An integrated approach to numerical modelling of metal fatigue crack propagation
File(s)
Author(s)
Soyemi, Adeleke
Type
Thesis
Abstract
Engineering structures, especially ones that are built-up such as steel bridges and offshore structures, usually contain imperfections in the form of material discontinuities or cracks, arising from the fabrication process, inherent material defects or weak welded or bolted joints. In many cases, this leads to the initiation and propagation of fatigue cracking under repeated loading cycles which require urgent attention. Thus, it has become increasingly necessary to undertake regular inspections in order to prevent the service breakdown of such structures that can potentially bring significant economic risk. Due to limited resources, these checks require an optimised inspection schedule, which can only be achieved through in-depth understanding of the fatigue crack behaviour of any identified cracks.
Realising the need for an efficient strategy to undertake such studies, this work proposes a mechanics-based framework that allows realistic cycle-by-cycle fatigue crack growth analysis to be undertaken through an accelerated computation of damage under cyclic loading. A novel material integration (MI) technique is proposed in this work, which computes an accelerated fatigue damage over a number of cycles by integrating the constitutive behaviour of a material point undergoing fatigue damage. Comparisons with existing techniques such as the linear extrapolation technique (LE) and the linear scaling (LS) technique demonstrates its capability to approximate the acceleration of fatigue damage over a sizeable block of constant amplitude loading cycles, thereby achieving a faster computation for an otherwise impracticable numerical cycle-by-cycle solution.
In addition, this study has also developed a new framework that is capable of adaptive simulation of fatigue crack growth studies with the purpose of saving computational time, hence targeted towards large scale analyses. This framework is based on a novel shell element formulation allowing for embedded discontinuities, which is coupled with criteria for crack propagation and orientation over a mesh of elements. Verification studies are undertaken on realistic components to demonstrate the accuracy of the proposed adaptive analysis procedure, and an application study on a representative steel deck substructure is presented to illustrate the applicability of the proposed approach to real problems.
Realising the need for an efficient strategy to undertake such studies, this work proposes a mechanics-based framework that allows realistic cycle-by-cycle fatigue crack growth analysis to be undertaken through an accelerated computation of damage under cyclic loading. A novel material integration (MI) technique is proposed in this work, which computes an accelerated fatigue damage over a number of cycles by integrating the constitutive behaviour of a material point undergoing fatigue damage. Comparisons with existing techniques such as the linear extrapolation technique (LE) and the linear scaling (LS) technique demonstrates its capability to approximate the acceleration of fatigue damage over a sizeable block of constant amplitude loading cycles, thereby achieving a faster computation for an otherwise impracticable numerical cycle-by-cycle solution.
In addition, this study has also developed a new framework that is capable of adaptive simulation of fatigue crack growth studies with the purpose of saving computational time, hence targeted towards large scale analyses. This framework is based on a novel shell element formulation allowing for embedded discontinuities, which is coupled with criteria for crack propagation and orientation over a mesh of elements. Verification studies are undertaken on realistic components to demonstrate the accuracy of the proposed adaptive analysis procedure, and an application study on a representative steel deck substructure is presented to illustrate the applicability of the proposed approach to real problems.
Version
Open Access
Date Issued
2020-01
Date Awarded
2020-05
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
Izzuddin, Bassam
Sponsor
Petroleum Technology Development Fund, Nigeria
Publisher Department
Civil and Environmental Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)