Strain-Dependent Cyclic Strength Ratio Model for Transitional Silts

Although the simplified method can be used to approximate the cyclic resistance of plastic soils, recent experience in the Pacific Northwest indicates that the SHANSEP-based framework implemented in this method can overestimate cyclic resistance for low plasticity silts. This paper presents a statistical model developed for the estimation of the cyclic strength ratio, τcyc/su,DSS, of silts for the number of uniform loading cycles, N, corresponding to cyclic shear strain failure criteria ranging from 1% to 10%. The database used for model development and statistical modeling procedure are described. The relative importance of overconsolidation ratio, OCR; plasticity index, PI; fines content, FC; and void ratio, e, on τcyc/su,DSS is evaluated, followed by nonlinear regression analysis to finalize fitted model parameters. OCR was excluded from the proposed model due to the strong correlation between OCR and undrained shear strength of silt. Strong correlation between e and PI lead to apparent multicollinearity and the need to consider just one of these two variables in the final statistical model. The τcyc/su,DSS estimates using the proposed model and those obtained from a previously developed cyclic resistance ratio-based model improve as the cyclic shear strain failure criterion increases. The statistical model returns τcyc/su,DSS increasing from 0.47 to 0.68 as the cyclic shear strain failure criterion increases from 1% to 10% for N = 30 and PI = 10, whereas for PI = 30, τcyc/su,DSS increases from 0.51 to 0.75. The proposed model may be used in a future performance-based earthquake engineering framework to link estimates of the strain-dependent τcyc/su,DSS to the consequences of cyclic failure.