Graduation Semester and Year




Document Type


Degree Name

Doctor of Philosophy in Civil Engineering


Civil Engineering

First Advisor

Xinbao Yu


The Federal Highway Administration (FHWA) released a policy in 2000 that required all new federally funded bridges to be designed using the AASHTO LRFD specifications by October 2007. The transition from ASD to LRFD posed a challenge due to the lack of area specific resistance factors. Though several studies were performed to calibrate area specific resistance factors, they did not improve from the resistance factors suggested in AASHTO 2012. The objective of this study is to analyze the uncertainties in LRFD calibration and to calibrate more accurate and improved resistance factors. A drilled shaft load test database from Mississippi and Louisiana has been selected to carry on the research. Osterberg Cell load test was a majority among the load tests in the database. Extrapolation is required in most Osterberg cell load test which may cause errors in the calibration. An analysis of the error due to the extrapolation can result in more accurate LRFD calibration of resistance factors. The analysis was performed on 8 drilled shaft cases from Louisiana and Mississippi. 4 of the 8 drilled shafts reached 5% of the shaft diameter (D) failure criterion and 4 of the drilled shafts were close to 5%D. For each of the cases, extrapolation was performed on tip and side resistance curves to get the equivalent top-down curve. Data points were removed systematically from the end of top and bottom movement curves and extrapolation was performed for each trial to get an equivalent top-down curve. Bias and error values was measured for each of the trial top down scurves for both 5% of the drilled shaft diameter (D). 80 extrapolation cases were achieved from this analysis. Finally, multiple linear regression analysis was performed on the extrapolated data set in order to reduce the effect of the extrapolation error on the resistance factor. Applying bounded bias distribution may also result in more accurate resistance factors, since the bias values have significant role in the calibration process. As the probability of failure significantly depends on the lower tail of the distribution of the resistance values and there is a physical presence of a lower limit of the resistance of a drilled shaft, introducing a lower bound to the resistance distribution will ensure more realistic calibration of the resistance factors. An analysis by simulation of load tests to failure on will also help to understand the reasons for low resistance factor values. The objective of this study is also to minimize effect of extrapolation by means of finite element modelling of the bidirectional load tests included in a database collected from Louisiana and Mississippi. The finite element modelling was performed in PLAXIS 2D until the top and the bottom movement curves reach the measured loads corresponding to the failure criteria. LRFD calibration of resistance factors was performed based on the simulated bidirectional load test database and the results were compared to the results from conventional approach.


LRFD calibration, Drilled shaft, Load testing, Resistance factors


Civil and Environmental Engineering | Civil Engineering | Engineering


Degree granted by The University of Texas at Arlington