摘要
为了更好地解决煤矿乳化泵故障难以排查和潜在的异常难以及早发现的问题,研发了综采面泵站实时在线监测和故障预警诊断系统,模拟了乳化泵常见的故障工况和正常工况来采集建模所需的训练集样本,应用了随机森林算法对乳化泵常见的四类工况进行诊断模型构建,使用乳化泵四类不同工况的真实数据对模型进行测试,准确率达到了97.1%的优良性能,并运用了工控上位机和下位机技术开发了一套在线监测系统,能够及时地发现乳化泵存在的故障类型、故障部件和尽早地发掘潜在的异常情况,查询设备历史故障数据、部件振动信号的趋势图谱、时频域特征值进行追溯分析,生成报表给出乳化泵检修维护和供应管理策略,对于泵站检修工作有重要的指导作用,对于保障工作面支护系统的正常运行、工作面安全高效开采有着重要意义。
In order to better solve the problems that emulsification pump faults are difficult to troubleshoot and potential anomalies are difficult to find early,a real-time fault diagnosis and online monitoring system for fully mechanized mining face pump station is developed,and the training set required for modeling is collected by simulating the common fault and normal working conditions of emulsification pump.The random forest algorithm is applied to construct the diagnosis model for four types of common normal and fault working conditions of emulsification pump.The model was tested with the real data of four different working conditions of the emulsion pump,and the accuracy rate reached 97.1%,and the use of industrial control upper computer and lower computer technology to develop an online monitoring system,can promptly find the emulsion pump fault type,fault parts and early discovery of potential abnormal conditions,query equipment historical fault data,component vibration signal trend map,time-frequency domain characteristic value for retrospective analysis,and generate a report to give emulsion pump maintenance and supply management strategy,for the pump station maintenance work has an important guiding role,for the normal operation of the face support system,safe and efficient mining of the face is of great significance.
作者
吴早阳
赖岳华
李丹宁
西成峰
杨程锦
WU Zaoyang;LAI Yuehua;LI Danning;XI Chengfeng;YANG Chengjin(Beijing Tianma Intelligent Control Technology Co.,Ltd.,Beijing100013,China)
出处
《煤炭技术》
CAS
2024年第5期284-287,共4页
Coal Technology
基金
中国煤炭科工集团有限公司科技创新创业资金专项项目(2023—TD—QN006)。
关键词
综采工作面
乳化液泵站
故障实时诊断
在线监测
fully-mechanized face
emulsion pump station
real-time fault diagnosis
on-line monitorin
