The need for an improved understanding of the influence of sample scale on the compressibility behaviour of fine-grained soils is crucial in many applications, such as roads, embankments and foundations. The effect of sample scale represents a challenge when obtaining engineering parameters in the laboratory compared to those obtained in the field. This research therefore aimed at contributing to existing knowledge through both experimental and numerical studies. The experimental study was completed via a series of consolidation tests on fully saturated fine-grained soil (i.e. kaolin clay) at various sample scales, which were subsequently explored analytically using the finite element software PLAXIS 2D. This type of clay was chosen as it is easily sourced and well known. A Scanning Electron Microscope (SEM) and X-Ray Diffractometer (XRD) were employed in studying the micro-structure of the soil.

The Oedometer apparatus was used to obtain the combined effect of sample scale and initial moisture content on the compressibility parameters of kaolin clay. Compressibility parameters such as coefficient of consolidation (cv), compression index (cc) and coefficient of volume compressibility (mv) were also investigated in this study. Three different methods were used to obtain cv: Casagrande, Taylor’s and Inflection methods. The sample scales were divided into three categories: soil sample height, diameter and diameter to height (D/H) ratio scale. Particular attention was given to the D/H ratio effect on the compressibility parameters due to the frictionless boundary conditions, and sample diameter scale due to drainage path length.

Based on the experimental data, Taylor’s method was considered more reliable in deriving cv as compared to the Casagrande and Inflection methods, due to the end of primary consolidation (EOP) observed at all sample scales. It was also observed that on average, sample scale has an effect on cv with a correlation factor of 0.451 and that friction was reduced by 35%. The effect of sample scale on cc and mv was found to be insignificant. The investigation also showed a correlation factor between the initial moisture content and cv of 0.546 and, 0.162 and 0.026 for cc and mv respectively. The new proposed model developed in PLAXIS 2D was found to show no significant difference with the laboratory data except where the calculated coefficient of permeability was used. A new proposed model was developed in PLAXIS to further study the effect of sample diameter scale on the behaviour of fine-grained soils.