Desarrollo de programación en Python para el cálculo de curvas de fragilidad de edificaciones mediante análisis no lineal en tiempo-historia.

Abstract

This work develops and implements a methodology to estimate seismic fragility curves using nonlinear time-history analysis. The structure is representing as a single degree of freedom (SDOF) system with bilinear behavior, defined by yield and ultimate displacement parameter. The response is obtained through step-by-step integration using the Newmark method, allowing the construction of the force vs displacement relationship (hysteresis) and the extraction, for each seismic record, of a performance point based on the maximum absolute response. From these results, a cloud of points is generated relating a selected intensity measure (PGA, elastic Sd, or elastic Sa) to the system performance. For each damage state, the point is classified according to exceedance of predefined damage thresholds, and the intensity range is statistically processed by grouping records, computing average intensities per group and exceedance fractions k/N. In addition, anchoring points are incorporated at intensity levels below and above the range, obtained from available record or, when insufficient data exist, from auxiliary values defined relative to the intensity limits; these anchors stabilize the fit and ensure a consistent probabilistic transition at the extremes. Finally, Baker lognormal functions are fitted to estimate the parameters θ and β of the fragility curves for each damage state. The procedure is automated in a computational code that integrates record Reading, statistical processing, and the export of tables and curves, providing a tool applicable to structural performance and seismic fragility assessments