【作者】 王丹；

【导师】 罗文波；

【作者基本信息】 哈尔滨工业大学 ， 人机与环境工程， 2006， 硕士

【摘要】 本文在研究大量相关文献的基础上,对非线性系统的辨识方法进行了综述,其中着重描述了频域非线性系统辨识方法,这些频率结果比相应的时域结果有更重要的意义,且它们易于计算和解释。现今铝蜂窝夹层结构广泛应用于工程领域,但由于铝蜂窝夹层结构复杂,使得应用传统的辨识方法难以对其动力学特性给出清楚和完整的描述,本文的目的就是应用随机数据的频域非线性分析与辨识方法对铝蜂窝夹层结构的动力学特性进行辨识。首先,系统总结了使用一般三阶非线性模型辨识的方法和步骤。得到的结果是用基本的和高阶的谱密度函数以及基本的和高阶的频率响应函数表示的频率函数。其次,深入阐述了前线性系统与后线性系统的特点与计算方法。对于这些系统它们的高阶频率响应函数将成为单一的频率变量函数。最后,将随机数据的频域非线性分析与辨识方法应用于某铝蜂窝夹层结构多功能电子机箱的非线性特性研究中。利用前线性系统与后线性系统中的6个模型来分别拟合该系统,从中选取辨识相对误差最小的模型来描述该非线性系统。这种方法能够更准确的分析非线性系统特性,得到该系统的非线性阶数及各阶非线性的频率响应函数,实现了对被试机箱的非线性系统辨识。在此基础之上,从结构材料与结构形式的角度进一步讨论了非线性产生的原因。利用本文的方法,实现了对铝蜂窝夹层结构动力学特性的辨识,这对铝蜂窝夹层结构的设计,实验和分析具有一定的实用价值。更多还原

【Abstract】 The identification approaches of nonlinear systems, especially the approaches in frequency domain, are discussed based on investigation of related articles. The frequency results are far more significant and easy to calculate and explain than the ones in time domain. The honeycomb sandwich structure has been widely utilized in engineering field nowadays. However, due to the complex construction of this structure, it seems not quite easy to describe its dynamic characteristic clearly and completely via conventional methods. The purpose of this thesis is to analyze and identify the dynamic characteristic of honeycomb sandwich structure based on nonlinear methods to stochastic data in frequency domain.First, the methods and procedures of general three orders nonlinear identification are summarized and the results are frequency functions described by basic and high order spectral density functions and frequency response functions.Secondly, the characteristics and procedures of before linear systems and after linear systems are discussed. It is found that the high order frequency response functions of these systems will become single frequency variable functions.Finally, the nonlinear identification method to stochastic data in frequency domain is used to analyze and identify the nonlinear characteristic of some multifunctional electrical case with honeycomb sandwich structure. This system is fitted separately by six models of before linear system and after linear system and then the optimal model with minimum relative error is chosen to describe the system. This method can not only analyze the system characteristic accurately, but also get the order and nonlinear frequency response function in every order, therefore realizes the nonlinear system identification of the case under study. The causes of nonlinear are also discussed later based on these results from the view of structure material and form.To identify the dynamic characteristic of honeycomb sandwich structure via method mentioned in this thesis can be valuable in practice for further design,更多还原