Thesis

Finite element methods for 2D induction heating problems

Creator
Rights statement
Awarding institution
  • University of Strathclyde
Date of award
  • 2026
Thesis identifier
  • T17594
Person Identifier (Local)
  • 202175209
Qualification Level
Qualification Name
Department, School or Faculty
Abstract
  • Numerical modelling of induction heating is a challenge for many reasons. For instance, there are difficulties in defining the functional setting, resolving non-linear material properties, and capturing steep boundary layers. More specifically, for example, in a simplified 2D domain, the variational formulation is not necessarily well-defined due to a term in the heat equation belonging to L1(Ω) only. In this Thesis, we have two main focuses: conducting mathematical analysis on this 2D induction heating problem that deals with the irregular right-hand side, and building code to simulate realistic induction heating processes. In our analysis, by showing that the coupling leads to a right-hand side that is more regular than L1(Ω), we prove existence of solutions in a setup that is more general than the existing literature. Under strict assumptions on the shape of the mesh, we prove convergence of the standard Galerkin finite element method. The main new result we present is a proof of convergence of a recently developed finite element method to a decoupled problem under no assumptions on the mesh. The fact that this method has no conditions on the mesh allows the resolution of boundary layers in more complicated geometries. We also implement the standard finite element method and this new method using the Python software FEniCSx. We study two different 2D models, and show that these codes produce results that are comparable to experimental data, and also improve upon some results from finite element codes used by industry. We conclude by running the model of induction heating over an irregular mesh, and explicitly demonstrate the results we have proven, that the newly developed method is robust over irregular meshes.
Advisor / supervisor
  • Barrenechea, Gabriel R.
Resource Type
DOI
Date Created
  • 2025

Relations

Items