基于Monte-Carlo方法的风电功率预测不确定性估计  

Uncertainty Evaluation of Wind Power Prediction Based on Monte-Carlo Method

作　　者：王勃[1] 刘纯[1] 张俊[2] 冯双磊[1] 李颖毅[2] 郭锋[2] 

机构地区：[1]中国电力科学研究院,北京100192 [2]国网浙江省电力公司,杭州310007

出　　处：《高电压技术》2015年第10期3385-3391,共7页High Voltage Engineering

基　　金：国家自然科学基金(51477156)~~

摘　　要：风电功率预测的不确定性估计是实现对风电场优化调度的基础。为此,提出了1种定性分析风电功率预测敏感气象因子的方法,采用2种方式分别对敏感气象因子的误差分布函数进行了估计,并采用Monte-Carlo随机抽样实现了对风电功率预测不确定性的估计,最后进行了算例验证。研究结果表明:不同预报时刻对应气象参数预报误差的不确定性差别较大;基于气象参数不确定性随机抽样的方法可实现对功率预测的不确定性估计,且在相同的置信度下,逐小时样本估计优于总体样本估计。与传统风电功率预测不确定性估计方法相比,该方法对风电场历史运行数据要求低,且不需要单独建模,具有较高的工程实用性。The uncertainty estimation of wind power prediction is the basis of wind farm optimal dispatching. Firstly, we proposed a qualitative analysis method of sensitive meteorological factors of wind power prediction, and quantitatively evaluated the prediction error distribution of sensitive factors using non-parametric regression and Gaussian fitting re- spectively. Then, we adopted Monte-Carlo random sampling to realize the uncertainty estimation of the wind farm power prediction. Finally, we verified the validity of the proposed algorithm by case study. The results show that there are sig- nificant differences between uncertainties of the forecast error corresponding to different forecast time, the uncertainty estimation of predicted power can be achieved through the Monte-Carlo method, and hourly sample estimation is superior to the overall sample estimation at the same confidence level. Compared with traditional uncertainty estimation methods of wind power forecasting, the proposed method has a low requirement for operating data, and does not need additional model.

关 键 词：风电场 功率预测 误差分布 不确定性估计 数值天气预报 Monte-Carlo 

分 类 号：TM614[电气工程—电力系统及自动化]