The Cone Penetration Test (CPT) is widely used for determining in-situ properties of soil because of its continuous data measurement and repeatability at relatively low cost. The test is even more attractive in cohesionless soils such as sands, silts and most tailings due to difficulties associated with retrieving undisturbed samples in such soils. Behaviour of soils is highly dependent on both density and stress level. The state parameter is widely accepted to represent the soil behaviour encompassing both density and stress effects. Hence, determining the in-situ state parameter from CPT is of great practical interest.

The CPT was analysed using a large strain spherical cavity expansion finite element code using a critical state soil model (NorSand) capable of accounting for both elasticity and plastic compressibility. The constitutive model was calibrated through triaxial tests on reconstituted samples. The state parameter was then interpreted from CPT tip resistance, and the results were verified against an extensive database of calibration chamber tests. The efficiency of the method was further investigated by analysing two well documented case histories confirming that consistent results could be obtained from different in-situ testing methods using the proposed framework. Consequently, cumbersome and expensive testing methods can be substituted by a combination of triaxial testing and finite element analysis producing soil specific correlations.

One of the difficulties in analysing the cone penetration problem is the less researched effect of high stresses developing around the cone on the behaviour of the soil. A hypothesis was put forward on the particle breakage process at the particle level and its implications for the behaviour of sands at higher stress levels were discussed. A series of triaxial tests were performed, focusing on the effects of particle breakage on the location of the critical state line. The hypothesis was used to explain the observed behaviour. Particle breakage was shown to cause additional compression and a parallel downward shift in the critical state line. The magnitude of the shift was linked to the amount of breakage and it was argued that significant breakage starts after the capacity for compression due to sliding and rolling is exhausted.