RADIOSS 2017 Tutorials and Examples

HyperWorks 2017
RADIOSS Tutorials and Examples
Learn more at altairhyperworks.com

Copyrights, Trademarks, Patents and Third Party Licenses
Intellectual Property Rights Notice:
Copyrights, Trademarks, Trade Secrets, Patents & Third Party Software Licenses
Altair® HyperWorks® v.2017
A Platform for Innovation®
Copyright© 1986-2016 Altair Engineering Inc. All Rights Reserved.
Special Notice: Pre-release versions of Altair software are provided ‘as is’, without warranty of any
kind. Usage is strictly limited to non-production purposes.
HyperWorks® 2017 Products

HyperMesh® 1990-2016; HyperCrash® 2001-2016; OptiStruct® 1996-2016; RADIOSS® 1986-2016;
HyperView® 1999-2016; HyperView Player® 2001-2016; HyperMath® 2007-2016; HyperStudy®
1999-2016; HyperGraph® 1995-2016; MotionView® 1993-2016; MotionSolve® 2002-2016;
HyperForm® 1998-2016; HyperXtrude® 1999- 2016; Process Manager™ 2003-2016; Templex™
1990-2016; TextView™ 1996-2016; MediaView™ 1999-2016; TableView™ 2013-2016; BatchMesher™
2003-2016; HyperWeld® 2009-2016; HyperMold® 2009-2016; Manufacturing Solutions™ 2005-2016;
solidThinking Inspire® 2017 2009-2016; solidThinking Evolve® 2017 1993-2016; Durability Director™
2009-2016; Suspension Director™ 2009-2016; AcuSolve® 1997-2016; AcuConsole® 2006-2016;
SimLab® 2004-2016; Virtual Wind Tunnel™ 2012-2016; FEKO® (©1999-2014 Altair Development S.A.
(Pty) Ltd.; ©2014-2016 Altair Engineering, Inc.); ConnectMe™ 2014-2016.
Additional Altair Products:

Multiscale Designer™ 2011-2016; Flux™ v.12.2 1983-2016; InCa3D v.3.1 1996-2016; CDE v.2
2012-2016; Got-It v.3 2002-2016; WinProp v.14 2000-2016
Altair Packaged Solution Offerings (PSOs) Copyright© 2008-2016

Automated Reporting Director™ 2008-2016; GeoMechanics Director 2011-2016; Impact Simulation
Director™ 2010-2016; Model Mesher Director™ 2010-2016; Model Verification Director™ 2013-2016;
NVH Director™ 2010-2016; Squeak and Rattle Director™ 2012-2016; Virtual Gauge Director™
2012-2016; Weight Analytics™ 2013-2016; Weld Certification Director™ 2014-2016
Altair Simulation Cloud Suite:

Simulation Manager™ 2003-2016; Compute Manager™ 2003-2016; Display Manager™ 2003–2016;
and Process Manager™ 2003-2016.
Altair PBS Works™:

Compute Manager™ 2012-2016; Display Manager™ 2013-2016; PBS™ 1994-2016; PBS
Pro™ 1994-2016; PBS Professional® 1994-2016; PBS Application Services™ 2008-2016; PBS
Analytics™ 2008-2016; and PBS Desktop™ 2008-2012; e-Compute™ 2000-2010; OpenPBS®
1994-2003 and Personal PBS® 2008-2012.
Software products of solidThinking, Inc., a wholly owned subsidiary of Altair Engineering:

solidThinking Inspire® 2017 2009-2016; solidThinking Evolve® 2017 1993-2016; solidThinking
Compose™ 2017 2007-2016; solidThinking Activate™ 2017 1989-2016; solidThinking Embed™ 2017
1989-2016; solidThinking Embed™ SE 2017 1989-2016; Click2Extrude™ Metal 2017 1996-2016;
Click2Extrude™ Polymer 2017 1996-2016; Click2Cast® 14.0 2011-2016; Click2Form™ 2017
1998-2016; Envision™ 2013-2016.
ALTAIR ENGINEERING INC. Proprietary and Confidential. Contains Trade Secret

Information.
Not for use or disclosure outside of Altair and its licensed clients. Information contained in Altair
software shall not be decompiled, disassembled, “unlocked”, reverse translated, reverse engineered, or
publicly displayed or publicly performed in any manner. Usage of the software is only as explicitly
permitted in the end user software license agreement. Copyright notice does not imply publication.
Third party software licenses

AcuConsole contains material licensed from Intelligent Light (www.ilight.com) and used by permission.
Software Security Measures:

Altair Engineering Inc. and its subsidiaries and affiliates reserve the right to embed software security
mechanisms in the Software for the purpose of detecting the installation and/or use of illegal copies of
the Software. The Software may collect and transmit non-proprietary data about those illegal
copies. Data collected will not include any customer data created by or used in connection with the
Software and will not be provided to any third party, except as may be required by law or legal process
or to enforce our rights with respect to the use of any illegal copies of the Software. By using the
Software, each user consents to such detection and collection of data, as well as its transmission and
use if an illegal copy of the Software is detected. No steps may be taken to avoid or detect the
purpose of any such security mechanisms.
Technical Support
Altair provides comprehensive software support via web FAQs, tutorials, training classes, telephone,
and e-mail.
Altair Support on the World Wide Web
The Altair web site is a valuable online companion to Altair software. Visit www.altairhyperworks.com
for tips and tricks, training course schedules, training/tutorial videos, and other useful information.
Altair Training Classes
Altair training courses provide a hands-on introduction to our products, focusing on overall
functionality. Courses are conducted at our main and regional offices or at your facility. If you are
interested in training at your facility, please contact your account manager for more details. If you do
not know who your account manager is, please send an e-mail to training@altair.com and your account
manager will contact you.
Telephone and e-mail
When contacting Altair support, please specify the product and version number you are using along
with a detailed description of the problem. Many times, it is very beneficial for the support engineer to
know what type of workstation, operating system, RAM, and graphics board you have, so please have
that information ready. If you send an e-mail, please specify the workstation type, operating system,
RAM, and graphics board information in the e-mail.
To contact an Altair support representative, reference the following table or the information available
on the HyperWorks website:
http://www.altairhyperworks.com/ClientCenterHWSupportProduct.aspx
Location Telephone e-mail
Australia 64.9.413.7981 anzsupport@altair.com
Brazil 55.11.3884.0414 br_support@altair.com
Canada 416.447.6463 support@altairengineering.ca
China 86.400.619.6186 support@altair.com.cn
France 33.1.4133.0992 francesupport@altair.com
Germany 49.7031.6208.22 hwsupport@altair.de
India 91.80.6629.4500 support@india.altair.com

1.800.425.0234 (toll free)
Italy 39.800.905.595 support@altairengineering.it
Japan 81.3.5396.2881 support@altairjp.co.jp
Korea 82.70.4050.9200 support@altair.co.kr
Mexico 55.56.58.68.08 mx-support@altair.com

New Zealand 64.9.413.7981 anzsupport@altair.com
North America 248.614.2425 hwsupport@altair.com
Scandinavia 46.46.460.2828 support@altair.se
Spain 34.910.810.080 support-spain@altair.com
South Africa 27.21.8311500 support@altair.co.za
United Kingdom 01926.468.600 support@uk.altair.com
For questions or comments about this help system, send an email to hwsupport@altair.com.
In addition, the following countries have resellers for Altair Engineering: Colombia, Czech Republic,
Ecuador, Israel, Russia, Netherlands, Turkey, Poland, Singapore, Vietnam, Indonesia
Official offices with resellers: Canada, China, France, Germany, India, Malaysia, Italy, Japan, Korea,
Spain, Taiwan, United Kingdom, USA
See www.altair.com for complete contact information.
Table of Contents
Tutorials and Examples ......................................................................................................................................... 3
Tutorials............................................................................................................................................................ 3
Accessing the Model Files.............................................................................................................................. 4
Introductory Tutorials ................................................................................................................................... 5
Examples....................................................................................................................................................... 272
List of Examples......................................................................................................................................... 273
Example 1 - Twisted Beam ......................................................................................................................... 281
Example 2 - Snap-thru Roof ....................................................................................................................... 287
Example 3 - S-beam Crash ......................................................................................................................... 301
Example 4 - Airbag .................................................................................................................................... 317
Example 5 - Beam Frame ........................................................................................................................... 328
Example 6 - Fuel Tank ................................................................................................................................ 335
Example 7 - Pendulums ............................................................................................................................. 353
Example 8 - Hopkinson Bar ........................................................................................................................ 371
Example 9 - Billiards (pool) ........................................................................................................................ 390
Example 10 - Bending ................................................................................................................................ 417
Example 11 - Tensile Test .......................................................................................................................... 426
Example 12 - Jumping Bicycle .................................................................................................................... 465
Example 13 - Shock Tube ........................................................................................................................... 486
Example 14 - Truck with Flexible Body ....................................................................................................... 506
Example 15 - Gears.................................................................................................................................... 531
Example 16 - Dummy Positioning .............................................................................................................. 540
Example 17 - Box Beam ............................................................................................................................. 573
Example 18 - Square Plate ......................................................................................................................... 640
Example 19 - Wave Propagation ................................................................................................................ 671
Example 20 - Cube..................................................................................................................................... 685
Example 21 - Cam...................................................................................................................................... 691
Example 22 - Ditching using SPH and ALE (Mono-Domain and Multi-Domain)............................................ 709
1 RADIOSS 2017 Tutorials Altair Engineering
Proprietary Information of Altair Engineering

Example 23 - Brake .................................................................................................................................... 732
Example 24 - Laminating ........................................................................................................................... 741
Example 25 - Spring-back .......................................................................................................................... 752
Example 26 - Ruptured Plate ..................................................................................................................... 771
Example 27 - Football (Soccer) Shots ......................................................................................................... 782
Example 37 - Analytical Beam .................................................................................................................... 789
Example 39 - Biomedical Valve .................................................................................................................. 801
Example 42 - Rubber Ring: Crush and Slide ................................................................................................ 810
Example 43 - Perfect Gas Modeling with Polynomial EOS .......................................................................... 820
Example 44 - Blow Molding with AMS ....................................................................................................... 839
Example 45 - Multi-Domain ....................................................................................................................... 846
Example 46 - TNT Cylinder Expansion Test ................................................................................................. 855
Example 47 - Concrete Validation .............................................................................................................. 876
Example 48 - Solid Spotweld...................................................................................................................... 897
Example 49 - Bird Strike on Windshield ..................................................................................................... 905
Example 50 - INIVOL and Fluid Structure Interaction (Drop Container) ...................................................... 914
Example 51 - Optimization in RADIOSS for B-Pillar (Thickness optimization) .............................................. 921
Example 52 - Creep and Stress Relaxation ................................................................................................. 929
Example 53 - Thermal Analysis .................................................................................................................. 935
Altair Engineering RADIOSS 2017 Tutorials 2

Tutorials and Examples
Tutorials
Below is a list of the RADIOSS tutorials available.

Introductory Tutorials
RD-0010: Running RADIOSS from HyperMesh
RD-0020: Running RADIOSS at the Command Line
Large Displacement Finite Element Analysis - HyperCrash
RD-3000: Tensile Test Setup using HyperCrash
RD-3030: Buckling of a Tube using Half Tube Mesh
RD-3050: Simplified Car Pole Impact using HyperCrash
RD-3060: Three Point Bending with HyperCrash
RD-3150: Seat Model with Dummy using HyperCrash
RD-3160: Setting up Multi-Domain Analysis using HyperCrash
HF-5000: Using Results Mapper in HyperCrash
Large Displacement Finite Element Analysis - HyperMesh
RD-3500: Tensile Test Setup using HyperMesh
RD-3510: Cantilever Beam with Bolt Pretension
RD-3520: Pre-Processing for Pipes Impact using RADIOSS
RD-3540: Front Impact Bumper Model using HyperMesh
RD-3550: Simplified Car Pole Impact
RD-3560: Bottle Drop
RD-3580: Boat Ditching (with and without Boundary)
RD-3590: Fluid Flow through a Rubber Clapper Valve
RD-3595: Three Point Bending with HyperMesh
RD-3597: Cell Phone Drop Test using HyperMesh
RD-3599: Gasket with HyperMesh

Accessing the Model Files
Required model files of the models you build in the tutorials are available online.
1. To access model files, visit Altair Connect or the Altair Client Center.
A user ID and password are required to access the model files. Follow the instructions at the
website to obtain login credentials.
2. Select the required file package and download it onto your system.
Note that the files may require unzipping before proceeding with the tutorials. When extracting
zipped files, preserve any directory structure included in the file package.

Introductory Tutorials
RD-0010: Running RADIOSS from HyperMesh

This tutorial demonstrates how to launch a RADIOSS job from within HyperMesh. A HyperMesh
database containing a fully defined RADIOSS finite element model is retrieved and a RADIOSS job is
launched from the RADIOSS panel in HyperMesh.
Exercise
Step 1: Load the User Profile

1. Launch HyperMesh. The User Profiles dialog appears upon start-up by default.
2. If the User Profiles dialog is not visible, select Preferences from the toolbar and choose User
Profiles.
3. Under Application:, select RADIOSS.
4. Click OK. This loads the appropriate User Profile. It includes the appropriate template, macro
menu, and import reader. It simplifies the menu systems to give access to only the functionality
of HyperMesh that is necessary.
Step 2: Retrieve the HyperMesh database

1. From the File menu on the toolbar, select Open. An Open file browser window opens.
2. Select the Radioss_Sample_Run.hm file you saved to your working directory from the
radioss.zip file. Refer to Accessing the Model Files.
3. Click Open. The Radioss_Sample_Run.hm database is loaded into the current HyperMesh
session, replacing any existing data.
Step 3: Launch the RADIOSS job

1. From the Analysis page, select the RADIOSS panel.
2. Click save as. A Save file browser window opens.
3. Select the directory where you would like to write the model file and enter the file name,
Radioss_Sample_Run.rad, in the File name: field. The .rad file name extension is the
suggested extension for RADIOSS input decks.
4. Click Save. The name and location of the Radioss_Sample_Run.rad file now displays in the
input file: field.
5. Set the memory toggle, located in the center of the panel, to memory default.
6. Set the run options toggle, located on the left side of the panel, to analysis.
7. Set the export options: toggle, underneath the run options switch, to all.

8. Click RADIOSS.
This exports the input file and launches the job. If the job is successful, new results files can be
seen in the directory where the model file was written. The Radioss_Sample_Run.out file is a
good place to look for error messages that will help to debug the input deck if any errors are
present.
The default files written to your directory are:
Radioss_Sample_Run.html HTML report of the analysis, giving a summary of the
problem formulation and the analysis results.
Radioss_Sample_Run.out ASCII output file containing specific information on the
file set up, the set up of your optimization problem,
estimate for the amount of RAM and disk space
required for the run, information for each optimization
iteration, and compute time information. Review this
file for warnings and errors.
Radioss_Sample_Run.res HyperMesh binary results file.
Radioss_Sample_Run.stat Summary of analysis process, providing CPU
information for each step during analysis process.
Radioss_Sample_Run.h3d HyperView binary result file.
Step 4: Post-process the RADIOSS job
While still in HyperMesh, you can launch HyperView after the job has finished from the RADIOSS
panel by clicking HyperView. HyperView will open and automatically load the H3D file from the
RADIOSS job for post-processing.

RD-0020: Running RADIOSS at the Command Line
The tutorial Running RADIOSS from HyperMesh demonstrates how RADIOSS could be launched
from within HyperMesh. RADIOSS also can be run at the command line (UNIX or MSDOS). This
tutorial assumes you already have the running file, Radioss_Sample_Run.rad, in either your UNIX
or MSDOS directory. This tutorial also assumes you know the location of the solver script.
In this tutorial, $HWSDIR describes the directory containing the RADIOSS executable. On UNIX
machines, the script is normally located in the HyperWorks installation directory under
<install_directory>/scripts/. On Windows, it is normally located in the HyperWorks installation
directory under <install_directory>/hwsolvers/scripts/.
Running RADIOSS from the Command Line (UNIX or MSDOS).
To run RADIOSS from the command prompt, enter:
$HWSDIR/<solver_name> Radioss_Sample_Run.rad
To check the current version of RADIOSS at the command prompt, enter:

$HWSDIR/<solver_name> -version
To execute a check run to validate your input deck and determine how much RAM and
disk space is necessary for the run, at the command prompt, enter:
$HWSDIR/<solver_name> Radioss_Sample_Run.rad -check
Information regarding memory requirements is written to the file Radioss_Sample_Run.out.
Refer to the Running RADIOSS section of the RADIOSS User's Guide for more detailed information.

Large Displacement Finite Element Analysis
HyperCrash
RD-3000: Tensile Test Setup using HyperCrash

This tutorial demonstrates how to simulate a uniaxial tensile test using a quarter size mesh with
symmetric boundary conditions.
The model is reduced to one-quarter of the total mesh with symmetric boundary conditions to
simulate the presence of the rest of the part.
Model Description
 UNITS: Length (mm), Time (ms), Mass (kg), Force (kN) and Stress (GPa)
 Simulation time Rootname_0001.rad [0 – 10.]
 Boundary Conditions:
o The 3 upper right nodes (TX, RY, and RZ)
o A symmetry boundary condition on all bottom nodes (TY, RX, and RZ)
 At the left side is applied a constant velocity = 1 mm/ms on -X direction.
 Tensile test object dimensions = 11 x 100 with a uniform thickness = 1.7 mm
Johnson-Cook Elastic Plastic Material /MAT/PLAS_JOHNS (Aluminum 6063 T7)
[Rho_I] Initial density = 2.7e-6 Kg/mm3

[E] Young’s modulus = 60.4 GPa
[nu] Poisson’s ratio = 0.33
[a] Yield stress = 0.09026 GPa
[b] Hardening parameter = 0.22313 GPa
[n] Hardening exponent = 0.374618
[EPS_max] Failure plastic strain = 0.75
[SIG_max] Maximum stress = 0.175 GPa
Input file for this tutorial: TENSILE_0000.rad

Exercise
Step 1: Import the mesh

1. Open HyperCrash and set the User profile: to RADIOSS V14 and the Unit system: to kN
mm ms kg.
1. Click Run.
2. From the menu bar, select File > Import > RADIOSS.
3. In the Select RADIOSS File(s) dialog, select TENSILE_0000.rad.
4. Click OK.
Setting up the Problem in HyperCrash
Step 2: Create and assign a material

1. From the menu bar, select Model > Material.
2. Right-click in the material list and select Create New > Elasto-plastic > Johnson-Cook (2).
3. For Title, enter Aluminum. Enter all the material data listed above.
4. In the bottom of the material window, right-click in the Support entry box and select Include
picked parts icon .
5. Select the part in the graphics area (left-click).

6. Right-click to validate the selection.
7. Press ENTER or click Save > Close.

Step 3: Create and assign a property
1. From the menu bar, select Model > Property.
2. Right-click in the property list and select Create New > Surface > Shell (1).
3. For Title, enter Pshell.
4. For Shell Thickness, enter 1.7.
5. In the bottom of the property window, right-click in the Support entry box and select Include
picked parts icon .
6. Select the part in the graphics area.
7. Right-click to validate the selection.
8. Click Save > Close.
Step 4: Define boundary conditions representing symmetry

1. From the menu bar, select LoadCase > Boundary Condition.
2. Right-click in the display list area and select Create New.
3. For Name, enter constraint1 and click Save. Expand the folders Translation and Rotation.
4. Right-click in the Support entry box, click Select in graphics and select Add/Remove nodes
by picking selection icon to select the nodes in the Graphic Window, as shown in the figure
below:
5. Click Yes in the Dialog menu bar to validate your selection.

6. To constrain the nodes, toggle Tx, Ry and Rz and click Save.
7. Repeat the same operations to create constraint2, as shown in the figure below:
8. Toggle Tx, Ty, Tz, Rx, Ry and Rz, and click Save.

9. Repeat the same operations to create constraint3, as shown in the figure below. Press SHIFT,
left-click and hold the mouse to draw a box to select the nodes.
10. Toggle Ty, Rx, and Rz.

Step 5: Define the imposed velocity

1. From the menu bar, select LoadCase > Imposed > Imposed Velocity.
3. Set the Title to imposed_velocity.
4. Right-click in the entry box for Time function and select Define Function. A Function
Window opens up.
5. For Function name, enter FUNC_VEL.
6. Enter the first point (0,1) and click Validate.
7. Enter the second point (1e30,1) and click Validate.
8. Click Save in the dialog.
9. Right-click in the Support entry box, click Select in graphics and select the Add nodes by
box selection icon , to select the nodes in the graphic window, as shown in the figure below:
10. Go to the Properties tab and enter a Y-Scale factor = -1.

11. Ensure Direction of the imposed velocity is set to X (translation).
Step 6: Select a node for time history output

1. From the menu bar, select Data History > Time History.
2. In the list display area, right-click and select Create New > TH of nodes.
3. Enter the title Node_79.

4. Click Add Row to add a new row. With that row selected, scroll down to the input section
and enter NODid as 79 and press ENTER.
As an alternative, use the Pick button to select a node in the graphic window.
Step 7: Create Control Cards, Export the Starter and Engine files
1. From the menu bar, select Model > Control Card:
2. Enter the values for the Control Cards, as shown in the images below, saving after every step:



3. Click File > Export > RADIOSS to export the solver file.
4. In the Write Block Format 140 RADIOSS File window that opens, navigate to your desired
run directory and create a new folder named TENSILE_TEST.
5. For filename, enter TENSILE and click OK.
6. Leave the Header window empty and click on Save Model. The file TENSILE_0000.rad is
written.
The model is now ready to run through the Starter and the Engine. It will produce the result files
TENSILEA* for animation in HyperView and TENSILE01 for time history plotting in HyperGraph.
RADIOSS Computing
Step 8: Run RADIOSS Starter and RADIOSS Engine

1. Launch RADIOSS from the Start menu. A HyperWorks Solver Run Manager window appears.

2. In the Input file field, select TENSILE_0000.rad. from the folder you created.
3. Click Run.
The HyperWorks Solver View window is opened. The RADIOSS Starter will run and on completion
the RADIOSS Engine will automatically run.
Step 9: Review the listing files for this run and verify the results
1. See if there are any warnings or errors in the .out files.
2. Using HyperView, plot the displacement and strain contour.
Exercise Expected Results
Total Displacement Contour (mm)

Plastic Strain Contour

Simulate buckling of a tube using half tube mesh with symmetric boundary conditions.
The figure illustrates the structural model used for this tutorial: a half tube with a rectangular
section (38.1 x 25.4 mm) and length of 203 mm.
Model
Model Description
 Simulation time: Engine [0 – 10 ms]
 The tube thickness is 0.914 mm.
 An imposed velocity of 13.3 mm/ms (~30 MPH) is applied to the right end of the tube
 Elasto plastic material using Johnson-Cook law /MAT/PLAS_JOHNS (STEEL).
[Rho_Initial] Initial density = 7.85e-6 Kg/mm3
[E] Young’s modulus = 210 GPa
[nu] Poisson coefficient = 0.3
[a] Yield Stress = 0.206 GPa
[b] Hardening Parameter = 0.450 GPa
[n] Hardening Exponent = 0.5

File needed to complete this exercise: BOXTUBE_0000.rad

Exercise
Step 1: Import the mesh

mm ms kg.
2. Set User interface style as New.
3. Set the working directory to <install_directory>/tutorials/hwsolvers/radioss/.
4. Click Run.
5. Click File > Import > RADIOSS.
6. In the input window, select BOXTUBE_0000.rad.
7. Click OK.

1. Click Model > Material.
2. In the window, right-click and choose Create New > Elasto-plastic > Johnson-Cook (2).
3. For Title, enter Steel.

Enter all the material data, as shown in the following figure.

4. Right-click in Support entry box and click Select in graphics and select Include picked
parts and select boxtube in the graphics area.
5. Press ENTER, or click Yes in the lower right corner.

1. Click Model > Property.
2. In the window, right-click and select Create New > Surface > Shell (1).
3. For Title, enter Pshell.

4. For Shell thickness, enter 0.914.
5. Right-click in the Support entry box and click Select in graphics and select Include picked
parts and select boxtube in the graphics area.
6. Press ENTER or click Yes in the lower right corner.
Step 4: Define Rigid Body

1. Click Mesh Editing > Rigid Body. Right-click in the display list area and select Create New.
2. Right-click in the graphic area and select Add nodes by box selection icon to select the
nodes in the graphic window, as shown below:

3. Press ENTER or click Save to validate.
Note: For the remainder of the tutorial, you need to have the ID of
the master node of the rigid body.
4. Click Show Node Info icon in the toolbar, and select the rigid body master node in the
graphic window. The Node ID appears in the message window (node ID: 803).
5. Click Cancel in the lower right corner to exit the picking loop.
6. Click Close.
Step 5: Define boundary condition applied on rigid body

1. Click LoadCase > Boundary Condition.
3. In the Boundary condition field, enter the name Rigid_BC.

4. In the Node by Id field, enter 803, then click Ok.
5. To constrain the nodes, toggle the buttons in Tx, Ty, Rx, Ry and Rz.
6. Click Save.
Step 6: Define boundary condition representing symmetry

1. In the Boundary condition display list area, select Create New. Name the new constraint set
symmetry.
2. Right-click in the Support entry box and click Select in graphics and select Add nodes by
box selection icon to select the nodes in the graphic window, as shown below:
3. Right-click to validate.
4. Toggle the buttons Tx, Ry and Rz.
Step 7: Define the imposed velocity

1. Click LoadCase > Imposed Velocity. Right-click in the display list area and select Create
New.
2. For Title, enter VELOCITY.

3. Right-click in the Time function parameter entry box and select Define New. A Function
Window opens.
4. For the function name, enter FUNC_VEL.
5. Enter the first point (0, 13.3) and click Validate.
6. Enter the second point (1e30, 13.3) and click Validate.
7. Click Save in the Function Window to accept the function.
8. Expand the Advanced selector at the bottom and in the Node by Id field, enter 803 and click
Ok, (or toggle Add RB master nodes).
9. Go to the Properties tab and enter a Y-Scale factor = -1.
10. Set the direction of the imposed velocity to Z (translation).
Step 8: Define a Rigid Wall

1. Click LoadCase > Rigid Wall > Create.
2. For the Select RWALL, select Infinite Plane.
3. For Title, enter RIGID WALL.
4. Enter the following values: M0: X= 0, Y= 38.1, Z= -204. M1: X= 0, Y= 38.1, Z= 1.
5. In the Distance to search slave nodes field, enter 20.
6. Toggle See option.

7. Click See to visualize it in the graphic window.
Step 9: Create a self contact for the tube

1. Click LoadCase > Contact Interface.
2. Right-click in the Contact Interface list and select Create New > Multi usage (Type 7).
3. Toggle Self impact.
4. Right-click in the graphic area, and select Include picked parts icon and select the part in
the graphic window.
5. Click Yes in the lower right corner of the main window to validate.
6. For Title, enter the name Contact.
7. Set Scale factor for stiffness as 1.
8. Set Min. gap for impact active to 0.900.
9. Set Coulomb friction to 0.200.

Step 10: Export the model
1. Under the Model menu, select Control Card.
2. Check Control Card to activate it.
Note: Make sure to save it before moving to the next Control Card.

