基于改进K-均值的面向风光海的配电网典型运行场景生成
|
摘要
首先对含风光海的电力系统进行建模,得到了可再生能源出力及负荷的时序曲线。其次,利用密度峰值聚类对K-均值聚类进行改进,避免了传统K-均值聚类初始聚类中心及聚类个数无法确定的问题。最后,基于改进的K-均值聚类算法对风力发电、光伏发电、潮汐发电及时序负荷进行同步聚类,在不同时间尺度下,生成了电力系统典型运行场景。结果验证了该算法的可行性,并且与传统K-均值聚类相比,改进后的算法聚类有效性和收敛性更强。
Firstly,this paper models the power system with wind power,solar energy and tidal power,and obtains the time series curve of renewable energy output and load. Secondly,the density peak clustering is used to improve K-means clustering,which avoids the problem that initial clustering centers and their number of the original K-means cannot be determined. Lastly,based on the improved K-means clustering algorithm,wind power generation,photovoltaic power generation,tidal power generation and time series load are synchronously clustered. At different time scales,typical operation scenarios of power system are generated. The results verify the feasibility of the improved algorithm,and it is more convergent than the traditional K-means clustering.
Firstly,this paper models the power system with wind power,solar energy and tidal power,and obtains the time series curve of renewable energy output and load. Secondly,the density peak clustering is used to improve K-means clustering,which avoids the problem that initial clustering centers and their number of the original K-means cannot be determined. Lastly,based on the improved K-means clustering algorithm,wind power generation,photovoltaic power generation,tidal power generation and time series load are synchronously clustered. At different time scales,typical operation scenarios of power system are generated. The results verify the feasibility of the improved algorithm,and it is more convergent than the traditional K-means clustering.
引文
[1]董凯,江辉,黄泽荣,等.运用成本效益分析的风/柴/储能系统规划方法[J].电力系统及其自动化学报,2010,22(3):67-72.
[2]刘文霞,蒋程,张建华,等.一种用于序贯蒙特卡罗仿真的风电机组多状态可靠性模型[J].电力系统保护与控制,2013,41(8):73-80.
[3]王苏,周健.考虑尾流效应的风电场可靠性模型研究[J].黑龙江电力,2015,37(3):211-215.
[4]袁晓玲,施俊华,徐杰彦.计及天气类型指数的光伏发电短期出力预测[J].中国电机工程学报,2013,33(34):57-64.
[5]刘明君.含潮汐流能和储能的发电系统概率建模及其可靠性评估[D].重庆:重庆大学,2016.
[6]刘明君,李文沅,王财胜.孤岛模式运行下含潮汐发电和电池储能的微电网可靠性评估[J].电力自动化设备,2016,36(11):33-39.
[7]陈凡,刘海涛,黄正,等.基于改进k-均值聚类的负荷概率模型[J].电力系统保护与控制,2013,41(22):128-133.
[8]王韶,周鑫.应用层次聚类法和蚁群算法的配电网无功优化[J].电网技术,2011,35(8):161-167.
[9]李敏,陈果,沈大千,等.基于改进凝聚层次聚类算法的变压器绕组及铁心故障诊断研究[J].高压电器,2018,54(1):236-242.
[10]程启明,张强,程尹曼,等.基于密度峰值层次聚类的短期光伏功率预测模型[J].高电压技术,2017,43(4):1214-1222.
[11]吴琴,钟庆,王钢,等.基于模糊C均值聚类的风电场多机等值方法[J].现代电力,2016,33(6):27-32.
[12]解蛟龙.风/光/负荷典型场景缩减方法及在电网规划中的应用[D].合肥:合肥工业大学,2017.
[13]丁明,解蛟龙,潘浩,等.微电网经济运行中的典型时序场景分析方法[J].电力自动化设备,2017,37(4):11-16.
[14]吴丽珍,蒋力波,郝晓弘.基于最优场景生成算法的主动配电网无功优化[J].电力系统保护与控制,2017,45(15):152-159.
[15]王群,董文略,杨莉.基于Wasserstein距离和改进K-medoids聚类的风电/光伏经典场景集生成算法[J].中国电机工程学报,2015,35(11):2654-2661.
[16]张宏宇,印永华,申洪,等.基于序贯蒙特卡洛方法的风电并网系统调峰裕度评估[J].电力系统自动化,2012,36(1):32-37.
[17]陈凡,卫志农,黄正,等.考虑风电并网的发电充裕度评估方法的比较[J].电力自动化设备,2014,34(6):30-35.
[18]汪新园.大规模风电接入电力系统的可靠性评估及储能策略研究[D].杭州:浙江大学,2013.
[19]蒋程,刘文霞,张建华,等.含电池储能风电场的发电系统风险评估[J].太阳能学报,2014,35(2):207-213.
[20]王海明.风电场稳态建模及应用研究[D].太原:太原理工大学,2015.
