Adaptive Finite Difference Methods and Implementation in the Software DUNE
Authors/Creators
Description
The diffusion equation is a partial differential equation that describes how
quantities such as heat, particles, or other diffusive substances spread out over
time and space. This type of equation can be, at often times, quite challenging
to solve analytically. Finding the exact solution is often difficult due to the com-
plexity of the functions involved. This thesis focuses on solving the diffusion
equation using an adaptive finite difference method. We will be exploring two
different methods that would produce a numerical solution. The concept of ad-
aptivity optimizes our solution in a way that enhances accuracy and decreases
both computational power and time. We will also be working with the DUNE
software which is particularly helpful in creating adaptive grids. The software
enables us to access the grid cells and it’s neighbors which will be useful when
creating our solver.
Abstract
As an artist, choosing the right brush is essential. Large brushes are useful for broad, uniform areas such as a blue sky, while smaller brushes allow for precise work and fine details. A similar principle can be applied to solving complex mathematical problems.
Partial differential equations (PDEs) describe many physical phenomena, but finding analytical solutions can be difficult or even impossible. The finite difference method provides a practical alternative by replacing continuous space with a grid of discrete points, allowing us to compute an approximate solution.
However, using a fine grid everywhere can be computationally expensive and unnecessary. Instead, we can use an adaptive approach that strategically places more grid points where the solution changes rapidly and fewer points where it remains relatively smooth.
In this project, we develop a solver for the diffusion equation that applies this idea of adaptive resolution. Much like an artist selecting the appropriate brush for each part of a painting, the solver determines where a coarse or fine grid is needed, aiming to achieve an accurate solution while reducing unnecessary computational work.
Files
Adaptive Finite Difference Methods and Implementation in the Software DUNE.pdf
Additional details
Software
- Repository URL
- https://github.com/Kraneee25/Adaptive-Finite-Difference-Method---Diffusion-Equation
- Programming language
- Python