Elmer GUI Tutorials

CSC – IT Center for Science

April 6, 2023
ElmerGUI Tutorials

About this document

The Elmer GUI Tutorials is part of the documentation of Elmer finite element software. Elmer GUI Tutorials
gives examples on the use of Elmer in different fields of continuum physics. Also coupled problems are
included.
All these tutorials assume the use of ElmerGUI, the graphical user interface of Elmer. There are also
older tutorials in the Elmer non-GUI Tutorials that may be useful for power users who want to understand
more deeply how models in Elmer are set up.
The present manual corresponds to Elmer software version 9.0.
Latest documentations and program versions of Elmer are available (or links are provided) at http:
//www.csc.fi/elmer.

Copyright information
This document is licensed under the Creative Commons Attribution-NonCommercial 3.0 License. To view
a copy of this license, visit http://creativecommons.org/licenses/by-nc/3.0/.
Initially these tutorials have been written by the Elmer team at CSC - IT Center for Science. However,
Also external contributions to the tutorials are welcome.

Contributors
We gratefully acknowledge Rich Bayless for proofreading, correcting and updating the tutorials to be com-
patible with Elmer version 9.0.

2
Contents

Table of Contents 3
1 Heat equation – 3D – Temperature field of a solid object 7

2 Heat equation – 2D – Temperature field of an L-shaped domain 29

3 Heat Equation –1D – Temperature of an idealized geological intrusion 34

4 Heat Equation – 2D – Axi Symmetric Steady State Radiation 40

5 Heat Equation – 2D – Active and Passive elements 45

6 Linear elasticity equation – 2D – Loaded elastic beam 51

7 Linear elasticity equation – 3D – Loaded elastic beam 55

8 Non-linear elasticity equation – 3D – Loaded elastic curve 59

9 Smitc solver – 2D – Deflection of a linear elastic plate 64

10 Smitc solver – 2D – Eigenmodes of an elastic plate 69

11 Electrostatic equation – Computation of fringe capacitance 74

12 Electrostatic equation – Capacitance of two balls 79

13 Electrostatic equation – Capacitance of perforated plate 85

14 Magnetic field induced by harmonic current in a wire 90

15 Magnetostatics – Magnetic field resulting from a permanent magnet 97

16 Harmonic magnetic field – 2D – Induction heating of a graphite crucible 102

17 Helmholtz – 2D – Acoustic Waves – Air in a Cavity 107

18 VectorHelmholtz – model wave propagation in bent waveguide 112

19 Navier-Stokes equation – Laminar incompressible flow passing a step 118

20 Navier-Stokes equation – 2D – Incompressible – Driven Cavity 122

21 Navier-Stokes equation – Turbulent incompressible flow passing a step 136

22 Navier-Stokes equation – Laminar compressible flow over a step 141

23 Vortex shedding – von Karman instability 148

3
24 Thermal flow in curved pipe 152

25 Interaction between fluid flow and elastic obstacle 158

26 Flow and Heat –2D – Transient – Rayleigh-Benard instability 165

27 Temperature distribution of a toy glacier 170

28 Temperature and velocity distributions of a toy glacier and bedrock 176

29 Generic scalar PDE on 3D angle domain 181

30 Electro-kinetics – 2D – Electro-osmotic flow and advected species 186

31 Asymmetrical Conductor with a Hole

TEAM Workshop Problem 7 193

Index 201
Instructions for the Elmer tutorials
using ElmerGUI

Here are some general instructions for following the Elmer tutorials
• The Elmer tutorials consist of the tutorial documentation and the tutorial folders. For example, these
ElmerGUI tutorials are described in the document ’ElmerTutorials.pdf’ and the working project files
are stored under the folder ’tutorials-GUI-files’.
• If you need a copy of the folder ’tutorials-GUI-files’, it can be downloaded from: https://www.
nic.funet.fi/pub/sci/physics/elmer/doc/

• Each sub-folder under ’tutorials-GUI-files’ contains all the files needed to run a particular tutorial,
such as the ElmerGUI project file, the geometry input file, and the generated mesh files. If you wish
to just run the existing project tutorial and examine the resulting output, start ElmerGUI and load the
existing project.
• If you are more interested in learning how to use ElmerGUI by constructing an ElmerGUI project
from the beginning, the first step will be to create your own sub-folder, and copy in the geometry input
file from the tutorial of interest. Then start ElmerGUI and create a new project, following the step by
step instructions in each tutorial.
• Many examples of geometry input files should be available among the ElmerGUI/samples direc-
tory that should have come with the Elmer installation. Look under a subdirectory named after the
suffix of the sample file, where the suffix indicates the type of geometry format.
• The instructions written in verbatim refer to operations with the GUI. Indentation means step in
the menu hierarchy. The ElmerGUI instructions should not be mixed with the statements in the .sif
command file.
• The menu structure for the default set of equations is located in directory edf, there are a few addi-
tional ones in directory edf-extra. These may be copied to the directory edf permanently, or may
be appended to the menus while running the ElmerGUI.
• The default menu structure may differ from the configuration used when writing the tutorial. Hence
the user is encouraged to check by herself whether the menu structure needed for a particular tutorial
has been loaded into ElmerGUI.

• After having once defined the case in ElmerGUI you may go to the working directory and launch
ElmerSolver from command-line. There you may manually edit the .sif file using a text editor to alter
the parameters.
• Manual alteration to the .sif file will not be communicated to the ElmerGUI project. All edits or
changes will be overwritten by ElmerGUI the next time the project is saved. Copy the manually
altered .sif files to another folder or change the name of the .sif file, to prevent losing your edits.

CSC – IT Center for Science

 6

• It is assumed that the default method for visualization uses .vtu format result files. ElmerPost
that uses the .ep format is no longer available in most installations. Fortunately, most of the tutorials
currently output .vtu format result files, so Paraview (or ElmerVTK) may be used for visualization.
• Since ElmerGUI v. 9.0 the ElmerVTK widget that comes built into ElmerGUI has been upgraded to
use .vtu files. ElmerVTK may be an ideal visualization tool for most straight-forward demonstra-
tions, since it is easy to use. More complicated applications, such as animation of transient studies,
may require the use of Paraview. Regardless of which post processing visualization tool is used, the
construction of an ElmerGUI project remains the same for both.
• These cases have been run a number of times but errors are still possible. Reporting them to elmer-
[email protected], for example, or posting in the Elmer Forum at http://www.elmerfem.org/forum/
is greatly appreciated.

CSC – IT Center for Science

Tutorial 1

Heat equation – 3D – Temperature field

of a solid object

Directory: TemperatureGeneric
Solvers: HeatSolver
Tools: ElmerGUI,netgen,OpenCascade
Dimensions: 3D, Steady-state
Author: Mikko Lyly, Peter Råback

Introduction
As the first tutorial in ElmerTutorials.pdf, the basic usage of ElmerGUI will be demonstrated. New users of
Elmer are invited to use this tutorial to learn how to use ElmerGUI and Elmer. Geometry will be provided,
since creating geometry can be a barrier to using Elmer. Basic steps for geometry creation, for Windows
users and for Linux users, are covered in the document GetStartedElmer.pdf.

Case definition
This tutorial tries to demonstrate how to solve the heat equation for a generic 3D object. The solid object
(see figure 1.1) is heated internally by a heat source.

Figure 1.1: Generic object being heated

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 8

At some part of the boundary the temperature is fixed. Mathematically the problem is described by the
Poisson equation 
−κ∆T = ρf in Ω
(1.1)
T = 0 on Γ
where κ is the heat conductivity, T is the temperature and f is the heat source. It is assumed that density and
heat conductivity are constants.
To determine the problem we assume that part of the boundary is fixed at T0 = 293 K, the internal heat
generation is, h = 0.01 W/kg, and use the material properties of aluminium. ElmerGUI has predefined
values for many constants and material properties defined with SI units, so we will use SI units throughout
this tutorial. The goal of the simulation is to find out the temperature distribution in the object.

Existing Project
The Elmer tutorials consist of the tutorial documentation and the tutorial folders. For example, these Elmer-
GUI tutorials are described in the document ’ElmerTutorials.pdf’ and the working project files are stored
under the folder ’tutorials-GUI-files’, and both can be downloaded from: https://www.nic.funet.
fi/pub/sci/physics/elmer/doc/
Each sub-folder under ’tutorials-GUI-files’ contains all the files needed to run a particular tutorial, such
as the ElmerGUI project file, the geometry input file, and the generated mesh files.
If you wish to just run the existing project tutorial and examine the resulting output, start ElmerGUI and
load the existing project file as follows:

File
Load Project...
Run
Start Solver

You should see the Convergence Monitor and the Solver log windows pop up, and the solution finish in
a few moments in most cases. The results can then be examined using ElmerVTK or Paraview, as follow:

Run
Start ElmerVTK

or
Run
Start Paraview

New Project
If you are more interested in learning how to use ElmerGUI by constructing an ElmerGUI project from
the beginning, then the next few sections will describe in detail the step by step actions needed to build an
ElmerGUI project. The first step will be to create your own project folder, and copy in the geometry input
file from the tutorial of interest. Then start ElmerGUI and create a new project, as follows:

Run
New Project...

Start at the top and select the project directory as shown in Figure 31.3. Then select the Geometry input
file, that was copied into your project folder. This could either be an elmer mesh or a geometry input file,
such as an elmergrid input file. Lastly, if additional Equation Definition Files are needed for the tutorial,
select as many as needed in the right hand box and add them to the left hand box. Finish by clicking on OK.

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 9

Figure 1.2: New Project

ElmerGUI Equation Definition Files

The New Project screen automatically locates and lists all of the available EDFs. The left side of the screen
lists the Default EDFs, which will be loaded into each ElmerGUI project. The right hand box lists all of the
extra EDFs, that are not normally loaded, allowing the option to add individual extra EDFs to a new project.

This tutorial will use the Default EDF, heatequation.xml, for the Heat Solver.

Solution procedure
Start ElmerGUI from command line or by clicking the icon in your desktop. Here we describe the essential
steps in the ElmerGUI by writing out the clicking procedure. Tabulation (indentation) generally means that
the selections are done within the window chosen at the higher level.
The geometry is given in step format in file pump_carter_sup.stp in the samples/step di-
rectory of ElmerGUI. This file is kindly provided at the AIM@SHAPE Shape Repository by INRIA. The
heat equation is ideally suited for the finite element method and the solution may be found even at meshes
that for some other problems would not be feasible. Therefore you may easily experiment solving the same
problem with different meshes. If your particular version of ElmerGUI lacks OpenCascade, you might
try to solve a similar problem with the grd files angle3d.grd, angles3d.grd, bench.grd, or
cooler.grd, for example.
The CAD geometry defined by the step file is transformed on-the-fly by OpenCascade library into an stl
file for which nglib creates tetrahedral volume discretization. You may also use the tetlib library (tetgen) if
you have installed it as a plug-in.
Load the input file:
File
Open -> pump_carter_sup.stp
The meshing will take a minute or two. You should obtain your mesh and may check for the number of
elements in the Model summary, as shown in Figure 1.3.

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 10

Figure 1.3: The computational mesh based on default settings

Visual inspection reveals that the mesh is not quite satisfactory in geometric accuracy. We choose to
modify the mesh by altering the settings in the following way, as shown in Figure 1.4. In order to affect the
mesh density we set the maximum element size, max h, and the minimum element size, min h, and select
the option to restrict mesh size on surfaces by the STL surface density. Note that one must click on Mesh,
Remesh in order for the new settings to be applied.

Figure 1.4: Refinement settings

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 11

View -> Cad model...

Model -> Preferences...
Restrict mesh size on surfaces by STL density = on
Apply
Mesh -> Remesh

The meshing operation takes a minute or two. The modified mesh should be more appealing to the eye.
For example, notice the improvement in the inside diameters of the bores, they are more circular in shape.
You may check for the number of elements in the Model summary, as shown in Figure 1.5.
Note that the remeshing operation will change the Elmer mesh.* files, but will not change the original
geometry input file. The original step CAD file is unchanged, so any time the original geometry input file is
input or reread, the meshing operations must be repeated.

Figure 1.5: Refined mesh

We want to set the temperature at the inside of the three bores, each of which have two interior surfaces.
We could set the boundary conditions by selecting all six surfaces, one at a time, as shown in Figure 1.6.
ElmerGUI has the capability to join boundary surfaces, so that we can unite all six boundaries into a
single boundary, which will make it easier to apply our boundary conditions for the simulation. We demon-
strate this optional step of ’Unify Surface’ next. Note that the Elmer mesh.* files will be updated with the
changed boundary surfaces, and the original geometry input file will be unchanged.

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 12

Figure 1.6: The computational mesh selecting one of the six interior boundaries

In order for ElmerGUI to unite surface boundaries, we must select the six pieces that constitute the
boundaries. Select the first surface by double clicking on the surface, then add the next five surfaces to the
selection by double clicking while holding down the Ctrl-key. Once all six surfaces are highlighted, use
the menus to unify the surfaces.

Mesh
Unify Surface

Figure 1.7: The computational mesh after uniting the six interior boundaries

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 13

After we have the mesh we start to go through the Model menu from the top to bottom. The EDF
menus include default settings for each entry, and we will highlight which particular entries should be set or
changed.
In the Setup we choose things related to the whole simulation such as file names, time stepping, con-
stants etc. The simulation is carried out in 3-dimensional Cartesian coordinates and in steady-state. Only
one steady-state iteration is needed as the case is linear.
For this particular case, we can accept all of the default settings under Setup. While the Setup window
is open, take a look at all of the possible entries, as shown in Figure 1.8. Note the two free text input boxes,
one at the top and the other at the bottom. Use the free text input boxes to add entries to the sif file such as
comments about the model, or define constants, enter MATC commands, etc.

Model
Setup
Simulation Type = Steady state
Steady state max. iter = 1

Choose Apply to close the window.

Figure 1.8: The Setup window

In the equation section we choose the relevant equations and parameters related to their solution. In this
case we’ll have one set only one equation – the heat equation, as shown in Figure 1.9.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use
mouse selection to assign them later. In this case we have just one body and therefore it is easier to assign
the Equation and Material to the body directly, whereas the active boundary is chosen graphically.

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 14

Model
Equation
Add
Heat Equation
Active = on
Apply to bodies = Body 1
Name = Heat Equation
Add
OK

For the linear system solvers we are happy to use the defaults. Click on Edit Solver Settings to
review the possible settings. One may however, try out different preconditioners (ILU1,. . . ), for example.

Figure 1.9: The Equation window

The Material section includes all the material parameters. The tabs near the top for each equation are
divided to generic parameters which are direct properties of the material without making any assumptions
on the physical model, such as the mass. Other properties assume a physical law, such as heat conductivity.
We choose Aluminium from the Material library which automatically sets the needed material properties.

Model
Material
Add
Apply to bodies = Body 1
Material library
Aluminium (generic)
Add
OK

Click on the tabs near the top of the window, such as ’General’ or ’Heat Equation’, to inspect the values
set by selecting the predefined material from the Material Library, as shown in Figure 1.10.

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 15

Figure 1.10: The Material window, showing the General tab

A Body Force represents the right-hand-side of a equation that in this case represents the heat source.
Model
Body Force
Add
Heat Equation
Heat Source = 0.01
Apply to bodies = Body 1
Name = Heating
Add
OK
Click on the tabs near the top of the window, to show the ’Heat Equation’ tab, and inspect the possible
values, as shown in Figure 1.11.

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 16

Figure 1.11: The Body Force window, showing the Heat Equation tab

No initial conditions are required in a steady state case. Initial conditions are usually only needed for
transient studies. Note that the default temperature setting for any model is zero, which doesn’t affect a
steady state study, but will be important for a transient study. The Initial Condition input window is similar
to the Body Force window, as shown in Figure 1.12.

Figure 1.12: The Initial Condition window, showing the Heat Equation tab

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 17

In this case we will set only one boundary condition, which will be a fixed surface temperature, set to
room temperature. You may have noticed that there are many surface boundaries listed for the geometry, as
previously shown in Figure 1.6. For the heat equation, the default boundary condition is zero heat flux, or
fully insulated, and for this example case we don’t need to create boundary conditions for the other surfaces.

First we create the boundary condition, as shown in Figure 1.13. Note that we do not assign the new
boundary condition to any particular boundary, that assignment will happen when we set the boundary
properties.

Model
BoundaryCondition
Add
Heat Equation
Temperature = 293.0
Name = RoomTemp
Add
OK

Figure 1.13: The Boundary condition window

Then we set the boundary properties, as shown in Figure 1.14. Select our desired boundaries and then
assign a boundary condition to those boundaries.

Model
Set boundary properties

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 18

Figure 1.14: The Boundary property window

For our case, we will select the united group of three interior bores by double clicking with the mouse
and then apply the condition for this boundary.
Boundary condition
RoomTemp
For the execution of the simulation, ElmerSolver needs the mesh files and the sif command file. We have
now basically defined all the information for ElmerGUI to write the command file.

Sif
Generate
Edit -> look how your command file came out

After writing the sif file, we may also visually inspect, and optionally edit, the command file, as shown
in Figure 1.15.

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 19

Figure 1.15: The Sif editor window

Before we can execute the solver we should save the files in a directory. In saving the project all the
necessary files for restarting the case will be saved to the destination directory.

File
Save Project
After we have successfully saved the files we may start the solver
Run
Start solver

The solver log and a convergence view automatically pops up showing relative changes of each iteration, as
shown in Figure 1.16.

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 20

Figure 1.16: The Solver log and convergence windows

As the case is linear only one iteration was required for the solution and the second one just is needed
to check the convergence. The norm of the solution should be around 432.4 K (with the default tetgen mesh
389.8 K, respectively).
Note: if you face problems in the solution phase and need to edit the settings, always remember to
regenerate the sif file and save the project before execution.
Once one gets used to generating sif files and running the solver, ElmerGUI includes a double green
arrow icon. The double green arrow icon combines all three steps into a single mouse click:
Generate and save sif, save project, and then run solver.

Postprocessing with ElmerVTK

ElmerVTK is a simple post processor to use, and makes sense for a first look at Elmer simulation results.
We will cover Paraview later, which is a more powerful visualization program and, of course, is also a little
harder to learn and use.

To view the results we will run ElmerVTK for the visualization.

Run
ElmerVTK

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 21

The initial window for ElmerVTK will pop up, with the surface of the geometry colored in blue, as
shown in Figure 1.17. Observe the row of tabs along the top of the window. To hide the surface coloring,
left click on the Surfaces tab.

Figure 1.17: The ElmerVTK window, Surfaces

We want to add 3D contouring, so click on the Isosurfaces tab. Click in the Contour control Variable
box, and select temperature from the drop down box. Continue and click in the Color control Color drop
box and again select temperature, as shown in Figure 1.18.

Figure 1.18: The ElmerVTK window, Isosurfaces selection

Click on Apply and Ok, to apply the settings and close the pop up selection window.

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 22

The isosurfaces will be displayed, as shown in Figure 1.19. Use the mouse to rotate the part, to see the
isosurfaces from different viewpoints.

Figure 1.19: The ElmerVTK window, Isosurfaces shown

Let’s add a color bar to the view, so we can see the range of temperatures in the simulation. Click on the
Colorbar tab. The Colorbar selection menu will pop up, as shown in Figure 1.20. Click in the Color map
drop box, and select Isosurface. Also, click on the radio button for Vertical, to locate the color bar along the
left side.

Figure 1.20: The ElmerVTK window, Colorbar selection

Click on Apply and Ok, to apply the settings and close the pop up selection window.

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 23

The color bar and the isosurfaces will be displayed, as shown in Figure 1.21.

Figure 1.21: The ElmerVTK window, with color bar

To demonstrate how to save a picture with ElmerVTK, click on File, Save picture as, then
select a file name and directory to store the picture, and ElmerVTK will save your simulation results, as
shown in Figure 1.22. If the picture is too small or too large for you, just resize the ElmerVTK window and
save the picture again.

Figure 1.22: The ElmerVTK window, final results

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 24

Next we will demonstrate clipping with ElmerVTK, which may be very helpful for 3D simulations.
Refer back to the Isosurfaces selection window in Figure 1.18, under the section Appearances, click on
Use clip plane, and then Apply. The results are shown in Figure 1.23.

Figure 1.23: The ElmerVTK window, clip plane applied

ElmerGUI supplies a default clip plane setting going in the X direction, through the center of the model.
To adjust the clip plane settings, click on the tab Preferences, and notice at the bottom are six settings.
Adjust the X, Y, or Z coordinates, or adjust the X, Y, or Z normals. Note that one may use a negative normal,
such as -1, to clip the model in the opposite direction.

Figure 1.24: The ElmerVTK window, Preferences and clip plane

One can now close ElmerVTK, or spend some time exploring a few more features of ElmerVTK, such
as Vectors. Also, if one runs a transient study, ElmerVTK can show each of the time steps, using the slider
control at the bottom right of the window.

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 25

Postprocessing with Paraview

Next we will cover the basics of using Paraview to visualize the simulation results.

Run
Paraview

After a short wait, Paraview will open and it should show our case_t0001.vtu file in the upper left window,
as shown in Figure 1.25. First step is to click on the light green Apply button.

Figure 1.25: The Paraview window, click on Apply

Once we have clicked on Apply, the geometry will appear in the layout port, and will be all grey. We will
then tell Paraview which simulation results we wish to display, by clicking on the button under the section
header Coloring located in the lower left corner, as shown in Figure 1.26. Select temperature in the drop
box, as shown in Figure 1.26.

Figure 1.26: The Paraview window, select variable

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 26

The temperature of the simulation will be displayed in the layout port, along with a color bar that is
automatically added to the layout, as shown in Figure 1.27.

Figure 1.27: The Paraview window, showing results

Lastly, let’s save a picture with Paraview. At the top of the window, click on File, then Save
Screenshot, as shown in Figure 1.28.

Figure 1.28: The Paraview window, Screenshot

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 27

Follow the prompts, such as selecting a file name and directory to store the picture, select the size in
pixels of the picture, and Paraview will save your simulation results, as shown in Figure 1.29.

Figure 1.29: The Paraview window, final results

Paraview has many options, as you can see from the many icons in the main window. As one example,
we will demonstrate clipping with Paraview, which may be very helpful for 3D simulations. At the top bar,
click on Filters, Common, Clip, as shown in Figure 1.30. As always with Paraview, click on the
light green Apply button to apply the action to the layout.

Figure 1.30: The Paraview window, Filter, Common, Clip

CSC – IT Center for Science

1. Heat equation – 3D – Temperature field of a solid object 28

One can rotate the model, along with moving the clip plane, or selecting direction with X, Y, or Z
normals. These steps will be left to the student to discover.

Figure 1.31: The Paraview window, clipped results

This is the end of the introduction to Paraview.

CSC – IT Center for Science

Tutorial 2

Heat equation – 2D – Temperature field

of an L-shaped domain

Directory: TemperatureAngleGUI
Solvers: HeatSolve
Tools: ElmerGUI
Dimensions: 2D, Steady-state
Author: Peter Råaback

Problem description
An L-shaped structure (see figure 2.1) is heated by an internal heat source, which magnitude is 1 W/m3 .
The density of the structure is 1 kg/m3 and the heat conductivity is 1 W/mK. All the boundaries Γi are
kept on constant temperature of 0 K. The problem is to solve the temperature distribution in the structure.
Mathematically the problem to be solved is

−κ∆T = ρf inΩ
(2.1)
T = 0 onΓ

where κ is the heat conductivity, T is the temperature and f is the heat source. It is assumed that density and
heat conductivity are constants.

Figure 2.1: L-shaped domain

CSC – IT Center for Science

2. Heat equation – 2D – Temperature field of an L-shaped domain 30

Figure 2.2: The finite element mesh in ElmerGUI

Solution procedure
Start ElmerGUI from command line or by clicking the icon in your desktop. Here we describe the essential
steps in the ElmerGUI by writing out the clicking procedure. Tabulation generally means that the selections
are done within the window chosen at the higher level.
The mesh is given in ElmerGrid format in file angle.grd in the samples directory of ElmerGUI, load
this file.
File
Open -> angle.grd
You should obtain your mesh and may check in the Model summary window that it consists of 341 nodes
and 300 bilinear elements. If the mesh was successfully imported your window should look something in
figure 2.2.
After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we
choose things related to the whole simulation such as file names, time stepping, constants etc. The simulation
is carried out in 2-dimensional Cartesian coordinates and in steady-state. Only one steady-state iteration is
needed as the case is linear.
Model
Setup
Simulation Type = Steady state
Steady state max. iter = 1

Choose Accept to close the window.

In the equation section we choose the relevant equations and parameters related to their solution. In this
case we’ll have one set only one equation – the heat equation.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use mouse
selection to assign them later. In this case we have just one body and one boundary and therefore its easier
to assign the Equation and Material to it directly.

CSC – IT Center for Science

2. Heat equation – 2D – Temperature field of an L-shaped domain 31

For the linear system solvers we are happy to use the defaults. One may however, try out different
preconditioners (ILU1,. . . ) or direct Umfpack solver, for example.
Model
Equation
Add
Name = Heat Equation
Apply to bodies = 1
Heat Equation
Active = on
Apply
OK
The Material section includes all the material parameters. They are divided into generic parameters
which are direct properties of the material without making any assumptions on the physical model, such as
the mass. Other properties assume a physical law, such heat conductivity.
Model
Material
Add
Name = Ideal
Apply to bodies = 1
General
Density = 1.0
Heat Equation
Heat Conductivity = 1.0
Apply
OK
A Body Force represents the right-hand-side of a equation that in this case represents the heat source.

Model
Body Force
Add
Name = Heating
Heat Source = 1.0
Apply to bodies = 1
Apply
OK
No initial conditions are required in steady state case.
In this case we have only one boundary and set it to zero.

Model
BoundaryCondition
Add
Heat Equation
Temperature = 0.0
Name = Zero
Apply to boundaries = 1
Apply
OK
For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.

CSC – IT Center for Science

2. Heat equation – 2D – Temperature field of an L-shaped domain 32

Figure 2.3: The output log of ElmerSolver when used under ElmerGUI

Sif
Generate
Edit -> look how your command file came out

Before we can execute the solver we should save the files in a directory. In saving the project all the
necessary files for restarting the case will be saved to the destination directory.
File
Save Project
After we have successfully saved the files we may start the solver

Run
Start solver
A convergence view automatically pops up showing relative changes of each iteration. As the case is linear
only one iteration was required for the solution and the second one just is needed to check the convergence.
The resulting output log is shown in figure 2.3.
Note: if you face problems in the solution phase and need to edit the setting, always remember to save
the project before execution.
To view the results we use Paraview.
Run
Start Paraview
Picture 29.2 shows the the surface mesh colored with temperature (the pictures were generated by the obso-
lete ElmerPost software).

CSC – IT Center for Science

2. Heat equation – 2D – Temperature field of an L-shaped domain 33

Figure 2.4: The temperature distribution of the L-shaped domain.

Results
As a reference result the maximum temperature in the structure is given. For a comparison the same problem
was solved six times with different element sizes using the -relh flag of ElmerGrid format located in the
Mesh menu under Configure. After choosing Remesh and saving the mesh the solver may be recalled
with the modified mesh.
The maximum temperature obtained by using different meshes is recorded in Table 2.1. From the results
one can see that the result converges. With a denser mesh the result is more accurate, but solving the problem
takes more calculation time. For reference, the CPU-time used in each case is also shown in the table (using
Lenovo 60 laptop).

Table 2.1: Results with different element sizes

Model mesh h [m] Elements max(T ) [K] cpu time [s]
0.2 75 0.1452 0.14
0.1 300 0.1477 0.19
0.05 1200 0.1489 0.41
0.025 5120 0.14925 1.22
0.01 30800 0.14937 7.77

CSC – IT Center for Science

Tutorial 3

Heat Equation –1D – Temperature of an

idealized geological intrusion

Directory: HeatIntrusion1D
Solvers: HeatSolve
Tools: ElmerGUI
Dimensions: 2D, Transient
Author: Peter Råback, Thomas Zwinger

Problem description
This is an extreme simplification of the intrusion process in related to geological processes. The case is
effectively 1D but it is treated as 2D for generality of the approach.
We study temperature distribution from -8 km to -4 km. It is assumed that before the intrusion there is a
linear temperature distribution from 320 to 160 C. At t = 0 an intrusion of 900 C replaces the temperature
from -6.5 km to -5.5 km. The initial temperature distribution may therefore be presented by a piecewise
linear function passing through the points depicted in table 3.1.

Table 3.1: Piecewise linear function defining the initial temperature

depth [m] T (C)

-8000 320.0
-6500 260.0
-6499 900.0
-5501 900.0
-5500 220.0
-4000 160.0

The material parameters for the rock are assumed to be those of ideal granite. We take k = 2.5 W/mK,
Cp = 1250 J/kgK, and ρ = 2800 kg/m3 .
As boundary conditions we use the initial temperature at the upper and lower end. Hence the problem
depends only on the depth direction.

Solution procedure
Start ElmerGUI from command line or by clicking the icon in your desktop. Here we describe the essential
steps in the ElmerGUI by writing out the clicking procedure. Indentation generally means that the selections
are done within the window chosen at the higher level.

CSC – IT Center for Science

3. Heat Equation –1D – Temperature of an idealized geological intrusion 35

Figure 3.1: The geometry of the mesh in ElmerGUI

The mesh is given in ElmerGrid format in file geoslab.grd in the samples directory of ElmerGUI,
load this file.
File
Open -> geoslab.grd
You should obtain your mesh and may check in the Model summary window that it consists of 303 nodes
and 200 quadrilateral surface elements. The mesh is very coarse in the x−direction intentionally since we
are looking just for 1D solution. If the mesh was successfully imported your window should look something
in figure 3.1.
After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we
choose things related to the whole simulation such as file names, time stepping, constants etc. The simulation
is carried out in 2-dimensional Cartesian coordinates and in transient. We will use 1000 time steps each of
size 10 years. For time stepping we will use 2nd order scheme.
Model
Setup
Simulation Type = Steady transient
Timestepping Method = BDF
BDF Order = 2
Timestep Intervals = 10000
Timestep Sizes = $10*365*24*3600
Output Intervals = 2
Choose Apply to accept the values and close the window.
In the equation section we choose the relevant equations and parameters related to their solution. In this
case we’ll have only one equation – the heat equation.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use mouse
selection to assign them later. In this case we have just one body and one boundary and therefore its easier
to assign the Equation and Material to the body directly.
For the linear system solvers we are happy to use the defaults. One may however, try out different
preconditioners (ILU1,. . . ) or direct Umfpack solver, for example.