3. Click File > Export > RADIOSS.
4. In the Write Block Format 140 RADIOSS File window that opens up, enter the name BOXTUBE
and click OK.
5. Leave the Header of RADIOSS File window empty and click Save Model.
The Starter file BOXTUBE_0000.rad is written.
The model is now ready to run through the Starter and the Engine.
Step 11: Open RADIOSS from Windows Start menu

Step 12: Review the listing files for this run and verify on the results
1. Using HyperView, plot the displacement and strain contour at 10 ms.
Total Displacement (mm) and Plastic Strain (Mid Layer and Average)

RD-3050: Simplified Car Pole Impact in HyperCrash
To simulate frontal pole test with a simplified full car.
Model Description
 UNITS: Length (mm), Time (s), Mass (ton), Force (N) and Stress (MPa)
 Simulation time: Engine file (_0001.rad) [0 – 0.06 ms]
 An initial velocity of 15600 mm/s is applied on the car model to impact a rigid pole of radius
250 mm.
 Elasto-plastic Material /MAT/PLAS_JOHNS (WINDSHIELD)
[Rho_Initial] Initial Density = 2.5x10-9 ton/mm3
[E] Young's Modulus = 76000 MPa
[nu] Poisson’s Ratio = 0.3
[ 0] Yield Stress = 192 MPa
[K] Hardening Parameter = 220 MPa
 Elasto-plastic Material /MAT/PLAS_JOHNS (STEEL)
[Rho_Initial] Initial Density = 7.9x10 -9 ton/mm3
[ 0] Yield Stress = 200 MPa
[K] Hardening Parameter = 450 MPa
[SIG_max] Maximum Stress = 425 MPa

 Elasto-plastic Material /MAT/PLAS_JOHNS (RUBBER)
[Rho_Initial] Initial Density = 2x10-9 ton/mm3
[ 0] Yield Stress = 1e30 MPa
[n] Hardening Exponent = 1
Exercise
Step 1: Retrieve the HyperMesh file

mm ms. kg.
2. Set User Interface style as New.
3. Set the working directory to <install_directory>/tutorials/hwsolvers/radioss.
4. Click Run.
5. Click File > Import > Nastran.
6. In the input window, select full_car.nas.
7. Click OK.
Step 2: Create WINDSHIELD material and assign to car windows

2. In the Material list, right-click and select Create New > Elasto-Plastic > Johnson-Cook (2).
3. For Title, enter WINDSHIELD.
4. Enter all the material data, as shown in the image below.

5. Click the Tree tab and select PSHELL3 and PSHELL16 in the tree.
6. Click to show only these parts.

7. Click the Material tab.
8. Right-click in the Support entry box and click Selected Parts of Tree . This icon allows
adding the part selected in the tree to the selection. The selected parts will be highlighted in the
graphic area.
9. Click Save.
Step 3: Create RUBBER material and assign to car tires

2. For Title, enter RUBBER. Enter all the material data, as shown in the image below.

3. Click the Tree tab and select PSHELL20 to PSHELL23 in the tree.

6. Right-click in the Support entry box and click Selected Parts of Tree . The selected parts
will be highlighted in the graphic area.
7. Click Save.
Step 4: Create STEEL material and assign to all other parts

2. For Title, enter STEEL.
3. Enter all the material data, as shown in the image below.

4. Click the Tree tab and select PSHELL3, PSHELL16 and PSHELL20 to PSHELL23 in the tree.
5. Click to invert the tree selection.

6. Click to show all the parts except the ones made with glass and rubber.

8. Right-click in the Support entry box and click Selected Parts of Tree . The selected parts
will be highlighted in the graphic area.
Step 5: Create a rigid wall to represent the ground

1. Click LoadCase > Rigid Wall > Create.
2. Under Rigid wall name > Select RWALL type, select Infinite Plane.
3. Enter the rigid wall name, Ground.
4. Enter the following values for M0 and M1:

5. In the Selection tab, set the Distance to search for slave nodes to 300.
6. Click See at the bottom of the panel to display the rigid wall.
7. Click Save.
Step 6: Create Pole Rigid Walls

1. Under Rigid wall name > Select RWALL type, select Cylinder.
2. Enter the rigid wall name, Pole.
3. Enter the following values for M0 and M1:

4. Set the Diameter to 500.
5. Set the Distance to search for slave nodes to 1500.
6. Click See at the bottom of the panel to display the rigid cylinder.
7. Click Save.
8. Click Close to close the Rigid Walls panel.
Step 7: Define interface with the whole car

2. In the window right-click and select Create New > Multi usage (Type 7).
3. Select the Self Impact box.
4. In the Title field, enter CAR_CAR.
5. Set [Istf] Stiffness definition to 2: (K=(Km+Ks)/2.
6. For [Gapmin] Min. gap for impact activ., enter 0.7.
7. For [Fric] Coulomb friction, enter 0.2.
8. Set [Iform] Friction penalty formulation to 2: (Stiffness).

9. In the Model Display toolbar, click Display All to display the entire model.

10. Click in the [Mast_id] Master field, move the cursor to the graphical window and right-click.
The menu shown in the image below should appear. Choose the option Add selected parts by
box and use the mouse to drag a box to select the entire car in the graphic window.
11. Click Save.
Step 8: Create an interface between engine and radiator

1. Right-click in the Contact list and select Create New > Multi usage (Type 7).
2. Check Create symmetric interface at saving box.
3. In the Title field, enter ENGINE_RADIATOR.
4. For [Istf] Stiffness definition, set to 2 (K=(Km+Ks)/2.
5. For [Gapmin] Min. gap for impact active, enter 0.7.
6. For [Fric] Coulomb friction, enter 0.2.
7. For [Iform] Friction penalty formulation, set to 2 [Stiffness].
8. In the Tree tab, highlight the part PSHELL28 (Radiator) and PSHELL30 (Engine) and
Isolate them.
9. In the Contact Interface tab, click in the [Slav_id] Slave nodes field, move the cursor to the
graphic window, right-click and select Include picked Part. Select the Radiator (PSHELL28).
10. In the Contact Interface tab, click in the [Mast_id] Master Surface field, move the cursor to
the graphic window, right-click and select Include picked Part. Select the Engine
(PSHELL30).

11. Click Save.
12. Click Close to close the Contact tab.
An additional symmetric interface is created.
Step 9: Define Initial Velocities

1. Click LoadCase > Initial Velocity.
2. In the Velocity list, right-click and select Create New.
3. In the Title field, enter 35MPH.
4. In the Tree window, highlight FULL_CAR.
5. In the [Vx] field, enter 15600.
6. In the Initial Velocity tab and click in the [Gnod_id] Support field. Move the cursor to the
graphic window, right-click and select Add selected parts of tree .
Step 10: Define Time History Nodes

1. Click Data History > Time History.
2. In the Time History list, right-click and select Create New > TH of nodes.
3. For Title, enter RAIL.
4. In the Tree tab, select PSHELL19.
5. Click Isolate Tree Selections .
6. Go back to the Time History panel and click Add/Remove nodes by picking selection in
the second table.

7. Select six nodes on the rails, for example as shown in the following image:
7. Click Yes in the lower right corner or right-click in the graphic window to exit the selection.

1. Create the Engine file: From the menu bar, select Model > Control Card.
2. Check the Control Cards, as shown in the images below.
Note: Make sure to save all control card before editing the next.


3. Under the Quality menu, click Model Checker to check the quality, then check All Solver
Contact interfaces, remove all the initial penetrations in the model.
4. Under the Mesh Editing menu, click Clean, then clean the model before exporting.
5. Click File > Export > RADIOSS, enter FULLCAR and click OK.

6. Leave the Header of RADIOSS File window empty and click Save Model. The Starter file
FULLCAR_0000.rad is written.
Step 12: Open RADIOSS from Windows Start menu
Step 13: Select the Starter file FULLCAR_0000.rad as Input file and Run
the model
Final deformation and energy balance plot

RD-3060: Three Point Bending with HyperCrash
This tutorial demonstrates how to set up a 3-point bending model with symmetric boundary
conditions in Y direction (across the XZ plane).
Model Description
 Simulation time: in Rootname_0001.rad [0 – 7.0E-2s]
 Only one half of the model is modeled because it is symmetric.
 The supports are totally fixed. An imposed velocity of 1000 mm/s is applied on the Impactor
in the (–Z) direction
 Model size = 370mm x 46.5mm x 159mm
 Honeycomb Material /MAT/LAW28: HONEYCOMB
[Rho_I] Initial density = 3.0e-10 ton/mm3

[E11], [E22] and [E33] Young’s modulus (E ij) = 200 MPa
[G11], [G22] and [G33] Shear modulus (Gij) = 150 MPa
 Elasto-Plastic Material /MAT/LAW36: Inner, Outer and Flat
[Rho_I] Initial density = 7.85-9 ton/mm3

[E] Young’s modulus = 210000 MPa
[nu] Poisson's ratio = 0.29


 Strain Curve:
0 1 2 3 4 5 6 7 8 9
STRAIN 0 0.010 0.013 0.015 0.020 0.025 0.030 0.035 0.040 0.045
STRESS8E- 300 310 320 330 340 350 360 370 380 400
 Elastic Material /MAT/PLAS_JOHNS: Impactor
[Rho_I] Initial density = 8e-9 ton/mm3

Step 1: Import the RADIOSS mesh model

1. Open HyperCrash 2017.
2. For User profile:, select RADIOSS V14.
3. For Unit system:, select N mm s T.
4. Select User interface style as New.
5. Click Run.
7. In the input window, navigate to the correct directory and select BENDING_0000.rad.
8. Click OK.

2. In the Window, right-click and select Create New > Elastic > Linear elastic (1) as shown
below:
3. For Title, enter Rigid Material.

4. Enter all the material data, as shown in the following image.
5. Right-click in the entry box Support and click Include picked parts and select the parts
Impactor and Support in the graphics area.
6. Click Yes in the lower right corner to validate.
Step 3: Create and assign a material for Inner, Outer, and Flat parts
1. In the Window, right-click and select Create New > Elasto-plastic > Piecewise linear (36).
2. For Title, enter Shell Material.
3. Enter all the material data, as shown in the following image:

4. Open the Strain rate folder and click to add a row.
5. Right-click in Yield stress function field and click Select in Model to select an existing
function in the model.
6. In the Function file window, select the function with an ID of 2, then click OK to import the
curve. The function can be edited, as shown in the image below.

7. Click the Tree tab and select the parts Inner, Outer, and Flat on the tree.
8. Click to isolate this part.

10. Right-click the entry box Support, and click Include picked parts in the graphic area, and
select the parts Inner, Outer and Flat in the graphics area as shown in the following image.
11. Click Yes in the lower right corner to validate.

Step 4: Create and assign a new material for HCFoam

1. In the Window, right-click and select Create New > Honeycomb > Honeycomb orthotropic
(28).
2. For Title, enter Foam.
3. Enter all the material data, as shown in the following image:

4. Right-click on the Yield stress function 11 field and click Select in Model to select a curve
already present in the model.
5. In the Function file window, select the function with ID of 5, then select OK.
6. Repeat this process for the Yield functions, as shown in the following image.

7. Click the Tree tab and select the part HCFoam (7) on the tree.
8. Click to show only this part.

10. Right-click in the entry box Support, and click Include picked parts to select the part
HCFoam in the graphics area as shown in the following image.

11. Click Yes in the lower right corner.

1. Click Model > Property.
2. In the Window, right-click and select Create New > Surface > Shell (1), as shown below.
3. For Title, enter Shell Property.

4. Enter Shell thickness and Shell element formulation values, as shown in the following
image.

5. Click the Tree tab and select the parts Inner, Outer and Flat on the tree.

7. Click the Property tab.
8. Right-click in the entry box Support, and click Include picked parts to select the parts
Inner, Outer and Flat in the graphics area to assign Shell property.
10. Click Save.
Step 6: Create and assign a property for Impactor and Support

1. For Title, enter Rigid Property.
2. Enter the Shell thickness value as .9119, as shown in the following image.

3. Click the Tree tab and select the parts Impactor and Support in the tree.

5. Click the Property tab.
6. Right-click in the entry box Support and click Include picked parts to select the parts
Impactor and Support in the graphics area to assign the Rigid property.
8. Click Save.
Step 7: Create and assign a property for HCFoam

1. In the Window, right-click and select Create New > Volume > General solid (14).
2. For Title, enter Foam.
3. Click the Tree tab and select the part HCfoam on the tree.
4. Click to show only this part.

5. Go back to the Property tab.
6. In the Flag for solid elements formulation, select HEPH.

7. Right-click in the entry box Support and click Include picked parts to select HCfoam in
the graphics area to assign Foam property.
Step 8: Create rigid body for Impactor

1. From the menu bar, click Mesh Editing > Rigid Body.
2. In the window, right-click to select Create New, enter the name Impactor.
3. Click the Tree tab and select the Impactor assembly on the tree.
4. Click to show all parts.

5. Click the Mesh Editing tab.
6. Right-click in the entry box Support and right-click in the graphic area (as shown below). Select
Include picked parts option to select Impactor in graphic area.

7. Click Yes > Save.
Step 9: Create rigid body for Support

1. In the Title field, enter the name Support.
2. Click in the entry box Support and right-click in the graphic area. Click Include picked parts
option to select Support in the graphic area.
3. Click Yes to complete the selection.
4. Click Save. The rigid body for Support should look like the following image.
5. Click Close.

Step 10: Define boundary conditions for the model
2. In the window, right-click to select Create New.
3. Press F6 to show the rigid bodies.
4. In the Title field, enter Boundary.
5. Right-click in the entry box Support and right-click in the graphic area. Click Add/Remove
nodes by picking selection and select the master node of the rigid body.
6. Constrain all DOF except translation in Z as shown in the following image. To constrain the
nodes, check the boxes for TX, TY, RX, RY and RZ.

7. Click Save.
8. Repeat the same process to create boundary conditions for the Support and Symmetry boundary
condition for the inner/outer/flat.
9. Click node selection icon to select master node of Support, as shown in the following
image.
10. Constrain all DOF by selecting TX, TY, TZ, RX, RY and RZ, as shown in the following image.

11. Click Save.
12. In the Boundary condition creation field, enter Symmetry.
13. Click the Tree tab and select the parts Inner, Outer, HCfoam and Flat on the tree.

15. Press the p key to change the perspective visualization.
16. Click the Boundary Condition tab.
17. From the Visualization toolbar, select the YZ View, as shown below.
18. Right-click in the entry box Support, right-click in the graphic window, and click Add nodes by
box selection to select the nodes, as shown in the image below

19. To constrain the nodes, select TY, RX and RZ.
Step 11: Define Impactor Velocity

1. Click LoadCase > Imposed > Imposed Velocity.
2. In the window, right-click to select Create New.
3. For Title, enter IMPOSED VELOCITY.

4. For Direction, select Z (translation) and -1000 for Y-Scale factor.
5. For Time function, use the predefined curve in the model Funct 1.
6. For Y Scale factor, enter -1000.
7. Press the F6 key to show the rigid bodies.
8. Click in the entry box Support and right-click in the graphic area. Click and select the master
node of Impactor.
9. Click Yes in the lower-right corner.
Step 12: Define contacts for the model

2. In the window, right-click and select Create New > Multi usage (Type 7).
3. Click on the check box next to Create symmetric interface at saving.
4. For Title, enter Support.
5. Click the Tree tab and select the parts Flat and Support on the tree.

7. Click the Contact Interface tab.
8. Set Min gap for impact active to 0.2.

10. Set [Iform] Friction penalty formulation at 2 [Stiffness].
11. Click in the Slave nodes entry box and right-click in the graphic window. A menu appears, click
Include Picked Parts and select the FLAT.
12. Press Y or click Yes at the bottom right of the screen. HyperCrash will automatically move to
the selection of the Master surface.
13. Right-click and click Include Picked Parts and select the Support.
14. Press Y or click Yes at the bottom right of the screen.
15. Click Save.

16. Repeat the same process to create contact between Outer and Impactor.
17. Click the Tree tab and select the parts Outer and Impactor on the tree.

19. Right-click in the window and select Create New > Multi usage (Type 7).
21. Click on the check box next to Create symmetric interface at saving.
22. In the Title, enter Imp_Outer.
23. Set Min gap for impact active to 0.2.
25. Set [Iform] Friction penalty formulation to 2 [Stiffness].
26. Select Outer Part as Slave and Impactor as Master, as shown in the following image.

27. Click Save.
28. Repeat the same process for self impact for Outer, Inner and Flat, as self impact.
29. Click the Tree tab and select the parts Outer, Inner and Flat on the tree.

32. Select Self-Impact.
33. Set Title as Self.
34. Set the Min gap for impact active to 0.7.
35. Set the Coulomb friction to 0.1.
36. Set [Iform] Friction penalty formulation to 2 [Stiffness].
37. Select components Outer, Inner and Flat, as shown in the following image.
38. Click Save.

Step 13: Clean the model
1. Click Mesh Editing > Clean.
2. Select All.
3. Click Clean > Close.

1. Click Model > Control Card and select the control cards in the images below.
Note: Make sure to save each control card before editing the next.

3. In the Output window that opens, enter the name 3PBENDING and click OK.
4. Leave the Header of RADIOSS File window empty and click Save Model.
The Starter file 3PBENDING_0000.rad is written.

Step 15: Open RADIOSS Manager from windows Start menu
Step 16: Run the model 3PBENDING_0000.rad using RADIOSS Manager in

the class_exercise folder
Plastic Strain Simple Average and Rigid Body Force (Impactor)

RD-3150: Seat Model with Dummy using HyperCrash
Introduction
This tutorial presents the different steps involved in building a simple Sled model using HyperCrash
pre-processing tool.
Exercise
Step 1: Model Import

Set the User Profile, units and interface.
2. For User Profile, select RADIOSS V14.
3. For Unit System, select N_mm_s_T.
4. For User Interface Style, select New.
5. Click Run.
Step 2: Import the seat model and merge all components, floor, seatbelt
and foam block
2. Select the file SEAT__00D00.rad.

3. Click OK.
Step 3: Model Merging

1. Click File > Import > RADIOSS. HyperCrash message window prompt.

2. Click Merge.
3. Select the file FLOORD00.rad.
4. Click OK.
5. In the Set all to field, enter the value 100000.
6. Click the Set all to button to offset the numbering of all the entities.
7. Click Merge to merge the floor model.

8. Redo the steps 1 to 7 for the cushion model:
 File: FOAMD00.rad
 Set all to offset: 200000
9. Redo the steps 1 to 7 for the seatbelt model:
 File: BELTD00.rad
 Set all to offset: 300000

Step 4: Model Hierarchy
1. In the Tree, select the subset of the seat named Seat model (300005).
2. Right-click and select Change Name.
3. In the Change Name window, enter the name Seatbelt.

4. Click Ok.
5. Click any item on the tree, right-click and select New Assembly.
6. Enter the name Frame and click Ok.
7. Select the parts Seat plate, Backseat plate, Feet, Seat frame, and Backseat frame using
the SHIFT or CTRL keys.
8. Press and hold the middle mouse button and drag the selected parts into the new assembly
Frame.
9. Select the Tree root (Seat) and right-click.
10. In the pop-up menu, select List Selection. The List Selection dialog opens.

11. In the displayed window, check if all parts have properties (PID) and materials (MID).
12. Click Close and Export the model to save.
Step 5: Connection
To add the feet of the seat and the seatbelt anchorage point to the floor rigid body.
1. Click Mesh Editing > Rigid Body.

2. Select the rigid body: Floor.
3. Click See selected rigid bodies ( ).
4. Click Display All and then Left View (F11).
5. Right-click in the Grnod_Id entry box and click Select in graphic, click Add nodes by box
selection and select all the nodes of the seat, feet and the anchorage points of the seatbelt.
6. Right-click to validate.
7. Select the Floor rigid body in the list, right-click and add the rigid body and master node to time
history.

Connect Seat Cushion to the seat frame with a tied interface (Type 2)

2. Right-click in the window and select Create New > Kinematic condition (Type 2).
3. Display only the cushion parts. Press F11 for XZ view, select Slave nodes section, and click
Add noes by box selection.
4. Holding down the SHIFT key, click to draw a polygon window around nodes on the backside of
cushion of the nodes.
Tip: Press the letter P for non-perspective view, if needed.
Press SHIFT and draw a closed polygon window around the

nodes to select. When finished, release the SHIFT key.
5. Display Frame Assembly in the Tree, pick Master surface section, click Add/Remove a
face and pick one element on each part of the frame facing the cushion. Then select the
Expand option on the lower right corner to pick select all.
6. Select the Expand option on the lower right corner to select all the elements of the seat
assembly facing the seat cushions.
7. Click Yes or Enter on the keyboard to end the selection.

Slave Master
8. For the Title of the contact, enter seat cushion fixation.

9. Click Save.
10. Click at the top of the interface panel, to check the interface. The created interface should be
displayed with green text, as shown below. Otherwise, the interface has to be modified.
11. Click Close.

12. Export the model to save.
Step 6: Dummy positioning

1. Click Safety > Dummy Positioner.


2. From the Dummy model list menu, select New dummy.
A DummyMng panel opens.

3. Select the File subpanel.
4. Select the file H350R12BD00. The dummy model is displayed in the small graphic window.


5. Click Validate.
6. Set Set all to value to 400000.
7. Click the Set all to button to offset the numbering of all entities.
8. Click OK to merge the Dummy model.
9. Click Import in the dummy positioning window and select the file H350R12B_Position.M00
and click OK.
Note: H350R12B_Position.M00 contains all parameters for the

automatic dummy positioning.
10. Close the Dummy positioner and Export the model to save.

Step 7: Seatbelt setting
1. Click Safety > Belt Generator.
2. Enter the name Upper belt and click OK to validate.
3. Click Seat belt reference points ( ).
4. Click Add nodes by picking ( ) and select three nodes, as shown in the following image (red
dots).

5. Click Yes on the right corner and OK to validate the node selection.
6. Click Add/Remove body parts ( ) and select the parts: torso, pelvis, upper legs, and the
seat cushion fabric, as shown in red in the image.

7. Click Yes to validate the selection.
8. Set the Gap value to 5.00 mm.
9. Set the Belt geometric width to 40.
10. Set the Element Size to 8.
11. Click Material ( ) and select the material file BELT.mat you saved to your working directory
from the radioss.zip file. Refer to Accessing the Model Files.
12. Click OK.

13. Click Property ( ) and select the property file BELT.prop you saved to your working
directory from the radioss.zip file.
14. Click OK.
15. Click the Preview button to display the proposed seat belt. Some intersections may exist
between the seat cushion and the seat belt.
16. Use the orientation tools to modify the angle of the Rigid Body 2.

17. Click Save to save the belt definition.
18. Redo the same operations in order to create the lower belt. Select nodes, as shown below:
19. Select the parts: pelvis, upper legs and seat cushion fabric.
20. Click Preview > Save > Close.

Seatbelt vs Dummy
Step 8: Contact interfaces

During the seatbelt creation, two contact interfaces between the seatbelt and the dummy have been
created. You will need to check and remove any remaining intersections and penetrations.
2. Select interface BELT ID 400038.
3. Click See selected ( ) to display.
4. Click in Master Surface, right-click in the graphic area, and click Include picked parts, to
select the Fabric backframe and the Backseat frame as they may come into contact with the
shoulder belt during the analysis.
5. Click Save.

6. Select interfaces BELT ID 400038 and BELT ID 400039.
7. Click See selected ( ) to display.

9. Set Friction penalty formulation to 2.
10. Click Save.
11. Select interfaces BELT ID 400038 and BELT ID 400039.
12. Click Check penetration selected interfaces ( ).

13. In the Quality panel remove the intersections and penetrations, using the Depenetrate Auto
( ).
14. Click Close in order to come back to the Contact Interface panel.
Seat structure
Creation of Self-Impact between different parts of the Seat.
1. In the Tree window, select subsets Frame, Floor and Foam. Click the Isolate icon .
2. Right-click in the Contact list and select Create New > Multi-usage (Type 7).
3. Click Self impact.
4. Set the Title to Self impact seat structure.
5. Set Gap/element option to Variable gap.

8. Right-click in the Master Surface entry box and click Select in graphics > Add selected
parts of tree ( ).
9. Click Save.
10. Select the self impact seat structure interface in the list.
11. Click Check penetration selected interfaces ( ). Some penetrations exist between the seat
cushion and the seat structure.
12. Switch to the Tree window, and select the subset named Frame.
13. Switch to the Quality window and click Fixed part ( ).

14. Press the ESC key to remove all selected parts.
15. Click Add selected parts of tree ( ).
16. Click Depenetrate Auto ( ).
Note: Only the nodes of the seat cushion are moved. The seat parts
are fixed.
17. Click Close twice.

Dummy vs Seat
Creation of Interface between Dummy and Seat.

2. Select interface Create/Modify > Multi usage (Type 7).
3. In the Tree window, select the Foams subset - the two cushion parts only.
4. Switch back to the Interface panel and right-click in the Slave Nodes entry box and click
Select in graphics > Add selected parts of Tree ( ).
5. Again switch to the Tree window.
6. Select the subset named HYBRID III 50% DUMMY FINE MESH V_1.2.
7. Switch back to the Contact interface panel and right-click in the Master Nodes entry box and
click Select in graphics > Add selected parts of Tree ( ).
8. Set the interface Title to Dummy - Seat.
10. Set GAP MIN to 3.00mm.
12. Click Save.

Dummy vs Floor
Creation of an interface between dummy feet and the floor.

1. Right-click in the Contact list and select Create New > Tied with void (Type 10).
2. Set the dummy feet as slave nodes.
3. Set the floor as master surface.

4. Set the interface Title to Feet – Floor.
5. Set Gap for impact activation to 3.0 mm.
Seat Deformer
Modifying the seat cushion mesh to conform to the dummy using the Seat Deformer tool.
Step 1: Edit Pre-simulation settings
To remove the intersection between the dummy and the set HyperCrash will generate a RADIOSS
input deck and run a pre-simulation step. The settings for the pre-simulation are defined in the
menu Option > Presimulation Parameters (for Seat Deformer). For this exercise, modify the
settings, as shown below:

Step 2: Select Seat Parts
1. In the Tree browser, select Foams assembly, Seat plate, Backseat plate, Seat frame, and
Backseat frame, as shown below.

2. Click Safety > Seat Deformer > Pre-simulation (new) and click Add selected parts of
Tree ( ).
Step 3: Select Fixed Nodes

1. In the Tree browser, select the Seat plate, Backseat plate, Seat frame and Backseat frame, as
shown below.
2. Switch back to the Seat Deformer Wizard and click Add selected parts of Tree ( ).
3. Click Next.

Step 4: Select Dummy Parts
1. Select the dummy parts, as shown below.
2. Click Run and the pre-simulation will start.

Step 5: Review the results and apply the deformed shaped
1. Once the pre-simulation is completed review the results in HyperView by opening the h3d file.
Create a cut section in the middle of the dummy and verify that the dummy does not
intersect/penetrate the seat foam.
2. If an intersection/penetration does not exist, go back to the HyperCrash window and load the
results by clicking Yes in the dialog.
3. When the job is completed, click Yes to load the results.
You can also load the results by clicking File > Import > .h3d node coordinates, then click
Yes to the "Warning: all the nodes coordinates will be replaced by the ones found in the
selected .h3d file."

3. Below is the deformed shape for the seat foam after the pre-simulation.

After the seat deformation, check if any initial penetrations remain between the seat and
the dummy.
1. Click LoadCase > Contact Interface to open the Contact Interface tab.
2. Select interface Dummy – Seat.
3. Click Check penetration selected interfaces ( ). Penetrations exist between the seat
beam and the dummy.
4. Click Select All ( ).
5. Click Highlight by Vector ( ).
6. Click Fixed part ( ).

