基于DFT灰色预测理论在日电量负荷预测中的应用

Application of DFT-based grey model in daily load forecast

提出了一种基于离散傅里叶变换(DFT)的灰色预测方法,首先将电量负荷序列采用离散傅里叶变换,分解出表征负荷规律性的各种频率分量,再对其进行组合重构,组合成周期T>7 d的低频分量和周期T≤7 d的剩余高频分量两部分,对低频分量采用灰色GM(1,1)模型预测,对高频分量单独进行处理,最后将两个处理结果结合起来,即得到最终结果。该方法从一定程度上屏蔽了随机因素对负荷预测造成的不利影响,提高了预测精度。计算结果表明了该方法的正确性和有效性。

A grey forecast method based on DFT（Discrete Fourier Transform）is presented. The load sequence is processed by DFT to decompose its frequency components,which are the token of load features. These frequency components are re-combined into low-frequency component （T〉7d） and high-frequency component（T≤7 d）. Forecast is carried out with grey forecast method for the low-frequency component and the high-frequency component is processed separately,the final result is the combination of both results. This method masks the influences of random factors on load forecast in a certain extent to improve the forecast precision. The calculative results show it correct and effective.