多时间尺度下光伏功率波动特性概率分布研究  

Probability Density Function of Photovoltaic Power Variations Considering Different Time Scales

作　　者：李芬 李春阳 闫全全[2] 徐静 孙丽兵 杨兴武 LI Fen;LI Chun-yang;YAN Quan-quan;XU Jing;SUN Li-bing;YANG Xing-wu(College of Electrical Engineering,Shanghai University of Electric Power,Shanghai 200090,China;Maintenance Company of SMEPC,Shanghai 200063,China;Hubei Provincial Meteorological Service Center, Wuhan 430205,China;Shanghai Solar Energy Science & Technology Co.,Ltd.,Shanghai 201108,China)

机构地区：[1]上海电力学院电气工程学院,上海200090 [2]上海市电力公司检修公司,上海200063 [3]湖北省气象服务中心,湖北武汉430205 [4]上海太阳能科技有限公司,上海201108

出　　处：《水电能源科学》2018年第12期215-218,共4页Water Resources and Power

基　　金：上海市高校教师培养资助计划(CXYsdl18012);上海市自然科学基金项目(16ZR1413000);国家自然科学基金青年项目(51307105)

摘　　要：基于武汉地区分布式光伏电站大量实测数据,运用广义高斯分布和有限学生t混合模型等多种概率模型对不同时间尺度下光伏功率波动特性建模,发现在10～15min时间尺度下广义高斯分布最适用于描述分布式光伏功率变化的概率分布,而在30～60min时间尺度下高斯混合模型拟合效果最好。在此基础上,建立了逐时光伏出力波动与辐射量波动模型,用于定量分析光伏电站能量输出波动,可有效降低光伏功率波动随机性和不确定性对电力系统运行造成的影响,有利于提高光伏并网渗透率。Based on a large amount of measured data of distributed PV station in Wuhan Region,the generalized Gaussian distribution,finite student t-mixture model and other models were used to analyze the PV power fluctuation characteristics in multiple time scales.The study reveals that the generalized Gaussian distribution is suitable for describing the probability distribution of distributed PV power variation in 10 to 15 minutes time scales,and Gaussian mixture model is best for 30 to 60 minutes time scales.And then the relationship between the variability of PV energy production and global radiation was established to quantitatively analyze the energy production variability of PV station,which can effectively reduce the impact of randomness and uncertainty of PV power fluctuation on the operation of power system, and is beneficial to improve the penetration level of PV power.

关 键 词：光伏功率波动特性 概率密度函数 广义高斯分布 有限学生t混合模型 高斯混合模型 

分 类 号：TM615[电气工程—电力系统及自动化]