Akademik Veri Yönetim Sistemi
6th International Academic Studies Conference, 26 - 28 Temmuz 2021, sa.1, ss.221, (Özet Bildiri)
In this study, it is aimed to model the mechanical behavior of the MR (magnetorheological) damper with artificial neural networks. In many studies involving semi-active controllers using MR dampers, the need for realistic modeling of MR dampers has arisen. Due to their nature, MR dampers produce different forces at different speeds, positions and electric current values, and there is no linear relationship between them. For this reason, a large number of parametric models are encountered in the literature. Although some of these models are preferred in terms of ease of use, they do not give results close to the experimental results, and it is difficult to determine the parameters in models like the modified Bouc-Wen model, which gives appropriate answers to the experimental results. In order to overcome these problems, MR dampers were modeled using artificial neural networks, and it was determined which responses they would give under the conditions, thanks to the experimental data obtained. The proposed model can predict the nonlinear behavior of the MR damper with 99% accuracy. Researchers who will use this model in their studies can easily add an MR damper that behaves close to realistic results.