带式输送机托辊故障诊断及协同管控研究综述  

Review on idler fault diagnosis and coordinated control in belt conveyors

作　　者：邢震 田野新 包建军 齐智峰 周李兵 叶柏松 张蓉 XING Zhen;TIAN Yexin;BAO Jianjun;QI Zhifeng;ZHOU Libing;YE Baisong;ZHANG Rong(CCTEG Changzhou Research Institute,Changzhou 213015,China;State Key Laboratory of Intelligent Coal Mining and Strata Control,Changzhou 213015,China;Tiandi(Changzhou)Automation Co.,Ltd.,Changzhou 213015,China;College of Computing,Georgia Institute of Technology,Bellevue 98005,America)

机构地区：[1]中煤科工集团常州研究院有限公司,江苏常州213015 [2]煤炭智能开采与岩层控制全国重点实验室,江苏常州213015 [3]天地(常州)自动化股份有限公司,江苏常州213015 [4]乔治亚理工学院计算机科学学院,华盛顿州贝尔维尤98005

出　　处：《工矿自动化》2025年第3期39-53,共15页Journal Of Mine Automation

基　　金：江苏省科技成果转化专项资金项目(BA2022040);天地(常州)自动化股份有限公司科研项目(2024TY0001);中国煤炭科工集团有限公司科技创新创业资金专项项目(2021-2-GH004)。

摘　　要：托辊作为带式输送机的关键部件,其故障频发严重影响煤矿生产效率与安全。目前国内外在托辊故障诊断技术和带式输送机管理控制策略方面开展了广泛研究,然而尚未形成一套被广泛认可且行之有效的监测与管控手段。通过分析托辊故障的类型及机理,指出井下带式输送机托辊故障诊断的特殊性及面临的挑战。梳理了托辊故障诊断技术及托辊故障后协同管控的研究现状:在故障状态感知技术方面,探讨了振动、声音、温度及图像信号感知技术的原理与应用;在数据处理及特征提取方面,探讨了各类信号的处理方法及特征提取策略;在故障识别方法方面,探讨了从传统方法到机器学习、深度学习及多源信息融合的托辊故障识别方法的技术演进过程;在托辊故障后协同管控方面,探讨了目前面临控制系统复杂性高、不同控制策略之间的兼容性差、状态监测数据的准确性和实时性难以保证等问题。基于上述研究,指出托辊故障诊断技术虽取得一定进展,但仍存在故障辨识度不高、覆盖范围有限、检测物理量单一、无法对故障进行分类及判断程度、未能评估故障可能引发的风险,以及缺乏全面的管控策略等问题,提出托辊故障诊断技术发展方向:从单一监测向多维度融合监测发展、从稀疏覆盖向密集全面覆盖迈进、从表象诊断向本质分析探究故障演化规律、从被动应对到主动预防的转变并推动从局部管控向全局协同管控的升级。As a critical component of belt conveyors,idlers are prone to frequent failures,significantly impacting the efficiency and safety of coal mine operations.Extensive research has been conducted worldwide on idler fault diagnosis techniques and coordinated control strategies for belt conveyors.However,a universally accepted and effective monitoring and control framework is still lacking.This paper provides a comprehensive review of idler fault types and failure mechanisms,emphasizing the unique challenges associated with diagnosing faults in underground belt conveyors.The current state of research on idler fault diagnosis and post-failure coordinated control is systematically analyzed in four key areas:①Fault State Perception Technologies:The principles and applications of vibration,acoustic,thermal,and image-based sensing technologies are discussed.②Data Processing and Feature Extraction:Various signal processing methods and feature extraction strategies are examined.③Fault Identification Methods:The evolution of idler fault identification techniques is reviewed,ranging from traditional approaches to advanced machine learning,deep learning,and multi-source information fusion.④Post-Failure Coordinated Control:Challenges such as the high complexity of control systems,poor compatibility between different control strategies,and difficulties in ensuring the accuracy and real-time performance of condition monitoring data are highlighted.Despite notable advancements in idler fault diagnosis technologies,several challenges persist,including low fault identification accuracy,limited monitoring coverage,single-parameter detection,and the inability to classify faults or assess their severity.Furthermore,there is inadequate evaluation of potential fault-induced risks and a lack of comprehensive management strategies.Based on these findings,future research directions are proposed:advancing from single-parameter monitoring to multidimensional integrated monitoring,transitioning from sparse coverage to dense and comprehen

关 键 词：带式输送机托辊 托辊故障诊断 带式输送机协同管控 筒体破损 托辊轴承失效 筒体变形 主轴变形 

分 类 号：TD634[矿业工程—矿山机电]