盾构机刀盘驱动液压马达故障诊断研究

发布时间：2018-04-08 17:02

本文选题：盾构机　切入点：BP神经网络　出处：《广东工业大学》2017年硕士论文

【摘要】：随着人工智能技术和信息技术在机械工业中的发展,机械故障诊断技术也朝着智能化的方向发展。盾构机刀盘直径大、工作环境恶劣和受力复杂,并且盾构刀盘驱动液压马达数量有8个,分布范围大。液压马达发生故障将直接导致盾构机无法正常工作,甚至引发事故。利用传统的依赖于维修者个人经验的手工或半自动的故障诊断方式对液压马达进行故障诊断,工作效率低、准确度不高,本文开展的盾构机刀盘驱动液压马达智能诊断系统的研究工作是要开发出效率更高,更智能的故障诊断的方式,因此对大型机械设备的维护具有重要意义。本文针对盾构机刀盘驱动液压马达的故障诊断,研究了基于BP神经网络和专家系统诊断方法的液压马达故障诊断系统。首先,研究了盾构机刀盘驱动液压马达的故障产生机理和故障发生规律,为液压马达故障诊断系统积累了专家知识。其次,研究了液压马达故障的模式识别,故障特征信号的选取及故障特征监测参数,确定了故障信号监测的位置、信号获取过程和信号的标度变换过程。分析了滤波方法在故障信号预处理中的应用,研究了故障特征提取的时域分析法和频域分析法。再次,研究了BP神经网络和专家系统结合的故障诊断方法,分析了人工神经网络的拓扑结构和BP神经网络的算法流程,同时研究了专家系统的结构,深入研究了专家系统知识库的知识表示和专家系统的知识获取模型以及液压马达故障知识库结构。在此基础上,研究了神经网络专家系统的结构及其盾构机刀盘驱动液压马达故障诊断模型,并深入的研究了神经网络专家系统的推理机制和解释机制。最后,研究盾构机刀盘驱动液压马达故障诊断系统的开发流程,同时构建了故障诊断系统的硬件系统,并选取了硬件设备搭建数据采集、数据处理和数据通信的硬件平台。在此基础上,利用模块化思想设计盾构机刀盘驱动液压马达故障诊断系统软件总体结构,然后采用C++语言编写各模块程序源代码,并在Visual Studio 2005软件中建立了盾构机刀盘驱动液压马达故障诊断系统软件中各模块的应用程序,包括液压马达故障诊断系统登录界面、状态监测与故障诊断功能、故障报警与诊断功能、BP神经网络训练设计、知识库管理等。通过本文的研究,所得研究结果对大型设备液压马达的故障诊断和维护具有指导作用,同时可以作为液压系统关键零部件的故障监测诊断研究的借鉴和参考。
[Abstract]:With the development of artificial intelligence and information technology in mechanical industry, mechanical fault diagnosis technology is also developing intelligently.The shield machine has large diameter of cutter head, bad working environment and complex force, and there are 8 hydraulic motors driven by cutter head of shield machine, which have a wide distribution range.Hydraulic motor failure will directly lead to shield machine can not work properly, or even cause an accident.The traditional manual or semi-automatic fault diagnosis method, which depends on the personal experience of the maintainer, is used to diagnose the fault of the hydraulic motor.The research work of the intelligent diagnosis system for hydraulic motor driven by cutter head of shield machine in this paper is to develop a more efficient and intelligent way of fault diagnosis, so it is of great significance for the maintenance of large mechanical equipment.In this paper, the fault diagnosis system of hydraulic motor driven by shield machine is studied based on BP neural network and expert system diagnosis method.Firstly, the mechanism of fault generation and the rule of fault occurrence of hydraulic motor driven by cutter head of shield machine are studied, and the expert knowledge is accumulated for the fault diagnosis system of hydraulic motor.Secondly, the fault pattern recognition of hydraulic motor, the selection of fault feature signal and the parameters of fault feature monitoring are studied, and the location of fault signal monitoring, signal acquisition process and signal scale transformation process are determined.The application of filtering method in fault signal preprocessing is analyzed. The time domain analysis method and frequency domain analysis method for fault feature extraction are studied.Thirdly, the fault diagnosis method combining BP neural network and expert system is studied, the topology structure of artificial neural network and the algorithm flow of BP neural network are analyzed, and the structure of expert system is also studied.The knowledge representation of expert system knowledge base, the knowledge acquisition model of expert system and the structure of hydraulic motor fault knowledge base are studied.On this basis, the structure of neural network expert system and the fault diagnosis model of hydraulic motor driven by cutter head of shield machine are studied, and the reasoning mechanism and explanation mechanism of neural network expert system are studied deeply.Finally, the development process of the hydraulic motor fault diagnosis system driven by the cutter head of shield machine is studied. At the same time, the hardware system of the fault diagnosis system is constructed, and the hardware platform of data acquisition, data processing and data communication is built.On this basis, the software structure of hydraulic motor fault diagnosis system driven by cutter head of shield machine is designed by using modularization thought, and the source code of each module program is compiled by C language.In the software of Visual Studio 2005, the application program of each module in the software of hydraulic motor fault diagnosis system driven by cutter head of shield machine is established, including the login interface of hydraulic motor fault diagnosis system, the function of condition monitoring and fault diagnosis.The function of fault alarm and diagnosis is BP neural network training design, knowledge base management and so on.Through the research in this paper, the results can be used as a guide for the fault diagnosis and maintenance of the hydraulic motor of large equipment, and it can also be used as a reference for the fault monitoring and diagnosis of the key parts of the hydraulic system.
【学位授予单位】：广东工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U455.39;TH137.51

【参考文献】