设备健康监测数据采集与分析模型构建  

Construction of equipment health monitoring data acquisition and analysis model

作　　者：王娜[1] WANG Na(School of Information Engineering.,Shaanxi Vocational and Technical College,Xi’an 710038,China)

机构地区：[1]陕西职业技术学院电子信息工程学院,西安710038

出　　处：《自动化与仪器仪表》2022年第8期42-45,49,共5页Automation & Instrumentation

基　　金：《陕西省教育科学"十三五"规划课题"校企融合"背景下基于"多元素融合"的计算机信息管理专业课程体系研究与实践》(SGH20Y1595)。

摘　　要：为提高滚动轴承等工业设备健康监测数据采集和分析的准确率,结合深度学习的特点,构建一个基于多重卷积神经网络融合的故障诊断模型,实现工业设备健康实时监测。首先,对卷积神经网络CNN和深度学习网络DNN的基本原理和网络结构进行具体分析,分别利用CNN和DNN多维度深层特征提取和挖掘的特点,提出多重CNN融合故障诊断算法;然后将此算法应用到工业设备监控监测模型中,对工业设备数据进行采集和处理分析。结果表明,对比于传统的CNN和DNN模型,融合后的模型在计算过程中的损耗较低,对工业设备故障诊断的准确率高达99.74%,可实现工业设备健康监测数据的有效采集和分析,模型性能优越。In order to improve the accuracy of health monitoring data collection and analysis of industrial equipment such as rolling bearings,combined with the characteristics of deep learning,a fault diagnosis model based on multiple convolutional neural network fusion is constructed to realize real-time health monitoring of industrial equipment.Firstly,the basic principle and network structure of convolutional neural network CNN and deep learning network DNN are analyzed,and the multi-dimensional deep features of CNN and DNN fault diagnosis algorithm is then applied to industrial equipment monitoring model to collect and process industrial equipment data.The results show that compared with the traditional CNN and DNN models,the loss of the integrated model is low,and the accuracy of industrial equipment fault diagnosis is 99.74%,which can realize the effective collection and analysis of industrial equipment health monitoring data,and the model performance is superior.

关 键 词：深度学习 卷积神经网络 工业设备 故障诊断 健康监测 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]