Akademik Veri Yönetim Sistemi
9th ECCOMAS Thematic Conference on Computational Methods in Structural Dynamics and Earthquake Engineering, COMPDYN 2023, Athens, Yunanistan, 12 - 14 Haziran 2023
It is possible to convert median response spectra estimated based on a selected Ground-motion Prediction Equation (GMPE) to its corresponding Fourier amplitude spectrum (FAS) by using the inverse vibration theory (IRVT). The IRVT uses extreme value statistics, properties of single-degree-of-freedom oscillator transfer functions, and spectral ratio correction to develop FAS that are compatible with input median response spectra. After obtaining a linear site amplification function in frequency domain corresponding to a specific site condition from analysis of FAS, the quarter-wavelength approximation (QWL) can be employed inversely and a 1D shear wave profile (VS profile) as well as a host kappa value representative of the site condition and GMPE under consideration can be derived. kappa is the spectral-decay parameter which is often used to account for the reduction of the high-frequency amplitude of ground-motion caused by attenuation within the site profile. As part of the first step, when the IRVT is used to compute FAS, utilized peak-factor model and duration of motion model influences the results. Likewise, as part of the second step, when a linear site amplification function is derived based on FAS in hand, a reference VS profile is employed that influences the linear site amplification function computed. And lastly as part of the third step, in computation of 1D VS profile and host kappa value from linear site amplification function representing the site condition and the GMPE under consideration by employing the inverse quarter-wavelength theory, the density model utilized influences the computed VS profile. Main objective of this study is to derive GMPE compatible VS profiles and host kappa values for the recent Turkish GMPE of [1]. In doing so, different peak factor models; duration models; randomly generated reference VS profiles and different density models were used. This allowed investigating the degree of effect of peak factor model, duration model, reference VS profile and density model on computed 1D VS profiles and host kappa values.