基于EEMD-样本熵和Elman神经网络的短期电力负荷预测

Short-term Power Load Forecasting Based on Ensemble Empirical Mode Decomposition-sample Entropy and Elman Neural Network

针对电力负荷序列的非线性、非平稳性等特点,提出了一种基于集总经验模式分解EEMD-样本熵和El-man神经网络的短期负荷预测方法。为了减小电力负荷序列局部分析的计算规模以及提高负荷预测的精度,先利用EEMD-样本熵将原始电力负荷序列分解成一系列复杂度差异明显的子序列;然后在综合考虑温度及日期类型等因素对各子序列影响的基础上,根据各子序列的特点构造不同的Elman神经网络对各子序列分别进行预测;最后将各子序列的预测结果叠加得到最终预测值,并对EUNITE国际电力负荷预测竞赛公布的数据进行仿真实验。仿真结果表明该方法能有效地提高负荷预测的精度。

According to the nonlinearity and non-stationarity of power load series, a short-term power load forecasting approach based on ensemble empirical mode decomposition （EEMD）-sample entropy and Elman neural network is proposed in the paper. In order to reduce the calculation scale of partial analysis for power load and improve the accuracy of load forecasting. Firstly, the power load time series is decomposed into a series of power load subsequences with obvi- ous differences in complex degree by using EEMD-sample entropy. Then, on the basis of considering the influence on each subsequences of temperature and day type, different models to forecast each component separately are constructed using Elman neural network according to the features of subsequences. Finally, these forecasting results of subsequenc- es are combined to obtain final forecasting result. The simulation example of the European network on intelligent tech- nologies for smart adaptive system （EUNITE）power load prediction competition verifies that the combined prediction model can effectively improve the power load forecasting precision.