摘要
光伏输出功率主要受太阳辐照度的影响,而天空中云的生成、运动以及消融会使太阳辐照度呈现随机性和波动性。地基云图可实时记录天空状况,因此获取地基云图的图像特征是光伏功率准确预测的关键步骤。对全天空成像仪采集的地基云图展开研究。首先,修复地基云图;然后,利用图像处理技术提取影响太阳辐照度变化的图像特征,包含光照强度、高频分量、透射率、天顶距离以及云因子特征;最后,将图像特征作为输入数据,光伏功率作为输出数据,利用梯度提升决策树算法构建光伏功率预测模型,实现光伏功率的预测。实验结果表明,采用从地基云图提取的图像特征构建的光伏功率预测模型,使得光伏功率预测的均方根误差可低于1%,为光伏功率的准确预测提供了一种技术手段。
The photovoltaic(PV)output power is mainly affected by solar irradiance,which will exhibit randomness and volatility due to the generation,movement and ablation of clouds in the sky.Considering that ground cloud images can record the sky conditions in real time,acquiring the features of ground cloud images is a key step for the accurate prediction of PV power.In this paper,the ground cloud image collected by a total sky imager is studied.First,the ground cloud image is repaired.Then,image processing techniques are used to extract the image features that affect the changes in solar irradiance,including light intensity,high-frequency components,transmittance,zenith distance,and the cloud factor.Finally,a gradient boosting decision tree(GBDT)algorithm is used to build a prediction model of PV power with the image features as input data and PV power as output data,thus achieving the prediction of PV power.Experimental results show that the PV power prediction model constructed based on the image features extracted from the ground cloud image can make the root mean squared error of PV power prediction less than 1%,providing a technical means for the accurate prediction of PV power.
作者
路志英
周庆霞
李鑫
王泽涵
LU Zhiying;ZHOU Qingxia;LI Xin;WANG Zehan(Key Laboratory of Smart Grid of Ministry of Education,Tianjin University,Tianjin 300072,China)
出处
《电力系统及其自动化学报》
CSCD
北大核心
2020年第8期70-76,共7页
Proceedings of the CSU-EPSA
基金
国家自然科学基金资助项目(51677123)。
关键词
太阳辐照度
光伏功率
地基云图
特征提取
预测模型
solar irradiance
photovoltaic power
ground cloud image
feature extraction
prediction model