7. Press the ESC key to remove all selected parts.
8. Click Fixed part ( ) and then select the displayed parts of the dummy.
9. Click Depenetrate Auto ( ). Only the nodes of the seat cushion are moved. The parts of the
dummy are fixed.
10. Click Close and then Export the model to save.
Loadcase Setting
Step 9.1: Initial velocity

Update the initial velocity defined in the model to include all the nodes in the model.
1. Click LoadCase > Initial Velocity to open the Initial Velocity tab.
2. Select the initial velocity All in the list.
3. Click See selected initial velocity ( ).

4. Right-click in the Support entry box and click Select in graphics > Add all nodes ( ).
5. Change [Vx] X Velocity from –10000 to –13000 mm/s.

Step 9.2: Imposed velocity

Update the imposed velocity on the floor to decelerate the car.
1. Click LoadCase > Imposed > Imposed Velocity.
2. Select Imposed velocity in the list.
3. Click See selected imposed velocity ( ). The floor rigid body is displayed on the screen. The
imposed velocity is defined on its master node.

4. Right-click the Time Function entry box and select Edit function. Check if the initial value of
the function is the same as the initial velocity.


Step 9.3: Boundary conditions
To simulate the Sled Test, you need to constrain all degrees of freedom on the floor except X-
direction.
2. Select Floor in the list.
3. Click See selected boundary condition ( ). The floor rigid body is displayed on the screen.
The boundary condition is defined on its master node.
4. Verify that the degree of freedom for Ty, Tz, Rx, Ry, and Rz are fixed.


Time History Data Setting
Step 10.1: Nodes

2. Select the node group H350MEF2D00_th_nodes.
3. Click See selected th ( ). These are the nodes of the dummy rigid bodies.
4. For the first 5 nodes of the group:
 Select the node in the list.
 Click See selected node ( ).

 Enter a name in the field Node name, as shown in the table.
 Click Ok.
5. When all labels are defined, click Save > Close.


Step 10.2: Parts
2. Select the second and third part group on the list.
3. Click Delete selected th ( ).

4. Click Yes to the question in the main window (Yes or Cancel). The selected parts groups are
deleted from the model.
5. Select the remaining part group in the list.
6. Click See selected th ( ).

7. Go to the Tree panel and select the root of the tree.

8. Switch back to the Data History panel and click Add parts by tree selection ( ).
9. Click Save and then Export the model to save.
Step 10.3: Interfaces

To add all interfaces to Time History.
1. Click LoadCase > Contact Interface to open the Contact Interface tab.
2. Select all interfaces in the list.
3. Right-click and select Data History > Yes.
Step 10.4: Final Check

1. Go to Quality Module.
2. Select Check All Solver Contact Interfaces.
3. Make sure there are no intersections and initial penetrations; if so, fix them.

4. Click Close.
5. Go to Mesh Editing and clean so all the unused materials and properties are removed.
Step 11: Create Control Cards and Export the Model

1. Click Model > Control Cards to create the Control Cards in the images below.
Note that the /DT/SHELL/DEL command is used to delete some of the rigid body shells to allow
the dummy’s joints to bend during the simulation.


3. Enter a name for the model in the file output window and click OK.
4. Write relevant information regarding your model in the Header window.

5. Click Save Model.
The model is now ready to be computed.

RD-3160: Setting up Multi-Domain Analysis using HyperCrash
The objective of this tutorial is to show how to use the Multi-Domain technique. For more
information on this technique, refer to Multi-Domain.
The model used is a low speed pole impact on a bumper system. Note that the model is finely
meshed (average mesh size = 2mm) in the region of the pole impact and coarsely meshed (average
mesh size = 10mm) elsewhere.
In order to run this analysis using Multi-Domain technique, we have to split this model into two
domains, one containing the finely meshed region and the other containing the rest. A node to node
link (/LINK/TYPE4) is then specified at the boundary between the two domains.
These domains will be created using a pre-processor (using HyperCrash in this tutorial) and the
options specific to Multi-Domain analysis will be added to the input decks through a text-editor. A
Multi-Domain master input file will also be created using a text editor.
For a list of Multi-Domain options, refer to Multi-Domain Input.
For information on how to create links or connections between domains, refer to Multi-Domain in
the User's Guide.
For more information on Multi-Domain Master Input, refer to Multi-Domain Master Input File.

Exercise
Step 1: Import Full Model

2. Set User profile: to RADIOSS V14 and Unit system: to kN mm ms kg.
3. Click Run.
4. Click File > Import > RADIOSS to import the model monodomain_0000.rad.
Step 2: Create Input Files for the Two Domains

1. Click Model > Control Card to set the Control Cards, as shown in the following images:

2. In the Tree, select the subsets of the fine-meshed region (subsets BB_fine1 (21), BB_fine2
(24), and fine_mesh (69)), then right-click, then click Export Selection.

2. In the Export Selection window, select the option to Add model’s control card not linked to
any part, toggle Export geometry and select ALL POSSIBLE RELATED ENTITIES.

3. Click Ok.
4. Save the file as fine_mesh. This will write the file fine_mesh_0000.rad.
5. Click Model > Control Card and enter the following Control Cards:

6. In the Tree, select the subsets/spotwelds of the coarse-meshed region, then right-click
Export Selection.
7. In the Export Selection window, select the option to Add model’s control card not linked to
any part, toggle Export geometry and select ALL POSSIBLE RELATED ENTITIES.

8. Click Ok.
9. Save the file as coarse_mesh. This will write the file coarse_mesh_0000.rad.
Step 3: Define Link between the Two Domains

1. In the original single model, the fine meshed region is connected to the coarse meshed region at
both ends. When this model is split into two domains, we have to create a set of nodes in both
the domains and link these node sets through the starter option (/EXTERN/LINK). This option
has to be added to the two Starter input files using a text editor.

2. Open the Starter file coarse_mesh_0000.rad and add the option /EXTERN/LINK, as shown
below:
Note: Two external links through node sets 1001 and 1002 have
been added to this domain. These node sets were already
defined in monodomain_0000.rad and exported to the two
domains in Step 2.
3. Open the Starter file fine_mesh_0000.rad and add the same options.
4. Create a RAD2RAD input file input.dat defining the two domains and specifying the connections
between them.

5. The input files are now ready to be run using the Multi-Domain technique. For information on
how to launch a Multi-Domain computation, refer to Multi-Domain.

HyperMesh
RD-3500: Tensile Test Setup using HyperMesh

This tutorial demonstrates how to simulate a uniaxial tensile test using a quarter size mesh with
symmetric boundary conditions.
The model is reduced to one-quarter of the total mesh with symmetric boundary conditions to
simulate the presence of the rest of the part.
Model Description
 Simulation time Rootname_0000.rad [0 – 10.]
 Boundary Conditions:
o The 3 upper right nodes (TX, RY, and RZ)
o The center node on left is totally fixed (T X, TY, Rx, RY, and RZ)
o A symmetry boundary condition on all bottom nodes (T Y, Rx, and RZ)
 At the left side is applied a constant velocity = 1 mm/ms on -X direction.

 Tensile test object dimensions = 11 x 100 with a uniform thickness = 1.7 mm
Johnson-Cook elastic plastic material /MAT/PLAS_JOHNS (Aluminum 6063 T7)

[E] Young’s modulus = 60.4 GPa
[SIG_max] Maximum Stress = 0.175 GPa
[EPS_max] Failure Plastic Strain = 0.75

Input file for this tutorial: TENSILE_000.rad

Exercise
Step 1: Load the RADIOSS (Block) User Profile

1. Launch HyperMesh Desktop.
2. From the Preferences menu, select the User Profiles or from the toolbar, click the icon.
3. Select RADIOSS (Block140) and click OK.
Step 2: Load the solver deck

1. Click File > Import > Solver Deck or click .
2. Click the Select File icon to open the TENSILE_0000.rad file you saved to your working
directory from the radioss.zip file. Refer to Accessing the Model Files.
3. Click Open.
4. Click Import.
5. Click Close to close the window.
Step 3: Define material for the tensile test object

1. In the Model browser, right-click and select Create > Material.
A Material with name material1 of card image M1_Elastic appears in the Entity Editor (EE) in
the bottom pane of the Model browser.
2. In the Entity Editor (EE), for Name, enter Mat_1 in the Value field.
3. Set Card Image to M2_PLAS_JOHNS_ZERIL.
4. Click Yes on the pop-up that warns of a card image change.
5. Input the values, as shown in the following image in the EE.

Step 4: Define property for the tensile test specimen
1. In the Model browser, right-click and select Create > Property.
A Property with name property1 of card image P1_SHELL appears in the Entity Editor (EE) in
the bottom pane of the Model browser.
2. For Name, enter sheet_1.7.
3. For Thick, enter 1.7. in the Value field corresponding to sheet thickness.

Step 5: Assign material and property to the test specimen
1. In the Model browser, select the SHELL_1 component. The Entity Editor opens for the
component.
2. For Name, enter Tensile_coupon.
3. Click Prop_Id, to activate the option.
4. Click Unspecified > Property.
5. In the Select Property dialog, select sheet_1.7 from the list and click OK.
6. Repeat steps 3 - 5 for Mat_Id and select Mat_1.

Step 6: Create boundary conditions
1. From the Utility browser, start the BCs Manager from the pull-down menu, select Tools > BCs
Manager.
2. For Name, enter constraint1, set Select type to Boundary Condition and set GRNOD to
Nodes.
3. Click on Nodes. A nodes selection appears. Select the three nodes, as shown in the figure
below and click proceed.
4. Fix degrees of freedom Tx, Ry and Rz.

5. Click Create to create the constraint. The created constraint appears in the table, and handles
appear in the graphics area.
Nodes.
7. Select the node, as shown in the image below.
8. Fix degrees of freedom Tx, Ty, Rx, Ry and Rz.
9. Click Create to create the constraint. The created constraint appears in the table, and a handle
appears in the graphics area.
Nodes.
11. Select the nodes, as shown in the image below.
12. Fix degrees of freedom Ty, Rx and Rz.

13. Click Create to create the constraint. The created constraint appears in the table, also handles
appear in graphics.
Step 7: Create Imposed Velocity

1. For Name, enter velocity, set Select type as Imposed Velocity and set GRNOD to Nodes.

2. Select the nodes, as shown in the image below.
3. Set the direction as X and Scale Y as -1.0.

4. Click Create/Select curve ID for Curve ID. An XY curve editor appears.
5. Click New to create a new curve.
6. For Name, enter Load and click proceed.
7. Enter the values, as shown in table below.
8. Click Update to update the curve with the new values.

9. Click Close to close the Curve editor, the created curve is assigned to this constraint.
10. Click Create to create the velocity boundary condition.
11. Click Close to close the BCs Manager.
Step 8: Create output requests and control cards

For this exercise the output request will be generated from the Engine file assistant which is located
in the Utility browser.
1. To start the Engine file assistant, select Tools > Engine File Assistant.
2. Input the values, as shown below:

The tool generates typical output requests, such as stress, strain, velocity, etc.
Step 9: Export the model as TENSILE_0000.rad

1. From the File menu, click Export > Solver Deck or click the Export Solver Deck icon .
2. For File:, click the folder icon and navigate to the destination directory where you want to
export to.
3. Enter the name TENSILE_0000.rad and click Save.
4. Click the downward-pointing arrows next to Export options to expand the panel.
5. Select Merge starter and engine file to export the Engine and Starter file as one file.

6. Click Export > Close.
Step 10: Open RADIOSS Manager from Start menu
Step 11: Run the model TENSILE_0000.rad using RADIOSS Manager

1. Select the TENSILE_0000.rad for the Input file.
2. Click Run.

Step 12: Review the listing files for this run and verify the results
1. See if there is any warning or errors on .out files.
Total Displacement Contour (mm)

RD-3510: Cantilever Beam with Bolt Pretension
This tutorial demonstrates how to simulate a simple cantilever problem with a concentrated load at
the free end, using Dynamic Relaxation (/DYREL) method to obtain a static solution.
Model Description
 Simulation time:
o CANTILEVER_0000.rad [0 – 25.1 ms]
 Steps to setup this model:
o Fix the Cantilever Beam to the support with a 10 kN pre-tension. The bolt attains 10 kN in
10 ms and remains constant thereafter.
o After pre-tension, a concentrated load of 0.2 kN is gradually applied at the free end of the
beam from 10 ms to 25 ms and it remains constant thereafter.
 Material used:
Elasto-plastic material /MAT/LAW2.

[c] Strain rate coefficient = 0.014 GPa
[EPS_0] Reference strain rate = 1

Input file for this tutorial: CANTILEVER_0000.rad

Exercise

2. From the Preferences menu, select the User Profiles or from the toolbar, click the icon.

2. Click the Select File icon to open the CANTILEVER_0000.rad file you saved to your working
3. Click Open.
4. Click Import.
Step 3: Create a rigid body connecting spring ends to Bolt Support

component
1. In the Model browser, right-click and select Create > Component.
A component is created and is shown in the Entity Editor (EE), below the Model browser.
2. Using the Entity Editor (EE), change the Name to Rigids.
3. Set the Card Image as None.
4. In the Model browser, hide the component 1.
5. Click the Mask icon in the toolbar.

6. In the graphics area, select one element from the bolt.

7. Click on elems >> by attached to select the whole bolt.
8. Click mask to hide them and click return.
9. From the 1D page, select the rigids panel.
10. Click the selector arrow nodes 2-n: and change it to multiple nodes.
11. In the rigids panel, for primary node, select the node at the end of spring, as shown in Fig 1
below, and for nodes 2-n, select the nodes, as shown in Fig 2.
Note: Be sure to set the selector to multiple nodes.
Fig 1

Fig 2
12. With all the DOF’s checked, click create to create the rigid body.
13. Click the Mask icon in the toolbar and click reverse to show remaining elements of the bolt.
14. Click return to exit the panel.
15. In the Model browser, rght-click the 3 components and click Show to display onscreen, as
shown below.
16. Use Steps 3.10 through 3.12 to create a rigid body with the nodes shown in the following image
with the other ends of the springs as the primary node and the nodes on the bolts as slave
nodes.

Step 4: Create and assign material, property to Plate and Support bolts
1. In the Model browser, click the component 1. The component appears in the Entity Editor.
2. Change the name of the component to Plate.
3. Set Card Image to Part.
5. For Name, enter Steel and set the Card Image to M2_PLAS_JOHNS_ZERIL and click Yes to
confirm.
6. Enter the values, as shown below.

8. For Name, enter Plate, and set the Card Image to P14_SOLID and click Yes to confirm.
9. In the Model browser, click the component 2, the EE for the component opens.
10. For Name, enter Bolt_Support.
11. Set the Card Image to Part.

12. For Prop_Id, click Unspecified > Property and select the property, Plate and click OK.
13. For Mat_Id, click Unspecified > Material and select the material, Steel and click OK.
Step 5: Create and update properties for Pre-tensioner Spring

1. In the Model browser, select the component 3, which opens the Entity Editor.
2. For Name, enter Spring.
3. Set the Card Image to Part.
4. In the Model browser, right-click and select Create > Property. A new property is created and
a dialog opens with the new property.
5. Change the Name to Spring.
6. Set the Card Image to P32_SPR_PRE and click Yes to confirm.
7. Fill in the other values, as shown below:
8. In the Model browser, click on the property Spring to open the Entity Editor.
9. Right-click on IFUN2 and select Create to create and attach a curve. A Create Curve dialog
opens.
10. Change the Name of the curve to Stiffness.
11. Click Close to exit the dialog.

12. In the Model browser, select the curve Stiffness, right-click and select Edit from context menu.
13. The XY curve editor appears. Fill in the values, as shown below.
14. Click Update > Close. The created curve is assigned to the property.
Step 6: Defining Boundary Conditions to fix bottom of the BOLT_SUPPORT

1. From the Tools menu, start the BCs Manager.
2. For Name, enter FIXED, set Select type to Boundary Condition and set GRNOD to Nodes.
3. Click on the nodes, the nodes selection appears; by window option, select the bottom layer of
the bolt support, as shown below and the selection should appear as shown below in the XY
Plane view:

4. Fix all translational degrees of freedom.
5. Click Create to create the constraint. The created constraint appears in the table and a handle
appears in graphics area.
Step 7: Defining the load (CLOAD) of the edge of the beam

1. For Name, enter LOAD, set Select type to Concentrated Load and set GRNOD to Nodes.
2. Select the nodes on the edge of the beam, as shown in the image below by window option.

3. For Direction, select Y.
4. Set Scale Y, to -1.0 to apply load in negative Y direction.
5. Click the Create/Select curve tab. A GUI to enter curve appears.
6. Create a curve with Name LOAD and enter the values, as shown below using the same
procedure explained in Step 5.
x = {0, 10, 25, 250}
y = {0, 0, 0.02, 0.02}
7. Click Update and Close in the XY curve editor GUI, the created curve is assigned to the BC.
8. Click Create to finish the creation of the load at the selected nodes.
Step 8: Define a contact interface between Plate and Support_Bolt

1. In the Model browser, right-click and select Create > Contact. A contact is created and the
Entity Edit opens.
2. Set Name as SELF.
3. Set Card Image to TYPE7 and click Yes to confirm.
4. Click on Grnod_id (S) in the EE and set the selector to Components.
5. Pick the components Plate and Support_Bolt using the list selection dialog.
6. Click on Surf_id (M) in the EE and set the selector to Components.
7. Pick the components Plate and Support_Bolt using the list selection dialog.
8. Set Igap to 0.
9. For FRIC, enter 0.1 and for GAPmin, enter 0.04.

Step 9: Create time history to obtain displacement at free end
1. In Model browser, right-click and select Create > Output Block from the Analysis page, select
the output block panel.
2. In the Entity Editor, set the name to Deflection and select the nodes on the free end of the
cantilever, as shown in the following image:
3. Set NUM_VARIABLES to 1 and click on the Data:Var icon . A table will open, enter the
variable name DEF.
4. Click edit and enter the variable name DEF.

Step 10: Create output request and control cards

Step 11: Run the model checker
1. Click Tools > Model Checker > RadiossBlock to open the Model Checker tab.
2. The Model Checker will display a list of perceived errors within the model.
For most of these issues, the Model Checker is equipped to auto-correct many issues, decreasing
the possibility of a solver error.
3. Click the Apply Auto Correction icon and click the Run icon to auto-correct issues within
the model.

1. Click File > Export or click the Export icon .
export to.
3. For Name, enter CANTILEVER and click Save.
5. Select Merge starter and engine file to export both the Starter and Engine file in one file.
6. Click Export to export the file.
Step 13: Run the model in the solver

1. Go to Start > Programs > Altair HyperWorks 2017 > RADIOSS.
2. For Input file, browse to the exercise folder and select the file CANTILEVER_0000.rad.

3. Post-process the results with HyperView.
4. Using HyperGraph, open the T01 file and plot the deflection at the free end of the cantilever.

RD-3520: Pre-Processing for Pipes Impact using RADIOSS
For this tutorial it is recommended to complete the introductory tutorial, HM-1000: Getting Started
with HyperMesh. Working knowledge of the creation and editing of collectors and card images are a
definite pre-requisite. Familiarity with the Interfaces panel, and the creation of boundary conditions
are useful, although not required.
Objective
In this tutorial you will learn how to set up a RADIOSS input file in HyperMesh for analyzing the
impact response between two pipes. The modeling steps that are covered are:
 Creating materials, sections, and parts for the model.
 Defining the contact between the two pipes using /INTER/TYPE7.
 Applying a translational initial velocity to a pipe using the /INIVEL card.
 Applying local constraints to the other pipe using the /BCS card.
Model Description
The units used in this tutorial are milliseconds, millimeters and kilograms (ms, mm, kg), and the
tutorial is based on RADIOSS 14.0.
Pipe model

Exercise

2. From the Preferences menu, select the User Profiles or click the icon on the toolbar.
Step 2: Import the solver deck

2. Click the Select File icon to open the pipesd00.rad file you saved to your working directory
3. Click Import. The model loads into the graphics area.
Note: On import of a RADIOSS deck, any HyperMesh warning and

error messages are written to a file named radiossblk.msg.
This file is created in the folder from which HyperMesh is
started. The content of the file is also displayed in a pop-up
window.
On import, any RADIOSS cards not supported by HyperMesh

are written to the control card unsupp_cards. This card is
accessed from the control cards panel on the BCs page and is a
pop-up text editor. The unsupported cards are exported with
the rest of the model.
Care should be taken if an unsupported card points to an entity

in HyperMesh. An example of this is an unsupported material
referenced by a /PART card. HyperMesh stores unsupported
cards as text and does not consider pointers.
On import, HyperMesh renumbers entities having the same ID

as other entities. In HyperMesh, for example, all elements
must have a unique ID. The message file radiossblk.msg
provides a list of renumbered elements and their original and
new IDs.
Step 3: Understand the relationships between the /PART, /SHELL, /MAT

and /PROP cards in HyperMesh
A /PART shares attributes such as section properties (/PROP) and a material (/MAT). A group of
shells (/SHELL) sharing common attributes generally share a common part ID (PID).

The figure below shows how these keywords are mapped to HyperMesh entities:
/SHELL elem_ID part_ID Organized into component
collectors
/PART part_ID prop_ID mat_ID Component collector with a

component card image
/PROP prop_ID Property collector with a property

card image
/MAT mat_ID Material collector with a material

card image
Map to HyperMesh Entities
Component, property and material collectors are created and edited from the Collectors panel.
For the RADIOSS keyword interface, there is only one component card image and it is named Part.
There are several property card images, such as P1_SHELL, P2_TRUSS, and P14_SOLID. There are
many material card images, such as M1_ELAST and M48_HONEYCOMB.
The complete list of card images is available from the Collectors panel, as you assign card images to
the various types of collectors.
A HyperMesh card image allows you to view the image of keywords and data lines for defined
RADIOSS entities as interpreted by the loaded template. The keywords and data lines appear in the
exported RADIOSS input file as you see them in the card images. Additionally, for some card
images, you can define and edit various parameters and data items for the corresponding RADIOSS.
Use the Entity Editor or card (card editor) panel from the permanent menu to review and edit card
images. Also, for many entities, their card image can be viewed and edited from the panels in which
they are created.
Step 4: Create a /MAT card

In HyperMesh, a /MAT card is associated to a material collector. To relate it to a /PART card, the
material needs to be assigned to a component.
You can assign the material to the component collector as you create the component using the
Create subpanel of the Collectors panel or from component create options in the pull-downs or from
the Model browser using the Entity Editor (EE). In situations where the material was not assigned to
the component at the time of creation (and in this case, a dummy material is created with the same
name as the component collector), update the component collector's definition by assigning the
material in the Update subpanel of the Collectors panel or from the Assign option in Model browser
or using the Entity Editor (EE) of the component.
In this step, create a material with the M1_ELAST card image using the Model browser. This
material will be assigned to both pipes.
1. In the Model browser, right-click and select Create > Material. A material is created and
displayed in the Entity Editor (EE) below the Model browser.

2. For Name, enter elast1.
3. Set Card Image to M1_ELAST.
4. In the Entity Editor (EE), click to activate the field.
Rho_Initial (density), enter 7.8E-6
E (Young's modulus), enter 208
nu (Poisson's ratio), enter 0.30
Note: If you have difficulties completing any task with the creation,
update or editing of materials in this tutorial, refer to the on-
line help for the materials by clicking Help from the menu.
Hint: Any material that was mistakenly created with wrong values
can be edited using the card image option.
In this step, the material created will be used for the analysis. The next step is to define the /PROP
card that will be used to define the properties of the elements in the model.
Step 5: Create a /PROP card

In HyperMesh, the /PROP card is assigned to a property. To generate this card, create a property
collector using either the Property icon in the toolbar or click Properties > Create from the pull-
down or from the Model browser, click Create > Property.
The model consists of two pipes modeled with shell elements. Create a property with a
/PROP/SHELL card that will be used for all the elements.

1. In the Model browser, right-click and select Create > Property. A property is created and
displayed in the Entity Editor (EE).
2. For Name, enter prop_shell.
3. Set Card Image to P1_SHELL.
4. Set Ishell to 24.
5. For shell thickness Thick, enter 2.5.
Step 6: Assign the /PART, /MAT and /PROP cards to the elements
Assign the /PART card to the component for the coarse pipe and specify the /PROP/SHELL card ID in
it.
1. In the Model browser, select the components Pipe1 and Pipe2. A combined Entity Editor (EE)
appears for both the selected components.
2. Set Card Image to PART.
3. For Prop_Id, click Unspecified > Property and select the property, prop shell and click OK.
4. For Mat_Id, click Unspecified > Material and select the material, elast1 and click OK.

Step 7: Create Interface/Contact cards
RADIOSS contacts can be created from Model browser, with a right-click Create > Contact, they
can also be created in the Interfaces panel from the Analysis page or from the menu, select BCs >
Create > Interfaces.
A RADIOSS contact is a HyperMesh group. When you want to manipulate a /INTER card, such as
delete it, renumber it, or turn it off, you need to work with HyperMesh group entities.
In this step, create a contact between the two pipes using /INTER/TYPE7. The pipe with the coarser
mesh (2) will be the master surface while the one with finer mesh (1) will be the slave surface.
RADIOSS has multiple ways to define master and slave entity types from which to choose; in this
example define the master and slave entities as components, by doing this, the master will be
exported as /SURF/PART and the slave as a /GRNOD/PART.
1. In the Model browser, right-click and select Create > Contact. A contact is created and the
Entity Editor (EE) opens.
2. For Name field, enter contact.
4. For Surf_id(M) that corresponds to the master selection, click on the drop-down arrow and
select Components.
5. Click Components and select component 2 in the selection or on the graphics window and click
OK.
6. For Grnod_id(S) that corresponds to the slave selection, click on the drop-down arrow and
select Components.
7. Click Components and select component 1 in the selection or on the graphics window and click
OK.
8. For static coefficient [Fric], enter 0.10.
In this step, you defined the contact between the two pipes as /INTER/TYPE7.

Step 8: Create boundary conditions
Boundary conditions for RADIOSS can be efficiently created using the BC’s Manager available on the
Utility browser. The BC’s Manager can be accessed from the Tools menu. RADIOSS boundary
conditions are mapped to load collector in HyperMesh.
In this step, you will apply a translational initial velocity along Z direction to the coarse pipe using
BC’s Manager.
1. In the BCs Manager, enter Name as tran_vel and set Select type as Initial Velocity under
the Create header.
2. Click Parts, select component 2 from the GUI, and click proceed. This creates the entity set of
type GRNOD, which is referred to in the /INIVEL card.

