基于时段聚类的峰谷电价时段确定方法研究  

Periods division method for peak-valley prices based on clustering algorithm

作　　者：赵明 董力[2] 李孟阳 高赐威[2] 喻洁[2] 

机构地区：[1]云南省电力试验研究院(集团)有限公司科学研究院,昆明650000 [2]东南大学电气工程学院,南京210096

出　　处：《电力需求侧管理》2016年第4期11-17,共7页Power Demand Side Management

基　　金：云南省电网公司科技项目(项目名称:基于用户能效特性的电网需响应关键技术研究);江苏省青蓝工程资助;国家自然科学基金项目(51577029);国家高技术研究发展计划(国家863计划)项目(2015AA050401)

摘　　要：合理的峰谷电价设置能够极大地改善负荷特性。在考虑用户需求响应的基础上,建立了峰谷时段动态调整模型框架,由时段划分、时段调整和效果评估三个模块组成。预处理日负荷数据后,通过排序法和K-means聚类法划分峰谷平时段,采用剔除或介绍友时段的方法处理孤立时段,比较时段调整前后设定指标的变化情况评判时段调整效果,发现Kmeans聚类法更能反映季节对用户用电行为产生的影响。某地区实例分析表明,当时段-电价曲线与峰谷时段切合度较高时,可以削峰填谷,提高机组利用率,有利于激励更多的需求响应资源;反之,则会使得负荷特性指标恶化。Reasonable periods division of peak-valley price can greatly improve load characteristics. Based on the customers＇ demand response, this dynamic adjustment model framework of peakvalley periods is established, which is composed by three modules of period division, adjustment and assessment. After preprocessing daily load data, all the periods are divided into three categories named as peak, normal and valley, with the use of sorting method and K-means clustering method. Considering the actual operating conditions, the method removes or introduces the friend periods to the isolated periods. After discussing the index changes to judge the implementation effect, the phenomenon shows that K- means clustering method can better reflect the impact of seasonal variation on the consumers＇ electricity consumption behavior. At last,this model is applied on the real data of some regions. The results demonstrate that when the period-price curve highly matches with the peak-valley periods, this model can effectively shift peak load,improve the efficiency of generator set and stimulate demand response through the adjustment of time period partition scheme. On the contrary, the load characteristic index will be worsened.

关 键 词：需求响应 峰谷电价 时段调整 排序分配 K-means聚类分配 

分 类 号：TM731[电气工程—电力系统及自动化]