Scholarlink

Dynamic soil properties, i.e. modulus reduction (G/Gmax) and damping (ξ) ratio, play a crucial role in seismic ground response analysis (GRA). The uncertainty associated with these dynamic soil properties may affect the outcomes of GRA substantially. To increase the reliability in the analyses, the associated uncertainties need to be considered while performing the GRA. In this work, the effect of uncertainty associated with G/Gmax and ξ curves have been assessed. For this purpose, the soil profile of Ultadanga site in Kolkata city has been used. The study considers the mean and mean ± std profiles shear wave velocity profiles (Vs) obtained from the randomized Vs profiles generated in the study by Roy (Indian Geotech J, 2024). The baseline G/Gmax and ξ curves and associated std for each layer have been generated following the model proposed by Darendeli and Stokoe (Development of a new family of normalized modulus reduction and material damping curves. University of Texas, Austin, Texas, 2001). Then for each layer, fifty randomized G/Gmax and ξ curves are generated by adopting a negative correlation between them. Now the equivalent linear ground response analysis (EQL-GRA) is conducted for each of the three Vs profile with 200 sets of G/Gmax and ξ curves. Each set consists of a random selection of a pair of G/Gmax and ξ curves from each individual layer. So, for a particular Vs profile, 200 number of simulations and for all the profiles total 600 numbers of simulation have been performed in this study. The results are prepared to represent the transfer function (TF), PGA profile and spectral acceleration (SA). A comparison of median and log standard deviations has also been conducted for all the adopted three Vs profiles. Finally, the cumulative probability curves of surface PGA variations have been generated for those profiles. A MATLAB (Computer software. MathWorks, Natick, MA) code has been developed to carry out the entire study. The outcome clearly exhibits significant effect of nonlinear soil properties, i.e. G/Gmax and ξ curves, on EQL-GRA. The study clearly reveals the higher values of median SA and PGA for mean + std profile. On the other hand, higher log standard deviations are observed for mean and mean − std profile.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Effect of Uncertainty in Nonlinear Soil Properties on Seismic Site Response Analysis

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Narayan Roy

摘要

Dynamic soil properties, i.e. modulus reduction (G/Gmax) and damping (ξ) ratio, play a crucial role in seismic ground response analysis (GRA). The uncertainty associated with these dynamic soil properties may affect the outcomes of GRA substantially. To increase the reliability in the analyses, the associated uncertainties need to be considered while performing the GRA. In this work, the effect of uncertainty associated with G/Gmax and ξ curves have been assessed. For this purpose, the soil profile of Ultadanga site in Kolkata city has been used. The study considers the mean and mean ± std profiles shear wave velocity profiles (Vs) obtained from the randomized Vs profiles generated in the study by Roy (Indian Geotech J, 2024). The baseline G/Gmax and ξ curves and associated std for each layer have been generated following the model proposed by Darendeli and Stokoe (Development of a new family of normalized modulus reduction and material damping curves. University of Texas, Austin, Texas, 2001). Then for each layer, fifty randomized G/Gmax and ξ curves are generated by adopting a negative correlation between them. Now the equivalent linear ground response analysis (EQL-GRA) is conducted for each of the three Vs profile with 200 sets of G/Gmax and ξ curves. Each set consists of a random selection of a pair of G/Gmax and ξ curves from each individual layer. So, for a particular Vs profile, 200 number of simulations and for all the profiles total 600 numbers of simulation have been performed in this study. The results are prepared to represent the transfer function (TF), PGA profile and spectral acceleration (SA). A comparison of median and log standard deviations has also been conducted for all the adopted three Vs profiles. Finally, the cumulative probability curves of surface PGA variations have been generated for those profiles. A MATLAB (Computer software. MathWorks, Natick, MA) code has been developed to carry out the entire study. The outcome clearly exhibits significant effect of nonlinear soil properties, i.e. G/Gmax and ξ curves, on EQL-GRA. The study clearly reveals the higher values of median SA and PGA for mean + std profile. On the other hand, higher log standard deviations are observed for mean and mean − std profile.