基于加权双高斯分布的广义自回归条件异方差边际电价预测模型.pdfVIP

第34 卷 第1 期 电 网 技 术 Vol. 34 No. 1 2010 年1 月 Power System Technology Jan. 2010 文章编号：1000-3673 （2010 ）01-0139-06 中图分类号：F 416.2 文献标志码：A 学科代码：790·625 基于加权双高斯分布的 广义自回归条件异方差边际电价预测模型 刘西陲，沈炯，李益国 （东南大学 能源与环境学院，江苏省 南京市 210096 ） A Generalized Auto-Regressive Conditional Heteroskedasticity Model for System Marginal Price Forecasting Based on Weighted Double Gaussian Distribution LIU Xi-chui, SHEN Jiong, LI Yi-guo (School of Energy Environment, Southeast University, Nanjing 210096, Jiangsu Province, China) ABSTRACT: The regularity of conditional variance variation 的基础，是电力市场研究的核心问题。电价研究的一 and the statistical distribution characteristic of residual error of 个重要内容是系统边际电价(system marginal price， system marginal price (SMP) in electricity market are SMP)的预测。由于电价变化受诸多非确定性因素(如 researched. On this basis, by means of leading in the 供需变化、燃料价格波动、网络阻塞、市场力等)影 generalized auto-regressive conditional heteroskedasticity 响，这些非确定性因素很难检测和量化，因此，建立 (GARCH) model a GARCH model based on the assumption of weighted double Gaussian (WDG) distribution is proposed to 有效的电价预测模型进行准确预测比较困难。 research the variation regularity of SMP. Taking the actual data 一般说来，电价预测的方法主要分为2 类[1-7]： from PJM market in USA and NSW market in Australia as 仿真法和时间序列法。仿真法需要建立详细而严格的 samples, both GARCH model and the proposed GARCH-WDG 模型，建模需要大量的私有信息，一般很难获得，所 model are tested, the testing results show that the GARCH 以仿真法在市场运行和监管领域应用较多；应用领域 model can offer good estimation and forecasting results of SMP,