- Title
- Bayesian back analysis for embankments on soft soils considering creep
- Creator
- Huang, Shan
- Relation
- University of Newcastle Research Higher Degree Thesis
- Resource Type
- thesis
- Date
- 2023
- Description
- Research Doctorate - Doctor of Philosophy (PhD)
- Description
- In coastal areas, embankments are inevitable to be constructed over soft soils and marine clays for supporting transportation infrastructures. Accurate time-dependent settlement predictions are crucial for mitigating the risks associated with unexpected delays, cost overruns, and potential stability concerns. Due to the complex time-dependent behaviours of soft soils, accurately predicting long-term settlement remains a challenging task. In this thesis, a settlement prediction framework combining the Bayesian back analysis approach and one-dimensional consolidation analysis incorporating creep is proposed and is applied in predicting the long-term response of soft soils to external loads. The simplified consolidation analysis with analytical solution, and the rigorous consolidation analysis with numerical solution are employed to perform consolidation analysis of PVD-improved soft soils incorporating creep. The effectiveness of different consolidation analyses within the settlement prediction framework for long-term settlement prediction of soft soils is investigated. The influence of the coupling effects of soil deformation and pore water dissipation on Class C prediction of soft soil settlement is also examined. Additionally, the impact of various types of monitoring data on the performance of the consolidation model with different solution schemes within the settlement prediction framework is explored. To improve the efficiency of the settlement prediction framework, an optimization process within the Bayesian back analysis approach is carried out. Due to the high nonlinearity of the Bayesian back analysis problem, it should be best solved by the sampling methods. The classical Markov chain Monte Carlo (MCMC), ensemble MCMC, nested sampling, and Bayesian updating with the structural reliability method (BUS) are employed to solve the Bayesian back analysis problem. A comparison of the sampling methods is conducted considering factors such as running time and the number of likelihood function calls. The optimization within the settlement prediction framework can provide practical guidance for selecting an appropriate soil consolidation model and in selecting an appropriate sampling method for solving Bayesian back analysis problems based on practical circumstances encountered. The proposed settlement prediction framework has been successfully applied to perform long-term settlement prediction for a trial embankment constructed on PVD-improved soft clays in Ballina NSW, Australia. The obtained results demonstrate the effectiveness of the proposed settlement prediction framework. It indicates that the proposed prediction framework can assist engineers and practitioners in achieving cost-effective outcomes throughout both the design and construction phases.
- Subject
- embankment settlement prediction; bayesian back analysis; soft soil creep; one-dimensional consolidation; PVDs; thesis by publication
- Identifier
- http://hdl.handle.net/1959.13/1496512
- Identifier
- uon:54170
- Rights
- This thesis is currently under embargo and will be available from 22.11.2024, Copyright 2023 Shan Huang
- Language
- eng
- Hits: 483
- Visitors: 430
- Downloads: 1
| Thumbnail | File | Description | Size | Format |
|---|