Assessment of geotechnical variability of Songdo silty clay

Abstract

The reliable estimation of the coefficient of variation (COV) and probability density function (PDF) is required to characterize the uncertainty and variability of soil properties. An outlier test, a randomness test (runs test), and a Goodness-of-Fit test are performed in sequence to determine the reliable COVs and PDFs of soil properties for Songdo silty clay. Soil properties such as C-c, e(0), C-c/(1+e(0)), C-alpha/C-c, gamma(t) and gamma(b) are found to fit to both normal and lognormal distributions, whereas c(v), k(v), and s(u)/p(o)' appear to fit to a lognormal distribution at the 5% significant level. It is observed that the COVs of C-c, C-c/(1+e(0)), c(v), k(v), C-alpha/C-c, and s(u)/p(o)' are larger than those of e(0), gamma(t) and gamma(b). The COVs of c(v) and k(v), which are associated with the time rate of consolidation, are found to be the largest among those of soil properties considered. It is also found that the N-sigma rule provides similar COVs to those of a statistical method when using N-sigma = 6 for soil properties with small COVs and N-sigma = 4.2-53 for soil properties with large COVs. The empirical method appears to provide conservative characteristic values for soil properties that have relatively large COVs. Finally, this study presented the lower and upper bound characteristic values for properties of Songdo silty clay. (C) 2012 Elsevier B.V. All rights reserved.