基于差分分解的有源配电网网供负荷预测方法

Load forecasting method for active distribution network based ondifferential decomposition

有源配电网在网供负荷预测过程中,容易受到天气突变影响,导致负荷预测结果不精准。为此,提出了基于差分分解的有源配电网网供负荷预测方法。采用差分分解方法,将负荷序列分解成负荷保持序列和负荷差分序列,使预测过程聚焦于序列的变化。根据时间序列间的相似度计算结果,评估预测负荷序列与实际负荷序列的相似程度,构建误差补偿模型,控制网供负荷预测。实验结果表明,在正常和连续降雨天气下,该方法负荷预测精准度最大值分别为95%、92%,说明使用该方法能够得到精准的预测结果。

Active distribution network are easy affected by sudden change of weather in the process of power supply and load forecasting,resulting in inaccurate load forecasting results.Therefore,an active distribution network load forecasting method based on difference decomposition is proposed.The load series is decomposed into load holding series and load difference series by using the difference decomposition method,so that the forecasting process can focus on the changes of the series.According to the similarity calculation results between time series,the similarity between the predicted load series and the actual load series is evaluated.The error compensation model is built to forecast the load supplied by the control network.The experimental results show that the maximum load forecasting accuracy of this method is 95%and 92%respectively under normal and continuous rainfall weather,indicating that this method can obtain accurate forecasting results.