深度学习在煤矿机电设备检测技术中的应用研究  

Application Study of Deep Learning in Coal Mine Electromechanical Equipment Inspection Technology

作　　者：张天峰 ZHANG Tianfeng(Guoneng Shuzhi Technology Development(Beijing)Co.,Ltd.,Beijing 100011,China)

机构地区：[1]国能数智科技开发(北京)有限公司,北京100011

出　　处：《自动化仪表》2025年第2期102-106,共5页Process Automation Instrumentation

基　　金：2019年工业和信息化领域公共服务能力提升专项基金资助项目(2019-00910-4-1)。

摘　　要：煤矿行业的传统胶带机轴承故障检测存在成本高、效率低和错误率高等问题。为了解决这些问题,构建了一种基于冲击脉冲传感器的胶带机故障检测模型。首先,利用冲击脉冲传感器对现有故障检测方法进行优化。然后,利用优化后的方法构建胶带机轴承故障检测模型。最后,验证模型的实际应用效果。创新性地将深度学习与冲击脉冲传感器结合,用于煤矿胶带机轴承故障的监测,以提高矿用胶带机故障检测的准确性,从而实现智能化、自动化。试验结果表明:该模型的振幅频率的准确率比对比方法分别高0.055和0.033;模型迭代至100次时趋于平稳,且收敛速度更快、损失函数降低得更小。该模型可应用于煤矿机电设备检测领域。The traditional tape machine bearing fault inspection has the problems of high cost,low efficiency and high error rate and other in coal mining industry,In order to solve these problems,a tape machine fault inspection model based on impact pulse sensor is constructed.Firstly,the existing fault inspection method is optimized by using the impact pulse sensor.Then,the optimized method is used to construct a tape machine bearing fault inspection model.Finally,the model effect of practical application is verified.The innovative combination of deep learning and impact pulse sensor is used for the monitoring of bearing faults of coal mine tape machine with improves the accuracy of fault inspection of mine tape machine and realizes intelligence and automation.The test results show that the amplitude frequency of the model is 0.055 and 0.033 more accurate than that of the comparison method seperately;it tends to be stable when model iterating up to 100 times,the convergence speed becomes faster,and the loss function reduces less.The model can be applied to the inspection of electromechanical equipment in coal mines.

关 键 词：冲击脉冲传感器 矿用胶带机 故障监测 轴承 自动化 生命周期 

分 类 号：TH13[机械工程—机械制造及自动化]