Effects of permeability and compressibility on liquefaction screening using cone penetration resistance

Empirical relationships between normalized cone penetration resistance (qc1N), cyclic resistance ratio (CRR), magnitude of earthquake, and silt content, derived from field observations, are widely used for liquefaction potential assessment of loose saturated sands and silty sands (Robertson and Wride 1998). This dissertation examines the effects of non-plastic fines content on the relationship between qc1N and CRR. A brief review of recent work on the effects of fines content on liquefaction resistance is presented. It indicates that liquefaction resistance of sands and silty sands are nearly the same when compared at the same 'equivalent' intergranular void ratios (Thevanayagam et al. 2002a). Stress-strain relationships are also nearly the same for sand and silty sand at the same 'equivalent' intergranular void ratio. However, permeability k and coefficient of consolidation ch of sand and silty sand are not the same at the same 'equivalent' intergranular void ratio. k and ch decrease significantly with an increase in silt content (Shenthan 2001). A brief literature review of explanations for the dependence of qc1N versus CRR on silt content is presented. One of the hypotheses in the literature (Thevanayagam and Martin 2002) proposes that qc1N of sands and silty sands, at the same equivalent' intergranular void ratio but with different k and compressibility mv, would be very sensitive to k and compressibility; Since k and mv are sensitive to fines content, the relationship between CRR and qc1N would depend on fines content. Numerical simulations of cone penetration in sands and silty sands conducted as a part of this dissertation indicate that silty sand shows smaller qc1N than qc1N for sand at the same 'equivalent' intergranular void ratio. The ch (=k/(mvgammaw)) has a major influence on qc1N. This is due to significant differences in concurrent pore pressure generation and dissipation around the cone tip during penetration in low permeable silty sands compared to highly permeable sand. In highly permeable sand cone resistance is a 'drained' response whereas it is an 'undrained' response in very low permeable silty sand, even if both soils have the same stress-strain constitutive relationships. At intermediate ch values, there is a transitional behavior between 'drained' and 'undrained' response. The respective cone resistances are very different. Further numerical simulations indicate that qc1N is also dependent on diameter of cone d, and velocity of penetration v. A relationship is found between liquefaction resistance, qc1N, and a normalized penetration rate T (=vd/ch). The effect of silt content appears in this relationship through ch. The numerical simulations results are expected to reflect the limitations of the material model. Yet the trends provide an understanding the effects of permeability and compressibility on cone penetration resistance. A limited number of large scale 1-g laminar box liquefaction experiments and cone penetration tests conducted in sands support this relationship. Additional numerical simulation studies and large scale liquefaction experiments are needed to further refine and validate this relationship for silty sands.