Design of a semi-active controlled magnetic damper for seismic mitigation of buildings

Magnetorheologic (MR) dampers are widely used around the world in various areas such as civil engineering, biomedical applications, automotive and defense industry, and it remains as an active research area. The control of structures via MR dampers under excitation is a leading topic in this area. In this study, the damping of concrete buildings via MR dampers under seismic excitation was investigated by using different steel building models. The analytical models for the flow and magnetic field inside the MR damper were examined. Different CFD approaches were investigated by considering a reference MR damper, and it was shown that the existing rheological approaches in the literature are not capable of modeling MR dampers. Rheological properties of MR fluid were defined by using a rheometer, and a new CFD approach was developed based on these data. The developed approach was then implemented for the design and optimization of the MR damper that was manufactured consistent to the scale of the building models. Different parametric dynamic models of the manufactured MR damper were investigated, and a non-parametric artificial neural network dynamic model was developed. In order to control the model buildings, three control algorithms were developed as two of them were based on the developed artificial neural network model. The building stock of Turkey consists of 5-8 storey reinforced-concrete structures. Four different steel building models were designed by considering their period parameters. Building models were semi-actively controlled by the MR damper on a shaking table under the earthquake acceleration records of El-Centro and Kocaeli. Experiments were re-conducted by using a commercial MR damper. The study shows that the base shear force indexes can decrease up to 50% regarding the passive-off condition and the manufactured MR damper has 23% better performance compared to a commercial MR damper.