基于多尺度特征匹配的矿井排水系统故障诊断算法  

Fault Diagnosis Algorithm of Mine Drainage System Based on Multi-scale Feature Matching

作　　者：郭激光 常树峰 张文锋[1] 王亮 田坤云[3] GUO Jiguang;CHANG Shufeng;ZHANG Wenfeng;WANG Liang;TIAN Kunyun(Anyang Zhujiao Coal Industry Co.,Ltd.,Anyang 455100,China;Dongpang Mine,Jizhong Energy Resources Co.,Ltd.,Xingtai 054200,China;School of Resources and Safety Engineering,Henan University of Engineering,Zhengzhou 451191,China)

机构地区：[1]安阳市主焦煤业有限责任公司,河南安阳455100 [2]冀中能源股份有限公司东庞矿,河北邢台054200 [3]河南工程学院资源与安全工程学院,河南郑州451191

出　　处：《金属矿山》2024年第8期146-151,共6页Metal Mine

基　　金：河南省高校科技创新团队计划项目(编号:22IRTSTHN009)。

摘　　要：矿井排水系统的高效运行对于煤矿生产具有重要意义,然而,故障的发生会导致排水系统运行异常,进而影响生产安全和效率。提出了一种基于多尺度特征匹配的矿井排水系统故障诊断算法。该算法首先通过传感器获取到实时的矿井排水系统数据,并对其进行预处理和特征提取。然后,采用多尺度特征匹配方法,将数据特征映射到不同的尺度空间中。在每个尺度空间中,通过构建特征匹配模型,将当前数据与已知的正常运行数据进行比对,以判断是否存在故障。为了提高故障诊断的准确性和鲁棒性,引入集成学习方法,通过集成多个尺度空间的特征匹配结果,并利用投票机制进行综合决策,最终确定矿井排水系统的故障类型和位置。研究表明:所提算法在矿井排水系统故障诊断方面具有良好性能,与传统方法相比,该算法能够高效地检测故障,有助于及时准确地发现和解决矿井排水系统问题,对于确保矿井安全生产具有一定的意义。The efficient operation of mine drainage system is of great significance to coal mine production.However,the occurrence of fault will lead to abnormal operation of drainage system,which will affect production safety and efficiency.A fault diagnosis algorithm of mine drainage system based on multi-scale feature matching is proposed.Firstly,The algorithm obtains the real-time mine drainage system data through the sensor,and carries on the preprocessing and feature extraction.Then,the multi-scale feature matching method is used to map the data features to different scale spaces.In each scale space,by constructing a feature matching model,the current data is compared with the known normal operation data to determine whether there is a fault.In order to improve the accuracy and robustness of fault diagnosis,an ensemble learning method is introduced to determine the fault type and location of the mine drainage system by integrating the feature matching results of multiple scale spaces and using the voting mechanism to make comprehensive decisions.The study results show that the proposed algorithm has good performance in fault diagnosis of mine drainage system.Compared with traditional methods,the proposed algorithm can efficiently detect faults,help to timely and accurately discover and solve the problems of mine drainage system,and has certain significance for ensuring the safety of mine production.

关 键 词：矿井排水系统 多尺度特征匹配 故障诊断 集成学习方法 

分 类 号：TD74[矿业工程—矿井通风与安全]