基于灰色关联投影法的短期负荷预测相似日选择算法

为了提高短期负荷预测的精度,合理有效地选取预测相似日是非常重要的,提出了基于灰色关联投影法的相似日选取算法。该算法首先针对影响负荷变化因素的复杂性和不确定性,以灰色理论计算出了相应的权重,接着,结合投影法计算历史日相对于待预测日的灰色投影值,比较投影值,可以得到与待预测日最接近的相似日。理论与实例均表明,该算法具有通用性,对任意包含多个因素的历史日都能够快速准确地算出其与待预测日的相似程度,具有很好的相似日选择效果。

A Novel Algorithm of Selecting Similar Days for Short-term Load Forecasting Based on Grey Relation Projection Method

To improve accuracy of the short-term load forecasting,the rationality and validity of selecting similar days is very important,thus,a novel algorithm of selecting similar days based on grey relation projection method is proposed.Firstly,this proposed algorithm,facing to complexity and uncertainty of the influencing factors to load change,applies the grey theory to calculate corresponding weights,then,the grey projection value of historical data are obtained,compared with the values,the most similar historical data with the forecasted day will be gained.Finally,the simulation results demonstrate that this algorithm is quite generic and effective,for historical data including any more factors,the similarity degree can be quickly and accurately obtained.