Characterization and Estimation of Geotechnical Variability in Ankara Clay: A Case History

An important component in reliability-based design is the geotechnical property variability. Generic estimates are used often, but calibration to a local geologic setting is preferable. In this case history, a methodology is shown that employs local geotechnical data to estimate the total variability, using Ankara Clay for illustration. A literature review is used to estimate the inherent variability, which is modeled as a random field with coefficient of variation (COV) and scale of fluctuation. The resulting inherent variability COVs are much smaller than the generic ranges. Local correlations between various laboratory and field tests and soil strength and compressibility parameters then are developed to quantify the transformation uncertainties. The various sources of uncertainty are combined through a second-moment method to estimate the total geotechnical variability as a function of the test type and correlation used. The results show: (1) the COVs for direct laboratory measurements are significantly smaller than those obtained through correlations, and (2) depending on the geotechnical data available, the local COVs can be very different from the generic guidelines. These could lead to unconservative designs. These issues are illustrated by a simple design example.