【作者】 陆冬；

【导师】 汤宝平；

【作者基本信息】 重庆大学 ， 机械电子工程， 2007， 硕士

【摘要】 试验模态分析技术是基于测试实验建立和描述结构振动特性的数学模型的过程。它提供了识别结构动力学特性参数的重要方法,在机械结构设计和结构修改中起着重要的作用,而试验模态分析的关键之处就是模态参数识别。因此,如何提高模态参数识别的精度以便更好的为结构设计和修改服务,一直是工程界积极探讨的课题。本文对试验模态分析基本原理进行了深入研究,从提高参数识别精度出发,分析了模态参数识别精度的影响因素。根据噪声对频响函数估计精度的影响,提出了频响函数估计的最小二乘优化算法;并分析了不同参数识别方法对参数识别精度的影响。同时,研究了应用最为广泛的分量分析法和正交多项式拟合法两种识别方法,并分别实现了两种方法的算法,通过实例对比了两种识别方法的优劣。为了得到更高精度的模态参数,作为尝试,本文还初步研究了基于小波去噪和HHT变换的模态参数识别方法,实例表明,该方法能够得到更高精度的模态参数。在研究了试验模态分析基本原理,以及国内外试验模态分析软件的基础上,采用面向对象的程序设计语言Visual C++开发了机械结构试验模态分析的软件系统。根据试验模态分析系统的构成和主要功能把整个软件系统划分为四大功能模块---结构模型的建立模块、数据采集模块、频响函数分析模块以及参数识别模块。软件采用多视图设计思想,一个功能模块对应一个视图,每个视图有自己独立操作的菜单和工具条。该软件系统提供了三种试验方法即单输入单输出的变激励方法和变响应方法以及单输入多输出的方法,参数识别模块提供了分量分析法和正交多项式拟合法,并预留了增加新的参数识别算法的接口。为了验证试验模态分析系统的正确性,对平板结构进行试验模态分析,并把模态参数识别结果与ME’scopeVES软件识别结果作对比,对比结果表明软件的设计不仅正确而且参数识别精度和ME’scopeVES软件识别精度一致。最后通过对振动台简支梁的实测试验,采用本软件进行了简支梁试验模态分析,得到了简支梁的固有频率、阻尼比和振型。更多还原

【Abstract】 Experiment modal analysis is the process to establish mathematic model of describing structure’s vibration characteristic based on experiment. It provides the key method for the dynamic’s characteristic parameters identification. And it plays the important role in structural designing and modifying. While the key step in the experiment modal analysis is modal parameters identification. So, How to improve the accuracy of the results of modal parameters identification in order to provide better serves to the structural designing and modifying is now being researched by many engineers.This paper has done many researches on the basis of the theories of modal analysis. In order to improve the accuracy of estimation results of frequency response function, this paper analyses some factors which can affect the accuracy of the results of modal parameters identification. According to the noise influence of the frequency response function, this paper puts forward an arithmetic for the optimization of frequency response function by using least-square method, and also analyses different parameters identification methods which can also affect the accuracy of the results of modal parameters identification. Based on these, two methods, component analysis method and orthogonal polynomial fitting method, which are widely used, have been done some researches, and arithmetic has been carried out. An experiment had been made to compare the two methods. To obtain the better accuracy of the results of modal parameters identification, this paper also dose some pilot study on modal parameters identification based on HHT and wavelet de-noising. The experiment shows that this method can improve the accuracy better.Based on the research on the basic theories of experimental modal analysis and analyzing the current research on corresponding software both at home and abroad, an experimental modal analysis software system is studied and developed by Visual C++, which is an object oriented programming software. According to the constitution of experimental modal analysis system and its main function, the software system has been divided into four function modules, which are structural building module, data acquisition module, frequency response function’s analysis module, and parameters identification module. The software system adopts multi-view idea, which has four views and one module one view. Each view has its own menu and toolbar. Three test methods have been provided in the system, that are single input single output and single input multi-output, and single input single output has two forms, changing excitation point and changing response point. Parameters identification module gives two methods which are component analysis method and orthogonal polynomial fitting method, and it also has interface for any other methods.In order to test the experimental modal analysis system’s correctness, a plane’s experimental modal analysis has been done by this system and compared the results with the software ME’scopeVES. The compared results show that the system has the same accuracy with the software ME’scopeVES. At the end, an actual measurement of freely supported beam in vibration platform has been done to identify its parameters using this system, and the freely supported beam ’s freedom frequencies, damping ratios and vibration mode have been identified.更多还原