Reliability assessment of laterally loaded single piles

The deterministic solution of laterally-loaded piles is implicit and often requires an iterative numerical procedure to obtain. On the other hand, parametric uncertainty is inherent in geotechnical engineering. While the uncertainties can be accounted for by adopting probabilistic methods, extendi...

Full description

Saved in:
Bibliographic Details
Main Author: Chan, Chin Loong
Other Authors: Low Bak Kong
Format: Theses and Dissertations
Language:English
Published: 2012
Subjects:
Online Access:https://hdl.handle.net/10356/48201
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:The deterministic solution of laterally-loaded piles is implicit and often requires an iterative numerical procedure to obtain. On the other hand, parametric uncertainty is inherent in geotechnical engineering. While the uncertainties can be accounted for by adopting probabilistic methods, extending the deterministic analysis to a probabilistic study remains complicated and challenging. The objective of this study is to present some practical approaches of performing reliability assessment of laterally-loaded piles. A direct stiffness algorithm for the laterally-loaded pile is developed, which is able to model both nonlinear soil behavior as well as nonlinear pile flexural rigidity. The deterministic framework underlies the implicit limit state functions used in subsequent reliability analysis. Deflection and bending moment failure modes are investigated. Single-failure-mode analyses are extended to system reliability analysis in which bi-modal system failure probability bounds are computed. Practical procedures for the second-order reliability method and for the Monte Carlo method with importance sampling are proposed. An efficient method to estimate probabilistic sensitivities is introduced. The reliability approach is integrated with standalone finite-element programs via the response surface and the artificial neural network methodologies.