Rafet Sisman

The consideration of consecutive seismic events remains a notable deficiency within current seismic design practices for engineering structures. The progressive accumulation of damage from an earthquake sequence poses a significant risk to the structural integrity of buildings, even those designed to exhibit ductile seismic behavior. The Kahramanmaras earthquake sequence in 2023 presented a unique opportunity to evaluate the contribution of multiple earthquakes on seismic damage in reinforced concrete (RC) buildings. The primary objective of this research is to quantify the variation in structural damage within low- to mid-rise RC buildings as a result of consecutive seismic events occurring within a short time period. Utilizing recorded ground motions and representative numerical models for RC buildings, the incremental damage sustained by RC building typologies subjected to the Kahramanmaras earthquake sequence is analyzed. This study examines the incremental increase in the probability of reaching predefined damage levels, comparing the effects of the first event alone with those of the complete earthquake sequence. The results demonstrate a significant contribution of subsequent events to structural damage accumulation, increasing overall collapse rates by approximately 2% on average, with the most severe impact observed in low code-compliant buildings where up to 9.2% of moderately damaged structures progress to collapse. Furthermore, conditional damage transition probabilities are derived, establishing a Bayesian updating framework to recalibrate post-sequence empirical fragility curves for single-event assessments.

Kinematic interaction (KI) effects arising from foundation embedment alter the motion transmitted from the free field to the foundation level and can therefore influence the seismic demands imposed on structures. Although several analytical formulations and guideline-based procedures exist to represent these effects, their accuracy when used directly to modify design and response spectra has not been thoroughly validated against real ground-motion data. This study provides a systematic evaluation of two widely used KI approaches; (i) the analytical transfer functions of Elsabee and Morray (1977) and (ii) the spectral-modification procedure of FEMA-440 (2005). Using a large suite of recorded motions, free-field and foundation input motions are computed through frequency-domain modification, and corresponding acceleration response spectra are obtained to form computed spectral ratios. These computed ratios serve as a benchmark for assessing the predictive performance of the simplified KI procedures across a range of embedment-to-stiffness conditions. The results show that analytical KI factors substantially underestimate short-period spectral ordinates, whereas FEMA-440 captures overall trends more effectively but loses accuracy for larger embedment depths. The study defines practical ranges in which simplified KI procedures remain reliable and identifies conditions that require more refined soil-structure interaction modelling. The findings provide engineers with clearer guidance for incorporating KI effects in seismic analysis and design.

In current engineering practice, the torsional effects of ground motions are only considered implicitly through “accidental eccentricity coefficients". In seismic design codes, these coefficients account for not only the unintended (or uncertain) asymmetric distributions of structural mass and stiffness in the plan but also the torsional components of ground motions. The torsional effects may develop due to wave passage effects or the incoherency of translational ground motions along the foundation systems. While it is possible to obtain torsional components of ground motions directly from seismic records with modern instruments, such sensors are uncommon. Therefore, several numerical procedures have been proposed in the literature for estimating torsional components of ground motions from translational measurements. In this study, the Single Station Procedure (SSP)—based on the deconstruction of translational ground motions into body waves and the subsequent reconstruction of rotational ground motions—is employed to obtain torsional ground motions for seven different seismic events in Turkey. Additionally, the corresponding accidental eccentricity coefficients that can cause the same response modification with torsional ground motions are determined. First, a conventional procedure is utilized to obtain the corresponding accidental eccentricity coefficients. Various drawbacks of the conventional procedure, such as negative eccentricity coefficients, are addressed. Then, a relatively obscure definition of accidental eccentricity is revived to propose a novel procedure for determining accidental eccentricity coefficients. It is observed that the proposed procedure yields higher values for the accidental eccentricity coefficients than the conventional procedure. It also observed that the torsional spectrum provided in Eurocode 8 provisions requires higher eccentricity values than those implied by the recorded ground motions from Turkey.