Calibration of Resistance Factors Needed in the LRFD Design of Drilled Shafts

LTRC Number: 07-2GT
State Project Number: 736-99-1408
The first report on Load and Resistance Factor Design (LRFD) calibration of driven piles in Louisiana (LTRC Final Report 449) was
completed in May 2009. As a continuing effort to implement the LRFD design methodology for deep foundations in Louisiana, this
report will present the reliability based analyses for the calibration of the resistance factor for LRFD design of axially loaded drilled
shafts. A total of 16 cases of drilled shaft load tests were available to authors from Louisiana Department of Transportation and
Development (LADOTD) archives. Out of those, only 11 met the Federal Highway Administration (FHWA) “5%B” settlement
criterion. Due to the limited number of available drilled shaft cases in Louisiana, additional drilled shaft cases were collected from
state of Mississippi that has subsurface soil conditions similar to Louisiana soils. A total of 15 drilled shafts from Mississippi were
finally selected from 50 available cases, based on selection criteria of subsurface soil conditions and final settlement. As a result, a
database of 26 drilled shaft tests representing the typical design practice in Louisiana was created for statistical reliability analyses.
The predictions of total, side, and tip resistance versus settlement behavior of drilled shafts were established from soil borings using
the FHWA O’Neill and Reese design method via the SHAFT computer program. The measured drilled shaft axial nominal resistance
was determined from either the Osterberg cell (O-cell) test or the conventional top-down static load test. For the 22 drilled shafts that
were tested using O-cells, the tip and side resistances were deduced separately from test results. Statistical analyses were performed
to compare the predicted total, tip, and side drilled shaft nominal axial resistance with the corresponding measured nominal
resistance. Results of this showed that the selected FHWA design method significantly underestimates measured drilled shaft
resistance. The Monte Carlo simulation method was selected to perform the LRFD calibration of resistance factors of drilled shaft
under strength I limit state. The total resistance factors obtained at different reliability index () were determined and compared with
those available in literature. Results of reliability analysis, corresponding to a target reliability index () of 3.0, reveals resistance
factors for side (side), tip (tip), and total resistance factor (total) are 0.20, 0.75, and 0.5, respectively.