PSI - Issue 78

Atilla Ansal et al. / Procedia Structural Integrity 78 (2026) 2133–2140

2134

A site-specific seismic hazard analysis requires the local seismic hazard study to estimate for the uniform hazard acceleration spectrum (UHS) on the engineering outcrop (Erdik, et al., 2004) The next stage in site-specific response analysis is the selection and scaling of the input acceleration records to match the UHS on the outcrop obtained from the probabilistic regional earthquake hazard analyses (Ansal et al., 2012, 2024). It is preferred to carry on a deaggregation to estimate possible source distance and fault type. A large number of acceleration records compatible with the local earthquake hazard and deaggregation findings recorded on stiff site conditions may be used for site response analyses to account for the variability in earthquake source and path factors. The selected acceleration records need to be properly scaled with respect to UHS on the bedrock at the microzonation site (Tönük and Ansal 2010). Site response analyses are conducted by adopting non-linear one-dimensional (Idriss and Sun, 1992) approach for all the soil profiles. The variability in the soil profiles is modelled by Monte Carlo simulations with respect to layer thickness and shear wave velocity as suggested by Tönük and Ansal (2023). The results of 1D and 2D (Shamekhi, et al., 2023, Shamekhi and Ansal, 2025,) analysis (including the vertical acceleration at the bedrock outcrop) are shown below in Figure 1. The results of 2D indicates better modeling for some stations and for some of the other stations 1D and 2D results are similar, however, in general 1D results are giving higher values in comparison to 2D analysis in modeling the recorded acceleration registered at the Istanbul Rapid Response Network stations that recorded the 19.05.2011 ML=5.9 Kütahya Earthquake.

40

EW

Recorded 1D Site Response 2D Site response

30

20

10

Peak Spec. Acc. (mg)

0

ISKIF

ISKIK

ITFME

YESIO

ZYAIO

AGPIO

ZYTAL

FATZA

KRTTP

ZOTML

BAHBL

BAGYT

HAVAV

ESNBB

ZAVKO

AVCEM

KAKML

AKUKO

GUNGR

EMNMR

Figure 1. 1D and 2D site response analyses results compared to recorded peak spectral accelerations during the 19.05.2011 Kütahya Earthquake

2. Site Response Analyses The approach adopted by the author and his colleagues was based (a) on probabilistic local hazard study conducted to determine the probabilistic acceleration spectra on the engineering bedrock; and (b) the selection of a large number hazard-compatible acceleration time histories based on the seismic deaggregation conducted to perform the site response analyses (Tönük and Ansal, 2022) taking into account stress and frequency dependence in the analysis. The selection of a set of input motions is an important step (Kottke & Rathje, 2011) in carrying out a site response analysis. Recorded time histories (PEER, 2024) are selected for stiff site conditions with Vs30 ≥ 760 m/s and within the range of ± 10% of expected Mw and ± 20% of the estimated fault distance based on hazard deaggregation. The main aim of scaling is to find the best method to match the uniform hazard acceleration spectrum on the engineer bedrock. One option (Ansal, et al., 2012) corresponds to scaling selected acceleration records individually to obtain the best fit for each period level with respect to the target acceleration spectrum by varying the PGA without modifying the frequency content. Another issue in site response is the number of acceleration time histories needed to account to a limited extent, for the variability and uncertainty due to source and path effects (Papaspiliou & Kontoe, 2013). Previous studies conducted on site response have indicated that the difference between the average acceleration spectra computed on the ground surface (Figure 2) starts to decrease significantly after 20 input motion. Therefore, it should be preferred