带有偏差单元的递归神经网络诊断方法在齿轮箱故障诊断中的应用

提出了带有偏差单元的递归神经网络方法，并将其应用到齿轮箱的故障诊断中．通过对齿轮箱的基本零件齿轮故障的设置，研究TBP神经网络的改进算法与带有偏单元的递归神经网络用于齿轮箱故障诊断的有效性。研究表明，带有偏单元的递归神经网络方法比BP网络训练速度快，精度要高，节省时间，能够精确实现故障定位。

Bias Unit with a Recurrent Neural Network Diagnosis Method in Gearbox Fault Diagnosis

In this paper,with a bias unit recurrent neural network method and its application to gearbox fault diagnosis. Through the basic components of the gear box gear fault set,study the improved algorithm of BP neural network with the unit with a partial recurrent neural network for fault diagnosis of the effectiveness of the gear box. Studies have shown that cells with a partial recursive neural network than the BP network training speed and accuracy must be high,save time,be able to accurately realize the faul...