3. In the BC’s Manager, enter the initial velocity components as 0, 0 and -30 for Vx, Vy and Vz
fields.
There is an option for creating/referring the initial velocity card to a local coordinate system.
However, if nothing is specified, the global coordinate system is selected by default.
4. Click Create. Cross check in the Model browser for your reference that a load collector and an
entity set are created.
This completes the creation of an initial velocity for the pipe in the negative global Z direction.
Step 9: Create a /BCS and constrain the finer mesh pipe

In this step, you will fully constrain the end nodes of the bottom pipe by using the Boundary
Conditions Manager.
1. In the BCs Manager, enter Name as SPC and set Select type as Boundary Condition.
2. Now specify the node set of type as GRNOD for the BCS card, switch the entity from Parts to
Nodes and select the end nodes of the bottom pipe, which are to be constrained.
3. Under the Boundary condition components subheading (as illustrated below) activate all the
translational and rotational check boxes. Click Create.
A load collector with a BCS card is created and applied the nodes as selected in the above steps.
A corresponding node set is created.
Step 10: Create output definitions and control cards


2. For File:, click the folder icon and then navigate to the destination directory where you want
to export to.
3. For Name, enter pipe and click Save.
5. Select Merge starter and engine file to export both the Starter and Engine file in one file.
6. Click Export to export the solver deck.
This concludes this tutorial. You may discard this HyperMesh model or save it for your own
reference.
In this tutorial some of the concepts that govern the HyperMesh interface to RADIOSS are
introduced. You also used numerous panels that allowed you to do basic modeling in terms of
RADIOSS, such as defining contacts or boundary conditions.
Step 12: Run the model in the solver

2. For Input file, browse to the exercise folder and select the file pipe_0000.rad.

Deformation and energy balance plot

This exercise simulates buckling of a tube using half tube mesh with symmetric boundary
conditions.
The figure illustrates the structural model used for this tutorial: a half tube with a rectangular
section (38.1 x 25.4 mm) and length of 203 mm.
Model Description
 Simulation time: Engine [0 – 10 ms]
 The tube thickness is 0.914 mm.
 An imposed velocity of 13.3 mm/ms (~30 MPH) is applied to the right end of the tube
 Elasto-plastic material using Johnson-Cook law /MAT/PLAS_JOHNS (STEEL).
[Rho_Initial] Initial density = 7.85e-6 Kg/mm3
[nu] Poisson coefficient = 0.33
File needed to complete this tutorial: tube_box.hm

Exercise
Step 1: Load the RADIOSS User Profile

2. From the Preferences menu, select the User Profiles or click the icon on the toolbar.
Step 2: Load the model

1. From the toolbar, click the Open Model icon to open the tube_box.hm file you saved to your
working directory from the radioss.zip file. Refer to Accessing the Model Files.
2. Click Open. The model loads into the graphics area.
Step 3: Create Material for the tube

1. In the Model browser, right-click and select Create > Material. The Entity Editor is displayed
below the Model browser.
2. For Name, enter Steel.
3. Set Card Image to M2_PLAS_JOHNS_ZERIL and click Yes to confirm.
4. Set Type as PLAS_JOHNS.

6. Click anywhere in the Model browser to exit the Entity Editor.
Step 4: Create Property for the tube

1. In the Model browser, right-click and select Create > Property. The Entity Editor is displayed
2. For Name, enter Pshell.
3. Set Card Image to P1_SHELL.

Step 5: Assign material and property to the component
1. Select the component Tube_box in the Model browser.
2. In the Entity Editor, for Prop_Id, click Unspecified > Property
3. In the Select Property dialog, select Pshell and click OK.
4. In the Entity Editor, for Mat_Id, click Unspecified > Material.
5. In the Select Material dialog, select Steel and click OK.

Step 6: Create Rigid Body
1. Create a component collector RBODY. Set Card Image to None in the Entity Editor.
2. In the 1D page, select rigids.
3. Set nodes 2-n to multiple nodes.
4. Set primary node tab to calculate node.
5. Select the nodes of one edge to tie all the degree’s of freedom, as shown in the image below:

6. Click create.
Step 7: Create Symmetry Boundary Conditions

1. Click Tools > BCs Manager to start the BCs Manager.
2. For Name, enter Symmetry, set Select type as Boundary Condition and set GRNOD to
Nodes.
3. Click on the nodes, nodes selection appears; by window option, select the top layer of the
channel as shown below and the selection should appear as below:

4. Fix the degrees of freedom for symmetry condition, as shown below:
5. Click Create to create the constraint. The created constraint appears in the table, and a handle
appears in graphics area.
Step 8: Create Imposed Velocity

1. For Name, enter Velocity, set Select type as Imposed Velocity and set GRNOD to Nodes.
2. Select the master node of the RBODY on which the boundary condition needs to be applied.

3. Set the Direction as Z.
4. Click Create/Select curve to create imposed velocity loading curve. A new GUI opens.
5. Click New to enter Load as the name of the curve.
6. Click proceed.
7. Enter the X values as 0, 1000.
8. Enter corresponding Y values as 13.3, 13.3.
9. Click the Create tab to create the constraint. The created constraint appears in the table and a
handle appears in graphics area.
Step 9: Create boundary condition on the rigid body

1. Enter Name as RBODY_constraint, set Select type as Boundary Condition and set the
GRNOD to Nodes.
2. Select the master node of the RBODY on which the boundary condition need to be applied.

3. Set the degrees of freedom to not allow movement in X and Y direction and no rotation about Y-
axis and Z-axis, as shown below.
4. Click the Create tab to create the constraint. The created constraint appears in the table and a
handle appears in graphics area.
Step 10: Create a Rigid Wall

1. In the Model browser, right-click and select Create > Rigid Wall.
2. Set the Geometry Type as Infinite plane.
3. Click on the Base node option and select extreme node opposite to rigid body edge.

4. Set the normal vector using the N1, N2, N3 option, as shown below. Ensure that N3 is not
active. Click Proceed.
Note: Keep N3 inactive.
5. Set d (distance) value to 20.
6. Go to Analysis > rigid walls panel.

7. Move to the Geometry page. Click on the Edit tab besides base node and change the Z value
to 10.0 to be away from the channel along the Z-axis.
8. Click update.

Step 11: Creating a Self Contact to avoid self penetration during impact
1. In the Model browser, right-click and select Create > Contact. The Entity Editor will open.
2. Enter the Name as Self_Interface and set the Card Image as TYPE7 and click Yes to
confirm.
3. Toggle the option to Components for Grnod_id (S) (slave entity), select Tube_box and click
OK.
4. Toggle the option to Components for Surf_id (M) (master entity), select Tube_box and click
OK.
5. Set STFAC = 1, FRIC = 0.20 and GAPmin = 0.90.

6. Click anywhere in the Model browser to exit the Entity Editor.
7. To review the created interface, go to the Analysis > Interface panel.
8. Go to the update subpanel, select created interface and click review. It will show master and
slave surface as blue and red.

1. Launch the HyperMesh Solver browser from View > Browsers > HyperMesh > Solver.
2. Right-click in the Solver browser general area to create the cards, shown below with the given
values for each parameter:
Keyword Type Keyword Parameter Parameter

Value
CONTROL CARDS TITLE Status [Checked]
CONTROL CARDS TITLE TITLE Box_Tube
ENGINE KEYWORDS RUN Status [Checked]
ENGINE KEYWORDS RUN Tstop 10.01
ENGINE KEYWORDS PRINT Status [Checked]
ENGINE KEYWORDS PRINT N_Print -100
ENGINE KEYWORDS ANIM/ELEM Status [Checked]
ENGINE KEYWORDS ANIM/ELEM EPSP [Checked]
ENGINE KEYWORDS ANIM/ELEM ENERGY [Checked]
ENGINE KEYWORDS ANIM/ELEM VONM [Checked]
ENGINE KEYWORDS ANIM/ELEM HOURG [Checked]
ENGINE KEYWORDS ANIM/VECT Status [Checked]
ENGINE KEYWORDS ANIM/VECT VEL [Checked]
ENGINE KEYWORDS ANIM/VECT FOPT [Checked]
ENGINE KEYWORDS ANIM/VECT CONT [Checked]
ENGINE KEYWORDS ANIM/DT Status [Checked]
ENGINE KEYWORDS ANIM/DT Tstart 0
ENGINE KEYWORDS ANIM/DT Tfreq 1

Value
ENGINE KEYWORDS ANIM/NODA Status [Checked]
ENGINE KEYWORDS ANIM/NODA DMAS [Checked]

export to.
3. Enter the name boxtube and click Save.
5. Select Merge starter and engine file to export the engine file with the model file.
6. Click Export to export the file.
Step 14: Run the solver using RADIOSS Manager

2. For Input file, browse to the exercise folder and select the file boxtube_0000.rad.

Step 15: Results analysis in HyperView
Total Displacement (mm) and Plastic Strain (Mid Layer, Simple Average)

RD-3540: Front Impact Bumper Model using HyperMesh
For this tutorial it is recommended to complete the introductory tutorial HM-1000: Getting Started
with HyperMesh, as well as RD-3520: Pre-Processing for Pipes Impact Using RADIOSS for the basic
concepts on the HyperMesh RADIOSS interface.
In this tutorial you will learn how to use HyperMesh to set up a RADIOSS input deck for analysis of
the impact of a bumper against a barrier behind rigid wall. The modeling steps that are covered are:
 Associating /PART, with /MAT and /PROP.
 Converting node-to-node connections (/RBODY) into a mesh-less welding formulation
(/INTER/TYPE2 with /SPRING) using HyperMesh connectors.
 Defining the contact for the elements in the bumper with an /INTER/TYPE7 card.
 Defining the interaction between bumper and barrier with an /INTER/TYPE7 card.
 Defining the interaction between barrier and rigid wall with the /RWALL/PLANE and
/BOX/RECTA cards.
 Specify the output of resultant forces for a plane on the left interior and exterior crash boxes
with /SECT.
 Creating a /TH/NODE card to output time history for nodes.
The units used in the model are millisecond, millimeter and kilogram (ms, mm, kg), and the tutorial
is based on RADIOSS Block 14.0.
Exercise
The model used consists of a simplified bumper model (see image below):
Bumper model

2. From the Preferences menu, select the User Profiles or click the icon in toolbar.

Step 2: Load the bumper.hm file
1. Click the Open Model icon to open the bumper.hm file you saved to your working directory
2. Click Open.
The model loads into the graphics area.
Step 3: Define vehicle mass component to partially take into account the
inertia properties and mass of the missing parts of the vehicle
1. In the Model browser, right-click and select Create > Component. The Entity Editor (EE) will
open.
2. For Name, enter Vehicle mass.
3. Set Card Image to None and click Yes to confirm.
4. Click Geometry > Create > Nodes > XYZ to open the Nodes panel.
5. In the X field, enter 700.
6. In the Y field, enter 0.
7. In the Z field, enter 170.
8. Click create to create the node.
9. Go to the 1D page, and click rigids.
10. Click the selector arrow nodes 2-n: and select sets.
11. For primary node, select the node created in the steps above.
12. Click sets and select the Constrain Vehicle set.
13. With all the DOF’s checked, click create to create the rigid body.
Note: A spider will be drawn connecting the created node to the

edge nodes of the structure modeled.
14. Click Card Edit in the toolbar, set the selector to elements and select the rigid body
created.
15. Click edit.
16. Fill the mass and inertia information in the card image, as shown in the table below:
Mass JXX JXY JXZ JYY JYZ JZZ
800 1.5E+07 -5.0E+03 -8.0E+06 5.0E+07 -900 6.0E+07
17. Set ICOG as 4 and set Ispher as 0.


Step 4: Create a GRNOD/BOX/RECTA that contains all nodes except
barrier nodes
1. Click View > Browsers > HyperMesh > Solver to activate the Solver browser, if it is not
active on your screen.
2. Right-click in the Solver browser and select Create > BOX > BOX/RECTA. The Entity Editor
opens.
3. For Name, enter box velocity.
4. Optionally, select a Color.
5. Enter Corner1 and Corner2 X, Y, and Z coordinates, as shown below.
Step 5: Create initial velocity on bumper except barrier

1. Click Tools > BCs Manager.
2. In the BCs Manager, enter Name as trans_vel.
3. Select the Select type as Initial Velocity under the Create header.
4. Set the entity selector to BOX under GRNOD.
5. Click on it and select box velocity.
6. Enter -10, 0, 0 for Vx, Vy and Vz fields, respectively.

7. After the above step, a set named InitialVelocity_grnodbox is created automatically or you
can create this set before the above step and then refer to this set in the above step, instead of
BOX.
8. Click the Create > Close.
Step 6: Define master surface for contact

1. Right-click in the Solver browser and select Create > SURF_EXT > PART. The Entity Editor
opens.
2. For Name, enter barrier_surface.
3. For Entity IDs, click on Components.
4. In the Select Components dialog, select barrier and click OK.
5. Right-click in the Solver browser and select Create > SURF > PART. The Entity Editor opens.
6. For Name, enter bumper_surface.
7. For Entity IDs, click on Components.
8. In the Select Components dialog, select bumper, exterior crashbox left, exterior crashbox
right, interior crashbox left, and interior crashbox right and click OK.
9. Right-click in the Solver browser and select Create > SURF > SURF. The Entity Editor opens.

10. For Name, enter barrier_bumper_surface.
11. For Entity IDs, select Sets.
12. Click on Sets and select barrier_surface and bumper_surface and click OK.
Step 7: Create self impact contact between parts of the bumper

1. Right-click in the Solver browser and select Create > INTER > TYPE7. The Entity Editor opens.
2. For Name, enter impact.
3. For Grnod_id (S) (slave entity), set the selector to Components.
4. Click Components, select bumper, interior crashbox (left and right) and exterior
crashbox (left and right) and click OK.
5. For Surf_id (M) (master entity), set the selector to Set.
6. Click Set, select barrier_bumper_surface and click OK.
7. Set Igap to 2.
8. For the static coefficient Fric, enter 0.15.

Step 8: Create a system that specifies the location and the cross section
plane normal
1. Click the numbering icon on the toolbar.

2. Click the nodes selector and select by id.
3. For the IDs enter 6224, 6227, and 5993.
4. Check the display check box on.
5. Click on.
Note: Node numbers will appear next to the node for selection in
further steps.
6. From the Analysis page, click systems.

7. Go to the create by node reference page.
8. Select Node ID 6224 for origin node.
9. Select Node ID 6227 for z- axis.
10. Select Node ID 5993 for yz plane.
11. Click create to create a system.
12. Click the Card Edit icon on the toolbar.
13. Set the entity selector to systs.

14. Select the system and click edit.
15. Change the option from Skew to Frame.
16. Click return.
Step 9: Create a set of elements that will contribute to the cross-sectional

force results
1. Right-click in the Solver browser and select Create > GRSHEL > SHEL. The Entity Editor
opens.
2. For Name, enter CrosssectionPlane-elements.
3. For Entity IDs, toggle to Elements selector active, select two rows of element on either side of
the system, as shown in figure below.
Step 10: Define a section

1. Right-click in the Solver browser and select Create > SECT > SECT.
2. For Name, enter Crosssection_Plane.
3. For Frame_ID, select the system defined in the previous step by clicking on the screen.
4. For grshel_ID, select the set CrosssectionPlane-elements which is defined in previous step,
as shown below.

Step 11: Select the section for time history output
1. Right-click in the Solver browser and select Create > TH > SECTIO.
2. For Name, enter Section_force.
3. For Entity IDs, toggle Crosssections and select Crosssection_Plane.
4. For NUM_VARIABLES, select 1 and for Data: Var, enter DEF.
This selects the default output for RADIOSS.

Step 12: Create a BOX/RECTA and GRNOD/BOX containing the nodes
making up the barrier and bumper’s left side
These nodes will be slave to the rigid wall.
1. Right-click in the Solver browser and select Create > BOX > BOXRECTA.
2. For Name, enter half model.
3. Optionally, select a Color.
4. Enter Corner1 and Corner2 X, Y and Z coordinates, as shown below:
5. Right-click in the Solver browser and select Create > GRNOD > BOX.
6. For Name, enter RigidwallSlave_grnodbox.
7. For Entity IDs, set the selector to Box and select the above created half model (BOX/RECTA).
Step 13: Define a rigid wall

1. Press the F8 key to enter the create nodes panel.
2. Select the XYZ ( ) subpanel.

3. For x=, y= and z=, enter the values –600, -750 and 90, respectively.

4. Click create.
5. Right-click in the Solver browser and select Create > RWALL > PLANE.
6. For Name, enter wall.
7. Set Geometry type as Infinite Plane.
8. With the Base node selector active, select the node that was created in step 4.
9. Set Normal to 1,0,0.
10. For grnod_id1 (S), toggle Set and select RigidWallSlave_grnodbox (GRNOD/BOX).
11. For fric, specify 1.0 for the friction coefficient.


Value
CONTROL CARDS TITLE TITLE Bumper_Impact
ENGINE KEYWORDS RUN Tstop 20
ENGINE KEYWORDS PARITH Status [Checked]
ENGINE KEYWORDS PARITH Keyword2 ON
ENGINE KEYWORDS TFILE Status [Checked]
ENGINE KEYWORDS TFILE Time 0.1

Frequency
ENGINE KEYWORDS ANIM/BRICK/TENS Status [Checked]
ENGINE KEYWORDS ANIM/BRICK/TENS STRESS [Checked]
ENGINE KEYWORDS ANIM/BRICK/TENS STRAIN [Checked]
ENGINE KEYWORDS ANIM/SHELL/TENS/STRESS Status [Checked]
ENGINE KEYWORDS ANIM/SHELL/TENS/STRESS MEMB [Checked]
ENGINE KEYWORDS ANIM/SHELL/TENS/STRAIN Status [Checked]
ENGINE KEYWORDS ANIM/SHELL/TENS/STRAIN MEMB [Checked]

Value
ENGINE KEYWORDS ANIM/VECT DISP [Checked]
ENGINE KEYWORDS ANIM/DT Tfreq 1
ENGINE KEYWORDS DT/NODA Status [Checked]
ENGINE KEYWORDS DT/NODA CST 0 – Tmin 3.6e-4

export to.
3. Enter the name bumper_impact and click Save.
5. Toggle Merge starter and engine file to export the engine file with the model file (or export
separately).
6. Click Export to export both model and engine file.

2. For Input file, browse to the exercise folder and select the file bumper_impact_0000.rad.

Step 17 (Optional): View the results in HyperView
The exercise is complete. Save your work to a HyperMesh file.

RD-3550: Simplified Car Front Pole Impact
This tutorial demonstrates how to simulate frontal pole test with a simplified full car.
Model Description
 Simulation time: Engine file (_0001.rad) [0 – 0.0601 ms]
 An initial velocity of 15600 mm/s is applied on the car model to impact a rigid pole of radius
250 mm.
 Elasto-plastic Material /MAT/LAW2 (Windshield)
[Rho_I] Initial Density = 2.5x10-9 ton/mm3

[a] Yield Stress = 192 MPa
[b] Hardening Parameter = 200 MPa
 Elasto-plastic Material /MAT/LAW2 (Rubber)
[Rho_I] Initial Density = 2x10-9 ton/mm3
[a] Yield Stress = 1e30 MPa
[n] Hardening Exponent = 1

 Elasto-plastic Material /MAT/LAW2 (Steel)
[Rho_I] Initial Density = 7.9x10-9 ton/mm3
[a] Yield Stress = 200 MPa
[b] Hardening Parameter = 450 MPa
[SIG_max] Maximum Stress = 425 MPa
Exercise

2. From the Preferences menu, select the User Profiles or click the icon on toolbar.

1. Click the Open Model icon to open the fullcar.hm file you saved to your working directory
2. Click Open.
The model loads into the graphics area.
Step 3: Create and assign the material for the windshield components
2. For Name, enter windshield.
3. Set Card Image as M2_PLAS_JOHNS_ZERIL and click Yes to confirm.

5. In the Model browser, select components COMP-PSHELL_3 and COMP-PSHELL_16.
6. Click Mat_Id in the EE, select the material windshield and click OK to update the selected
components with the created material.
Step 4: Create and assign the material for the rubber components
1. In the Model browser, right-click and select Create > Material. The Entity Editor is displayed.
2. For Name, enter rubber.
3. Set Card Image to M2_PLAS_JOHNS_ZERIL and click Yes to confirm.
5. In the Model browser, select components COMP-PSHELL_20 through COMP-PSHELL_23.

6. For Mat_Id, select the material rubber and click OK to update the selected components with
the created material.

Step 5: Create Steel material and assign to all other parts
1. In the Model browser, right-click and select Create > Material. The Entity Editor is displayed.
2. For Name, enter steel.
3. Set Card Image to M2_PLAS_JOHNS_ZERIL.
5. In the Model browser select all components labeled with COMP-PSHELL and COMP-PROD,
except COMP-PSHELL_3, COMP-PSHELL_16 and COMP-PSHELL_20 to COMP-PSHELL_23.
6. For Mat_Id, select the material steel and click OK to assign the material to the selected
components.
Step 6: Create a Rigid Wall

1. In the Model browser, right-click and select Create > Rigid Wall. The Entity Editor is displayed.
2. For Name, enter ground.
3. Set Geometry type as Infinite plane.
4. Click Base node and select 'any node' from the model.
5. Define the normal vector Z = -1.
6. Set distance d = 300.

7. Go to the Analysis > rigid walls panel.
8. Move to the geom page.
9. Click name and select Ground from the list.
10. Click the edit tab besides base node and change values of the coordinates as indicated below.
X = -2300, Y = 1200, and Z = -1.
11. Click update > return.
Step 7: Create a Cylindrical Rigid Wall to represent pole

1. In the Model browser, right-click and select Create > Rigid Wall. The Entity Editor will display.
2. For Name, enter pole.
3. Set the Geometry type as Cylinder.
4. Click Base node and select ‘any node’ from the model.
5. Define the normal vector Z= 1.
6. For Radius node, do not select anything. Leave it as <Unspecified>.
7. Set distance d= 1500.

8. Go to Analysis > Rigid Walls panel.
9. Move to the geom page.
10. Click name and select Pole from the list.
11. Click the edit tab besides base node and change values of the coordinates as indicated below.
X = -320, Y = 1250, and Z = 0.
12. Set Radius = 250.
13. Click update > return.
Step 8: Defining Contact using TYPE 7 interface (Self Contact)

1. Hide all the 1D (TRUSSES) and 3D (SOLID) parts in the model by going to the Solver browser
PROP > SHELL, Isolate only. Return to the Model browser.
2. In the Model browser, right-click and select Create > Contact. The Entity Editor will display.
3. For Name, enter CAR_CAR.
5. For Surf_id (M) (master entity), set the option to Components and select displayed
components and click OK.
6. Input other parameters, as shown below.

Step 9: Defining Contact using TYPE 7 interface between Engine and
Radiator
1. In the Solver browser, right-click and select Create > SURF_EXT > PART.
2. For Name, enter engine.
3. Click on Components and select COMP-PSOLID_24.
4. In the Model browser, right-click and select Create > Contact.
5. For Name, enter ENGINE_RADIATOR and set the Card Image as TYPE7 and click Yes to
confirm.
6. For Grnod_id (S) (slave entity), set the selector switch to Components and click
Components, select COMP-PSOLID_26.
7. For Surf_id (M) (master entity), set the selector switch to Set and click Set, select engine.

Step 10: Defining initial velocity
1. Click Tools > BCs Manager to start the BCS Manager.
2. For Name, enter 35MPH, set Select type as Initial Velocity and set GRNOD to Parts.
3. Click comps and select all of the parts in the model.
4. Set the Vx as 15600.

5. Click Create to create the boundary condition and boundary condition appears in the table.
6. Click Close.
Step 11: Create Time History Nodes

1. In the Model browser, isolate COMP-PSHELL_19.
2. Click Tools > Create Cards > TH > NODE.
3. For Name, enter RAIL and select nodes on the Rail, as shown below.

4. For NUM_VARIABLES, select 1 and for Data: Var, enter the following:

2. Right-click in the Solver browser general area to create the cards shown below with the given
Keyword Type Keyword Parameter Parameter Value
CONTROL CARDS TITLE TITLE Car_Analysis
ENGINE KEYWORDS RUN Run Number 1
ENGINE KEYWORDS TFILE Time Frequency 9e-5
ENGINE KEYWORDS ANIM/ELEM HOURG [Checked]

ENGINE KEYWORDS ANIM/VECT FOPT [Checked]
ENGINE KEYWORDS ANIM/DT Tfreq 0.003

2. Enter a filename in the destination directory where you want to export to.
3. Enter the name FULLCAR and click Save.
5. Click Merge starter and engine file to export the engine file with the model in one file.
6. Click Export to export both model and engine file.

2. For Input file, browse to the exercise folder and select the file FULLCAR_0000.rad.


RD-3560: Bottle Drop
This tutorial demonstrates how to simulate a Bottle Drop Test containing water and air. The
objective is to evaluate the diffusivity of water and air in the bottle on drop.
Exercise

Step 2: Load the bottle.hm file

1. Click the Open Model icon to open the bottle.hm file you saved to your working directory

Step 3: Define Materials for Air and Water
2. For Name, enter Air.
3. For Card Image, select M37_BIPHAS and click Yes to confirm.
4. Input the values, as shown below. Remember to select ALE under ALE CFD Formulation.
5. Similarly create a material with the name Water using Steps 3.1 to 3.4.
6. Input the values, as shown below.

Step 4: Load stress-strain curve from a file
To create the material for bottle (plastic) you need a stress strain curve that is available in a file
from test.
1. Click XYPlots > Create > Plots.
2. Enter the plot= name as stress-strain and click create plot > return.
3. Click XYPlots > Edit > Curves.
4. Toggle the create radio button.
5. Click the load button to load the stressstrain_curve.txt file.
6. With the x radio button selected, click the green + to the right of comp= and set it to x.
7. Select the y radio button, click the green + to the right of comp= and set it to y.
8. Click create > return.
9. In the Model browser, click on curve.

10. In the Entity Editor, rename it as stress_strain. The data in the file is loaded as a curve in
HyperMesh.

Step 5: Define Material for Bottle
2. For Name, enter Bottle.
3. For Card Image, select M36_PLAS_TAB and click Yes to confirm.
Select the stress-strain curve created for fct_ID1.
Step 6: Define property and assign material for Air

2. For Name, enter Air.
3. For Card Image, select P14_SOLID and click Yes to confirm.

4. Enter parameters, as shown below.
5. In the Model browser, click on the air component.

6. Select material and property created for Air in the Entity Editor.
Step 7: Define property and assign material for Water

2. For Name, enter Water.

5. In the Model browser, click on the water component.
6. Select material and property created for Water in the Entity Editor.
Step 8: Define property and assign material for Bottle

3. For Card Image, select P1_SHELL.
N=5
Thick = 0.3
5. In the Model browser, click on the bottle component.
6. Select material and property created for Bottle in the Entity Editor.

Step 9: Define an Interface between Bottle and Water
1. In the Model browser, right-click and select Create > Set.
2. For Name, enter ALE_Surf.
3. Set Card Image to SURF_EXT and click Yes to confirm.
4. For Entity IDs, set the entity selector to Components.
5. Click Components and select water and air.
6. Click OK to complete the selection.

8. For Name, enter Bottle_Water, and for Card Image, select TYPE1.
9. For ls2(S) (slave entity), set the selector to Set.
10. In the Select Set dialog, select ALE_surf and click OK.
11. For ls1(M) (master entity), set the selector to Components.
12. In the Select Components dialog, select Bottle and click OK.

Step 10: Create Initial Velocity for Bottle
2. Set the Select type to Initial Velocity.
4. Click Parts and bottle.
5. Set the Vz velocity to -5468.200 (Negative direction indicating opposite to Global Z-axis).
6. Click Create to create the imposed velocity boundary condition.
Step 11: Create Initial Velocity for Water and Air

1. Set the Select type to Initial Velocity.
2. For Name, enter Liquid.
3. Click Parts and select water and air.
4. Set the Vz velocity to -5468.200 (Negative direction indicating opposite to Global Z-axis).

5. Click Create to create the imposed velocity boundary condition.
6. Select the Liquid initial velocity in the table, right-click and select Card Edit.
7. Change the Type to T+G and click return to complete the definition.
Step 12: Create Rigid Wall

1. In the graphics area, press the F8 key, and create the node at the coordinates: X= 0, Y= 0, Z=
-50 and create node.
3. For Name, enter GROUND with Geometry type as Infinite plane.
4. Select node created in Step 12.1 as base node and make sure the normal vector is in the z-
direction, as shown below.
5. Set the d to 250.0.

