【作者】 陈昊； 张钊； 陈强； 李翔；

【Author】 Chen Hao Zhang Zhao Chen Qiang Li Xiang Nanjing Power Supply Company,Jiangsu Nanjing 210002;2.Shanghai Electric power Company,Shanghai 200002;3.Electrical Engineering School of South East University,Jiangsu Nanjing;210096

【机构】 南京供电公司； 上海电力公司； 东南大学电气工程学院；

【摘要】 提出了一种基于非高斯分布广义自回归条件异方差模型的短期负荷预测新方法。在论证 ARCH 效应的存在性基础上,推广标准 GARCH 模型,建立了非高斯分布假设 GARCH 模型(GARCH-t,GARCH-GED)。最后比较了 ARMA,标准 GARCH,非高斯分布 GARCH 模型的预测能力,从平均预测误差,最大预测误差能力等指标上来看,GARCH-GED 模型表现最为出色。算例表明,非高斯分布 GARCH 模型是有效而可行的。更多还原

【Abstract】 In this paper,a new method of load forecast based on ARCH model with non-Gaussian distributions is proposed.By discussing the ARCH effect,the standard ARCH model is generalized.Improved ARCH models with non-Gaussian distributions,such as GARCH-t and GARCH-GED,are afforded as extense models.Finally,the forecasting performances of ARMA,standard GARCH,GARCH-t,GARCH-GED are compared.According to the index of mean error and maximum error,the GARCH-GED is superior to other models.The empirical example shows that the ARCH models with non-Gaussian distributions are efficient and feasible.更多还原