Due to the advancing possibilities of modeling, computing technology and data acquisition, today a variety of simulation models and data is available that is nearly unmanageable. This diversity in turn opens new possibilities for the development of computational methods in which information from different sources and of different quality is merged to further increase the efficiency of computational procedures. This is particularly important for reliability estimation and sensitivity analysis, which not only require many model evaluations, but also a careful consideration of the mutual influences between model parameters and model refinement on the one hand and the quality of the estimation on the other.

Several multilevel and multifidelity information fusion methods are presented and compared for the computation of the failure probability in the context of reliability assessment. Furthermore, the potential of machine learning to provide surrogate models and to guide the information update procedure is investigated.

The particular challenge for civil engineers in the structural design of timber floors to reach the requirements for serviceability is well known and normally decisive for the dimensioning. Extensive modeling of timber slabs in combination with active vibration elements has shown that, compared to simple passive vibration dampers, a significant improvement in vibration behavior is possible also with a considerable reduction in mass. The interpretation of these results suggests to intensify the research in the field of active vibration control of slabs in residential buildings. The active damping systems are intended to positively affect slender slab structures and their vibrational characteristics. The current work includes the vibrational behavior of cross laminated timber floors and how these characteristics can be affected. The dynamic parameters are determined for a single-span cross laminated timber slab, which can be used for the following investigations with active control. Comparative values of different control systems are already available. The current control systems use the acceleration values of the continuous vibration measurement for the counteraction and levels of intensity. Random and dynamic motions caused by people moving on the slab systems increase the difficulty to solve this task using common control techniques. Both acceleration feedback as well as velocity feedback are tested and a significant reduction in vibration acceleration of the timber slabs can be reached. Important for the validation of the results are the analytical investigations and the software-supported verification of the measurement data according to the vibration excitation including active damping. Implementation of active control in the FE-simulations and comparisons between experimental and numerical investigations are currently in progress. This enables the comparison of results from different types of excitation and particularly the very important person-induced excitation. The aim is the possibility to use the FE-simulations and analyze entire slab systems with active vibration control.

Mass dampers are a widely accepted control technique for seismically-excited tall buildings. When these mass dampers are optimally tuned to the primary natural frequencies of buildings, the structural responses (i.e., floor displacements and accelerations) can be effectively mitigated. However, the tuned mass dampers may introduce a large displacement when buildings are subjected to intensive earthquake loadings. To address this shortcoming, some researchers suggested adding nonlinear restoring forces to mass dampers, such as forming a track nonlinear energy sink. Still, a sufficiently large mass in this nonlinear mass damper is a critical issue. Therefore, this research develops a track nonlinear energy sink with a mass moment of inertia. The proposed mass damper not only has the feature of nonlinear restoring forces but also increases effective mass by rotational components. This study first derives the equation of motion for a building with the proposed mass damper. A design method based on the frequency-domain input-output relationship is established. To further verify the damper performance, a prototype track nonlinear energy sink with a mass moment of inertia is fabricated and experimentally evaluated by real-time hybrid simulation. The experimental results exhibit that the proposed mass damper outperforms the conventional track nonlinear energy sink. Moreover, only adequately effective mass, i.e., sufficient momentum to maintain static friction, is feasible to generate control performance against input ground motion.

The use of coupled shear walls has been one of the potential options as an earthquake-resistant system in reinforced concrete high-rise buildings in recent times. When the coupled shear walls are subjected to earthquake motion, energy dissipates in the coupling beams depending on the design at the base of the shear walls. This paper aims to simplify the behaviour of coupling beams with feasible boundary conditions using analytical, numerical, and experimental analyses. Considering Galano and Vignoli's (2000) scaled model as a reference point, analytical equations have been developed and validated with numerical analysis using ATENA 2D 2006 software to study the behavior of coupling beams. Nonlinear static pushover analyses have been considered as one of the methods used in this software. A series of specimens were cast to evaluate the actual response of a prototype model, with an L_w1 and L_br equal to 200 mm, L_w equal to 500 mm and a thickness of 100 mm. Based on the analytical results, the prototype model can be considered an appropriate scale model to conduct numerical and experimental investigation of the behavior of the reinforced concrete coupling beam. Numerical analyses based on feasible boundary conditions render good approximations to experimental deflections and thus give confidence in the use of this technique in the design office.