【作者】 时岩；

【导师】 常思勤；

【作者基本信息】 南京理工大学 ， 车辆工程， 2006， 博士

【摘要】 本文针对车辆及其他机电产品降噪和故障诊断的技术需求，对基于声强测量技术的噪声自动分析系统进行了深入的研究。在噪声自动分析系统的总体方案设计、硬件设计、软件设计、声强信号处理与算法研究及系统实际工程应用等方面本文做了全面、细致的工作，所研制的噪声自动分析系统已获国家实用新型专利，为汽车等机电产品的噪声控制、质量监测及故障诊断等提供了有效的分析工具和手段。噪声自动分析系统是用于自动分析与监测车辆及其他机电产品运转噪声的机电一体化设备，涉及声强测量技术、虚拟仪器技术、计算机控制与分析技术、信号处理技术及人工智能技术等的应用与集成。它针对现有噪声分析设备自动化、智能化程度不高等不足之处，建立了以计算机为控制中枢的测控平台。通过计算机控制下的二维自动导轨的运动带动声强传感器在竖直二维平面内快速移动、准确定位，从而对测量面上若干测点进行自动扫描，并将扫描过程中所拾取的噪声信号经动态信号采集卡的调理和数模转换后输入计算机。在计算机中通过软件算法进行声强的计算与分析、自动绘制声强分布云图、噪声源的自动辨识、异常声响的自动判定以及故障的自动识别等工作。整个测量过程具有测量速度快、定位精度高及声强计算准确等特点，在噪声源辨识与分析和基于噪声的故障诊断等方面自动化、智能化程度较高。本文首先对噪声自动分析系统的总体方案进行了设计，满足其自动化、智能化要求；基于总体方案的设计要求，分别对系统的硬件、软件进行了详细的设计工作，并对系统进行了标定与调试；在此基础上，成功地实现了整个噪声自动分析系统，为进一步的研究工作奠定了坚实的基础。运用基于互谱分析的声强算法，对噪声自动分析系统在噪声监测和主噪声源识别方面的性能进行了研究。研究结果表明：在噪声监测方面，系统具有快速自动扫描的能力，满足了噪声监测对快速化测量的要求；在主噪声源识别方面，通过运用复合型优化算法，系统可以在计算机的控制下对主噪声源进行自动优化搜索，提高了主噪声源的识别定位精度。基于互谱分析的声强算法是分析平稳噪声信号的有效工具，但对于短时非平稳噪声信号其分析能力有所不足。本文针对此问题研究了基于STFT谱的声强联合时—频分析方法，该方法可以将非平稳的一维噪声信号以二维的时间频率函数形式表示出来，有效地提取噪声信号随时间和频率的变化特征，从而分析噪声信号非平稳性的成因。利用上述特性，本文对某型号发动机加速噪声的非平稳性进行了分析。针对汽车等机电产品所辐射的噪声普遍存在非平稳的瞬时冲击噪声的特点，在分析小波和小波包变换原理的基础上，研究了一种单小波包重构算法，并利用该算法对瞬时冲击信号的检测和特征提取进行了仿真分析。在此基础上，研究了一种基于小波包分析的声强算法，并对其声强测量精度进行了验证。基于小波包分析的声强计算方法融合了小波包分析和声强两种技术，可有效地分析故障发生时瞬时冲击噪声的反常现象，为基于噪声信号的故障诊断提供了一种有效的方法。本文利用该方法对某型号发动机单缸失火故障诊断指标的选取问题进行了分析。针对传统声强测量中容易受到强大背景噪声干扰的问题，本章从软件和硬件两个方面对提高声强测量抗干扰性的方法进行了探索性研究。在软件方面，将选择性声强技术应用到噪声源辨识当中，可以将目标噪声源从强大的背景噪声中有效地识别出来。在硬件方面，通过一种抗干扰屏蔽罩的设计，可以在一定程度上屏蔽某些方向上非目标噪声源辐射声波的干扰，提高了对目标噪声源的辨析能力。本文的研究着重于噪声自动分析系统解决实际应用问题的能力，所进行的研究工作都进行了相应的实验分析，验证了其可行性。本文最后针对某型号微型车车外辐射噪声较大的问题，利用噪声自动分析系统对其车身外表面进行了声强测试，识别出了其主要噪声源，解决了此工程实际问题。更多还原

