一种电力负荷预测混合模型研究
    点此下载全文
引用本文:田珂1?覮,丁博2,马文栋1,赵卫华2,王坤2.一种电力负荷预测混合模型研究[J].计算技术与自动化,2020,(4):39-44
摘要点击次数: 317
全文下载次数: 0
作者单位
田珂1?覮,丁博2,马文栋1,赵卫华2,王坤2 (1.国网河南省电力公司河南 郑州450000 2.国网河南省电力公司 电力科学研究院客户服务中心河南 郑州450000) 
中文摘要:为了提高短期负荷预测(STLF)的精度问题,采用了新的信号分解和相关分析技术,结合改进的经验模态分解法(IEMD)将负荷需求时间序列分解为若干个规则的低频分量。为了补偿信号分解过程中的信息损失,通过使用T-Copula进行相关分析来合并外部变量的影响。通过T-Copula分析,可从风险值(VaR)得出峰值负荷指示二进制变量,以提峰值时间负荷预测的准确性。将IEMD和T-Copula得到的数据应用于深度置信网络(DBN)来预测特定时间的未来负荷需求。
中文关键词:短期负荷预测  经验模态分解  T-Copula  峰值负荷  风险值  深度置信网络
 
A Hybrid Model for Power Load Forecasting
Abstract:In order to improve the accuracy of short-term load forecasting (STLF),this paper uses new signal decomposition and correlation analysis technology,combined with improved empirical mode decomposition (iemd) to decompose the load demand time series into several regular low-frequency components. In order to compensate for the information loss during signal decomposition,T-Copula is used for correlation analysis to merge the effects of external variables. Through T-Copula analysis,the binary variable indicating peak load can be obtained from the value of risk (VaR) to improve the accuracy of peak time load forecasting. The data from IEMD and T-Copula are applied to deep confidence network (DBN) to predict future load demand at a specific time.
keywords:short-term load forecasting  empirical mode decomposition  T-Copula  peak load  VaR  DBN
查看全文   查看/发表评论   下载pdf阅读器