应用递归人工神经网络预测电力短期负荷

Using Recurrent Artificial Neural Net to Predict Short-term Load

介绍了神经网络的基本原理,使用递归人工神经网络模型对电力短期负荷进行预测,采用了梯度下降法,来提高训练的收敛速度,预测仿真结果表明,使用递归神经网络预测比传统的预测方法更准确.

This paper introduced the principles of neural networks and used Recurrent Artificial Neural Network（RANN） to forecast short term load. In order to improve the training speed, grads descension was adopted. Simulation result shows that the RANN is better than the traditional method.