【作者】 段虎明；

【导师】 秦树人；

【作者基本信息】 重庆大学 ， 机械电子工程， 2008， 博士

【摘要】 实验模态分析是随着计算机技术、振动测试技术、虚拟仪器技术的发展,从上个世纪60年代开始发展起来的一项重要的工程应用技术。到目前为止,与其相关的理论模型研究,已经较为完善,并逐步转向各种行业的工程应用中。在机械、建筑、航空航天、船舶、汽车等多个领域的应用非常广泛,并且已经日趋成熟。在国内外也出现了多种不同功能、不同层次的模态分析软件。随着科技的进步,很多工程结构对模态分析软件的分析结果精度要求越来越高,然而由于结构本身、传感器、数据采集调理、参数识别等环节都可能引入误差,所以模态参数识别结果的精度往往不高。本文正是基于以上背景,完整地开发了一套虚拟式模态分析仪,并且对其前端的频响函数估计精度问题进行了深入的研究和探讨。文章首先介绍了实验模态分析的相关研究背景和国内外的研究现状,然后较为系统地描述了实验模态分析的基础知识。从物理模型出发,推导出了模态分析的理论模型,从正反两个角度说明了实验模态分析的原理。根据实验模态分析的运作流程,完整地介绍了实验模态分析的过程,包括实验前的实验准备(结构的安装、调试和校准,激励测点、响应测点布置等)、几何结构模型的建立、时域原始振动数据的采集、频响函数的估计、模态参数的识别、模态模型的验证、结果参数和振型动画的输出等。针对提高频响函数的估计精度问题使用两种方法进行了专门的研究和探讨,一种方法是使用频率抽取校正方法首先对采集的原始数据信号进行频谱校正,提高其频谱精度,待各个测试点各个方向的数据都进行校正以后,再使用其进行频响函数估计,然后在频响函数的基础上进行模态参数识别,得到较高精度的模态参数;另一种方法是采用高阶谱的方法,使用高阶谱优良的噪声不敏感性,首先对原始采样数据信号进行双谱估计,再使用双谱重构原始信号的Fourier变换,达到去噪的目的,然后使用重构信号估计频响函数,进而进行模态参数识别。通过仿真数据和实际的简支梁采样数据分别验证了两种方法的有效性,得到了精度较高的模态参数识别结果。文章还重点介绍了QLVMA-2型虚拟式模态分析仪的研制研发过程和其各个部分的功能结构。该模态分析仪采用图形化的显示操作界面,简洁明了、用户操作简单、上手快,真正体现了“虚拟”仪器的特点。该模态分析仪可以实现整个实验模态分析过程,分别为几何结构建模、结构数据采集、频响函数估计、模态参数识别、结果参数和振型动画显示、模态模型验证等。在介绍仪器功能的基础上通过简支梁的实验模态分析、钢板结构的实验模态分析、工字形梁的实验模态分析、刚性框架结构的实验模态分析、摩托车车架结构的实验模态分析等各种形状、大小的结构实验模态分析实例,验证了该仪器实验模态分析结果的正确性和适用性,并且与国内外著名的商用模态分析软件作了对比实验、与ABAQUS有限元分析软件的计算模态参数和振型作了对比实验,都充分的验证了虚拟式模态分析仪分析结果的可靠性和准确性。文章最后还对所做工作进行了总结,并且对下一步研究方向进行了展望。更多还原

【Abstract】 With the development of computer technology, vibration measurement and virtual instrument, the experimental modal analysis has become an important engineering application technique since 1960s. So far, the correlative theoretical mode research has been perfected and gradually turned to the engineering application in various industries. The technique has become more and more popular, especially in the fields of mechanism, architecture, aviation, spaceflight, shipping, and automobile and so on. A variety of modal analysis softwares with different functions and levels have emerged inland and overseas. With the advancement of science and technology, the request of the precision of analysis result to the softwares has become higher in many engineering structures. However, due to the introduction of errors in structure itself, transducer, data acquirement, data processing, and parameter identification and so on, the result is usually not that precise. Hence, on the above situations, this dissertation produces a set of virtual instrument for modal analysis, with a deep and thorough research into the precision problem of frequency response function measurement of front signal processing.The dissertation first introduces the correlative research background and study status about experimental modal analysis inland and overseas, together with a systematical depiction of the basic knowledge about this analysis. Started from mechanics mode, the theoretical mode of modal analysis has been drawn, illustrating the principle of experimental modal analysis from both the positive and negative viewpoint. Based on the operation procedure, a complete process of experimental modal analysis is introduced, including the preparation before the experiment (the allocation, debugging and calibration of structure, the allocation of test nodes about impact and response, etc), the construction of geometrical structure mode, the original data acquisition of time field, the frequency response function measurement, the modal parameter identification, the validity of modal mode, the output of result parameter and mode shape and so on.Two means have been applied to the specific research on the precision of frequency response function measurement. The spectrum correction method of frequency extracting is used to first have spectrum correction of the original data of time field so as to improve the precision of spectrum. After the original data of time field of all test nodes and test directions are corrected, they are then used to measure the frequency response functions where the identification of modal parameters are based and a better precision result is received. The other is the method of higher-order spectral analysis, of which the high quality of noise non-sensitivity is used. After the bispectrum of the original data is measured, it is then used to reconstruct the DFT of the original data in the aim of reducing noise. Finally, the frequency response functions are measured, using the DFT data of reconstruction, and modal parameters are identified. The validity of the two methods has been testified respectively by the simulation result and the actual result of a simple supported beam, where a better precision result of modal parameters is gained.The dissertation also gives an emphasis on the programming process on the functional configuration of the QLVMA-2 virtual instrument of modal analysis. Concise and intelligible, and easy and quick to handle, the instrument adopts a graphical operation interface, embodying perfectly the characteristic of a‘virtual’apparatus. The whole process of experimental modal analysis can be carried out on the QLVMA-2, including constructing mode of geometrical structure, the data acquisition of nodes on structure, frequency response functions measurement, modal parameter identification, display of result parameters and mode shapes, validity of modal mode, and so on. Based on the introduction, the validity and applicability of the QLVMA-2 virtual instrument of modal analysis is confirmed through a series of structure analysis examples, such as a simple supported beam, an armor plate, an I-Beam, a steel frame, and an autocycle frame. Compared with some distinguished modal analysis softwares inland and overseas and Abaqus which is a software of finite element analysis, the testified results also prove the reliability and accuracy of the QLVMA-2.The dissertation finally closes with a conclusion of the writer’s research work and presents an expectation of the following research direction.更多还原