Uncertainty analysis for drilled shaft axial behavior using CYCU/DrilledShaft/143

Drilled shaft axial capacities and their uncertainties were evaluated in this paper. A large load test database of drilled shafts labeled as CYCU/DrilledShaft/143 was compiled for this evaluation. All load tests were performed in the field under drained and undrained soil conditions. The database was made available at ISSMGE TC304 database sharing platform 304 dB. Drained and undrained analyses were conducted separately. These load test data were first analyzed using representative interpretation criteria to determine their “capacities”. The results of each of the interpretation criteria were next normalized by a standard interpretation criterion and a predicted capacity to establish the relationships between these various interpretation criteria and to characterize model factors, respectively. The analyses of 143 drilled shafts indicated that L1 and DeBeer interpreted loads can be recommended for the serviceability limit state design. The criteria van der Veen, 4%B, Terzaghi and Peck, slope–tangent, L2, and Fuller and Hoy yielded values that are at the transition region of the load–displacement curve, in which the “ultimate” capacity can be reasonably estimated. The Chin method is not conservative, because it produced a value larger than all of the above interpreted “ultimate” capacities with displacements exceeding 60 mm on the average and beyond the range of the measured load–displacement curve. The scatter in the load–displacement curves is studied by applying different normalization schemes and fitting the normalized curves to a hyperbolic model. The hyperbolic model (curve fitting) parameters are generally negatively correlated. Statistics of these hyperbolic model parameters are shown to be comparable to values reported in previous studies. This hyperbolic model is widely used for reliability-based design at the serviceability limit state. Hence, the statistics of the hyperbolic model parameters presented in this study are useful.