Automated Seismic Fragility and Vulnerability Assessment for Buildings (ASFRAVA-B): Integrating probabilistic seismic design into performance-based engineering practices

Practical assessment of seismic fragility and vulnerability is critical for effective earthquake engineering and risk mitigation. This study introduces ASFRAVA-B, an open-source, automated framework designed to derive seismic fragility and vulnerability functions for building assessments. ASFRAVA-B employs the capacity spectrum method and built-in statistical procedures to efficiently generate seismic fragility and vulnerability curves. The framework supports various statistical fitting methods, including multiple stripes analysis with maximum likelihood estimation, generalized linear models with logit and probit links, and machine-learning logistic regression, with three regularization options. It offers flexibility for users to choose the fragility function. The applicability of ASFRAVA-B is demonstrated through two case studies: (1) an evaluation of how masonry infill affects the seismic fragility and vulnerability of a two-story school building compared to an open frame; and (2) a comparative exercise against Nonlinear Time History Analysis (NLTHA) of a four-story, three-dimensional multiple-degree-of-freedom building model using incremental dynamic analysis. Results show that the infilled frame generally exhibits a higher probability of exceedance than the open frame, despite demonstrating greater stiffness and strength in the pushover curve. This result highlights the necessity of considering masonry infill walls in structural models, particularly for low-rise structures where such infill significantly contributes to overall stiffness. In the comparative analysis, ASFRAVA-B tends to slightly overestimate damage probabilities at lower intensity measures, yet provides substantial computational efficiency compared to the detailed NLTHA model. Sensitivity analyses demonstrate that the configuration of ground motion records exerts a moderate influence on fragility curves, emphasizing the importance of careful record selection for reliable outcomes.