CSC – IT Center for Science

3. Heat Equation –1D – Temperature of an idealized geological intrusion 36

Model
Equation
Add
Name = Heat Equation
Apply to bodies = 1
Heat Equation
Active = on
Update
OK
The Material section includes all the material parameters. They are divided to generic parameters which
are direct properties of the material without making any assumptions on the physical model, such as the
mass. Other properties assume a physical law, such heat conductivity.
Model
Material
Add
Name = Granite
Apply to bodies = 1
General
Density = 2700.0
Heat Capacity = 1250.0
Heat Equation
Heat Conductivity = 2.5
Update
OK
We will need an initial condition for the case.
Model
Initial Condition
Add
Name = Initial State
Apply to bodies = 1
Heat Equation
Temperature -> click in the box, press enter to open a text box and write the
Apply
OK
Now the expression to be written uses an internal linear interpolation feature of Elmer. The interpolation
expression will be entered into the ‘Variables’ box named ‘Temperature’. Click in the ‘Temperature’ box and
press enter. A pop up window that will accept text input will open. Enter the below text, and click on ‘close’
to accept. Note, do not try to enter the expression into the ‘Free text input’ box right below ‘Temperature’,
it will not work.
Variable coordinate 2
Real
-8000 320.0
-6500 260.0
-6499 500.0
-5501 500.0
-5500 220.0
-4000 160.0
End
In this case we only need boundary conditions for the upper and lower boundary. First we create the
boundary conditions

CSC – IT Center for Science

3. Heat Equation –1D – Temperature of an idealized geological intrusion 37

Model
BoundaryCondition
Add
Heat Equation
Temperature = 320.0
Name = Down
OK
Add
Heat Equation
Temperature = 160.0
Name = Up
OK
Then we set the boundary properties
Model
Set boundary properties
Choose the defined group of three boundaries by clicking with the mouse and apply the condition for the
lower boundary.
Boundary condition
Down
and similarly for the upper boundary.
For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.
Sif
Generate
Edit... -> look how your command file came out
Before we can execute the solver we should save the files in a directory. In saving the project all the
necessary files for restarting the case will be saved to the destination directory.
File
Save Project
After we have successfully saved the files we may start the solver
Run
Start solver
A convergence view automatically pops up showing relative changes of each iteration. As the case is linear
only one iteration was required for the solution and the second one just is needed to check the convergence.
The resulting output log is shown in figure 3.2.
Note: if you face problems in the solution phase and need to edit the setting, always remember to save
the project before execution.
To view the results we use Paraview.
Run
Start ParaView
If your configuration is OK a Paraview window should pop up. Choose temperature for the surface to be
plotted. As we have a time series we may run the sequence by pressing the play icon.
You may also choose the Plot Over Line filter to see the data plotted over a line for better numerical
inspection.
The final maximum temperature of the analysis is around 600 C. Whether this is close to correct could
be studied by increasing further the time and space resolution.

CSC – IT Center for Science

3. Heat Equation –1D – Temperature of an idealized geological intrusion 38

Figure 3.2: The output log of ElmerSolver when used under ElmerGUI

Figure 3.3: The data visualized in Paraview as a surface and line plot

CSC – IT Center for Science

3. Heat Equation –1D – Temperature of an idealized geological intrusion 39

3.0.1 Variation accounting for latent heat release

For intrusion processes the latent heat often plays on important role. For that the internal phase change
model can be used. It requires that all the internal energy is expressed using specific enthalpy rather than
heat capacity and latent heat.
We now eliminate the heat capacity of the original case (1250 J/kgK) and create a specific enthalpy that
includes heat capacity of 1000 J/kgJ and latent heat release of 200 kJ/kg release between interval 700–800
C.
The expression for Specific Enthalpy accounting for these two now yields
Variable Temperature
Real
0.0 0.0
700 7.0e5
800 9.0e5
900 10.0e5
End
For the phase change model we select

Model
Equation
Heat Equation
Phase Change Model = spatial 2
Update
OK

With these changes the maximum temperature at the end of simulation cycle becomes around 624 C.

CSC – IT Center for Science

Tutorial 4

Heat Equation – 2D – Axi Symmetric

Steady State Radiation

Directory: Radiation
Solvers: HeatSolve
Tools: ElmerGUI
Dimensions: 2D, Axi-Symmetric
Author: Juha Ruokolainen, Rich Bayless

Case definition
At high temperature the radiation heat transfer between walls is often the dominating heat transfer mecha-
nism. In this tutorial we examine how radiation heat transfer between concentric cylinders is modelled.

The problem is a pure heat transfer problem that may be solved with HeatSolve. The view and
Gebhart factors associated with the radiation are solved as a first step in solving the equations. Thereafter
the non-linear heat equation is solved until convergence is reached.

Solution procedure
The mesh is given in ElmerGrid format in file radiation.grd, load this file.
File
Open -> radiation.grd
You should obtain your mesh and may check Model Summary... that it consists of 1231 surface ele-
ments. Your geometry and mesh should look something like as shown in figure 31.1

After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we
choose things related to the whole simulation such as file names, time stepping, constants etc.
The simulation is carried out in Axi Symmetric coordinates. The only constant required is the Stefan-
Boltzmann constant that gives the relationship between temperature and radiation power, and this constant
is predefined in the Setup menu.
Model
Setup
Coordinate System = Axi Symmetric
Simulation Type = Steady State
Steady State Max Iterations = 1
Output Intervals = 1
Apply

CSC – IT Center for Science

4. Heat Equation – 2D – Axi Symmetric Steady State Radiation 41

Figure 4.1: Geometry and mesh

In the equation section we choose the relevant equations and parameters related to their solution. In this case
we’ll have the Heat equation.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use
mouse selection to assign them later. In this case we have just two bodies and therefore its easier to assign
the Equation and Material to it directly.
For the linear system solvers we are happy to use the defaults. One may however, try out different
preconditioners (ILU1,. . . ) or direct Umfpack solver, for example.
Model
Equation
Name = Radiation
Apply to Bodies = 1 2
Heat Equation
Active = on
Edit Solver Settings
Steady State
Convergence Tolerance = 1.0e-8
Non-linear system
Convergence Tolerance = 1.0e-8
Max Iterations = 50
Relaxation Factor = 0.7
Newton After Iterations = 1
Newton After Tolerance = 1.0e-4
Linear system
Convergence Tolerance = 1.0e-12
Preconditioning = ILU1
Add

CSC – IT Center for Science

4. Heat Equation – 2D – Axi Symmetric Steady State Radiation 42

OK
The Material section includes all the material parameters. They are divided into generic parameters which
are direct properties of the material without making any assumptions on the physical model, such as the
mass. Other properties assume a physical law, such as conductivities and viscosity.
The material properties differ only in heat conductivity. Heat capacity is not actually needed since the
case is not transient. The inner body has ten times higher conductivity than the outer body.

Model
Material
Name = Inner
Apply to Bodies = 1
General
Density = 1.0
Heat capacity = 1.0
Heat Equation
Heat Conductivity = 10.0
Add
New

Name = Outer
Apply to Bodies = 2
General
Density = 1.0
Heat capacity = 1.0
Heat Equation
Heat Conductivity = 1.0
Add
OK

A Body Force represents the right-hand-side of a equation. The body force is the heating power in units
W/kg.

Model
Body Force
Name = Power
Apply to Bodies = 1
Heat Equation
Volume Heat Source = 10000
Add
OK

Initial conditions are needed in this case. We choose a constant Temperature field of 250C.
Model
Initial Condition
Name = Initial
Apply to Bodies = 1 2
Heat Equation
Temperature = 250.0
Add
OK

Only one boundary condition may be applied to each boundary and therefore all the different physical
BCs for a boundary should be grouped together.

CSC – IT Center for Science

4. Heat Equation – 2D – Axi Symmetric Steady State Radiation 43

The radiation boundary conditions are set for two different boundaries. The first one is for the internal
heated object and the second one for the outer insulating body. The normal direction of the surfaces is impor-
tant since a wrong direction may result to a badly set problem for the view factor computation. Internal and
external surfaces are very different. The normal direction may be switched with the input box Radiation
Target Body. A good sign of a properly set case is that the view factors add up to about one.
The third boundary condition is the Dirichlet condition for the external boundary. Dirichlet conditions
boost up the convergence even though the heat equation is basically well defined also with external radiation
conditions.
Model
BoundaryCondition
Name = Inner
Apply to Bodies = 1
Heat Equation
Radiation Settings = Diffuse Gray
Emissivity = 0.6
Radiation Target Body = -1
Add
New

Name = Outer
Apply to Bodies = 2
Heat Equation
Radiation Settings = Diffuse Gray
Emissivity = 0.1
Radiation Target Body = -1
Add
New

Name = Exterior
Apply to Bodies = 3
Heat Equation
Temperature = 100.0
Add
OK
The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking
with the mouse. Here we use the latter approach as that spares us of the need to know the indexes of each
boundary.
Model
Set boundary properties
Choose Inner -> set boundary condition Inner
Choose Outer -> set boundary condition Outer
Choose Exterior -> set boundary condition Exterior
Update
For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.
Sif
Generate
Edit -> look how your command file came out
Before we can execute the solver we should save the files in a directory. The ElmerGUI project includes
all the files needed to restart the case.

CSC – IT Center for Science

4. Heat Equation – 2D – Axi Symmetric Steady State Radiation 44

Figure 4.2: Temperature distribution in the radiation heat transfer problem

File
Save Project
After we have successfully saved the files we may start the solver
Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration.

When there are some results to view we may start the postprocessor also
Run
Start ParaView

Results
With the given computational mesh the problem is solved in a few seconds. With 1,231 second order 9-noded
rectangular elements the maximum temperature is 565.7 K. The corresponding results are shown in Fig. 4.2.

CSC – IT Center for Science

Tutorial 5

Heat Equation – 2D – Active and Passive

elements

Directory: PassiveElements
Solvers: HeatSolve
Tools: ElmerGUI
Dimensions: 2D,Transient
Author: original author not known

Case definition
This tutorial shows an example of using passive elements in Elmer. This feature allows the activation and
deactivation of parts of the geometry during the simulation. This tutorial uses the heat equation solver to
demonstrate this capability. Use with other solvers is analogous.
The geometry of the problem consists of two parts. The lower edge of the lower part is held at constant
temperature of 1 degrees. The upper body is heated with a constant heating power. Between time steps 5
and 6 the two bodies are connected by two heat conductors, and the heat is conducted from the higher body
to the lower one. The goal of the simulation is to model the temperature distribution in the system over time.
The problem is a pure heat transfer problem that may be solved with HeatSolve.
The geometry for this tutorial looks like as shown in figure 31.1.

Figure 5.1: Geometry

CSC – IT Center for Science

5. Heat Equation – 2D – Active and Passive elements 46

ElmerGUI Equation Menu

We will be using the Heat Solver, which is one of the default, pre-loaded GUI definitions, so it does not need
to be manually activated. For reference, the heatequation.xml definition file is, by default, located in
Linux in:
$ELMER_HOME/share/ElmerGUI/edf
and in Windows, located in:
C:/Program Files/Elmer 9.0-Release/share/ElmerGUI/edf

Solution procedure
The mesh is given in ElmerGrid format in file tmesh.grd, load this file.

File
Open -> tmesh.grd

You should obtain your mesh and may check Model Summary... that it consists of 2018 surface
elements. Your mesh should look like as shown in figure 31.5

Figure 5.2: Mesh

After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we
choose things related to the whole simulation such as file names, time stepping, constants etc.
The simulation is carried out in 2-dimensional Cartesian coordinates.
2nd order bdf time stepping method is selected with 15 steps and with step size of 1 seconds.
The heat equation is solved with an iterative method. The system is linear, thus multiple non-linear
iterations are not needed.

Model
Setup

CSC – IT Center for Science

5. Heat Equation – 2D – Active and Passive elements 47

Simulation Type = Transient

Steady state max. iter = 1
Time Step Intervals = 15
Time Step Sizes = 1
Apply

In the equation section we choose the relevant equations and parameters related to their solution.
In this case we’ll have one equation (named “Heat”) which consists of the heat equation.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use mouse
selection to assign them later. In this case we have just three bodies and therefore its easier to assign the
Equation and Material to it directly.
For the linear system solvers we are happy to use the defaults. One may however, try out different
preconditioners (ILU1,. . . ) or direct Umfpack solver, for example.

Model
Equation
Name = Heat
Apply to Bodies = 1 2 3
Heat Equation
Active = on
Add
OK

The Material section includes all the material parameters. They are divided into generic parameters which
are direct properties of the material without making any assumptions on the physical model, such as the
mass. Other properties assume a physical law, such as conductivities and viscosity.
The material properties of the system are artificial. The following three properties are needed for each
material.

Model
Material
Apply to Bodies = 1 3
General
Density = 1
Heat Capacity = 1
Heat Equation
Heat Conductivity = 1
Add
OK

Material
Apply to Bodies = 2
General
Density = 1
Heat Capacity = 10
Heat Equation
Heat Conductivity = 1
Add
OK

A Body Force represents the right-hand-side of a equation. It is generally not a required field for a body.
There are three bodies with the same equation but different material properties. Body 3 is heated by a
constant body force. Body 2 forms the connecting parts of the system. An initial condition as well as a body
force is defined for this body. The body force contains the initial deactivation, and later activation, of the
connecting part. Note that this part is included in the geometry all the time, but the passive command is used
to define when they are included into the simulation.

CSC – IT Center for Science

5. Heat Equation – 2D – Active and Passive elements 48

Now the passive condition for the connecting part is defined. When the parameter Temperature
Passive has a value larger than zero, the current element is excluded from the solution, otherwise it is
included as a normal element. The parameter may depend on variables, coordinates or time. Here it is
defined to depend on time using a tabular format.
The section ‘Free text input’ is used, because the option to define a passive element has not been added
to the ElmerGUI menu file ‘heatequation.xml’. For any solver, one can always add additional commands by
using the Free text input box.

Model
Body Force
Name = Heat Source
Apply to Bodies = 3
Heat Equation
Volume Sources
Heat source = 10
Add
OK

Body Force
Name = Passive
Apply to Bodies = 2
Heat Equation
Free text input
Temperature Passive = Variable time
Real
0.0 1.0
5.0 1.0
5.2 -1.0
8.0 -1.0
End
Add
OK

Initial conditions should be given to transient cases, and probably are not needed for steady state solu-
tions.
In this case we choose a constant Temperature for the passive element. The initial condition for the
initially passive elements is taken to be 5 degrees; a temperature halfway between the colder part and the
hotter part of the system.

Model
Initial Condition
Name = Initial Guess
Apply to Bodies = 2
Heat Equation
Temperature = 5.0
Add
OK

Only one boundary condition may be applied to each boundary and therefore all the different physical
BCs for a boundary should be grouped together.
The boundary conditions are simple. The lower boundary of the lower body is held at 1 degrees and the
upper boundary of the upper body at 10 degrees. The side walls are assumed to be adiabatic.

Model
Boundary Condition
Name = Bottom

CSC – IT Center for Science

5. Heat Equation – 2D – Active and Passive elements 49

Heat Equation
Temperature = 1.0
Add
New

Boundary Condition
Name = Top
Heat Equation
Temperature = 10.0
Add
OK

The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking on
each boundary with the mouse. Here we use the latter approach as that spares us of the need to know the
indexes of each boundary.

Model
Set boundary properties
Choose Bottom -> set boundary condition Bottom
Choose Top -> set boundary condition Top
OK

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.

Sif
Generate
Edit -> look how your command file came out
Before we can execute the solver we should save the files in a directory. The ElmerGUI project includes
all the files needed to restart the case.

File
Save Project

After we have successfully saved the files we may start the solver.

Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration.

When there are some results to view we may start the postprocessor also.

Run
Start ParaView

Results
Due to the number of the time steps the simulation may take around 15 seconds.

You may inspect the results with Paraview or with ElmerVTK.

In Figure 22.4 the obtained temperature distribution is presented, where the left image shows the passive
condition, and the right image shows the active condition.

CSC – IT Center for Science

5. Heat Equation – 2D – Active and Passive elements 50

Figure 5.3: Temperature distribution at 4 and 6 seconds

CSC – IT Center for Science

Tutorial 6

Linear elasticity equation – 2D – Loaded

elastic beam

Directory: ElasticBeam2D
Solvers: StressSolve
Tools: ElmerGUI
Dimensions: 2D, Steady-state
Author: Peter Råback

Case definition
A homogenous, elastic beam (Ω) is rigidly supported on one end (boundary Γ4 ). On boundary Γ3 the beam
is subjected to a load q(x), which grows linearly from zero to q0 (see figure 6.1). The length of the beam is
1 m and the thickness 0.1 m. Material properties of the beam are those of iron (in the database) Poisson ratio
0.29 and Young’s modulus 193 · 109 N/m2 . Problem is to solve the displacement of the beam.
y
q(x)

Γ3
Ω Γ2
Γ4 x

Γ1

Figure 6.1: Beam and loading.

Problem is solved according to linear elasticity theory. Mathematically the problem to be solved is


 −divσ = 0 in Ω
σ = λtr[ε(u)]I + 2µε(u) in Ω



u = 0 on Γ4 (6.1)
σn = 0 on Γ ∪ Γ

1 2



σn = −q on Γ3


where λ and µ are the Lamé constants (which can be expressed in terms of the Poisson ratio and Young’s
modulus), ε is the linearised strain tensor, u is the displacement vector, q is the given surface traction and n
is the outward unit normal to the boundary.

CSC – IT Center for Science

6. Linear elasticity equation – 2D – Loaded elastic beam 52

Solution procedure
The mesh is given in ElmerGrid format in file beam.grd, load this file.
File
Open -> beam.grd
You should obtain your mesh, Model, Summary... and may check that it consists of 1000 biquadratic
elements.

Figure 6.2: The mesh used in the computations

After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we
choose things related to the whole simulation such as file names, time stepping, constants etc. The simulation
is carried in steady-state in 2-dimensional Cartesian coordinates.
Model
Setup
Simulation Type = Steady state
Steady state max. iter = 1
In the Equation section we choose the relevant equations which in this case only includes the Linear
elasticity equation which solves the problem according to linear elastic theory. In this case we assume
that the beam is thin in the z-direction and hence assume plane stresses. When defining Equations and
Materials it is possible to assign to the bodies immediately, or to use mouse selection to assign them later. In
this case we have just one body and therefore its easier to assign the Equation and Material to it directly. For
the linear system solvers we are happy to use the defaults. One may however, try out different preconditioners
(ILU1,. . . ) or direct Umfpack solver, for example.
Model
Equation
Name = Elasticity
Apply to Bodies = 1
Linear elasticity
Active = on
Plane Stress = True
Add
OK
The Material section includes all the material parameters. They are divided to generic parameters which are
direct properties of the material without making any assumptions on the physical model, such as the mass.
Other properties assume a physical law, such as conductivities and viscosity.
Here we choose the generic iron from the material library. You may click through the material parameters
of the various solvers to ensure that the properties are indeed as they should be. Any consistent set of units
may be used in Elmer. The natural choice is of course to perform the computations in SI units.

Model
Material

CSC – IT Center for Science

6. Linear elasticity equation – 2D – Loaded elastic beam 53

Material library
Iron (generic)
Apply to Bodies = 1
Add
OK
There are no body forces and convergence should be easily obtained with the default initial condition i.e.
zero for all fields.
The first boundary condition fixes the beam rigidly at the wall. The second boundary condition may
seem confusing. Basically it represents the form using a linear interpolation between two points. At x = 0
the force fy = 0 and at x = 1 the force fy = −1.0e7, respectively. The semicolon is an alternative separator
to line break. To fill in the second boundary condition press Enter in the Force 2 checkbox.
Model
BoundaryCondition
Name = Wall
Linear elasticity
Displacement 1 = 0.0
Displacement 2 = 0.0
Add
New

Name = Top
Linear elasticity
Force 2 = Variable Coordinate 1; Real; 0 0; 1 -1.0e7; End
Add
The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking
with the mouse. Here we use the latter approach as that spares us of the need to know the indexes of each
boundary.
Model
Set boundary properties
Choose left-hand-side -> set boundary condition Wall
Choose top -> set boundary condition Top
For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.
Sif
Generate
Edit -> look how your command file came out
Before we can execute the solver we should save the files in a directory. The project includes all the files
needed to restart the case.
File
Save Project
After we have successfully saved the files we may start the solver
Run
Start solver
A convergence view automatically pops up showing relative changes of each iteration. Two iterations are
performed, the second one only to ensure convergence at the nonlinear level. The norm of the finished
computation should be around 0.01808.
When there are some results to view we may start the postprocessor also
Run
Start ParaView

CSC – IT Center for Science

6. Linear elasticity equation – 2D – Loaded elastic beam 54

Results
As a result the absolute value of maximum displacement is given. The displacements calculated with differ-
ent load values q0 are tabulated in table 6.1. Note that the absolute value of the displacement varies linearly
with respect to the load since the model is linear.

Figure 6.3: The displacement of an elastic beam with two different loads using a linear model

Table 6.1: Displacements with different load values

q0 [N/m2 ] max |u| [m]

-1.0e7 0.05755
-1.0e8 0.5755
-1.0e9 5.755

If you look at the results you can see that the displacement values become relatively large. The linear
theory is valid only to small displacements. From Fig 6.3 you can also notice that the beam does not maintain
its original form. This means that the linear elasticity theory can not take into consideration all the necessary
phenomena that are related to the problem, any more. To be able to solve the problem we must use general
elasticity theory.

Extra task: transient loading

If you have time you may try to solve the case using transient loading. You may try with the following
settings.
Model
Setup
Simulation Type = Transient
Time Stepping Method = bdf
BDF Order = 2
Time Step Intervals = 200
Time Step Sizes = 1.0e-4
Having a too large time step will not accurately describe the transient behaviour while having a too small
time step consumes unnecessary resources. There is only numerical damping in the system and hence it will
oscillate for some time.

CSC – IT Center for Science

Tutorial 7

Linear elasticity equation – 3D – Loaded

elastic beam

Directory: ElasticBeam3D
Solvers: StressSolve
Tools: ElmerGUI
Dimensions: 3D, Steady-state
Author: Peter Råback

Case definition
Assume a homogenous, elastic beam being rigidly supported on one end. On the other end it is subjected
with a load of 2000 N resulting from an attached object in the gravitational field. The gravity affects also
the beam itself. The length of the beam is 1 m and the thickness is 0.05 m, and the width 0.1 m. Material
properties of the beam are those of dry pine timber: Poisson ratio 0.37, Young’s modulus 10 · 109 N/m2 ,
and density 550 kg/m3 . The problem is to solve the displacement and stress field of the beam. Here the
StressSolve routine based on the linear theory of elasticity is applied.

Solution procedure
The mesh is given in ElmerGrid format in file beam3d.grd, load this file.
File
Open -> beam3d.grd

You should obtain your mesh and may check that it consists of 6073 nodes and of 1200 quadratic hexahedral
elements. The second order elements give improved accuracy compared to the first order elements as they
avoid the phenomenon known as locking.

Figure 7.1: The mesh used in the computations

CSC – IT Center for Science

7. Linear elasticity equation – 3D – Loaded elastic beam 56

After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we
choose things related to the whole simulation such as file names, time stepping, constants etc. The simulation
is carried in steady-state in 3-dimensional Cartesian coordinates.
Model
Setup
Simulation Type = Steady state
Steady state max. iter = 1
In the Equation section we choose the relevant equations which in this case only includes the Linear
elasticity equation which solves the problem according to linear elastic theory. We also want to com-
pute the stresses as a post-processing step. For the linear system solvers we change the default settings
in order to obtain a better convergence in this case. As the equation is fully linear we also eliminate the
non-linear iteration loop.
Model
Equation
Name = Elasticity
Apply to Bodies = Body 1
Linear elasticity
Active = on
Calculate Stresses = on
Edit Solver Setting
Linear System
Method = Iterative / GCR
Preconditioning = ILU1
Nonlinear system
Max. iterations = 1
Apply
Add
OK
The Material section includes all the material parameters. They are divided into generic parameters
which are direct properties of the material without making any assumptions on the physical model, such as
the mass. Other properties assume a physical law, such as Young’s modulus and Poisson ratio.

Model
Material
Name = Pine
General
Density = 550
Linear Elasticity
Youngs Modulus = 10.0e9
Poisson ratio = 0.37
Apply to Bodies = Body 1
Add
OK

In this case there is a body force i.e. the gravity acting on the beam. We assume that the gravity points
to the negative y direction.
Model
BodyForce
Name = Gravity
Linear Elasticity
Force 2 = $ -9.81 * 550
Apply to Bodies = Body 1

CSC – IT Center for Science

7. Linear elasticity equation – 3D – Loaded elastic beam 57

Add
OK
Here we use a MATC expression for computing the volume force. This expression is constant and is computed
when the command file is interpreted.
Convergence should be obtained with the default initial condition i.e. zero for all fields, hence no initial
condition is applied.
The first boundary condition fixes the beam rigidly at the wall. The second boundary condition distributes
the load of 2000 N uniformly on the area of 5.0e-3 m2 .
Model
BoundaryCondition
Name = Wall
Linear elasticity
Displacement 1 = 0.0
Displacement 2 = 0.0
Displacement 3 = 0.0
Add
New

Name = Mass
Linear elasticity
Force 2 = -4.0e5
Add
The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking
with the mouse. Here we use the latter approach as that spares us of the need to know the indexes of each
boundary.
Model
Set boundary properties
Choose the wall end of the beam -> set boundary condition Wall
Choose the other end of the beam -> set boundary condition Mass
For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.
Sif
Generate
Edit -> look how your command file came out

Before we can execute the solver we should save the files in a directory. The project includes all the files
needed to restart the case.
File
Save Project
After we have successfully saved the files we may start the solver

Run
Start solver
The simulation may take a minute or so depending on the speed of the processor. This time the convergence
monitor does not have a meaningful output since of the different steps only one is related to the actual
solution and the six other ones to the computation of stresses with the Galerkin method.

CSC – IT Center for Science

7. Linear elasticity equation – 3D – Loaded elastic beam 58

Results
When there are some results to view we may start the postprocessor, this time we use Paraview.
Run
Start Paraview
Choose the displacement as the color field to plot. The maximum displacement is 6.36 cm You may also
choose various stress components, or the von Mises stress. Also you may choose to see Surface With
Edges to visualize the mesh also. The resulting picture is shown in Fig 7.2

Figure 7.2: The displaced shape of the elastic beam colored with the von Mises stresses.

Note that the displacements are so large that the assumption of linearity may be severely questioned.
When further increasing the loading one should resort to a solver that is able to catch the geometric non-
linearities – the ElasticSolver.

Extra task: Gravity in x direction

The beam should be more rigid if the beam is oriented differently. For that aim, change the direction of
gravity to orient in the negative x. Change the body force
Model
BodyForce
Linear Elasticity
Force 1 = $ -9.81*550
Update
OK
and the boundary condition
Model
BoundaryCondition
Linear elasticity
Force 1 = -4.0e5
Update
OK
The rigidity should scale as dh3 and hence the maximum displacement should be reduced roughly to one
quarter of the original.

CSC – IT Center for Science

Tutorial 8

Non-linear elasticity equation – 3D –

Loaded elastic curve

Directory: ElasticHookNonlinear
Solvers: ElasticSolve
Tools: ElmerGUI
Dimensions: 3D, Transient
Author: Peter Råback

Case definition
An elastic U-shaped cylinder is pressed from its ends such that the object faces large displacements. The
material properties of stainless steel are used. Problem is to gradually increase the displacement and visualize
the increase of stresses. The problem requires the use of a solver capable of dealing with large displacements.

Solution procedure
The definitions for the relevant equation are not loaded into ElmerGUI by default. Hence, one needs to load
these before starting the simulations.

File
Definitions
Append -> nonlinearelasticity.xml
The additional definitions should reside in the directory edf-extra within the distribution.

The mesh is given in ElmerGrid format in file u_turn.grd, load this file.
File
Open -> u_turn.grd
You should obtain your mesh and may check that it consists of 12288 trilinear elements and 13585 nodes.
After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup
we choose things related to the whole simulation such as file names, time stepping, constants etc. The
simulation is carried as a transient problem in 3-dimensional Cartesian coordinates. The inertial forces are
of no importance so we could as well scan through a set of steady state loading conditions. We choose the
time step such that the total time being simulated is 1 s. We assume that the original mesh (with diameter
0.6 is given units of cm. Hence we need to scale the system down to work in SI units of meters.

Model
Setup

CSC – IT Center for Science

8. Non-linear elasticity equation – 3D – Loaded elastic curve 60

Figure 8.1: The mesh used in the computations as shown in ElmerGUI.

Simulation Type = Transient

Timestep Sizes = 0.05
Timestep Intervals = 20
Coordinate Scaling = 0.01
In the Equation section we choose the relevant equations which in this case only includes the Nonlinear
elasticity equation. We have just one body and therefore its easy to assign the Equation and Material
to it directly. We can use linear system solvers and we are happy to use the defaults. One may however, try
out different preconditioners (ILU1,. . . ) or more efficient iterative methods (BiCGStabl), for example.
Model
Equation
Name = Elasticity
Apply to Bodies = 1
Nonlinear elasticity
Active = on
Edit Solver Settings
Calculate Stresses = on
Calculate Principal = on
Add
OK
Here we choose the stainless steel from the material library. You may click through the material parameters
of the various solvers to ensure that the properties are indeed as they should be. Any consistent set of units
may be used in Elmer. The natural choice is of course to perform the computations in SI units.
Model
Material
Material library
Austenitic stainless steel (AK Steel 201)
Apply to Bodies = 1
Add
OK
There are no body forces and convergence should be easily obtained with the default initial condition i.e.
zero for all fields. Hence we don’t need to toggle these subitems.

