基于EMD和RBFNN的冷负荷组合预测模型

针对冷负荷预测问题，提出了一种基于相空间重构（PSR）、经验模态分解(EMD) 和径向基神经网络(RBFNN) 的 冷负荷组合预测模型。该模型首先利用经验模态分解方法，把冷负荷序列分解为少数模态分量，然后利用分组分量法将其分 为多个高频子分量、总低频分量和残余量，最后以PSR为基础对各分量利用RBFNN方法建模并将预测结果重构。该方法应 用于实际冷站负荷预测后，与单一RBFNN、SVM、LSSVM及基于EMD的SVM、基于EMD的RBFNN5类方法进行比较，结果 表明该方法对冷负荷预测精度有明显提高。

An Ensemble Prediction Model of Cooling Load Based on Empirical Mode Decomposition and Radial Basis Function Neural Network

In order to predict Cooling Load, an Empirical Mode Decomposition(EMD), Phase Space Reconstruction(PSR) based on Radial Basis Function Neural Network(RBFNN) ensemble learning paradigm is proposed. The original Cooling Load series are decomposed into a finite number of independent intrinsic mode functions with different frequencies, and then grouped by component method into various sub-components of the high-frequency, low-frequency component of the total, the remainder. Then different RBFNN models are used to model based on PSR, forecast the all sub-series, according to the intrinsic characteristic time scales. All fore casting results are combined to output the ultimate result. This model is applied to Cooling Load tendency forecasting. The results prove that the finally forecasting performance outperforms the RBFNN, SVM, LSSVM, SVM based on EMD and SVM based on EMD ahead forecasting.