摘要:动车组是城市和市郊铁路实现小编组、大密度的高效率运输工具。这么方便快捷的交通工具如今变得不可或缺。转向架则是高速动车组的行走装置,它决定了列车运营的速度和运行的品质。动车组转向架轴承的工作状况是影响铁路运输安全的重要因素之一因此,开展动车组转向架轴承可靠性分析与故障诊断的研究,对保证运营安全、提高文修效率和避免不必要的损失等都具有重要的意义。21700
应用故障树分析法建立了动车组转向架轴承故障模型,并提出了要提高其可靠性的要求,并简单介绍了动车组转向架轴承振动机理、故障特征频率等特点。在轴承故障的监测技术中,本文利用振动监测技术监测动车组转向架轴承,并深入研究了故障诊断领域比较先进的理论与方法。两种方法对轴承故障进行诊断和监测。一种是时频域参数指标诊断方法,另一种方法是:智能诊断方法,先对振动信号进行小波包消噪提高其信噪比,再采用基于EMD(经验模态分解)的方法来提取轴承故障特征,把故障信号分解得到IMF,对几个重要的IMF进行分析,获得每个IMF分量的能量,作为BP神经网络的输入向量;根据遗传算法寻优的特点,结合改进遗传算法对BP神经网络的参数进行优化,再利用其对轴承的故障进行诊断,分析了该方法诊断的效果。系统是以软件为核心的虚拟仪器开发,使得系统具有扩展性强、灵活定义、性能高和文护费用低等优势。
神经网络经过半个多世纪的发展,如今已经成为一门比较成熟的学科。但由于其在实际应用中还有着巨大的潜力和前景。因此,关于神经网络的理论和实际应用的研究仍然是目前研究的热点和前沿。
本系统的所有开发和设计都以MATLAB软件为平台,并大量使用了MATLAB中的神经网络工具箱。MATLAB及其神经网络工具箱在神经网络系统的开发和设计中有着强大的功能,能够极大地简化开发和设计的过程。利用神经网络对动车组转向架的轴承进行故障分析。
毕业论文关键词: 转向架轴承;故障分析;诊断方法;神经网络
The EMU bogies Bearing Fault Diagnosis
Abstract: EMU is to achieve inter-city and suburban railway marshalling small, high density and efficient means of transport. Bogie is a high-speed dynamic running gear EMU determines
the speed of the train operators and operational quality. One of the important factors EMU bogie bearings working conditions that affect the safety of railway transportation Therefore, to study the EMU reliability analysis and fault diagnosis bogie bearings for ensuring safe operations, improving maintenance efficiency and avoiding unnecessary losses has important significance.
Application of fault tree analysis to establish the EMU bogie bearing fault model , and proposed to improve its reliability requirements , and briefly describes the EMU bogie vibration mechanism bearing fault characteristic frequency. In bearing failure in monitoring technology , we use technology to monitor vibration monitoring EMU bogie bearings, and in-depth study of the fault diagnosis of more advanced theories and methods . Both methods for diagnosis and monitoring of bearing failure . An index is a diagnostic method domain parameter frequency Another method is : intelligent diagnostic method, first the vibration signal wavelet packet de-noising improve its signal to noise ratio , then using EMD ( empirical mode decomposition ) based method to extract bearing fault feature , the fault signal decomposition IMF, IMF to analyze several important , get the energy of each component of the IMF , as BP neural network input vector ; according to the characteristics of the genetic algorithm optimization , combined with improved genetic algorithm BP neural networks to optimize the parameters , and then use it on the bearing fault diagnosis, diagnostic analysis of the effect of this method . System software is the core of the virtual instrument development , the system has scalability, flexible definition , high performance and low maintenance cost advantages. 动车组转向架轴承故障诊断+源程序:http://www.youerw.com/tongxin/lunwen_14171.html