中文摘要：

为达到国家电网要求的风电场短期预测（日前）误差不超过25%的技术规定,文章以15 min时间间隔的风电实测和预测功率历史数据为基础,从横向误差和纵向误差的角度,提出了一种利用横向一次修正和储能系统二次修正的有功分级平滑来提高风电功率短期预测精度的方法。通过对误差时间序列进行统计分析,得出所需的横向时间向量将0-24 h预测功率前后平移,实现初级修正;运用储能系统装置,制定储能系统能量管理策略,有效地补偿风电功率的纵向误差,实现二级修正。以风电场的典型历史数据为例,经过两级有功分级修正,对风电功率预测进行较精确评估,同时做了简易的经济评价。

英文摘要：

In order to meet the technical regulations of State Grid that short-term wind power prediction error（day-ahead）does not exceed 25% of the technical provisions, this article is based on measurement and prediction wind power historical data in the time scale 15 minutes, from the perspective of the transverse and longitudinal error, put forword a kind of methods to improve the wind power short-term forecast accuracy by Active grading smooth through first transverse correction and Energy storage secondary correction. Through the statistical analysis of error time series, concluded the need horizontal time vector, and then translation 0-24 hours predictive power back and forth, achieve primary revision. design energy storage system energy management strategy to meet effective vertical wind power compensation error, implementation secondary revision. Taking the typical wind farm historical data as an example, after two stages of active classification correction, realization of wind power prediction for a more accurate assessment, at the same time, do a simple economic evaluation.