CSC – IT Center for Science

8. Non-linear elasticity equation – 3D – Loaded elastic curve 61

We need to type now boundary conditions that we will later assign with the mouse to some surfaces.
We set boundary conditions for both ends of the hook such that they close with respect to each other with
time. The distance travelled in 1 s will be set to 0.006 m i.e. to same as the radius. The other displacement
components are set to zero. To type in the multiline expressions for the boundary condition just press Enter
in the Displacement 1 checkbox. Note that the semicolon is an alternative separator to line break.
Model
BoundaryCondition
Name = MovingRight
Nonlinear elasticity
Displacement 1 = Variable "time"
Real MATC "0.006*tx"
Displacement 2 = 0.0
Displacement 3 = 0.0
Add
New

Name = MovingLeft
Nonlinear elasticity
Displacement 1 = Variable "time"
Real MATC "-0.006*tx"
Displacement 2 = 0.0
Displacement 3 = 0.0
Add
The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking
with the mouse. Here we use the latter approach as that spares us of the need to know the indexes of each
boundary.
Model
Set boundary properties
Choose negative x end -> set boundary condition "MovingRight"
Choose positive x end -> set boundary condition "MovingLeft"
For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.
Sif
Generate
Edit -> look how your command file came out
Before we can execute the solver we should save the files in a directory. The project includes all the files
needed to restart the case.
File
Save Project
After we have successfully saved the files we may start the solver
Run
Start solver
A convergence view automatically pops up showing relative changes of each iteration. As there are 20 time
steps the convergence will be shown for each time step. You can start looking at the results already after a
few time steps. The whole simulation takes a few minutes.
Run
Start ParaView
In reality the material could break at some point. Here we have assumed a linear material law even
though the geometric displacement makes the equation non-linear.

CSC – IT Center for Science

8. Non-linear elasticity equation – 3D – Loaded elastic curve 62

Figure 8.2: The final state of the deformed curve. Color shows the resulting von Mises stresses.

Alternative task: Rotating square profile

You may perform the tutorial with an alternative geometry: square_profile.grd. Follow almost the
same logic except now we rotate the ends of the beam.
We need to enforce rigid body rotations to the ends of the mesh. The following types of Dirichlet
conditions need to be implemented

ux = (cos(φ) − 1)x − sin(φ)y

(8.1)
uy = (cos(φ) − 1)y + sin(φ)x

These can be set using the following MATC function

Displacement 1 = Variable "time, Coordinate"
Real MATC "(cos(tx(0)*pi)-1.0)*tx(1)-sin(tx(0)*pi)*tx(2)
Displacement 2 = Variable "time, Coordinate"
Real MATC "(cos(tx(0)*pi)-1.0)*tx(2)+sin(tx(0)*pi)*tx(1)
Note that the argument to MATC is always called tx and it holds the parameters in the given order. In this
case component "0" refers to time, component "1" to Coordinate 1 (i.e. x), and component "2" to
Coordinate 2 (i.e. y). The multiplier for the trigonometric functions is π which means that a revolution
of 180 degrees is performed in 1 second.

CSC – IT Center for Science

8. Non-linear elasticity equation – 3D – Loaded elastic curve 63

Figure 8.3: A square profile being rotated from one end by 90 and 180 degrees. Color shows the von Mises
stresses.

CSC – IT Center for Science

Tutorial 9

Smitc solver – 2D – Deflection of a linear

elastic plate

Directory: ElasticPlateLinear
Solvers: SmitcSolver
Tools: ElmerGUI
Dimensions: 2D, Eigenmode
Author: Peter Råback

Problem description
This tutorial demonstrates how to use the Smitc solver to solve small deflections of plates. The Smitc solver
is for elastic linear plates and uses the theory of Reissner and Mindlin.
The case under investigation is a L-shaped steel plate under pressure. The plate is shown in figure 9.1
The longer sides have the length of 2 m and the shorter 1 m. So the area of the plate is 3 m2 . The plate has
a thickness of 1 cm. We assume that on the plate there is about 15300 kg of sand. The sand is uniformly
distributed on the plate and the sand stays uniformly distributed even if the plate undergoes small deflection.
The sand exerts to the plate a pressure of 50000 P a. The plate is clamped from all sides meaning that both
deflection and rotation are zero on all edges.

Figure 9.1: The geometry of plate and the numbering of edges.

For more details on the solver we refer to the documentation of Smitc solver in the Elmer Models Manual.

CSC – IT Center for Science

9. Smitc solver – 2D – Deflection of a linear elastic plate 65

Solution procedure
Start ElmerGUI from command line or by clicking the icon in your desktop. Here we describe the essential
steps in the ElmerGUI by writing out the clicking procedure. Indentation generally means that the selections
are done within the window chosen at the higher level.
Before we can start the set-up we should make sure that the menus for Smitc solver are present. If not,
they may be found in file

$ELMERHOME/bin/edf-extra/elasticplate.xml
To load these definitions do the following
File
Definitions
Append -> choose the file
To see what kind of new menu structures were loaded you may play around with viewer collapsing and
opening. Note that if you want to load an existing project you should load the xml-definitions that were used
in creating the project. Therefore it may be best to place all actively used menu definitions in directory
$ELMERHOME/bin/edf

When the menu structures for plate solver are there then we are ready to continue. The first thing to do
is create a mesh with ElmerGrid. The definition of mesh is in the file simple_plate.grd. The mesh is
about uniform and consist of 1000 linear square elements.
File
Open -> simple_plate.grd
You should obtain a L shaped figure with 6 sides. You may check in the Model summary... window
that it consists of 1082 nodes and 1008 surface elements. If the mesh was successfully imported your window
should look something like figure 9.2.

Figure 9.2: The plate mesh

After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup
we choose things related to the whole simulation such as file names, time stepping, constants etc.
The simulation uses 2D Cartesian geometry. The simulation is not time dependent i.e. Steady State.
There are no coupled solvers so only one iteration is needed.

CSC – IT Center for Science

9. Smitc solver – 2D – Deflection of a linear elastic plate 66

Model
Setup
Simulation Type = Steady state
Steady state max. iter = 1
Apply
In the equation section we choose the relevant equations and parameters related to their solution. When
defining Equations and Materials it is possible to assign to the bodies immediately, or to use mouse selection
to assign them later. In this case we have just one body and therefore its easier to assign the Equation and
Material to it directly.
Model
Equation
Add
Name = Elastic Plate
Apply to bodies = 1
Elastic Plates
Active = on
Add
OK
The Material section includes all the material parameters. They are divided into generic parameters
which are direct properties of the material without making any assumptions on the physical model, such as
the mass. Other properties assume a physical law, such heat Young’s modulus. As our problem is academic
in nature we choose some simple ideal parameters, similar to steel, but data from material database could
also be used instead.
Model
Material
Add
Name = Ideal
Apply to bodies = 1
General
Density = 7800.0
Elastic Plates
Youngs Modulus = 209.0e9
Poisson ratio = 0.3
Thickness = 1.0e-2
Tension = 0.0
Add
OK
A Body Force represents the right-hand-side of a equation i.e. external forces. In Body Force block we
give the equations right hand side. It is the uniform pressure from the weight of the sand and it is the same
constant in every point.
Model
BodyForce
Name = Pressure
Apply to Bodies = Body 1
Elastic Plates
Pressure = 5.0e4
Add
OK
In this case all the boundaries are rigidly fixed we set all the components of the solution field to be
zero. The 1st component is the displacement in the normal direction while the 2nd and 3rd components are
actually the components of rotation vector

CSC – IT Center for Science

9. Smitc solver – 2D – Deflection of a linear elastic plate 67

Model
BoundaryCondition
Add
Elastic Plates
Deflection 1 = 0.0
Deflection 2 = 0.0
Deflection 3 = 0.0
Name = Fixed
Apply to boundaries = 1 2 3 4 5 6
Add
OK

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.
Sif
Generate
Edit -> look how your command file came out
Before we can execute the solver we should save the files in a directory. In saving the project all the
necessary files for restarting the case will be saved to the destination directory.
File
Save Project
After we have successfully saved the files we may start the solver
Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration. In this case there is
just one iteration and thus no curve appears. The problem is solved in few seconds.

Results
To view the results start Paraview
Run
Paraview
and select the 1st component of the deflection field (confusingly named the x-component).
Currently you would use Paraview, ElmerVTK, or something that can read the VTU files. Here the
results are shown from Paraview.
Result for displacement (deflection x) is shown in figure 9.3.
Note that the 1st component of Deflection is the displacement to normal direction whereas the 2nd and
3rd components are the x- and y-components of rotation vector, and are shown in figure 9.4.

Extra task
You may test the effect of pre-stressing by altering the Tension material parameter.

CSC – IT Center for Science

9. Smitc solver – 2D – Deflection of a linear elastic plate 68

Figure 9.3: The deflection of the plate.

Figure 9.4: The rotation of the plate in x and y components.

CSC – IT Center for Science

Tutorial 10

Smitc solver – 2D – Eigenmodes of an

elastic plate

Directory: ElasticPlateEigenmodesGUI
Solvers: SmitcSolver
Tools: ElmerGUI
Dimensions: 2D, Eigenmode
Author: Peter Råback

Problem description
For thin elastic structures it is often advisable to use dimensionally reduced models i.e. study plates or shells.
In this tutorial we compute the few lowest eigenmodes of an elastic plate. Our geometry is a simple pentagon
which (compared to a square) eliminates some of the trivial symmetries. The pentagon is rigidly fixed at all
boundaries.
For more details on the solver we refer to the documentation of Smitc solver in the Elmer Models Manual.

Solution procedure
Start ElmerGUI from command line or by clicking the icon in your desktop. Here we describe the essential
steps in the ElmerGUI by writing out the clicking procedure. Tabulation generally means that the selections
are done within the window chosen at the higher level.
Before we can start the set-up we should make sure that the menus for Smitc solver are present. If not,
they may be found in file

$ELMERHOME/bin/edf-extra/elasticplate.xml
To load these definitions do the following
File
Definitions
Append -> choose the file
To see what kind of new menu structures were loaded you may play around with viewer collapsing and
opening. Note that if you want to load an existing project you should load the xml-definitions that were used
in creating the project. Therefore it may be best to place all actively used menu definitions in directory
$ELMERHOME/bin/edf

When the menu structures for plate solver are there then we are ready to continue. The mesh is given in
2d netgen format in file pentagon.grd in the samples directory of ElmerGUI, load this file.

CSC – IT Center for Science

10. Smitc solver – 2D – Eigenmodes of an elastic plate 70

Figure 10.1: The finite element mesh in ElmerGUI

File
Open -> pentagon.in2d

You should obtain a pentagon consisting of 5 triangles. To increase the number of elements change the
parameters passed on to the nglib library by going to
Mesh
Configure
nglib / Max H: 0.05

You may check in the Model summary window that it consists of 1199 nodes and 2276 linear triangles.
If the mesh was successfully imported your window should look something in figure 10.1.
After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we
choose things related to the whole simulation such as file names, time stepping, constants etc. The simulation
is carried out in 2-dimensional Cartesian coordinates and in steady-state (also used for eigenmodes). Only
one steady-state iteration is needed as the case is linear.

Model
Setup
Simulation Type = Steady state
Steady state max. iter = 1
Apply

In the equation section we choose the relevant equations and parameters related to their solution. When
defining Equations and Materials it is possible to assign to the bodies immediately, or to use mouse selection
to assign them later. In this case we have just one body and therefore its easier to assign the Equation and
Material to it directly.

CSC – IT Center for Science

10. Smitc solver – 2D – Eigenmodes of an elastic plate 71

For the solver setting we need to activate the eigen mode computation. We also choose the direct umfpack
solver which for small 2D problems often performs great.
Model
Equation
Add
Name = Plate Equation
Apply to bodies = 1
Elastic Plates
Active = on
Edit Solver Settings
Solver Specific Options
Eigen Analysis = on
Eigen System Values = 10
Linear System
Direct = on
Umfpack
Add
OK
The Material section includes all the material parameters. They are divided into generic parameters
which are direct properties of the material without making any assumptions on the physical model, such as
the mass. Other properties assume a physical law, such heat Young’s modulus. As our problem is academic
in nature we choose some simple ideal parameters but data from material database could also be used instead.
Model
Material
Add
Name = Ideal
Apply to bodies = 1
General
Density = 1000.0
Elastic Plates
Youngs Modulus = 1e9
Poisson ratio = 0.3
Thickness = 0.001
Tension = 0.0
Add
OK
A Body Force represents the right-hand-side of a equation i.e. external forces. In eigenmode analysis no
body forces are used. Nor are any Initial conditions required.
In this case all the boundaries are rigidly fixed we set all the components of the solution field to be zero.
The 1st component is the displacement in the normal direction while the 2nd and 3rd components are its
derivatives in x and y directions.
Model
BoundaryCondition
Add
Elastic Plates
Deflection 1 = 0.0
Deflection 2 = 0.0
Deflection 3 = 0.0
Name = Fixed
Apply to boundaries = 1 2 3 4 5
Add
OK

CSC – IT Center for Science

10. Smitc solver – 2D – Eigenmodes of an elastic plate 72

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.
Sif
Generate
Edit -> look how your command file came out
Before we can execute the solver we should save the files in a directory. In saving the project all the
necessary files for restarting the case will be saved to the destination directory.
File
Save Project
After we have successfully saved the files we may start the solver
Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration. In this case there is
just one iteration and thus no curve appears.

Results
The resulting eigenvalues are shown in table 10.1. Note that some eigenmodes are degenerated but as the
finite element mesh is not perfectly symmetric there will be minor differences in the eigenvalues.

Table 10.1: Ten lowest eigenvalues for the pentagon plate

No ω2
1 18.9
2,3 81.3
4,5 214.5
6 281.1
7, 8 472.5
9, 10 621.0

Note: if you face problems in the solution phase and need to edit the setting, always remember to save
the project before execution.
To view the results we here start Paraview
Run
Paraview

and select the 1st component of the deflection field (confusingly named the x-component). If one chose
for vtu output the mode Eigen Analysis = True then each file includes one eigenmode. Then the
eigenmodes may be treated as time steps. In figure 10.2 some of the lowest eigenmodes are depicted.

Extra task
You may test the effect of pre-stressing by altering the Tension material parameter.
There are other similar geometries that you could use i.e. hexagon.in2d, heptagon.in2d, octagon.in2d.
When the number of vertices is increased the eigenvalues should slightly decrease.

CSC – IT Center for Science

10. Smitc solver – 2D – Eigenmodes of an elastic plate 73

Figure 10.2: The 1st, 2nd, 4th, 6th, 7th and 9th eigenmode of the plate. The displacement in the normal
direction (first component) is shown.

CSC – IT Center for Science

Tutorial 11

Electrostatic equation – Computation of

fringe capacitance

Directory: FringeCapacitance
Solvers: StatElecSolver
Tools: ElmerGUI
Dimensions: 2D, Steady-state
Author: Peter Råback

Case definition
This case presents solving the Laplace equation for electric potential.

− ∇ · ε∇φ = 0 ∈ Ω (11.1)

where the electric potential φ is given at conducting surfaces. This is a standard type of equation with many
variants in physics, electrostatics being just one of them. From the solution one may calculate derived fields,
and capacitance which is obtained from the total electric energy
Z
1
E= ε|∇φ|2 dΩ (11.2)
2

and the relation E = CU 2 /2.

Figure 11.1: The geometry of the capacitor

The geometry studied is a 2D plate capacitor having the well known approximation for Capacitance
Capp = εr ε0 A/d. With the help of the simulation one may evaluate the fringe capacitance resulting from
the end effects of the capacitor geometry. The measurements of the capacitor are 10 × 1, and the distance

CSC – IT Center for Science

11. Electrostatic equation – Computation of fringe capacitance 75

to ground is also 1. Defining the permittivity of vacuum to be 0 = 1 the comparison to the analytical
approximation becomes trivial since then then Capp = 10.
The derived fields in the StatElecSolver are computed by averaging the fields over elements – not using
the Galerkin method which would provide optimal accuracy. The fields may, however, be sufficient for
visualization purposes.

Figure 11.2: Computational mesh used in the simulation

Solution procedure
The definitions for the electrostatic equation are not loaded into ElmerGUI by default. Hence, one needs to
load these before starting the simulations.
File
Definitions
Append -> electrostatics.xml
The additional definitions should reside in the directory edf-extra within the distribution. Moving the
desired xml files to the edf-directory enables automatic loading of the definitions at start-up. By inspecting
the definitions in the Elmer Definitions File editor one may inspect that the new definitions
were really appended.
The mesh is given in ElmerGrid format in file disc.grd, load this file.
File
Open -> disc.grd
You should obtain your mesh and may check that it consists of 30484 nodes and of 30348 bilinear elements.

Model
Summary...
After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we
choose things related to the whole simulation such as file names, time stepping, constants etc. The steady-
state simulation is carried out in 2-dimensional Cartesian coordinates. For convenience we also set ε equal
to one.
Model
Setup
Simulation Type = Steady state
Vacuum Permittivity = 1.0

CSC – IT Center for Science

11. Electrostatic equation – Computation of fringe capacitance 76

In the equation section we choose the relevant equations and parameters related to their solution. In this case
we’ll have only the electrostatics solver.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use
mouse selection to assign them later. In this case we have just one body and therefore its easier to assign the
Equation and Material to it directly.
In the solver specific options we want to activate some flags that are needed to invoke the computation
of derived fields. For the linear system solvers we are happy to use the defaults. One may however, try out
different preconditioners (ILU1,. . . ) or direct Umfpack solver, for example.
Model
Equation
Name = Electrostatics
Apply to Bodies = 1
Electrostatics
Active = on
Edit Solver Settings
Solver specific options
Calculate Electric Field = True
Calculate Electric Energy = True
Add
OK
The Material section includes all the material parameters. In this case we only have the relative permittivity
which we set to one.
Model
Material
Name = Ideal
Electrostatics
Relative Permittivity = 1.0
Apply to Bodies = 1
Add
OK
We have two boundary conditions for the potential at the ground and at the capacitor. For other bound-
aries the do-nothing boundary results to zero flux over the boundary.
Model
BoundaryCondition
Name = Ground
Electrostatics
Potential = 0.0
Add
New

Name = Capacitor
Electrostatics
Potential = 1.0
Add
The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking
with the mouse. Here we use the latter approach as that spares us of the need to know the indexes of each
boundary.
Model
Set boundary properties
Choose Ground -> set boundary condition Ground
Choose Capacitor -> set boundary condition Capacitor

CSC – IT Center for Science

11. Electrostatic equation – Computation of fringe capacitance 77

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.
Sif
Generate
Edit -> look how your command file came out
Before we can execute the solver we should save the files in a directory. The project includes all the files
needed to restart the case.
File
Save Project
After we have successfully saved the files we may start the solver
Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration. The equation is
fully linear and hence only two iterations are needed – the second one just ensures that convergence of the
non-linear level was really obtained. The norm of the solution should be 0.3055.
When the solution has finished we may start the postprocessor to view some results.
Run
Start ParaView

Results
From the output of the simulation one may see that the capacitance in this case was 13.697 compared to the
analytical estimate of 10. Hence the fringe capacitance in this case increases the capacitance by 37 %.

Figure 11.3: The electrostatic potential

CSC – IT Center for Science

11. Electrostatic equation – Computation of fringe capacitance 78

Figure 11.4: The electrostatic energy density, a close-up

CSC – IT Center for Science

Tutorial 12

Electrostatic equation – Capacitance of

two balls

Directory: CapacitanceOfTwoBalls
Solvers: StatElecSolver
Tools: netgen,ElmerGUI
Dimensions: 3D, Steady-state
Author: Peter Råback

Case definition
This case presents the solution of the capacitance of perfectly conducting balls in free space. A voltage
difference between the balls results to electric charge being introduced to the system. The balls have also
self-capacitance that comes from the voltage difference with the far field. Therefore a symmetric capacitance
matrix with of size 2 × 2 needs to be solved. The capacitances may be computed from two different voltage
configurations. For both the electrostatic equation is solved automatically.
The problem does not have an analytical solution in a closed form. However, the cross-capacitance
between the balls may be approximated from the series solution in 1 :
a2 a2 a4
 
C12 = 4πε 1+ 2 + 4 + ... (12.1)
d d − 2a2 d − 4d2 a2 + 3a4
and the self-capacitance from
a2 a3
 
a
C10 = C20 = 4πεa 1 − + 2 − + . . . (12.2)
d d − a2 d3 − 2da2

Let’s mark C̃ = C/ε. In this case C̃12 ≈ 1.191 and C̃10 ≈ 5.019. Unfortunately the error bounds are not
given.
In this particular case the balls are assumed to have a radius of a = 0.5 and they are placed at distance
d = 2 apart from each other (measured from the ball origins).

Meshing with Netgen (Optional)

If you have Netgen installed and running, then you can generate the mesh before starting ElmerGUI, follow-
ing the below instructions.
If you don’t have Netgen available, don’t worry, an Elmer mesh has been provided in the ‘Capacitance-
OfTwoBalls’ subfolder under ‘tutorials-GUI-files’. Just load the ElmerGUI project in that subfolder, and it
will load the mesh for you. After loading the mesh, skip ahead to the next section, ‘Solution Procedure’.
1 TUW-101191, PHD Dissertation by Christoph Wasshuber, About Single-Electron Devices and Circuits, Institut für Mikroelek-

tronik, 1997, Appendix A.3, Capacitance of Two Spheres, https://www.iue.tuwien.ac.at/phd/wasshuber/

CSC – IT Center for Science

12. Electrostatic equation – Capacitance of two balls 80

If you have Netgen installed, then meshing is performed with the graphical user interface of Netgen.
Netgen creates tetrahedral quality meshes and provides a native output for Elmer. At the time of writing
this tutorial the quadratic elements had some problems with numbering but these should not affect the linear
elements.
The file is given as netgen geometry format in file TwoBallsInBall.geo. The geometry definition
includes the two smaller balls inside a bigger ball. Ultimately the bigger ball would be infinitely large.
As this is impossible here we choose a modest radius of 5. The larger this value, the better the far-field
approximation of the electrostatic solution is.
The content of the file is given below:
#
# a large ball with two smaller balls cut off
#
algebraic3d
solid smallballs = sphere (-1.0, 0.0, 0.0; 0.5)
or sphere (1.0, 0.0, 0.0; 0.5);
solid bigball = sphere (0.0, 0.0, 0.0; 5.0);
solid rest = bigball and not smallballs;
tlo rest -col=[0,0,1] -transparent;
Open the file and apply the default meshing. In this example two consecutive uniform refinements were
performed (choose Refine Uniform under Refinement) so that the final mesh consisted of 41 693
nodes and 238 976 linear tetrahedrons.
To save the mesh first choose under File the Export Filetype to be Elmer. Then choose
Export Mesh and save the mesh into a suitable directory to be opened by ElmerGUI.

Figure 12.1: Surface mesh for the two inner balls as seen in Netgen

The order of the mesh using nodal elements may be increased by ElmerGrid. Assuming the mesh
would reside in directory meshlin a mesh consisting of quadratic elements may be performed with the
following command:

CSC – IT Center for Science

12. Electrostatic equation – Capacitance of two balls 81

ElmerGrid 2 2 meshlin -increase -out meshquad

This will maintain the number of elements but the number of nodes will, in this case, increase to 359 009.

Solution procedure
The definitions for the electrostatic equation may not have been loaded into ElmerGUI by default. If this is
the case one needs to load these before starting the simulations.
File
Definitions
Append -> electrostatics.xml
The additional definitions should reside in the directory edf-extra within the distribution. Moving the
desired xml files to the edf-directory enables automatic loading of the definitions at start-up. By inspecting
the definitions in the Elmer Definitions File editor one may inspect that the new definitions
were really appended.
The mesh is already created, load it from the directory that was created above.
File
Load Mesh -> mesh

Figure 12.2: The mesh with one highlighted ball as seen in ElmerGUI

The display in ElmerGUI upon loading the mesh will show the outside of the large ball. To view the
inner two smaller balls, double click on the outside of the large ball, then click on ‘Hide/show selected’.

View
Hide/show selected Ctrl+H

CSC – IT Center for Science

12. Electrostatic equation – Capacitance of two balls 82

After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup
we choose things related to the whole simulation such as file names, time stepping, constants etc. The
steady-state simulation is carried out in 3-dimensional Cartesian coordinates. For convenience we also set
the permittivity of vacuum ε0 equal to one. This makes it easier to compare the results to the analytical
expressions.

Model
Setup
Simulation Type = Steady state
Vacuum Permittivity = 1.0
In the equation section we choose the relevant equations and parameters related to their solution. In this case
we’ll have only the electrostatics solver.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use
mouse selection to assign them later. In this case we have just one body and therefore its easier to assign the
Equation and Material to it directly.
In the solver specific options we want to activate some flags that are needed to invoke the computation
of derived fields. For the linear system solvers we are happy to use the defaults. One may however, try out
different preconditioners (ILU1,. . . ) or direct Umfpack solver, for example.
Model
Equation
Name = Electrostatics
Apply to Bodies = 1
Electrostatics
Active = on
Edit Solver Settings
Solver specific options
Calculate Capacitance Matrix = True
Calculate Electric Field = True
Calculate Electric Energy = True
Add
OK
The Material section includes all the material parameters. In this case we only have the relative permittivity
εr which we set to one.
Model
Material
Name = Ideal
Electrostatics
Relative Permittivity = 1.0
Apply to Bodies = 1 2
Add
OK
We have two boundary conditions for the potential at the ground and at the capacitor. For other bound-
aries the do-nothing boundary results to zero flux over the boundary.

Model
BoundaryCondition
Name = Farfield
Electrostatics
Electric Infinity BC = True
Add
New

CSC – IT Center for Science

12. Electrostatic equation – Capacitance of two balls 83

Name = CapBody1
Electrostatics
Capacitance Body = 1
Add
New

Name = CapBody2
Electrostatics
Capacitance Body = 2
Add

The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking
with the mouse. Here we use the latter approach as that spares us of the need to know the indexes of each
boundary.
Model
Set boundary properties
Choose Outer sphere -> set boundary condition Farfield
Choose one inner sphere -> set boundary condition CapBody1
Choose the other inner sphere -> set boundary condition CapBody2
For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.

Sif
Generate
Edit -> look how your command file came out
Before we can execute the solver we should save the files in a directory. The project includes all the files
needed to restart the case.
File
Save Project
After we have successfully saved the files we may start the solver

Run
Start solver
A convergence view automatically pops up showing relative changes of each iteration. The equation is
fully linear and hence only two iterations are needed – the second one just ensures that convergence of the
non-linear level was really obtained. The norm of the solution should be 0.36356324.
When the solution has finished we may start the postprocessor to view some results.
Run
Start ParaView

Results
The essential result of this case are the values of the capacitance matrix. In this case C̃12 ≈ 1.691 and
C̃10 ≈ 5.019. For linear elements the obtained figures are 1.6983, 5.0793 and 5.0812, for quadratic Lagrange
elements 1.6641, 5.0340 and 5.0340, respectively, and finally for quadratic p-elements 1.6856, 4.9863 and
4.9884.
The values are rather satisfactory with a difference less than 2% from the series approximation.
Note that the derived fields in the StatElecSolver are computed by averaging the fields over elements
– not using the Galerkin method which would provide optimal accuracy. To get optimal accuracy, use
FluxSolver, for example.

CSC – IT Center for Science

12. Electrostatic equation – Capacitance of two balls 84

Figure 12.3: The electrostatic potential on the clipping plane. This is the latter of the two symmetric config-
urations where the unit voltage is applied to one ball and zero voltage to the other, respectively.

CSC – IT Center for Science

Tutorial 13

Electrostatic equation – Capacitance of

perforated plate

Directory: CapacitanceOfPerforatedPlate
Solvers: StatElecSolver
Tools: ElmerGUI
Dimensions: 3D, Steady-state
Author: Peter Råback

Case definition
This case presents solving the Poisson equation for electric potential and calculating appropriate derived
quantities, such as capacitance, based on the result. The geometry under study is a perforated plate.
The shape of the holes is assumed to be square with size 3 × 3 mm2 . The holes cover the other plate
uniformly so that the size of each unit cell is 10 × 10 mm2 . The thickness of the plate is assumed to be 1.5
mm and the distance from the reference plate 1.0 mm. The plate may be assumed to be infinitely large. Due
to symmetry considerations it suffices to study only one quarter of a unit cell.
The results may be compared to the ideal plate capacitor without holes. For a plane capacitor, the
capacitance is
A
C = εr ε0 , (13.1)
d
where εr is the relative permittivity, ε0 is the permittivity of vacuum, A is the area of a capacitor plate, d is
the separation of the capacitor plates, and Φ is the potential difference between the plates. For the case under
study we get an approximation C = 221.36 f F .

Preliminaries
The definitions for the electrostatic equation might not have been loaded into ElmerGUI by default. Check
the Model/Equation menu to verify the situation. If electrostatics is not present one needs to load
definitions for it before starting the simulations.
File
Definitions
Append -> electrostatics.xml

The additional definitions should reside in the directory edf-extra within the distribution. Moving the
desired xml files to the edf-directory enables automatic loading of the definitions at start-up. By inspecting
the definitions in the Elmer Definitions File editor one may inspect that the new definitions
were really appended.

CSC – IT Center for Science

13. Electrostatic equation – Capacitance of perforated plate 86

Meshing
In this case meshing is performed using ElmerGrid format in file hexhole.grd and the ElmerGrid plugin
within ElmerGUI. The default mesh is OK and therefore no modifications are needed by the user.
Elmer does not operate an any particular units but usually SI-units are preferred. We therefore choose to
scale the problem with 0.001 so that these measurements will be in mm.
Load the mesh file.

File
Open -> hexhole.grd
You should obtain your mesh and may check that it consists of roughly of 33 159 nodes and of 29 610 linear
hexahedral elements.

