基于负荷点相似的地区短期负荷预测新方法  

A New Method to Forecast Regional Short-Term Load Based on Similar Load Point

作　　者：罗军[1] 何光宇[1] 张思远[1] 万源[2] 李小锐[2] 

机构地区：[1]电力系统及发电设备控制和仿真国家重点实验室(清华大学电机系),北京市海淀区100084 [2]江西省电力公司,江西省南昌市330029

出　　处：《电网技术》2007年第6期67-71,共5页Power System Technology

摘　　要：提出了一种基于相似点的地区短期负荷预测新方法并形成了基于该算法的专家支持系统,该系统可定量地考虑气象信息、小水电、工作日类型等负荷相关因素。采用该系统对某基荷较小的地区电网进行负荷预测,结果说明了本文算法的正确性和有效性。基于本文算法而编制的软件系统即将投入实际应用。On the-basis of similar load point a new algorithm to forecast short-term loads in small area is proposed, besides an expert supporting system based on classification and regression tree （CART） and genetic algorithm （GA） is formed, in which the factors related to loads such as meteorological information, small hydropower plants, workday type and so on can quantitatively considered. The pilot calculation for a small area with lower base load is conducted and the calculation results show that the proposed method is correct and efficient. A software system based on the proposed method will be soon applied in a certain power system.

关 键 词：短期负荷预测 数据挖掘 决策树 遗传算法 

分 类 号：TM715[电气工程—电力系统及自动化]