中文摘要：

为了提高Elman网络的动态性能，有效地解决高阶系统的辨识问题，对Elman网络的结构及学习算法进行了改进，提出了一种新的Elman网络，建立了相应的神经网络载荷识别模型，并用于齿轮箱的载荷识别研究。试验结果表明，该网络模型具有收敛速度快、识别精度高的特点，为载荷识别研究提供了一种新的思路，具有一定的实用价值。

英文摘要：

To enhance the dynamic capability of the Elman network and to solve effectively the problem in identifying high order system, a new Elman network is developed by improving the network structure and learning arithmetic. Corresponding neural network load identification model was established, which was used to identify the load of a gearbox. The Experiment results showed that the improved neural network model has the characteristics of an extremely fast convergence and a relatively high accuracy. The investigation made in this paper provides a new approach for load identification research and is of certain practical usefulness.