【Abstract】 Aiming at noise reduction and fault diagnosis of vehicle and other mechatronicsproducts an automatic noise analysis system is studied based on sound intensity technologyin this paper. The research of the paper includes scheme design, hardware design, softwaredesign, sound intensity processing algorithm, engineering application and so on. Theautomatic noise analysis system provides a tool for noise control, quality monitoring andfault diagnosis of vehicle and other mechatronics products.The automatic noise analysis system is a mechatronics equipment used to analyzeand monitor noise of vehicle and other mechatronics products in working. It comes downto sound intensity, virtual instrument, computer control and analysis, signal processing andartifical intelligent technologies. Existing noise analysis equipments have insufficiencieson automatization and intelligentization. So the automatic noise analysis system usescomputer as the central workbench. By a computer controlled planar robot sound intensityprobe can be drived to scan measurement plane automatically with high moving speed.Then the sound intensity probe can scan the measurement plane automatically and measurenoise signals simultaneously. The measured noise signals are transmited into computerafter conditioned and acquired by the dynamic signal acquisition card. In the computer themeasured noise signals are analyzed and processed by the software algorithm such assound intensity computing and analyzing, plotting equal intensity contour plans, noisesource identification, distinguishing abnormal sound, fault diagnosis and so on. Hige speed,exact location precision and accurate sound computing are characteristic of the system. Thesystem can be used for noise source identification and fault diagnosis with automatizationand intelligentization.The goal of the scheme design of the automatic noise analysis system is to realizeautomatization and intelligentization. Based on above scheme the hardware and thesoftware of the system are designed detailedly. Calibratation and debugging is alsocompleted. Consequently the whole system is realized. The realization of the automaticnoise analysis system lay the foundation for the further research.Using cross spectrum based sound intensity processing algorithm noise monitoringand main noise source identification is studied in the paper. On the aspect of noisemonitoring the automatic noise analysis system has the ability to scan noise source fleetlyand automatically. Using multiple model optimization algorithm the system can search for main noise source controlled by computer and the location precision is high.Cross spectrum based sound intensity processing algorithm is effective on stationarynoise signals. But it can’t deal with nonstationary noise signals. Aiming at this problem ajoint time—frequency analysis method is studied in the paper based on STFT spectrum.This method can express noise signals as time—frequency form. The variety characteristicof noise signals along with time—frequency can be extracted by this method. Soundintensity joint time—frequency analysis method based on STFT spectrum provides a toolto analyze nonstationary characteristic of noise signals. The acceleration noise of an engineis analyzed with the method.Nonstationary instantaneous shock noise signals is ubiquitous in vehicle and othermechatronics products. Aiming at this problem a single wavelet packet reconstructionalgorithm is studied in the paper based on theory of wavelet transform. Using thisalgorithm a simulated instantaneous shock signal is analyzed to extract its instantaneouscharacteristic. Upon above work a sound intensity computing algorithm is studied based onwavelet packets analysis and the precision of this method is validated. The sound intensityalgorithm based on wavelet packets analysis incorporates sound intensity technology withwavelet packets technology so it can analyze abnormal phenomena of instantaneous shocknoise signals when faults take place. Meanwhile a tool for fault diagnosis is provided basedon noise signals. Using this method this paper analyzes the chosen of fault diagnosis indexof single cylinder misfire.In the sound intensity measurement the interference of background noise is a realmatter. From the aspects of software and harkware the anti-interference method is studiedfor sound intensity measurement. Using selective sound intensity technology the targetnoise source can be identified from strong background noise. An anti-interference shield isstudied, which can screen sound wave of background noise so the ability of sound intenstyto discriminate target noise source is improved.The research of the paper emphasizes on the engineering application. The feasibilityof researches in the paper is confirmed through experiments. To reduce exterior noise of aminibus the automatic noise analysis system is used to measure sound intensity of itssurface noise. According to the test result the maim noise source is separated, whichprovides reference for noise control of the minibus.更多还原