中文摘要：

为解决综合负荷模型的随机时变性和地域分散性，提出了一种基于日负荷曲线的变电站用电行业负荷构成比例在线修正方法。将每年分为夏大、夏小、冬大、冬小和一般5个负荷水平期，每个时期选取典型日负荷曲线并定义负荷率、最小负荷率、峰谷差率、最大负荷出现时间等4个特征参数，应用模糊C均值聚类算法对变电站分类，并获得各聚类中心特征向量，运用模式识别原理，建立聚类中心变电站特征向量和由负控点典型日负荷曲线聚类得到的行业特征矩阵之间的典型隶属关系，并构建其和统计综合法获得的变电站行业负荷构成比例之间的典型映射关系。由变电站实测日负荷曲线获得当日的隶属关系，并与典型映射结合实现对变电站当日用电行业负荷构成比例的在线实时修正。

英文摘要：

To cope with the random time variation and regional divergence of synthetic load model, a daily load curve-based online method to modify substation’s structural proportion of synthetic load for power consuming-industries is proposed. According to different load levels, each year is divided into five load level periods, i.e., peak-load period in summer, valley-load period in summer, peak-load period in winter, valley-load period in winter and common-load period; typical daily load curve for each period is chosen and in each period four characteristic parameters, i.e., the load rate, maximum load rate, peak-valley difference rate and the occurrence time of maximum load are defined; by use of fuzzy C means clustering algorithm, the substations are classified and characteristic vectors of various types are obtained. Utilizing pattern recognition, the typical membership relation between characteristic vectors of various types of substation cluster center and industrial characteristic matrix obtained by clustering typical daily load curve of load control center is built and the typical mapping relation between the membership relation and substation’s structural proportion of synthetic load for power consuming-industries obtained by statistical synthesis is constructed. Thus, through combining the membership relation obtained by measured daily load curve of substation with typical mapping, online real-time modification of substation’s intraday structural proportion of synthetic load for power consuming-industries can be realized.