毕业设计（论文）-汽车发动机智能故障诊断技术.doc

汽车发动机智能故障诊断的研究
摘要
本论文对国内外汽车故障诊断技术的发展、汽车故障的种类与特点、汽车故障的诊断方法、故障系统的基本诊断过程和理论方法进行了深入的分析。根据神经网络的特点,指出神经网络与故障诊断结合的可行性和必然性。在讨论了神经网络的基本理论基础上,明确神经网络应用于故障诊断的二种途径。前馈型BP网络具有极强的模式识别和分类能力。本文从应用角度分析了网络设计中的网络层数、隐含层的神经元数、初始权值、学习速率、期望误差的选取问题。以MATLAB ,重点研究了应用神经网络对汽车系统的控制仿真和故障诊断问题。
本文采用振动诊断法,在对汽车发动机进行结构及其典型故障分析,以及对振动信号的时域、频域及小波包进行深入分析的基础上,针对现场实测的EQ6102汽油型发动机机体表面振动信号与气缸盖固紧螺栓振动信号,提出了该型发动机的故障诊断流程,即对所测振动信号进行相关分析。
根据发动机机体振动信号的频率特性,确定出故障气缸;然后对该故障气缸进行时域分析,得出峭度参量是汽油发动机故障的敏感时域参数;接着对该故障信号进行频域分析,由随转速增加的频率图及柴油发动机的典型故障定性分析确定出该发动机的故障类型;最后对该故障信号进行小波包分析,确定该种故障的特征频带。通过上述分析确定的发动机故障敏感参量,可以为神经网络等模式识别提供较为准确的特征参量。
关键词:汽车故障诊断;神经网络;系统仿真
Automotive Engine Fault Diagnosis of Intelligent
Abstract
This paper analyses the development of fault technique in domestic and international, the category and characteristics of auto fault, the method of auto fault diagnosis, the basic diagnosis process and theories method of the fault diagnosis system. According to characteristics of the work, it is indicated that bination between the work and the fault diagnosis have possibility and inevitability. Basing on discussing the basic theories of work, it is defined that work can be applied for fault diagnosis in three ways. The work of forward type has the very strong mode identification and classifies ability. Form appliance aspect paper work layer number, nerve cell number of inside layer, original weigh, train speed, expecting error when designing are used for the work's training and to establish the simulation of the auto A/C fault diagnosis.
The vibration diagnosis method is adopted by this paper. Based on anglicizing structure and typical diagnosis of diesel engine, and anglicizing the vibration signals by time domain, frequency domain and wavelet packet deeply, the paper puts forward a fault diagnosis flow of the diesel engine for the vibration signals of the surface and cylinder head bolts on the EQ6102 diesel engine. The measured vibration signals are carried on correl