Figure 13.1: The mesh with one highlighted backplate as seen in ElmerGUI

Case setup
After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we
choose things related to the whole simulation such as file names, time stepping, constants etc. The steady-
state simulation is carried out in 3-dimensional Cartesian coordinates. We want to work in mm so we need
to scale the mesh with a factor 0.001.

Model
Setup
Simulation Type = Steady state
Coordinate Scaling = 0.001
In the equation section we choose the relevant equations and parameters related to their solution. In this case
we’ll have only the electrostatics solver.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use
mouse selection to assign them later. In this case we have just one body and therefore its easier to assign the
Equation and Material to it directly.

CSC – IT Center for Science

13. Electrostatic equation – Capacitance of perforated plate 87

In the solver specific options we want to activate some flags that are needed to invoke the computation
of derived fields. For the linear system solvers we are happy to use the defaults. One may however, try out
different preconditioners (ILU1,. . . ) or direct Umfpack solver, for example.
Model
Equation
Name: Electrostatics
Apply to Bodies: 1
Electrostatics
Active: on
Edit Solver Settings
Solver specific options
Calculate Electric Field: True
Calculate Electric Energy: True
Add
OK
The Material section includes all the material parameters. In this case we only have the relative permit-
tivity εr which we could set to one. Alternatively, we can obtain it automatically by choosing the properties
of air.
Model
Material
Add
Material library
Air (room temperature)
Apply to bodies: Body 1
Add
OK
We have two boundary conditions for the potential at the ground and at the capacitor. For other bound-
aries the do-nothing boundary results to zero flux over the boundary.
Model
BoundaryCondition
Name: Ground
Electrostatics
Potential: 0.0
Add
New

Name: Capacitor
Electrostatics
Potential: 1.0
Add
The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking
with the mouse. Here we use the latter approach as that spares us of the need to know the indexes of each
boundary.
Model
Set boundary properties
Choose the backplate -> set boundary condition Ground
Choose the four pieces of the perforated plate -> set boundary condition Capacitor
For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.

CSC – IT Center for Science

13. Electrostatic equation – Capacitance of perforated plate 88

Sif
Generate
Edit -> look how your command file came out
Before we can execute the solver we should save the files in a directory. The project includes all the files
needed to restart the case.

File
Save Project
After we have successfully saved the files we may start the solver

Run
Start solver
A convergence view automatically pops up showing relative changes of each iteration. The equation is
fully linear and hence only two iterations are needed – the second one just ensures that convergence of the
non-linear level was really obtained. The norm of the solution should be roughly 0.8846.

Numerical results
The numerical results obtained for capacitance and electric force are compared to those of a complete plane
capacitor.
The results of the simulation as well as the comparison to the complete plane capacitor values are shown
in Table 13.1.

Table 13.1: Comparison of numerical results to analytic values

relh C(fF) ratio

1.0 216.37 0.977
0.7 216.34
0.5 216.32

Visualization
When the solution has finished we may start the postprocessor to view some results. Here we assume that
we chose the .vtu format for Paraview as output. The opening of Paraview can be done automatically from
the ElmerGUI menu, or alternatively manually from Paraview.
File
Open: case0001.vtu
Apply
You may now choose the fields to be depicted etc. For more information on the use of of Paraview look at
the documentation of the software. In figure 13.2 some examples of visualizations with Paraview are given.

CSC – IT Center for Science

13. Electrostatic equation – Capacitance of perforated plate 89

Figure 13.2: The electrostatic potential and the electric energy density of the quarter of a unit cell of a
perforated plate capacitor visualized by Paraview.

CSC – IT Center for Science

Tutorial 14

Magnetic field induced by harmonic

current in a wire

Directory: MagneticFieldWire
Solvers: WhitneyAVHarmonicSolver
Tools: ElmerGrid, ElmerGUI
Dimensions: 3D, Harmonic
Author: Peter Råback

Case definition
This case demonstrates the simplest case on how to utilize the edge element based solvers for computation
of magnetic and electric fields.
Consider a simple copper wire with radius R = 0.001 m and length R = 0.01 m. A potential difference
of 1 V/m is applied over the wire. We want to know the magnetic field strength around the wire.
For steady state we have a simple analytical solution for reference.
 Iµ0
~ − 2π ln(r/R)~ez , r>R
A= Iµ0 2 2 (14.1)
− 4πR 2 (r − R )~
e z , r ≤R

resulting to a magnetic flux density

Iµ0

~ =
B 2πr ~
eφ , r>R
(14.2)
Iµ0 r
2πR R ~eφ , r≤R

We can solve the current from Ohms law to obtain the maximum field value at r = R to be

~ = 1 ~
|B| µ0 Rσ|E| (14.3)
2
Using electric conductivity of copper (σ=59.59e6 A/m2 V) we obtain 0.0374 T.
We are interested in what happens when the voltage is applied sinusoidally with a frequency of 100 kHz.
The harmonic case is unfortunately much more difficult to solve and no simple analytical expression exists.
Therefore we need to solve the problem numerically.

Solution procedure
This tutorial will need the additional EDF, magnetodynamics.xml, for the WhitneyAVHarmonic Solver.
The definitions for the relevant equation are not loaded into ElmerGUI by default. Hence, one needs to load
these before starting the simulations.

CSC – IT Center for Science

14. Magnetic field induced by harmonic current in a wire 91

File
Definitions
Append -> magnetodynamics.xml
The additional definitions should reside in the directory edf-extra within the distribution. Moving the
desired xml files to the edf-directory enables automatic loading of the definitions at start-up. By inspecting
the definitions in the Elmer Definitions File editor one may inspect that the new definitions
were really appended.
The mesh is already defined in ElmerGrid format as file wire.grd. Load it from the tutorial directory
where it resides.
File
Open -> wire.grd
The ElmerGrid plug-in of ElmerGUI will read the mesh and create the mesh for ElmerGUI using the Elmer-
Grid plug-in. Note that the user could also create the mesh directly on the command line by
ElmerGrid 1 2 wire.grd

and then use the Load Mesh option in ElmerGUI to take the mesh into use. If the user wants to modify
the default mesh that can be done by editing the file directly. The mesh has been constructed so that it can
capture some boundary layer phenomena that we expect to take place on the surface of the wire. In principle
the mesh could be even shorter since the results do not really depend on the axial direction.

Figure 14.1: The mesh for the wire and surrounding air as seen in ElmerGUI

After we have the mesh we start to go through the Model menu from the top to bottom.
In the Setup we choose things related to the whole simulation such as file names, time stepping, constants
etc. The steady-state simulation is carried out in 3-dimensional Cartesian coordinates. The initial size of the
mesh is in millimeters, which is 1000 times too large. Hence we apply scaling in this menu, Coordinate

CSC – IT Center for Science

14. Magnetic field induced by harmonic current in a wire 92

Scaling = 1.0e-3. Enter a single scaling number, and the scaling will be applied equally to all coordinates,
whether in 2D or 3D.
Also we want to use a harmonic equation that requires the frequency to be set somewhere, in this case
we will define the frequency in the Simulation section.
SetUp
Coordinate Scaling
1.0e-3
Angular Frequency
1.0e5
The permeability of vacuum is predefined in SI Units in the Constants section. To set other than the default
value for it (in case using a different unit system), change the value in the Constants section.
In the equation section we choose the relevant equations and parameters related to their solution. In
this case we’ll have the MgHarm solver (there exists also a steady-state/transient version of the solver as
MgDyn), as well as the postprocessing solver MgDynPost.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use mouse
selection to assign them later. In this case we know that body 1 is the wire and body 2 is the surrounding
air and we can apply the definitions directly. They will have the same set of solvers but different material
properties.
The solver specific options may need some alternations. We choose a more suitable linear solver strategy
for the AV equation.
For the postprocessing we add some fields to be computed and skip the computation of nodal fields
since the elemental fields are often preferable for discontinuous fields. We also want to ensure that the
postprocessing solver is run just before saving the results, after the primary solver has been computed.
In the following the correct equation are chosen and the suitable solver specific options are chosen:
Model
Equation
Name = MgHarm
Active = on
Apply to Bodies = 1 2
Edit Solver Settings
Linear System
Iterative = BiCGStabL
Preconditioning = none
BiCGStabl order = 4
Name = MgDynPost
Active = on
Edit Solver Settings
Solver Specific Options
Calculate Magnetic Field Strength = on
Calculate Joule Heating = on
Skip Nodal Field = on
Discontinuous Bodies = on
General
Execute Solver = before saving
Add
OK
The Material section includes all the material parameters. In this case we basically have two different
materials – the copper wire and the surrounding air. We use the definitions from the small material database
that comes with the code, and the material properties are in SI Units.

Model
Material

CSC – IT Center for Science

14. Magnetic field induced by harmonic current in a wire 93

Material Library
Copper
Apply to Bodies = 1
Add
New

Material
Material Library
Air (room temperature)
Apply to Bodies = 2
Add
OK
We need to set the boundary conditions both for the scalar potential associated to the nodal degrees of
freedom, av, and to the vector potential associated to the edge degrees of freedom, av {e}. Both of these
have both real and imaginary components. We set the scalar potential only for the conductors and the vector
potential for all external boundaries.
Model
BoundaryCondition
Name = Ground
MgHarm
AV re = 0
AV re e 1 = 0
AV re e 2 = 0
AV im = 0
AV im e 1 = 0
AV im e 2 = 0
Add
OK
For the other of the wire we have exactly the same BCs except that the scalar potential is set to 0.01 V to
give the desired electric field of 1 V/m, since the length of the wire is 0.01 m.
Model
BoundaryCondition
Name = Voltage
MgHarm
AV re = 0.01
AV re e 1 = 0
AV re e 2 = 0
AV im = 0
AV im e 1 = 0
AV im e 2 = 0
Add
OK
and finally for all other external BCs we set the boundary non-axial components of the vector potential to
zero.
Model
BoundaryCondition
Name = AxialField
MgHarm
AV re e 1 = 0
AV re e 2 = 0
AV im e 1 = 0

CSC – IT Center for Science

14. Magnetic field induced by harmonic current in a wire 94

AV im e 2 = 0
Add
OK
The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking
with the mouse. Here we use the latter approach as that spares us of the need to know the indexes of each
boundary.
Model
Set boundary properties
Choose one end of the wire -> set boundary condition to Ground
Choose the other end of the wire -> set boundary condition to Voltage
Choose all other external BCs -> set boundary condition to AxialField
For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.

Sif
Generate
Edit -> look how your command file came out
Before we can execute the solver we should save the files in a directory. The project includes all the files
needed to restart the case.

File
Save Project
After we have successfully saved the files we may start the solver
Run
Start solver
A convergence view automatically pops up showing relative changes of each iteration. The equation is fully
linear and hence only two iterations are needed – the second one just ensures that convergence of the non-
linear level was really obtained. The convergence monitor also plots the postprocessing steps but they do not
really converge as each field is computed just once.
When the solution has finished we may start the postprocessor to view some results.
Run
Start ParaView

Results
Here we present some results of the computations. The visualization is done using Paraview and the vtu
format. For optimal visualization turn on the Discontinuous Bodies flag on both on the calculation
of the fields and while saving the results in vtu format. These flags ensure that the solution is continuous
within the bodies while maintaining jumps between bodies. No other formats in Elmer can currently support
these features.
The resulting absolute values of the imaginary and real part of the magnetic flux density are depicted in
Figure 14.2 and Figure 14.3. The imaginary part dominates in the amplitude (im 0.00522 T vs. re 0.00111 T).
The corresponding vector field of the magnetic flux density is depicted in Figure 14.5.
Why are the results so far from the steady state solution? If you run the case again with 100 Hz you can
see that the solution is very close to the static one as shown in Figure 14.4. The maximum magnetic flux
density from the computations is 0.0366 T which is very close to the analytical solution of 0.0374 T. So the
frequency really has an significant effect on the results!
You may study what happens when you increase the frequency further. At some point the mesh cannot
capture the solution properly any more and the results become questionable.

CSC – IT Center for Science

14. Magnetic field induced by harmonic current in a wire 95

Figure 14.2: Imaginary part of the magnetic flux density at 100k Hz.

Figure 14.3: Real part of the magnetic flux density at 100k Hz.

CSC – IT Center for Science

14. Magnetic field induced by harmonic current in a wire 96

Figure 14.4: Real part of the magnetic flux density at 100 Hz.

Figure 14.5: Imaginary part of the magnetic field strength plotted with vectors scaled by the field values.

CSC – IT Center for Science

Tutorial 15

Magnetostatics – Magnetic field

resulting from a permanent magnet

Directory: HorseshoeGUI
Solvers: MagnetoDynamics2D
Tools: Gmsh, ElmerGUI
Dimensions: 2D, Steady-state
Author: Peter Råback

Case definition
This case roughly reproduces the case of a permanent magnet as demonstrated in the following link:
http://www.strek.strefa.pl/students/meslec/lab06.pdf
Consider a horseshoe-shaped permanent magnet. It consists of a ferromagnetic material but the two end
sections are premagnetized in opposite directions. This results to a familiar magnetic field pattern. The
horseshoe consists of three different regions and additionally there is the surrounding air. There is a circular
outer boundary in order to conveniently allow for far field conditions. The material is assumed to have a
constant relative permeability of 5000 and the magnetization is set to 750 kA/m.
Note that as this is a 2D case the resulting fields actually are those of an infinitely long horseshoe which
of course does not make much sense in real life.

Meshing
The computational mesh is predefined in Gmsh format in file horseshoe.msh. If the user wants to modify
the default mesh that must be done with Gmsh. The geometry of the file is given in file horseshoe.geo.

Solution procedure
The definitions for the relevant equation are not loaded into ElmerGUI by default. Hence, one needs to load
these before starting the simulations.
File
Definitions
Append -> magnetodynamics2d.xml

The additional definitions should reside in the directory edf-extra within the distribution.

Optionally, one may copy the desired xml files to the edf-directory from the directory edf-extra
which will enable automatic loading of the definitions every time ElmerGUI is started. By inspecting the
definitions in the Elmer Definitions File editor one may verify that the new definitions were

CSC – IT Center for Science

15. Magnetostatics – Magnetic field resulting from a permanent magnet 98

really appended.

The mesh is already defined, load it from the samples directory where it resides.
File
Open -> horseshoe.msh

The ElmerGrid plug-in of ElmerGUI will read the mesh and convert it to a format understood by Elmer.

Figure 15.1: The mesh for the horseshoe and surrounding air as see in ElmerGUI

After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we
choose things related to the whole simulation such as file names, time stepping, constants etc. The steady-
state simulation is carried out in 2-dimensional Cartesian coordinates. Nothing needs to be changed here.
Currently the permeability of vacuum is not given in the ElmerGUI. To set other than the default value for
it, the free text box can be used.
In the equation section we choose the relevant equations and parameters related to their solution. In this
case we’ll have the MgDyn2D solver, as well as the postprocessing solver MgDyn2DPost.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use
mouse selection to assign them later. In this case the equations need to be solved in all the bodies and hence
clicking all from 1 to 4 here is most convenient. We give a higher priority to the actual solver so that the
vector potential will be computed before the derived fields. In this case solver specific options should be OK
but they could also be changed in this context.
Model
Equation
Name = MgDyn2D
Active = on
Priority = 1
Apply to Bodies = 1 2 3 4

CSC – IT Center for Science

15. Magnetostatics – Magnetic field resulting from a permanent magnet 99

Name = MgDyn2DPost
Active = on
Add
OK
The Material section includes all the material parameters. In this case we basically have two different
materials but the different magnetization must also be given as a material property. Hence we actually need
to define four materials.
Model
Material
Name = Air
MgDyn2d
Relative Permeability = 1.0
Add
New

Name = Iron
MgDyn2d
Relative Permeability = 5000.0
Add
New

Name = IronPlus
MgDyn2d
Relative Permeability = 5000.0
Magnetization 1 = Real 750.0e3
Add
New

Name = IronMinus
MgDyn2d
Relative Permeability = 5000.0
Magnetization 1 = Real -750.0e3
Add
OK
We may now assign the material properties by selecting with the mouse. This spares us of the need to
know the indexes of each body.
Model
Set body properties
Choose air -> set Material to Air
Choose curved part of horseshoe -> set Material to Iron
Choose upper straight part of horseshoe -> set Material to IronPlus
Choose lower straight part of horseshoe -> set Material to IronMinus

We have just one boundary condition i.e. the outer boundary for which we use the far field condition.
Model
BoundaryCondition
Name = Farfield
MgDyn2D
Infinity BC = True
Add
OK

CSC – IT Center for Science

15. Magnetostatics – Magnetic field resulting from a permanent magnet 100

The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking
with the mouse. Here we use the latter approach as that spares us of the need to know the indexes of each
boundary.
Model
Set boundary properties
Choose the 4 pieces of the outer sphere -> set boundary condition Farfield
For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.

Sif
Generate
Edit -> look how your command file came out
Before we can execute the solver we should save the files in a directory. The project includes all the files
needed to restart the case.

File
Save Project
After we have successfully saved the files we may start the solver
Run
Start solver
A convergence view automatically pops up showing relative changes of each iteration. The equation is
fully linear and hence only two iterations are needed – the second one just ensures that convergence of the
non-linear level was really obtained. The norm of the solution should be 0.3679.
When the solution has finished we may start the postprocessor to view some results.

Run
Start ParaView

Results
The resulting z-component of the vector potential is depicted in Figure 15.2. The corresponding postpro-
cessed magnetic field intensity is depicted in Figure 15.3. Note that the derived fields is enforced to be
continuous by default which is not optimal for visualization. For optimal results use Discontinuous Galerkin
(DG) method for the postprocessing. Note that when using DG the postprocessing should be done with .vtu
files and Paraview. The postprocessing tools of Elmer cannot deal with elementwise-fields.

CSC – IT Center for Science

15. Magnetostatics – Magnetic field resulting from a permanent magnet 101

Figure 15.2: The vector potential of the magnetic field.

Figure 15.3: A closeup of the vector potential combined with the magnetic field intensity vectors. Note that
the fields in the horseshoe itself have been masked away to demonstrate the well known field shape in the
free space.

CSC – IT Center for Science

Tutorial 16

Harmonic magnetic field – 2D –

Induction heating of a graphite crucible

Directory: InductionHeatingGUI
Solvers: MagnetoDynamics2DHarmonic
Tools: Gmsh, ElmerGUI
Dimensions: 2D, Axi-Symmetric
Author: Peter Råback

Case definition
At high temperatures the most practical method to heat up the crucible is by electromagnetic induction. The
induction coil generates an alternating current that flows through the crucible. The Ohmic resistance en-
countered by this current dissipates energy, thereby directly heating the crucible via internal heat generation.
The tutorial case is a simple axi-symmetric crucible that could be used, for example, to grow silicon
carbide (SiC) with physical vapour deposition. The crucible is made of dense graphite and isolated by
porous graphite. At the bottom of the crucible there is some SiC powder. The physical properties of the
material are given in Table 16.1. The dimensions of the induction heating crucible are given in Table 16.2.
We neglect the helicity of the coil and assume an average current density that may be computed easily
when the area of the coil is known, j0 = nI/A, where A is the coil area. Here we assume a current density
of 1.0e6 A/m2 . The frequency of induction heating f is assumed to be 50 kHz.
The permeability of vacuum is 4π10−7 if the other variables are in SI-units. Relative permeability is
assumed to be one in all materials.

Solution procedure
The definitions for the relevant equation are not loaded into ElmerGUI by default. Hence, one needs to load
these before starting the simulations.
File
Definitions

Table 16.1: Material parameters of the crucible

material ε κ [W/mk] σ (1/Ωm)

graphite 0.7 10.0 2.0E4
insulation 0.9 1.0 2.0E3
powder 0.5 25.0 1.0E4

CSC – IT Center for Science

16. Harmonic magnetic field – 2D – Induction heating of a graphite crucible 103

Table 16.2: Dimensions of the crucible

body part rinner router hinner houter
graphite 2.0 2.5 6.0 8.0
insulation 2.5 4.0 8.0 12.0
coil 5.0 5.5 8.0

Append -> magnetodynamics2d.xml

The additional definitions should reside in the directory edf-extra within the distribution. Moving the
desired xml files to the edf-directory enables automatic loading of the definitions at start-up. By inspecting
the definitions in the Elmer Definitions File editor one may inspect that the new definitions
were really appended.
The mesh is already defined, load it from the samples directory were it resides.

File
Open -> crucible.msh
The ElmerGrid plug-in of ElmerGUI will read the mesh and convert it to a format understood by Elmer.

Figure 16.1: The mesh for the crucible and coil

After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup
we choose things related to the whole simulation such as file names, time stepping, constants etc. Currently
the permeability of vacuum is not given in the ElmerGUI. To set other than the default value for it, the free

CSC – IT Center for Science

16. Harmonic magnetic field – 2D – Induction heating of a graphite crucible 104

text box can be used. The steady-state simulation is carried out in rotationally symmetric 2-dimensional
coordinates which must be changed.

Model
Setup
Coordinate system -> Axi symmetric

In the equation section we choose the relevant equations and parameters related to their solution. In this
case we’ll have the MgDyn2DHarmonic solver, as well as the postprocessing solver MgDyn2DPost.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use
mouse selection to assign them later. In this case the equations need to be solved in all the bodies and hence
clicking all from 1 to 6 here is most convenient. We give a higher priority to the actual solver so that the
vector potential will be computed before the derived fields. In this case solver specific options should be OK
but they could also be changed in this context.
Model
Equation
Solver -> MgDyn2DHarmonic
Active = on
Priority = 1
Angular Frequency = 50.0e3
Apply to Bodies = 1 2 3 4 5 6
Solver -> MgDyn2DPost
Active = on
Edit Solver Settings
Solver Specific Options
Target Variable Complex = on
Calculate Joule Heating = on
Name = Induction
Add
OK
The Material section includes all the material parameters. In this case we basically have two different
materials but the different magnetization must also be given as a material property. Hence we actually need
to define four materials. The thermal properties are not needed at this stage as we are only solving for the
induction. The internal losses of the coil are omitted and it is treated as an pure current source with material
properties of air.
Model
Material
Name = Air
MgDyn2dHarmonic
Relative Permeability = 1.0
Electric Conductivity = 0.0
Add
New

Name = Graphite
MgDyn2dHarmonic
Relative Permeability = 1.0
Electric Conductivity = 2.0e4
Add
New

Name = Insulation
MgDyn2dHarmonic

CSC – IT Center for Science

16. Harmonic magnetic field – 2D – Induction heating of a graphite crucible 105

Relative Permeability = 1.0

Electric Conductivity = 2.0e3
Add
New

Name = Powder
MgDyn2dHarmonic
Relative Permeability = 1.0
Electric Conductivity = 1.0e4
Add
New

We may now assign the material properties by selecting with the mouse. This spares us of the need to
know the indexes of each body.
Model
Set body properties
Choose external air -> set Material to Air
Choose coil -> set Material to Air
Choose insulation (outermost body) -> set Material to Insulation
Choose graphite (actual crucible) -> set Material to Graphite
Choose powder at the bottom of crucible -> set Material to Powder
Choose internal air above the powder-> set Material to Air

We need to provide a current source in order for the equation to have a non-trivial solution.
Model
BodyForce
Name = CurrentSource
MgDyn2DHarmonic
Current Density = 2.5e5
Add
OK
This must be also joined with the coil

Model
Set body properties
Choose coil -> set Body force to CurrentSource
We have just one boundary condition i.e. the outer boundary for which we use the far field condition.
Model
BoundaryCondition
Name = Farfield
MgDyn2DHarmonic
Infinity BC = True
Add
OK
The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking
with the mouse. Here we use the latter approach as that spares us of the need to know the indexes of each
boundary.
Model
Set boundary properties
Choose the 2 pieces of the exterior -> set boundary condition Far field

CSC – IT Center for Science

16. Harmonic magnetic field – 2D – Induction heating of a graphite crucible 106

Figure 16.2: Induction heating of a simple crucible. a) in-phase component of the magnetic field intensity
b) out-of-phase component of the magnetic field intensity c) Joule losses in the conductors

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.

Sif
Generate
Edit -> look how your command file came out
Before we can execute the solver we should save the files in a directory. The project includes all the files
needed to restart the case.
File
Save Project
After we have successfully saved the files we may start the solver

Run
Start solver
A convergence view automatically pops up showing relative changes of each iteration. The equation is
fully linear and hence only two iterations are needed – the second one just ensures that convergence of the
non-linear level was really obtained. The norm of the solution should be 0.3679.
When the solution has finished we may start the postprocessor to view some results.
Run
Start ParaView
With the given computational mesh the problem is solved in a few seconds. With the 6 542 linear triangles
the heating efficiency is estimated to be 18.9 W. The corresponding results are shown in Fig. 16.2.
It can be noted that would the estimated heating efficiency be different from the known one the user
may give the postprocessing solver the Desired Heating Power in order to scale the potential of the
solution. The solution, on the other hand, may be used as a source term in heat equation, for example.

CSC – IT Center for Science

Tutorial 17

Helmholtz – 2D – Acoustic Waves – Air

in a Cavity

Directory: AcousticWaves
Solvers: Helmholtz
Tools: ElmerGUI
Dimensions: 2D,Steady State
Author: original author not known

Introduction
Elmer provides two alternative ways of conducting acoustic analyses in the frequency domain. Firstly, one
may simply use the Helmholtz equation which is based on the assumption of lossless flow, i.e. the effects of
viscosity and heat conduction are assumed to be negligible. More refined analyses where these effects are
taken into account may be carried out by using the specific solver for the set of time-harmonic dissipative
acoustic equations. The aim of this tutorial is to demonstrate the usage of the solver for the basic Helmholtz
equation, which is frequently taken as the starting point in acoustic analyses.

Case definition
In this problem the fluctuations of the pressure in an air-
filled cavity shown in Figure 31.1 are considered. The
cavity is connected with the surrounding air by an open
narrow pipe. The pressure fluctuations are generated by
a vibrating membrane on the boundary ΓS with the fre-
quency of the motion being f = 100 Hz. The remaining
parts of the boundary are assumed to be rigid walls. In
addition, the effects of gravity are assumed to be negligi-
ble.
Suitable boundary conditions in terms of the pressure
must be given. On the rigid walls the pressure flux is
prescribed to vanish which corresponds to the assump-
tion that there is no velocity in the direction normal to
the boundary. At the open end Γ0 the impedance bound-
ary condition suitable for forward travelling plane waves
is given by setting Z = −c with c being the sound speed.
We assume that c = 343 (m/s). Finally, the wave source
Figure 17.1: Geometry
is given by defining a non-vanishing pressure flux on

CSC – IT Center for Science

17. Helmholtz – 2D – Acoustic Waves – Air in a Cavity 108

the corresponding part of the boundary. We take sim-

ply ∇P · ~n = 1 where P is the (complex) amplitude of the pressure and ~n is the outward unit normal to the
boundary.

ElmerGUI Equation Menu

We will be using the Helmholtz Solver, which is one of the default, pre-loaded GUI definitions, so it does
not need to be manually activated. For reference, the helmholtz.xml definition file is, by default, located
in Linux in:
$ELMER_HOME/share/ElmerGUI/edf

and in Windows, located in:

C:/Program Files/Elmer 9.0-Release/share/ElmerGUI/edf

Solution procedure
The mesh is given in ElmerGrid format in file domain.grd,
load this file.

File
Open -> domain

You should obtain your mesh and may check Model

Summary... that it consists of 10050 surface elements.
Your mesh should look like as shown in figure 31.5
After we have the mesh we start to go through the Model
menu from the top to bottom. In the Setup we choose things
related to the whole simulation such as file names, time step-
ping, constants etc.
The simulation is carried out in 2-dimensional Cartesian
coordinates.

Model
Setup Figure 17.2: Mesh.
Simulation Type = Steady state
Steady state max. iter = 1
Apply

In the equation section we choose the relevant equations

and parameters related to their solution.
In this case we’ll have one equation (named “Helmholtz”).
When defining Equations and Materials it is possible to
assign to the bodies immediately, or to use mouse selection to
assign them later. In this case we have just one body and there-
fore its easier to assign the Equation and Material to it directly.

For the linear system solvers we are happy to use the defaults. One may however, try out different
preconditioners (ILU1,. . . ) or direct Umfpack solver, for example.

Model
Equation
Name = Helmholtz
Apply to Bodies = 1
Helmholtz Equation
Active = on

CSC – IT Center for Science

17. Helmholtz – 2D – Acoustic Waves – Air in a Cavity 109

Angular Frequency = 628.3

Convection Velocity 1 = 0.0
Convection Velocity 2 = 0.0
Add
OK

The Material section includes all the material parameters. They are divided into generic parameters which
are direct properties of the material without making any assumptions on the physical model, such as the
mass. Other properties assume a physical law, such as conductivities and viscosity.

Model
Material
Apply to Bodies = 1
General
Density = 1.224
Helmholtz
Damping coefficient = 0.0
Sound speed = 343
Add
OK

A Body Force represents the right-hand-side of a equation. It is generally not a required field for a body,
and is not needed for this example.
Initial conditions should be given to transient cases, and probably are not needed for steady state solu-
tions, and is not needed for this example.
Only one boundary condition may be applied to each boundary and therefore all the different physical
BCs for a boundary should be grouped together.
The conditions may be assigned to boundaries in the Boundary condition menu, or by clicking on each
boundary with the mouse. Here we use the first approach since there are only three boundaries.

Model
BoundaryCondition
Name = Constraint1
Boundary 1 is checked
Helmholtz Equation
Flux conditions
Real part of the flux = 1
Imag part of the flux = 0
Add
New

Name = Constraint2
Boundary 2 is checked
Helmholtz Equation
Flux conditions
Real part of the flux = 0
Imag part of the flux = 0
Add
New

Name = Constraint3
Boundary 3 is checked
Helmholtz Equation
Flux conditions
Real part of the impedance = -343

CSC – IT Center for Science

17. Helmholtz – 2D – Acoustic Waves – Air in a Cavity 110

Imag part of the impedance = 0

Add
OK

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined
all the information for ElmerGUI to write the command file. After writing it we may also visually inspect
the command file.
Sif
Generate
Edit -> look how your command file came out
Before we can execute the solver we should save the files in a directory. The ElmerGUI project includes
all the files needed to restart the case.

