Model updating method for hybrid simulation based on global sensitivity analysis

The hybrid simulation (HS) is a cost-effective and promising method for investigating the seismic performance of structures. This paper proposes a model updating method for HS to improve incomplete boundary conditions and model error, and develops criteria to select the updated parameters. First, an actual HS model updating is conducted to verify the effectiveness of uniform design as a parameter estimation method. This experiment shows that parameter sensitivity has a significant relationship with the update effect. Then, the Sobol global sensitivity analysis (GSA) method is applied to the HS model updating to obtain the sensitivity indices of parameters that change with time, which provides a sound basis for parameter selection and the determination of the search range. Following this, a whole-structure module for dynamic time history analysis is proposed to improve the boundary conditions, simplify the implementation, and refine the numerical models. The errors in six boundary conditions of the traditional HS are then analyzed. The proposed method of model updating based on GSA is then obtained by integrating the above ideas. In the case of multiple inputs of ground motion, four strategies are developed to verify the effectiveness of the method. The results of a total of 216 numerical cases under different working conditions show that the proposed method can improve the accuracy of HSs and the identified parameters.