摘要
新能源储能系统具有能量的时空平移性,光伏发电已成为未来发电的必然趋势,但是新能源储能自身的波动性使得光伏发电预测不精准。为此,提出了新能源储能波动下的光伏发电预测改进算法。采用储能技术平抑新能源储能波动,缓解波动给光伏发电预测带来的影响。构建光伏发电预测模型,通过光伏发电系统改进结构参数,优化连接权值和阈值。采用升压转换器,选取合适开关控制信号,获取光伏发电最大功率点追踪结果。分析光伏系统实时发电量与时序关联性,采用改进分数阶算法,有效地跟踪光伏发电状态变化,计算光伏发电预测量。由实验结果可知,该算法与实际值存在最大为0.2 kW的误差,其余时间均一致,具有精准预测效果。
The new energy storage system has the time⁃space translation of energy,and photovoltaic power generation has become the inevitable trend of future power generation.However,due to the volatility of new energy storage itself,photovoltaic power generation prediction is not accurate.Therefore,an improved algorithm for forecasting photovoltaic power generation under the fluctuation of new energy storage is proposed.Energy storage technology is adopted to stabilize the fluctuation of new energy storage and mitigate the impact of fluctuation on photovoltaic power generation forecast.Build a photovoltaic power generation prediction model,improve the structural parameters through the photovoltaic power generation system,and optimize the connection weights and thresholds.The boost converter is used to select the appropriate switch control signal to obtain the maximum power point tracking results of photovoltaic power generation.Analyze the correlation between real-time generation and time sequence of photovoltaic system,adopt improved fractional order algorithm,effectively track the change of photovoltaic power generation state,and calculate the photovoltaic power generation pre measurement.It can be seen from the experimental results that the maximum error between the algorithm and the actual value is 0.2 kW,and the rest of the time is consistent,with accurate prediction effect.
作者
陆沈雄
范亚伟
屠永辉
戚成健
温世超
LU Shenxiong;FAN Yawei;TU Yonghui;QI Chengjian;WEN Shichao(Zhejiang Huayun Information Technology Co.,Ltd.,Hangzhou 310008,China)
出处
《电子设计工程》
2024年第3期152-156,共5页
Electronic Design Engineering
关键词
新能源储能波动
光伏发电
预测算法改进
功率点追踪
fluctuation of new energy storage
photovoltaic power generation
improvement of prediction algorithm
power point tracking
