基于物联网和卷积神经网络的拖拉机轴承故障诊断

Fault Diagnosis of Tractor Bearing Based on Internet of Things and Convolutional Neural Network

摘要: 建立了物联网网络，对拖拉机变速箱轴承工作时运行的振动数据进行实时采集，并基于卷积神经网络算法，建立了拖拉机变速箱轴承故障诊断分析模型。实验结果表明：卷积神经网络整体训练过程比较稳定，收敛速度比较快，通过20次的迭代训练后，实际的验证精度在99.5%以上，验证了该模型的可行性。

Abstract: It firstly establishes the Internet of things network,which can collect the vibration data of tractor gearbox bearing in real time,and then establishes the fault diagnosis and analysis model of tractor gearbox bearing based on convolution neural network algorithm.Experiments show that the overall training process of convolutional neural network is relatively stable and the convergence speed is relatively fast.After 20 iterative training,its actual verification accuracy is more than 99.5%,which verifies the feasibility of the model.