Recursive Hybrid GA and Moving Window Hybrid GA to Parameter Identification of Structural Systems with Passive Control Devices

It is intended to identify the change of parameters of a structural model with passive control devices. Two methods were proposed to the parametric identification in this study. One is the recursive hybrid Genetic Algorithm and the other is the moving window hybrid Genetic Algorithm. In the development of these two algorithms, the time histories of the measurements were divided into a series of time intervals, and then the model of equivalent linear systems were employed to identify the modal parameters of the systems for each time interval. These two methods have different ways in dividing the time histories of the measurements. For the second one, the time histories of the sequential intervals are overlapped. Finally, these two methods are applied to the three-story structural models with added-damping-and-stiffness devices mounted on the shaking table. The ground motion records used for these models are time histories of El Centro earthquakes adjusted to different intensities. The results showed that the frequencies vary with the intensity of the ground motion and reflect the nonlinear behavior of the systems. The comparison of the results of the both methods showed that the change of the parameters for each interval are not smoother when compared to the results of the first method.