基于反应力向量灵敏度的模型参数化方法

模型参数化是建立识别方程的关键。本文在虚拟结构向量转换方法的基础上提出了“基于反应力向量灵敏度的模型参数化方法”。该方法首先将反应力向量对单元物理参数作一阶展开，对线性参数系统来说，即为反应力向量灵敏度与单元物理参数的乘积，其次，在线性参数系统的假定下，对一个单元中包含多个单元物理参数的情况提出了一个简便的求反应力向量灵敏度的算法，利用该方法可直接由常规的有限元代码获得反应力向量灵敏度矩阵，而无需仿制专门的有限元代码集成分析代码，文中通过算例阐述了该方法的应用。

Parameterization Method Based on Sensitivities of Response Force Vectors

Parameterization of dynamic equilibrium equation is an important procedure to establish the idertification equation.A parameterization method based on sensitivities of response force vectors is presented in the paper.Derived on the basis of virtual structure vector transforning method,the presented method expends the response vectors with respect to elemental parameters using first order Taylor expansion.With regard to linear parameters system the expansion results in the multiplication of sensitivities of response vectors by element parameters.Under the assumption of linear parameters system,the presented method can handle easily the situation when there are several parameters in one element.The method can be implemented conveniently when using available finite element analysis source codes.Some examples are given in the paper to illustrate the method.