【作者】 吕中亮；

【导师】 杨昌棋；

【作者基本信息】 重庆大学 ， 工程力学， 2010， 硕士

【摘要】 随着大型工程机械工作环境的复杂化,其结构所受到的激励载荷也变得多样化,激励载荷由传统的窄带、稳态信号激励已逐渐走入了宽带、非稳态信号激励,对于宽带、非稳态信号激励下结构的动态响应分析也已经变得越来越迫切;另外,随着各种测试设备精度的大幅提高,结构在这种复杂激励下动力响应的测量也变为了可能。计算机技术的发展,使有限元计算越来越多的应用于各种动力学领域的模拟、分析和计算中。文章第一部分研究了各种常用的数值方法对于宽带、非稳态信号激励下的结构响应计算的适应性,并得出以下结论:其对于窄带、稳态信号激励的计算能够得到高精度的结果,然而对于宽带、非稳态信号激励下的计算,还有很多值得改进的地方,有时甚至会出现错误的结果。针对这种问题,提出了基于信号分解和重组的方法,即把宽带、非稳态信号分解成不同频率段的窄带信号,而不同频率段的窄带激励信号采用不同的计算方法,分别得到其响应,进而得到原宽带、非稳态信号激励下结构的响应。小波理论是近年来兴起的一种崭新的信号分析理论,一种可达到时间域或频率域局部化的时频分析方法,被称为数学上的“显微镜”。在文章第二部分,研究了如何利用小波变换的方法把宽带、非稳态信号分解成不同频率段的窄带、稳态信号。通过小波分解后不同频率段的信号,可以分别通过不同的计算方法求得其响应。对于高频信号还可以展开写成高频小波基的线性组合,利用系统对高频小波基的响应很快就可以得到系统的高频响应,进而通过叠加法原理就可以得到系统的响应。文章的第三部分研究了在宽带、非稳态信号激励下,结构动力响应的数值模拟计算中时间步长的选取原则,以及把高频信号激励下的响应转化成对高频小波基的响应时,为了保证计算精度,计算时间截断的选取原则。通过计算发现,对于宽频、非稳态载荷激励的响应计算,文中提出的方法在满足高精度要求的前提下,能够大大提高计算效率。文章最后,通过实验验证了:在宽频、非稳态信号激励下,通过小波分解的方法把激励信号分解到不同的频段上,进而通过叠加法得到原线性系统的响应的方法是确实可行的。更多还原

【Abstract】 Along with complication of large-scale engineering machinery’s working environment,exciting load of its structure has become diversification.Signals of exciting load has become wideband and non-stationary from narrow band and steady in tradition.The dynamic responseof structure in the excitation of wideband and non-stationary signals becomes more and more urgently. In addition, with a substantial increase in accuracy of test equipment, the measurement of dynamic response of structure in this complex excitation load signal also becomes possible.With the development of computer technology, more and more finite element method applied to the variety simulation, analysis and calculation of dynamics.The first part discuss a variety of numerical methods commonly used for broadband, non-stationary signals excited by adaptive structural response calculations, and comes to the following conclusions: for the band, the calculation of steady-state excitation signal can get highly accurate results, however, for broadband and non-stationary signals incentive, it is difficult to get good results, even an error result. It is proposed based on signal decomposition and restructuring methods. The broadband, non-stationary signal is decomposed into narrowband signals of different frequencies, different frequency of excitation signals with different methods of calculation, the response and the original broadband, non-steady-state response excitation of the structure has obtained.Wavelet theory is the rise in recent years, a new signal analysis theory, one can achieve the time domain or frequency domain analysis of time-frequency localization method, known as the mathematical "microscope." In the second part, the research studies how to use wavelet transform method decompose broadband, non-stationary signal into different frequency bands of narrow-band, steady signal. Through wavelet signals’decomposition of different frequencies can be separately obtained by different methods of calculation of its response. The high-frequency signal also can be written to a linear combination of high-frequency wavelet basis, using the system response on wavelet basis, high-frequency response of the system can be quickly obtained. And then through the superposition principle the response can be measured.The third part research on with a broadband, non-stationary signal excitation, structural dynamic response of the numerical simulation time step selection principles, and the high frequency signal excitations translate into high-frequency response of wavelet bases, in order to ensure accuracy, computation time truncation selection principle.Finally, validated by experiment: in the excitation of non-stationary broadband signals, through the wavelet method, decompose the excitation signal into different frequency bands, and then obtained by superposition of the response of the linear system is indeed feasible .更多还原