经验模态分析综合法在负荷预测中的应用

Application of Synthetic Method Based on Empirical Mode Decomposition to Load Forecasting

针对电力负荷的周期相似性,提出一种基于经验模态分析法的综合负荷预测方法。先对原始数据进行统计,使用EMD将统计的时间序列分解为有限个固有模态分量,对固有模态分量使用模糊C均值聚类,再采用ARMA将聚类后的固有模态分量预测,最后把每个分量预测值求和得到负荷预测值。实例仿真计算表明,该算法比直接使用ARMA模型进行预测具有更高的预测精度,是一种有效短时预测方法。

According to the similarity of power load, an integrated load forecasting method based on empirical mode decomposition （EMD） is proposed. Firstly, an artificial statistical is done for the raw data, and the statistical time series is decomposed into different intrinsic modes by EMD, then the intrinsic mode components are clustered by fuzzy clustering. Then, these different clustered components are predicted by aoturegressive moving average （ARMA） model. Finally, the forecasted load is obtained by adding together the predicted values of each component. The experiment simulations show that the proposed algorithm has a higher forecasting accuracy than the direct use of ARMA model, which is an effective short - term forecasting method.