File
Save Project

After we have successfully saved the files we may start the solver.

Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration.

When there are some results to view we may start the postprocessor also.

Run
Start ParaView

Results
The difference of the maximum of (0.32) and the minimum (0.21) value of the pressure is found to be
∆p ≈ 0.11
You may inspect the results with Paraview or with ElmerVTK.

In Figure 17.3 the obtained pressure wave magnitude is presented. In Figure 17.4 the obtained pressure
wave in the x-direction and y-direction is presented.

CSC – IT Center for Science

17. Helmholtz – 2D – Acoustic Waves – Air in a Cavity 111

Figure 17.3: Pressure wave magnitude

Figure 17.4: Pressure wave, x-direction, and y-direction

Extra task:
If you have time you may try to solve the case with different parameters.

CSC – IT Center for Science

Tutorial 18

VectorHelmholtz – model wave

propagation in bent waveguide

Directory: WaveguideGUI
Solvers: VectorHelmholtz
Tools: ElmerGUI
Dimensions: 3D, Steady-state
Author: Juhani Kataja

Case definition
In this tutorial we model propagation of a guided wave through a bend in a rectangular waveguide. Let us
choose the primary wave to be a T E10 mode at f = 2.5 GHz propagating in the positive z axis direction
(time-harmonic convention e−iωt ).
The electric field of the primary wave is given
by
~ p = ~uy ik sin(kc x)eiβz ,
E
kc
√
where kc = π/a, k = ω ε0 µ0 and β =
p
k 2 − kc2 . The dimensions of the waveguide’s
cross section are a = 10 cm and b = 5 cm, the
bend is a piece of circular torus with outer radius
of 12 cm and inner radius of 2 cm. The waveguide
geometry is depicted in Fig. 18.1.
The input port is that part of the boundary which
is normal to z -axis, the output port that which is
normal to x -axis and the rest of the boundary is a
PEC boundary.
The PEC boundary condition is given by ~n ×
E~ = 0, the input boundary condition by

~ − iβ~n × (~n × E)
~n × ∇ × E ~ = i2β~n × (~n × E
~ p ),

and the output boundary condition by

~ − iβ~n × (~n × E)
~n × ∇ × E ~ = 0.
Figure 18.1: Geometry of the waveguide
Here ~n is the exterior normal of the domain.

CSC – IT Center for Science

18. VectorHelmholtz – model wave propagation in bent waveguide 113

Load ElmerGUI Equation Menu

The GUI definitions of VectorHelmholtz solver utilized in this tutorial are located in the edf-extra folder
and thus need to be manually activated in the ElmerGUI.

Note that the extra definition needs to be loaded first, before any other steps in building any tutorial.
Once the extra definition has been loaded, and the ElmerGUI project has been saved, then the instructions
for the extra definition will be stored in the project, and you won’t have to load the extra definition again.

File
Definitions
Append -> vectorhelmholtz.xml
Close

The vectorhelmholtz.xml definition file is, by default, located in Linux in:

$ELMER_HOME/share/ElmerGUI/edf-extra

and in Windows, located in:

C:/Program Files/Elmer 9.0-Release/share/ElmerGUI/edf-extra

We will also be using the Result Output Solver, which is one of the default, pre-loaded GUI definitions,
so it does not need to be manually activated. For reference, the resultoutput.xml definition file is, by
default, located in Linux in:
$ELMER_HOME/share/ElmerGUI/edf

and in Windows, located in:

C:/Program Files/Elmer 9.0-Release/share/ElmerGUI/edf

After reading over several of the preceding tutorials that wrote results to the case.vtu file, and none
of those tutorial used Result Output Solver, you may be wondering why we now all of a sudden need to load
a special solver to output results?

Actually, the Result Output Solver has been loaded in each of the preceding tutorials, but it was done in
the background, without notifying the user. All one needs to do, is to enter a file name into the Post file field
in the Setup menu, as follows.

Model
Setup
Post file -> case.vtu
Apply

This will cause Result Output Solver to be loaded in the background and will run after saving the simu-
lation. To get back to your question, why add an equation entry for Result Output, the answer will be that
we want to demonstrate how to use some of the advanced features of the Result Output Solver. Please take a
look at the Elmer Models Manual and review the Result Output Solver. One of the features that we will use,
is to ask for the .vtu file to be output in ASCII, instead of in binary.

Another feature of Result Output Solver is the ability to save the results in formats other than .vtu, in-
cluding vtk, Dx, Gmsh, and GiD formats.

One last point, either enter a ‘case.vtu’ name into the Post file field of Setup, or add Result Output Solver
to an Elmer project. Don’t add both, it won’t hurt anything, but it will output two .vtu files, one will be valid
and the other will be empty.

CSC – IT Center for Science

18. VectorHelmholtz – model wave propagation in bent waveguide 114

ElmerGUI Solution Procedure

Load geometry:
File
Open -> waveguide_bend.step
As we are modelling a wave phenomenon at 2.5GHz the maximum H (maximum element size) in the
1
mesh must be around 10λ , where λ ≈ 8.3, so set Max H = 0.012. We mentioned the waveguide cross section
is 10 cm by 5 cm, which equals 0.100 meters by 0.050 meters. Taking the smaller dimension and dividing
by Max H, gives 0.050/0.012 = approximately four elements across the smaller dimension.

Within ElmerGUI, since the geometry input file is in Step format, Nglib from Netgen is used to mesh the
model into linear tetrahedral elements. Nglib by default will generate a mesh where at least two elements
will span the smaller dimension of the model. In this case, setting Max H = 0.012 to obtain about four
elements across the smaller dimension is a necessary step.

Note that one could use Elmergrid to increase the order of the tet mesh generated by nglib from linear to
quadratic, (from a command prompt, enter elmergrid 2 2 -increase) although in this case we keep the linear
mesh as generated by nglib.

Mesh
Configure..
Max H: 0.012
Apply
Mesh
Remesh

Add the VectorHelmholtz and VectorHelmholtzCalcFields solver modules:

Model
Equation -> Add..
Vector Helmholtz Post Process
Active = on
Priority 4
Result Output
Active = on
Priority 2
Edit solver settings
Solver specific options
Uncheck the box ‘Binary Output’
Apply
Vector Helmholtz Equation
Active = on
Apply to Bodies: Body 1
Angular Frequency = $ 2*pi*2.5e9
Priority 5
Free text input
$ a = 10e-2
$ b = 5e-2
$ c0 = 1/sqrt(8.854e-12*4*pi*10^-7)
$ omega=2*pi*2.5e9
$ k0 = omega/c0
$ kc = pi/a
$ beta0 = sqrt(k0^2-kc^2)
Edit Solver Settings
Linear system

CSC – IT Center for Science

18. VectorHelmholtz – model wave propagation in bent waveguide 115

Iterative = BiCGStabl
Preconditioning = vanka
BiCGStabl order = 6
Convergence tol = 1e-6
Apply
Add
OK
Here in the Free text input part we defined some constants that will make definition of the rest of
the model easier.

Then add boundary PEC conditions:

Model
Boundary Condition -> Add..
Name = PEC
VectorHelmholtz Equation
E re {e} = 0
E im {e} = 0
Apply to boundaries = 2...13
Add
OK
Note that because the mesh is tetrahedral, the E re/im {f} Dirichlet condition is unnecessary. The
hexahedral and pyramidal Piola transformed elements have DOFs on element faces rendering the {f}
Dirichlet conditions meaningful.
Add the input port boundary condition:
Model
Boundary Condition -> Add..
Name = Inport
Apply to boundaries = 14
VectorHelmholtz Equation
Magnetic Boundary Load 2 [enter]
Variable Coordinate 1
Real MATC "-2*beta0*k0/kc*sin(kc*(tx+a/2))"
Close
Electric Robin Coefficient im = $ beta0
Add
OK

Next add the output port boundary

Model
Boundary Condition -> Add..
Name = Outport
Apply to boundaries = 1
VectorHelmholtz Equation
Electric Robin Coefficient im = $ beta0
Add
OK
Assign material to the body

Model
Material -> Add..
Apply to bodies: Body 1

CSC – IT Center for Science

18. VectorHelmholtz – model wave propagation in bent waveguide 116

Relative Permittivity = 1
Add
OK
Now navigate to choose what results are to be saved:

Model
Equation -> Equation 1
Vector Helmholtz Post Process
Edit Solver Settings
Solver specific options
Calculate Electric Field = on
Calculate Magnetic Field Strength = on
Calculate Poynting Vector = on
Calculate Energy Functional = on
Apply
Update
OK
Save the project
File
Save project -> [choose location]

And solve:
Run
Start Solver
The resulting fields should now appear in case0001.vtu file ready for further post processing.

Results
We are interested in the Energy Functional Value in the solver log. It should read (prepended with
“VectorHelmholtzSolver:”)

Energy Functional value: -11284.937620324963 453999.53923919413

The first number is the real part and second the imaginary part. Denoting this with l(E) it holds that in this case

iβk02 ab
l(E) = (1 + ρ),
µkc2
where ρ is the reflection coefficient of electric field. Thus ρ ≈ −0.022 + 0.024i, which translates to roughly 29.7 dB
return loss.
In Fig. 18.2 the Poynting vector and the real part of the electric field of the solution field is shown.

Extra task:
Try outputting results using the .vtk format, and loading the .vtk file using Paraview. Also try saving the results in binary
as well as in ASCII, and compare the resulting .vtu files using a text editor.

CSC – IT Center for Science

18. VectorHelmholtz – model wave propagation in bent waveguide 117

Figure 18.2: The resulting field solution: The y -component of the real part of the electric field and the real
part of Poynting’s vector

CSC – IT Center for Science

Tutorial 19

Navier-Stokes equation – Laminar

incompressible flow passing a step

Directory: FlowStepGUI
Solvers: FlowSolve
Tools: ElmerGUI
Dimensions: 2D, Steady-state
Author: Peter Råback

Case definition
This tutorial represents the canonical step flow of viscous fluid. A fluid, flowing past a step (see figure 19.1), has the
density 1 kg/m and viscosity 0.01 kg/ms. The velocity profile at the inlet is parabolic with a mean velocity < vx >=
1.0 m/s and vy = 0.0 m/s. At the outlet only the vertical component is defined, vy = 0.0 m/s. At all other walls the
no-slip boundary condition, ~v = 0, is applied. Thus the Reynolds number for the case is around 100.

Figure 19.1: Geometry of the step flow problem

Mathematically the problem to be solved is


−∇ · (2µε) + ρ~
u · ∇~
u + ∇p = 0 in Ω
(19.1)
∇·~u = 0 in Ω
with the boundary conditions

 ux = 1 on Γinlet
ux = 0 on Γno−slip (19.2)
uy = 0 on Γinlet ∪ Γoutlet ∪ Γno−slip


where µ is the viscosity, ε is the strain tensor, ρ is the density, ~

u is the velocity and p is the pressure. It is assumed that
the density and viscosity are constants.

Solution procedure
The mesh is given in ElmerGrid format in file step.grd, load this file.

CSC – IT Center for Science

19. Navier-Stokes equation – Laminar incompressible flow passing a step 119

File
Open -> step.grd

You should obtain your mesh and may check that it consists of 9696 nodes and of 9442 bilinear elements.

Model
Summary...

After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we choose
things related to the whole simulation. The steady-state simulation is carried out in 2-dimensional Cartesian coordinates,
which are also the defaults.

Model
Setup
Simulation Type = Steady state
Coordinate system = Cartesian

In the equation section we choose the relevant equations and parameters related to their solution. In this case the only
the Navier-Stokes equation is needed.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use mouse selection
to assign them later. In this case we have just one body and therefore its easier to assign the Equation and Material to
it directly. One could also edit the solver setting in order to try different strategies for solving the non-linear or linear
system. Initially the Navier-Stokes solver uses the more robust Picard iteration which is changed to Newton iteration
after few initial steps. For the given viscosity the default values are ok, but may need tuning when going into higher
Reynolds numbers.

Model
Equation
Name = Navier-Stokes
Apply to Bodies = Body 1
Navier-Stokes
Active = on
Edit Solver Setting
Nonlinear System
Max. iterations = 20
Newton after iterations = 3
Add
OK

The Material section includes all the material parameters. They are divided into generic parameters which are direct
properties of the material without making any assumptions on the physical model, such as the density. Other properties
assume a physical law, such as viscosity.

Model
Material
Name = Ideal
General
Density = 1.0
Navier-Stokes
Viscosity = 0.01
Apply to Bodies = Body 1
Add
OK

The current case does not have any body forces. Convergence should also be obtained using the default initial
condition which sets all field values to zero. Hence no setting for initial condition are needed.
Only one boundary condition may be applied to each boundary and therefore all the different physical BCs for a
boundary should be grouped together. In this case the Temperature and Velocity. The side walls are assumed to be
adiabatic.
The parabolic inlet-profile is achieved using the MATC environment. To be able to edit the content of the inlet
profile click Enter to open an edit box for the Velocity 1. The given expression will be interpreted at run-time so
that vx = 6(y − 1)(2 − y). As y ∈ [1, 2] thereby creating a parabolic velocity profile with a mean velocity of unity.

CSC – IT Center for Science

19. Navier-Stokes equation – Laminar incompressible flow passing a step 120

Model
BoundaryCondition
Name = Inlet
Navier-Stokes
Velocity 1 = Variable Coordinate 2; Real MATC "6*(tx-1)*(2-tx)"
Velocity 2 = 0.0
Add
New

Name = Outlet
Navier-Stokes
Velocity 2 = 0.0
Add
New

Name = Walls
Navier-Stokes
Noslip wall BC = on
Add
OK

The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking with the mouse.
Here we use the latter approach as that spares us of the need to know the indexes of each boundary.

Model
Set boundary properties
Choose Inlet -> set boundary condition Inlet
Choose Outlet -> set boundary condition Outlet
Choose Walls -> set boundary condition Walls

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined all the
information for ElmerGUI to write the command file. After writing it we may also visually inspect the command file.

Sif
Generate
Edit -> look how your command file came out

Before we can execute the solver we should save the files in a directory. The project includes all the files needed to
restart the case. Create a suitable directory for the case if needed.

File
Save Project

After we have successfully saved the files we may start the solver

Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration. The problem should converge in
about ten iterations to a norm of 0.4347 visible on the output.
When there are some results to view we may start the postprocessor also

Run
Start ParaView

Results
The results may be viewed using the postprocessor as shown in Figure 19.2 and 19.3. One may also register specific
values, for example the pressure difference is 0.388 Pa, the minimum and maximum lateral velocities are -0.1666 m/s
and 1.5 m/s, respectively. One special result of interest is the point, on the x-axis, at which the direction of the flow
changes. In this case its position is about 5.0 m after the step.

CSC – IT Center for Science

19. Navier-Stokes equation – Laminar incompressible flow passing a step 121

Figure 19.2: Absolute value of the velocity field

Figure 19.3: Pressure field

Extra task: Decreasing the viscosity

Try what happens if the viscosity is further decreased by a factor 10. Convergence may be difficult to obtain. Some
tricks that may be tested include
• Introducing a relaxation factor (typically in the range 0.5–0.7)
• Increasing number of non-linear iterations
• Favoring Picard iteration over Newton
• Increasing mesh density (and length of domain)
Don’t be worried if you fail to find convergence. This task will mainly act as a motivator in using turbulence models for
higher Reynolds numbers.
Remember to re-perform the following phases in order to get the updated results

Sif
Generate
File
Save Project
Run
Start solver

You may just reload the results in the postprocessor rather than closing and opening the program.

CSC – IT Center for Science

Tutorial 20

Navier-Stokes equation – 2D –
Incompressible – Driven Cavity

Directory: DrivenCavity
Solvers: FlowSolve
Tools: ElmerGrid, ElmerGUI
Dimensions: 2D, Steady-state
Author: Evangelos Voyiatzis

Introduction
This simple tutorial illustrates the application of Elmer to computational fluid dynamics problems. The computational
mesh is generated using the ElmerGrid utility while the model definition and its solution are carried out via ElmerGUI.
The results are visualized with ParaView.

Case definition
The geometrical domain, which is shown in Figure 20.1, is a square with a side length of 1 m. The square is filled
with an ideal fluid with a density of 1 kg/m3 and a viscosity of 0.01 kg/m s. The boundary ΓA−B is moving with
a constant speed of 4 m/s in the x Cartesian direction while the no slip boundary condition is applied to the rest
of the boundaries. The flow is laminar with a Reynolds number of 400. This system has been studied in detail
by Ghia, Ghia and Shin in J. Comp. Phys. 48 (1982) 387-411 and it is also part of the NASA repository https:
//www.grc.nasa.gov/WWW/wind/valid/cavity/cavity.html.

The study by Ghia, et al, is run with a range of Reynolds numbers, from 100 up to 10,000. We will examine the case
of Reynolds number equal to 400. Using SI units, the driven lid moves at 4 m/s, the viscosity equals 0.01 Pa-s, and the
density equals 1.0 kg/m3̂. The values are chosen to provide the desired Reynolds numbers from the Ghia article simply
by changing the driven lid velocity. For example, with the lid moving at 4 m/s we have a Reynolds number of 400, if the
lid moves at 1 m/s we will have a Reynolds number of 100, and so forth.

The study by Ghia, et al, presents numerical results in two tables. The tabular results are the u velocity (in the
x-direction) along a vertical line through the geometrical center and the v velocity (in the y-direction) along a horizontal
line through the geometrical center. The two lines are shown in Figure 20.1. A comparison of the tabulated results from
Ghia against the results generated by Elmer will be presented.

CSC – IT Center for Science

20. Navier-Stokes equation – 2D – Incompressible – Driven Cavity 123

Figure 20.1: Geometry of the driven cavity problem

Mathematically the problem to be solved is


−∇ · (2µε) + ρ~
u · ∇~
u + ∇p = 0 in Ω
(20.1)
∇·~u = 0 in Ω
with the boundary conditions 
 ux = 4 on ΓA−B
ux = 0 on ΓB−C ∪ ΓC−D ∪ ΓD−A (20.2)
uy = 0 on Γ


where µ is the viscosity, ε is the strain tensor, ρ is the density, ~

u is the velocity and p is the pressure. It is assumed that
the density and viscosity are constants.

Solution procedure
The first step is the generation of a uniform computational mesh with 128 elements in each of the x and the y coordinate
directions. This is achieved by the following input file for ElmerGrid which is stored in the "DrivenCavity.grd" file,
which may be found in the tutorials-GUI-files/DrivenCavity folder:

*** ElmerGrid input file for structured grid generation ***

Version = 210903
Coordinate System = Cartesian 2D
Subcell Divisions in 2D = 1 1
Subcell Limits 1 = 0 1
Subcell Limits 2 = 0 1
Material Structure in 2D
1
End
Materials Interval = 1 1
Boundary Definitions
! type out int double of the boundaries
1 -1 1 1
2 -2 1 1
3 -3 1 1
4 -4 1 1
End
Element Degree = 2
Surface Elements = 16384

CSC – IT Center for Science

20. Navier-Stokes equation – 2D – Incompressible – Driven Cavity 124

Then start ElmerGUI and create a new project, as follows:

Run
New Project...

Start at the top and select the project directory as shown in Figure 31.3. Then select the Geometry input file,
DrivenCavity.grd. Lastly, the SaveLine Equation Definition File is needed for the tutorial, so select SaveLine in the right
hand box and add it to the left hand box. Click on OK to accept the changes.

Figure 20.2: New Project

ElmerGUI Equation Definition Files

The New Project screen automatically locates and lists all of the available EDFs. The left side of the screen lists the
Default EDFs, which will be loaded into each ElmerGUI project. The right hand box lists all of the extra EDFs, that are
not normally loaded, allowing the option to add individual extra EDFs to a new project.

This tutorial will use the Default EDF, navier-stokes.xml, for the Navier-Stokes Solver. In addition, we will
add the Extra EDF, saveline.xml, for the SaveLine Solver.
Finish by clicking on OK. ElmerGUI will generate a mesh which is visualized in Figure 20.3.

CSC – IT Center for Science

20. Navier-Stokes equation – 2D – Incompressible – Driven Cavity 125

Figure 20.3: The mesh for the driven cavity problem

You can verify that the mesh consists of 16384 surface elements by inspecting the dialog box that appears when
selecting

Model
Summary

We can now define the model going through the Model menu from the top to bottom. In the equation section we
choose the relevant equations and parameters related to their solution. We have only one set of equations which consists
of the incompressible Navier-Stokes equation. We also will use the SaveLine tab in the equations menu.

Model
Equation -> Add ..

We select the Navier-Stokes tab, click on the ‘Active’ box, and click to apply the equation to ‘Body 1’. We will
accept the default settings for the Navier-Stokes equation for now.
Since we have two solvers active for Body 1, you may ask, which solver will run first and which solver will run
second? Elmer has the ability to control when each solver will run. For example, we could add a priority number to each
solver, where the higher priority number solver will run first. In this case, we want the Saveline solver to run after the
Navier-Stokes solver has run. While on the Navier-Stokes tab, enter ‘2’ in the priority box, then switch to the Saveline
tab and enter ‘1’ in the priority box. Now the Navier-Stokes solver has the higher (numerically larger) priority number,
so it will execute before the Saveline solver.

While on the SaveLine tab, click on the ‘Active’ box. Verify that ‘Body 1’ is already selected. Don’t click on the
‘OK’ button just yet, we still need to add some info into the Saveline solver.

The equation tabs should look like Figure 20.4.

CSC – IT Center for Science

20. Navier-Stokes equation – 2D – Incompressible – Driven Cavity 126

Figure 20.4: Activate Equations, Navier-Stokes on left, Saveline on right

While the SaveLine tab is active, click on the ‘Edit Solver Settings’ tab, followed by the ‘Solver specific options’
tab, and enter the information below as shown. Saveline needs to know what name to use when it creates the results file,
in this case, we will enter ‘save-400.dat’. For Polyline Coordinates and for Polyline Divisions, enter

Size(4,2); 0.5 1.0 0.5 0.0 1.0 0.5 0.0 0.5

Size(2);20 20

The SaveLine tab should look like Figure 20.5. Now finish up by clicking on the ‘OK’ button to close the Equation
menu.
In a later section we will discuss the details of creating the SaveLine settings.

Figure 20.5: SaveLine Specific Settings

CSC – IT Center for Science

20. Navier-Stokes equation – 2D – Incompressible – Driven Cavity 127

The Material section includes all material parameters. They are divided into generic parameters which are direct
properties of the material without making any assumptions on the physical model, such as the density. Other properties
assume a physical law, such as the viscosity. In the material section, we select

Model
Material -> Add ..

The Material menu allows entry of any desired material properties, or one could select pre-defined materials from
the ‘Material Library’ button. In our case, we will enter the material values as stated under ‘Case Definition’ above.
In the dialog box that appears, we first click on the General tab and we set the density equal to 1 and apply it to "Body
1" as shown in left panel of Figure 20.6. Subsequently, we select the "Navier-Stokes" tab where we set the viscosity to
0.01, the "Compressibility model" to "Incompressible", apply it to "Body 1" and click ok as shown in the right panel of
Figure 20.6.

Figure 20.6: Material properties

For the linear system solvers we are happy to use the defaults. One may however, try out different preconditioners
(ILU1,. . . ) or direct Umfpack solver, for example.
The appropriate boundary conditions can be set via

Model
Boundary condition -> Add ..

In the dialog box that appears, we select the "Navier-Stokes" tab. For boundary condition 1, we apply the no slip
condition on boundaries 1, 2 and 4 by clicking on the "No wall slip BC" option in the Navier-Stokes tab as shown in the
left panel of Figure 20.7. For boundary condition 2, we apply the lid velocity with 4 m/s in the x-direction and 0 m/s in
the y-direction in the same tab after selecting "Boundary 3" as shown in the right panel of Figure 20.7.

CSC – IT Center for Science

20. Navier-Stokes equation – 2D – Incompressible – Driven Cavity 128

Figure 20.7: Boundary conditions for the driven cavity

For the execution, ElmerSolver needs the mesh files and the command file. We can write and view the command file
by selecting

Sif
Generate
Edit -> look how your command file came out

Before we can execute the solver we should save the files in a directory. The ElmerGUI project includes all the files
needed to restart the case.

File
Save Project

After we have successfully saved the files we may start the solver.

Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration. The problem should converge
in less than ten iterations. The variation of the residuals should be similar to the one in Figure 20.8. The required
computational time should be a couple of minutes.

CSC – IT Center for Science

20. Navier-Stokes equation – 2D – Incompressible – Driven Cavity 129

Figure 20.8: Convergence of the solver.

When there are some results to view we may start the postprocessor:

Run
Start ParaView

or

Run
Start ElmerVTK

Graphical Results
You may visualize the results with Paraview and/or ElmerVTK. Results from each of the postprocessors is shown, with
the Paraview results on the left and the ElmerVTK results on the right. In Figure 20.9 the magnitude of the obtained
velocity field is presented. The x- and y- cartesian components of the velocity vectors are shown in Figure 20.10 and
20.11, respectively.

The results for the case of Reynolds number = 400 are shown in Figure 20.9, Figure 20.10, and Figure 20.11.

CSC – IT Center for Science

20. Navier-Stokes equation – 2D – Incompressible – Driven Cavity 130

Figure 20.9: Re 400: Distribution of the magnitude of the velocity vector.

Figure 20.10: Re 400: Distribution of x component of the velocity vector.

CSC – IT Center for Science

20. Navier-Stokes equation – 2D – Incompressible – Driven Cavity 131

Figure 20.11: Re 400: Distribution of y component of the velocity vector.

SaveLine
The SaveLine subroutine contained in the SaveData solver is used to save Elmer solution data in ASCII format along
a line segment running through a 2D or 3D geometry model. The SaveLine subroutine and keywords are described in
detail in the Elmer Models Manual.
The ElmerGUI equation tab for SaveLine was previously shown in Figure 20.5. We entered data in several of the
fields, such as Filename, Polyline Coordinates, and Polyline Divisions.
For Filename, enter a descriptive name such as ‘save-400.dat’, which indicates the Reynolds number 400 case, with
an extension such as ‘dat’ or ‘csv’. The file name and extension can be freely chosen.
Filename Numbering and File Append are not used for this case, but can be used when multiple simulation runs are
performed, such as a transient simulation.
Polyline Coordinates and Polyline Divisions are described in the Elmer Models Manual as follows:

Polyline Coordinates(n,DIM) Real

Save the line consisting of line segments defined by two points (n = 2).
There can be more than one set of points (n = 2; 4; 6; : : :) but as a
line segment is defined by two points there must be an even number of points.
Polyline Divisions(n/2,DIM) Integer
The user may give the number of divisions for each polyline. This allows
also the proper saving of discontinuous data. The size of this vector should
be such that it is compatible with the number of lines.

Polylines could also be called ‘Multiple Line Segments’, because two end points must be defined for each line
segment. For one line segment, two points, for two line segments, four points, etc. Each point has the number of
coordinates defined by the simulation, two coordinates for 2D and three coordinates for 3D.
In this case, we are dealing with a 2D simulation, so there will be two coordinates with DIM = 2. The Ghia, et al,
article presents tabular data for a vertical line and a horizontal line, both of which pass through the geometric center.
We will need to define two line segments with four points, so ‘n’ = 4. If we multiply n * DIM, (4*2=8) then we get the
number of individual real numbers that must be listed.
Refer back to Figure 20.5 and in the Polyline Coordinate entry box you will see ‘Size(4,2);’ which defines ‘n’ and
‘DIM’, followed by a series of eight numbers.
The Polyline Divisions entry box contains ‘Size(2);’, where we enter ‘n/2’, (4/2=2), followed by two numbers, one
for each line segment. Each number indicates how many division are desired for each line segment. So if we want data
results printed for a desired number of points along each line segment, enter that number here. In this case, we want
twenty points along each line segment. On the other hand, if there is a sharp curve in the line segment that we want to
capture, dividing the line segment into 50 or 100 pieces, or splitting the single line segment into three line segments with
different divisions, will help add detail when we want to graph the results.

CSC – IT Center for Science

20. Navier-Stokes equation – 2D – Incompressible – Driven Cavity 132

SaveLine Results

Figure 20.12: Results comparison for Re = 400

The left side of the spreadsheet presents the u velocity data from Table 1 in Ghia, et al, and the right side of the
spreadsheet presents the results calculated by Elmer.
The Elmer results must be normalized to between 0.0 and 1.0, by dividing by the driven wall velocity, in this case
divided by 4.
There is decent agreement between the two sets of results with a Reynolds number of 400.

High Reynolds Numbers

The article by Ghia, et al, presents data over a range of Reynolds number. The actual values in the article are 100, 400,
1000, 3200, 5000, 7500, and 10,000. You may try to solve the cases with parameters corresponding to the given list of
Reynolds numbers. This can be achieved by modifying the velocity of the upper boundary from 1 m/s up to 100 m/s,
respectively.
Furthermore, although not shown in this tutorial, you can compare the obtained results with the data from the NASA
repository at https://www.grc.nasa.gov/www/wind/valid/cavity/cavity.html.
When running the range of Reynolds numbers, one will notice that starting at 50 m/s driven velocity the solution
will no longer converge when using the default solver settings. Convergence can be achieved up to the maximum speed
of 100 m/s, Reynolds number of 10,000, by changing the Navier-Stokes solver settings.
The Elmer Models Manual describes the details behind each of the documented solvers. The Navier-Stokes Flow
Solver is used for this driven cavity tutorial. Increasing the Reynolds number to 10,000 leads to issues with convergence.
The Linearization section of the Flow Solver chapter mentions: ‘... to first take a couple of Picard iterations, and switch
to Newton iteration after the convergence has begun.’ In this case, increasing the number of Picard iterations from 3 to
20, allows for convergence of the solution. One other setting change is needed for convergence in this situation, and that
is to change from iterative to direct (umfpack) on the Linear tab.