[21]王群.风/光经典场景集的生成方法及应用[D].杭州:浙江大学,2015.
[22]李朝晖,尹晓博,杨海晶,等.基于改进的k-means聚类算法的季节性负荷特性分析[J].电网与清洁能源,2018,34(2):53-59.
[23]王杨.基于时序蒙特卡洛模拟的微电网可靠性分析[D].重庆:重庆大学,2014.
[24] RODRIGUEZ A,LAIO A. Clustering by fast search and find of density peaks[J]. Science,2014,344(6191):1492-1496.
[25] National Renewable Energy Laboratory(NREL)[EB/OL]. https://www. nrel. gov/gis/data. html.
[26] National Oceanic and Atmospheric administration.Tidal and currents[EB/OL]. https://tidesandcurrents. noaa. gov/cdata/Station List? type=Current%20Data&filter=survey.
[27]周开乐,杨善林,丁帅,等.聚类有效性研究综述[J].系统工程理论与实践,2014,34(9):2417-2431.
[2]刘文霞,蒋程,张建华,等.一种用于序贯蒙特卡罗仿真的风电机组多状态可靠性模型[J].电力系统保护与控制,2013,41(8):73-80.
[3]王苏,周健.考虑尾流效应的风电场可靠性模型研究[J].黑龙江电力,2015,37(3):211-215.
[4]袁晓玲,施俊华,徐杰彦.计及天气类型指数的光伏发电短期出力预测[J].中国电机工程学报,2013,33(34):57-64.
[5]刘明君.含潮汐流能和储能的发电系统概率建模及其可靠性评估[D].重庆:重庆大学,2016.
[6]刘明君,李文沅,王财胜.孤岛模式运行下含潮汐发电和电池储能的微电网可靠性评估[J].电力自动化设备,2016,36(11):33-39.
[7]陈凡,刘海涛,黄正,等.基于改进k-均值聚类的负荷概率模型[J].电力系统保护与控制,2013,41(22):128-133.
[8]王韶,周鑫.应用层次聚类法和蚁群算法的配电网无功优化[J].电网技术,2011,35(8):161-167.
[9]李敏,陈果,沈大千,等.基于改进凝聚层次聚类算法的变压器绕组及铁心故障诊断研究[J].高压电器,2018,54(1):236-242.
[10]程启明,张强,程尹曼,等.基于密度峰值层次聚类的短期光伏功率预测模型[J].高电压技术,2017,43(4):1214-1222.
[11]吴琴,钟庆,王钢,等.基于模糊C均值聚类的风电场多机等值方法[J].现代电力,2016,33(6):27-32.
[12]解蛟龙.风/光/负荷典型场景缩减方法及在电网规划中的应用[D].合肥:合肥工业大学,2017.
[13]丁明,解蛟龙,潘浩,等.微电网经济运行中的典型时序场景分析方法[J].电力自动化设备,2017,37(4):11-16.
[14]吴丽珍,蒋力波,郝晓弘.基于最优场景生成算法的主动配电网无功优化[J].电力系统保护与控制,2017,45(15):152-159.
[15]王群,董文略,杨莉.基于Wasserstein距离和改进K-medoids聚类的风电/光伏经典场景集生成算法[J].中国电机工程学报,2015,35(11):2654-2661.
[16]张宏宇,印永华,申洪,等.基于序贯蒙特卡洛方法的风电并网系统调峰裕度评估[J].电力系统自动化,2012,36(1):32-37.
[17]陈凡,卫志农,黄正,等.考虑风电并网的发电充裕度评估方法的比较[J].电力自动化设备,2014,34(6):30-35.
[18]汪新园.大规模风电接入电力系统的可靠性评估及储能策略研究[D].杭州:浙江大学,2013.
[19]蒋程,刘文霞,张建华,等.含电池储能风电场的发电系统风险评估[J].太阳能学报,2014,35(2):207-213.
[20]王海明.风电场稳态建模及应用研究[D].太原:太原理工大学,2015.
[21]王群.风/光经典场景集的生成方法及应用[D].杭州:浙江大学,2015.
[22]李朝晖,尹晓博,杨海晶,等.基于改进的k-means聚类算法的季节性负荷特性分析[J].电网与清洁能源,2018,34(2):53-59.
[23]王杨.基于时序蒙特卡洛模拟的微电网可靠性分析[D].重庆:重庆大学,2014.
[24] RODRIGUEZ A,LAIO A. Clustering by fast search and find of density peaks[J]. Science,2014,344(6191):1492-1496.
[25] National Renewable Energy Laboratory(NREL)[EB/OL]. https://www. nrel. gov/gis/data. html.
[26] National Oceanic and Atmospheric administration.Tidal and currents[EB/OL]. https://tidesandcurrents. noaa. gov/cdata/Station List? type=Current%20Data&filter=survey.
[27]周开乐,杨善林,丁帅,等.聚类有效性研究综述[J].系统工程理论与实践,2014,34(9):2417-2431.