节点文献

内燃机的非平稳信号分析方法及其噪声源小波识别技术的研究

Research on Processing Method of Non-stationary Signal and Wavelet Identification Technology of Acoustical Sources in I. C. Engines

【作者】 韩军

【导师】 郝志勇;

【作者基本信息】 天津大学 , 动力机械及工程, 2004, 博士

【摘要】 内燃机是动力、交通、舰船、石化等国民经济重要领域的关键设备。随着全球范围内环保要求的日益提高,内燃机的噪声控制已成为目前国内外的热点研究问题。解决内燃机的噪声控制问题首先必须在内燃机上找到最大噪声辐射的部位或部件,即主要的噪声源,然后针对主要噪声源的声源特性采取相应的措施,才能够收到预期的效果。目前的研究已经表明,从内燃机表面辐射的声信号中包含有大量的振动噪声源信息。但是由于内燃机复杂的机械结构及能量转换过程所导致的复杂多激励、非稳态和非线性的多重耦合系统特性,使得采用传统的信号分析理论及方法难以完全适用于对一些有用信息进行捕捉和深入分析,因而有关的研究和应用也受到了很大的限制。本文在分析国内外研究现状的基础上,围绕几个难点问题进行了深入和系统的理论和方法研究,并在工程问题进行了实际应用。其中主要内容如下: 1.对内燃机的基本动力学特性进行了系统全面的分析与介绍。通过缸内燃烧气体波动方程的建立和求解,揭示了这一主要冲击激励源的瞬态振荡特性和变化规律;通过发动机内不同性质的激励源机理和路径特征分析,建立了切合实际的冲击响应结构模型,为整机的动力响应分析和实用噪声源识别方法研究奠定了理论基础。2.讨论了声强法的物理意义及其在声场分析和声源识别中的作用,并利用试验仪器建立了声强测量系统。基于先进的试验分析设备,分别对YC6108G 型柴油机、JX493ZQ 型柴油机、WD615 型柴油机和一辆中型乘用车进行了声强在噪声源识别和声场分析中的应用研究。3.由于采集到的内燃机的声信号具有非平稳特性,而传统的基于傅里叶变换的分析方法由于其局限性,不能完全适合于分析非平稳信号,也很难揭示信号更多的本质特征。相对于傅里叶变换而言,小波变换在非平稳信号分析方面具有很多独特优点。本文首先对傅里叶变换,短时傅里叶变换和小波变换进行了比较介绍,分析了它们各自的特点,并详细讨论了小波变换的一些问题。包括集中性和小波函数中心频率对分辨率的影响等,还给出了如何确定小波变换中最大和最小尺度参数的方法。

【Abstract】 Research on the noise reduction of internal combustion engines is being paid more and more attentions recently because of its importance in the essential industrial branch such as power, transportation, shipping, petrochemistry etc. The first task for noise control purposes is identifying which surfaces on his power unit will radiate noise well and the major vibroacoustical sources. Then in order to control the noise level, it is important to take the corresponding methods depending on the characteristics of major engine vibroacoustical sources. The acoustic signals obtained from a diesel engine contain many useful vibroacoustical sources information to reflect the engine operating conditions. Such information often exists in the form of short duration transients. Because of the complex multi-excitations caused by engine structural and energy transform procedure, its non-steady working process and non-liner system behavior, the traditional analytic theory and methods in acoustic signals are not always suitable to describe this complex situation. Based upon a detailed analysis of the present worldwide research status, an in-depth and systemic theory and methodology investigation is made to the some problems. Research achievements are put into the practical engineering application. Among the total, the major creative contributions and achievements are as follows: 1. A detailed theoretic analysis is made to the fundamental dynamic behavior of the reciprocating engine. Via the development and resolution of the inner-cylinder combustion gas oscillating equation, its transient oscillating behavior and changing tendency are investigated. A practical shock-induced transient response model is developed to reveal the engine dynamic response and proceed the relative research of applicable identification methodology. 2. The physical meaning of sound intensity and its application in noise field analysis and noise sources identification were discussed. A sound intensity measurement system was set up. Based on those, research on application of sound intensity in IC engine noise source identification and acoustic field analysis for a 6108 diesel engine, 493ZQ diesel engine, WD615 diesel engine and a medium bus was made. Influences of measurement methods on measurement accuracy were discussed. 3. The acoustic signals obtained from a IC engine are non-stationary transients. Unfortunately, the Fourier transform (FT) based methods are not suitable for non-stationary signal analysis and are not able to reveal the inherent information of non-stationary signals. Compared with the FT, the wavelet transform has many distinct advantages for acoustic signal analysis. In this part, the dissertation starts with a comparison among the Fourier transform, the Short Time Fourier transform and the wavelet transform, whose properties are given. Some problems about the wavelet transform are discussed in detail, including the concentration, the border distortion and the effect of the wavelet function’s central frequency on the resolution of the wavelet transform. A method to determine the maximal and the minimal scale parameters in the wavelet transform calculation is put forward. 4. Based on theory and test, research on application of the wavelet transform in IC engine acoustic source identification and acoustic field analysis for a 6108 diesel engine, 493ZQ diesel engine, WD615 diesel engine was made. Some important acoustic sources are identified. 5. Although the IC engine noise sources have distinctive time instances, it is still difficult to resolve them accurately based on noise measurement. This is because the occurrences of each noise source are too close together. And some noise sources such as the injection of fuel and operation of inlet and exhaust valves produce low-level noise. Hence, the signal energy conservation based traditional methods are unable to recognize such noise, which contain relatively small energy. The independent component analysis (ICA) brings a different strategy in dealing with the problems of blind source separation (BSS). This dissertation focuses on the study of the acoustic signals generated from a test diesel engine using the ICA in an effort to identify its noise sources. The principles, properties and implementation issues of the ICA are presented. Then research on acoustic field analysis for a single engine and 493ZQ diesel engine was made and illustrates the separation results from the measured acoustic signals.

  • 【网络出版投稿人】 天津大学
  • 【网络出版年期】2006年 11期
节点文献中: 

本文链接的文献网络图示:

本文的引文网络