IRUS Total

Modelling solid/fluid interactions in hydrodynamic flows: a hybrid multiscale approach

File Description SizeFormat 
RamosFernandez-E-2021-PhD-Thesis.pdfThesis39.59 MBAdobe PDFView/Open
Title: Modelling solid/fluid interactions in hydrodynamic flows: a hybrid multiscale approach
Authors: Ramos Fernandez, Eduardo
Item Type: Thesis or dissertation
Abstract: With the advent of high performance computing (HPC), we can simulate nature at time and length scales that we could only dream of a few decades ago. Through the development of theory and numerical methods in the last fifty years, we have at our disposal a plethora of mathematical and computational tools to make powerful predictions about the world which surrounds us. From quantum methods like Density Functional Theory (DFT); going through atomistic methods such as Molecular Dynamics (MD) and Monte Carlo (MC), right up to more traditional macroscopic techniques based on Partial Differential Equations (PDEs) discretization like the Finite Element Method (FEM) or Finite Volume Method (FVM), which are respectively, the foundation of computational Structural Analysis and Computational Fluid Dynamics (CFD). Many modern scientific computing challenges in physics stem from combining appropriately two or more of these methods, in order to tackle problems that could not be solved otherwise using just one of them alone. This is known as multi-scale modeling, which aims to achieve a trade-off between computational cost and accuracy by combining two or more physical models at different scales. In this work, a multi-scale domain decomposition technique based on coupling MD and CFD methods, has been developed to make affordable the study of slip and friction, with atomistic detail, at length scales otherwise impossible by fully atomistic methods alone. A software framework has been developed to facilitate the execution of this particular kind of simulations on HPC clusters. This have been possible by employing the in-house developed CPL_LIBRARY software library, which provides key functionality to implement coupling through domain decomposition.
Content Version: Open Access
Issue Date: Nov-2020
Date Awarded: Sep-2021
URI: http://hdl.handle.net/10044/1/99788
DOI: https://doi.org/10.25560/99788
Copyright Statement: Creative Commons Attribution NonCommercial NoDerivatives Licence
Supervisor: Dini, Daniele
Heyes, David
Sponsor/Funder: BP-ICAM
Engineering and Physical Sciences Research Council (EPSRC)
Department: Materials
Publisher: Imperial College London
Qualification Level: Doctoral
Qualification Name: Doctor of Philosophy (PhD)
Appears in Collections:Materials PhD theses

This item is licensed under a Creative Commons License Creative Commons