Machine learning-aided peak displacement and floor acceleration-based design of hybrid self-centering braced frames

The hybrid self-centering braced frame (HSBF) is an emerging seismic resilient structural system with the capacity of controlling peak inter-story drifts and floor accelerations and eliminating residual inter-story drift by combining displacement- and velocity-proportional damping. The shape memory alloy-based braces (SMABs) and viscous dampers (VDs) are the main nonlinear components for obtaining the expected displacement- and velocity-proportional damping, respectively. This research focuses on developing a practical multiple performance objectives-based design method on the basis of the developed artificial neural network (ANN) models for controlling the desired peak inter-story drifts and floor accelerations in low-to mid-rise rise multi-story HSBFs. Four multi-story HSBFs with 3, 7, 11, and 15 stories were designed using the proposed multiple performance objectives-based method. The dynamic responses of the designed HSBFs were studied using nonlinear dynamic analysis under different seismic intensities. It can be observed from the analysis results that the designed HSBFs can obtain the prescribed performance objectives of peak inter-story drifts and floor accelerations under considered earthquakes. Moreover, negligible residual inter-story drifts were observed in the designed HSBFs even under the maximum considered earthquakes, confirming the excellent post-earthquake repairability of the HSBFs. For practical application, the software ‘SmartHSBF’ was developed to automatically obtain the design information of SMABs and VDs by defining the designed performance objectives.