基于RS-SVM数据融合的全液压钻机智能故障诊断研究

发布时间：2018-04-01 01:23

本文选题：钻机　切入点：振动测量节点　出处：《成都理工大学》2016年硕士论文

【摘要】：钻机是一个机电液一体化的复杂控制系统,其故障类型繁多。全液压钻机在运行过程中容易发生剥落、磨损、松动、疲劳、冲击、泄漏等故障,当故障发生时将导致钻机某些物理参数的改变,如动态信号非平稳性的出现,液压回路的压力、流量波动等,因此,无论何种故障,都与系统的物理参数有着直接或间接的联系,只要掌握钻机监测参数的特征信息,就能实现钻机的故障诊断。针对全液压坑道钻机的液压系统和减速器齿轮故障,从机电设备的实时状态监测和故障诊断两个重要方向入手,对钻机故障诊断展开了设计与研究。涉及钻机故障机理、MEMS振动测量节点、多传感器数据融合方法及钻机故障模拟实验等四方面内容。首先研究了全液压钻机的故障产生机理。提出多传感器故障信号采集方案。针对旋转机械监测困难区域的振动测量,开发研制基于MEMS加速度计的无线振动测量节点,实现节点软硬件设计。其次根据钻机故障特点,通过研究小波包能量特征向量提取、粗糙集属性约简、支持向量机等理论,提出基于RS-SVM的多传感器数据融合方法,将其应用于钻机故障诊断中,提高钻机故障分类诊断的准确率。最后,搭建振动标定实验台,进行振动标定校准实验。并利用断齿齿轮、节流阀泄漏支路模拟钻机的减速器齿轮断齿、液压马达内泄漏故障,进行钻机故障模拟实验。综上所述,论文研究的振动实时状态监测节点和多传感器数据融合故障诊断方法具有一定的参考和应用价值。
[Abstract]:Drilling rig is a complex control system of electromechanical and hydraulic integration, and its fault types are various.Fully hydraulic drilling rig is prone to spalling, wear, loosening, fatigue, impact, leakage and other faults during operation. When the failure occurs, it will lead to the change of some physical parameters of drilling rig, such as the non-stationary dynamic signal.The pressure and flow of hydraulic circuit fluctuate, so no matter what kind of fault, it has direct or indirect relation with the physical parameters of the system, so long as we master the characteristic information of the monitoring parameters of drilling rig, we can realize the fault diagnosis of drilling rig.Aiming at the faults of hydraulic system and reducer gear of full hydraulic tunnel drilling rig, the design and research of fault diagnosis of drilling rig are carried out from two important directions of real time state monitoring and fault diagnosis of electromechanical equipment.This paper deals with the fault mechanism of drilling rig and MEMS vibration measurement node, multi-sensor data fusion method and drilling machine fault simulation experiment.First of all, the mechanism of fault generation of full hydraulic drilling rig is studied.A multi-sensor fault signal acquisition scheme is proposed.A wireless vibration measurement node based on MEMS accelerometer is developed for vibration measurement in difficult areas of rotating machinery monitoring. The hardware and software of the node are designed.Secondly, according to the fault characteristics of drilling rig, by studying the theory of wavelet packet energy feature vector extraction, rough set attribute reduction, support vector machine and so on, a multi-sensor data fusion method based on RS-SVM is proposed and applied to the fault diagnosis of drilling rig.Improve the accuracy of fault classification and diagnosis of drilling rig.Finally, the vibration calibration test bench is built, and the vibration calibration experiment is carried out.The fault simulation experiment of drilling machine is carried out by using broken gear and throttle valve leakage branch to simulate reducer gear breaking tooth and hydraulic motor leakage fault.To sum up, the method of vibration real-time monitoring node and multi-sensor data fusion fault diagnosis studied in this paper has certain reference and application value.
【学位授予单位】：成都理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TH17

【相似文献】