Cyclic backbone curves (CBCs) are critical for assessing the seismic performance of structural systems as they characterize the hysteretic behavior and energy dissipation capacity under cyclic loading. However, traditional methods for generating CBCs rely on an experimental program through multiple testing on the same configurations, which are burdensome and accompanied by some limitations and inaccuracies. In this paper, an automated data-driven platform that employs machine learning methods to predict CBCs for self-centering post-tensioned (SCPT) piers. By eliminating the need for labor-intensive simulations, this tool accelerates the design of new SCPT piers and the evaluation of retrofit strategies for existing ones. The framework aligns with the push toward automated data-driven construction practices, offering a scalable solution for global seismic risk mitigation. The platform is powered by a comprehensive database developed through a large number of randomized quasi-static cyclic analyses on diverse SCPT configurations, as well as explainable artificial intelligence options.
Hosseini et al. (Tue,) studied this question.