\

CONTROL CARDS TITLE TITLE Bottle_drop
CONTROL CARDS MEMORY Status [Checked]
CONTROL CARDS MEMORY NMOTS 40000
CONTROL CARDS SPMD Status [Checked]
CONTROL CARDS IOFLAG Status [Checked]
CONTROL CARDS ANALY Status [Checked]
ALE-CFD-SPH ALE_CFD_SPH_CARD Status [Checked]
ALE-CFD-SPH ALE_CFD_SPH_CARD ALE_Grid_Velocity [Checked –]

ENGINE KEYWORDS RUN Tstop 1.5e-2
ENGINE KEYWORDS PARITH Keyword2 OFF
ENGINE KEYWORDS TFILE Time Frequency 0.00015
ENGINE KEYWORDS ANIM > ANIM/ELEM Status [Checked]
ENGINE KEYWORDS ANIM > ANIM/ELEM VONM [Checked]
ENGINE KEYWORDS ANIM > ANIM/ELEM PRES [Checked]
ENGINE KEYWORDS ANIM > ANIM/DT Status [Checked]
ENGINE KEYWORDS ANIM > ANIM/DT Tstart 0
ENGINE KEYWORDS ANIM > ANIM/DT Tfreq 1.5e-3
ENGINE KEYWORDS DT > DT Status [Checked]
ENGINE KEYWORDS DT > DT Tscale 0.5
ENGINE KEYWORDS DT > DT Tmin 0.0

1. Click File > Export Solver Deck or click the Export Solver Deck icon .
export to.
3. For Name, enter bottle and click Save.
5. Click Merge starter and engine file to export all the data in one file (or export separately).
6. Click Export to export solver deck.

2. For Input file, browse to the exercise folder and select the file bottle_0000.rad.


RD-3580: Boat Ditching
Boat Ditching with Boundary Elements

The objective of this tutorial is to simulate Boat Ditching with Boundary Elements to represent
continuous water using bi-phase material law (Law 37). In this model, the top chamber is air, lower
chamber is water surrounded by boundary elements. Law 37 is used for air, water and boundary.
Boundary conditions are applied on each surface of boundary in the normal direction. An interface
between fluid and boat (CEL) is defined to manage the contact.
Exercise

Step 2: Load the boat_ditching_1.hm file

1. From the toolbar, click the Open Model icon to open the boat_ditching_1.hm file you saved
to your working directory from the radioss.zip file. Refer to Accessing the Model Files.
Step 3: Define and assign Material, Property to component AIR

1. In the Model browser, right-click and select Create > Material. The new material shows up in
the Entity Editor (EE).
2. For Name, enter air.
3. For Card Image, select M37_BIPHAS.
Remember to select ALE under ALE CFD Formulation.

5. Create a new property named Air with a Card Image of P14_SOLID by right-clicking in the
Model browser.
6. Click on the component Air and assign Air as the Prop_Id and air as the Mat_Id in the Entity
Editor.
Step 4: Define and assign Material, Property to component WATER

2. For Name, enter water.

5. In the Model browser, create a new property named Water with a Card Image of P14_SOLID.
6. Click on the component Water and assign Water as the Prop_Id and water as the Mat_Id in
the Entity Editor.
Step 5: Define and assign Material, Property to component BOAT

2. For Name, enter boat.
3. For Card Image, select M1_ELAST.

5. In the Model browser, create a new property named Boat with a Card Image of P1_SHELL and
assign the new property with the values shown below:
6. Click on the component Boat and assign Boat as the Prop_Id and boat as the Mat_Id in the
Entity Editor.
Step 6: Define and assign Material, Property to component Air-BC

the Entity Editor.
2. For Name, enter air-bc.

5. Click on the component Air-BC and assign Air as the Prop_Id and air-bc as the Mat_Id in
the Entity Editor.
Step 7: Define and assign Material, Property to component Water-BC

the Entity Editor.
2. For Name, enter water-bc.

5. Click on the component Water-BC and assign Water as the Prop_Id and water-bc as the
Mat_Id in the Entity Editor.
Step 8: Define an Interface between Boat and Fluid

1. Click Tools > Create Cards > ALE-CFD-SPH > INTER_TYPE18. The new interface opens in
the Entity Editor.
2. For Name, enter Boat-Fluid.
3. Enter the parameter values, as shown below for Stfval and GAP.
4. Set the Surf_id (M) for master selection to Components and select the boat component.
5. Set the Grnod_id (S) for slave selection to Components and select all the components, except
boat.

Step 9: Create RBODY for the Boat and assign mass to the Master Node
1. Isolate the boat part using the Model browser.
2. From the pull-down menu, select Tools > Rbody Manager.
3. For Title:, enter RIGID-BOAT, verify that Master node: is set to Calculate Node and set Slave
node(s): to Parts and select the Boat.
4. Click Create to create the RBODY. The created RBODY appears in the table.
5. Select the created RBODY in the table and right-click and select Edit card to open the card
image panel.
6. Assign a mass of 23.04 kg to the boat.
7. Click return to return from the card image panel.
8. Click Close to close the RBODY Manager.
Step 10: Create Initial Velocity

2. For Name, enter Boat.
3. For Select type, select Initial Velocity.

4. Set GRNOD to Nodes.
5. Click the Node tab and select the master node of the RBODY created in the previous step.
6. Set Z velocity (VZ) to -11.0 indicating velocity opposite to global Z-axis.
7. Click Create to create the initial velocity boundary condition.
Step 11: Create Boundary Conditions on outermost faces

1. In the Model browser, right-click on the Components sub-folder and select Show to display all
components.
2. Enter a new boundary condition in the BCs Manager named Constraint-x.
3. For Select type, select Boundary condition.
5. Click the Node tab and select a node on both faces normal to x-axis.
6. Then click the nodes yellow tab and select By face.
HyperMesh will automatically select nodes on the face, as shown in figure.

7. Check Tx box to constraint translation in X direction.
8. Click Create to create the constraint.

9. Follow the same procedure (step 1-8) but create constraint in Y direction on the sides parallel to
Y plane of global axis.
10. Follow the same procedure (step 1-8) but create constraint in Z direction on the sides parallel to
Z plane of global axis.
Step 12: Creating control cards and output requests

CONTROL CARDS TITLE TITLE Boat-Ditch-1
ALE-CFD-SPH ALE_CFD_SPH_CARD ALE_Grid_Velocity [Checked]
ENGINE KEYWORDS RUN RunName Boat-Ditch-1

ENGINE KEYWORDS ANIM/ELEM DENS [Checked]
ENGINE KEYWORDS ANIM/ELEM PRES [Checked]

export to.
3. For name, enter boatditching_1 and click Save.
5. Click Merge starter and engine file to export one solver deck (or export separately).

2. For Input file, browse to the exercise folder and select the file boatditching_1_0000.rad.


Boat Ditching without Boundary Elements
The objective of this tutorial is to simulate Boat Ditching without Boundary Elements. So there is no
boundary to represent continuous water. Basically, you are simulating Boat-Ditching in an enclosed
volume. In this model, the top chamber is air (including its outer layer) and the lower chamber is
water (including its outer layer). Bi-Phase material Law 37 was used to model air and water.
Boundary conditions are applied on each surface of boundary in the normal direction. An interface
between fluid and boat (CEL) is defined to manage the contact.
Exercise

Step 2: Load the boat_ditching_2.hm file

1. From the toolbar, click the Open Model icon to open the boat_ditching_2.hm file you saved
to your working directory from the radioss.zip file. Refer to Accessing the Model Files.
Step 3: Define and assign Material, Property to component AIR

1. In the Model browser, right-click and select Create > Material. The new material shows in the
Entity Editor.
3. For Card Image, select M37_BIPHAS and click Yes to confirm.

5. In the Model browser, create a new property named Air with a Card Image of P14_SOLID.
6. Click on the component Air and assign as the Prop_Id and air as the Mat_Id in the Entity
Editor.
Step 4: Define and assign Material, Property to component WATER

Entity Editor.
2. For Name, enter water.

5. In the Model browser, create a new property named Water with a Card Image of P14_SOLID.
6. Click on the component Water and assign Water as the Prop_Id and water as the Mat_Id in
the Entity Editor.
Step 5: Define and assign Material, Property to component BOAT

Entity Editor.
2. For Name, enter boat.
4. Input the values, as below.

5. In the Model browser, create a new property named Boat with a Card Image of P1_SHELL and
assign the new property with the values shown below:
6. Click on the component Boat and assign Boat as the Prop_Id and boat as the Mat_Id in the
Entity Editor.
Step 6: Define an Interface between Boat and Fluid

1. Click Tools > Create Cards > ALE-CFD-SPH > INTER_TYPE18. The new interface opens in
the Entity Editor.
2. For Name, enter Boat-Fluid.

3. Enter the parameter values, as shown below for Stfval and GAP.
3. Set the Surf_id (M) for the master selection to Components and select the boat component.
4. Set the Grnod_id (S) for the slave selection to Components and select all the components,
except boat.
Step 7: Create RBODY for the Boat and assign mass to the Master Node
1. In the Model browser, isolate the boat part.
2. From the pull-down menu, select Tools > Rbody Manager.
3. For Title, enter RIGID_BOAT. Verify that the Master node is set to Calculate Node and set the
Slave node(s) to Parts and select the Boat.

4. Click Create to create the RBODY. The created RBODY appears in the table.
5. Select the created RBODY in the table and click Edit Card to open the Card Image panel.
6. Assign a mass of 23.04 kg to the boat.
7. Click return to return from the Card Image panel.
8. Click Close to close the RBODY Manager.
Step 8: Create Initial Velocity

1. Click BCs Manager in the Utility panel or click Tools > BCs Manager.
2. For Name, enter Boat.
3. For Select type, select Initial Velocity.
5. Click the Node tab and select the master node of the RBODY created in the previous step (ID:
690501).
6. Set Z velocity (VZ) to -11.0, indicating velocity opposite to global Z-axis.

7. Click Create to create the initial velocity boundary condition.
Step 9: Create Boundary Conditions on outermost faces

1. In the Model browser, right-click on the Components subfolder and select Show to display all
components.
2. Enter a new boundary condition in the BCs Manager, named Constraint-x.
3. For Select type, select Boundary Condition.
5. Click the Node selector and select a node on both faces normal to x-axis.
6. Click the nodes selector and select by face. HyperMesh will automatically select nodes on the
face, as shown in figure.

7. Check Tx box to constraint translation in X direction.

9. Follow the same procedure (step 1-8) but create constraint in Y direction on the sides parallel to
Y plane of global axis.
10. Follow the same procedure (step 1-8) but create constraint in Z direction on the sides parallel to
Z plane of global axis.

CONTROL CARDS TITLE TITLE Boat-Ditch-2
ENGINE KEYWORDS RUN RunName Boat-Ditch-2
ENGINE KEYWORDS ANIM > ANIM/ELEM Status [Checked]
ENGINE KEYWORDS ANIM > ANIM/ELEM VONM [Checked]
ENGINE KEYWORDS ANIM > ANIM/ELEM DENS [Checked]
ENGINE KEYWORDS ANIM > ANIM/ELEM PRES [Checked]

ENGINE KEYWORDS ANIM > ANIM/VECT Status [Checked]
ENGINE KEYWORDS ANIM > ANIM/VECT VEL [Checked]
ENGINE KEYWORDS ANIM > ANIM/VECT CONT [Checked]
ENGINE KEYWORDS ANIM > ANIM/DT Status [Checked]
ENGINE KEYWORDS ANIM > ANIM/DT Tstart 0
ENGINE KEYWORDS ANIM > ANIM/DT Tfreq 1.0

export to.
3. For name, enter boatditching_2 and click Save.
5. Click Merge starter and engine file to export the one solver deck (or export separately).

2. For Input file, browse to the exercise folder and select the file boatditching_2_0000.rad.


RD-3590: Fluid Flow through a Rubber Clapper Valve
The objective of this tutorial is to simulate the flow of water through a rubber valve using an inlet
option in multi-phase material law (Law 51). In this model the top chamber is air, the lower
chamber is water, and the bottom row of elements is the inlet. Law 51 is used for air, water and
inlet. Boundary conditions are applied on each surface of fluid in its normal direction. An interface
between fluid and rubber (CEL) is defined to manage the contact.
Exercise

Step 2: Load the valve.hm file

1. From the toolbar, click the Open Model icon to open the valve.hm file you saved to your
Step 3: Creating curves for pressure_inlet

2. In the Solver browser, right-click and select Create > FUNCT. The Curve editor dialog box
opens.
3. In the Curve editor window, click New.
4. For the Name, enter pressure_inlet and click proceed.
5. From the Curve editor window, select pressure_inlet from the curve list.
6. Enter the X and Y coordinates, as shown below.
7. Click Update.

8. Follow Steps 3.3 - 3.7 to create a curve named density, with the values shown below.
9. Click Close.
Step 4: Define and assign Material, Property to component inlet

1. In the Model browser, right-click and select Create > Material. The new material appears in
the Entity Editor.
2. For Name, enter inlet-water.
3. For Card Image, select MLAW51 and click Yes to confirm.

5. In the Model browser, right-click and select Create > Property to create a new property.
6. For Name, enter solids.
7. For Card Image, select P14_SOLID. Keep all the default settings.
8. Click Yes to confirm.
9. In the Model browser, click on the inlet component and assign solids as the Prop_Id and
inlet-water as the Mat_Id.
Step 5: Define and assign Material, Property to component Air

the Entity Editor.
3. For Card Image, select MLAW51 and click Yes to confirm.

5. Click on the air component in the Model browser and assign solids as the Prop_Id and air as
the Mat_Id.
Step 6: Define and assign Material, Property to component Water

1. In the Model browser, right-click on the material air and click Duplicate. Edit the material
parameters and table data with the following changes.
2. Change the Name to water.
3. Set C0(1) to 1.0e-04.

4. Change the value for Alpha(1) to 1.0 and Alpha(2) to 0.0.
5. Change Rho_Initial to 1.000e-06.
6. In the Model browser, right-click on the water component and select Assign. Assign solids as
the Prop_Id and water as the Mat_Id.
Step 7: Define and assign Material, Property to component Rubber

4. Enter the following properties:
Rho_Initial = 1e-6 kg/mm3
E = 0.7
Nu = 0.4
7. For Card Image, select P14_SOLID.
8. Set ISOLID to 12.
9. In the Model browser, right-click on the rubber component and select Assign. Assign rubber
as the Prop_Id and rubber as the Mat_Id.
Step 8: Define an Interface between Rubber and Fluid

1. Open the Solver browser and right-click to select Create > ALE-CFD-SPH > INTER_TYPE18.
2. For Name, enter rubber-fluid, and for Card Image, select TYPE18.

3. To set the Surf_id (M), change the selector to Components and select the rubber component.
4. To set the Grnod_id (S), change the selector to Components and select all the comps, except
rubber.
Step 9: Create Boundary Conditions on outermost faces of solid comps

2. For Name, enter constraint-X, set Select type as Boundary Condition and set the GRNOD
to Nodes.
3. Click Nodes and select a node for each outer face parallel to x-axis.
4. Click Nodes in the panel and select by face.
HyperMesh will automatically select all nodes in the face.

5. Click Create.
6. Repeat Steps 9.1 to 9.5 to create Boundary conditions on Y and Z faces (see image below for
reference).

7. Check the box Ty in order to constrain the translational d.o.f in Y-direction, as shown below:
Boundary conditions for Y-axis
8. Check the box next to Tz in order to constrain the translational d.o.f in Z-direction, as shown
below:
Boundary conditions for Z-axis

Step 10: Create Boundary Condition to fix one end of the rubber
1. For Name, enter Fix-rubber, set Select type to Boundary Condition and set the GRNOD to
Nodes.
2. Select all the nodes on the edge of the clapper, as shown below.
3. Constraint all the translational degree’s of freedom.

Step 11: Create output requests a control card

Value
CONTROL CARDS TITLE TITLE CLAPPER
ALE-CFD-SPH ALE_CFD_SPH_CARD GridVel_Gamma 100.00
ALE-CFD-SPH ALE_CFD_SPH_CARD GridVel_Cwx 1.00
ALE-CFD-SPH ALE_CFD_SPH_CARD GridVel_Cwy 1.00
ENGINE KEYWORDS RUN RunName CLAPPER

Value
ENGINE KEYWORDS DT Status [Checked]
ENGINE KEYWORDS DT Tscale 0.5
ENGINE KEYWORDS DT Tmin 0.0

export to.
3. For Name, enter clapper and click Save.
5. Click Merge starter and engine file to export solver deck as one file (or export separately).
6. Click on Export to export solver deck.

1. Launch Start > Programs > Altair HyperWorks 2017 > RADIOSS.

2. For Input file, browse to the exercise folder and select the file clapper_0000.rad.

RD-3595: Three Point Bending with HyperMesh
This tutorial demonstrates how to set up 3-point bending model with symmetric boundary
conditions in Y direction.
Model Description
 Simulation time: in Engine file [0 – 6.601e-002 s]
 Only one half of the model is modeled because it is symmetric.
 The supports are totally fixed. An imposed velocity of 1000 mm/s is applied on the Impactor
in the (–Z) direction
 Model size = 370mm x 46.5mm x 159mm
 Honeycomb Material /MAT/LAW28: HONEYCOMB
[Rho_I] Initial density = 3.0e-10 ton/mm3

[E11], [E22] and [E33] Young’s modulus (E ij) = 200 MPa
[G11], [G22] and [G33] Shear modulus (Gij) = 150 MPa
 Elasto-Plastic Material /MAT/LAW36: Inner, Outer and Flat
[Rho_I] Initial density = 7.85-9 ton/mm3


 Strain Curve:
0 1 2 3 4 5 6 7 8 9
STRAIN 0 0.012002 0.014003 0.018003 0.022002 0.026003 0.030006 0.032 0.033005 0.033523
STRESS 325 335.968 343783 349.245 358.649 372.309 383.925 388.109 389.292 389.506
 Elastic Material /MAT/PLAS_JOHNS: Impactor
[Rho_I] Initial density = 8e-9 ton/mm3

Exercise

Step 2: Retrieve the RADIOSS file

2. Click the Select File icon to open the BENDING_0000.rad file you saved to your working
3. Click Import.
Step 3: Create and Assign material and property for HCFOAM

the Entity Editor.
2. For Name, enter Foam.
3. For Card Image, select M28_HONEYCOMB and click Yes to confirm.
4. Input values, as shown below:

6. For Name, enter Foam and set the new property Card Image as P14_SOLID. Leave all the
settings as default, except for ISOLID which should be set to 24.
7. In the Model browser, right-click on the component HCFoam and select Assign. Assign Foam as
the Prop_Id and Foam as the Mat_Id.
8. Click Apply.

Step 4: Create and Assign material and property for the component Inner
the Entity Editor.
2. For Name, enter Inner.
3. For Card Image, select M36_PLAS_TAB and click Yes to confirm.
6. For Name, enter Inner and set Card Image as P1_SHELL. Leave all the settings as default,
except for Ishell which should be set to 4 and Thick which should be set to 9.119e-01.
7. In the Model browser, right-click on the component Inner and select Assign. Assign Inner as
the Prop_Id and Inner as the Mat_Id.
Step 5: Create and Assign material and property for the component Outer
1. In the Model browser, right-click on the material Inner and select Duplicate. Name the new
material Outer. This creates a new material that is identical to the source material.
2. In the Model browser, right-click on the property Inner and select Duplicate. Name the new
property Outer. This creates a new property that is identical to the source property.

3. In the Model browser, right-click on the component Outer and select Assign. Assign Outer as
the Prop_Id and Outer as the Mat_Id.
Step 6: Create and Assign material and property for the component Flat
Follow the procedure described in Step 5 with Outer replaced by Flat.
Step 7: Create and assign material and property for Impactor

the Entity Editor.
2. For Name, enter Impactor.
5. In the Model browser, right-click on the property Inner and select Duplicate. Name the new
property Impactor. This creates a new property that is identical to the source property.
6. In the Model browser, right-click on the component Impactor and select Assign. Assign
Impactor as the Prop_Id and Impactor as the Mat_Id.
Step 8: Create and assign material and property for Support

Follow the same procedures as in Step 5. Create a copy of Impactor property and material with
name support and assign it to component support.
Step 9: Create a rigid body to make Impactor and Support Rigid

2. For Name, enter Impact rigid.

3. Select any color for easy visualization.
4. Set Card Image to None.
5. Go to the 1D page, select the rigids panel.
6. Verify that you are in the create subpanel.
7. For dependent switch to comps.
8. For primary node switch to calculate node.
9. Click comps.
10. Select Impactor, then click select.
11. Click create.
13. Similarly, create rigid body for Support component in a collector with the name “Support
rigid” using Steps 9.1 to 9.12.
Step 10: Define imposed velocity and boundary condition for the impactor
1. From the Utility page, start the BCs Manager.
2. For Name, enter IMPOSED_VELOCITY, set Select type to Imposed Velocity and set the
GRNOD to Nodes.
3. Click nodes and select the master node of the rigid body of the Impactor, as shown in the
following image.

4. Set the Direction as Z.
5. Set Scale Y to -1000.0 as the direction of velocity is opposite to the global Z-axis.
6. Set the Curve ID to Select curve.
7. Select the predefined curve to Func1.
8. Click create to create the imposed velocity.

9. For Name, enter Impactor_constraints, set Select type to Boundary Condition and set the
GRNOD to Nodes.
10. Click nodes and select the master node of the rigid body.
11. Check all the degrees of freedom to constrain, except Tz.
12. Click create to create the boundary condition.
Step 11: Define fixed boundary condition for the support

2. For Name, enter Support_fixed, set Select type to Boundary Condition and set the GRNOD
to Nodes.
3. Select the master node of the rigid body created on Supporter, as shown in the following
image.
4. Check all the degrees of freedom.

Step 12: Define symmetry boundary condition for the foam, inner, outer
and flat
2. For Name, enter SYMMETRY_XZ, set Select type to Boundary Condition and set the GRNOD to
Nodes.
3. Select the nodes of the foam, inner, outer and flat, as shown in the following image.
4. Check the degrees of translational degrees of freedom Y and rotational degrees of freedom X
and Z to constraint.

6. Click close to exit the BC Manager.
Step 13: Define contacts between the beam and the support
2. In the Solver browser, right-click and select Create > INTER > TYPE7.
3. Enter the values, as shown below:

4. Set the Surf_id (M) for the master selection to Components and select the Support
component.
5. Set the Grnod_id (S) for the slave selection to Components and select the Flat component.

6. Similarly create the contact for Impactor with Outer, as shown below.

Step 14: Define the self contact between the beam components
1. Using the directions in Step 13, create a new Type 7 interface named Self with the components
Outer, Inner, and Flat as Master and the same components Outer, Inner, and Flat as Slave.
This will make the components self-contact instead of self-penetrate. Verify that the interface
has a Fric of 0.1 and Gapmin of 0.2.

Step 15: Create Interface time history
1. Right-click in the Solver browser and select Create > TH > INTER.
2. For Name, enter IMPACTOR.
3. Switch the entity selector to groups.
4. Click groups and select the interfaces Impactor and Support from the list.
5. Click OK.
6. Set NUM_VARIABLES to 1 and Data: Var to DEF.

CONTROL CARDS TITLE TITLE 3PBENDING
ENGINE KEYWORDS RUN RunName 3PBENDING
ENGINE KEYWORDS RUN RunNumber 1

ENGINE KEYWORDS RUN Tstop 7.01e-2
ENGINE KEYWORDS TFILE Time_frequency 0.0001
ENGINE KEYWORDS ANIM/DT Tfreq 2.5e-3
ENGINE KEYWORDS DT/NODA Status [Checked]
ENGINE KEYWORDS DT/NODA CST_0 [Checked]
ENGINE KEYWORDS DT/NODA/CST_0 Tscale 0.9
ENGINE KEYWORDS DT/NODA/CST_0 Tmin 7e-7
ENGINE KEYWORDS DT/NODA DEL [Checked]
ENGINE KEYWORDS DT/NODA/DEL Tscale 0.9
ENGINE KEYWORDS DT/NODA/DEL Tmin 3.5e-8
ENGINE KEYWORDS RBODY_ENGINE Status [Checked]

RBODY/ON

ENGINE KEYWORDS RBODY_ENGINE NUM_rbnodes 2
ENGINE KEYWORDS RBODY_ENGINE Data: Nodes 29664

29665

2. For File:, navigate to the destination directory where you want to export to.
3. For name, enter 3BENDING and click Save.
Step 18: Open RADIOSS Manager

2. For Input file, browse to the exercise folder and select the file 3PBENDING_0000.rad.
1. See if there are any warnings or errors in .out files.
2. Using HyperView, plot the displacement, strain contour and vectors.

von Mises Stress Contour (MPa)

Contact Force for Impactor Interface

RD-3597: Cell Phone Drop Test using HyperMesh
This tutorial demonstrates how to simulate a free fall of a cell phone due to gravity from a height of
1001mm using 2nd order tetra elements.
Model Description
 Simulation time: in Engine [0 – 3.3e-3]
 This is a very simple cell phone model used to demonstrate how to set up a drop test. The
model is an assembly of two solid parts meshed with Tetra 10 elements, connected with
spring elements, and contact defined between them.
 To reduce the simulation time, the cell phone is dropped 1 mm from the ground with an initial
velocity of -4429.4469 mm/s representing the velocity that it would have attained from a free
fall of 1000 mm.
 Boundary Conditions: Gravity load + initial velocity of -4429.4469 mm/s on the cell phone.

 Elasto-plastic Material /MAT/LAW36 (Plastic)
[Rho_I] Initial density = 1.16E-9 ton/mm3

[E] Young's modulus = 1000 MPa
STRAIN 0 16
STRESS 1 17
Exercise

Step 2: Load the cellphone.hm file

1. From the toolbar, click the Open Model icon to open the cellphone.hm file you saved to
your working directory from the radioss.zip file. Refer to Accessing the Model Files.
Step 3: Creating the material curve

1. Click XYPlots > Curve Editor.
3. For the curve name, enter stress_strain_curve.
4. Click proceed.
5. From the Curve editor window, select stress_strain_curve from the Curve List.
6. Enter the X and Y coordinates, as shown below.

7. Click Update > Close.
Step 4: Create material and properties for the cell phone parts
1. In the Model browser, right-click and select Create > Material to create a new material.
2. For Name, enter cell_phone.
3. For Card Image, select M36_PLAS_TAB and click Yes in the confirmation window.

5. Select N_func and set to 1.
6. Click fct_ID1 and select stress_strain_curve (the function curve previously created).
7. In the Model browser, right-click and select Create > Property to create a property.
8. For Name, enter cell_phone.
10. Set the variable I_tetra to a value of 1.

11. In the Model browser, expand the Components folder and highlight the components
Cellphone_bottom and Cellphone_top and right-click to Assign (or use the Entity Editor) the
newly created property and material.
Step 5: Create property for the spring links

2. For Name, enter spring.
3. Set Card Image to P13_SPR_BEAM and click Yes to confirm.
4. Enter the following values:
Mass (MASS): 2e-6 ton
Inertia (Inertia): 2e-4 mm4
Translation stiffness (K_Tensn, K_ShrY, and K_ShrZ): 50
Rotation stiffness (K_Tor, K_FlxY, and K_FlxZ): 100N
5. Click return to return to component panel.
6. In the Model browser, select the component Connection_springs and right-click Assign (or
use the Entity Editor) to assign the newly created property to the spring component.
Step 6: Define the Interface between cell phone parts

1. In the Model browser, right-click and select Create > Contact Surface.
2. For Name, enter self.
3. Click on Elements.
4. Switch from add shell elements to add solid faces.
5. Select elements by collector and select Cellphone_bottom and click select.

