Due to lack of soil sampling during a conventional cone penetration testing (CPT), it is necessary to classify soils based on recorded tip and sleeve friction and pore pressure (if available) values. However, currently available soil classification models are based on deterministic and judgemental determination of soil classification boundaries which do not address the uncertainties intristic to the problem. Moreover, size and quality of databases used in the development of these soil classification models are undocumented and thus questionable. Similar limitations do also exist in the development of SPT-CPT correlations which are widely used in SPT dominated design such as soil liquefaction triggering. To eliminate these discussed limitations, within the confines of this study it is attempted to present (1) a new probabilistic CPT- based soil classification methodology, and (2) new SPT-CPT correlations which address the uncertainties intrinsic to the problems. For these purposes, a database composed of 400 CPT/SPT boring data pairs was compiled. It is intended to develop probabilistic models, which will correlate CPT tip and sleeve friction values to actual soil classification and CPT tip resistance to SPT blow count N. The new set of correlations, model parameters of which estimated by implementing maximum likelihood methodology, presented herein are judged to represent a robust and defensible basis for (1) prediction of soil type based on CPT data and, (2) estimation of SPT-N value for given CPT data.

As we know the most popular inference methods for order restricted model are likelihood inference. In such models, the Maximum Likelihood Estimation (MLE) and Likelihood Ratio Testing (LRT) appear some suspect behaviour and unsatisfactory. In this thesis, I review the articles that focused in the behaviour of the Likelihood methods on Order restricted models. For those situations, an alternative method is preferred. However, likelihood inference is satisfactory for simple order cone restriction. But it is unsatisfactory when the restrictions are of the tree order, umbrella order, star-shaped and stochastic order types.

One of the most critical problems geotechnical engineers face with is the determination of the amount of consolidation settlement that will occur at a site as a result of the construction of a structure. The compressibility behavior of the soil is an important parameter in determining the amount of consolidation settlement. The goal of this study is to develop probabilistically based correlation between the compressibility behavior of soil and in-situ test data. Within the scope of this research effort, performed CPT tests and the recorded settlement case histories where consolidation settlements at the field under various surcharge loads were compiled from the Bursa East and West Waste Water Treatment Plant soil investigation projects. A database was composed of the results of 45 CPT and 57 settlement plate recordings. For the compilation of this database, a series of finite difference software FLAC-3D analyses were carried out to calculate the change in stress distribution under the settlement plates. A maximum likelihood framework was used for the development of compressibility behavior of soils. As a result of careful processing of available data, the cone tip resistance (qc), soil behavior type index (Ic) were selected as two important parameters effecting the value of the one-dimensional constraint modulus, M. The regression analysis which uses the settlement values recorded at the site and those computed using the change in the stress distribution, the thickness of the sublayers and the proposed one-dimensional constraint modulus were carried out to calculate the values of these model parameters. Two correlations based on the cone tip resistance and soil behavior type index were developed for the computation of the one-dimensional constraint modulus, M.