【作者】 杜宪峰；

【导师】 李志军； 毕凤荣；

【作者基本信息】 天津大学 ， 动力机械及工程， 2012， 博士

【摘要】 柴油机的振动噪声控制是一个涉及多个学科的、综合性的、复杂的系统工程。在柴油机振动噪声控制研究领域，振动噪声源及其特性识别、振动噪声源与整机振动噪声的关系是两个最关键的环节。为了达到良好的振动噪声控制效果，需要识别柴油机的主要振动噪声源及其特征，建立振动噪声源与整机振动噪声的关系，并在此基础上进行低振声柴油机结构优化与振声信号能量评价，从而实现在设计阶段预测和控制柴油机振动噪声的目的。论文以天津市自然科学基金重点项目“基于低噪声目标的内燃机噪声源特性识别及结构优化”和国家自然科学基金项目“基于热机耦合的柴油机活塞敲击噪声研究与控制及活塞系统优化设计”为研究背景，以柴油机振动噪声的特征提取、贡献度计算、信号能量特征评价、信号特性预测与机体的传递特性及结构改进设计为主要研究内容，开展以低振声柴油机为目标的柴油机振动噪声源识别与整机振动噪声控制的研究工作。应用时频技术对柴油机振动噪声进行了信号特征提取、信号源识别与主要噪声源贡献度计算，提出了柴油机主要振动噪声源识别与贡献度计算的分析方法；在信号特征提取研究的基础上，采用信号能量特征评价指标分析了柴油机振动信号的主要频率成分，并依据所建立的关于灵敏度因子的计算模型，分析了与供油参数变化密切相关的柴油机振动源及振动源能量特征的变化规律；建立了柴油机整机振动噪声的预测流程，分析了柴油机机体的传递特性并对其进行了以低振声为目标的结构改进设计研究，取得了良好的减振降噪效果。全文的主要研究成果如下：1．在小波分析（WA）与经验模态分解（EMD）的基本原理及算法研究的基础上，结合信号相关性分析、组合模态函数（CMF）方法与小波包技术（WPT），提出了基于EMD与WA协同的信号特征提取方法与分析流程，并指明了该方法在信号特征提取时需注意的问题。柴油机振动信号特征提取的应用研究结果验证了该方法的可行性与有效性。2．针对单通道测量信号在信号源分离与识别方面的不足，以及干扰噪声会严重影响信号分解过程等问题，提出了基于EMD与独立成分分析（ICA）协同的信号源特征提取方法，克服了源信号数不超过观测信号数的假设，以及避免了干扰噪声会造成本征模态函数（IMF）模态混叠的现象。同时，针对柴油机测量噪声的噪声源贡献度问题，提出了基于EMD-ICA盲源分离模型的单一测量噪声的噪声源贡献度计算方法，定义了独立元的贡献度与贡献度矩阵，并通过噪声源贡献度计算的应用研究结果很好的验证了该方法的可行性与有效性。3．建立了基于多体动力学(MBS)－有限元法(FEM)－表面振动速度法的柴油机整机振动噪声仿真分析平台，计算了柴油机主要结构部件的辐射噪声声功率与贡献度的大小，分析了模拟工况下机体振动响应信号及其时频特征，并通过整机平均声压级谱的试验测试结果来验证仿真预测方法的计算精度。同时，在该柴油机仿真平台的基础上，采用传递函数与谱相关技术探讨了机体结构的传递特性，并依据振动源贡献度计算流程确定了影响机体各部位振动状况的主要振动源及大小排序。该仿真平台与分析流程为探讨柴油机激励载荷、机体传递特性、振声信号特征之间的相互关系提供了一个理想的分析研究平台。4．在信号特征提取的基础上，采用能量特征评价指标对柴油机振动信号进行分析，识别出柴油机振动信号的主要频率成分，并采用能量变化指标计算得出了转速与负荷变化对柴油机机体与缸盖测点振动信号能量特征的影响程度。同时，建立了关于灵敏度因子的计算模型，确定了与柴油机供油参数变化密切相关的振动源以及该振动源信号能量特征的变化规律，有助于实现柴油机供油参数的合理选择以及振动噪声控制。5．建立了以低振声柴油机为目标的机体结构改进设计与评价的分析流程。首先以振动速度最小化为目标对柴油机机体结构进行改进设计，并分析各设计参数对振动速度的灵敏度；然后以某四缸柴油机机体为研究对象，探讨拓扑与形状优化方法在低振动柴油机机体结构设计中的应用，并通过试验分析验证计算模型改进的有效性；最后在信号特征提取研究的基础上，建立了基于信号特征的结构改进方案评价方法，用以评价柴油机机体改进前后能量与频率的变化。该分析流程的采用可以方便的用来设计合理的低振动柴油机结构部件，并对设计方案提供可靠的设计评判依据，在提高计算精度的同时改善了设计自动化程度。更多还原

