基于改进的D-S证据理论的中长期负荷预测方法  

The Medium and Long-Term Load Forecasting Based on Improved D-S Evidential Theory

作　　者：吴耀武[1] 娄素华[1] 卢斯煜[1] 乔惠 康福填 

机构地区：[1]华中科技大学强电磁工程与新技术国家重点实验室,武汉430074 [2]合肥供电局,合肥230000 [3]北京市供电局,北京100031

出　　处：《电工技术学报》2012年第8期157-162,共6页Transactions of China Electrotechnical Society

基　　金：国家重点基础研究发展计划(973项目)(2012CB215102);国家自然科学基金重点项目(50937002)资助

摘　　要：提出一种基于改进的证据理论的中长期负荷预测方法。该方法采用证据理论中Dempster合成法则和Yager组合规则,融合多种基础预测方法的预测结果,并将基于D-S证据理论的预测模型与专家经验进行无缝结合,使得改进后的预测方法更具实际意义。以某省级电力系统为例进行了负荷预测的分析和对比,结果证明了本文方法的有效性。A medium and long-term load forecasting model is proposed for power system,which is based on improved evidence theory.Evidential theory is able to integrate lots of information provided by different evidential source,which helps to make correct analysis and decisions.The combination rules of Dempster and Yager are used here to mix various basic forecasting methods.Expert advice is considered to be part of the model,so that the improved forecasting model is more meaningful practically,and the load forecasting accuracy has been improved apparently.Using this method to study one provincial grid in china,the results prove the validity of the method presented in this paper.

关 键 词：中长期负荷预测 证据理论 信度分配函数 专家经验 

分 类 号：TM715[电气工程—电力系统及自动化]