Dependency of unit shaft resistance on in-situ stress: Observations derived from collected field data

In this study, a database comprised of 30 pullout pile load tests was collected from geotechnical literature and analyzed to investigate the dependency of unit shaft resistance on effective vertical stress. The collected database consists of steel pipe, timber, and concrete piles, with varying normalized penetration depth with respect to pile diameter, driven into loose to very dense sand. Different correlations for the uplift lateral earth pressure coefficient K, Bjerrum-Burland ratio , and the average unit shaft resistance f _ave were derived using different assumptions. A comparison between measured and predicted capacities of the collected piles using the developed correlations indicated that the assumption of values of K and that were constant with depth did not provide a reasonable fit for the measured capacities of the collected piles and thus this assumption is inappropriate. The best correlations for K and that yield a reasonable fit to the measured capacities of the collected piles were found to be functions of sand relative density, pile diameter, and level of effective vertical stress. This indicates that average unit shaft resistance does not reach a limiting value, but rather continues to increase with depth. Moreover, the correlations for K and in terms of effective stress revealed that average unit shaft resistance increases as pile diameter decreases and this increase depends on initial sand relative density. Comparisons of measured and predicted pullout capacities of the collected piles using the best-obtained correlations for K and were made and compared to predictionsobtained from other methods. On the basis of these comparisons, it is concluded that the correlations for K and in terms of effective stress give results comparable to those obtained from other methods, without stipulating limiting values for the average unit shaft resistance.