Model
Equation
Equation 1
select the Navier-Stokes tab
Edit Solver Settings
select the Linear system tab

CSC – IT Center for Science

20. Navier-Stokes equation – 2D – Incompressible – Driven Cavity 133

click on the direct button under Method

select Umfpack as the direct solver
select the Nonlinear system tab
Newton After iterations
change from 3 to 20
Apply
Update
OK

For the case of Reynolds number = 10,000, velocity profiles are shown in Figures 20.13, 20.14 and 20.15.

Figure 20.13: Re 10,000: Distribution of the magnitude of the velocity vector.

Figure 20.14: Re 10,000: Distribution of x component of the velocity vector.

CSC – IT Center for Science

20. Navier-Stokes equation – 2D – Incompressible – Driven Cavity 134

Figure 20.15: Re 10,000: Distribution of y component of the velocity vector.

CSC – IT Center for Science

20. Navier-Stokes equation – 2D – Incompressible – Driven Cavity 135

Saveline results for the case with Reynolds number = 10,000 are shown in Figures 20.16 and 20.17. Some deviation
between Elmer results and Ghia results can be observed at the high Reynolds number case. Overall, the shape of the
curves of the results show good agreement.

Figure 20.16: Re 10,000 u velocity profiles

Figure 20.17: Re 10,000 v velocity profiles

Extra task:
For additional credit, try increasing the Reynolds number above 10,000. For example, the solution will converge at Re =
20,000, if the non-linear relaxation factor is reduced from 1.0 to 0.7.

CSC – IT Center for Science

Tutorial 21

Navier-Stokes equation – Turbulent

incompressible flow passing a step

Directory: FlowStepKe
Solvers: FlowSolve,KESolver
Tools: ElmerGUI
Dimensions: 2D, Steady-state
Author: Peter Råback, Juha Ruokolainen

Case definition
This tutorial is a natural continuation of the tutorial 19 where the same case was solved with a smaller Reynolds number.
It is advisable to study that case before working with this case.
When the Reynolds number increases, the Navier-Stokes equations do not possess any steady-state solution. Ba-
sically the solution can be averaged simulating the transient flow over time. However, the computational cost of this
approach is often very heavy particularly while at high Reynolds numbers the computational mesh in direct numerical
simulation needs to be very dense. Instead its customary to solve time averaged equations. Unfortunately these equations
include unknown correlations between quantities that need to modelled in some way.
The workhorse of turbulence modelling is the k − ε model which is used in this tutorial. The k − ε model is a two-
equation model that introduces two additional variables – the turbulent kinetic energy k and the turbulent dissipation ε
which determines the scale of the turbulence.
The case under study is the canonical step flow of viscous fluid. A fluid, flowing past a step has the density 1 kg/m3
and viscosity 1.0e − 4 kg/ms. The velocity profile at the inlet is defined by a parabolic profile with mean velocity
vx = 1.0 m/s and vy = 0.0 m/s. This way the Reynolds number will be 10000. At the outlet only the vertical
component is defined, vy = 0.0 m/s. At all other walls the no-slip boundary condition, ~v = 0, is applied.
Also the new turbulent variables require boundary conditions. In the inlet the condition should reflect the values
for developed turbulent profile. Here we roughly estimate the turbulent kinetic energy and elongate the inlet distance
before the step to have a fully developed turbulent profile. From the literature it is the the turbulent intensity i.e. the
kinetic energy of turbulence vs. the kinetic energy of mean flow scales as 0.16Re−1/8 . For our current configuration an
estimate for the turbulent kinetic energy is 0.00457. For the walls a no-slip condition is applied which also set the values
of the turbulent parameters accordingly. Also a boundary layer model could be used but here our mesh should be able to
capture even the boundary phenomena quite well.

Solution procedure
The definitions for the relevant equation are not loaded into ElmerGUI by default. Hence, one needs to load these before
starting the simulations.

File
Definitions
Append -> k-epsilon.xml

CSC – IT Center for Science

21. Navier-Stokes equation – Turbulent incompressible flow passing a step 137

The additional definitions should reside in the directory edf-extra within the distribution.

Optionally, one may copy the desired xml files to the edf-directory from the directory edf-extra which will
enable automatic loading of the definitions every time ElmerGUI is started. By inspecting the definitions in the Elmer
Definitions File editor one may verify that the new definitions were really appended.

The mesh is given in ElmerGrid format in file steplong.grd, load this file.
File
Open -> steplong.grd
You should obtain your mesh and may check that it consists of 14584 nodes and of 14175 bilinear elements.
Model
Summary...
After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we choose
things related to the whole simulation. The steady-state simulation is carried out in 2-dimensional Cartesian coordinates,
which are also the defaults. The coupled system converges unfortunately quite slowly and hence we need to increase the
number of maximum iterations.
Model
Setup
Simulation Type = Steady state
Coordinate system = Cartesian
Steady state max. iter = 100
In the equation section we choose the relevant equations and parameters related to their solution. In this case the
Navier-Stokes and k − ε equations are needed. We want to solve the Navier-Stokes and k − ε equations iteratively
using only one non-linear iteration for optimal convergence of the coupled system. Some relaxation is needed in order to
achieve convergence at all. We also relax a little bit on the steady state convergence tolerance. Initially the Navier-Stokes
solver uses the more robust Picard iteration which may be changed to Newton iteration after the iteration progresses.
However, here we want to suppress the use of Newton linearisation since it seems to cause problems with the k − ε
equation.
Model
Equation
Name = Flow equations
Apply to Bodies = Body 1
Navier-Stokes
Active = on
Edit Solver Setting
Nonlinear System
Max. iterations = 1
Relaxation factor = 0.5
Newton after tolerance = 0.0
Steady state
Convergence tol. = 1.0e-4
K-Epsilon
Active = on
Edit Solver Setting
Nonlinear System
Max. iterations = 1
Relaxation factor = 0.5
Steady state
Convergence tol. = 1.0e-4
Add
OK
The Material section includes all the material parameters. They are divided into generic parameters which are direct
properties of the material without making any assumptions on the physical model, such as the density. Other properties
assume a physical law, such as viscosity. For the model parameters of the turbulent equations we are happy with the
defaults.

CSC – IT Center for Science

21. Navier-Stokes equation – Turbulent incompressible flow passing a step 138

Model
Material
Name = Ideal
General
Density = 1.0
Navier-Stokes
Viscosity = 1.0e-4
Viscosity Model = K-Epsilon
Apply to Bodies = Body 1
Add
OK

The current case does not have any body forces. To help in the convergence we make a rude initial guess.

Model
Initial Condition
Name = Initial Guess
Navier-Stokes
Velocity 1 = 0.0
Velocity 2 = 0.0
K-Epsilon
Kinetic Energy = 0.00457
Kinetic Dissipation = 1.0e-4
Apply to Bodies = Body 1
Add
OK

When defining Boundary conditions it is possible to assign to the boundaries immediately, or to use mouse selection
to assign them later. In this case we use the latter approach since we do not necessarily know the numbering of boundaries
by heart. There is a special boundary condition that takes care of the Boundary conditions for the no-slip walls for both
the Navier-Stokes and k − ε equation. Additionally there are inlet and outlet conditions. For the inlet click Enter
to open an edit box for the Velocity 1 when typing in the expression. which will be evaluated at run-time so that
vx = 6(y − 1)(2 − y).

Model
BoundaryCondition
Add
Name = Inlet
Navier-Stokes
Velocity 1 = Variable Coordinate 2; Real MATC "6*(tx-1)*(2-tx)"
Velocity 2 = 0.0
K-Epsilon
Kinetic Energy = 0.00457
Kinetic Dissipation = 1.0e-4
Add
New

Name = Outlet
Navier-Stokes
Velocity 2 = 0.0
Add
New

Name = Walls
Navier-Stokes
Noslip Wall BC = on
Add

The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking with the mouse.
Here we use the latter approach as that spares us of the need to know the indexes of each boundary.

CSC – IT Center for Science

21. Navier-Stokes equation – Turbulent incompressible flow passing a step 139

Model
Set boundary properties
Choose inlet -> set boundary condition Inlet
Choose outlet -> set boundary condition Outlet
Choose walls -> set boundary condition Walls

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined all the
information for ElmerGUI to write the command file. After writing it we may also visually inspect the command file.

Sif
Generate
Edit -> look how your command file came out

Before we can execute the solver we should save the files in a directory. The project includes all the files needed to
restart the case. Create a suitable directory for the case if needed.

File
Save Project

After we have successfully saved the files we may start the solver

Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration. The convergence if far from
monotonic but computation should terminate when sufficient convergence is reached after 30 iterations.

Results
When there are some results to view we may start Paraview for the postprocessing.

Run
Paraview

The results may be viewed using the postprocessor as shown in Figures 21.1 and 21.2. One may also register specific
values, for example the pressure difference is 0.302 Pa, the minimum horizontal and vertical velocities are -0.213 m/s
and -0.0834 m/s, respectively. One special result of interest is the point, on the x-axis, at which the direction of the flow
changes. In this case its position is about 5.1 m after the step.

Figure 21.1: Variables of the Navier-Stokes solver: absolute velocity on top and pressure on bottom

CSC – IT Center for Science

21. Navier-Stokes equation – Turbulent incompressible flow passing a step 140

Figure 21.2: Variables of the k − ε solver: kinetic energy on top and its dissipation on bottom

CSC – IT Center for Science

Tutorial 22

Navier-Stokes equation – Laminar

compressible flow over a step

Directory: FlowStepCompressible
Solvers: HeatSolve, FlowSolve
Tools: ElmerGUI
Dimensions: 2D, Steady-state
Author: Juha Ruokolainen, Rich Bayless

Case definition
This tutorial demonstrates how to simulate compressible air flowing past a step. The whole step has length of 1.4 m and
the height of 0.2 m and the first part of it has length of 0.4 m and the height of 0.1 m (Figure 22.1). The model has three
sets of boundary conditions. The air flows into the step from the inlet region and withdraws from the outlet region. The
other edges of the step compose the third boundary.

The flowing air is considered as an ideal gas in this case, and its density ρ depends on
the pressure p and temperature T through the equation of state where R is the gas constant. p
ρ= ,
The following table lists the material parameters for the gas used in this tutorial. Note that the RT
parameters being used in this tutorial are different from the parameters that would be retrieved
from the Material Library, if one selected the entry for ‘Air (Room Temperature).

Table 22.1: Material parameters.

parameter value
viscosity 16.7e-6 Ns/m2
heat conductivity 0.026 W/(m·K)
heat capacity 1.01e3 J/(kg·K)
specific heat ratio 1.4
reference pressure 1e5 Pa

The geometry looks as shown in figure 22.1

CSC – IT Center for Science

22. Navier-Stokes equation – Laminar compressible flow over a step 142

Figure 22.1: Geometry of the Step.

Check ElmerGUI Equation Menu

We will be using the FlowSolver and the HeatSolver, which are the default, pre-loaded GUI definitions, so they don’t
need to be manually activated. For reference, the heatequation.xml and navier-stokes.xml definition files
are, by default, located in Linux in:
$ELMER_HOME/share/ElmerGUI/edf

and in Windows, located in:

C:/Program Files/Elmer 9.0-Release/share/ElmerGUI/edf

Solution procedure
The mesh is given in ElmerGrid format in file mesh.grd, load this file.
File
Open -> mesh.grd
You should obtain your mesh and may check Model Summary... that it consists of 531 surface elements. Your
mesh should look like as shown in figure 22.2.

Figure 22.2: Mesh

After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we choose
things related to the whole simulation such as file names, time stepping, constants etc. The simulation uses 2D Cartesian
geometry and the problem is solved in steady state using no more than twenty steady state iterations.
Model
Setup
Simulation Type = Steady State
Steady state max. iter = 20
Apply
In the equation section we choose the relevant equations and parameters related to their solution.
In this case we’ll have one set of equations (named “Heat and Flow”) which consists of the Heat equation and of the
Navier-Stokes equation.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use mouse selection
to assign them later. In this case we have just one body and therefore its easier to assign the Equation and Material to
it directly. It is important to select the convection to be computed since that couples the velocity field to the heat equation.

For the linear system solvers we are happy to use the defaults. One may however, try out different preconditioners
(ILU1,. . . ) or direct Umfpack solver, for example.

CSC – IT Center for Science

22. Navier-Stokes equation – Laminar compressible flow over a step 143

Model
Equation
Name = Heat and Flow
Apply to Bodies = 1
Heat Equation
Active = on
Convection = Computed
Edit Solver Settings
General
Numerical techniques
Stabilize = uncheck
Bubbles = check
Apply (to exit Edit Solver Settings)
Navier-Stokes
Active = on
Edit Solver Settings
General
Numerical techniques
Stabilize = uncheck
Bubbles = check
Apply (to exit Edit Solver Settings)
Add
OK

The Material section includes all the material parameters. They are divided into generic parameters which are direct
properties of the material without making any assumptions on the physical model, such as the mass. Other properties
assume a physical law, such as conductivities and viscosity.

Our intention is to model compressible flow and that is why we have to set the value ”Perfect Gas” for the Compressibility
Model entry. Furthermore, because perfect gas model has been chosen the settings Reference Pressure and
Specific Heat Ratio must also be given. The Navier-Stokes equation also needs the value of viscosity and the
heat equation needs the values of heat capacity and heat conductivity. We will not use the Material Library, instead we
will enter our own values. The needed material parameters for an ideal gas for this tutorial are listed in table 22.1.

Model
Material
Name = Our Ideal Gas
Apply to Bodies = 1
General
Heat Capacity = 1.01e3
Specific Heat Ratio = 1.4
Reference Pressure = 1.0e5
Heat Equation
Heat Conductivity = 0.026
Navier-Stokes
Viscosity = 16.7e-6
Compressibility models = Perfect Gas
Add
OK

A Body Force represents the right-hand-side of a equation. It is generally not a required field for a body.

Initial conditions should be given to transient cases, and probably are not needed for steady state solutions. For the
initial condition of temperature we have chosen 300 K.

Model
Initial Condition
Name = Initial Guess
Apply to Bodies = 1
Heat Equation

CSC – IT Center for Science

22. Navier-Stokes equation – Laminar compressible flow over a step 144

Temperature = 300
Add
OK

Only one boundary condition may be applied to each boundary and therefore all the different physical BCs for a
boundary should be grouped together.
There are three sets of boundary conditions, so three Boundary Condition sections are needed. The first one is
used to prescribe the boundary conditions in the inlet region. Note that we have defined the x-velocity and temperature
as a variable of y-coordinate. This is done by setting different values for the x-velocity and temperature (the numerical
values of the second column between the words Real and End) in the different y-points (the numerical values of the
first column between words Real and End) of the inlet region. This kind of procedure prevents singularities from
occurring in the corner points of the inlet region. In addition, this kind of definition is more realistic than a constant
condition, in which the values of the x-velocity and temperature remain the same in the whole inlet region.

To enter the three pairs of numbers in the Heat Equation menu tab, click into the box labelled Temperature under
Dirichlet Conditions, then press the enter key. An editor window will pop up and you can enter all of the text from
‘Temperature’ to ‘Navier-Stokes’ into the window. Repeat the same steps in the Navier-Stokes menu tab, and enter all
the text from ‘Velocity 1’ to ‘Velocity 2’. Refer to Figure22.3, where the left side shows the Temperature boundary
condition for Heat and the right side shows the Velocity 1 boundary condition for Flow.

Figure 22.3: Left side: Heat BC, Right side: Flow BC

CSC – IT Center for Science

22. Navier-Stokes equation – Laminar compressible flow over a step 145

Model
BoundaryCondition
Name = Inlet
Heat Equation
Temperature = Variable Coordinate 2
Real
0.1 300
0.15 350
0.2 300
End
Navier-Stokes
Velocity 1 = Variable Coordinate 2
Real
0.1 0
0.15 0.02
0.2 0
End
Velocity 2 = 0.0
Add
New

After the rest of the boundary conditions have been defined the problem is ready to solve.

Name = Outlet
Navier-Stokes
Velocity 2 = 0.0
Add
New

Name = Walls
Heat Equation
Temperature = 300
Navier-Stokes
Noslip wall BC = on
Add
OK

The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking on each boundary
with the mouse. Here we use the latter approach as that spares us of the need to know the indexes of each boundary.

Model
Set boundary properties
Choose Inlet -> set boundary condition
Choose Outlet -> set boundary condition
Choose Walls -> set boundary condition
OK

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined all the
information for ElmerGUI to write the command file. After writing it we may also visually inspect the command file.

Sif
Generate
Edit -> look how your command file came out

Before we can execute the solver we should save the files in a directory. The ElmerGUI project includes all the files
needed to restart the case.

File
Save Project

After we have successfully saved the files we may start the solver

CSC – IT Center for Science

22. Navier-Stokes equation – Laminar compressible flow over a step 146

Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration.

When there are some results to view we may start the postprocessor also

Run
Start ParaView

Results
The simulation may take about a minute. You may inspect the results with Paraview.

Figure 22.4: Temperature distribution

Figure 22.5: Velocity distribution

Figure 22.6: Pressure distribution

CSC – IT Center for Science

22. Navier-Stokes equation – Laminar compressible flow over a step 147

Extra task:
If you have time you may try to solve the case with different parameters, such as doubling the peak inlet velocity in the
inlet boundary condition from 0.02 to 0.04. See how much you can increase the peak inlet velocity and still converge to
a solution.

CSC – IT Center for Science

Tutorial 23

Vortex shedding – von Karman

instability

Directory: VonKarmanGUI
Solvers: FlowSolve
Tools: ElmerGUI
Dimensions: 2D, Transient
Author: Juha Ruokolainen, Peter Råback

Case definition
This tutorial is about simulating the development of vortex shedding i.e. the von Karman instability. For full details
on the problem look at the benchmark case definition in the 1996 article by M. Schäfer and S. Turek, "Benchmark
computations of laminar flow around a cylinder".
The geometry is a 2D channel with a circular obstacle. The channel height is 0.41 meters and the channel length is
2.2 meters. The circular obstacle is 0.1 meters in diameter, located at x = 0.2 meters and y = 0.2 meters. The obstacle is
slightly off centerline of the channel.
The flow enters from x = 0 and exits at x = 2.2 meters. The upper and lower walls and the circular obstacle are set
to the no-slip condition, U = V = 0.0. The inlet flow has a parabolic profile, with a maximum flow velocity at centerline
of 1.5 m/s. The mean inlet flow velocity is then 1.0 m/s.
The Reynolds Number equals the mean velocity times the obstacle diameter, divided by the kinematic viscosity, Re
= U ∗ D/nu. Re = 1.0 m/s ∗ 0.1 m/0.001 m2 /s = 100. The fluid density equals 1.0 kg/m3 .

Solution procedure
The mesh is given in 2d netgen format in file circle_in_channel.in2d, load this file.

File
Open -> circle_in_channel.in2d

You should get a mesh consisting of 749 nodes and 1328 triangles, as shown in Figure 23.1. Note the increased mesh
density near the circular obstacle.

Figure 23.1: Computational mesh of the problem.

CSC – IT Center for Science

23. Vortex shedding – von Karman instability 149

This is a rather sparse mesh. To decrease the element size we can refine the mesh using ElmerGUI commands.

Mesh
Configure
nglib / Max H: 0.02
Mesh
Remesh

This refined mesh now includes 3464 nodes and 6506 triangles.
After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we choose
things related to the whole simulation such as file names, time stepping, constants etc. The simulation is carried out in
2-dimensional Cartesian coordinates. 2nd order bdf time stepping method is selected with 200 steps and we want the
total simulation time to be 8 seconds. Rather than calculating the time step manually, we will use MATC to perform
the calculation. MATC statements that start with a ‘$’ will be evaluated once at the program start. When added to an
ElmerGUI project the statement should be entered without internal spaces.

Model
Setup
Simulation Type = Transient
Steady state max. iter = 1
Time Stepping Method = bdf
BDF Order = 2
Time Step Intervals = 200
Time Step Sizes = $8/200

For the solver specific settings we are quite happy to use the defaults. However, we relax a little bit the convergence
tolerances to get speedier simulation.

Model
Equation
Name = Navier-Stokes
Apply to Bodies = 1
Navier-Stokes
Active = on
Edit Solver Settings
Nonlinear system
Convergence tol. = 1.0e-4
Linear System
Convergence tol. = 1.0e-6
Add
OK

The Material section includes all the material parameters. Here we choose simple parameters for the academic test case

Model
Material
Name = Ideal
General
Density = 1
Navier Stokes
Viscosity = 0.001
Apply to Bodies = 1
Add
OK

The system does not need any body forces. We are also happy with the default initial condition of zero and therefore
no initial conditions are applied either. Any other initial condition would require the values to be explicitly set.
We have three different kinds of boundaries: inlet, no-slip walls, and outlet. The inlet has a parabolic fully developed
laminar profile with a maximum velocity of 1.5 m/s. Additionally for the inlet the vertical velocity component is assumed
zero. The circle and the lower and upper walls are given the no-slip treatment. For the outlet only the vertical component
is set to zero since the default discretization weakly imposes a zero pressure condition if the normal velocity component
is not defined.

CSC – IT Center for Science

23. Vortex shedding – von Karman instability 150

MATC statements that start with ‘MATC’ will be evaluated repeatedly while the simulation is running. For more
details on using MATC, refer to ‘GetStartedElmer.pdf’ that contains the old Elmerfem wiki explanation.

Model
BoundaryCondition
Name = Inlet
Navier-Stokes
Velocity 1 = Variable Coordinate 2; Real MATC "4*1.5*tx*(0.41-tx)/0.41^2"
Velocity 2 = 0.0
Add
New

Name = Walls
Navier-Stokes
Velocity 1 = 0.0
Velocity 2 = 0.0
Add
New

Name = Outlet
Navier-Stokes
Velocity 2 = 0.0
Add
Ok

The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking with the mouse.
Here we use the latter approach as that spares us of the need to know the indexes of each boundary.

Model
Set boundary properties
Choose inlet -> set boundary condition Inlet
Choose both horizontal walls and circle -> set boundary condition Walls
Choose outlet -> set boundary condition Outlet

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined all the
information for ElmerGUI to write the command file. After writing it we may also visually inspect the command file.

Sif
Generate
Edit -> look how your command file came out

Before we can execute the solver we should save the files in a directory. The project includes all the files needed to
restart the case.

File
Save Project

After we have successfully saved the files we may start the solver

Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration. The norm after the first time step
should be around 0.695, and after last 0.749, respectively.
When there are some results to view we may start the postprocessor also

Run
Start ParaView

CSC – IT Center for Science

23. Vortex shedding – von Karman instability 151

Results
Due to the number of the time steps the simulation will take a few minutes. You may inspect the results with Paraview as
the time-steps are computed, or wait until all time steps have been computed. You need to reload the files if the number
of time steps are increased.
In Figure 23.2 the velocity field is presented for three different time steps. The maximum velocity in the system
should be about 2.1724 m/s. Note that here visualization was not performed with Paraview and may therefore look quite
different.

Figure 23.2: Velocity distribution at steps 20, 100 and 200

Effect of Reynolds number

The Reynolds number in this case is around 100 resulting to unsteady flow. The critical Reynolds number is around 90
and reducing the flow velocity so that Reynolds number becomes, say 20, makes the system to possess a steady-state
solution. On the other hand, increasing the velocity will make the von Karman vortices even more pronounced until they
break into fully chaotic motion. This specific finite element mesh will allow only a minor increase in Reynolds number
to be able to capture the phenomena.

CSC – IT Center for Science

Tutorial 24

Thermal flow in curved pipe

Directory: CurvedPipeGUI
Solvers: HeatSolve, FlowSolve
Tools: ElmerGUI
Dimensions: 3D, Steady-state
Author: Thomas Zwinger, Peter Råback

Case definition
This tutorial demonstrates how to set up a coupled case of thermal flow in curved pipe with a finite thickness. Within
the pipe both the flow and heat transfer equations need to be solved while on the solid section only heat transfer needs
to be considered.
The inner diameter of the pipe is 0.01 m and the outer 0.012 m, respectively. It is bent to a 135 degree angle with a
radius of 0.02 m. Beyond the bend 0.03 m of the direct part is also accounted for. The fluid flowing in the pipe is water
with an original temperature of 350 K. The outer temperature of the iron pipe is 300 K making the water gradually to
cool.
The water is injected with a parabolic velocity profile with a maximum of 0.01 m/s. In reality the laminar analytic
profile is described by the Bessel’s function. Here the flow is treated as laminar and steady-state even though at these
Reynolds number (about 100) the unsteady nature of the flow should probably be considered. This would enhance the
heat transfer. The steady-state case, however, will achieve the educational goals of the tutorial.

Solution procedure
The mesh is defined in ElmerGrid format in file curved_pipe.grd, load this file.

File
Open -> curved_pipe.grd

You should obtain your mesh and may check that it consists of 23670 trilinear bricks and 25245 nodes. The density
of the mesh may be varied by altering the Reference Density in the file. For further information on the mesh
definition look at the ElmerGrid manual. Often it is desirable to use some professional mesh generation tool in CFD
and translate it into Elmer format. For educational purposes we are quite happy to use this simple geometry.
After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we choose
things related to the whole simulation such as file names, time stepping, constants etc. The simulation is carried out in
3-dimensional Cartesian coordinates in steady-state. There is nothing really to be done here, but you may verify that the
defaults are correct.

Model
Setup
Coordinate system = Cartesian
Simulation type = Steady state
Steady state max. iter = 1
...

CSC – IT Center for Science

24. Thermal flow in curved pipe 153

Figure 24.1: The mesh of the curved pipe as seen in ElmerGUI

In the Equation section we choose the relevant equations and parameters related to their solution. In this case we’ll
have two different sets of solvers (called as Equation in Elmer slang). The other consists of heat and flow solvers, while
the other includes just the heat solver. We’ll name them appropriately.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use mouse selection
to assign them later. In this case we know that the fluid body has the index 1 and the solid body has the index 2. Therefore
it is easy to assign the Equation and Material to the bodies directly.
Here we neglect the effect of natural convection. Therefore there is just one-directional coupling from the flow to
heat transfer. In order to follow the direction of causality we address the flow solver with a higher priority than the heat
solver (default is zero).
Here we are quite happy with the default solver settings of the individual equations. However, for the flow solver we
change the default preconditioner ILU0 to ILU1 to enhance convergence (with increased memory consumption). For
3D cases the direct solvers are usually not feasible so it is better to stick with the iterative BiCGstab linear solver.
The equation for the fluid

Model
Equation
Add
Name = Heat and Flow
Apply to Bodies = 1
Heat Equation
Active = on
Convection = Computed
Navier-Stokes
Active = on
Priority = 1
Edit Solver Setting
Linear System
Preconditioning = ILU1
OK

and then for the solid

Model
Equation
Add
Name = Just Heat
Apply to Bodies = 2
Heat Equation
Active = on

CSC – IT Center for Science

24. Thermal flow in curved pipe 154

Convection = None
OK

The Material section includes all the material parameters. They are divided into generic parameters which are direct
properties of the material without making any assumptions on the physical model, such as the density. Other properties
assume a physical law, such as conductivities and viscosity.
Here we choose water and iron from the material library. You may click trough the material parameters of the various
solvers to ensure that the properties are indeed as they should be. Any consistent set of units may be used in Elmer. The
natural choice is of course to perform the computations in SI units.

Model
Material
Add
Material library
Water (room temperature)
Apply to Bodies = 1
OK

Add
Material library
Iron (generic)
Apply to Bodies = 2
OK

The Body force section usually represents the right-hand-side of an equation. It could be used to account for the
natural convection, for example. In this case, however, we do not apply any body forces.
Also an Initial condition could be given in steady-state case to enhance convergence. However, in this case
convergence is pretty robust with the default guess of zero.
We have four different boundary conditions: thermal inflow, internal no-slip, outflow, and external fixed temperature.
Otherwise natural BCs are assumed. As it is tedious to know the indexes by heart we first define the different BCs and
only afterwards apply them to the existing boundaries with the mouse.

Model
BoundaryCondition
Name = HotInflow
Heat Equation
Temperature = 350.0
Navier-Stokes
Velocity 1 = 0.0
Velocity 2 = 0.0
Velocity 3 = Variable Coordinate
Real MATC "100.0*(1.0e-4-tx(0)^2-tx(1)^2)"
Add
New

The condition for Velocity 3 above may easiest be typed by pressing Enter-key in the edit box which will open a
larger window for editing.

Name = Outflow
Navier-Stokes
Use normal-tangential coordinate system = on
Velocity 2 = 0.0
Velocity 3 = 0.0
Add
New

Name = NoSlip
Navier-Stokes
NoSlip Wall BC = on
Add
New

CSC – IT Center for Science

24. Thermal flow in curved pipe 155

Name = Troom
Heat Equation
Temperature = 300.0
Add

When choosing the boundaries it is important that you choose the right inlet. For that purpose you may activate the
compass,

View
Compass = on

Now the inlet is the one with normal pointing at the z-direction. Now we are ready to choose the boundaries

Model
Set boundary properties
Choose inlet face -> set boundary condition HotInflow
Choose outlet face -> set boundary condition Outflow
Choose outer side -> set boundary condition Troom

Unfortunately we cannot see the internal boundary. For that purpose click on the outer boundary and choose

View
Hide/show selected

The outer boundary should vanish and you can proceed with the last BC,

Model
Set boundary properties
Choose internal side -> set boundary condition Noslip

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined all the
information for ElmerGUI to write the command file. After writing it we may also visually inspect the command file.

Sif
Generate
Edit -> look how your command file came out

Before we can execute the solver we should save the files in a directory. The project includes all the files needed
to restart the case. It’s a good idea to give the project an illuminating name. Avoid paths which includes empty spaces
since they may cause problems later on.

File
Save Project
Make New Folder -> curved_pipe
OK

After we have successfully saved the files we may start the solver

Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration. The simulation may take a few
minutes depending on your platform. After the simulation has terminated we may study the results.

Results
The computed norms should be 3.255 for the Navier-Stokes equation and 324.66 for the heat equation. If there is some
discrepancy the setup of the system was probably different from the one in the tutorial.
After the simulation has terminated we may open the postprocessor to view the results.

