OPTIMIZATION OF PILE LENGTH USING DECISION ANALYSIS WITH INFORMATION FROM LOAD TESTS DURING CONSTRUCTION

摘要

Pile load tests are regularly used to verify the capacities of piles. They can also be used to optimize pile designs. This paper presents a case study based on a real project, in which the information obtained from load tests performed during construction was used to shorten the design pile lengths without a formal decision analysis. The design pile lengths were determined using the information obtained from subsurface explorations, based on the adopted design method for the project and engineering judgment. In this study, a simple probabilistic model was used to access the probabilities of failure, defined as the loss of serviceability of the pile-supported structure due to excessive settlement and differential settlement, for a range of pile lengths based on design information. These probabilities were updated using the Bayesian approach given the additional information from load tests. The probabilities of failure were then used in a formal decision analysis to determine the optimal pile lengths based on design information and based on the additional information from load tests. These optimal pile lengths were compared with those implemented in the project without a formal decision analysis to demonstrate its value.