对神经网络在故障诊断中的应用进行了探讨，针对潜油电泵机组故障的多态性

和多层次性，结合振动数据的轴向频域特征，建立了基于高斯-牛顿法的 BP 神经网络，

通过对其训练和学习，实现了三种常见故障的有效诊断。-The neural network fault diagnosis in the paper discusses the application, esp for fault the polymorphism of unit

And more layered, combined with the data of the axial vibration frequency characteristics, based on gauss-Newton method of the BP neural network,

Through their training and learning, to realize the three common fault diagnosis of effective.