Back-Calculation of the Shear Strength Parameters of Subgrade Soil Based on Lightweight Deflectometer Test Using 3D Numerical Modeling
Künye
Kadakci Koca, T. Back-Calculation of the Shear Strength Parameters of Subgrade Soil Based on Lightweight Deflectometer Test Using 3D Numerical Modeling. Geotech Geol Eng (2022). https://doi.org/10.1007/s10706-022-02290-8Özet
The undisturbed sampling from natural carbonate gravel-bearing subgrade soil for shear strength tests is a challenging process, or in-situ tests are costly and require time. In addition, mechanical properties of soils, provided through laboratory testing are not commonly in good harmony with the in-situ characteristics of the same soil. For these reasons, back-calculation of the in-situ shear strength parameters and Poisson's ratio of a 2 m thick subgrade soil was addressed in this study. In this context, a total of nine lightweight deflectometer tests (LWDTs) were performed. Back analyses were performed considering the in-situ measured settlement under the plate with the aid of 3D numerical models based on elastic theory. The dynamic deformation moduli derived from LWDT were used as input for the numerical models. The soil physical properties were obtained from laboratory tests. The subgrade soil was characterized by the Mohr-Coulomb failure criterion. Initially, a site-specific Poisson's ratio was back-calculated. Afterward, the finite element models were operated for trial sets of internal friction angle and cohesion values until the amount of settlement from iterative numerical solution converges to the one derived from the LWDT. Hence, the test results were simulated with 0.001 mm precision. Additionally, settlement under the plate was calculated empirically considering the elastic, half-space medium. The empirical settlement values were well correlated with the test results. As a result of this study, it is recommended to investigate correlations between the LWDT results and physical properties to accurately evaluate the performance of numerical simulation models.