Date of Award
deep foundation, scour prediction, reliability analysis
This research project studied the reliability of driven pile foundations subjected to scour, gravity loads including dead loads and live loads, and their combinations. The design of the foundation followed the procedures of the AASHTO LRFD Bridge Design Specification. The design methods were the HEC-18 pier scour equation used to determine the scour depth, and the α-Method and Nordlund-Thurman Method used to design the driven pile. The limit state considered was the axial bearing resistance of the piles in the soil. The reliability indexes for bridges subjected to maximum expected scour depth and gravity loads were found through Monte Carlo Simulations. First, the reliability of the HEC-18 pier scour equation alone was investigated. It was found that the uncertainty of the model bias for the HEC-18 equation has a significant impact on the reliability index; the reliability index is 1.47 in average when using the NCHRP 489 model bias, and 2.32 when using the NCHRP 24-34 model bias. The research found that the upper limit provision of the HEC-18 pier scour equation can reduce the reliability index by a significant amount. The research also looked into the effect of the channel shape on the reliability index. It was found that for a channel with isosceles trapezoidal cross section, the channel side slope has negligible influence on the reliability index, while for a channel built by trapezoidal main channel with over banks, the reliability index can change when the channel slopes change. Next, the reliability of the α-Method and Nordlund-Thurman Method for pile analysis was evaluated, for the case when the foundation is subjected to gravity loads alone. Two different simulation methods were applied to check the reliability index accounting for the uncertainties in estimating the foundation capacities and the loads. Simulation Method 1 uses the probability that the pile resistance is larger than the applied loads as benchmark; while in Simulation Method 2, the probability that the design pile depth selected to carry the design loads exceeds the required pile depth needed to carry the expected simulated loads is used to determine the reliability index. The two simulations methods produced similar results. For the α-Method, the reliability index was found to be 2.26 for the AASHTO resistance factor of 0.35. For the Nordlund-Thurman Method, the reliability index was found to be 2.10 for the resistance factor of 0.45. These results were in agreement with those of the NCHRP 507 report. Finally, the combined reliability of piles subjected to the combination of scour and gravity loads was evaluated base on the Ferry-Borges model. Again, two simulation methods were used to check for the reliability index. It was found that the α-Method combined with HEC- 18 equation has a reliability index of 2.22 to 2.24 when Simulation Method 1 is used, and 2.10 to 2.14 when Simulation Method 2 is used. For the Nordlund-Thurman Method, the reliability indexes were 2.04 for Method 1 and 2.22 for Method 2. The scour factor of 1.00, 0.75, 0.50, and 0.25 were successively applied in the simulations to check for the effects of a scour factor on the reliability index. It was found that the reliability index reduced by 0.20 in average for Method 1 and 0.40 for Method 2 when the scour factor was reduced from 1.00 to 0.25. The sensitivity of the reliability to the change in dead and live load factors was also checked. It was found that when the live load factor was changed from the current 1.75 to 0.75, the reliability index reduced by 0.20; and when the dead load factor was reduced from the current 1.25 to 0.50, the reliability index reduces by 1.40 to 1.50 in average.
Yan, Jun, "Reliability Analysis of the HEC-18 Scour Equation and AASHTO Deep Foundation Design Codes" (2013). CUNY Academic Works.