info:eu-repo/semantics/conferenceProceedings Šaponjić, Jovan Divac, Nikola Stojanović, Boban Bačanin, Vladimir Radovanović, Slobodan Divac, Dejan Mirković, Uroš Mirković, Nikola eng All rights reserved Serbian Association for Tunnels and Underground Structures (ITA Serbia); Jaroslav Černi Water Institute, Belgrade application/pdf 1648210 bytes Parameter determination of the hardening soil model with small strain stiffness: an optimization approach coupling finite element simulations and genetic algorithms 2025 Southeastern Europe Tunnelling Conference (SETC-2025), 1-3 October 2025, Belgrade, Serbia startpage: 229 endpage: 241 In the scope of Phase I of Line 1 of the Belgrade Metro project, a large laboratory and in-situ testing database of the soils and rock masses of Belgrade was used for determining the parameters of the hardening soil model with small strain stiffness (HSS). Due to the complexity of parameter determination of the HSS model based on conventional laboratory testing results, an optimization approach coupling finite element analyses and genetic algorithms was taken in the parameter determination process with the goal of defining a framework for full, reliable, and unified implementation of the HSS model in all necessary numerical analyses. For the selected geotechnical soil layer, each of the available laboratory tests (direct shear test and CD triaxial test) and in-situ tests (pressuremeter) were numerically simulated and the HSS parameters were calibrated so that each individual simulated test, consisting of multiple test specimens, could achieve its respective experimental recordings as best as possible by minimizing the value of the applied criterion function. The resulting parameters from the calibration process, alongside the analytically derived parameters for laboratory results, were compared and analyzed, and it is realized that while analytically derived parameters provide a reasonable first estimate, numerical calibration is essential for accurately capturing soil stiffness behavior and stress dependency and therefore creating reliable design solutions based on realistic soil properties. hardening soil model, optimization, genetic algorithms, direct shear, triaxial, pressuremeter https://unilib.phaidrabg.rs/o:8716 doi:10.5937/SETC25022S ISBN: 978-86-82565-56-7