6. For face nodes, select nodes by collector and select cellphone bottom and click select >
add > return.
8. For Name, enter Self.
10. For Grnod_id (S), select nodes > by collector and select Cellphone_top and click select >
add and click return.
11. For Surf_id (M), switch to Contactsurf, click on Contactsurf and select self.
12. Click OK.
13. Set Fric to 0.1.
14. Set Gapmin to 0.3.

Step 7: Create a rigid wall
2. For Name, enter GROUND.
3. Set the Geometry type to Infinite plane.
4. Click in the graphics area and press the F8 key on the keyboard. Enter the node coordinates:
X=0, Y=0, and Z=19.
5. Click create.
7. In the Entity Editor, select the created node as Base node.
8. Make sure the normal vector is set to z-axis, as shown below.

9. For d, enter 50.
10. To review, go to the Solver browser, select the RWALL folder.
11. Right-click on GROUND and click Review.

12. Click return to exit from the panel.
Step 8: Define gravity load

1. In the Model browser, right-click and select Create > Set.
2. For Name, enter Gravity, set Card Image as GRNOD and click Yes to confirm.
3. Select Nodes of all three parts.
4. In the Model browser, right-click and select Create > Load Collector.
5. For Name, enter loadcol1, set Card Image as GRAV_Collector and click Yes to confirm.
6. Set Direction to Z.
7. For Grnod_id, select Gravity from the Select Set dialog and click OK.
8. Set scale_y to -9810.0 indicating gravity in opposite Z direction.

9. From the XYPlots pull-down, click Curve Editor.
11. For Name =, enter gravity.
12. Click proceed.
13. In the Curve editor window, select gravity from the Curve List.
14. Enter X and Y, as shown in the following image:

15. Click Update > Close to close the Curve editor window.
16. Back in Gravity load collector, update Ifunc to the curve just created.
Step 9: Apply an initial velocity to the Cell Phone

1. In the Model browser, right-click and select Create > Load Collector.
2. For Name, enter Initial_velocity, set Card Image to INIVEL_Collector.
3. For Grnod_id, select the same set (Gravity) previously used.
4. For Vz =, enter the value -4429.4469.

Step 10: Creating output request and control cards
CONTROL CARDS TITLE TITLE Cellphone_drop
CONTROL CARDS MEMORY NMOTS 40000 Not needed

ENGINE KEYWORDS RUN Tstop 3e-3
ENGINE KEYWORDS ANIM/DT Tstart 0.0
ENGINE KEYWORDS ANIM/DT Tfreq 2e-4

1. Click File > Export or click the Export Solver Deck icon .
3. For Name, enter Cellphone and click Save.


2. For Input file(s), browse to the exercise folder and select the file cellphone_0000.rad.
2. Using HyperView plot the strain and stress contour.
Von Mises Stress Contour (MPa)

Plastic Strain (mm/mm)

RD-3599: Gasket with HyperMesh
This tutorial demonstrates how to simulate a rubber gasket in sequential loading, given the
following load sequence:
 Translation Transverse (10 mm)
 Translation Longitudinal (5 mm)
 Torsion (20 Degrees)
Model Description
 Simulation time:
o Engine [0 – 1.501] in steps of 0.5 ms for each load case
 The outer circumference area is fixed on all degrees of freedom (V X, VY, VZ) and the center
node is fixed on X direction and the X and Y rotation (VX, WX, Wy)
 The gasket dimensions are: Thickness = 100 mm, External Diameter = 200 mm and Internal
Diameter = 50 mm.
 Hyper-Elastic Material /MAT/LAW42 (Rubber)
[Rho_I] Initial density = 6.0-6 Kg/mm3

[mue1] ( 1) = 0.6
[alfa1] ( 1) =2
(alfa2] ( 2) = -2

Exercise

Step 2: Load the gasket.hm file

1. From the toolbar, click the Open Model icon to open the gasket.hm file you saved to your
Step 3: Define and assign material, property to Rubber

1. In the Model browser, right-click and select Create > Material to create material.
3. For Card Image, select M42_OGDEN and click Yes in the confirmation window.

5. In the Model browser, right-click and select Create > Property to create property.
6. For Name, enter gasket.
8. In the Model browser, expand the Component folder and select GASKET. Right-click and Assign
(or use the Entity Editor) the newly created property and material.
Step 4: Create a component for the rigid body at center of Gasket

2. For Name, enter center and switch Card Image to None and click Yes to confirm.
3. Select any color for easy visualization.
Step 5: Create a rigid body at center of Gasket

1. From the 1D page, select the rigids panel.
2. For primary node, switch to calculate node.

3. For nodes 2-n, switch to multiple nodes.
4. Click the nodes and select a node in the inner face.
5. Click nodes and select by face. HyperMesh will select all nodes on the inner face.
6. Click create.
Step 6: Create gasket inner fixed boundary conditions

2. For Name, enter Inner_BC, set Select type to Boundary Condition and set the GRNOD to
Nodes.
3. Select the master node of rigid body created in Step 5 and click proceed.
4. Check the Tx translational and Rx, Ry rotational degrees of freedom.
5. Click Create to create the inner fixed boundary condition.

Step 7: Create gasket inner Y displacement boundary conditions
2. For Name, enter DISP_Y, set Select type to Imposed Displacement and set the GRNOD to
Nodes.
3. Select the master node of rigid body created in Step 5.
4. Set Direction as Y.
5. Click Create/Select curve to go to the XY curve editor.
6. Click New and enter Name as DISP_Y. Click proceed.
7. Enter the following values for X and Y:
X = {0, 0.5, 1.0}
Y = {0, 10, 10}
8. Click Update and Close the XY curve editor GUI.

9. Click Create to create the boundary condition.
Step 8: Create gasket inner Z displacement boundary conditions

2. For Name, enter DISP_Z, set Select type to Imposed Displacement and set the GRNOD to
Nodes.
4. Set Direction as Z.
6. Click New and enter Name as DISP_Z. Click proceed.
7. Enter the following vales for X and Y:
X = {0, 0.5, 1, 1.5}
Y = {0, 0, 5, 5}

Step 9: Create gasket inner Z rotation boundary conditions

2. For Name, enter ROT20DEG_Z, set Select type to Imposed Displacement and set the GRNOD
to Nodes.
4. Set Direction as ZZ.
6. Click New and enter Name as ROT20DEG_Z. Click proceed.
7. Enter the following vales for X and Y:
X = {0, 1, 1.5, 2}
Y = {0, 0, 0.349, 0.349}

Step 10: Create gasket outer boundary conditions

2. For Name, enter OUTER_BC, set Select type to Boundary Condition and set the GRNOD to
Nodes.
3. Click Nodes and select a node on the outer surface.

4. Click Nodes on the panel and then select by face to select all nodes on the outer surface.
5. Check all the translational and rotational degrees of freedom.
6. Click Create to create the outer fixed boundary condition.
Step 11: Create output request and control cards

CONTROL CARDS TITLE TITLE GASKET
CONTROL CARDS MEMORY NMOTS 40000 Not needed

ENGINE KEYWORDS RUN RunName GASKET
ENGINE KEYWORDS TFILE Time frequency 1.5e-3


3. For name, enter GASKET and click Save.

2. For Input file, browse to the exercise folder and select the file GASKET_0000.rad.
2. Using HyperView plot the displacement and strain contour and vectors.
Displacement Contour for the 3 load steps (mm)

Von Mises Stress Contour at the end of the simulation

Examples
This manual illustrates examples solved using the RADIOSS software with regard to common
problem types.
The main purpose of this manual is:
 First, is to illustrate examples for validation using uncommon models for carrying out various
RADIOSS functionalities. Whenever possible, the results provided by RADIOSS are compared
with experimental data or analytical solutions. Furthermore, when the computation time is
significant, different types of formulations are compared in order to provide users with an
overall idea of the cost for a given option or formulation.
 Second, as a guide for new RADIOSS users or for users interested in a type of problem with
which they are not familiar. The data is provided to enable a detailed understanding of the
options used for modeling. The reader can load data files or process the results obtained at a
later stage. The techniques for modeling can be applied to similar problems.
 Third, to explain about the additional functions included in the RADIOSS data files, thus
providing helpful options when using the RADIOSS data input manuals.

List of Examples
1 – Twisted Beam
Torsion - bending coupling
Sensitivity study on mesh and
element formulations.
2 – Snap-through Roof
Snap-through problem solved
by explicit and implicit solvers.
Results are compared with
experiments.
3 – S-beam Crash
Sensitivity study on element
formulations, plasticity
treatment and boundary
conditions for impact.
4 – Airbag
Airbag deployment using
monitored volumes with
communications.
Perfect gas modeling.
5 – Beam Frame
Transient dynamic analysis
using beam elements.
6 – Fuel Tank
Fluid-structure coupling and
fluid flow are studied using ALE
formulation.
Two analyses are performed:
sloshing and fuel tank
overturning.

7 – Pendulums
Momentum transmission,
contact modeling, bi- and tri-
dimensional analysis.
8 – Hopkinson Bar
Study of the stress wave
propagation and the strain rate
effect on the Hopkinson bar.
9 – Billiards (Pool)
Impact between balls,
trajectory study and treatment
with several interfaces
(Penalty/Lagrangian method).
10 – Bending
Pure bending test.
3- and 4-nodes shell.
11 – Tensile Test (Material

Characterization)
Correlations between
simulations and experimental
results. Treatment of the
necking point and the failure.
12 – Jumping Bicycle
A sequence of events managed
using "sensors".

13 – Shock Tube
Analysis with SPH, Lagrangian
and Eulerian formulations.
Correlation with theory.
Perfect gas modeling.
14 – Truck with Flexible

Body
Creating an overall flexible
body.
Quasi-static treatment
(gravity).
Eigen analysis (flexible body
inputs)
Dynamic analysis (bump).
15 – Gears
Contact modeling for quadratic
surface with interfaces 16 and
17.
16 – Dummy Positioning
Quasi-static analysis by explicit
solver with different
convergence options.
Static analysis by implicit solver
(linear and nonlinear problem).
17 – Box Beam
Crash test.

18 – Square Plate
Torsion and tension-
compression tests.
19 – Wave Propagation
Bi-dimensional wave
propagation.
Lagrangian and ALE
formulations.
Infinite domain modeling.
20 – Cube
Demonstrative problem.
Contact modeling.
Co-rotational formulation
elements.
21 – Cam
Contact modeling.
Linear and quadratic surface.
Comparison of fine and coarse
meshes.
22 – Ditching
Fluid simulation using the
Smooth Hydrodynamic Particles
formulation.
Comparison with experimental
data.
23 – Brake
Frictional contact modeling.
Lagrangian formulation.

24 – Laminating
Study of the number of
elements with regard to
thickness, the large/small
strain formulation, plastic strain
formulation and temperature
dependency.
25 – Spring-back
Explicit stamping simulation
followed by an implicit/explicit
spring-back simulation. Final
shape of the sheet metal is
compared with experiments.
26 – Ruptured Plate
Perforation of a thick plate by a
rigid sphere. Different failure
models integrated in material
law (2 and 27) or independent
(/FAIL options) are used.
27 – Football (Soccer) Shots

Simulations of football (soccer)
shooting impacts on a round or
a square bar. The airbag
modeling is used.
37 – Analytical Beam
Illustrates how to prepare a
RADIOSS deck for linear
analysis, and demonstrates a
high quality of RADIOSS finite
elements to resolve linear and
nonlinear problems.
39 - Biomedical Valve
A Fluid-Structure-Interaction
(FSI) problem is studied.

42 - Rubber Ring
Crush and Slide
43 - Perfect Gas
Polynomial EOS is used to
model Perfect Gas.
44 - Blow Molding with AMS

Blow molding with Advanced
Mass Scaling (AMS).
45 - Multi-Domain
Separate the whole model into
master domain and sub-
domain.
46 - TNT Cylinder Expansion

Test
An experimental test used to
characterize the adiabatic
expansion of detonation
products.

47 - Concrete Validation
Three kinds of tests are
performed in order to evaluate
the simulation/experiment
correlation.
48 - Solid Spotweld
Solid spotweld connects two
metal sheets with tied contact.
49 - Bird Strike on
Windshield
Introduce how to simulate a
bird hitting a windshield.
50 - INIVOL and Fluid

Structure Interaction (Drop
Container)
Introduces /INIVOL for initial
volume fractions of different
materials in multi-material ALE
elements, /SURF/PLANE for
infinite plane, and fluid
structure interaction (FSI) with
a Lagrangian container.

51 - Optimization in
RADIOSS for B-Pillar
The optimization objective is to
minimize the mass of the B-
Pillar by changing the shell
thickness.
52 - Creep and Stress

Relaxation
How to use typical visco-elastic
material to simulate creep and
stress relaxation tests.
53 - Thermal Analysis
A heat source moved on one
plate. Heat exchanged between
heat source and plate through
contact also between plate and
atmosphere (water) through
convective flux.

Example 1 - Twisted Beam
Summary
This example deals with a clamped beam subjected to a coupled torsion-bending loading. This
simple test being particularly severe for shell elements, a sensitivity study is performed on the
mesh and element formulation. An analytical solution validates the accuracy of results. The problem
under analysis consists of a concentrated load being applied to the extremity of the beam with the
static approach requiring a convergence method to enable fast convergence towards equilibrium.
The dynamic relaxation option allows for an efficient quasi-static response to be obtained.
The results are compared using two separate views:
 Shell element formulations (BATOZ, QEPH, DKT18 and BT hourglass type 4).
 Influence of the mesh (Triangular and quadrilateral meshes are compared using three
different element densities: 4x24, 2x12 and 1x3).
Several results can be extracted:
 X-displacement of the loaded point
 Y-displacement of the loaded point
 Z-displacement of the loaded point
 Error on energy
 CPU time
Comparisons are made between theoretical displacements and those by simulations.
Results show that QEPH and BATOZ element formulations provide the most accurate results and the
more the mesh is fine, the more accurate the results will be. To pass this test, a good curvature
representation of element formulation is needed; the BT hourglass type 4 formulation does not
satisfy this condition. QEPH offers a good ratio in terms of precision-cost, and is useful for quasi-
static analysis. DKT18 is a costly element formulation.

Title
Twisted beam
Number
1.1
Brief Description
Bending test on a twisted beam modeled with triangular and quadrilateral meshes and different
element formulations (QEPH, BT hourglass type 4, BATOZ, DKT).
Keywords
 4-node shell (Q4) and 3-node shell (T3)
 QEPH, BT (Hourglass type 4), BATOZ and DKT18
 Density mesh, elasticity, and dynamic relaxation
 Linear problem
RADIOSS Options
 Concentrated load (/CLOAD)
 Dynamic relaxation (/DYREL)
Compared to / Validation Method

 Analytical solution
Input File
QEPH: <install_directory>/demos/hwsolvers/radioss/01_Twisted_Beam/QEPH/TWISBEAM*
BATOZ: <install_directory>/demos/hwsolvers/radioss/01_Twisted_Beam/BATOZ/TWISBEAM*
BT-TYPE4: <install_directory>/demos/hwsolvers/radioss/01_Twisted_Beam/BT-
type4/TWISBEAM*
DKT18: <install_directory>/demos/hwsolvers/radioss/01_Twisted_Beam/DKT18/TWISBEAM*
Technical / Theoretical Level

Medium
Overview
Aim of the Problem

The purpose of this example is to compare element formulations concerning mesh density with
regard to a coupled torsion-bending problem.
Physical Problem Description
Units: In, s, lbs-s2/in
A twisted beam is clamped at one end, and subjected to a concentrated load at the other end.

The material used follows a linear elastic law (/MAT/LAW1) and has the following geometrical
characteristics with no specific measurement unit:
Initial density: 7.34x10-4
Young’s modulus: 2.9x107 [MPA]
Poisson ratio: 0.22
Thickness: 0.32
Length: 12
Width: 1.1
Load case:
Fx = 0
Fy = 1.0
Fz = 0
Fig 1: Initial mesh (4x24).
This simple test is particularly severe for shell element behaviors, due to the torsion-bending
coupling. Users appreciate the qualities/restrictions of the shell element formulations in RADIOSS.
The following points are:
 Displacements are very low. Thus, you are faced with a linear problem.
 Another load case, using Fy = 0 and Fz = 1, is considered, but does not give concern to
additional conclusions.

Analysis, Assumptions and Modeling Description
Modeling Methodology
The beam is modeled with 4-node shell and 3-node shell meshes.
The following are tested for each model:
 Four shell formulations:
- QEPH formulation (4-node shell element, Ishell = 24)
- BT (Hourglass type 4) formulation (4-node shell element, Ishell = 4)
- QBATOZ formulation (4-node shell element, Ishell = 12)
- DKT18 formulation (3-node shell element, Ish3n = 30)
 Three mesh densities in each shell formulation:

- Mesh A: 4 x 24 elements
- Mesh B: 2 x 12 elements
- Mesh C: 1 x 3 elements
4-node Shell Mesh 3-node Shell Mesh
RADIOSS Options Used

One concentrated load is applied at the extremity, on central node M (for mesh A and mesh B), two
concentrated loads must be applied to the corner points of the beam end (for mesh C).
A static solution provides the steady state part of the transient response. In this example, dynamic
relaxation is used to obtain a static result. Static loading is considered a dynamic resolution
method. Using /DYREL in the *_0002.rad file, the dynamic loading is damped by introducing a
diagonal damping matrix.
Relaxation factor = 1; period to be damped = 0.0025
The displacement of node M is stabilized at the static response: t = 0.035.
For further details, refer to the RADIOSS Theory Manual.

Simulation Results and Conclusions
Curves and Animations

The displacement components regarding X, Y and Z of node M are compared to the beam theory in
order to understand the performance of the various elements when several mesh densities are
used.
Reference results [Batoz & Dhatt, "Structural Modeling Finite Elements", Vol. 3, Hermès, Paris,
1992]:
X-displacement: UM = 0
Y-displacement: VM = 0.00175
Z-displacement: WM = -0.00179
The chart below shows the displacement oscillations of point M until reaching stabilization in the
direction of the static solution.
Fig 2: Time history plots of point M displacements (Mesh A/QEPH)

 Energy margin error at t = 0.05:
QEPH BT (type 4) BATOZ DKT18
Energy 99.9%
0.1% 0% 0%
margin error (diverge)
 CPU (normalized):
QEPH BT (type 4) BATOZ DKT18
CPU 1 diverge 1.1042 1.1430
 Nodal displacements of node M:
 Ratio (Displacement by simulation/ Displacement by theory):
Conclusion
 QBAT and QEPH provide good results (precision).

 Good results provided by DKT18 when the mesh is fine, though no better than QBAT and
QEPH.
 BT (Hourglass type 4) does not pass this test (due to the flat facet approach).
 QEPH: the best element formulation for quasi-static analysis. Very good precision-cost ratio.
 QBAT: good curvature representation. For quasi-static analysis, the cost is 4% higher
compared to using the QEPH formulation.
 DKT18 represents the highest cost for this test.

Example 2 - Snap-thru Roof
Summary
A snap-through problem is studied on a shallow cylindrical roof upon which an imposed velocity is
applied at its mid-point. The characteristic curve, caused by the limit load and achieved by
simulation is compared to a reference. This example is considered a static problem.
Only one-quarter of the structure is taken into consideration and adequate boundary conditions are
applied on the model sides.
The problem is solved using two different approaches:
 An analysis by an explicit solver
 An analysis by an implicit solver
The implicit strategy uses the arc-length method with a time step limitation. The RADIOSS implicit
options are defined in the modeling description.
The simulations using explicit and implicit methods provide accurate results with a good evaluation
of the limit load experimentally observed. A time step control with a low value is required in order
to describe the nonlinear path of the load displacement curve. Both computations converge toward
a single solution.

2.1 - Snap Roof: Explicit
Title
Snap Roof - Explicit
Number
2.1
Brief Description
An imposed velocity is applied onto a shallow cylindrical roof at its midpoint. The analysis uses an
explicit approach.
Keywords
 Explicit solver
 T3 Shell
 Elasticity and quasi-static analysis
 Stability, snap-through problem, and limit load
RADIOSS Options
 Boundary conditions (/BCS)
 Imposed velocity (/IMPVEL)
 Rigid body (/RBODY)

 Experimental results
Input File
Explicit solver: <install_directory>/demos/hwsolvers/radioss/02_Snap-
through/Explicit_solver/SNAP_EXP*

Beginner

Overview
Aim of the Problem

The purpose of this example is to study a snap-thru problem with a single instability. Thus, a
structure that will bend when under a load is used. The results are compared to the references
contained in: Finite Element Instability Analysis of Free Formed Shells. Report 77−2, 1977,
Norwegian Institute of Technology, Trondheim, HORRIGMOE G.
This static analysis is performed with an explicit approach.
A shallow cylindrical roof, pinned along its straight edges upon which an imposed velocity is applied
at its mid-point.
Units: mm, ms, g, N, MPa
Geometrical data are provided in Fig 1, with the following dimensions:
 l = 254 mm
 R = 2540 mm
 Shell thickness: t = 12.7 mm
 = 0.1 rad
Fig 1: Geometrical data of the problem
The material used follows a linear elastic law and has the following characteristics:
 Initial density: 7.85x10-3 g/mm3
 Young modulus: 3102.75 MPa
 Poisson ratio: 0.3

The structure is considered perfect, having no defects. To take account of the symmetries, only a
quarter of the shell is modeled (surface ABCD).
A regular mesh with a total of 72 3-node shells (Fig 2)
Fig 2: T3 mesh
The shells have the following properties:

 Thickness 12.7 mm
 BT Elasto-plastic Hourglass formulation (Ishell = 3).

Node time histories do not indicate the pressure output. In order to obtain such output at point C, a
rigid body must be created at this point. Point C has a constant imposed velocity of -0.01 ms-1 in
the Z direction. Its displacement is linked proportionally to time.
Boundary conditions are:
 Edge BC is fixed in an X translation, and in Y and Z rotations (symmetry conditions).
 Edge CD is fixed in a Y translation, and in X and Z rotations (Idem).
 Edge DA is fixed in X, Y, Z translations, and in X and Z rotations.
 Point C is fixed in X, Y translations, and in X, Y, Z rotations.
Fig 3: Boundary conditions


Only a quarter of the total load is applied due to the symmetry. Therefore, force Fz of the rigid
body, as indicated in the Time History, must be multiplied by 4 in order to obtain force, P.
Figure 4 represents a characteristic load displacement curve for a snap-through. This diagram plots
the reaction at point C of the shell as a function of its vertical displacement.
Fig 4: Load P versus displacement of point C: snap-thru instability.
The displacement of point C is indicated in its absolute value. The curve illustrates the characteristic
behavior of the instability of a snap-thru. Beyond the limit load, an infinite increase in load Fz will
cause a considerable increase in displacement q due to the collapsing of the shell.
The first extreme defines the limit load =2208.5 N (displacement of point C = 10.5 mm).
The increase in the curve slope after the snap-thru, shows that the deformed configuration becomes
more rigid.

Fig 5: Comparison between a reference curve and a curve obtained using RADIOSS
The difference between the two curves is approximately 10% for reduced displacements (up to 5
mm) and slightly more (15%) for the higher nonlinear part of the curve (between 5 and 20 mm).
For displacements exceeding 20 mm, the curves are shown much closer together.
The accuracy of the RADIOSS results in comparison to those obtained from the reference is ideal for
this explicit approach.

Deformed Mesh (profile view) – Displacement Norm
Initial configuration
Start of snap-thru
Large motion phase
Stable configuration
Loading with a new structural rigidity

2.2 - Snap Roof: Implicit
Title
Snap Roof - Implicit
Number
2.2
Brief Description
A shallow cylindrical roof upon which an imposed velocity is applied at its mid-point. Analysis uses
an implicit approach.
Keywords
 Implicit solver, time step control by arc-length method
 Static nonlinear analysis
 Stability, snap-thru, and limit load
 T3 Shell
RADIOSS Options
 Implicit options (/IMPL)

Input File
Implicit solver: <install_directory>/demos/hwsolvers/radioss/02_Snap-thru/
Implicit_solver/SNAP_IMP*

Advanced

Overview
Aim of the Problem

The purpose of this example is to study a snap-thru problem with a single instability. Thus, a
structure that will bend when under a load will be used. The results are compared to a reference
solution [1]. This analysis is performed using an implicit approach. An implicit strategy using an arc-
length method is illustrated.
A shallow cylindrical roof, pinned along its straight edges, upon which an imposed velocity is applied
at its mid-point.
Geometrical data are indicated in Fig 6, with the following dimensions:
 l = 254 mm
 R = 2540 mm
 Shell thickness: t = 12.7 mm
 = 0.1 rad
Fig 6: Geometrical data of the problem
The material used follows a linear elastic law and has the following characteristics:
 Young modulus: 3102.75 MPa

The modeling problem described in the explicit study remains unchanged.
The implicit computation requires specific implicit parameters that must be defined in the Engine file
*_001.rad using the options beginning with /IMPL.
Fig 7: Description of the problem (one quarter of the shell is modeled)
The imposed velocity is considered using the implicit method. Thus, the constant input curve is
converted into an imposed displacement according to the computation time.
Fig 8: Imposed velocity curve

The limit point causes major non-linearities. Therefore, a static nonlinear analysis is performed
using the arc-length displacement strategy. The time step is determined by a displacement norm
control. In order to exceed the limit point characterized by a null tangent on the load displacement
curve and to describe the increasing and decreasing parts of the nonlinear path, a small time step is
required, which is ensured by setting a maximum value.
The nonlinear implicit parameters used are:
Implicit type: Static nonlinear
Nonlinear solver: Modified Newton
Tolerance: 2x10-4
Update of stiffness matrix: 3 iterations maximum
Time step control method: Arc-length
Initial time step: 10 ms
Minimum time step: 0.5 ms
Maximum time step: 10 ms
Desired convergence iteration number: 6
Maximum convergence iteration number: 20
Decreasing time step factor: 0.8
Maximum increasing time step scale factor: 1.05
Arc-length: Automatic computation
Spring-back option: no
A solver method is required to resolve Ax=b in each iteration of a nonlinear cycle. It is defined in
/IMPL/SOLVER. The linear implicit options used are:
Linear solver: Direct solver MUMPS
Precondition methods: Factored approximate Inverse
Maximum iterations number: System dimension (NDOF)
Stop criteria: Relative residual in force
Tolerance for stop criteria: Machine precision
The input implicit options set in *_001.rad are:
/IMPL/PRINT/NONL/-1 Printout frequency for nonlinear iteration
/IMPL/SOLVER/2 Solver method (solve Ax=b)

5 0 0 0.0
/IMPL/NONLIN Static nonlinear computation
3 1 0.20e-3
/IMPL/DTINI Initial time step determines the initial loading increment
10
/IMPL/DT/STOP Min Max values for time step
0.5 10
/IMPL/DT/2 Time step control method 2 - Arc-length+Line-search will be used
6.0 20 0.8 1.05 with this method to accelerate and control convergence
Refer to RADIOSS Starter Input for more details about implicit options.


