标么值加权平均组合变量生成法用于中长期电力负荷预测  

Per-unit value weighted average combined variable generated method for mid-long term power load forecasting

作　　者：张筱慧[1] 

机构地区：[1]中国农业大学信息与电气工程学院,北京100083

出　　处：《电力系统保护与控制》2009年第12期40-43,共4页Power System Protection and Control

基　　金：国家科技支撑计划子课题"农村电网高可靠性供电及供用电节电关键技术研究"资助项目(2006BAJ04B06-001)

摘　　要：将标么值概念引入电力负荷预测,将预测中使用的部分自变量做标准化处理,可以消除变量取值单位对回归系数的影响,同时寻找变量间及变量在不同地区间的相似性。利用变量的加权线性组合生成新变量,利用相似性对一些变量的发展规律作出估计,使偏最小二乘回归方法能够应用于我国农村电网。The concept of per unit value is introduced into power load forecasting. Some independent variables used in forecasting is standardized to eliminate the effect of different unit variables to regress coefficients. At the same time, the comparability between different variables and variables in different areas are studied. New variables are created using the weighted linear combination of original variables, and the variable progress rule is estimated based on the comparability, and partial least-square regression can be used in power network of our country.

关 键 词：中长期电力负荷预测 组合变量 偏最小二乘回归 

分 类 号：TM715[电气工程—电力系统及自动化]