【作者】 廖庆斌；

【导师】 李舜酩；

【作者基本信息】 南京航空航天大学 ， 车辆工程， 2008， 博士

【摘要】 车辆乘坐室的振动和噪声问题已成为制约其快速发展的一个重要方面。目前,车辆中振动和噪声信号的特征提取方法以及振动响应的评估手段都还很不成熟,因此深入的研究车辆中振动和噪声信号的特征提取方法以及振动响应的评估手段对车辆中振动和噪声的控制具有重要的理论意义和工程价值。本文针对车辆工程中NVH的研究内容,对车辆振动和噪声信号的特征提取与结构振动响应的统计估计方法进行了深入的研究,本文研究的主要内容归纳如下:(1)提出了一种适用于强噪声背景下振动信号特征提取的方法。针对目前应用于车辆中信号的特征提取方法,根据车辆中振动和噪声信号的特点,对它们进行对比研究,提出了一种名为时序多相关-经验模式分解的方法,该方法对强噪声背景下,车辆中非线性、非平稳信号的特征提取具有较好的效果,对信号中混叠的噪声类型和强度没有特别的限制,具有很强的适应性;该方法通过对信号的多相关处理后再做经验模式分解,然后对分解得到的本征模态函数做相关性选择,增加的计算量很小。(2)提出了一种分析随机质量板振动响应的方法并对其振动响应进行了统计分析。为了分析随机质量板的振动响应,以Lagrange动力学方程和Rayleigh-Ritz能量法为基础,给出了一种计算随机质量板振动响应的分析方法,对其振动响应进行了统计分析,并应用Monte Carlo数值仿真方法进行了验证。(3)基于本征正交分解方法,提出了相关激励作用时,结构响应的统计估计方法。给出了相关激励作用下,结构响应统计估计的相对偏差的计算表达式,以及相关激励作用下,结构响应统计估计时载荷参数的确定方法。研究表明:响应统计分析与输入激励的相关性及特征频率的相关性直接相关联,随着结构振动频率的增加,响应的相对偏差逐渐减小,考虑激励的相关性及特征频率的相关性得到的相对偏差表达式与试验结果能够更好的吻合。(4)给出了SEA在复杂耦合系统中的响应统计估计方法及其载荷参数的确定方法。研究结构表明:当有载荷作用在子系统上时,载荷参数与子系统上的输入能量和整个系统的输入能量的比值以及子系统上的载荷参数有关;不相关输入是相关输入时的一种特殊情况,它们可以用统一的表达式来表示。复杂系统响应统计估计的试验表明,考虑相关激励的影响,能够使得计算得到的平均能量与试验值更好的吻合,并且随着频率的增加,相关激励的影响逐渐减弱,并且随着传递路径的增加,子系统响应的相对偏差逐渐增大。综上所述,本文较系统地研究了车辆NVH中涉及到的两个方面的内容:车辆振动和噪声信号的特征提取以及振动响应的统计估计方法。研究结果对于提高车辆NVH性能,指导车辆NVH设计以及后期修改等方面具有重要的理论意义和工程应用价值。更多还原

【Abstract】 The vibration and noise problems of passenger compartment in vehicles having been an important restriction for its rapid development. At present, the processing methods of vibration and noise signal and the assessment tools of vibration response in vehicles are still immature. Therefore, the systematical studies on the feature extraction method of vibration and noise signal and the method of the statistical estimation method of vibration response in vehicles has not only important theoretical significance, but also practical value.For the research of noise, vibration and harshness (NVH) in vehicle, a further study into the feature extraction method of vibration and noise signal and the statistical estimation method of structural vibration response have been made in this paper. The main contents of this paper are summarized as follows:(1) A novel method for suppressing the strong background noises in the feature extraction of vibration signal is proposed. According to the characteristics of vehicle vibration and noise signal, after comparing some presently applied feature extraction methods, it proposed a new method: multi-correlation of time series and empirical mode decomposition. Which can eliminate the effect of zero-average noise in sampling series, get over the difficulties of frequency identification in the subsequence analysis that caused by strong background noise, protrude the feature components of original signal, is of preferable for nonlinear, non-stationary vibration signals. This method has no special restrictions for the type and strength of noise, and a strong adaptability. After the processing of multi-correlation of time series, it provides a wonderful former data for spectrum analysis; then from using empirical mode decomposition, it can sufficiently put up the needed feature signal, and plus little increase in computational time.(2) A method which can analyses the vibration response of the variable mass plate is presented. Based on the Lagrange’s dynamics equation and Rayleigh-Ritz energy method, the response analysis method is described and the statistical character is pointed out. It demonstrated the proposed method using Monte Carlo number simulation.(3) Based on the proper orthogonal decomposition (POD), the statistical estimation method of vibration response, when the correlative excitations acting on the structure, is proposed,and the relative deviation calculation expression and calculation method of load parameters in response statistical estimate is given. The results indicate that the statistical analysis of response is directly connected with the relativity of the input excitation and the correlation of nature frequency, the relative deviation decreases gradually with the increment of vibration frequency, the calculation expression of relative deviation which considered the excitation relativity and correlation of nature frequency can in concordance with the result of experimentation better.(4) The statistical estimation method of response and the calculation method of load parameters in the complex coupling system which using statistical energy analysis (SEA), The results show that: the load parameters is connected with the ratio of the input energy of the subsystem and the input energy of the whole system, and with the load parameters of the subsystem; the irrelevant excitation is a special case of correlative excitation, they can be expressed in a unified expression. The experiments of complex system response statistical estimation have shown that: considering the correlative excitations, the calculation results of the average energy are better consistent with the result of experiment. With the increment of vibration frequency, the influence of correlative excitation gradually weakened, when the number of transmission paths increased, the relative deviation of subsystem response gradually increased.In conclusion, the two areas, the feature extraction method of vibration and noise signal and the statistical estimation method of vibration response of NVH in vehicle, are systematically studied in this paper. The results has an important theoretical significance and engineering applications for improving vehicle NVH performance, guiding vehicles NVH designs, as well as the later revisions of the structure.更多还原