综合权重的模糊时间序列的电力负荷预测方法  

Weighted Fuzzy Time-Series Method for Power Load Forecasting

作　　者：刘晓娟[1,2] 方建安[1] 

机构地区：[1]东华大学信息与科学技术学院,上海200051 [2]上海电力学院数理学院,上海200090

出　　处：《华东电力》2012年第4期518-520,共3页East China Electric Power

基　　金：国家自然科学基金项目(60874113)~~

摘　　要：电力负荷预测受诸多因素的影响,针对短期电力负荷的复杂性和不确定性,结合历史负荷数据,提出了一种综合权重的模糊时间序列预测方法。该方法首先对历史负荷数据进行预处理;然后利用模糊集和模糊时间序列的方法将历史负荷数据模糊化,考虑到负荷变化的趋势,借助于最优化理论给出了趋势权重,同时考虑到近期数据影响大于远期数据,给出了时间占优权重,从而得到了综合权重的模糊时间序列预测方法。最后的数值实验结果表明,该方法比传统的模糊时间序列方法具有更高的预测精度。Power load forecasting can be influenced by various factors.So taking into account the complication and uncertainty of the short-term load prediction,and combined with the historical load data,this paper proposes a comprehensively weighted fuzzy time-series method for short-term load forecasting.First,this method was applied to preprocess the historical load data;then,the fuzzy sets and the fuzzy time series were used to make the historical data fuzzy.In terms of the trend of the load change,the trend weight was obtained based on the optimization theory;meanwhile,given that the recent data has greater influence than the past data,the chronologically-dominant weight is proposed,and hence the comprehensively weighted fuzzy time-series forecasting method is formulated.Finally experimental results show the proposed method has higher forecasting accuracy than the traditional fuzzy time-series method.

关 键 词：预处理 电力负荷预测 模糊时间序列 综合权重 

分 类 号：TM715[电气工程—电力系统及自动化] TP391[自动化与计算机技术—计算机应用技术]