Stochastic simulation of earthquake ground motions for the seismic assessment of monumental masonry structures: Source‐based vs site‐based approaches

Abstract

Earthquakes are among the most destructive natural disasters and have resulted in a massive number of fatalities and economic losses all over the world. Simulated ground motion records are valuable, particularly for regions lacking seismic stations or with a limited history of large‐magnitude earthquakes. Notably, a significant percentage of monumental masonry buildings are located in regions with limited access to real records; hence, simulated records play a paramount role in their seismic protection. However, few studies have investigated the structural response of heritage buildings via response history analyses to assess the performance of simulated earthquakes against real ones. To accomplish this, this study simulates the recorded time‐series of the 9th of July 1998 Faial earthquake in the Azores (Mw = 6.2) at four available stations, using two different simulation approaches, that is, a source‐based stochastic finite‐fault method and a site‐based broadband stochastic method. First, two masonry facades with sidewalls characterized by different slenderness levels are adopted to conduct this research. Moreover, the proposed approach is also applied to an existing monumental structure, that is, São Francisco Church, located at Horta, which was affected by damage during the Faial earthquake. Results demonstrate that both simulation approaches provide similar results in terms of structural response prediction. The proposed framework also demonstrates that a small mismatch in terms of predicted damage patterns can result in a significant relative error in terms of displacement predictions.

Backbone ground motion model through simulated records and XGBoost machine learning algorithm: An application for the Azores plateau (Portugal)

“…The initial evaluation of the simulations involves a validation procedure in which the simulated records of the 1998 Faial event are compared against the real seismic records. 57,58 The input-model parameters have been further calibrated herein for the simulation of scenario earthquakes at the bedrock. Next, the simulation results of scenario events are compared against the well-known empirical GMMs suitable for the tectonic characteristics of the region.…”

“…Karimzadeh and Lourenço 57 simulated the 1998 Faial (M w = 6.2) event and provided region-specific input-model parameters based on simulation validations against observed motions from this event. In this study, the validated input-model parameters of Karimzadeh and Lourenço 57 and Karimzadeh et al 58 are calibrated for the scenario events. To account for uncertainty in the parameters representing source and attenuation effects, however, they are here assumed to be random variables.…”

“…These parameters have already been validated for the 1998 Faial event with a magnitude of 6.2 in. 57,58 The authors utilized alternative sets of input-model parameters to optimize the alignment with recorded motions of the 1998 Faial event. Their analysis determined that "Set2" yielded the most favorable fit, as evidenced by the highest Goodness of Fit (GOF) scores.…”

“…29 The authors consistently reported very good GOF scores for most available stations that recorded the 1998 Faial earthquake (M w = 6.2). Table 9 displays the GOF scores calculated by 57,58 for the four available stations that recorded the Faial 1998 event, along with their respective information. It is noted that in their study, a wide range of IMs were employed to calculate GOF scores.…”

A multi‐level approach to predict the seismic response of rigid rocking structures using artificial neural networks

“…Due to the stochastic nature of earthquakes 27 and the sensitivity of the dynamic rocking response to different seismic signal characteristics, most studies investigated one or two ground motion parameters, formulating scalar or bivariate IMs, respectively. However, a scalar or bivariate IM could be insufficient to represent the impact of the seismic signal on demand estimation.…”