一种用于短期电价预测的分时段时间序列传递函数模型  

Electricity price forecasting based on transfer function models for period-decoupled time series

作　　者：陈友[1] 王晛[1] 李渝曾[1] 

机构地区：[1]上海大学自动化系,上海200072

出　　处：《电力系统保护与控制》2008年第16期1-4,102,共5页Power System Protection and Control

摘　　要：在电力市场环境下,准确的价格预测可为市场参与者制定合理的竞争策略提供重要信息。本文提出了一种基于时间序列法的分时段传递函数模型来预测短期电价,该模型考虑了负荷因素对电价的影响,同时利用累积式自回归滑动平均模型(ARIMA)对电价序列和负荷序列的非平稳性进行处理,并且对一天24个小时时段分别建立了预测模型。采用加州历史电价数据进行算例研究,结果表明,利用本文模型进行电价预测能够提高预测的准确性。In electricity markets, accurate price forecasting provides crucial information for market participant to make reasonable competing strategies so as to maximize their benefits. This paper presents a price forecasting model based on Transfer Function Models for period-decoupled time series. The effect of the load on the electricity price can be fully taken into account. The ARIMA model is employed to deal with the nonstationary price and load series, and forecasting models for every hour period are developed independently. The numerical example based on the California market data shows that the proposed model could improve the accuracy of forecasting.

关 键 词：电力市场 电价预测 时间序列法 传递函数模型 

分 类 号：TM744[电气工程—电力系统及自动化]