基于CAR模型的用电量需求预测研究  

Research on Electricity Demand Forecast Based on CAR Model

作　　者：周小琳[1] Zhou Xiaolin(School of Economics and Management,Northeast Electric Power University,Jilin Jilin 132012)

机构地区：[1]东北电力大学经济管理学院,吉林吉林132012

出　　处：《东北电力大学学报》2018年第6期86-89,共4页Journal of Northeast Electric Power University

摘　　要：以日均用电量需求为输出变量,国内生产总值、第二产业产值比重、第三产业产值比重和城镇化率为输入变量,建立了基于多变量时间序列CAR的用电量需求预测模型.预测结果显示,所有年份用电量需求拟合值与观测值之间的相对误差均不超过1%,说明该模型在用电量需求预测领域具有较高的适用性.Taking the average daily electricity demand as output variable and the GDP,the proportion of thesecondary industry’s output value,the proportion of the tertiary industry’ s output value and the urbanizationrate as input variables,the electricity consumption demand prediction model based on the multi-variable timeseries was established. The predicted results show that the relative error between the fitting value and the ob-served value of electricity consumption demand is no more than 1% in all years,which indicates that this mod-el has high applicability in the forecasting field of electricity demand.

关 键 词：用电量需求 多变量时间序列 CAR模型 

分 类 号：TM715[电气工程—电力系统及自动化]