Fragility based evaluation of different code based assessment approaches for the performance estimation of existing buildings

Abstract

Seismic performance of existing buildings constructed before modern seismic design codes is one of the important problems of earthquake prone countries. Increasing life and economic losses after strong earthquakes shows the necessity of seismic performance assessment studies and methods. Therefore, in this study assessment methods recommended by different seismic codes were investigated and probabilistically assessed by comparing the fragility curves of existing low- to mid-rise RC structures. Building performances were determined according to TEC-2007, TBEC-2018, EC8/3 and ASCE 41-17. Eight different occupied buildings, classified according to construction dates (old and new) and story numbers (3 to 6), were selected and investigated by using these codes. At first, building capacity curves and damage limits, defined by the different code methods, were determined via pushover analyses. Structures were then subjected to more than 300 real strong ground motion (GM) records employing nonlinear dynamic analysis and fragility curves were obtained. Static analysis results showed that strength, stiffness and deformation capacity of new buildings are significantly higher than old ones according to all seismic codes considered and it is observed that TBEC-2018 gives the most conservative capacity estimations with respect to others. Comparison of fragility curves indicated that damage probabilities of TBEC-2018 and EC8/3 are higher than other seismic codes and this situation is valid for all story number and building age classes. In most cases TBEC-2018 and TEC-2007 results represent the upper and lower bounds of damage probabilities. It was also found that building age is the most effective parameter that affect the fragilities of existing buildings. Distribution of fragility curves imply that different code methods can give significantly different damage estimations under identical seismic demand levels.