基于声振信号的激励源重构方法研究

发布时间：2018-04-29 22:28

本文选题：振动 + 声辐射　；参考：《哈尔滨工程大学》2016年博士论文

【摘要】：振动噪声是设备状态监测与故障诊断中常用的重要参数。振动噪声是设备内部激励源与声振传递路径共同作用的结果,不同安装条件下同一设备同一工况下的振动噪声响应不同,这给设备状态监测与故障诊断特征分析和阈值判定带来了困难。本文拟剥离振动噪声传递途径的影响,重构激励源,从激励源特征来判定设备状态,为设备状态监测、故障诊断及振动噪声控制提供新的方法与手段。激励的频域识别法一般要求系统是线性的,且难以有效辨别激励是随机还是脉冲形式。本文提出了基于逆系统的激励源识别的时域方法,该方法采用横向滤波器模拟逆系统,通过自适应算法调整滤波器的权向量,数值和实验结果表明,该方法不仅可以识别谐频激励源,还可以重构随机激励源和脉冲激励源。该识别方法原理简单,不需要对系统特性矩阵进行求逆运算,避免了病态矩阵问题。该识别方法还不依赖于系统状态空间模型,即使当缺少系统先验知识时,该方法较现有逆系统方法具有较大优势。针对工程中不适宜安装接触式振动传感器的工况,本文提出了采用空气声信号重构激励源的方法。该方法以辐射声压作为拾取信号,采用近场声全息技术重构弹性结构表面的法向振速,进而根据重构振速求解作用于弹性结构的激励。对自由声场中的薄板结构进行了可行性探索,推导了基于近场声全息技术的激励重构理论,分析了重构激励的唯一性,并分别从解析及数值方面对作用于薄板的面力、线力、点力以及组合力进行了重构。本文研究了全息孔径大小、传声器间距、全息面距离和结构模态阶数等参数对激励重构精度的影响。数值结果表明:低频时,增大全息孔径、减小全息面距离均能有效提高重构精度;高频时,全息孔径大小、全息面距离对重构精度无影响;减小传声器间距能有效提高重构精度;重构采用的结构模态阶数越多,则重构精度也相应提高。针对拾取声信号中的随机干扰,本文采用了一种基于广义交叉验证准则的改进Tikhonov正则重构激励方法,结果表明,该方法能提高激励源重构精度。最后,本文研究了激励源、结构振动以及声辐射三者之间的相互关系,阐明了结构自身动力学特征及辐射效率也是影响采用声重构激励精度的重要因素。
[Abstract]:Vibration noise is an important parameter commonly used in equipment condition monitoring and fault diagnosis. The vibration noise is the result of the interaction of the internal excitation source and the acoustic and vibration transfer path of the equipment. The vibration and noise response of the same equipment under the same working condition is different under different installation conditions. This makes it difficult to analyze the characteristics of equipment condition monitoring and fault diagnosis and to determine the threshold value. This paper intends to peel off the influence of vibration and noise transmission path, reconstruct the excitation source, judge the equipment state from the characteristics of the excitation source, and provide a new method and means for equipment state monitoring, fault diagnosis and vibration and noise control. The frequency-domain identification method of excitation generally requires the system to be linear, and it is difficult to distinguish whether the excitation is random or impulsive. In this paper, a time-domain method of excitation source identification based on inverse system is proposed. In this method, the transversal filter is used to simulate the inverse system, and the weight vector of the filter is adjusted by adaptive algorithm. The numerical and experimental results show that, This method can not only identify harmonic excitation sources, but also reconstruct random excitation sources and impulse excitation sources. The method is simple in principle and does not need inverse operation to the characteristic matrix of the system, thus avoiding the ill-conditioned matrix problem. The recognition method does not depend on the system state space model, even when the prior knowledge of the system is lacking, the method has a great advantage over the existing inverse system method. In view of the unsuitable condition of installing contact vibration sensor in engineering, this paper presents a method of reconstructing excitation source by using air acoustic signal. In this method, the acoustic pressure is used as the pickup signal, and the normal velocity of the elastic structure surface is reconstructed by using the near-field acoustic holography technique, and then the excitation of the elastic structure is solved according to the reconstruction velocity. The feasibility of thin plate structure in free sound field is explored. The theory of excitation reconstruction based on near-field acoustic holography is deduced, and the uniqueness of reconstruction excitation is analyzed. The surface force and line force acting on thin plate are analytically and numerically analyzed. The point force and the combined force are reconstructed. In this paper, the effects of holographic aperture size, microphone spacing, holographic plane distance and structural modal order on the precision of excitation reconstruction are studied. The numerical results show that the reconstruction accuracy can be improved effectively by increasing holographic aperture and decreasing holographic plane distance at low frequency, while at high frequency, holographic aperture size and holographic plane distance have no effect on reconstruction accuracy. Reducing the distance between microphones can effectively improve the reconstruction accuracy, and the more the structural modal order is used, the higher the reconstruction accuracy will be. In this paper, an improved Tikhonov canonical reconstruction excitation method based on generalized cross validation criterion is proposed for random interference in pick-up acoustic signals. The results show that this method can improve the reconstruction accuracy of excitation sources. Finally, the relationship among excitation source, structural vibration and acoustic radiation is studied. The dynamic characteristics of the structure and the radiation efficiency are also important factors affecting the excitation accuracy of acoustic reconstruction.
【学位授予单位】：哈尔滨工程大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TB535

【相似文献】