摘要
不同气象情况会影响太阳辐射强度,使光伏发电输出功率波动,为实时掌握光伏发电功率,有必要在考虑气象因素的基础上构建光伏发电功率预测模型。利用灰色关联度方法对历史气象数据进行气象类型量化处理后筛选样本数据;建立了光伏AI发电功率预测模型,将筛选后的气象样本数据作为模型输入并进行训练,输出训练后的光伏发电功率预测结果。实际应用结果表明:该技术在预测光伏发电功率时不存在过拟合情况,筛选气象数据样本时的关联度数值最高为0.90,可有效预测一天内不同时刻的光伏发电功率,且误差值较小。
Different meteorological conditions could affect solar radiation intensity,causing fluctuations in the output power of photovoltaic power generation.In order to get the real time generating power of the photovoltaic power generation,it was necessary to build a photovoltaic power prediction model on the basis of considering meteorological factors.After quantifying the meteorological types of historical meteorological data,the grey correlation method was used to screen the sample data.The photovoltaic AI generation power prediction model was established,taking the screened meteorological sample data as the model input and training it to output the predicted photovoltaic power generation results.The practical application results showed that there was no over fitting in the prediction of photovoltaic generation power,and the correlation value when screening meteorological data samples was up to 0.90.It could effectively predict photovoltaic generation power at different times of the day,with very little error.
作者
杨焱
刘蔚
王永刚
Yang Yan;Liu Wei;Wang Yonggang(Guizhou Qianyuan Electric Power Co.,Ltd.,Guiyang Guizhou 550001,China;China Huadian Engineering Co.,Ltd.,Beijing 100160,China;Guizhou Power Grid Co.,Ltd.,Guiyang Guizhou 550000,China)
出处
《煤化工》
CAS
2023年第2期107-110,115,共5页
Coal Chemical Industry
关键词
光伏发电
气象因素
发电功率
功率预测
气象类型
photovoltaic power generation
meteorological factor
generating power
power prediction
meteorological type
