摘要
提出了一种基于神经网络的光储动态优化运行策略。首先,采用长短期记忆神经网络的方法预测光伏出力;其次,根据功率差调节和光伏出力预测结果,将以日为时长的负荷分为4个区间,不同区间分别对应特定控制策略;最后,根据各区间的控制策略实时优化储能充放电功率。该方法把多余的光伏能量储存起来,在用户电网需要时再以电能的形式释放出来,既能有效地抑制光伏发电的功率波动,提高光伏利用率,还能对用户电网进行削峰填谷调节。
A dynamic optimization operation method based on neural network was proposed.First,the method of long and short memory neural network is used to predict the PV output,and then the load of one day is divided into four parts according to the power difference adjustment and PV output,and each part corresponds to a control strategy.Finally,the real-time optimization is carried out according to each control strategy.This method is of great significance for smoothing the output fluctuation of PV system and improving PV utilization rate,and also has certain effect of peak shaving and valley filling.
作者
方陈
修晓青
周健
陈学良
王皓靖
黄华炜
FANG Chen;XIU Xiaoqing;ZHOU Jian;CHEN Xueliang;WANG Haojing;HUANG Huawei(Electric Power Research Institute,State Grid Shanghai Municipal Electric Power Company,Shanghai 200437,China;China Electric Power Research Institute,Beijing 100192,China)
出处
《电器与能效管理技术》
2019年第1期69-74,81,共7页
Electrical & Energy Management Technology
基金
国网上海市电力公司科技项目(520940160027)
关键词
光伏预测
储能系统
神经网络
动态优化
photovoltaic prediction
energy storage
neural network
dynamic optimization
