基于LMBP神经网络的农用车传动系齿轮故障诊断研究

The Research of Agricultural Vehicle Transmission Gear Fault Diagnosis Based on LMBP Neural Network

针对农用车传动系齿轮故障多,不适合拆卸检查等实际情况,提出用LMBP神经网络来进行该系统的故障诊断。介绍了LMBP神经网络的结构、算法。通过对传动系振动信号特点的分析,归纳出峰值、有效值等作为输入特征向量的典型网络训练和故障诊断样本。应用Matlab人工神经网络工具箱进行了网络的构建、训练和仿真。分析结果表明,LMBP神经网络训练速度快,能较准确地诊断出农用车传动系齿轮的各种故障。

LMBP neural network was proposed to diagnose the agricultural vehicle transmission gear failure,for there were too many defaults in the system and not suitable for disassembly.The structure and algorithm of LMBP neural network was introduced.By analyzing the characteristics of vibration signal reducer,the typical network training and fault diagnosis samples were summarized.The Matlab artificial neural network toolbox was applied for the network building,training and simulation.From the analysis result,LMBP neural network training speed was fast,and the failures of agricultural vehicle transmission gears can be accurately diagnosed.