【Abstract】 Vibration noise control of diesel engine is a systematic and complexinterdisciplinary project. In this research area, there are two key parts. The first part isidentification of vibration noise source and its characteristics while the second part isrelation between vibration noise source and vibration noise of the whole diesel engine.In order to better control vibration and noise, based on two parts, structuraloptimization and measuring of vibration noise signal energy distribution have to beapplied to predict and control vibration and noise from the initial design period.Based on Vibration Control and Optimization Design of Coupling Engine Piston,a National Natural Science Foundation of China project and Identification of NoiseSource and Structural Optimization of Gas Engine with the Purpose of LoweringNoise, a Natural Science Foundation Project of Tianjin, concentrating on the featureextraction, calculation of contribution degree, signal energy appraisal, characteristicprediction, transfer characteristic and structural improvement design of block, thispaper carries out research work on source identification and overall control of dieselengine vibration with the goal of creating engine with low vibration noise. Withtime-frequency technique, this paper conducts the feature extraction of vibration andnoise, sources identification and contribution degree calculation, and then renders ananalysis method for vibration and noise sources identification and contribution degreecalculation of diesel engine. Based on signal features extracting research, this paperintroduces in signal energy appraisal index, analyses the main vibration signalfrequency components of diesel engine, establishes sensitivity factor calculationmodel, analyses vibration signal source which is related closely to oil supplyparameters, and the change rule of the vibration source energy characteristics. Thispaper establishes the prediction procedure of the whole diesel engine, analysesvibration transmission and carries out diesel engine structure research with lowvibration noise as the goal, which has been proved to be effective on this virtualplatform. The research results are listed below:1. On the basis of wavelet analysis&empirical mode decomposition principlesand their calculation research, combining signal relativity analysis, combined modefunction and wavelet packet technology, this paper works out signal abstraction method and further points out questions to be noticed during the process. The methodis proved to be practical and effective in the application research of signal characterextraction.2. Aiming at the shortcomings of single channel and low signal noise ration ofdiesel engine in extracting signal source, this paper renders the coordinated signalsource identification method based on empirical mode decomposition andindependent component analysis, overcoming the assumption of less signal sourcethan observed ones and avoiding intrinsic mode function mode mixing due to noiseinterference. Meanwhile, aiming at contribution degree problem of different signalsource, this paper defines independent component contribution degree andcontribution matrix based on blind source separation model based on ICA andEMD-ICA, works out the contribution degree by various noise source which has beentested to be correct.3. This paper builds a set of vibration noise prediction procedure, namelyMultibody dynamics (MBS)-Finite element method (FEM)-Method of ExternalVibration Velocity. Through the analysis procedure, this paper works out the value ofvibration severity and radiation noise acoustic power of main engine parts, analyzesthe relation between excitation and engine vibration signal, which have been provedto be accurate in the test of spectrum of average sound pressure level. Meanwhile, thispaper explores into engine transmission characters using transfer function and spectralcorrelation technique, and calculates vibration source’s contribution degree to engineblock using contribution degree calculation method. The application of the predictionprocedure allows us to predict vibration and noise of diesel engine under differentworking conditions, decreasing the testing work, providing an ideal research platformfor diesel engine excitation, transmission characters and mutual-relation betweenvibration signals.4. Based on diesel engine vibration signal extraction, introducing in signalenergy evaluation index (instantaneous energy, spectral analysis of instantaneous,amplitude acceleration energy, spectral analysis of amplitude acceleration energy andenergy change index), this paper analyses the main vibration signal frequencycomponents of diesel engine, and evaluate the speed and load condition influencing tovibration signal energy feature by the energy change index; meanwhile, builds thecalculation model of sensitive model. this paper analyses and identifies the vibrationsource of diesel engine, which is closely to the supply oil parameters (fuel supply pressure), and according to the influence principles on engine vibration signal energyunder different fuel supply pressure, which will favor proper selection of engine fuelsupply pressure and realize vibration and noise control.5. This paper establishes the virtual simulation platform and structuraloptimization procedure to decrease vibration and noise of diesel engine. Firstly, thispaper carries out multi-target optimization to engine parts based on the slowestvibration speed and then analyzes different design parameters’ influence on vibrationspeed and displacement. Secondly, this paper explores into the application of topologyand shape optimization in diesel engine structural design, which turns to be rational intest. Finally, this paper works out structural characters evaluation methods based onsignal features to evaluate the change of vibration signal energy and frequency beforeand after structural improvement. The procedure enables us to design rationallow-vibration engine parts, enhance calculation accuracy and improves automationdegree of design.更多还原