【作者】 黄志强；

【导师】 黄敬党； 翁红林；

【作者基本信息】 福建农林大学 ， 机械设计及理论， 2006， 硕士

【摘要】 近年来，随着机械设备故障诊断技术的发展，作为机械设备一部分的小型柴油机，其故障诊断技术的研究也越来越得到人们的重视。声信号诊断技术作为一种非接触式诊断方法，由于其操作的便利性也日益受到人们的关注。本课题在总结和汲取前人研究成果的基础之上，结合实际情况，将声信号诊断技术应用到小型柴油机故障诊断上，引入了适用于非平稳信号分析的方法——小波包分析方法，从小型柴油机缸盖声信号中获取小型柴油机故障信号特征信息，并结合模式识别方法有效地实现了对小型柴油机故障的智能诊断。 本课题主要从应用的角度出发，以小型柴油机作为研究对象，在测试技术、信号处理、小波分析、模式识别等理论的基础上，深入研究小型柴油机的故障机理、故障声信号特征提取及其诊断方法。研究结果表明，在不解体的情况下，利用柴油机缸盖声信号对小型柴油机进行故障诊断是可行的。 本研究课题作了如下几个方面的研究工作： 1)从测量对象、测量位置及参数和方法的选择等方面对小型柴油机故障诊断系统的设备及其相关设备参数进行选择。同时，本研究还从时序分析及频谱分析角度对KM178F小型柴油机常见的几种故障如：排气间隙异常、进气压力低、低压油路供油不畅的故障机理及其声信号的特性作了具体的分析。研究表明，在不解体的情况下，利用小型柴油机缸盖声信号对其进行故障诊断是可行的。 2)针对小型柴油机声信号的一些具体特点，提出了将小波包分析的方法应用于声信号特征的提取，并将该方法具体应用到KM178F小型柴油机故障信号的分析上。在研究过程中，利用傅氏分析方法具体分析了KM178F柴油机在不同故障下的声信号的时频特点，并以此为依据对小波包分解的相关参数作了选择。在研究过程中，针对KM178F柴油机的这些特点确定采用Db4小波基对其进行小波包分析。通过对分析结果的比较表明利用各故障信号小波包分析得出的特征向量能够有效地实现对故障的区分。 为了实现对小型柴油机故障的智能诊断，本研究课题利用欧氏距离判别方法对小波包分析的结果进行归类以实现对小型柴油机故障类型的自动识别；并将该方法具体应用到KM178F小型柴油机随机声信号的智能识别上。通过对正常信号、气门间隙异常、进气压力低、低压油路供油不畅等几种的不同工况进行识别，正确率达到84.375％。表明效果良好。 3)本研究从小型柴油机故障诊断实际应用需要出发，设计了小型柴油机故障诊断软件系统的总体架构。在此基础上，完成了信号采集、样本信号分析、样本库的建立以及随机信号的分析等功能模块的开发。并将该诊断系统实际应用到KM178F小型柴油机的故障诊断中，通过对其声信号的实际识别表明效果良好。 从实际应用结果表明，将声信号诊断技术应用到小型柴油机故障智能诊断中，利用KM178F小型柴油机系统缸盖声信号的特性可以有效地实现对其工作状态的识别，从而实现对小型柴油机故障的智能诊断。同时，从柴油机的一些共性可以得出，该方法同样可以应用到其他型号的柴油机上。更多还原

【Abstract】 Application of Noise Signal in Fault Diagnosis of the Mini-dieselIn recent years, the technique of mechanical fault diagnosis is quickly developed in domestic and overseas. As an important part of machine, it is significant to diagnose faults of the mini-diesel. At the same time, acoustical diagnosis technique has been paid more and more attention because of its maneuverability. In this paper, characteristic information of faults is accessed from diesel engine noise signal by using new method, wavelet-package, on the basis of former achievements and valid results.This paper attempts to research profoundly fault mechanism, fault feature extraction and fault diagnosis method of diesel engine from the view of practical application. Its research object is mini-diesel and its academic foundation is test technology, signal processing, wavelet analysis, pattern recognition and etc. the results show that it is effective and feasible to diagnose faults of mini-diesel engine by means of un-disassembly using surface acoustical signal. The paper consists following parts:1. The feasibility of the fault diagnosis by the noise signal of mini-diesel is demonstrated from the points of the choice of measure objects, measure position and relative parameters. And the relative equipments and parameters of the lab are chose. At the same time, I also studied the familiar fault of the KM178F mini-diesel such as the abnormal of the drain tap, the low pressure of the in gas, the obstruction of the oil-tube from the analysis of time series and frequency. The study indicates that the noise signal can be used to diagnose the fault of the mini-diesel without disassemble the diesel.2. The technology of wavelet-package is applied to abstract the character of the noise signals considering the special trait of the mini-diesel. And the method is applied to the KM178F mini-diesel. During the studying, the FFT is used to analyze the trait of frequency of the noise signal. The parameters of the wavelet-package analyze are chose base on the result of the FFT analysis. Considering the specialty of the KM178F diesel, the Db4 wavelet function is chose. The result of analysis testified that the identification of noise signal could be realized by using the characteristic vectors of the wavelet-package analysis result.In order to diagnose the fault of the mini-diesel automatically, the study uses the Euclidean to analyze and class the result of the wavelet-package. Finally, this method is applied in the fault diagnosis of the KM178F mini-diesel. The rate of the correction can reach 84.375%, and the effect is testified.3. Finally, the blue print of the fault diagnosis system of mini-diesel based on noise signal is put forward. The truss of the software is also advanced. The effect is also testified by putting it into the use on the fault diagnosis of the KM178F mini-diesel.The application testified that it can identify the working state of the mini-diesel by using the noise signal of the KM178F mini-diesel, thus, the automatic fault diagnosis of the KM178F mini-diesel is realized. As the更多还原