,同时利用ISTA改进GPR的超参数优化选择过程;并将由此生成的误差序列带入到ISTA优化的GPR中进行二级预测,通过所得误差预测值对原始预测值进行校正并得到最终预测结果。案例分析表明,本文所提出的模型在短期风速预测中具有较高的预测精度。
Accurate wind speed prediction plays an active role in promoting the expansion of wind power grid-connected scale.In the light of the fluctuations and randomness of wind speed,the paper proposes a short-term wind speed prediction model based on empirical mode decomposition(EMD),Gaussian process regression(GPR)and improved state transition algorithm(ISTA).Firstly,the original wind speed sequence is decomposed by EMD,and the subsets of decomposed sequences are predicted by GPR for principal prediction,meanwhile ISTA is used to optimize the hyper-parameters selection of GPR.Then the resulting error sequence is brought into the optimized GPR for subordinate prediction,and the original prediction value is corrected by the obtained error prediction value to achieve the final prediction result.The case study shows that the proposed model has high prediction accuracy in the short-term wind speed prediction.
作者
黄元生
杨磊
高冲
刘诗剑
王光丽
HUANG Yuansheng;YANG Lei;GAO Chong;LIU Shijian;WANG Guangli(School of Economics and Management,North China Electric Power University,Beijing 102206,China;School of Economics and Management,North China Electric Power University,Baoding 071003,China;State Grid Jibei Electric Power Company Engineering Management Company,Beijing 100053,China)
出处
《智慧电力》
北大核心
2020年第1期35-41,共7页
Smart Power
基金
国家自然科学基金资助项目（61973117）~~
关键词
风速预测
集合经验模态分解
误差校正
高斯过程回归
改进状态转移算法
wind speed prediction
ensemble empirical mode decomposition
error correction
Gaussian process regression
improved state transition algorithm
