Fragility curves for seismic damage assessment in regular and irregular MRFs using improved wavelet-based damage index

In the conventional performance-based earthquake engineering framework, fragility curves are mapped to a function of story drift ratio or spectral acceleration corresponding to the first mode. Based on the fragility curves, the probability of structural healthy status can be estimated. Practically, it is difficult to measure story drift and expensive to develop an automative measuring system. Therefore, this paper strives to develop fragility curves by using wavelet-based refined damage-sensitive feature (rDSF), considering higher mode contributions. These fragility curves can be obtained from absolute acceleration responses recorded at each story of buildings. As irregularities have important and significant effects on behaviour and seismic performance of Moment Resisting Frames (MRFs) especially in the field of structural health monitoring, in the next step, we extend the fragility curves for irregular MRFs. To this end, an efficient and more precise rDSF using Morlet and cmorfb-fc wavelets are extended to map fragility curves. The correlation coefficient between rDSF and maximum story drift ratio is evaluated as a criterion to determine the effects of irregularity. The 6-story steel benchmark MRF was manufactured to consider a stiffness reduction irregularity caused by the deference in story height and ductility of beam-column members. To assemble rDSFs, acceleration responses recorded both from the regular and irregular MRFs were analyzed by using incremental dynamic analysis, subjected to different ground motion sets. Moreover, a system identification method is used to identify the natural frequencies of each MRF. Based on the acquired results, it can be concluded that the MRFs including higher first story and weak columns-strong beams changes with a higher probability of damages lead to damage occurrence at lower seismic intensities and also lower maximum story drift ratios as compared with those of reference MRF. Moreover, the results show that the fragility curves estimated by wavelets-based rDSF, especially by using cmorfb-fc wavelet-based rDSF, due to lower probability of damages (varied between 27% and 81% for different damage states) have more efficiency than the fragility curves derived from Morlet wavelet-based DSF, considering only the first mode, for all the MRFs of interest. Furthermore, global fragility curves lead to a total assessment of the MRFs, while for a more accurate damage diagnosis, the rDSF-base fragility curves computed for each story can be used.