Computational investigation of fracture in polymer composites and heterogeneous dry adhesives
File(s)
Author(s)
Fan, Wei
Type
Thesis
Abstract
Epoxy composites are widely used because of their excellent properties, but understanding the toughening mechanisms is of great significance to their successful application. Computational solid mechanics utilizing numerical approaches to simulate the mechanical behaviour of materials have become fundamental to materials design and investigation of mechanisms. The finite element method is the most widely used of these, but recent advances in artificial intelligence bring a new possibility to solid mechanics, the data-driven approaches.
In this thesis, the crack phase field method is adopted to investigate the mechanical properties and toughening mechanisms in epoxy based interpenetrating phase composites and core-shell particle reinforced composites, and a generalized approach is used so that the results may be applied to other composites. Various tougheners are applied to identify their effect on the mechanical performance, and the fracture process is discussed from the stress-strain aspect and the crack topology aspect.
A deep learning based framework to predict crack propagation patterns and stress-strain curves in composite materials is introduced. By implicitly encapsulating the relationship between the microstructure, crack behaviours, and material properties, this tailored approach substantially reduces the demand for training dataset.
Finally, this thesis applies simulations to thoroughly analyse the adhesion process of bio-inspired dry adhesives on rough surfaces. Through a comprehensive comparison with experimental results, the adhesive mechanisms of the heterogeneous dry adhesives are elucidated.
In this thesis, the crack phase field method is adopted to investigate the mechanical properties and toughening mechanisms in epoxy based interpenetrating phase composites and core-shell particle reinforced composites, and a generalized approach is used so that the results may be applied to other composites. Various tougheners are applied to identify their effect on the mechanical performance, and the fracture process is discussed from the stress-strain aspect and the crack topology aspect.
A deep learning based framework to predict crack propagation patterns and stress-strain curves in composite materials is introduced. By implicitly encapsulating the relationship between the microstructure, crack behaviours, and material properties, this tailored approach substantially reduces the demand for training dataset.
Finally, this thesis applies simulations to thoroughly analyse the adhesion process of bio-inspired dry adhesives on rough surfaces. Through a comprehensive comparison with experimental results, the adhesive mechanisms of the heterogeneous dry adhesives are elucidated.
Version
Open Access
Date Issued
2023-10
Date Awarded
2024-05
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
Taylor, Ambrose
Publisher Department
Mechanical Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)