摘要
针对胶带机托辊传统故障诊断存在的故障特征频率提取困难、故障诊断误差较大、效率较低等问题,提出了一种基于数据驱动的胶带机托辊故障诊断方法。通过对胶带机托辊常见故障特征的分析,搭建了胶带机托辊故障诊断模拟实验模型,用以采集胶带机托辊故障集。采用交替卷积层和池化层构建了特征提取块,优化了托辊故障诊断算法,现场应用结果表明,采用此套诊断方法能实现从简单到精确的分级诊断,能够抵抗外部环境的影响,使托辊故障平均识别准确率达到了99.15%,可为后期带式输送机托辊故障的自主识别和胶带机关键部件的智能化故障诊断提供参考。
In view of the problems such as difficulty in extracting fault feature frequency,large fault diagnosis error and low efficiency due to the influence of external interference on the traditional fault diagnosis method of belt conveyor roller,a data-driven fault diagnosis method for roller of belt conveyor is presented in this paper.By analyzing the common fault characteristics of the tape machine roller,the experimental model for fault diagnosis of the tape machine roller is constructed to collect the tape machine roller faults.The feature extraction block is constructed by using alternating convolution layer and pooling layer,and the roller fault diagnosis algorithm is optimized.The field application results show that the hierarchical diagnosis from simple to accurate has been achieved after adopting the proposed diagnosis method;thus influence of the external environment can be resisted,and the average identification accuracy of the roller fault has reached 99.15%.It can provide a reference for the intelligent fault diagnosis of key components of the tape conveyor.
作者
寇金成
KOU Jin-cheng(Tunlan Coal Preparation Plant of Shanxi Coking Coal Xishan Coal Power Group Co.,Ltd,Taiyuan Shanxi 030206,China)
出处
《机械研究与应用》
2023年第3期136-139,共4页
Mechanical Research & Application
关键词
胶带机
托辊
故障诊断
特征提取
卷积神经网络
tape machine
roller
fault diagnosis
feature extraction
convolutional neural network
