Scribd
Upload
20 views3 pages

Course-Content2024 Deal

Uploaded by

zawarikhan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
20 views3 pages

Course-Content2024 Deal

Uploaded by

zawarikhan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

FINITE ELEMENT METHOD IN SCIENTIFIC COMPUTING: A SHORT COURSE

1.0 Overview

The finite element method is a conceptual framework for the numerical approximation of solutions of
partial differential equations (PDE). It is based on the reformulation of the PDE in weak form, and
then uses a piecewise polynomial approximation of the exact solution. To this end, it subdivides the
domain on which the PDE is posed (i.e., in which you want to describe the physical process) into a
“mesh” of “finite elements” – i.e., “cells” of finite size (as opposed to infinitesimally small). On each
of these cells, one then approximates the solution by a polynomial. The result is that we can describe
the solution by the finitely many expansion coefficients of these polynomials. Because finitely many
unknowns can be stored and solved for in finite time, this method is now amenable to computation. In
this short course, the students will be introduced to some of the ideas of how this conversion from an
unsolvable, infinite dimensional problem (the PDE) to a solvable, finite dimensional problem (the
finite element approximation) works for some typical equations, and how all of this is represented on
a computer. The focus of this course will be on the practical side of the finite element method and it
will emphasize the modern algorithms and approaches for its implementation using deal.II as an
example.

An important part of this course is to show students how finite element methods look like in
implementation practice today. For this, the deal.II library will be used (see http://www.dealii.org/).
deal.II is a C++ software package that provides building blocks from which one can assemble finite
element solvers in much the same way as one writes MATLAB programs: it offers a wide variety of
data structures and algorithms for everything one typically needs in finite element codes, but how
these are put together is left to the user – just like MATLAB offers matrix and vector data structures
along with things like singular value decompositions, Fourier transforms, etc., but it is left to the user
what to do with all of this.

deal.II is a project that is today also the largest and most widely used open source finite element
software. It has grown to more than one million lines of C++, has thousands of users, and a developer
community that contributes hundreds of patches every month. It also has excellent documentation,
and the course will show students how to navigate it and in particular walk them through some of the
many tutorial programs that explain how deal.II can be applied to particular partial differential
equations in any domain of interest.

The short course will also have plenty of time for the students to play with deal.II: explore what
happens if one changes the equation, the boundary values, the domain, the right hand side; how best
to visualize the solution.

2.0 Objectives

The primary objectives of the course are as follows:


i) Train students on writing modern finite element codes,
ii) Train students on using several modern software tools that will help them with writing finite
element codes,
iii) Describe the approach to building a finite element approximation for whatever problem one is
interested in,
iv) Demonstrate deal.II as a general purpose finite element library to solve partial differential
equations.

3.0 Course details

Dates: April 29 – May 03, 2024 (5 days): 12 lecture hours and 9 tutorial hours
Lecture schedule
Monday, April 29, 2024

Lecture 1: 1 hrs:
• Course overview; why we use software libraries
• An introduction to the finite element method, part 1

Lecture 2: 1 hrs:
• An introduction to the finite element method, part 2

Lecture 3: 1 hrs:
• A brief introduction to deal.II

Tutorial 1: 1 hrs:
• Lab activity: Working on the command line;
• Lab activity: Getting started with installing deal.II

Lecture 4: 1 hrs:
• Generating meshes in deal.II

Tuesday, April 30, 2024

Lecture 5: 1 hrs:
• Allocating degrees of freedom in deal.II

Tutorial 2:1 hrs:


• Executing step-2 tutorial program in deal.II and visualize the DoF locations and sparsity
patterns for a given mesh

Lecture 6: 1 hrs:
• Solving Laplace’s equation using the Finite Element Method

Tutorial 3: 1 hrs:
• Executing step-3 tutorial program in deal.II to solve Laplace’s equation and visualize
solutions in Paraview

Wednesday, May 1, 2024

Lecture 7: 1 hrs:
• Using templates in C++
• Applying templates to solve Laplace’s equation in deal.II in a dimension independent way

Tutorial 4: 1 hrs:
• Executing the step-4 tutorial program in deal.II to solve Laplace’s equation in both two-
dimensional and three-dimensional domains

Lecture 8: 1 hrs:
• Generating adaptively refined meshes, treatment of hanging nodes in the finite element
method, and constraint enforcement

Tutorial 5: 1 hrs:
• Executing the step-6 tutorial program in deal.II to solve Laplace’s equation using
adaptively refined meshes
Thursday, May 2, 2024

Lecture 9: 1 hrs:
• Solving problems with more than one solution variable (“vector valued problems”) with
linear elasticity as an example

Tutorial 6: 1 hrs:
• Executing the step-8 tutorial program in deal.II to solve the linear elasticity equation, set
up and run patch tests, and visualize the results

Lecture 10: 1 hrs:


• Solving the mixed Laplace problem
• Using the block structured solvers

Tutorial 7: 1 hrs:
• Executing the step-20 tutorial program in deal.II to learn the use of block structured
solvers for the mixed Laplace problem

Friday, May 3, 2024

Lecture 11: 1 hrs:


• Discussing best programming practices
Using linear solvers and preconditioners

Tutorial 8: 1 hrs:
• Experimenting with different linear solvers and preconditioners and examining their
influence on code performance

Lecture 12: 1 hrs:


• Beyond computing: workflows in scientific computing and software development

Tutorial 9: 1 hrs:
• Wrap-up lab activities and final exam

4.0 Who can attend

• Researchers and practicing engineers in civil, mechanical, and aerospace engineering,


engineers and researchers from government organizations including R&D laboratories.
• Students at all levels (B. Tech. (7th/8th semester)/M. Tech./MS/PhD) or Faculty from
reputed academic institutions and technical institutions.

You might also like