An accurate numerical model that is able to represent real structural behaviors and reproduce structural responses with high fidelity is critical to numerous engineering applications such as damage detection, diagnosis, and prognosis, and data assimilation. While a wide variety of methods have been developed in the past decades for finite element model updating, a widely adopted concept is to solve a constrained optimization problem to minimize the prediction errors, such as those formulated by modal properties, frequency response functions, and static and dynamic responses, among others. This paper considers model updating for earthquake-excited building structures using incomplete acceleration measurements. In this case, due to the transient nature and limited duration of earthquake responses, as well as incomplete instrumentation and measurement noise, identifying accurate and adequate number of modal parameters for model updating is challenging. Furthermore, in presence of static nonlinear functions, model updating is a nonlinear inverse problem solved using nonlinear optimization methods. This often faces divergences due to issue of nonconvexity, resulting in unreasonable parameter estimations. Therefore, identifying an accurate model of building structure under earthquake excitation is still a challenge. In this paper, we propose a maximum a posteriori (MAP)- based approach using measured earthquake response time histories, which renders model updating as a regularized nonlinear optimization problem. The prior knowledge of structural parameters is incorporated to constrain the estimation. One main advantage of the proposed approach is that it makes no assumption of the upper and lower bounds while ensuring the physical meaning of the structural parameters. The proposed approach is validated through a numerical example.
|