考虑预测误差的风电场储能配置优化方法  

Research on Battery Storage Sizing for Wind Farm Considering Forecast Error

作　　者：兑潇玮[1] 朱桂萍[1] 刘艳章[2] 

机构地区：[1]电力系统国家重点实验室(清华大学电机系),北京市海淀区100084 [2]中国电力科学研究院,江苏省南京市210003

出　　处：《电网技术》2017年第2期434-439,共6页Power System Technology

基　　金：国家863高技术基金项目(2012AA050207);国家重点研发计划项目(2016YFB0900105)~~

摘　　要：利用储能降低风电预测误差能够降低风电出力不确定性,从而减少弃风,储能的容量和功率配置则直接影响到补偿预测误差的效果和经济性。基于T location-scale(TLS)分布建立了风电预测误差概率分布模型,使用几何分布描述风电误差持续性,并利用粒子群算法对模型中的参数进行了极大似然估计。以风电场收益最大为优化目标,将补偿风电预测误差的置信度以机会约束的形式加入储能配置优化模型,并在模型中考虑了对风电预测出力曲线的修正。仿真计算结果表明,所提方法能够很好地拟合预测误差的概率密度函数和误差持续时间,储能配置的经济性优于不考虑误差持续时间和出力修正的模型,验证了所提方法能够提高储能配置的准确性和经济性。Using battery storage can mitigate wind farm output uncertainty and decrease wind farm forecast error, contributing to decreasing wind curtailment. Installation capacity and power of battery storage has great influence on effect of decreasing forecast error and economy. This paper builds probability distribution model of wind farm forecast error based on T location-scale （TLS） distribution and simulates forecast error persistence with geometric distribution. Then model parameters are obtained with maximum likehood estimation in particle swarm optimization. Wind farm profit maximization is taken as optimization objective and confidence of decreasing forecast error as chance constraint in the optimization model. Correction of wind farm forecast output is included in the model. Results show that the proposed method fits probability density function and duration of wind farm forecast error correctly. Total profit of wind farm and battery storage is better than that not considering forecast error duration and forecast output correction. It is verified that the method can improve accuracy and economy of battery storage sizing.

关 键 词：储能配置 预测误差 机会约束 随机模拟 

分 类 号：TM715[电气工程—电力系统及自动化]