Probabilistic Analysis of Bored Pile Foundations in the Design Phase: An Application Example

Abstract

This paper presents a probabilistic analysis of pile foundations in the design phase considering Reliability Theory in order to improve the final decision about the foundation system. Two semi-empirical predicting methods of pile bearing capacity were used and the Standard Penetration Test (SPT) was applied for site characterization and definition of the uncertainties on the soil strength parameters. The First Order Reliability Method (FORM) was adopted for assessing the probability of failure in the reliability analyses phase. The design of the foundation elements for a three floors hypothetical building designed in reinforced concrete and installed in a sandy soil profile is presented to show how the proposed probabilistic approach can be applied in a practical case. The distributions of the random variables defined by the NSPT values as well as the shaft and base resistances of the piles were calculated by the two mostly used semi-empirical method in Brazil for bearing capacity prediction. The reliability analysis allowed to identify the differences in the prediction methods and provided the probability estimative for each isolated pile. It was observed that the values for reliability indexes and failure probabilities of isolated foundation elements depend directly on the approach for determining the variability of the pile-soil resistance parameters. The main advantage of this approach is to adapt the foundation design, in which the number of piles, as well as its diameter or length can be redefined to achieve the chosen target probability of failure.