Preparation of K0-consolidated reconstituted samples in the laboratory

Abstract

The formation of natural soils is often idealized by assuming <i>K₀</i> cross-anisotropic conditions. For laboratory testing, natural samples are not always available and these are commonly prepared from slurry using different types of apparatus. During an investigation into the behavior of stiff clays around tunnels, a consolidation device was designed for the preparation of one-dimensionally consolidated Speswhite Kaolin samples that would be subsequently trimmed to the required size and tested in a computer-controlled triaxial testing equipment. Although certain non-uniformity along the length of the samples was initially envisaged due to the relative dimensions of the consolidation chamber (i.e. length to diameter ratio) and potential friction issues, this was expected to be negligible based on a review of some of the different techniques used in the last twenty-five years. However, after preparing the first group of samples, which were intended to give relationships between the maximum applied vertical effective stress (σ_{<i>v max</i>}) and the final liquidity index (<i>LI</i>) and undrained shear strength (<i>c_u</i>), significant non-homogeneities were observed along their length. The reduction in the vertical stress applied to the slurry during consolidation using the particular design of the present device was quantified. The corresponding variations and patterns in <i>w</i> and <i>c_u</i> were also identified and a relationship between both was given for discrete values of <i>w</i> along the samples. Although most of these issues seem to be obvious to most geotechnical engineers with laboratory experience, these are generally either not investigated in detail (i.e. quantified) or not published. Most justify this by stating that additional consolidation stages in the triaxial cell during conventional laboratory practice improve homogeneity. Nevertheless, depending on the final use for the specimens this may not be the case and all the issues above can easily result in wrong correlations between the different parameters (as was the case at the beginning of the present study) and erroneous interpretation of some triaxial results. The authors believe that the aim should be to improve the quality of the reconstituted specimens prior to any laboratory testing. By publishing the results from this study the authors hope that other engineers appreciate the problem and to encourage them to publish their experiences with different materials/dimensions/arrangements when consolidating samples. This will help future researchers to make more informed evaluations of the different consolidation techniques and work more efficiently.