结合误差校正的北京市中长期电力需求预测  

Medium-and-long-term Forecasting of Beijing Electric Demand Based on Error Adjustment

作　　者：牛彦涛[1] 王辉[2] 

机构地区：[1]华北电力大学能源与环境研究院,北京102206 [2]华北电力大学经济与管理学院,北京102206

出　　处：《现代电力》2011年第1期90-94,共5页Modern Electric Power

摘　　要：中长期电力负荷预测是电力规划中一项战略性工作。为明晰揭示电力需求与经济社会发展的密切关系,基于GDP因素分解提出以人口、城市化、经济发展及产业结构建立电力需求非线性预测模型。由于未来一年的电力需求需要特别关注,为消除回归预测可能产生的较大误差,引入GARCH模型进行误差校正。将该方法应用于北京市中长期电力需求预测,并建立相关自变量预测模型;将所得预测结果、弹性系数法预测结果及北京市规划值进行比较,初始预测值与校正后的预测值进行比较,验证该方法具有较高可信性。该方法可为其他相关研究提供借鉴。Medium-and-long-term forecasting of electric load is the essential to power systems planning.Based on the factor decomposition of GDP,a nonlinear forecasting model of electricity demand,which can describe the relationship between electricity demand and economical social development,is built based on population,urbanization,economical development and industrial structure adjustment.Due to the special concern on electricity demand in the next year,the GARCH model is adopted to reduce errors caused by regression forecasting technique,which shows its effectiveness on forecasting medium-and-long-term electric demand in Beijing.Then the forecast model of relative self-variable is built.By comparison of initial forecasting value with that of with corrected forecasting value,the proposed method is verified to have high accuracy.The proposed approach could also be extended to other research area related to energy systems planning.

关 键 词：电力需求预测 误差修正 回归模型 分解研究 

分 类 号：TM715[电气工程—电力系统及自动化]