Calibration of resistance factors for load and resistance factor design of driven piles for bridge foundations.

摘要

The resistance factors are required in the Load and Resistance Factor Design (LRFD) of driven piles. In this study, the focus is on the design of driven piles for bridge foundations, and thus the load factors are assumed to follow those recommended by AASHTO Bridge Design Specifications. The resistance factors for various pile capacity prediction models are calibrated based on the concept of reliability theory using a database of 125 pile load tests. The statistics of the bias factor for a given empirical pile capacity model, defined as the ratio of the predicted capacity over the measured capacity, can be calculated using the database of pile load test cases. In this study, the following pile capacity prediction models were calibrated for the corresponding resistance factors: the AASHTO method, the API method, the Briaud method, the Coyle method, the Meyerhof method, the SPT-97 method, the BPNN method, and the GRNN method. First Order Reliability Method (FORM) and First Order Second Moment (FOSM) method are employed for calculation of reliability indices and calibration of resistance factors. The resistance factors are determined for the target reliability index values.; If additional information such as within-site pile load tests is available for a specific site, the prior statistics about the bias factor for a given pile capacity prediction model may be updated using Bayes' theorem, which has been well recognized as an effective tool for combining new data with previous data to revise the assessment of uncertainty and reliability. Bayesian updating method can be employed to improve the statistics of the bias factors of a given pile capacity prediction model. In this study, the statistics of the bias factor determined for a given pile capacity prediction model are used to form the prior distribution, and the typical pile load test statistics are used to form the likelihood distribution. The Bayesian updating is then applied to obtain the posterior distribution of the bias factor for a given pile capacity prediction model. The significance of the updated (posterior) distribution is interpreted.; This dissertation presents the results of a series of fundamental examination of various issues related to design of driven piles for bridge foundations using LRFD. The results of the study improve the understanding of LRFD and provide calibrated resistance factors that can readily be used in conjunction with each of the eight pile capacity prediction methods for a given target reliability index. The results of this study also clearly demonstrate the significance and advantage of within-site pile load tests. Being able to raise the resistance factor at the same target reliability index (i.e., at the same safety level) due to the availability of within-site pile load tests allows the engineer to reduce pile length or number of piles in a pile group, and thus helps to reduce the project costs.