悬臂式掘进机油液系统故障诊断方法研究  

Research on fault diagnosis method of boom roadheader oil-liquid system

作　　者：刘国鹏 Liu Guopeng(CCTEG Taiyuan Research Institute Co.,Ltd.,Taiyuan 030006,China;Shanxi Tiandi Coal Mining Equipment Co.,Ltd.,Taiyuan 030006,China;National Engineering Laboratory for Coal Mining Machinery,Taiyuan 030006,China)

机构地区：[1]中国煤炭科工集团太原研究院有限公司,山西太原030006 [2]山西天地煤机装备有限公司,山西太原030006 [3]煤矿采掘机械装备国家工程实验室,山西太原030006

出　　处：《煤炭科技》2025年第2期46-52,共7页Coal Science & Technology Magazine

基　　金：山西省专利转化计划项目(202404004);山西省重点研发计划项目(202202020101005);天地科技股份有限公司科技创新创业资金专项项目重点项目(2022-2-TD-ZD001);山西天地煤机装备有限公司科技重大项目(M2023-ZD12)。

摘　　要：悬臂式掘进机油液系统在长期运行过程中,其稳定性和可靠性对于整个设备的安全和效率至关重要。传统的单一传感器监测方法存在数据维度有限、互补性不足的局限,无法全面反映油液系统工作中的真实状态。为了解决这一问题,提出了基于卷积神经网络(Convolutional Neural Networks,CNN)的掘进机多传感器油液系统故障诊断方法。首先,进行润滑油的监测与数据采集,这一步骤主要关注润滑油的理化性能指标,如黏度和酸值等,以及油液中携带的磨粒信息。其次,利用多传感器采集的数据作为输入,构建并训练一个卷积神经网络(CNN)模型进行故障诊断。最后,为验证所提方法的有效性,将其与其他机器学习方法在故障状态分类的准确性方面进行比较。结果表明,所提出的方法在诊断精度上明显高于其他传统机器学习方法。这一结果不仅证明了CNN模型在处理此类复杂数据时的优势,还进一步验证了该方法的高效性和准确性。The stability and reliability of the oil-liquid system of boom roadheader is crucial to the safety and efficiency of the whole equipment during long-term operation.The traditional single-sensor monitoring method has the limitations of limited data dimension and insufficient complementarity,which cannot comprehensively reflect the real state of the oil-liquid system in operation.In order to solve this problem,a multi-sensor oil-liquid system fault diagnosis method for roadheader based on Convolutional Neural Networks(CNN)was proposed.Firstly,lubricant monitoring and data acquisition were carried out.This step focuses on the physical and chemical performance indicators of the lubricant,such as viscosity and acid value,and the information of abrasive particles carried in the oil-liquid.Second,a CNN model was constructed and trained for fault diagnosis using the data collected from multiple sensors as input.Finally,in order to verify the effectiveness of the proposed method,it was compared with other machine learning methods in terms of the accuracy of fault state classification.The experimental results show that the proposed method is significantly higher than other traditional machine learning methods in terms of diagnostic accuracy.This result not only proves the advantages of the CNN model in dealing with such complex data,but also further validates the efficiency and accuracy of the method.

关 键 词：悬臂式掘进机 油液 卷积神经网络 理化指标 磨粒信息 

分 类 号：TD421.5[矿业工程—矿山机电]