Evaluation of New and Existing Shear Modulus Reduction Models to Predict Measured In Situ G/G max in Transitional Silts

In this study, the shear modulus nonlinearity and degradation (or G/Gmax) predicted by a novel model developed for transitional silts is compared with other available G/Gmax models to the nonlinear elastic and inelastic in situ responses measured in several instrumented silt deposits. The newly developed G/Gmax model for transitional silts is developed using a suite of cyclic torsional shear (TS) and cyclic direct simple shear (CDSS) test data that were corrected to a common frequency using a new plasticity- and shear strain-dependent strain-rate correction method. Dynamic in situ testing was accomplished using a vibroseis-shaker and blast-induced ground motions to achieve maximum shear strains exceeding 1%. Moderately, overconsolidated silt deposits with varying degrees of plasticity and confinement from two field testing programs are considered for comparison. The effect of in situ loading frequency on stiffness is directly evaluated using different methods and accounted for in the comparison of nonlinear stiffness. Overall, comparison of frequency-adjusted in situ G/Gmax with widely-accepted curves for clays, strain-rate corrected curves recently developed for clayey soils, and the newly proposed silt models yielded considerable scatter in predictions, consistent with prior studies. The silt-specific models predicted the response of the low plasticity silt deposit best with mean point biases of 0.98 and 1.01, although the in situ G/Gmax of the medium plasticity silt was more or less predicted equally well using the three models considered. Differences between the model-estimated and observed in situ threshold shear strains to exceed the linear-elastic threshold shear strain, γte, are discussed, and implications for forward analyses are summarized.