【作者】 杨勇；

【导师】 周丽；

【作者基本信息】 南京航空航天大学 ， 工程力学， 2010， 硕士

【摘要】 与广义卡尔曼滤波、最小二乘等方法相比,序贯非线性最小二乘方法具有输入数据少、滤波稳定、计算效率高和易于实际操作的优点。但是对于复杂结构,尤其是自由度较多的结构,仍然存在大量数据需要处理,计算效率低下的问题。为了使该方法能够用于复杂结构,可以利用子结构分析方法和模型缩减方法来减少复杂结构的自由度数,实现该结构的损伤识别研究。本文利用基于子结构分析方法和模型缩减方法的序贯非线性最小二乘方法对杆板结构和梁板结构进行了仿真研究,通过比较自由度减少前后可知,利用子结构分析方法和模型缩减方法大大的减少了计算量,提高了计算效率。加筋板结构由于其强度高,在航空航天领域中被广泛运用。为了模拟加筋板结构的动态特性并且在损伤发生后能够立即对它的健康状态进行正确的评估,首先要为加筋板结构建立一个合理的有限元模型。本文以壳单元和考虑剪切自锁效应的梁单元为基础,建立了加筋板结构的有限元模型。作者首次将序贯非线性最小二乘方法运用于加筋板结构的参数估计和损伤识别中,加筋板结构的仿真结果证明了该方法在加筋板结构损伤识别中的可行性和准确性。最后,本文进行了两个实验。第一个实验通过对比有限元模型和锤击法实验得到的结构的固有特性验证了有限元模型的合理性,第二个实验通过比较有限元模型和利用序贯非线性最小二乘方法得到的加筋板结构梁单元的等效刚度来证明序贯非线性最小二乘方法在加筋板结构损伤识别中的有效性和准确性。更多还原

【Abstract】 Contrary to the extended Kalman filter and the least square estimation, the sequential non-linear least square estimation has significant advantages in terms of the stability and convergence of the solution as well as the computational efforts involved. For the complex structure, especially the m-dof structure, the technique also has many disadvantages in terms of the computational efforts and efficiency involved. In order to apply the technique to the complex structure, we can make use of substructure analysis method and model reduction method to reduce the degree of freedom of the complex structure. Based on substructure analysis method and model reduction method, the sequential non-linear least square estimation is used to simulate rod-plate structure and beam-plate structure. Simulation results demonstrate that the computational effort and efficiency can be largely improved by using substructure analysis method and model reduction method.Stiffened plate is widely adopted in the aerospace structure for its high strength. In order to simulate the dynamic characteristics of the stiffened plate and to evaluate the possible damage of stiffened plate during the normal operation or right after a major disaster, it is important to develop a suitable finite element model. Based on shell element and beam element which considers shear self-lock effect, the paper establishes a finite element model. The accuracy and effectiveness of the approach have been demonstrated using simulation results. Finally, the paper establishes two experiments. According to compare the difference of the natural characteristics between finite element model and the experiment, the first experiment demonstrates the rationality of the finite element model. To compare the difference of equivalent stiffness between the experiment and the theoretical value, the second experiment demonstrates the efficiency and accuracy of the approach in damage identification of stiffened plate.更多还原