Behaviour Investigation of SMA-Equipped Bar Hysteretic Dampers Using Machine Learning Techniques

Most isolators have numerous displacements due to their low stiffness and damping properties. Accordingly, the supplementary systems vital roles in enhancement lower isolation system displacement. Nevertheless, many cases, even by utilising additional dampers systems, occurrence of residual displacement is inevitable. To address this issue, study, a new smart type bar hysteretic equipped with shape memory alloy (SMA) bars recentring features, as damper, introduced investigated. In regard, 630 numerical models SMA-equipped (SMA-BHDs) were constructed based on experimental samples different lengths, numbers, cross sections SMA bars. Furthermore, hysteresis curves corresponding ideal bilinear curves, role geometrical mechanical parameters cyclic behaviour SMA-BHDs was examined. Due deficiency existing analytical models, proposed previously for steel (SBHDs), estimate first yield point post-yield ratio accurately, developed artificial neural network (ANN) group method data handling (GMDH) approaches. The results showed that, although ANN outperform GMDH ones, both ANN- GMDH-based can accurately linear nonlinear pre- parts errors high accuracy consistency.