基于支持向量机的加工中心主轴系统状态监测技术研究

发布时间：2018-04-02 03:00

本文选题：主轴系统　切入点：状态监测　出处：《吉林大学》2017年硕士论文

【摘要】：随着信息化时代的到来,制造业与信息科技进一步融合与发展,基于信息系统的智能装备、智能工厂等正引领着制造方式的变革。传统的数控机床越来无越法满足加工过程高效化、自动化的要求,实现数控机床的状态与故障监测,对于提高生产效率、降低人力成本、提升生产过程的自动化水平具有重要意义。本文以VDL-1000加工中心主轴系统为监测对象,采用多个传感器及数据采集卡等搭建硬件系统,利用支持向量机等算法编制软件系统,建立了加工中心主轴状态监测系统,并通过相关试验验证了监测系统的可行性。该监测系统不仅提高了加工中心的自动化水平,也为后续进行机床加工过程的故障预警与诊断奠定了基础。论文首先对VDL-1000加工中心及其主轴系统的主要结构进行了研究,对主轴系统的常见故障进行了统计分析。在此基础上,确定了主轴状态监测系统的监测参数,并提出了监测系统的总体方案。论文选择电流互感器、贴片式温度传感器、加速度传感器、频率变送器等分别对主轴系统的电流、温度、振动、转速等参数进行测量,利用各传感器、数据采集卡及工控机等完成了监测系统的硬件系统搭建。监测系统的软件系统则采用MATLAB与Lab VIEW混合编程,该软件系统具有信号采集、显示、滤波和分析等诸多功能,其中基本信息管理模块可通过密码验证登录系统,同时具有用户信息管理、系统信息管理、修改密码等功能;而数据采集模块则通过动态链接库PCI8602.dll的库函数实现信号的采集;数据存储模块则可实现数据的实时存储并将其保存为电子表格的形式,便于后续的查看与使用;状态监测模块利用Lab VIEW的显示控件实现对各监测信号的实时显示。信号处理与分析模块是主轴状态监测系统软件的核心模块:模块采用低通滤波与基于Birge-Massart阈值的小波降噪相结合的方式对不同信号进行滤波处理,以获得更加真实的信号,便于后续的信号分析。该模块还可通过信号的时域分析获取信号的时域特征,利用信号的自相关分析获取信号的周期特征,通过信号的频谱分析与小波分析来获得信号的频域特征。此外,该模块提供的基于支持向量机的状态识别算法可用于加工中心主轴系统故障状态的识别。主轴状态监测系统搭建完成后,本文针对该系统的功能进行了验证性试验,证明了系统的可行性,并利用轴承故障监测案例与松动故障监测案例验证了系统具有的故障监测功能。本文所开发的基于支持向量机的主轴状态监测系统具有简单实用、功能完备等特点,实现了预期的设计目标,在数控机床状态监测的研究方面具有较高的应用价值。
[Abstract]:With the arrival of information age, manufacturing and information technology are further integrated and developed. Intelligent equipment based on information system and intelligent factory are leading the transformation of manufacturing mode.The traditional NC machine tool can meet the requirements of high efficiency and automatization in the process of machining, and realize the condition and fault monitoring of NC machine tool, which can improve the production efficiency and reduce the labor cost.It is of great significance to improve the automation level of production process.In this paper, the main axis system of VDL-1000 machining center is taken as the monitoring object, the hardware system is built by using several sensors and data acquisition cards, and the software system is compiled by using support vector machine and other algorithms, and the monitoring system of spindle status of machining center is established.The feasibility of the monitoring system is verified by relevant experiments.The monitoring system not only improves the automation level of the machining center, but also lays a foundation for the fault warning and diagnosis of the machine tool.Firstly, the main structure of VDL-1000 machining center and its spindle system are studied, and the common faults of spindle system are analyzed statistically.On this basis, the monitoring parameters of the spindle state monitoring system are determined, and the overall scheme of the monitoring system is put forward.In this paper, current transformer, mount temperature sensor, acceleration sensor, frequency transmitter and so on are selected to measure the current, temperature, vibration, rotational speed of the spindle system, and the sensors are used to measure the current, temperature, vibration and rotational speed of the spindle system.Data acquisition card and industrial computer have completed the hardware system construction of the monitoring system.The software system of the monitoring system uses MATLAB and Lab VIEW mixed programming, the software system has many functions, such as signal acquisition, display, filtering and analysis, among which the basic information management module can verify the login system by cipher.At the same time, it has the functions of user information management, system information management and password modification, while the data acquisition module realizes the signal acquisition through the library function of dynamic link library (PCI8602.dll).The data storage module can store the data in real time and save it in the form of spreadsheet, which is easy to view and use. The state monitoring module uses the display control of Lab VIEW to realize the real-time display of each monitoring signal.The signal processing and analysis module is the core module of the spindle state monitoring system software. The module uses the low-pass filter and wavelet denoising based on Birge-Massart threshold to filter the different signals to obtain more real signals.Facilitate subsequent signal analysis.In this module, the time-domain features of the signals can be obtained by the time-domain analysis, the periodic characteristics of the signals can be obtained by the autocorrelation analysis of the signals, and the frequency-domain features of the signals can be obtained by the spectrum analysis and wavelet analysis of the signals.In addition, the state recognition algorithm based on support vector machine provided by this module can be used to identify the fault state of machining center spindle system.After the main shaft condition monitoring system has been built, the feasibility of the system is proved by the verification test on the function of the system.The fault monitoring function of the system is verified by bearing fault monitoring cases and loosening fault monitoring cases.The main axis state monitoring system based on support vector machine developed in this paper has the characteristics of simple and practical, complete function and so on. It has achieved the expected design goal, and has high application value in the research of NC machine tool condition monitoring.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG659

【相似文献】