Physical experiments on CPTU thin-layer effects of thin clay layers embedded in sand - With analysis and possible correction of cone resistance in layered profiles

Abstract: The cone penetrometer is an excellent tool to characterize stratigraphy and have good abilities for parameter interpretation in homogenous soils. However, in soils with small layer thickness, in terms of cone diameters, this thickness combined with the soil properties of the thin layer compared to the surround layers will greatly impact the measured values. Measured cone resistance is a result of complex flow mechanisms that is dependent on the soils property ahead of and behind the advancing cone. The effect of layer thicknesses of sand with varied density has received significant attention and some studies have also been performed on layered clays. There is however little available research on accurate detection of thin clay layers in sand with the CPTU. Cone resistance measurements in sand normally yields significantly larger values than in clay. Thin clay layers in sand is greatly affected by the surrounding soil and the measurements in these thin layers does therefore not only reflect the clay. Interpretation of properties in clay such as undrained shear strength without taking this effect into consideration will cause overprediction. The magnitude of the overprediction depends primarily on layer thickness and the contrast in strength and stiffness between the clay layer and sand layers above and below. A large scale chamber model was established to allow for CPTU testing of thin clay layers in saturated sand. The chamber had a diameter of 1.2mm and a height of 2.5mm. Procedures to create uniform loose to medium dense sand samples were developed. The samples consisted of homogenous, uniformly graded sand and non-sensitive clay with an undrained shear strength of 27.5kPa. Two samples were made and tested, one of which had clay layers of thicknesses 4, 8 and 12cm. The other had a single thick clay layer to determine the cone resistance unaffected by layering effects, called the characteristic cone resistance. CPTUs were performed using a standard 10cm2 cone. A vertical stress of 11.4kPa was added on top of the sample and vertical stress was recorded in the sample during construction. The stress development in the sample resulted in good approximation of the stress level, which showed significant silo effects by frictions along the chamber walls. The characteristic cone resistance in the clay was between 10% and 20% of that of the sand. In the 4cm clay layers the minimum measured cone resistance was 3.5 times higher than the characteristic value. While the difference decreased with increasing thickness, the effect of thin layers was significant for the 8 and 12cm thick clay layers as well. The cone resistance in sand was influenced by thin clay layers three cone diameters prior to the layer interface. When the cone entered the thin clay layers the cone resistance appeared to be significantly more affected by the sand layer above the clay layer than the one below. This dependence on the layer above is an important factor for the thin layer effects in clays. Detection of clay layers is typically done through interpretation of the pore pressure measurements in practice. However, the results from the experiments showed how excess pore pressure measurements did not reflect the clay layer thickness. In fact, no significant response was seen in the 4cm clay layer. This was due to deformations of the clay layer during penetration. A procedure for simplified numerical cone resistance analysis in layered profiles was developed. It consisted of spatial weighting filters for advanced averaging of the characteristic cone resistance profiles in order to approximate measured profiles. This procedure was applied to the determined characteristic profiles of the physical experiments of this study and the results compared well. Possible methods of correction of cone resistance measurements in thin clay layers in sand was estimated from the procedure together with the results of the physical experiments. Examples of correction is presented. The experimental work performed in this MSc study has provided high quality CPTU data together with detailed information on material properties and soil stresses. The data can be used in further interpretations aiming to accurately interpret measurements of thin clay layers in sand. The experimental work has been comprehensive, but further testing is recommended. Repeating the tests with other combinations of sand density, thin layer thickness and clay strengths would be of interest