LIST researcher awarded by Chinese Government

Published on 03/03/2015

 

LIST has the pleasure to announce that Dr Fan Xu from its "Materials Research and Technology" (MRT) department has obtained the 2014 Chinese government award "Outstanding self-financed PhD thesis abroad". His thesis was about the numerical study of instability patterns of film-substrate systems and was successfully defended in December 2014.

About the thesis

Surface wrinkles of stiff thin layers attached on soft materials have been widely observed in nature and these phenomena have raised considerable interests over the last decade. The post-buckling evolution of surface morphological instability often involves strong effects of geometrical nonlinearity, large rotation, large displacement, large deformation, loading path dependence and multiple symmetry-breakings. Due to its notorious difficulty, most nonlinear buckling analyses have resorted to numerical approaches since only a limited number of exact analytical solutions can be obtained. This thesis proposes a whole framework to study the film/substrate buckling problem in a numerical way: from 2D to 3D modeling, from classical to multi-scale perspective. The main aim is to apply advanced numerical methods for multiple-bifurcation analyses to various film/substrate models, especially focusing on post-buckling evolution and surface mode transition. The models incorporate the Asymptotic Numerical Method (ANM) as a robust path-following technique and bifurcation indicators well adapted to the ANM to detect a sequence of multiple bifurcations and the associated instability modes on their post-buckling evolution path. The ANM gives interactive access to semi-analytical equilibrium branches, which offers considerable advantage of reliability compared with classical iterative algorithms. Besides, an original nonlocal coupling strategy is developed to bridge classical models and multi-scale models concurrently, where the strengths of each model are fully exploited while their shortcomings are accordingly overcome. Discussion on the transition between different scales is provided in a general way, which can also be seen as a guide for coupling techniques involving other reduced-order models. Lastly, a general macroscopic modeling framework is developed and two specific Fourier-related models are derived from the well-established classical models, which can predict the pattern formation with much fewer elements so as to significantly reduce the computational cost.

Share this page:

Contact

Luxembourg Institute of Science and Technology (LIST)

5, avenue des Hauts-Fourneaux
L-4362 Esch-sur-Alzette

Send an e-mail