Statistical analysis of O-cell test data for nominal load capacities of drilled shafts

The Osterberg cell (O-cell) test has been increasingly used for evaluating load capacities of drilled shafts. In this test, the pre-installed O-cell jack simultaneously produces an upward force to the upper portion of the shaft and a downward force to the lower portion of the shaft at an equal magnitude, which can be used to estimate the side resistance and the tip resistance of the shaft separately. In this study, 25 O-cell test datasets were collected for the test shafts installed in rock or weak rock. Seven interpretation methods ("Davisson's, Chin's, Butler and Hoy's, FHWA 0.05D, Fuller and Hoy's, Brinch-Hansen's 80%, and creep limit method) were used to estimate the nominal bearing capacities of all drilled shafts based on the equivalent "top-down" curves. The mean value of the nominal load capacities based on these seven methods was considered as the representative capacity. A "bias" was calculated as the ratio of the nominal load capacity by a specific interpretation method to the representative load capacity. Statistical analysis was performed for all seven methods. It is concluded that the FHWA 0.05D and Butler and Hoy's methods resulted in the mean bias close to 1.00 and the small COV value. The FHWA 0.05D method is recommended for the reliability analysis for the resistance factor of drilled shafts based on O-cell test data. Copyright ASCE 2008.