动载荷时域Wilson-θ识别方法和PATRAN二次开发-工程力学专业论文.docx

摘 要 由载荷识别理论可知，用广义正交多项式拟合载荷函数的方法可以识别出结构的分布激励， 如一维的线分布和二维的面分布。以一维梁和二维薄板模型为例，推导出相应的时域识别数学 模型和识别原理。 PCL 语言为 PATRAN 用户提供了很好的二次开发工具。本文根据上述的分布载荷识别理论， 结合开发的 PATRAN 界面的窗口输入必要参数，在软件中嵌入一个动载荷时域识别的模块，实 现了该方法在 PATRAN 软件的集成，同时，根据输入的参数说明和操作流程，分别给出相应的 仿真算例，验证该模块的操作性和识别精度。 Wilson-θ 法作为一种常用的数值算法，当 θ 值大于 1.4 时，可以保证计算结果绝对收敛。 根据这一特性，通过反推该方法得到了一种基于 Wilson-θ 反分析法的载荷识别方法。同时，由 模态叠加原理和单位脉冲法推导出第二种载荷识别方法，通过相应的算例分析比较了它们的识 别精度、发散性以及抗噪性能。结果表明，较传统的时域方法这两种方法都具有更好的识别精 度、收敛性以及抗噪性。 关键词：载荷识别，时域，分布动载荷，模块，Wilson-θ 法 i  ABSTRACT According to the load identification theory，the distributed dynamic load of structures can be identified by the function fitted with the generalized orthogonal polynomial, such as a line-dimensional distribution and two-dimensional surface distribution. This paper takes one-dimensional beam and two-dimensional sheet model for example, derive the mathematical model and theory of time-domain identification. The software PATRAN provides PCL language for their users as a good secondary development tools. Based on the above theory, combining with the developed PATRAN interface window for the use of entering the necessary parameters, a dynamic load module of the time-domain identification is embedded and integrated in the software. At the same time, the instructions of the input parameters and operating procedures are illustrated, and the corresponding simulation examples are given respectively to verify the modules operational and accuracy. Frequently used as a numerical algorithm, the Wilson-θ method can ensure that the results converge when the value of θ is greater than 1.4. According to this, a counter-based method of load identification is obtained by the backstepping of the method of Wilson-θ. What’s more, another load identification method was derived from the mode superposition principle and the unit impulse theory, and their recognition accuracy, divergent and anti-noise performance are analyzed by the comparison with the corresponding examples. The results sh