Run
Start ParaView

CSC – IT Center for Science

24. Thermal flow in curved pipe 156

Figure 24.2: Convergence history of the case

Figure 24.3: Temperature distribution at the outlet of the pipe

A standard way of visualizing is to choose ColorMesh and there choose Surface and the desired field variable,
for example Velocity_abs or Temperature. In this case only the outflow cross section contains any information.
It may be seen in Figure 24.3 that the symmetry around pipe origin is lost in the bend.
Alternatively we may visualize the cross section at y = 0. To that aim choose Isocontours and there set
the Number Of Isosurfaces to 1, choose Surface, set Contour Variable to nodes_y, and Color
Variable to Temperature etc. Now you may nicely see how the velocity profile affects the temperature distri-
bution in the pipe.
In Figures 24.5 and 24.4 the obtained velocity and temperature distributions are presented.

CSC – IT Center for Science

24. Thermal flow in curved pipe 157

Figure 24.4: Velocity distribution at the cross section y = 0.

Figure 24.5: Temperature distribution at the cross section y = 0.

CSC – IT Center for Science

Tutorial 25

Interaction between fluid flow and

elastic obstacle

Directory: FsiObstacleGUI
Solvers: FlowSolve,ElasticSolve,MeshSolve
Tools: ElmerGUI
Dimensions: 2D, Steady-state
Author: Peter Råback

Case definition
This tutorial demonstrates how to set up a coupled case of fluid-structure interaction. Flow is initiated at one end of a
channel which has an elastic obstacle that bends under the influence of fluid forces. The resulting displacements modify
the domain thereby affecting the flow of the fluid.
The channel is assumed to be 2D. The length of the channel is 10 m and the height is 2 m. At the distance of 2 m
from the inlet sits an elastic beam with a rounded top the height of which is 1.2 m and width 0.4 m. A parabolic velocity
profile with maximum velocity of 1 m/s is assumed.
Material properties are assumed to be rather simple: For the structure the density is 1000 kg/m3 , Young’s modulus
is 1000 Pa, and Poisson ratio 0.3. For the fluid the density is 1 kg/m3 and viscosity is 0.1 Pas. Additionally the fluid has
elastic properties that are used to extend the displacement of the elastic beam to the fluid.
The idea of the case definition is to demonstrate simple yet strongly coupled FSI case without getting into turbulence
modelling. Realistic material parameters for the given size of the domain would easily result to turbulence and just small
displacements.
The case is solved using standard weak coupling with some relaxation to boost the convergence. The solution is
steady-state so only the final results are will be studied.

Solution procedure
The non-linear elasticity equation is not by default active in the menu structures ElmerGUI. Hence, the user must load
these before starting the simulations.

File
Definitions
Append -> nonlinearelastity.xml

The additional definitions should reside in the directory edf-extra within the distribution. Moving the desired xml
files to the edf-directory enables automatic loading of the definitions at start-up. By inspecting the definitions in the
Elmer Definitions File editor one may inspect that the new definitions were really appended.
The mesh is defined in .in2d format, the 2D format of netgen, in file obstacle_in_channel.in2d, load this
file.

File
Open -> obstacle_in_channel.in2d

CSC – IT Center for Science

25. Interaction between fluid flow and elastic obstacle 159

The default mesh is obviously too sparse. To make the mesh more dense set

Mesh -> Configure -> nglib -> Max H: 0.1

and choose

Mesh -> Remesh

You should obtain a denser mesh and may check Model Summary... that it consists of around 4140 nodes and 7890
linear triangles.

Figure 25.1: The mesh of the obstacle in channel case as seen in ElmerGUI

After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we choose
things related to the whole simulation such as file names, time stepping, constants etc. The simulation is carried out in
2-dimensional Cartesian coordinates in steady-state. There is not much to do here, just increase the number of iterations
needed for the convergence of the coupled system. We also set the output interval to zero which means that results are
written only at the end of the case.

Model
Setup
Coordinate system = Cartesian
Simulation type = Steady state
Steady state max. iter = 100
Output Intervals = 0
...

In the Equation section we choose the relevant equations and parameters related to their solution. In this case
we’ll have two different sets of solvers (called as Equation in Elmer slang). The fluid domain consists of flow and
mesh deformation solvers, while the elastic domain just includes the non-linear elasticity solver. We’ll name them
appropriately.
To enhance the convergence and efficient use of resources we set relaxation factors of the primary solvers to 0.5 and
the number of non-linear iterations to 1. The mesh deformation solver just extends the displacements of the elasticity
solver to the fluid domain, so no relaxation is needed here. For the linear systems we are quite happy with the defaults.
To honor the causality the flow solver should be solved first, then the elasticity solver and at last the mesh deformation
solver. We set the priorities accordingly.
The equation for the fluid flow + mesh deformation

Model
Equation
Add
Name = Flow and mesh deform
Apply to Bodies = 1
Navier-Stokes
Active = on
Priority = 2
Edit Solver Setting
Nonlinear System
Max.iterations = 1
Nonlinear System Relaxation Factor = 0.5

CSC – IT Center for Science

25. Interaction between fluid flow and elastic obstacle 160

Mesh Update
Active = on
Priority = 0
OK

and then for the solid

Model
Equation
Add
Name = Elasticity
Apply to Bodies = 2
Nonlinear Elasticity
Active = on
Priority = 1
Edit Solver Setting
Nonlinear System
Max.iterations = 1
Nonlinear System Relaxation Factor = 0.5
OK

Next we set our rather simple material parameters. The Material section includes all the material parameters. They
are divided into generic parameters which are direct properties of the material without making any assumptions on the
physical model, such as the density. Other properties assume a physical law, such as conductivities and viscosity.

Model
Material
Add
Name = Ideal fluid
General
Density = 1.0
Navier-Stokes
Viscosity = 0.1
Mesh Update
Elastic Modulus = 1.0
Poisson Ratio = 0.3
Apply to Bodies = 1
OK

Add
Name = Ideal structure
General
Density = 1.0e3
Nonlinear Elasticity
Youngs Modulus = 1.0e3
Poisson Ratio = 0.3
Apply to Bodies = 2
OK

The Body force section usually represents the right-hand-side of an equation. In this case we do not need any
body forces.
Also an Initial condition could be given in steady-state case to enhance convergence. However, in this case
convergence is pretty robust with the default guess of zero.
We have five different boundary conditions: inflow, outflow, lateral walls with no-slip conditions, fsi conditions, and
the beam base. As it is tedious to know the indexes by heart we first define the different BCs and only afterwards apply
them to the existing boundaries with the mouse.

Model
BoundaryCondition
Name = Inflow
Navier-Stokes

CSC – IT Center for Science

25. Interaction between fluid flow and elastic obstacle 161

Velocity 1 = Variable Coordinate 2

Real MATC "tx*(2-tx)"
Velocity 2 = 0.0
Mesh Update 1 = 0.0
Add
New

The condition for Velocity 1 above may easiest be typed by pressing Enter-key in the edit box which will open a
larger window for editing.

Name = Outflow
Navier-Stokes
Velocity 2 = 0.0
Mesh Update
Mesh Update 1 = 0.0
Add
New

Name = Walls
Navier-Stokes
NoSlip Wall BC = on
Mesh Update
Mesh Update 1 = 0.0
Mesh Update 2 = 0.0
Add
New

Name = Base
Nonlinear Elasticity
Displacement 1 = 0.0
Displacement 2 = 0.0
Add
New

The essence of fluid-structure interaction is in the following boundary condition. When the Fsi BC is active the fluid
forces are automatically within ElasticSolver. The back coupling to Navier-Stokes is achieved through the change in
fluid geometry which is enforced by the conditions for the MeshSolver.

Name = FSI
Nonlinear Elasticity
FSI BC = on
Navier-Stokes
NoSlip Wall BC = on
Mesh Update 1 = Equals Displacement 1
Mesh Update 2 = Equals Displacement 2

Now we are ready to choose the boundaries

Model
Set boundary properties
Choose inlet side -> set boundary condition Inflow
Choose outlet side -> set boundary condition Outflow
Choose upper and lower sides (three pieces) -> set boundary condition Walls
Choose obstacle base -> set boundary condition Base
Choose interface between fluid and solid (two pieces) -> set boundary condition FSI

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined all the
information for ElmerGUI to write the command file. After writing it we may also visually inspect the command file.

Sif
Generate
Edit -> look how your command file came out

CSC – IT Center for Science

25. Interaction between fluid flow and elastic obstacle 162

Before we can execute the solver we should save the files in a directory. The project includes all the files needed
to restart the case. It’s a good idea to give the project an illuminating name. Avoid paths which includes empty spaces
since they may cause problems later on.

File
Save Project
Make New Folder -> fsi_obstacle
OK

After we have successfully saved the files we may start the solver

Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration. The simulation may take around
10 seconds depending on your platform.
The computed norms should be around 0.514 for the Navier-Stokes solver, 0.108 for the elasticity solver, and 0.0548
for the mesh update solver. These are reached after 18 iterations using the rather strict default settings.
If there is some discrepancy the setup of the system was probably different from the one in the tutorial. If the results
are in agreement we are ready to look at the results.

Results
To visualize the results open the postprocessor, in this case ParaView After the simulation has terminated we may
open the postprocessor to view the results.

Run
Start ParaView

For flow problems we can visualize pressure (or velocity amplitudes) separately with the vector presentation of the
velocity field. In Paraview the filter to apply the displacement field to the nodal coordinates is called WarpByVector.
The maximum speed in the system is around 2.328 and the maximum displacement 0.2636. Note that for the saved
results the displacement and mesh update fields have been merged into one. In Figures 25.2, 25.3, and 25.4 the obtained
velocity, pressure and displacement distributions are presented in the deformed mesh. A close up showing the deformed
mesh is shown in figure 25.5

Figure 25.2: Velocity distribution of the obstacle in channel case.

CSC – IT Center for Science

25. Interaction between fluid flow and elastic obstacle 163

Figure 25.3: Pressure distribution of the obstacle in channel case.

Figure 25.4: Displacement distribution of the obstacle in channel case.

CSC – IT Center for Science

25. Interaction between fluid flow and elastic obstacle 164

Figure 25.5: Deformed mesh of the obstacle in channel case.

CSC – IT Center for Science

Tutorial 26

Flow and Heat –2D – Transient –

Rayleigh-Benard instability

Directory: RayleighBenardGUI
Solvers: HeatSolve, FlowSolve
Tools: ElmerGUI
Dimensions: 2D, Transient
Author: Juha Ruokolainen, Peter Råback

Case definition
This tutorial is about simulating the developing of the Rayleigh-Benard instability in a rectangular domain (Figure 26.1)
of dimensions 0.01 m height and 0.06 m length. The simulation is performed with water and the material parameters
of water required by the Elmer model are presented in Table 26.1 and can be loaded from the Material Library. The
temperature difference between the upper and lower boundary is set to 0.5 so that lower one has the temperature of 293.5
K and the upper one has the temperature of 293 K.
The density of water is inversely proportional to its temperature. Thus, heated water starts to flow upwards, and
colder downwards due to gravity. In this case we assume that the Boussinesq approximation is valid for thermal incom-
pressible fluid flow. In other words, the density of the term ρf~ in the incompressible Navier-Stokes equation can be
redefined by the Boussinesq approximation

ρ = ρ0 (1 − β(T − T0 ))

where β is the heat expansion coefficient and the subscript 0 refers to a reference state.

Figure 26.1: Domain.

CSC – IT Center for Science

26. Flow and Heat –2D – Transient – Rayleigh-Benard instability 166

Table 26.1: Material parameters for water

parameter value
density 998.3 kg/m3
viscosity 1040e-6 Ns/m2
heat capacity 4183 J/(kg·K)
heat conductivity 0.58 W/(m·K)
heat expansion coefficient 2.07e-4 K−1
reference temperature 293 K

Solution procedure
The mesh is given in ElmerGrid format in file rectangle.grd, load this file.

File
Open -> rectangle.grd

You should obtain your mesh and may check Model Summary... that it consists of 3036 bilinear elements. The
geometry and mesh should look like figure 26.1.
There is a possibility to divide and unify edges to simplify the case definition in the future.

Choose (left wall + right wall (Ctrl down)) -> unify edge

After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we choose
things related to the whole simulation such as file names, time stepping, constants etc. The simulation is carried out in
2-dimensional Cartesian coordinates. 2nd order bdf time stepping method is selected with 200 steps and with step size
of two seconds. Gravity is needed for the buoyancy force and it is defined by a vector with four components. The first
three components define a unit vector and the fourth its magnitude.

Model
Setup
Simulation Type = Transient
Steady state max. iter = 20
Time Stepping Method = bdf
BDF Order = 2
Time Step Intervals = 200
Time Step Sizes = 2.0
Gravity = 0 -1 0 9.82

In the equation section we choose the relevant equations and parameters related to their solution. In this case we’ll
have one set of equations (named “Natural Convection”) which consists of the heat equation and of the Navier-Stokes
equation.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use mouse selection
to assign them later. In this case we have just one body and therefore its easier to assign the Equation and Material to it
directly. It is important to select the convection to be computed since that couples the velocity field to the heat equation.
The system may include non-linear iterations of each equation and steady state iterations to obtain convergence of
the coupled system. It is often a good idea to keep the number of non-linear iterations in a coupled case low. Here we
select just one non-linear iteration for both equations.
For the linear system solvers we are happy to use the defaults. One may however, try out different preconditioners
(ILU1,. . . ) or direct Umfpack solver, for example.

Model
Equation
Name = Natural Convection
Apply to Bodies = 1
Heat Equation
Active = on
Convection = Computed
Edit Solver Setting

CSC – IT Center for Science

26. Flow and Heat –2D – Transient – Rayleigh-Benard instability 167

Nonlinear System
Max. iterations = 1
Navier-Stokes
Active = on
Edit Solver Setting
Nonlinear System
Max. iterations = 1
Add
OK

The Material section includes all the material parameters. They are divided into generic parameters which are direct
properties of the material without making any assumptions on the physical model, such as the mass. Other properties
assume a physical law, such as conductivities and viscosity.
Here we choose water at room temperature from the Material Library. You may click through the material parameters
of the various solvers to ensure that the properties are indeed as they should be. Any consistent set of units may be used
in Elmer. The natural choice is of course to perform the computations in SI units.
Apart from the properties from the material database, we enter a reference temperature for the Boussinesq approxi-
mation.

Model
Material
Apply to Bodies = 1
Material library
Water (room temperature)
General
Reference Temperature = 293
Add
OK

A Body Force represents the right-hand-side of a equation. It is generally not a required field for a body. In this
case, however, we apply the buoyancy resulting from heat expansion as a body force to the Navier-Stokes equation.

Model
Body Force
Name = Buoyancy
Apply to Bodies = 1
Navier-Stokes
Boussinesq = on
Add
OK

Initial conditions should be given to transient cases. In this case we choose a constant Temperature field and an
small initial velocity that initializes the symmetry break.

Model
Initial Condition
Name = Initial Guess
Heat Equation
Temperature = 293
Navier-Stokes
Velocity 1 = 1.0e-9
Velocity 2 = 0.0

Only one boundary condition may be applied to each boundary and therefore all the different physical BCs for a
boundary should be grouped together. In this case the Temperature and Velocity. The side walls are assumed to be
adiabatic.

Model
BoundaryCondition
Name = Bottom
Heat Equation

CSC – IT Center for Science

26. Flow and Heat –2D – Transient – Rayleigh-Benard instability 168

Temperature = 293.5
Navier-Stokes
Velocity 1 = 0.0
Velocity 2 = 0.0
Add
New

Name = Top
Heat Equation
Temperature = 293
Navier-Stokes
Velocity 1 = 0.0
Velocity 2 = 0.0
Add
New

Name = Sides
Navier-Stokes
Velocity 1 = 0.0
Velocity 2 = 0.0
Add

The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking with the mouse.
Here we use the latter approach as that spares us of the need to know the indexes of each boundary.

Model
Set boundary properties
Choose Bottom -> set boundary condition Bottom
Choose Top -> set boundary condition Top
Choose Sides -> set boundary condition Sides

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined all the
information for ElmerGUI to write the command file. After writing it we may also visually inspect the command file.

Sif
Generate
Edit -> look how your command file came out

Before we can execute the solver we should save the files in a directory. The ElmerGUI project includes all the files
needed to restart the case.

File
Save Project

After we have successfully saved the files we may start the solver

Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration.

When there are some results to view we may start the postprocessor also

Run
Start ParaView

Results
Due to the number of the time steps the simulation may take around ten minutes. You may inspect the results with
Paraview as the time steps are computed, or wait until all time steps have been computed. You must reload the files if
their number has changed. The time series can be automatically animated and even saved to an animation file.
In Figures 26.2 through 26.3 the obtained temperature and velocity distribution are presented. The maximum veloc-
ity in the system should be about 0.516 mm/s.

CSC – IT Center for Science

26. Flow and Heat –2D – Transient – Rayleigh-Benard instability 169

Figure 26.2: Temperature distribution

Figure 26.3: Velocity magnitudes

Extra task: Sensitivity to temperature difference

If you have time you may try to solve the case with different parameters. Changing the temperature difference is one
way of affecting the instability of the system. Decreasing the temperature differences the system eventually becomes
steady state and the convection rolls vanish altogether. Increasing the temperature difference may increase the number
of convection rolls and eventually the system becomes fully chaotic. Note that changing the temperature difference also
affects to the time scale of the wake.

CSC – IT Center for Science

Tutorial 27

Temperature distribution of a toy glacier

Directory: ToyGlacierTemperature
Solvers: HeatSolver
Tools: ElmerGUI,nglib
Dimensions: 2D, Steady-state
Author: Peter Råback, Thomas Zwinger

Introduction
The purpose of this simple tutorial is to be an introduction into Elmer for people dealing with computational glaciology.
This tutorial shows how to apply one equation and related boundary conditions to just one domain.

Problem description
Consider a 2D toy model of a glacier with length of 7000 m and thickness of about 1000 m. There is a slight declination
in the geometry which will make the glacier flow to the left. The left-hand-side is rounded to imitate a true glacier while
the right-hand-side is cut off.
We solve for the temperature distribution T of the glacier. A heat flux of q = 0.02 W/m2 is applied at the bottom of
the glacier while the surface stays at a fixed temperature of T0 = −10 C. The material properties of ice are used for the
heat conductivity κ(T ). The temperature distribution in the glacier may be solved from

 −κ∆T = 0 in Ω
T = T0 on ΓD (27.1)
κ ∂T = q on Γ

∂n N

Starting and meshing

Start ElmerGUI from command line or by clicking the icon in your desktop (or in the /bin directory of your installation).
Here we describe the essential steps in the ElmerGUI by writing out the clicking procedure. Indentation generally means
that the selections are done within the window chosen at the higher level.
The mesh is given in ElmerGrid format in file glacier_toy.in2d in the samples directory of ElmerGUI, load
this file.

Figure 27.1: The shape of the toy glacier to be studied

CSC – IT Center for Science

27. Temperature distribution of a toy glacier 171

File
Open -> glacier_toy.in2d

You should obtain a mesh consisting of just two triangular elements. The mesh is created by the netgen plugin and
in order to increase the mesh density the in-line parameters of netgen must be defined in ElmerGUI. Here we set the
maximum element size to 50.

Mesh
Configure...
nglib -> Max H: 50

The resulting mesh should consist of 3335 nodes and 6355 triangles as may be checked in the Model summary
window.

Command file definition

After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we choose
things related to the whole simulation such as file names, time stepping, constants etc. The simulation is carried out in
2-dimensional Cartesian coordinates and in steady-state. Only one steady-state iteration is needed as the case is linear.

Model
Setup
Simulation Type = Steady state
Steady state max. iter = 1

Choose Apply to close the window.

In the equation section we choose the relevant equations and parameters related to their solution. In this case we’ll
have one set only one equation – the heat equation.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use mouse selection
to assign them later. In this case we have just one body and one boundary and therefore its easier to assign the Equation
and Material to it directly.
For the linear system solvers we are happy to use the defaults. One may however, try out different preconditioners
(ILU1,. . . ) or direct Umfpack solver, for example.

Model
Equation
Add
Name = Heat Equation
Apply to bodies = 1
Heat Equation
Active = on
Apply
OK

The Material section includes all the material parameters. They are divided into generic parameters which are direct
properties of the material without making any assumptions on the physical model, such as the mass. Other properties
assume a physical law, such heat conductivity. We choose ice from the Material library which automatically sets for the
needed material properties.

Model
Material
Add
Material library
Water (frozen)
Apply to bodies = Body 1
Add
OK

This includes, for example, temperature dependent heat conductivity as may be seen under the HeatEquation page of the
material properties. MATC language is used here to define the functional form.
A Body Force represents the right-hand-side of a equation that in this case represents the heat source. In this case
there are no internal heat sources so we do not need one. Also no Initial Condition is required in steady state case.

CSC – IT Center for Science

27. Temperature distribution of a toy glacier 172

Figure 27.2: Defining boundary conditions in ElmerGUI session

We set three different kinds of boundary conditions. A fixed temperature, a fixed flux and natural boundary condition
(zero flux). As there are several boundaries we first define the different boundary types, and thereafter assign them using
the mouse. A screenshot of the case when setting the BCs is shown in figure 27.2.

Model
BoundaryCondition
Add
Name = Tsurface
Heat Equation
Temperature = -10.0
OK
Add
Name = Tflux
Heat Equation
Heat Flux = 0.02
OK
Add
Name = Tnatural
OK

Then we set the boundary properties

Model
Set boundary properties

Choose the correct boundary by clicking with the mouse and apply the condition for this boundary.

Boundary condition
Click top boundary -> choose Tsurface
Click bottom boundary -> choose Tflux
Click r.h.s. boundary -> choose Tnatural

Saving and solution

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined all the
information for ElmerGUI to write the command file. After writing it we may also visually inspect the command file.

CSC – IT Center for Science

27. Temperature distribution of a toy glacier 173

Figure 27.3: The convergence ElmerGUI

Sif
Generate
Edit -> look how your command file came out

Before we can execute the solver we should save the files in a directory. In saving the project all the necessary files
for restarting the case will be saved to the destination directory.

File
Save Project

After we have successfully saved the files we may start the solver

Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration. The heat conductivity of ice is
set to be dependent on temperature and this results to a non-linear iteration. The resulting output is shown in figure 27.3.
Note: if you face problems in the solution phase and need to edit the setting, always remember to save the project
before execution.

Results
To view the results we use Paraview,

Run
Start Paraview

Picture 27.4 shows the surface mesh coloured with temperature. Note that these results were carried out with the obsolete
VTK based tool within ElmerGUI, and therefore look different than in Paraview.
The maximum temperature obtained with the above choices is 0.11166 C. With a denser mesh the result is naturally
more accurate, but solving the problem takes more calculation time.
You may study the effect of mesh refinement by choosing a different value for the Max H parameter. under
Configure. After choosing Remesh and saving the mesh the solver may be recalled with the modified mesh.

Transient simulation
We use the steady-state simulation presented above as our starting point and solve a transient version of it. Initially the
glacier is assumed to be at -10 C and it is gradually heated from below.

CSC – IT Center for Science

27. Temperature distribution of a toy glacier 174

Figure 27.4: The temperature distribution of the toy glacier.

We use 2nd order bdf time stepping method is selected with 100 steps and with step size of 100 years - melting the
ice from below with such a small flux would take quite a few years. The mathematical expression followed by “$” is
evaluated when the command file is read. Here it is used to transform the time in years to those in one second.

Model
Setup
Simulation Type = Transient
Time Stepping Method = bdf
BDF Order = 2
Time Step Intervals = 100
Time Step Sizes = $ 3600 * 24 * 365.25 * 100
Gravity = ...

Initial conditions should be given to transient cases. In this case we choose a constant Temperature of -10 C. This is
consistent with the boundary condition at the top wall.

Model
Initial Condition
Name = Initial Guess
Heat Equation
Temperature = -10.0

CSC – IT Center for Science

27. Temperature distribution of a toy glacier 175

Figure 27.5: Temperature distribution after 1, 20, 50 and 100 time steps. The temperature scale is the same
that is used in the steady-state case. The maximum temperature at end should be about -3.7611 C.

CSC – IT Center for Science

Tutorial 28

Temperature and velocity distributions

of a toy glacier and bedrock

Directory: ToyGlacierTemperatureAndFlow
Solvers: HeatSolver,FlowSolver
Tools: ElmerGUI,nglib
Dimensions: 2D, Steady-state
Author: Thomas Zwinger, Peter Råback

Introduction
The purpose of this simple tutorial is to be an introduction into Elmer for people dealing with computational glaciology.
The tutorial is a continuation of the previous one with slightly more complex geometry. This tutorial shows how to apply
two different solvers to two different bodies.

Problem description
Consider a 2D toy model of a glacier sitting an a piece of bedrock. With the bedrock present it is possible to study more
accurately the temperature profiles.
Now we solve for the temperature distribution T for the combined system. A heat flux of q = 0.02 W/m2 is applied
at the bottom of the bedrock while the surface stays at a fixed temperature of T0 = −10 C. For ice the properties from
the database are assumed while for the bedrock density is assumed to be 2500 kg/m3 and heat conductivity 3 W/mK.
We also solve for the velocity distribution ~v of the glacier. The velocity field is solved from the Stokes equation
and is assumed to be affected only by the Gravity. As boundary conditions we apply a no-slip condition at the ice-rock
interface and symmetry condition at the right-hand-side of the glacier.

Starting and meshing

Start ElmerGUI from command line or by clicking the icon in your desktop (or in the /bin directory of you installation).
Here we describe the essential steps in the ElmerGUI by writing out the clicking procedure. Tabulation generally means
that the selections are done within the window chosen at the higher level.
The mesh is given in ElmerGrid format in file glacier_on_bedrock_toy.in2d in the samples directory of
ElmerGUI, load this file.

File
Open -> glacier_on_bedrock_toy.in2d

When netgen is ready with the meshing. You should obtain a mesh consisting of 4329 nodes and 8406 triangular
elements. Now the mesh density was predefined in the in2d file and therefore no command-line arguments are needed
to refine the mesh. If you got more elements check the value of Max H in the netgen parameter window.
If the mesh was successfully imported your window should look something in figure 28.1.

CSC – IT Center for Science

28. Temperature and velocity distributions of a toy glacier and bedrock 177

Figure 28.1: The finite element mesh in ElmerGUI

Command file definition

After we have the mesh we start to go through the Model menu from the top to bottom. Again we are happy with the
definitions in the Setup window.
In the Equation section we choose the relevant equations and parameters related to their solution. In this case we’ll
have two different sets of solvers (called as Equation in Elmer slang). The first consists of heat and flow solvers, while
the other includes just the heat solver. We’ll name them appropriately.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use mouse selection
to assign them later. In this case we know that the fluid body has the index 1 and the solid body has the index 2. Therefore
it is easy to assign the Equation and Material to the bodies directly.
Here we neglect the effect of convection to the temperature distribution. Therefore there is no coupling from velocity
field to energy equation. However, the temperature is affecting the viscosity of ice and therefore it needs to be solved
first. This is ensured by giving higher priority to the heat solver (default is zero). In order to obtain the Stokes equation
convection of momentum is omitted from the Navier-Stokes equations.
Here we are quite happy with the default solver settings of the individual equations of the linear systems. However,
the user may play around with different linear system settings to study their effect on convergence and computation time.
The equation for the ice

Model
Equation
Add
Name = Heat and Flow
Apply to Bodies = 1
Heat Equation
Active = on
Priority = 1
Convection = None
Navier-Stokes
Active = on
Convect = off
OK

and then for the solid

Model
Equation
Add
Name = Just Heat
Apply to Bodies = 2
Heat Equation
Active = on
Convection = None
OK

CSC – IT Center for Science

28. Temperature and velocity distributions of a toy glacier and bedrock 178

The Material section includes the material parameters. We choose ice from the Material library which automatically
sets for the needed material properties. For the bedrock we define the two parameters that are required.
Model
Material
Add
Material library
Water (frozen)
Apply to bodies = Body 1
Add
OK
Add
Name = Bedrock
Apply to bodies = Body 2
General
Density = 2500.0
Heat Equation
Heat Conductivity = 3.0
Add
OK
A Body Force represents the right-hand-side of a equations. For the heat equation there are no source terms. For the
Stokes equation we apply gravity which points to the negative y direction.
Model
BodyForce
Name = Gravity
Navier-Stokes
Force 2 = -9.81
Apply to Bodies = Body 1
Add
OK
We do not need any Initial Condition since the zero temperature (in Celsius) is a good initial guess for the heat
equation. For the Stokes equation a better solution could be used since if convergence problems would arise since the
non-Newtonian material laws do actually depend on the initial velocity.
We set four different kinds of boundary conditions. A fixed temperature, a fixed flux, no-slip condition and symmetry
condition. As there are several boundaries we first define the different boundary types, and thereafter assign them using
the mouse.
Model
BoundaryCondition
Add
Heat Equation
Temperature = -10.0
Name = Tsurface
OK
Add
Heat Equation
Heat Flux = 0.02
Name = Tflux
OK
Add
Navier-Stokes
No-slip Wall BC = on
Name = NoSlip
OK
Add
Navier-Stokes
Velocity 1 = 0.0
Name = Symmetry
OK

CSC – IT Center for Science

28. Temperature and velocity distributions of a toy glacier and bedrock 179

Figure 28.2: The convergence ElmerGUI

Then we set the boundary properties

Model
Set boundary properties

Choose the correct boundary by clicking with the mouse and apply the condition for this boundary.

Boundary condition
Click top boundary of ice -> choose Tsurface
Click the bare part of bedrock -> choose Tsurface
Click bottom boundary of bedrock -> choose Tflux
Click bottom boundary of ice -> choose NoSlip
Click r.h.s. boundary of ice -> choose Symmetry

Saving and solution

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined all the
information for ElmerGUI to write the command file. After writing it we may also visually inspect the command file.

Sif
Generate
Edit -> look how your command file came out

Before we can execute the solver we should save the files in a directory. In saving the project all the necessary files
for restarting the case will be saved to the destination directory.

