Back to search

National Technical University of Athens (NTUA)

Advanced high performance computing methods for the solution of crack propagation and material design problems using the extended finite element method (XFEM)

Abstract

dc:description

The need for advanced high-performance materials in the industry led to the development of various innovative solutions over the years, designed to possess application-specific properties, such as improved thermal conductivity. To model heat transfer in composite materials, their complex micro-structure, as well as the thermal resistance at the interfaces between materials must be taken into account. The standard finite element treatment requires very fine meshes to conform to the complex geometry of these interfaces. This thesis proposes an eXtended Finite Element Method (XFEM) formulation that captures the temperature jump by enriching the polynomial approximation around the material interfaces with appropriate discontinuous functions. Specifically, a new XFEM enrichment scheme is developed to address the issue of multiple-phase junctions, namely areas where multiple interfaces with different resistance properties intersect. In addition, a double-mesh LSM technique is developed for describing the geometry of material interfaces. A very fine mesh is employed by a Level Set Method (LSM) to represent complex interface geometries with high accuracy, whereas XFEM uses a mesh that does not conform to the material interfaces, but is instead a coarser version of the LSM mesh, to reduce the computational cost of the analysis. The combined numerical model is first validated against existing results from the literature on polycrystalline materials. Then, it is applied for heat conduction analysis of polymers reinforced with carbon-nanotubes. The unknown thermal resistance between these materials is inferred by calibrating the numerically predicted effective conductivity to corresponding experimental measurements. The proposed XFEM model can be straightforwardly extended to other similar problem types, such us elasticity or electrical conduction. XFEM is also an attractive choice for modeling crack propagation, by enriching the polynomial displacement field of FEM with specialized non-smooth functions, without the need of remeshing in the vicinity of the crack at each propagation step. However, this enrichment causes the stiffness matrix to become strongly ill-conditioned, rendering the convergence of iterative solvers very slow. On the other hand, direct solvers are inefficient in 3D problems, due to the increased bandwidth of the system matrix. In this thesis, two domain decomposition solvers, namely FETI-DP and P-FETI-DP, are proposed for solving the linear systems resulting from XFEM crack propagation analysis in large-scale 3D problems. By modifying the coarse problem of both solvers, any singularities caused by the crack propagation are avoided and the XFEM-related ill-conditioning is completely eliminated, ensuring the scalability of FETI-DP and P-FETI-DP as the number of subdomains is increased. Finally, an efficient implementation in high performance computing systems, specifically computer clusters is developed, by altering the original FETI-DP and P-FETI-DP equations to minimize communication and computation bottlenecks in distributed memory environments.

Degree

thesis:*
Grantor dc:publisher
National Technical University of Athens (NTUA)
Year dc:date
2022

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Bakalakos, Seradeim
  • Μπακαλάκος, Σεραφείμ

Subjects

dc:subject × 22

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/52392

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Bakalakos, Seradeim; Μπακαλάκος, Σεραφείμ. Advanced high performance computing methods for the solution of crack propagation and material design problems using the extended finite element method (XFEM). National Technical University of Athens (NTUA), 2022. http://hdl.handle.net/10442/hedi/52392