Elman神经网络在农网短期负荷预测中的应用

Application of Elman Neural Network to Short-Term Load Forecasting in Rural Power System

为提高负荷预测精度,更好地反映负荷的动态特性,提出了一种基于Elman神经网络的预测方法。先对负荷样本进行数据预处理,消除伪数据,然后把不同日各时刻的负荷序列作为样本,对未来时刻的负荷进行短期预测。预测实例表明,基于Elman神经网络的预测方法比基于BP神经网络的预测方法具有更好的预测效果。

To enhance the load forecasting precision and reflect the dynamic characteristics of the short-term load changes of rural power system,the method of load forecasting based on Elman neural network was proposed.Firstly,the load samples were filtered and processed generally.Then,the load series of each time of different days were chosen as the training samples.And finally,the short-term load was forecasted by historical data after eliminating the pseudo-data.As the results of an example of factual forecasting show,the effect of method of load forecasting based on Elman neural network better than the method of load forecasting based on BP neural network.