基于Elman神经网络的电力负荷预测模型研究

Research of Power Load Forecasting Based on Elman Neural Network

为提高电力负荷预测精度,文章采用Elman神经网络建立模型,提出了一种基于Elman神经网络的电力负荷预测方法,采用自适应学习速率动量梯度下降反向传播算法进行网络训练,对乌鲁木齐电网的实际历史数据进行了仿真,仿真结果表明,Elman神经网络对电力负荷进行预测具有收敛速度快,预测精度高的优点。

In order to improve the precision of forecasting of power load, in this paper a Elman artifical neural network （ANN） approach for load forecasting is proposed and the model based on Elman neural network. In the training algorithm of the network, a back-propagation algorithm with adaptive learning speed and momentum gradient-falling is used, the forecasting model tested by actual data from Urmqi electric network, simulation results indicate that the forecasting for power load based on Elman neural network features quick convergence speed and high forecasting precision.