Only a quarter of the total load is applied due to the symmetry. Thus, force Fz of the rigid body as
indicated in the Time History must be multiplied by 4 in order to obtain force, P.
Figure 9 represents the characteristic load displacement curve for a snap-thru. This diagram plots
the reaction at point C of the shell as the function of its vertical displacement. The implicit results
are compared with the experimental data.
Fig 9: Load P versus displacement of point C.
For a time step equal to or less than 10 ms (maximum value set in the implicit /IMPL/DT/STOP
option), agreement with RADIOSS is achieved, with good results obtained using the reference.
Accuracy is improved by decreasing the maximum time step, even though the CPU time is
increased.
Fig 10: Deformed configurations during the snap-thru.

Comparison between Implicit and Explicit Results
The load displacement curves achieved through implicit computations (time step limit set to 10 ms)
and explicit computations are very close. A maximum time step of 100 ms does not allow the
nonlinear path of the load displacement curve to be described accurately. However, the final static
solution is correct.
Fig 11: Load displacement curve obtained by implicit and explicit solvers.
Comparison of the computation time between the explicit and implicit (maximum time step set to
10 ms) approaches is shown in the table below:
Implicit solver Explicit solver
Normalized CPU 1 2.45
Cycles (normalized) 1 237

In comparison with the implicit computation, which uses a maximum time step of 10 ms, the saved
CPU time using a maximum time step fixed at 100 ms, approximately corresponds to factor 4.
Reference
[1] Finite Element Instability Analysis of Free Formed Shells. Report 77−2, 1977, Norwegian
Institute of Technology, Trondheim, HORRIGMOE G.

Example 3 - S-beam Crash
Summary
A sensitive study is performed on a crushed S-beam. The modeling includes a material law using
the elasto-plastic model of Johnson-Cook and a self-impacting interface based on the Penalty
method in order to model the buckling of the beam. An initial velocity is applied on the left section
via a kinematic condition: either a rigid body or a rigid link. The impacting condition is sliding and is
secured by specific boundary conditions in the right section. Half of the structure is modeled.
The results are compared according to three different views:
 Shell element formulations (BATOZ, QEPH and BT hourglass type 3)
 Plasticity options (global and progressive plasticity)
 Influence of the initial velocity (5 and 10 ms -1)
Several criteria are used to compare the results:
 Deformation configuration
 Crushing force versus displacement (via momentum integration)
 Energy assessment
 Displacement of the left section
 Hourglass energy
 Kinetic energy
 Internal energy
 Maximum force
 Maximum plastic strain
BATOZ and QEPH element formulations provide accurate results. The BT hourglass type 3
formulation is a low-cost method and the QEPH formulation provides a good precision/cost ratio
(the cost is three times lower than the BATOZ formulation). BATOZ and QEPH are element
formulations which do not have hourglass energy.
The results show an over-estimation of the plastic strain in the case of the global plasticity use.

Title
S-Beam
Number
3.1
Brief Description
An S-beam is crushed against a rigid wall with initial velocity.
Keywords
 Shell, type 3 Q4 Hourglass, QEPH, and BATOZ
 Type 7 interface, self-impacting, plasticity, and /MAT/LAW2
 MODIF files
RADIOSS Options
 Initial velocities (/INIVEL)
 Rigid link (/RLINK)
Input File
QEPH: <install_directory>/demos/hwsolvers/radioss/03_S-Beam/QEPH/
Global_plasticity/QEPH*
BATOZ: <install_directory>/demos/hwsolvers/radioss/03_S-Beam/BATOZ/
Global_plasticity/BATOZ*
BT_type3_NiP0: <install_directory>/demos/hwsolvers/radioss/03_S-Beam/BT-type3/
Global_plasticity/Q4_NIP0*
BT_type3_NiP5: <install_directory>/demos/hwsolvers/radioss/03_S-Beam/BT-type3/
NiP5/Q4_NIP5*

Advanced

Overview
Aim of the Problem

The purpose of this example is to study the behavior of a crashed S-beam using various shell
formulations and a number of different integration points. This test also compares the initial velocity
influence on results. A MODIF file is used to introduce a self-impacting interface.
An S–beam is crushed at an initial rate of 5 ms -1 against a rigid wall. The section is an empty
square-shaped tube (each side measuring 80 mm). The thickness is 1.5 mm. The tube is made of
steel, and plasticity is taken into account, but not failure. Using symmetry, half of the cross-section
is modeled.
Fig 1: Problem description and beam cross-section.
The following system is used: mm, ms, g, N, MPa

The material used follows an isotropic elasto-plastic Johnson-Cook law.
Material properties:
 Young’s modulus: 199355 MPa
 Poisson’s ratio: 0.3
 Density: 7.9x10-3 g/mm3
 Yield stress: 185.4 MPa
 Hardening parameter: 540 MPa
 Hardening exponent: 0.32
 Maximum stress: 336.6 MPa
All other properties are set to the default values.

The mesh is a regular shell mesh. Each shell measures approximately 10 mm x 10 mm.
A sensitive study is performed on:
 Shell element formulations: BATOZ, QEPH and Belytschko hourglass type 3
 Plasticity options: global and progressive plasticity model
 Influence of the initial velocity: 5 and 10 ms -1
Fig 2: Structure’s overall mesh
The rigid wall is modeled with boundary conditions on the right section of the beam (X, Z
translations and all rotations fixed).
The left section undergoes the following conditions:
 Fixed in the Z direction.
 Initial velocity of 5 m/s in the X direction.
 All nodes are rigidly connected in X, Y and Z directions.
 A 500 Kg mass is added on the left end.
Block format input specifications:
 Hierarchy organization: there is only one subset made up of three parts, one for each side of
the beam, and one for the top. The materials and properties are identical for each part.
 Node groups: there are three node groups, one for each end of the beam, and one for the
symmetry plane. The boundary conditions are set on the left end.
 TH selection: DX is saved for node 1 (the node used to display displacement at the left end).

Taking account of symmetry, half of the structure is modeled. The symmetry plane covers the y
axis = 0 mm. Boundary conditions are also set at the right end to simulate a rigid wall (slide).
Two equivalent possibilities are available for generating kinematic conditions attached to the left
extremity of the beam. The first consists of creating a rigid body to connect all of the left section
nodes to the gravity center of the beam cross-section, with a mass being introduced on a master
node. The second type of modeling retained uses the rigid link option, which rigidly connects the left
section nodes in the X, Y and Z directions. A 500 kg mass is added to the master node.
Both models provide identical results; the rigid link will be used for this example.
 An initial velocity of 5 ms-1 is used for the master node of the rigid link or for the rigid body.
 MODIF file:
A MODIF file enables to add option(s) during a run. The MODIF files carry the name
ROOTNAMErun*.rad. Where, run# is the RADIOSS run number four digits from 0000 to 9999
and run# is the name of the last Restart file + 1.
For example, to run a MODIF file after the first run (restart file ROOTNAME_0001.rad), the run
number for the MODIF file must be 2: ROOTNAME_0002.rad. MODIF files use the same inout
format as the RADIOSS deck. Put all the input decks in one folder and with Irun=2 RADIOSS
will automatically recognize the MODIF file.

After 20 ms, a self-impacting interface is required to deal with the buckling of the beam. This is
added using a MODIF file where the interface is defined and saved for the TH. This type of interface
corresponds to 7; all values are set to "default". To define the master side, a surface is defined
using three parts of the model (/SURF/PART). The safest and easiest method for defining the slave
side of a self-impacting interface consists of defining a node group with the master surface
(/GRNOD/SURF).
The MODIF file is ROOTNAME_0002.rad.
The next Engine file is ROOTNAME_0003.rad (final time = 30 ms).

Fig 3: Contact force at the start of self-impacting.
The MODIF file options used in Engine file ROOTNAME_0002.rad are:

Fig 4: Deformed mesh for Belytschko hourglass type 3 formulation (V=5 m.s -1)

The crushing force is obtained by time derivation of the X-momentum. The maximum displacement
over a 20 ms long computation corresponds to 96.4 mm.
Fig 5: Crushing force (X-direction) versus displacement for different element formulations (V=5 m.s -1)

Fig 6: Energy assessment for Belytschko hourglass type 3 (V=5 m.s -1).
The structure does not absorb a lot of energy and that you should check the hourglass energy,
which may be relatively high compared with the total energy.

The following table shows the results obtained using different element formulations and plasticity
options:
Global plastification (NiP = 0) NiP = 5
Q4 Hourglass Q4 Hourglass
BATOZ QEPH
type 3 type 3
Initial energy (mJ) 6.25012x106 6.25012x106 6.25012x106 6.25012x106
Kinetic energy (mJ) 5.499x106 5.47964x106 5.55602x106 5.55641x106

t = 30 ms (0.877) (0.875) (0.889) (0.888)
Internal energy (mJ) 750374 770384 684100 691255

t = 30 ms (0.123) (0.125) (0.109) (0.110)
Hourglass energy (mJ) 0 0 28385 33341.6

t = 30 ms (0.0016) (0.002)
Displacement (mm) 144.0 144.6 144.6 144.7
t = 30 ms
CPU 1.082 0.99 1 0.988
(Normalized)
Error on energy (%) 0% 0% -0.1% -0.1%
t = 30 ms
Maximum force (N) 42459.4 42688.5 35949.4 35387.2
Maximum plastic strain 0.462 0.448 0.414 0.323
Initial velocity = 5 ms-1 (Values in brackets are the energy percentages compared with the initial energy)

Plastic Strain - Time = 10.00 ms
Global plastification
Progressive plastification (Nip = 5)

Global plastification
Progressive plastification (Nip = 5)

Initial Velocity Influence
The following table indicates the influence of the crushing velocity (5 ms -1 and 10 ms-1).
Initial Velocity = 5 ms-1 Initial Velocity = 10 ms-1
Initial energy (mJ) 6.25012x106 2.5 x107
Kinetic energy (mJ)
X – displacement = 70 mm: 5.79897x106 (0.928) 2.44581x107 (0.978)

X – displacement = 140 mm: 5.57192x106 (0.891) 2.41546 x107 (0.966)
Internal energy (mJ)
X – displacement = 70 mm: 444848 (0.0711) 538142 (0.0215)

X – displacement = 140 mm: 666704 (0.107) 840622 (0.0336)
Hourglass energy (mJ)
X – displacement = 70 mm: 4879.87 (0.0009) 5969.83 (0.0005)

X – displacement = 140 mm: 9530.27 (0.002) 12702.4 (0.0004)
Maximum force (N) 35949.4 41704.3
Error on energy (%) -1.09% -1.11%
(Values in brackets refer to the energy percentages compared with the initial energy)
BT hourglass type 3 formulation is used in this section.

The amount of internal energy stored in the beam during a crash is relatively higher when the initial
velocity is set to 10 ms-1, instead of 5 ms-1. The hourglass energy is quite low with either initial
velocity.

Fig 7: Crushing force versus displacement for the different initial velocities

Fig 8: Kinetic energy normalized for the different initial velocities
First self-contact:
 Initial velocity = 5 ms-1: displacement = 120 mm;
 Initial velocity = 10 ms-1: displacement = 94.15 mm.

Example 4 - Airbag
Summary
This example deals with the deployment of a chambered airbag modeled by monitored volumes
using communications. The airbag is initially folded along four fold lines. The fabric is meshed with
shell elements which undergo an elastic orthotropic behavioral test. Perfect gas is injected into a
central chamber via an inflator with the air flow through the connected chambers being simulated.
The chambers inflate while the airbag is deploying.
In the self-impacting interface definition, the action of the Inacti flag to deactivate stiffness in the
case of initial penetration is studied in order to significantly increase the time step. An adequate gap
enables to pass from a kinematic interface time step to a higher element time step.

Title
Airbag
Number
4.1
Brief Description
A chambered airbag folded along four fold lines is deployed.
Keywords
 Orthotropic shell
 Monitored volumes and communicating airbags
 /MAT/GAS
 /PROP/INJECT1
 Material law 0 and type 7 interface
 Hierarchy organization
RADIOSS Options
 Monitored volume with communications (/MONVOL/COMMU1)
 Interface (/INTER/ with Inacti flag)
Input File
Inactiv_0_Gap0.1:
<install_directory>/demos/hwsolvers/radioss/04_Airbag/Inacti0_Gap01/AIRFIX*
Inactiv_5_Gap0.3:
<install_directory>/demos/hwsolvers/radioss/04_Airbag/Inacti5_Gap03/AIRBAG*
Inactiv_5_Gap1.5:
<install_directory>/demos/hwsolvers/radioss/04_Airbag/Inacti5_Gap15/AIRBAG2*

Beginner

Overview
Aim of the Problem

The purpose of this example is to deal with monitored volumes using communications on a simple
airbag model. Methods for increasing the time step are considered.
A 30-liter airbag is folded along the four fold lines. The following examples illustrate the airbag
folded and deployed.
Fig 1: Folded airbag Fig 2: Deployed airbag
The fabric thickness is 0.33 mm and is modeled using an elastic orthotropic material law
(/MAT/LAW19) with the following properties:
 Young’s Modulus: 500 MPa in both directions
 Shear Modulus: 10 MPa
 Reduction factor: 0.001
The property set is /PROP/SH_ORTH (shell orthotropic, type 9), using one integration point.

Fig 3: Overall mesh of the structure (folded and deployed).
The model is divided into two subsets: the fabric layers and the communication surfaces.
The fabric surface is then divided into 9 subsets, one for each monitored volume. Each "monitored
volume" is further divided into two parts. All the parts of the layer of fabric have the same Type and
MID.
The same properties apply for the communication surfaces.
Fig 4: Folder airbag with communications.
The airbag is modeled using 9 communicating volumes in order to simulate the air flow through the
folds and the behavioral differences within the airbag when unfolding. The communicating surfaces
between the volumes are simulated using dummy membranes. The dummy membranes are
modeled using shells with fictitious material (/MAT/LAW0).
A monitored volume is defined as a surface area having one or more shell property sets and where
the surface must be closed. The monitored volume used is a COMMU1 type for airbags using
communications (chambered, with communications, of the folder airbag type). For further details
about monitored volumes, see the RADIOSS Theory Manual.

The main properties for this type are:
 Volumetric damping factor: 0.001 g.mm-1ms-1 .
 External pressure: 0.1 MPa
 Constant perfect gas: 1.4
 Specific heat at constant pressure: 926 mJ/g (This is the specific heat coefficient related to
mass)
 Temperature: 780 K.
 Communication area: total (Acom =1 and Scom =0)
 Time to deflate vent hole: 1030 ms

The gas molecular weight and specific heat coefficients are defined in /MAT/GAS (using MASS type):
 Molecular weight of gas MW: 30.09204 g/kmol (AIR)
 Specific heat at constant pressure: 926 mJ/g (specific heat coefficient related to mass)
Specific input for the central chamber one (inflator):
 Vent hole membrane surface area is 1000 mm 2 (Avent =0) and is immediately activated.
 Relative vent deflation pressure: 0.0002

 Number of injectors: 1 (Njet =1; Ijet =0)
Using /PROP/INJECT1 to described mass inject:

 Final injected mass is 46 g injected into the central chamber (Fscale M and FscaleT =1). Two
functions define the mass and temperature of the injected gas compared with time (function
identifiers: fct_IDM and fct_IDT).
Time (ms) 0 2 4 5 6 8 11 12 15 19 28 30 106
Mass (g) 0 6 11 14 17 22 29 31 36 41 45 46 46
Injected mass function.
Time (ms) 0 106
Temperature (K) 780 780
Temperature of injected gas function.
Interface
Taking into account the fabric is self-impacting with itself, a self-impacting interface must be used.
The interface’s Block Format definition is made: defining the master surface (/SURF/PART), then
defining the slave nodes for all nodes on this surface (/GRNOD/SURF).
The distance between the fabric layers before unfolding is very small. In order to avoid initial
penetration, the gap required is approximately 0.1 mm, thus enabling the time step to considerably
decrease when such a gap is chosen.
By using Inacti =5, a 0.3 mm gap is chosen. Any initial penetration below 0.2 mm (two-thirds of the
input gap) is ignored (it is strongly recommended to verify that no initial penetration is above this
value).


In order to demonstrate the interest of the Inacti flag, the same model was run with Inacti at a
value of 0, with a gap of 0.1 mm (no initial penetration).
Fig 5: Comparison between option Inacti = 5 and Inacti = 0
Using Inacti = 5, the minimum time step is around 10 -3 ms. When not using this option, the
minimum time step is around 2x10-4 ms. For the full model, the number of cycles may be divided
up into 10 or more. Furthermore, the model is numerically less sensitive.
The time step is monitored by the interface time step (kinematic) for up to 40 ms despite the
unfolding and the fact that there is no energy contact from 7.8 ms. In order to transfer into the
element time step and to reduce computation time, it is advisable to increase the gap so the
kinematic step becomes higher than the element step.
Time-stepkinetic < 0.9 x GAP / Nodal_velocityrelative (using scale factor = 0.9)

The time step is only low during the unfolding phase (before 10 ms) with a gap equal to 1.5 mm.
Inacti flag = 5 Inacti flag = 0
GAP = 0.3 mm GAP = 1.5 mm GAP = 0.1 mm
Error on energy -16.3% -19.6% -15.5%
Elapsed Time [s] 57.5 47.8 78.58
Airbag deploy completely deploy completely deploy incompletely deploy
Fig 6: Time step obtained with GAP = 0.3 mm and GAP = 1.5 mm (Inacti = 5).

Fig 7: Contact energy with GAP = 0.3 mm and GAP = 1.5 mm (Inacti = 5).
It is obvious that a gap of 1.5 mm generates an increase in the contact force. However, the
additional error on energy remains quite low and is acceptable.

Fig 8: Time history of pressure.
Fig 9: Time history of volume.

Animations
Fig 10: Central chamber is inflating.
Fig 11: All chambers are inflating.

Fig 12: Airbag is deployed

Example 5 - Beam Frame
Summary
A beam frame with clamped extremities receives an impact at its mid-point from a pointed mass
having initial velocity. The material is subjected to the elasto-plastic law of Johnson-Cook. The
model is meshed with beam elements. An infinite rigid wall with only one slave node, including the
impacted node, is subjected to the initial velocity. This example is considered a dynamic problem
and the explicit solver is used.
The explicit approach leads to finding a quasi-static equilibrium of the structure after impact.

Title
Beam-frame
Number
5.1
Brief Description
A beam frame receives an impact from a mass having initial velocity.
Keywords
 Beam
 Rigid wall
 Plasticity and Johnson-Cook material (/MAT/LAW2)
RADIOSS Options
 Beam element (/PROP/BEAM)
 Rigid wall (/RWALL)
Input File
Beam_frame: <install_directory>/demos/hwsolvers/radioss/05_Beam-frame/FRAME*

Beginner
Overview
Aim of the Problem

The purpose of this example is to perform a static analysis using beam elements.
A pointed mass (3 kg) makes an impact at point O of a beam frame (see Fig 1 for the geometry)
using a speed of 10 ms-1 in the Z direction. The beams are made of steel and each beam section is
square-shaped (each side being 6 mm long).
Fig 1: Geometry of the frame.

Dimensions are: AB = BC = CD = BE = BF = E’C = CF’ = 90 mm.
Points A, D, E, F, E’, and F’ are fixed.
The beams have the following properties:
 Cross section: 36 mm2
 Moments of inertia in Y and Z: 108 mm 4
 Moments of inertia in X : 216 mm4
The steel material used has the following properties:
 Density: 0.0078 g/mm3
 Young’s modulus: 200 000 MPa
 Yield stress: 320 MPa
 Hardening parameter: 134.65 MPa
All other coefficients are set to default values. Plasticity is taken into account using Law 2 without
failure.
The mesh is a regular beam mesh, each beam being 9 mm long (total = 70 beams).
Fig 2: Mesh of the frame showing the position of the nodes.

The impacting mass is simulated using a sliding rigid plane wall (/RWALL) having an initial velocity
of 10 ms-1 and a mass of 3000 g. Only one slave node exists: the node O to simulate a point
impact.
Points A, F, F', D, E and E' are fully fixed.

Fig 3: Boundary conditions Fig 4: Rigid wall type infinite plane

The main results refer to the time history of points B and O with regard to displacements and
velocities.
Fig 5: Displacements of points B and O.

Fig 6: Velocity of points B and O (stabilization).
Fig 7: Normal and shear force on beam element 15 (near to point O).

Fig 8: Energy assessment (stability reached at in 6 ms).
Fig 9: Node displacement (max. = 30.96 mm).

Fig 10: Plastic strain (max. = 20.1%).

Example 6 - Fuel Tank
Summary
The fluid-structure interaction and the fluid flow are studied in cases of a fuel tank sloshing and
overturning. A bi-phase liquid-gas material with an ALE formulation is used to define the interaction
between water and air in the fuel tank.
In the case of sloshing, the fuel tank is subjected to a horizontal deceleration. The fuel tank
container is modeled with a Lagrangian formulation and undergoes an elasto-plastic material law.
Fluid structure coupling is taken into account.
The overturning of the fuel tank is studied by applying a variable deceleration. The tank container is
not modeled as the boundary nodes are fixed. The Eulerian formulation is used.

6.1 - Fluid Structure Coupling
Title
Fuel tank - Fluid
Structure Coupling
Number
6.1
Brief Description
Sloshing inside a fuel tank by simulating the fluid structure coupling. The tank deformation is
achieved by applying an imposed velocity on the left corners. Water and air inside the tank are
modeled with the ALE formulation. The tank container is described using a Lagrangian formulation.
Keywords
 Fluid structure coupling simulation, and ALE formulation
 Shell and brick elements
 Hydrodynamic and bi-phase liquid gas material (/MAT/LAW37)
RADIOSS Options
 ALE boundary conditions (/ALE/BCS)
 J. Donea Grid Formulation (/ALE/GRID/DONEA)
 Gravity (/GRAV)
 ALE material formulation (/ALE/MAT)
Input File
Fluid_structure_coupling: <install_directory>/demos/hwsolvers/radioss/06_Fuel_tank/
1-Tank_sloshing/Fluid_structure_coupling/TANK*

Advanced

Overview
Aim of the Problem

A numerical simulation of fluid-structure coupling is performed on sloshing inside a deformable fuel
tank. This example uses the ALE (Arbitrary Lagrangian Eulerian) formulation and the hydrodynamic
bi-material law (/MAT/LAW37) to model interaction between water, air and the tank container.
A rectangular tank made of steel is partially filled with water, the remainder being supplemented by
air. The initial distribution pressure is known and supposed homogeneous. The tank container
dimensions are 460 mm x 300 mm x 10 mm, with thickness being at 2 mm.
Deformation of the tank container is generated by an impulse made on the left corners of the tank
for analyzing the fluid-structure coupling.
Fig 1: Problem description.
The steel container is modeled using the elasto-plastic model of Johnson-Cook law (/MAT/LAW2)
with the following parameters:
The material air-water bi-phase is described in the hydrodynamic bi-material liquid-gas law
(/MAT/LAW37). Material law 37 is specifically designed to model bi-material liquid gas.

The equations used to describe the state of viscosity and pressure are:
 Viscosity:
 Liquid EOS:
where,
 Gas EOS:
The equilibrium is defined by: Pl = Pg
Where, Sij is the deviatoric stress tensor and eij is the deviatoric strain tensor.
Material parameters are:

 For liquid:
l 0
Liquid reference density: 0.001 g/mm3
Cl Liquid bulk modulus: 2089 N/mm2
al Initial mass fraction liquid proportion: 100%
l 0
Shear kinematic viscosity (= / ): 0.001 mm2/ms
 For gas:
g0
Gas reference density: 1.22x10-6 g/mm3
g0
g Shear kinematic viscosity (= / ): 0.00143 mm2/ms
Constant perfect gas: 1.4
P0 Initial pressure reference gas: 0.1 N/mm2
The main solid type 14 properties for air/water parts are:
 Quadratic bulk viscosity/linear bulk viscosity: 10-20
 Hourglass bulk coefficient: 10-5

Air and water are modeled using the ALE formulation and the bi-material law (/MAT/LAW37). The
tank container uses a Lagrangian formulation and an elasto-plastic material law (/MAT/LAW2).
Fig 2: Air and water mesh (ALE brick elements).
Fig 3: Tank container mesh (shell elements).

Using the ALE formulation, the brick mesh is only deformed by tank deformation the water flowing
through the mesh. The Lagrangian shell nodes still coincide with the material points and the
elements deform with the material: this is known as a Lagrangian mesh. For the ALE mesh, nodes
on the boundaries are fixed in order to remain on the border, while the interior nodes are moved.
Velocities (/IMPVEL) are imposed on the left corners in the X direction.
Table 1: Imposed velocity versus time curve

Velocity (ms-1) 0 5 0 0
Time (ms) 0 12 12.01 50
Fig 4: Kinematic condition: imposed velocities.
Regarding the ALE boundary conditions, constraints are applied on:

 Material velocity
 Grid velocity
All nodes, except those on the border have grid (/ALE/BCS) and material (/BCS) velocities fixed in
the Z-direction. The nodes on the border only have a material velocity (/BCS) fixed in the Z-
direction.
Both the ALE materials air and water must be declared ALE using /ALE/MAT. Lagrangian material is
automatically declared Lagrangian.
The /ALE/GRID/DONEA option activates the J. Donea grid formulation to compute the grid velocity.
See the RADIOSS Theory Manual for further explanations about this option.


Fluid – Structure Coupling
Fig 5: X – momentum variation for each part.
Kinematic conditions generate oscillations of the structure.
Fig 6: Density attached to the various brick elements.

Fluid Structure Coupling Time = 0 ms
Density
Velocity

Density
Velocity

Density
Velocity

6.2 - Fluid Flow
Title
Fuel tank - Fluid flow
Number
6.2
Brief Description
Fuel tank overturning with simulation of the fluid flow. The reversing tank is modeled using
horizontally-applied gravity. The tank container is presumed without deformation and only the
water and air inside the tank are taken into consideration using the ALE formulation.
Keywords
 Fluid flow simulation and ALE formulation
 Brick elements
 Hydrodynamic and bi-phase liquid gas (/MAT/LAW37)
RADIOSS Options
 J. Donea Grid Formulation (/ALE/GRID/DONEA)
Input File
Fluid_flow_gravity_1: <install_directory>/demos/hwsolvers/radioss/06_Fuel_tank/
2-Tank_overturning/Fluid_flow_1/PFTANK*
Fluid_flow_gravity_2: <install_directory>/demos/hwsolvers/radioss/06_Fuel_tank/
2-Tank_overturning/Fluid_flow_2/PFTANK*

Advanced

Overview
Aim of the Problem

The fluid flow is studied during the fuel tank overturning. This example uses the ALE (Arbitrary
Lagrangian Eulerian) formulation and the hydrodynamic bi-material law (/MAT/LAW37) to simulate
interaction between water and air. The tank container is presumed without deformation and it will
not be modeled.
A rectangular tank is partially filled with water, the remainder being supplemented by air. The tank
turns once around itself on the Y-axis. The overturning is achieved by defining a gravity field in the
X direction, which is parallel to the liquid gas interface. All gravity is applied in other directions. The
initial distribution pressure is already known and supposed homogeneous. The tank dimensions are
460 mm x 300 mm x 10 mm.
The example deals with two loading cases: an instantaneous rotation of the fuel tank by 90 degrees
(gravity function 1) and a progressive rotation (gravity function 2).
The main material properties for the ALE bi-phase air-water are:
 Air density: 1.22x10-6 g/mm3
 Water density: 0.001 g/mm3
 Gas initial pressure: 0.1 MPa

The bi-material air-water is described in the hydrodynamic material law (/MAT/LAW37). See
previous section for information about this law, including full input data.
This loading case does not require a tank container mesh and the model, air and water are only
comprised of the brick element using an ALE formulation.
Fig 8: Air and water mesh (ALE bricks).
Using the ALE formulation, brick mesh is only deformed by the tank deformation, the water flowing
through the mesh. The Lagrangian shell nodes still coincide with the material points, while the
elements are deformed with the material: this is the Lagrangian mesh. For the ALE mesh, nodes on
boundaries are fixed to remain on the border, while the interior nodes are moved.
Regarding the ALE boundary conditions (/ALE/BCS), constraints are applied on:
All nodes inside the border have grid and material velocities fixed in the Z direction; the nodes on
the left and right sides have a material velocity fixed in the X and Z directions, while the nodes on
the high and low sides have a material velocity fixed in the Y and Z directions. The grid velocity is
fully fixed on the border, just as the material velocity is fixed on the corners.

