基于相位相关的天空图像云团运动速度计算方法研究  

CLOUD VELOCITY CALCULATION METHOD OF SKY IMAGES BASED ON FOURIER PHASE CORRELATION THEORY

作　　者：王飞[1,2] 甄钊[1] 米增强[1] 孙宏斌[2] 刘纯[3] 王勃[3] 卢静[3] 

机构地区：[1]新能源电力系统国家重点实验室(华北电力大学),保定071003 [2]电力系统及发电设备控制和仿真国家重点实验室(清华大学电机系),北京100084 [3]中国电力科学研究院,北京100192

出　　处：《太阳能学报》2016年第10期2539-2546,共8页Acta Energiae Solaris Sinica

基　　金：国家自然科学基金(51577067);北京市自然科学基金(3162033);河北省自然科学基金(E2015502060);中央高校基本科研业务费专项资金资助重点项目(2014ZD29)

摘　　要：光伏电站天空图像中云团运动速度的快速准确计算是进行光伏功率分钟级超短期预测的前提基础。针对目前采用图像分割匹配计算方法耗时较长的缺点,基于分钟级时间尺度下相邻天空图像中云团形变较小的条件,根据图像相位差的周期分布特性,提出一种基于傅里叶相位相关理论的云团分钟级运动速度计算方法,并定义针对其计算结果准确性的评估指标。实际算例对比结果表明,在相同计算精度前提下,该方法较现有方法可大幅减少运算时间,且鲁棒性与抗干扰能力更强。The rapid and accurate calculation of cloud velocity is the prerequisite for PV power forecast at minute time scale using sky images. Aiming at long time consuming process of existing image segmentation and matching method, as the cloud shape change is smaller between two adjacent sky images at minute time scale, and according to the periodicity of the phase difference between two sky images, the cloud velocity calculation method was proposed based on Fourier phase correlation theory. An evaluation index to measure the accuracy of this method was also presented. The results show that the proposed calculation method can significantly reduce the computing time and maintain same calculation accuracy, and has greater robustness and anti-jamming capability.

关 键 词：天空图像 云团运动速度 相位相关 光伏功率预测 

分 类 号：TM615[电气工程—电力系统及自动化]