文章介绍的Elman神经网络的短期负荷预测模型,具有模型简单、运算效率高的特点,并具有较好的全局最优性能,从而很好地克服了传统BP算法容易陷入局部极小点的缺陷。文章还比较了Elman网络和BP网络结构的建模效果。通过仿真实验证明了Elman神经网络具有动态特性好、逼近速度快、精度高等特点,是一种实用、可靠的负荷预测方法。 The short-term load forecasting model of Elman neural network introduced in this paper has the characteristics of simple model and high computational efficiency, and has better global optimal performance, so it overcomes the defect that the traditional BP algorithm is easy to fall into local minimum. The article also compares the modeling effects of Elman network and BP network structure. The simulation results show that Elman neural network is a practical and reliable load forecasting method with good dynamic characteristics, fast approaching speed and high precision.