有限测点下结构复合反演的混合法

Multiple inversion method with responses on limited degrees of freedom in time domain and frequency domain

回顾了近年来输入信息未知条件下结构识别的复合反演算法,并对输出信息测试不完备条件下的复合反演问题做了深入的研究,提出了时频域混合的复合反演算法。该算法首先在时域内利用子结构上的测试信息反演未知输入,然后在频域内依据反演得到的未知输入和有限测点上的响应信息构造目标函数来识别结构参数。该算法适用于具有多个未知输入的情况,与仅适用于单个未知输入条件下的复合反演算法相比,具有更广泛的应用价值。最后用一数值算例说明了本文算法的可行性。

Multiple Inversion method in recent years is reviewed. And then, a new method based on time domain and frequency domain for both input and system identification is put forward. This method does not require responses of all dynamic degrees of freedom and can be used to the situation while multiple inputs acting on structure are unknown. This method can be divided into two steps. In the first step, by using limited responses, the multiple inversion based on sub-structure is applied. And then in the second step, by using the inputs obtained before, an optimization problem is solved and through which the vest structural parameters are identified. Compared with other method, that is useable only for solving one unknown input information （such as ground motion）, this method has a wider range of application background. At last, a numerical example is employed to show the feasibility of this method.