两种风电功率多步预测方式的分析及评价  

The Analysis and Evaluation of Two Ways for Multi-step Wind Power Prediction

作　　者：严干贵[1] 王东[1] 杨茂[1] 熊昊[1] 宋薇[1] 

机构地区：[1]东北电力大学电气工程学院,吉林吉林132012

出　　处：《东北电力大学学报》2013年第1期126-130,共5页Journal of Northeast Electric Power University

基　　金：国家自然科学基金(60934005;51177009;50877009);吉林省产业技术研究与开发专项项目(JF2012C018);国家科技支撑计划课题三(2011BAA07B03)

摘　　要：目前对风电功率预测的研究主要集中在预测方法,而风电功率预测一般为多步预测,缺乏对多步预测方式的分析。分析了滚动多步预测方式和多采样尺度多步预测方式。多步预测方式的预测精度与预测方法和风电功率数据有关,不同的预测方法或风电功率数据下两种多步预测方式的预测精度高低不同。以风电功率实时预测为例,基于吉林省某风电场实测风电功率数据,使用国家能源局所规定的风电功率实时预测评价指标对两种多步预测方式进行评价。Wind power prediction is significance for the operation of wind power grid integration. The study of wind power prediction mainly forcus on the prediction method, but Wind power prediction is awalys multi-step prediction, lacking analysis of the way for multi-step prediction. This paper take analysis on rolling and multi- sampling scale multi-step wind power prediction. The precision of the way of multi-step wind power prediction relate to the prediction method and the wind power data. The precision level between the two ways is different with different prediction method or wind power data. Take ultra-short term wind power forecast as example, the evaluation of two ways for multi-step wind power prediction is based on the evaluating indicator of national en- er~＇~ administration about wind oower orediction, usin~ the real data from a wind farm of .lilin Province.

关 键 词：风电功率预测 多步预测方式 滚动 多采样尺度 

分 类 号：TM733[电气工程—电力系统及自动化]