基于粗糙特征量的短期电力负荷预测  

Short-term Load Forecasting Based on Rough Characteristic-component Algorithm

作　　者：马立新[1] 郑晓栋 尹晶晶[1] 

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《电子科技》2016年第1期40-43,共4页Electronic Science and Technology

基　　金：国家自然科学基金资助项目(6120576);国家科技部政府间科技合作基金资助项目(2009014)

摘　　要：针对负荷特征一直是实际电力负荷预测中的重大问题。提出了基于粗糙特征量的约简算法。通过对天气及负荷历史数据进行挖掘,找到负荷的关键特征,并与径向基网络结合建立了负荷预测模型。算例结果表明,与按经验选取输入的传统网络相比,预测准确度有了明显的提高,更适用于电力负荷预测。The key characteristic of mining influence the load is always an important problem in power load forecasting. A reduction algorithm through rough characteristic-component algorithm is introduced. The key characteristics of the date of weather and history load data are discussed,and then a model combined with radical basis function neural network is established. Forecasting results of calculation examples show that the forecasting accuracy is obviously improved and more suitable for short-term load forecasting compared with traditional radical basis function neural network model that chooses input parameters in the light of experience.

关 键 词：电力系统 径向基 粗糙特征量 负荷预测 

分 类 号：TM715[电气工程—电力系统及自动化]