采用动态灰聚类算法的风电场动态等值方法  

Dynamic Equivalence of Wind Farm Based on Dynamic Gray Cluster Algorithm

作　　者：方瑞明[1] 吴敏玲[1] 王彦东[1] 尚荣艳[1] 彭长青[1] 

机构地区：[1]华侨大学信息科学与工程学院,福建厦门361021

出　　处：《华侨大学学报（自然科学版）》2017年第2期218-224,共7页Journal of Huaqiao University(Natural Science)

基　　金：国家自然科学基金资助项目(51577050);福建省厦门市重大科技创新平台资助项目(3502Z20111008);华侨大学研究生科研创新能力培育项目(1400201013)

摘　　要：针对风电机组运行状况间具有动态灰色关联性的特点,提出一种基于动态灰聚类算法的风电场动态等值方法.首先,根据实测运行数据对风电机组间的关联性进行分析,并确定数据样本的跨度选取时长.然后,采用动态灰关联分析,构造一个可以体现风电机组运行状况间动态灰色关联性的关联度矩阵G;进而以G中的样本组作为聚类指标进行K均值聚类,得出更合理的机群划分结果.最后,采用容量加权法计算机群等值参数,完成风电场的动态等值.仿真实验结果表明:所建立的动态等值模型与详细模型较接近,能够较准确地反映风电场的动态响应特性.The characteristics of dynamic gray correlation among the different operation conditions of wind turbines are considered, and then a method of dynamic equivalence for wind farm based on dynamic gray cluste-ring algorithm is proposed. Firstly, the correlation relationship among the wind turbines is analyzed by using the measured data, and the time span of data sample is determined. Secondly, a correlation matrix, named G, is constructed based on dynamic gray correlation analysis method, which can describe the characteristics of dy-namic gray correlation among the different conditions of wind turbines. Thirdly, the K-means cluster method is adopted to divide all wind turbines in the wind farm into several groups. Meanwhile？ the sample groups in G are used as clustering index. Finally, the equivalent parameters of each group are calculated by using the ca-pacity weighting method, and a dynamic equivalent model of wind farm is obtained. Simulation results carried on a real wind farm indicate that the obtained model can describe the dynamic response characteristics of the wind farm with the accuracy close to the detailed model.

关 键 词：动态灰关联分析 聚类算法 风电场 机群划分 动态等值 

分 类 号：TM715[电气工程—电力系统及自动化]