摘要
针对新能源电站功率预测误差导致的考核成本问题,本文提出一种基于“两个细则”考核要求的风光储电站功率优化上报策略。通过分析预测误差分布特性,采用改进K-means算法和非参数经验分布,建立预测误差分箱模型,结合拉丁超立方抽样和Cholesky分解构建,考虑时间相关性的功率场景集。基于场景集优化,建立日前和日内上报模型,设计储能容量配置方法。实验结果表明,该策略可以显著提升预测准确率,风电场日前考核电量降低53.10%,光伏电站降低79.66%;储能配置进一步改善日内预测效果,准确率提升0.6%,考核电量减少1.5MWh,验证了该方法的有效性。
To address the assessment cost issues caused by power prediction errors in new energy power plants,this paper proposes a power optimization reporting strategy for wind-solar-storage power plants based on the assessment requirements of the'Two Detailed Rules.'By analyzing the distribution characteristics of prediction errors,a prediction error binning model is established using an improved K-means algorithm and non-parametric empirical distribution.Combined with Latin Hypercube Sampling and Cholesky decomposition,a power scenario set considering temporal correlations is constructed.Based on scenario set optimization,day-ahead and intraday reporting models are established,and a method for energy storage capacity configuration is designed.Experimental results show that this strategy can significantly improve prediction accuracy,reducing the day-ahead assessed power of wind farms by 53.10%and photovoltaic power plants by 79.66%.Energy storage configuration further improves intraday prediction performance,with accuracy increased by 0.6%and assessed power reduced by 1.5 MWh,validating the effectiveness of the method.
作者
魏巍
WEI Wei(Huadian Gansu Energy Co.,Ltd.,Lanzhou Gansu 730000)
出处
《中国科技纵横》
2025年第21期151-153,共3页
China Science & Technology Overview
关键词
新能源电站
功率预测偏差
两个细则
优化上报
储能容量配置
new energy power plants
power prediction deviation
two detailed rules
optimization reporting
energy storage capacity configuration
