【作者】 刘中华；

【导师】 盛赛斌； 肖大雏；

【作者基本信息】 武汉大学 ， 控制理论与控制工程， 2005， 硕士

【摘要】 风机是电厂重要的辅机设备,其运行状态直接关系到电厂的安全经济性。本文将虚拟仪器技术与人工神经网络技术相结合用于风机故障诊断,利用计算机网络从对设备进行远程监控的角度,进行了建立一个基于虚拟仪器技术LabVIEW的远程状态监测和故障诊断系统的尝试。 本文在电厂现有的监测系统基础上,提出合理的监测方案,利用LabVIEW强大的网络通信功能,将TCP/IP技术应用于风机监测方案中,从而实现设备的远程状态监测和故障诊断。系统采用服务器—客户机的形式,服务器主要负责振动信号的数据采集和发送,客户机主要负责完成数据接收和存储、信号分析处理和设备的故障诊断任务。 本文研究基于数据采集卡的虚拟仪器技术,整个系统软件按模块化设计,保证了良好的独立性和集成性,其中重点开发了数据采集模块、网络通信模块、监测报警模块、频域分析模块、数据管理模块和智能诊断模块等。 文章论述了基于LabVIEW的常用信号幅域、时域、频域分析方法,重点讨论了复调制细化分析方法、倒频谱分析方法、全息谱分析方法及其在LabVIEW中的实现,有效地解决了频率细化等问题。并将人工神经网络技术与LabVIEW相结合,利用LabVIEW的MATLAB Script节点研究自组织特征映射(Self-Organizing Feature Map)神经网络技术应用于风机的智能故障诊断方法;从采集的振动信号频谱分析中提取故障征兆,经神经网络智能模块诊断后,给出风机的故障类型,实现风机的故障诊断任务。 本文为研究虚拟仪器技术在电厂辅机状态监测和故障诊断上的应用提供了有益的探讨,具有一定的实用价值。更多还原

【Abstract】 Fan is one of essential auxiliaries, which running status is directly concerned with the safety and economy of whole power plant. The thesis introduces Virtual Instrument and technology and artificial neural network technology into fault diagnosis, constructs a remote condition monitoring and fault diagnosis based Virtual Instrument technology, from the perspective of remote monitoring by the computer network.The thesis brings forward a reasonable project for condition monitoring of fan with the current system. For LabVIEW has network communication function, the TCP/IP technology is used in the project to accomplish remote condition monitoring and fault diagnosis. The system is made of server-client model, the server is responsible for data acquisition of vibration and transform. While the client is responsible for receiving and saving data, accomplish signal processing and fault diagnosis of device.The thesis researches the implementation of Virtual Instrument based on DAQ card. The system software is designed by modules, each module can be developed separately and used to finish a subtask, this makes the system easy to be rebuild. Date acquisition module, network communication module, monitoring and warning module, frequency spectrum analysis module, date management module and intelligence diagnosis module are developed.The thesis discusses the amplitude-domain, time-domain and frequency-domain analysis for signal based on LabVIEW. The author attaches key importance to making a further research on Zoom FFT, Quefrency, Holograph and their implements in LabVIEW, which solves the problems of low frequency resolution. The thesis researches SOM neural network on the intelligent fault diagnosis of fan, using the node technology of the MATLAB Script in LabVIEW. The system draw the fault symptom from the frequency spectrum of vibration signal, diagnose the fault type of fan and accomplish the task of fault diagnosis by the neural network intelligence diagnosis module.The thesis provides a useful way for the application of virtual instrument to the condition monitoring and fault diagnosis for auxiliary in power plants. It is proved that the system had highly practical value.更多还原