基于轻量化网络的起重机运行机构监测及诊断  

Monitoring and diagnosis of crane operating mechanism based on lightweight network

作　　者：陈仕涛 杨恒[2] 王震 王全伟[2] Chen Shitao;Yang Heng;Wang Zhen;Wang Quanwei

机构地区：[1]广东省特种设备检测研究院东莞检测院,东莞523000 [2]太原科技大学,太原030024

出　　处：《起重运输机械》2023年第8期38-42,共5页Hoisting and Conveying Machinery

摘　　要：对于需要时刻关注起重机健康状态以及故障排查耗时费力的情况,单一监测目标已经不足以描述整体的运行状态。文中提出一种综合多传感器信息的轻量化网络的诊断方法,以起重机运行机构整体为研究对象,分析运行机构各环节频发故障特征,采取远程检测手段帮助掌握起重机运行机构各个环节的健康状况。通过一维卷积网络实现运行机构某环节故障特征提取;通过搭建可分离卷积网络结构,进行多路通道的特征提取和融合,应用于运行机构多路传感器检测框架下的故障分析。所提方法结构参数小、诊断响应快,可为解决类似机构的多传感器检测及诊断提供参考。For the situation that it is necessary to pay attention to the health status of cranes at all times and the troubleshooting is time-consuming and laborious,a single monitoring target is no longer enough to describe the overall running status.A diagnosis method of lightweight network based on multi-sensor information was proposed.Taking the whole crane operating mechanism as the research object,the characteristics of frequent faults in each link of the operating mechanism were analyzed,and the status of each link of the crane operating mechanism was monitored through remote detection.The fault characteristics of a link of the operating mechanism were extracted by one-dimensional convolution network;by building a separable convolution network structure,the feature extraction and fusion of multi-channel were carried out,and was used for fault analysis under the framework of multi-sensor detection of operating mechanism.This method has small structural parameters and fast diagnosis response,and can serve as a reference in the solving of the multisensor detection and diagnosis of similar mechanisms.

关 键 词：起重机 运行机构 一维卷积网络 故障监测 轻量化模型 

分 类 号：TH215[机械工程—机械制造及自动化]