【摘要】 针对滚动轴承故障的在线监测与智能诊断问题,建立了一种基于深度神经网络(DNN)的嵌入式轴承故障智能诊断系统,采用轻便的可移动硬件设备完成滚动轴承的信号采集监测,可结合多种智能神经网络算法训练故障模型,快速诊断滚动轴承故障类型。采用凯斯西储大学数据集进行仿真实验,DNN模型平均测试准确率能够达到97.7%;对某型号直升机主减速器轴承进行测试实验,实验表明该智能诊断系统可以实现四种基础功能,完整采集滚动轴承信号并且进行故障诊断及结果显示,DNN模型平均测试准确率能够达到96.7%,在轴承智能检测方面取得了满意的效果。更多还原

【Abstract】 Aiming at the problem of online monitoring and intelligent diagnosis of rolling bearing faults,an embedded bearing fault intelligent diagnosis system based on a deep neural network is proposed in this paper.Lightweight and movable hardware equipment are used to complete the signal acquisition and monitoring of rolling bearing,and it can be combined with a variety of intelligent neural network algorithms to train fault models and quickly diagnose the type of rolling bearing faults.This paper uses the Case Western Reserve University data sets for algorithm verification,and the average test accuracy can reach 97.7%.更多还原