File
Save Project

After we have successfully saved the files we may start the solver

Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration. The heat conductivity of ice is
set to be dependent on temperature and this results to a non-linear iteration. The resulting output is shown in figure 28.2.
Note: if you face problems in the solution phase and need to edit the setting, always remember to save the project
before execution.

CSC – IT Center for Science

28. Temperature and velocity distributions of a toy glacier and bedrock 180

Figure 28.3: Temperature (upper figure) and velocity (lower figure) distributions of the toy glacier sitting on
a bedrock.

Results
To view the results we use Paraview,

Run
Start Paraview

The resulting temperature and velocity distributions are shown in figure 28.3 (these are generated with the obsolete
VTKPost tool within ElmerGUI).
The maximum temperature obtained with the above choices is 12.955 C. For the velocity the maximum absolute
value is 6.3337 mm/s which is actually unreasonably high since it corresponds to yearly movement of around 200 km.
This just shows that the shape of the toy glacier under study is very unrealistic.

CSC – IT Center for Science

Tutorial 29

Generic scalar PDE on 3D angle domain

Directory: ModelPDE3D
Solvers: ModelPDE
Tools: ElmerGUI
Dimensions: 3D, Steady-state
Author: Peter Råback

Problem description
This tutorial demonstrates the use of the generic advection-diffusion-reaction equation through ElmerGUI. The solver
may be found in module ModelPDE. The purposes of the model pde and also this tutorial is to help those who want to
understand Elmer from a mathematical perspective to be able to carry out some of their own code development.
The problem is a simple 3d structure winkel.grd that can be characterized by a 2 × 2 × 2 topological grid
where entries (1, 1, 1), (1, 2, 1), (2, 1, 1) and (2, 1, 2) are meshed. This is the simplest Cartesian structure with full 3D
solution.
We can rather freely play with the parameters of the Model PDE. The equation is generic and the parameters are
assumed to be unit free. For detailed description of the problem see the description in Elmer Programmers Tutorial.
As a first suggestion, we will show how to make a simple case where the two extreme edges are set to zero using
Dirichlet boundary conditions, and constant unity source term is applied to the body. This is the simple Poisson equation
with constant coefficient.

Menu structures for Model PDE

The menu structures for the case are defined in model-pde.xml. If after starting you cannot find the menu structures
add the file to the edf directory of your installation, or append the menu structures within ElmerGUI.
The following material parameters may be defined
Diffusion Coefficient Real
Diffusion coefficient, µ.
Reaction Coefficient Real
Reaction coefficient, λ.
Time Derivative Coefficient Real
Multiplier of the time derivative, ρ.
Convection Coefficient Real
Multiplier of convection coefficient, κ.
Convection Velocity 1 Real
Convection velocity in direction x, ax .
Convection Velocity 2 Real
Convection velocity in direction y, ay .
Convection Velocity 3 Real
Convection velocity in direction z, az .
The following parameter defines the heat source f on the right-hand-side

CSC – IT Center for Science

29. Generic scalar PDE on 3D angle domain 182

Figure 29.1: The finite element mesh in ElmerGUI

Field Source Real

The menu structures defines the following parameters for boundary conditions:
Field Real
Dirichlet BC for the scalar field under study, u.
Field Flux Real
Neumann boundary condition for the field, q.
Robin Coefficient Real
External Field Real
Coefficient α and external field value g for Robin boundary condition.
In transient cases the user may also give an initial condition for u,
Field Real

Solution procedure
Start ElmerGUI from command line or by clicking the icon in your desktop. Here we describe the essential steps in
the ElmerGUI by writing out the clicking procedure. Indentation generally means that the selections are done within the
window chosen at the higher level.
The mesh is given in ElmerGrid format in file winkel.grd in the samples directory of ElmerGUI, load this file.

File
Open -> winkel.grd

You should obtain your mesh and may check in the Model summary window that it consists of 35 721 nodes and
32 000 trilinear elements. If the mesh was successfully imported your window should look something in figure 29.1.
The figure also shows the two extreme boundary patches that we intend to use Dirichlet conditions for.
After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we choose
things related to the whole simulation such as file names, time stepping, constants etc. The simulation is carried out in
2-dimensional Cartesian coordinates and in steady-state. Only one steady-state iteration is needed as the case is linear.

Model
Setup
Simulation Type = Steady state
Steady state max. iter = 1

CSC – IT Center for Science

29. Generic scalar PDE on 3D angle domain 183

Choose Apply to close the window.

In the equation section we choose the relevant equations and parameters related to their solution. In this case we’ll
have one set only one equation – the Model PDE.
When defining Equations and Materials it is possible to assign to the bodies immediately, or to use mouse selection
to assign them later. In this case we have just one body and one boundary and therefore its easier to assign the Equation
and Material to it directly.
For the linear system solvers we are happy to use the defaults. One may however, try out different preconditioners
(ILU1,. . . ) or direct Umfpack solver, for example.

Model
Equation
Add
Name = Model PDE
Apply to bodies = 1
Model PDE
Active = on
Add
OK

The Material section includes all the material parameters. If material parameter is not defined. It is assumed to be
zero. Here we just set the diffusivity to one.

Model
Material
Name = Ideal
Apply to bodies = 1
Model PDE
Diffusion Coefficient = 1.0
Add
OK

A Body Force represents the right-hand-side of a equation,

Model
Body Force
Name = Source
Apply to bodies = 1
Model PDE
Field Source = 1.0
Add
OK

No initial conditions are required in steady state case.

Finally, for the BCs first define them and then use the mouse to apply them to the correct boundary patches. Refer
to figure 29.1 and select the two surfaces as shown.

Model
BoundaryCondition
Name = Zero
Model PDE
Field = 0.0
Add
OK

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined all the
information for ElmerGUI to write the command file. After writing it we may also visually inspect the command file.

Sif
Generate
Edit -> look how your command file came out

Before we can execute the solver we should save the files in a directory. In saving the project all the necessary files
for restarting the case will be saved to the destination directory.

CSC – IT Center for Science

29. Generic scalar PDE on 3D angle domain 184

Figure 29.2: The field values of the structure as visualized with Paraview. Isosurfaces are defined for field
values 0.5, 1.0, 1.5, 1.8 and 1.9, respectively.

File
Save Project

After we have successfully saved the files we may start the solver

Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration. As the case is linear only one
iteration was required for the solution and the second one just is needed to check the convergence.
The norm of the results at convergence should be 1.4243820.
Note: if you face problems in the solution phase and need to edit the setting, always remember to save the project
before execution.
To view the results we may visualize them with Paraview.

Further possibilities
Here are some things you could try out, or are at least possible. The menus of the GUI are just elementary so more
advanced features may need that the keywords are added by hand to the command file. No coding is needed to implement
these features though.
• You can also solve the problem with an unstructured tetrahedral mesh using Gmsh format file winkel.msh.
• Play with the reaction, diffusion, convection coefficient, and also the advection velocity. As far as they remain
constant the equation should be solvable with one sweep.
• You could try to use Neumann or Robin BCs as well. Remember though that a steady state equation needs
definitions that uniquely define the solution.
• Make the problem time-dependent. Note that then you most likely need to define the coefficient for the time
derivative.

CSC – IT Center for Science

29. Generic scalar PDE on 3D angle domain 185

• When the advection increases in size the solution may eventually become oscillatory. To eliminate that some
bubbles or p-elements may be needed. You may play around with element settings, for example use Element
= p:2 or Element = n:1 b:1 etc.
• You could try to increase the number of elements either by using -relh parameter in ElmerGUI, or setting Mesh
Levels = 2 in Simulation section.
• You could try to use MATC to make the coefficients parameter dependent. Dependence on the solution itself
introduces non-linearities that might not be well handled by the fixed point iteration scheme. For more demanding
non-linearities Newton linearisation or other techniques may be needed.
• Also some periodicity could be introduced to this problem by letting the two extreme surface patches have a
dependence between them.
• You could introduce sort of contact conditions for the surface or bulk values of the problem by defining minimum
or maximum values for the field.

CSC – IT Center for Science

Tutorial 30

Electro-kinetics – 2D – Electro-osmotic
flow and advected species

Directory: Electrokinetics
Solvers: StatElecSolve, FlowSolve, AdvectionDiffusion, Electrokinetics
Tools: ElmerGUI
Dimensions: 2D, Transient
Author: Original author not known

Case definition
This tutorial is an example of setting up a simulation for (microfluidic) electro-osmotic flow advecting a passive scalar
quantity. Diffusion of the species is also included. The geometry of the system is a simple 2D microchannel with T
crossing. The flow is induced by the applied electric field and the electric double layer at the channel walls. The analyte
(species) is inserted into the system from the left hand side inlet.
More details on the electro-kinetic capabilities of Elmer are found on the Models Manual, chapter “Electrokinetics”,
authored by T. Zwinger.
This is an old case of analyte transport with Helmholtz-Smoluchowski Boundary Conditions. Not perhaps particu-
larly accurate as the advection-diffusion equation suffers from high numerical diffusion. Particularly for purely advective
transport the particle based methods might be more favourable.
The geometry for this tutorial looks like as shown in figure 31.1.

Figure 30.1: Geometry

CSC – IT Center for Science

30. Electro-kinetics – 2D – Electro-osmotic flow and advected species 187

ElmerGUI Equation Menu

The GUI definitions of Advection Diffusion solver utilized in this tutorial are located in the edf-extra folder and
thus need to be manually activated in the ElmerGUI.

Note that the extra definition needs to be loaded first, before any other steps in building any tutorial. Once the extra
definition has been loaded, and the ElmerGUI project has been saved, then the reference to the extra definition will be
stored in the project, and you won’t have to load the extra definition again.

File
Definitions
Append -> advection-diffusion.xml
Close

The advection-diffusion.xml definition file is, by default, located in Linux in:

$ELMER_HOME/share/ElmerGUI/edf-extra
and in Windows, located in:
C:/Program Files/Elmer 9.0-Release/share/ElmerGUI/edf-extra

We will also be using the Stat Elec Solver, and the Flow Solver which are two of the default, pre-loaded GUI defini-
tions, so they do not need to be manually activated. For reference, the electrostatics.xml and the navier-stokes.xml
definition files are, by default, located in Linux in:
$ELMER_HOME/share/ElmerGUI/edf
and in Windows, located in:
C:/Program Files/Elmer 9.0-Release/share/ElmerGUI/edf

The Electrokinetics solver is used for a boundary condition and will be called when needed. The Electrokinetics
solver is located in the $ELMER_HOME/bin folder.

Solution procedure
The mesh is given in ElmerGrid format in file Tcross.grd, load this file.

File
Open -> Tcross.grd

You should obtain your mesh and may check Model Summary... that it consists of 1200 surface elements. Your
mesh should look like as shown in figure 31.5

CSC – IT Center for Science

30. Electro-kinetics – 2D – Electro-osmotic flow and advected species 188

Figure 30.2: Mesh

After we have the mesh we start to go through the Model menu from the top to bottom. In the Setup we choose
things related to the whole simulation such as file names, time stepping, constants etc.
The simulation is carried out in 2-dimensional Cartesian coordinates. 2nd order bdf time stepping method is selected
with 120 steps and with step size of 1e-5 seconds. Coordinate scaling reduces the size of the mesh and produces a T-
shaped micro channel.

Model
Setup
Simulation Type = Transient
Steady state max. iter = 20
Timestepping method = BDF
BDF order = 2
Time Step Intervals = 120
Timestep sized = 1e-5
Output intervals = 2
Coordinate Scaling = 1e-5 1e-5 1e-5
Apply

In the equation section we choose the relevant equations and parameters related to their solution. In this case we’ll
have one set of equations (named “Equation”) which consists of the Stat Electric equation, the Navier-Stokes equation,
and the Advection Diffusion equation.

All three solvers are active in this equation set. Further definitions include, first, that the convection of the species is
switched on (besides diffusion). Then that for Navier-Stokes equations the convective term may be left out resulting in
laminar Stokes flow, and finally that the electric field is computed by the electrostatics rather than given by the user.

Note that the entry under Electrostatics of “Electric Field = Computed” may not exist in early versions of the
ElmerGUI menu definition in the electrostatics.xml file. When the Solver is run, an error may be reported, such as
“electrokinetics helmholtz smoluchowski: No external electric field defined for Equation 1”. If this occurs, either update
your version of Elmer, or generate the .sif file manually, then add this statement Electric Field = Computed in
the Equation block, and run the solver manually.

CSC – IT Center for Science

30. Electro-kinetics – 2D – Electro-osmotic flow and advected species 189

When defining Equations and Materials it is possible to assign to the bodies immediately, or to use mouse selection
to assign them later. In this case we have just one body and therefore its easier to assign the Equation and Material to it
directly.
The electrostatics equation is linear and thus no non-linear iterations are needed. The equation is solved using a fast
direct method UMFPack.
The system may include non-linear iterations of each equation and steady state iterations to obtain convergence of
the coupled system. It is often a good idea to keep the number of non-linear iterations in a coupled case low. Here we
select just a maximum of 30 non-linear iterations for the Navier-Stokes equation.
The advection-diffusion equation does not affect either the electrostatic field or the flow, thus it may be solved only
after a converged solution for the previous two equations is available. This is achieved with the After timestep
definition below. The advected quantity is called Concentration. The advection-diffusion solver uses bubble stabilization
method to avoid numerical problems associated with convection type equations.

Model
Equation
Name = Equation 1
Apply to Bodies = 1
Electrostatics
Active = True, checked
Electric Field = Computed
Edit Solver Settings
General
Before timestep
Solver specific options
Calculate Electric Field = True
Linear system
Method
Direct = Umfpack
Preconditioning = ILU1
Apply
Navier-Stokes
Active = True, checked
Options
Convect = unchecked, False
Edit Solver Settings
Linear system
Preconditioning = ILU1
Nonlinear system
Max. iterations = 30
Newton
After iterations = 10
Apply
Advection Diffusion
Active = True, checked
Concentration Units = Absolute Mass
Convection = Computed
Edit Solver Settings
General
After timestep = True
Stabilize = False
Bubbles = True
Linear system
Preconditioning = ILU2
Nonlinear system
Max. iterations = 1
Apply
Add
OK

CSC – IT Center for Science

30. Electro-kinetics – 2D – Electro-osmotic flow and advected species 190

The Material section includes all the material parameters. They are divided into generic parameters which are direct
properties of the material without making any assumptions on the physical model, such as the mass. Other properties
assume a physical law, such as conductivities and viscosity.

We use a material similar to water for this example.

Model
Material
Name = Material 1
Apply to Bodies = 1
General
Density = 1e3
Electrostatics
Relative Permittivity = 1.0
Navier-Stokes
Viscosity = 1e-03
Advection Diffusion
Concentration Diffusivity = 1e-10
Add
OK

A Body Force represents the right-hand-side of a equation. It is generally not a required field for a body. No body
force is included in this example.
Initial conditions should be given to transient cases, and probably are not needed for steady state solutions. No initial
condition is used for this example.
Only one boundary condition may be applied to each boundary and therefore all the different physical BCs for a
boundary should be grouped together.
Finally the boundary conditions are defined. The first BC is given for the channel walls. Here, tangential velocity
(velocity components 1 and 2) is computed by the Helmholtz-Smoluchowski slip velocity condition, which means that
the velocity is computed using the computed electric field and the electro-osmotic mobility as inputs.

Model
BoundaryCondition
Name = channel-walls
General
Free text input
EO Mobility = Real 5e-08
Navier-Stokes
Velocity 1 = Variable Pressure
Real Procedure "Electrokinetics" "helmholtz_smoluchowski1"
Velocity 2 = Variable Pressure
Real Procedure "Electrokinetics" "helmholtz_smoluchowski2"
Add
OK

The next BC is the inlet condition. We give a potential of 100 Volts and define that there is no flow in y-direction.
The analyte concentration at the inlet is defined as a function of time using table format. The concentration is 1.0 up
until time instant 3 · 10−5 , is zero after 4 · 10−5 s and decreases linearly between these two time instants.

Model
BoundaryCondition
Name = inlet-el-A
Electrostatics
Potential = 100
Navier-Stokes
Velocity 2 = 0.0
Advection Diffusion
Dirichlet Conditions
Concentration
Variable Time

CSC – IT Center for Science

30. Electro-kinetics – 2D – Electro-osmotic flow and advected species 191

Real
0.0 1.0
3.0e-5 1.0
4.0e-5 0.0
0.5 0.0
End
Add
OK

The final two boundary conditions are for the outlets. Different potentials for these are defined as well as a condition
for velocity component.

Model
BoundaryCondition
Name = outlet-el-B
Electrostatics
Potential = 30
Navier-Stokes
Velocity 1 = 0.0
Add
OK

Model
BoundaryCondition
Name = outlet-el-C
Electrostatics
Potential = 0
Navier-Stokes
Velocity 1 = 0.0
Add
OK

The conditions may also be assigned to boundaries in the Boundary condition menu, or by clicking on each boundary
with the mouse. Here we use the latter approach as that spares us of the need to know the indexes of each boundary.

Model
Set boundary properties
Choose Inlet -> set boundary condition inlet-el-A
Choose Top Outlet -> set boundary condition outlet-el-B
Choose Bottom Outlet -> set boundary condition outlet-el-C
Choose Walls -> set boundary condition channel-walls
OK

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined all the
information for ElmerGUI to write the command file. After writing it we may also visually inspect the command file.

Sif
Generate
Edit -> look how your command file came out

Before we can execute the solver we should save the files in a directory. The ElmerGUI project includes all the files
needed to restart the case.

File
Save Project

After we have successfully saved the files we may start the solver.

Run
Start solver

CSC – IT Center for Science

30. Electro-kinetics – 2D – Electro-osmotic flow and advected species 192

A convergence view automatically pops up showing relative changes of each iteration.

When there are some results to view we may start the postprocessor also.

Run
Start ParaView

Results
Due to the number of the time steps the simulation may take around 15 seconds.

You may inspect the results with Paraview or with ElmerVTK.

In Figure 30.3 the obtained concentration distribution is presented.

Figure 30.3: Concentration at time: 0, 2, 15, and 45 ticks

CSC – IT Center for Science

Tutorial 31

Asymmetrical Conductor with a Hole

TEAM Workshop Problem 7

Directory: TEAM7
Solvers: CoilSolver, WhitneyAVSolver,MagnetoDynamicsCalcFields
Tools: ElmerGUI
Dimensions: 3D, Transient
Author: Jonathan Velasco

Case definition
The Testing Electromagnetic Analysis Methods (TEAM) Problem 7 is a validation case consisting of a thick aluminium
plate with a hole, which is placed eccentrically in a non-uniform magnetic field. The field is produced by a sinusoidal
current flowing through a coil that hovers over the plate as shown in Fig. 31.1 . Due to asymmetry a simplification of
the model becomes unavailable and a full 3D model is needed. The goal of the model is to demonstrate the basic setup
of a 3D eddy current problem.

Figure 31.1: TEAM Problem 7 Geometry Description

The full description of the given problem can be found in the COMPUMAG website under:
TEAM Problem 7
Mesh and tailored .sif file can be found in our Github repository under the TEAM7 directory:
Elmer Electromagnetics Tutorials

CSC – IT Center for Science

31. Asymmetrical Conductor with a Hole
TEAM Workshop Problem 7 194

Meshing: From Gmsh to ElmerGrid

The TEAM7 Mesh can be found in our Github repository under the TEAM7 directory:
Elmer Electromagnetics Tutorials

However, you may also want to import your own mesh from another software. Under the same GitHub repository you’ll
find the .geo file created in Gmsh (gmsh/TEAM7.geo). Note, this mesh is generated with Gmsh but ElmerGrid parses
several formats into ElmerSolver format. (e.g., .ansys, .inp, .fil, .mphtxt ...).

To mesh .geo file please run the following cmd/terminal command

gmsh TEAM7.geo -3

where the −3 implies 3D meshing. This will save the mesh as TEAM7.msh
If further mesh refinement is needed:

gmsh TEAM7.msh -refine

The command performs uniform mesh refinement. You can also perform barycentric refinement with command: −barycentric_refine.

Once the mesh is saved you can parse it into ElmerSolver by using the following command

ElmerGrid 14 2 TEAM7.msh

where 14 denotes the input mesh format (Gmsh) and 2 the output mesh format (ElmerSolver).
The mesh is ready to be used by ElmerSolver either using command line or ElmerGUI.

Necessary Solver Modules

The TEAM7 magnetodynamics problem can be solved using two modules: the Coil Current Solver and the WhitneyAV-
Solver. The field solution can be retrieved using the MagnetoDynamicsCalcFields module.

The Coil Current Solver solves for the irrotational electric field associated with the gradient of the electric scalar
potential. Hence, the current density is calculated and may be used as an external source (Body Force) when solving the
magnetodynamics problem. The magnetodynamics problem is then solved throughout the whole domain (coil, plate and
air) using the WhitneyAVSolver and the field solution such as the magnetic flux density, magnetic field strength among
other variables can be obtained for postprocessing visualization using the MagnetoDynamicsCalcFields module.

ElmerGUI Project
Start ElmerGUI from command line or by clicking the icon in your desktop. Here we describe the essential steps in the
ElmerGUI by writing out the click-by-click procedure. Indentation generally means that the selections are done within
the window chosen at the higher level.
Before setting up the model, let us configure ElmerGUI’s menu and necessary settings related to this particular
model by using the top ribbon in the GUI (Fig. 31.2)

Figure 31.2: Top Menu

For simplicity the project directory is presented under ElmerGUI_TEAM7. The model is solved using modules
Coil Current Solver, WhitneyAVSolver and MagnetoDynamicsCalcFields, which are found under the coilsolver.xml and
magnetodynamics.xml menus. As of the current version of the code, these menus are not added automatically and need
to be added manually. Let us setup the project directory, load the mesh (TEAM7) in ElmerSolver format (.mesh.*) and
load the necessary menus. See Fig

File
New project ...
Project directory

CSC – IT Center for Science

31. Asymmetrical Conductor with a Hole
TEAM Workshop Problem 7 195

Select project dir -> ElmerGUI_TEAM7

Geometry input
Elmer mesh -> TEAM7
Extra EDFs to be added
Add -> coilsolver.xml
Add -> magnetodynamics.xml
OK

Figure 31.3: New Project Window

Adding additional menus can be avoided by permanently moving the .xml files from $ELMER_HOME/install/share/ElmerGUI/edf
to $ELMER_HOME/install/share/ElmerGUI/edf

Model Preferences
The TEAM Problem 7 is solved by imposing a transient sinusoidal current that generates a magnetic field. Due to the
time-dependency nature of the problem let us setup the model time preferences by clicking on ElmerGUI on the top
ribbon

ElmerGUI
Preferences
Simulation Type = Transient
Timestep interval = 16
Timestep sizes = 0.0025
Close

The timestep interval and size used are meant to accommodate two periods at a frequency of 50Hz.

CSC – IT Center for Science

31. Asymmetrical Conductor with a Hole
TEAM Workshop Problem 7 196

Figure 31.4: Preferences Setup Window

Solution procedure
At this point the mesh (Fig. 31.5) has been loaded, the model setup has been configured and the necessary menus have
been added to the GUI. We may now proceed to set up the physics of the TEAM Problem 7.

Figure 31.5: TEAM 7 Mesh.

Let us configure the equations. Two sets of equations are needed in this problem. The first set of equations will apply
to the Coil domain alone whereas the second set applies to the Plate and Air domain. The reasoning behind it lies in the

CSC – IT Center for Science

31. Asymmetrical Conductor with a Hole
TEAM Workshop Problem 7 197

fact that we will solve the current direction before the simulation begins, using it as a source in the magnetodynamics
problem.
The coil equation are setup in the following manner

Model
Equation -> Add
Name = Equation Coil
Apply to Bodies = 1

CoilSolver
Active = on
Edit Solver Settings
Solver specific options
Coil Closed=on
Desired Coil Current = -2742
Normalize Coil Current=on
Coil Normal (3-vector) = 0 0 1
Narrow Interface=on
Fix Input Current Density=on
General
Execute solver -> Before Simulation
Apply

MgDyn
Active = on
Edit Solver Settings
Solver specific options
Use Elemental CoilCurrent=on
Apply

MgDynPost
Active=on
Edit Solver Settings
Solver specific options
Specify fields to compute
Calculate Current Density=on
What kind of fields to compute
Discontinuous bodies=on
General
Execute solver -> Before Saving
Apply
Add
OK

CSC – IT Center for Science

31. Asymmetrical Conductor with a Hole
TEAM Workshop Problem 7 198

The configuration of the menus will appear also in any new equation added. For this reason, we only need activate
the solvers needed and apply them to the respective bodies

Model
Equation -> Add
Name = Equation Plate and Air
Apply to Bodies = 1 2

MgDyn
Active = on

MgDynPost
Active=on

Add
OK

Elmer contains an extensive material library however, for this exercise we will define some user defined material
properties. In this example we only need two materials: Air and Aluminium which are applied on. Due to the fact that
the current is known a priori, the current density will be used as an external current source enabling the coil to be part of
the non-conducting domain (Air).

Model
Material -> Add
Name = Air
Apply to Bodies = 1 3
MgDyn
Relative Permeability=1
Update
OK

Model
Material -> Add
Name = Aluminium
Apply to Bodies = 2
MgDyn
Relative Permeability=1
Electric Conductivity =3.526e7
Update
OK

A Body Force represents the right-hand-side of a equation. In this case we want to re-use the computed current den-
sity (by the Coil Solver) and use it in our MagnetoDynamics problem. By activating the Use Elemental CoilCurrent=on
checkbox in the MgDyn solver menu, this action is performed. However, since the current is sinusoidal, we will set up
a feature that is found within the Body Force menu.

Model
Body Force
Name = External CoilCurrent Source
Apply to Bodies = 1
Current Density Multiplier= Variable "time"; Real LUA "cos(2*math.pi*50*tx[0])"
Add
OK

Note that the frequency is 50Hz. Hence cos(2 π f t), where f is the excitation frequency.

CSC – IT Center for Science

31. Asymmetrical Conductor with a Hole
TEAM Workshop Problem 7 199

Lastly, at infinity the magnetic vector potential is set to zero

Model
BoundaryCondition -> Add
Name = Infinity
Apply to Boundaries = 6
MgDyn
AV {e} = 0.0

Add
Ok

For the execution ElmerSolver needs the mesh files and the command file. We have now basically defined all the
information for ElmerGUI to write the command file. After writing it we may also visually inspect the command file.

Sif
Generate
Edit -> Double check file has been generated accordingly to specs

Before we can execute the solver save the the project.

File
Save Project

After we have successfully saved the files we may start the solver.

Run
Start solver

A convergence view automatically pops up showing relative changes of each iteration.

When there are some results to view we may start the postprocessor also.

Run
Start ParaView

CSC – IT Center for Science

31. Asymmetrical Conductor with a Hole
TEAM Workshop Problem 7 200

Results
Note that this model’s purpose is two fold. First, our goal is to setup a basic 3D eddy current problem. Second, we use
this model for validation purposes.
You may inspect the results with Paraview or with ElmerVTK.
The following image have been created using Paraview.

Figure 31.6: Postprocessing: Color map including induced currents on plate (Left), Line Plot to validate
Elmer’s results against TEAM7’s

CSC – IT Center for Science

Index

advection-diffusion.xml, 187 MagnetoDynamics2DHarmonic, 102

AdvectionDiffusion, 186 MagnetoDynamicsCalcFields, 193
MATC, 149
Boussinesq, 165 MeshSolve, 158
MgDyn, 92
capacitance, 74, 85 MgDynPost, 92
clipping with ElmerVTK, 24 MgHarm, 92
clipping with Paraview, 27 ModelPDE, 181
CoilSolver, 193
Convection Coefficient, 181 Navier-Stokes Solver, 124
Convection Velocity 1, 181 navier-stokes.xml, 124, 142, 187
Convection Velocity 2, 181 netgen, 7, 79
Convection Velocity 3, 181 nglib, 170, 176
nonlinearelasticity.xml, 59
Diffusion Coefficient, 181
Discontinuous Bodies, 94 OpenCascade, 7

ElasticSolve, 59, 158 Paraview, 25

electro-osmotic mobility, 190 passive elements, 45
Electrokinetics, 186 Polyline Coordinates, 126
Electrokinetics solver, 187 Polyline Divisions, 126
electrostatics.xml, 187
ElmerGrid, 90, 122 Rayleigh-Benard, 165
ElmerGUI, 7, 29, 34, 40, 45, 51, 55, 59, 64, 69, 74, 79, 85, Reaction Coefficient, 181
90, 97, 102, 107, 112, 118, 122, 136, 141, 148, Result Output Solver, 113
152, 158, 165, 170, 176, 181, 186, 193 resultoutput.xml, 113
ElmerVTK, 20 Reynolds Number, 148
External Field, 182 Robin Coefficient, 182

Field, 182 save a picture with ElmerVTK, 23

Field Flux, 182 save a picture with Paraview, 26
Field Source, 182 SaveLine Solver, 124
Flow Solver, 187 saveline.xml, 124
FlowSolve, 118, 122, 136, 141, 148, 152, 158, 165, 186 SmitcSolver, 64, 69
FlowSolver, 142, 176 Smoluchowski, 190
Stat Elec Solver, 187
Gmsh, 97, 102 StatElecSolve, 186
StatElecSolver, 74, 79, 85
Heat Solver, 9, 46
StressSolve, 51, 55
heatequation.xml, 9, 46, 142
HeatSolve, 29, 34, 40, 45, 141, 152, 165 Time Derivative Coefficient, 181
HeatSolver, 7, 142, 170, 176
Helmholtz, 107 VectorHelmholtz, 112
Helmholtz Solver, 108 vectorhelmholtz.xml, 113
Helmholtz-Smoluchowski, 186 von Karman, 148
helmholtz.xml, 108 vortex shedding, 148
KESolver, 136 WhitneyAVHarmonicSolver, 90
WhitneyAVSolver, 193
magnetodynamics.xml, 90
MagnetoDynamics2D, 97

CSC – IT Center for Science