A function defines gravity acceleration in the X direction compared with time in order to simulate
the rotation effect. Gravity is activated by /GRAV. Two cases are studied depending on the
acceleration function selected:
Fig 9: Variable acceleration function 1 Fig 10: Constant acceleration function 2
Gravity is considered for all nodes.

 Both ALE materials air and water must be declared as ALE using /ALE/MAT. Lagrangian
material is automatically declared as Lagrangian.
 The /ALE/GRID/DONEA option activates the J. Donea grid formulation in order to compute grid
velocity. See the RADIOSS Theory Manual for further explanation about this option.

Model with Constant Acceleration

Time = 170 ms
(Gravity function 1)
Density
Velocity

Model with Constant Acceleration
Time = 280 ms
Density
Velocity

Model with Variable Acceleration
Time = 50 ms
Density
Velocity

Model with Variable Acceleration
Time = 70 ms
Density
Velocity
Conclusion
This example studied hydrodynamic bi-material using Law 37 in RADIOSS, using ALE and Eulerian
formulations. The application of boundary conditions in ALE formations and handling the fluid-
structure interaction were discussed. Furthermore, the results obtained correctly represent the
physical problem.

Example 7 - Pendulums
Summary
The purpose of this example is to simulate the oscillation and wave propagation of a group of
pendulums, arranged in a line, when impacted at one end. The material is described as being
elastic. Two models are used to simulate two different physical problems:
 The 2D model represents the infinite cylindrical mass for pendulums
 The 3D model is necessary for determining the spherical mass
The quality of the model first depends on how contact is managed. For the 2D model, a simple type
5 interface with a plane facet is used. For the 3D model, however, a type 16 interface using the
Lagrange Multipliers method is used.

Title
Pendulums
Number
7.1
Brief Description
Five pendulums in line, initially in contact with each other, are struck by a sixth one. The shock
wave and oscillating motion are observed.
Keywords
 Tri-dimensional analysis, truss, brick, and 16-node thick shell
 Type 16 interface (Node to brick contact)
 Elasticity, momentum transmission, shock wave propagation, and multiple-impacts
 Bi-dimensional analysis, plane strain, type 5 interface, and quad element
RADIOSS Options
 Bi-dimensional analysis (/ANALY)
 Type 16 interface (/INTER/LAGMUL/TYPE16) and type 5 (/INTER/TYPE5)

 Experimental and analytical results
Input File
Tri-dimensional_analysis:
<install_directory>/demos/hwsolvers/radioss/07_Pendulums/3D_model/PENDULUMS_3D*
Bi-dimensional_analysis:
<install_directory>/demos/hwsolvers/radioss/07_Pendulums/Plan_strain_model/
PENDULUMS_2D*

Medium

Overview
Aim of the Problem

The purpose of this example is to study the shock wave propagation and the momentum transfer
through several bodies, initially in contact with each other, subjected to multiple-impact. The
process of collision and the energetic behavior upon impact are delineated using a tri-dimensional
model. A plane strain assumption can be used as a compliment to this study, whereby a bi-
dimensional model using fine mesh enables shock wave propagation and the mechanics contact to
be shown in a qualitative manner.
A metal ball strikes a line of five balls, initially in contact with each other. The momentum is
transferred from pendulum to pendulum until reaching the last one at the opposite end. The system
is subjected to gravity. This results in the end pendulums alternate oscillating for half the time
period.
The following system is used: mm, ms, g, N, MPa.
Fig 1: Description of the problem.
The left pendulum has an initial angle of 45 degrees in relation to the vertical. The material used is
aluminum alloy which behaves like a linear elastic law (/MAT/LAW1) during impact.
The properties are defined as follows:
 Density: 0.0027 g.mm-3
The geometrical characteristics of the balls and trusses are:
 Truss:
- Length: 124.6 mm
 Ball:
- Radius: 25.4 mm (massball = 182.5g)

Two approaches 2D and 3D are used to provide complementary simulation results.

Modeling Methodology: 3D Model
Brick and thick shell elements are used to create the 3D model for balls. The quadratic 16-node
thick shell element is used to model the external surface of the balls. However, the core of each ball
is modeled using 8-node solid elements.
Fig 2: Tri-dimensional mesh in initial state.
The modeling technique used enables to ensure contact between the quadratic surfaces.
Figure 3 shows the mesh used for balls. The mesh uses a hypercube mesh topology combining brick
and 16-node thick shell elements.
Fig 3: Mesh for balls (brick and 16-node thick shell).

The type 16 interface using the Lagrange Multipliers method is employed to model contacts between
the nodes and the quadratic elements’ surface. An interface must be defined for each ball (five
interfaces).
Fig 4: Slave nodes and master surfaces defined for the type 16 interface.
No gap is required for the type 16 interface, enabling the contact condition to be exactly satisfied.
Gravity is applied to all nodes. A function defines the gravity acceleration in the Z direction
compared with time. Gravity is activated by the /GRAV option.
Fig 5: Gravity loading (-0.00981 mm.ms -2 ).

The upper extremities of the trusses are fixed in Y and Z translations and in Y and Z rotations.
Fig 6: Boundary conditions on the upper extremities of trusses.
Modeling Methodology: 2D Model

By adopting a plane strain approach, a 2D model is used (N2D3D = 2 in the /ANALY option set in the
input file). The plane strain analysis defines the X-axis as the plane strain direction.
The mesh consists of 2D solid elements (quads). The dimension of the quad is about 0.5 mm for
balls.
Fig 7: 2D mesh in the initial state.
Normal vectors of quad elements should have the same orientation to avoid negative volumes.
Quad elements undergo a type 14 general solid property.
The contact between the external segments of the quads is modeled five times using a type 5
interface.

Fig 8: Master segments and slave nodes defined for type 5 interfaces.
Type 5 interface uses the Penalty method for a master segment contact (blue side) to the slave
node (red side). The gap is set to 0.1 mm as the initial interval between the masses. The contact is
sliding using a Coulomb friction coefficient that is equal to zero.
Type 7 general interface is not available in a 2D analysis.
The upper extremities of the trusses are fixed in Y and Z translations. The 2D conditions are
automatically taken into account with N2D3D = 2 in /ANALY.
Gravity is applied to all nodes. A constant function (-0.00981 mm.ms-2) defines the gravity
acceleration in the Z direction compared with time. Gravity is activated by /GRAV.
For the 2D analysis, the rigid body /RBODY option is not available.
For the purpose of this example, the following numbers are assigned to the balls:
Fig 9: Ball numbers.

2D Approach: Dynamic System Analysis

Maintaining Energy and Oscillations
Figure 10 shows the kinetic energy variation of the model. When considering energy, the system
behaves as a simple pendulum.
Fig 10: Global energy assessment.
When the pendulum mass is released at time t=0, the No. 6 end ball has maximum potential energy
and null kinetic energy. Ball 6 achieves maximum velocity before striking the five other pendulums.
For a moderate case that is without loss, you have:
Where, h is the vertical displacement of the ball’s center, V is the velocity and m is the mass.
The maximum kinetic energy is reached for: h = hmax = I(1 - cos(45)) = 43.934 mm
EKINETICmax = mghmax = 182.5 * 0.00981 * 43.934 = 78.656 mJ
Analytical solution:
EKINETICmax = 78.655 mJ (time = 203.33 ms, impact balls 6 and 5)
Simulation results:
EKINETICmax = 72.478 mJ (time = 612.5 ms, impact balls 1 and 2)

Maintaining the kinetic energy in the system is not entirely satisfactory, due to the energy contact
being dissipated during impact.
The two extreme pendulums alternate, oscillating for half of the time period. The velocity of the
middle balls in comparison to time is shown in Fig 11.
Fig 11: Velocity transmission between the end balls 1 and 6.
Velocity is transferred from pendulum to pendulum until reaching the end one.
Equation of Motion
The relative motion of a simple pendulum can be described using the equation:
where, is the system’s pulsation:

Fig 12: Pendulum motion.
Such analytical equation can be corroborated with regard to the end balls No. 1 and 6.
Rotations and rotational accelerations are indicated from the nodes located at the upper end of
the trusses.
Fig 13: Verification of the equation for ball 6.

Fig 14: Verification of the equation for ball 1.
The numerical results have an average correlation in relation to the analytical solution, due to the
dynamic response of the nodal acceleration saved in the Time History.
Energetic Behavior upon Impact
Lets consider the interval [203,33 ms and 204,11 ms] where multiple impacts occur from balls No.
6 to 1.
As shown in Fig 15, the internal energy stored in the system is released after each impact, in line
with the defining balls linear material law. The kinetic energy is transferred from pendulum to
pendulum.

Fig 15: Global energy assessment during multiple impacts.
The 16-node thick shells are elements, which do not suffer hourglass deformation. Therefore, the
low kinetic energy lost during multiple impact is due to the dissipated contact energy (-2.47mJ).
The external work of the gravity remains constant (78.655mJ).
The following animations separately illustrate:
 the motion of the pendulums
 the kinetic energy transmission
 the stress wave propagation

Balls Motion (Oscillations)

Momentum transmission from pendulum to pendulum (cutting plane X = 0):
Velocity Norm
Time for total transmission: 0.78 ms Maximum = 1.08588 m.s -1.

Shock wave propagation during multiple impact (cutting plane X=0):
von Mises Stress Wave
Time for transmission: 0.78 ms Maximum = 17.7062 MPa.

2D Approach: Multiple Impact Analysis
In this section covers the mechanics contact across a 6-ball chain.
The plane strain assumption changes the physical problem. Nevertheless, this case study is an
interesting example of a system undergoing several shocks.
Fig 16: A 6-ball chain system.
The force between balls compared with time is shown in Fig 17. Existence of a time interval where
forces’ contacts are not at zero.
Fig 17: Forces’ contact between balls compared with time (contact starts at t’=0 ms).
This process leads to multiple impacts. It corroborates the experimental observations, where the
theory was well estimated. Based on an Impulse Correlation Ratio (ICR), a regularized system of an
N-ball chain using an elastic contact spring gives similar results.

 Reference results: [V. Acaray, B. Brogliato/Second MIT Conference on Computational Fluid
and solid Mechanics]
von Mises stress wave propagation from ball to ball during the multiple impact period (isostep
values):

Conclusion
The impact between several pendulums in line was studied using RADIOSS. Two models
representing physical problems were studied:
(i) a global analysis using a relatively coarse mesh with 3D elements
(ii) a 2D model using a fine mesh
In the first case, the energy assessment and the wave propagation are studied. The mesh used is
not fine enough for studying the contact effects, due to the fact that 3D represents a high cost
model and using a fine mesh dramatically increases the computation time. The results are
compared to an analytical solution where the pendulum system is assimilated to a simple
pendulum.
The 2D analysis concentrates on contact between the balls. There still exists an analytical solution
though for a chain of three balls, but which can be generalized for the purpose of this example. The
results obtained by simulation and theory demonstrate the validity of the numerical results obtained
by RADIOSS.

Example 8 - Hopkinson Bar
Summary
Precise data for high strain rate materials is necessary to enable the accurate modeling of high-
speed impacts. The high strain rate characterization of materials is usually performed using the split
Hopkinson Pressure Bar within the strain rate range 100-10000 s-1. Using the one-dimensional
analysis of the Hopkinson bar experiment, it is assumed that the object deforms under uni-axial
stress, the bar object interfaces remain planar at all times, and the stress equilibrium in the object
is achieved using travel times. The RADIOSS explicit finite element code is used to investigate these
assumptions.

Title
Split Hopkinson
pressure bar testing
Number
8.1
Brief Description
The high strain rate tensile behavior of the 7010 aluminum alloy is studied using the Hopkinson
pressure bar technique (stress wave).
Keywords
 Axisymmetrical analysis and quad elements
 High strain rate and Split Hopkinson Pressure Bar (SHPB)
 Wave propagation and stress pulse
 Elastic model (/MAT/LAW1) and Johnson-Cook elasto-plastic model (/MAT/LAW2)
RADIOSS Options
 Axisymmetrical analysis (/ANALY)
 Imposed velocities (/IMPVEL)

 Experimental data
Input File
High_strain_rate:
<install_directory>/demos/hwsolvers/radioss/08_Hopkinson_Bar/High_strain_rate/
SHPB_H*
Low_strain_rate:
<install_directory>/demos/hwsolvers/radioss/08_Hopkinson_Bar/Low_strain_rate/SHPB_L*

Advanced

Overview
Aim of the Problem

In order to model and predict the behavior of material during impact, the responses at very high
strain rates should be studied. The Split Hopkinson Bar is an inexpensive device for performing high
strain-rate experiments [1]. This equipment consists of four long pressure bars:
 the striker bar
 the incident bar
 the transmission bar
 the drop bar
The object is sandwiched between the transmission and the incident bar. Assuming that the wave
propagation in the bar is non-dispersive, the force and displacement upon contact between the bar
and the object can be obtained from the strains measured through experience. In this example, the
dynamic tensile behavior, achieved through experience of the 7010 aluminum alloy with a Split
Hopkinson Pressure Bar (SHPB) is compared to numerical simulations. Two cases are studied at the
strain rates of 80 s-1 (low rate) and 900 s-1 (high rate) respectively. At high strain rates, experience
shows that the stress flow significantly increases by more than 30% with the strain rate increasing;
thus demonstrating strain rate dependence in aluminum alloys in general. For the strain rates’
range applied here, an existing Johnson-Cook model is used to describe the stress flow as a strain
and strain rate function. Failure is not taken into account.
The Split Hopkinson Pressure Bar technique corresponds to a high strain rate deformation of the
aluminum alloy at high stress. Figure 1 shows a diagram of the basic Hopkinson bar setup. It
consists of two cylindrical bars of the same diameter, respectively called Input and Output bars.
Fig 1: Hopkinson bar device.
The objects material undergoes an isotropic elasto-plastic behavior which can be reproduced using a
Johnson-Cook model (/MAT/LAW2). The steel bars and the striker follow a linear elastic law
(/MAT/LAW1).
The following system is used: mm, ms, g, N, MPa.

Fig 2: Object geometry and cross-section (dimensions in mm).
Johnson-Cook Model
The Johnson-Cook model describes the stress in relation to the plastic strain and the strain rate
using the following equation:
where:
is the strain rate
0 is the reference strain rate
p is the plastic strain (true strain)

a is the yield stress
b is the hardening parameter
n is the hardening exponent
c is the strain rate coefficient
The two optional inputs, strain rate coefficient and reference strain rate, must be defined for each
material in /MAT/LAW2 in order to take account of the strain rate effect on stress, that is the
increase in stress when increasing the strain rate. The constants a, b and n define the shape of the
strain-stress curve.
In the documents entitled CRAHVI, G4RD-CT-2000-00395, D.1.1.1, Material Tests – Tensile
properties of Aluminum Alloys 7010T7651 and AU4G Over a Range of Strain Rates, the behavior of
the 7010 aluminum alloy can be described according to the relations:
for strain rates below 80 s-1
for strain rates exceeding 80 s-1 up to 3000 s-1

Fig 3: Yield curve of the Johnson-Cook model:
The material properties of the object are:

The material used for the bars and projectile is type 1 (linear elastic) with the following properties:
The geometrical characteristics of the bars and projectile are:
Bars:
 Length: 4 m
 Diameter: 12 mm

Projectile:
 Radius: 12 mm
 Weight: 170 g
High Strain Rate Test Method

The object is screwed in between the incident and transmission bars. A stress pulse is introduced
into the input bar through impact from a steel projectile on the steel disc attached to one end of the
input bar. The impact generates a tensile wave which propagates along the input bar. Part of the
wave is reflected and a part is transmitted via the object’s interface. The stress pulse continues
through the object and into the transmitted bar. The wave reflections inside the sample enable the
stress to be homogenized during the test. The strain associated with the output or transmitted
stress wave is measured by the strain gauges on the output or transmitted bar. The strain gauges
attached to the object gauge length, provide direct measuring of the true strain, and the true plastic
strain in the object during the experiment. The transmitted elastic wave provides a direct force
measurement to the bar object interfaces by way of the following relation:
Where, Ebar is the modulus of the output bar, T is the strain associated with the output stress wave
and the Sbar is the cross-section of the output bar.
If the two bars remain elastic and wave dispersion is ignored, then the measured stress pulses can
be assumed to be the same as those acting on the object.
The engineering stress value in the object can be determined by the wave analysis, using the
transmitted wave:
Engineering stress can also be found by averaging out the force applied by the incident that is the
reflected and transmitted wave, as shown in the equation:
Where, I and R are the strains associated with input stress wave and T is the strain associated
with output stress wave.
True stress in the object is computed using the following relation (refer to Example 11 - Tensile Test
for further details):
The true strain rate is given by:
True stress and true strain are evaluated up to the failure point.

Interface 1: F1 = Sbar ( I(t) + R(t)) = SbarEbar( I(t) + R(t))
Interface 2: F2 = Sbar T(t) = SbarEbar T(t)
Balance in object: F1 = F2 ; 1(t) + R(t) = T(t)
Engineering stress in object: object (t) = F1 / Sobject = F2 / Sobject

Fig 4: 1D analysis.
Strain Rate Filtering

Because of the dynamic load, strain rates cause high frequency vibrations which are not physical.
Thus, the stress-strain curve may appear noisy. The strain rate filtering option enables to dampen
such oscillations by removing the high frequency vibrations in order to obtain smooth results. A cut-
off frequency for strain rate filtering (Fcut) is used with a value less than half of the sampling
frequency (1/ This or 1/ Tsampling) defined in the Engine file (*_0001.rad) using the /TFILE option.
Refer to Example 11 - Tensile Test for further details.
The cut-off frequency is set at 100 kHz in this example.

Taking into account the geometry’s revolution symmetry the material and the kinematic conditions,
an axisymmetrical model is used (N2D3D = 1 in /ANALY set up in the Starter file). Y is the radial
direction and Z is the axis of revolution.
The mesh is made of 12054 2D solid elements (quads). The quad dimension is about 2 mm.
Fig 5: Mesh of the axisymmetrical model with imposed velocities on the top of the input bar.

Low extremity nodes of the output bar are fixed in the Z direction. The axisymmetrical condition on
the revolutionary symmetry axis requires the blocking of the Y translation and X rotation.
The projectile is modeled using a steel cylinder with a fixed velocity in the direction Z. The required
strain rate is taken into account by applying two imposed velocities, 1.7 ms-1 and 5.8 ms-1 in order
to produce strain rate ranges in the of 80 s-1 and 900 s-1 (low and high rates) object.

True Stress, True Strain and True Strain Rate Measurement from Time History
Fig 6: Nodes and quads saved for Time History.
In the experiment, the strain gauge is attached to the object. In simulation, the true strain will be
determined from 9040 and 6 nodes’ relative Z displacements (l 0 = 3.83638 mm).
The true stress can be given using two data sources. The first methodology consists of using the
equation previously presented, based on the assumption of the one-dimensional propagation of bar-
object forces. The engineering strain t associated with the output stress wave is obtained from the
Z displacement of nodes located on the output bar. The true plastic strain is extracted from the
quads on the object, saved in the Time History file. True stress can also be measured directly from
the Time History using the average of the Z stress quads 6243, 6244, 6224 and 6235. It should be
noted that the section option is not an available option with the quad elements.
The strain rate can be calculated from either the true plastic strain of quads saved in /TH/QUAD or
from the true strain true.

Table 1: Relations used in the analysis
High Rate Testing
Z stress average from quads saved in

True stress
/TH
True strain
True strain rate
The purpose of this test is to obtain the results observed in experiments with a Johnson-Cook
model. The increase of stress is expected to equal approximately 30% compared to the low strain
rate test.
Experimental Data
Experimental results show that the variation of the true tensile flow stress compared with the true
strain is approximately equivalent to a strain rate between 80 s -1 and 100 s-1. The reference strain,
in the Johnson-Cook model is set to 0.08 ms-1. At higher rates, the true flow stress increases
significantly compared with the strain rate. The 7010 aluminum alloy exhibits an increase in the
flow stress by a typical 30% at high strain rates (900 s -1 to 3000 s-1) compared to static values.
Results are given at the specific true strains of 0.02, 0.05 and 0.10. The influence of the strain rate
on stress can be seen in Fig 7 [1].

Fig 7: Variation of true stress compared with true strain for 7010 alloy using two different rates (experimental data).
For the test performed with a strain rate of 900 s-1, the flow stress reaches 850 MPa at a 0.25
strain.
Table 2: True stress at specific strains using both strain rates (experimental data).
Strain rate: 80 s-1 Strain rate: 900 s-1
True strain 0.02 0.05 0.1 0.02 0.05 0.1 0.25
True stress (MPa) 550 600 610 625 775 800 850
Johnson-Cook Model
Figure 8 shows the variation of true stress in time in relation to the wave propagation along the
bars. Stresses are evaluated on the input bar, the object and the output bar.
Fig 8: Stress measurement localizations.

Fig 9: Stress waves in the input bar, the output bar and the object (imposed velocities = 5.8 ms -1 ).
The stress-time curve shows the incident, reflected and transmitted signals.
Fig 10: Diagram of SHPB showing the motion in time of the tensile pulse.

Fig 11: von Mises stress wave propagation along bars (imposed velocities = 5.8 ms -1 ).
The speed of wave, C along the bars is calculated using the relation:
C = 5189 ms-1
Where, E is the Young’s modulus and is the density of the bars.
The time step element is controlled by the smallest element located in the object. It is set at 5x10 -5
ms. The stress wave thus reaches the object in 0.77 ms and travels 0.26 mm along the bar for each
time step. Obviously, it remains lower than the element length of the smallest dimension (0.88
mm).
An imposed velocity of 5.8 ms-1 produces a strain rate in the object of approximately 900 s -1, while
a strain rate of approximately 80 s-1 is achieved using an imposed velocity of 1.7 ms-1. A simulation
is performed for each velocity value. It should be noted that the study on low rates is more limited
in time than on high rates due to the reflected wave generated on top of the output bar.
Figure 12 shows the true stress and true strain as a function of the strain rate.

Fig 12: Variation of true stress with true strain for high and medium strain rates.
At a high strain rate (900/s), an increase in the flow stress is observed, being approximately 30%
higher than the stress obtained for a low strain rate (80/s). The Johnson-Cook model used provides
precise results compared with the experimental data.

Fig 13: Stress Z and plastic strain on object at 0.6 ms.
The true stresses determined from both methodologies are shown side-by-side. This validates the
analysis based on a transmitted wave. Typical curves for a model having imposed velocities equal to
5.8 ms-1 are shown below:

True stress comparison in the object
True strain rate in the object

(using both computations)

Either data sources used to evaluate the strain rate give similar results.
The following results show:
 the strain rate effect on stress, with or without the cut-off frequency for smoothing (100 kHz);
 the influence of the strain rate coefficient (comparison with experimental data).
Strain rate effect

Influence of the strain rate coefficient c
These studies are performed for the high strain rate model ( = 900 s-1).
Figure 14 compares the distribution of the von Mises stress on the object, with and without the
strain rate filtering at time t=0.6 ms.

Fig 14: Comparison of the distribution of the von Mises stress at time t=0.6 ms.
More physical flow stress distribution is obtained using filtering. Explicit is an element-by-element
method, while the local treatment of temporal oscillations puts spatial oscillations into the mesh.
Reference
[1] CRAHVI, G4RD-CT-2000-00395, D.1.1.1, Material Tests – Tensile properties of Aluminum Alloys
7010T7651 and AU4G Over a Range of Strain Rates.

Example 9 - Billiards (pool)
Summary
The impact and rebound between balls on a small billiard table is studied. This example deals with
the problem of defining interfaces and transmitting momentum between the balls. The study is
divided into three parts:
At first, a general study is used to see the results of a cue ball when coming into contact with the 15
other balls arranged in a triangle. The balls are meshed for the purpose using 16-node shell
elements (for the curvature) and a type 16 interface between each ball as well as between the balls
and the table. The results show that the momentum is not homogenously transmitted: the balls on
the table are not being evenly spread out.
Secondly, the collision between two balls is studied. All parameters are the same as in the first part.
The reaction of those two balls is then compared to the analytical results.
Finally, six different interfaces are compared: types 16 and 17 tied or sliding interfaces using the
Lagrange Multipliers method and a type 7 tied or sliding interface using the Lagrange Multipliers or
the Penalty method. The study is also initiated using a quasi-static gravity application prior to
dynamic behavior. When comparing the kinetic energy transmission, the results show that
interfaces without the tied option provide better results than the others, and that the type 16
interface seems to be the best.

9.1 - Billiards (Pool)
Title
Billiards (Pool)
Number
9.1
Brief Description
A pool game is modeled to show the transmission of momentum between one impacting ball and 15
impacted balls.
Keywords
 16-node thick shell and sphere mesh
 Type 7 interface using the Lagrange Multipliers method and the Penalty method
 Type 16 sliding and tied interface, and quadratic surface contact
 Elastic shock
 Momentum transmission and shock wave
RADIOSS Options
 Type 7 interface (/INTER/TYPE7) and type 16 (/INTER/LAGMUL/TYPE16)
 16-node thick shell property type 20 (/PROP/TSHELL)
Input File
Billiard_game/Interface_16:
<install_directory>/demos/hwsolvers/radioss/09_Billiards/Billiards_model/BILLARD*
<install_directory>/demos/hwsolvers/radioss/09_Billiards/Billiards_model/Supplement_
Interface7Lag/BILLARD*

Advanced

Overview
Aim of the Problem

The purpose of this example is to investigate the transmission of momentum between several balls.
Contact with the various interfaces using the Penalty and Lagrange Multipliers’ method is analyzed.
Pool is a game consisting of 16 balls, each 50.8 mm in diameter. It is played on a small billiard
table measuring 1800 mm x 900 mm. Fifteen (15) balls are placed in a triangle to enable their tight
grouping. The initial velocity of the shooting (cue) ball is presumed equal to 1.5 ms -1. Elastic
rebounds are observed.
Fig 1: Pool game.
Units: mm, g, N, MPa.

The material is subjected to a linear elastic law (/MAT/LAW1) with the following properties:
Balls: Frame: Plate:
phenolic resin polymer slate
Initial density 0.00137 g.mm-3 0.001 g.mm-3 0.0028 g.mm-3

Young's modulus 10500 MPa 1000 MPa 62000 MPa
Poisson ratio 0.3 0.49 0

The balls are meshed with 16–node solid shells (quadratic elements) in order to improve the
conditions of contact by taking into account the curvatures. The frame of the table is made of 16–
node solid shells to comply with the interface used. The plate is modeled using only one solid
element. The 16–node thick shells are considered as solid elements. They are defined by a thick
type 20 shell property (number 16 solid formulation for quadratic 16-node thick shells, fully-
integrated with 2x2x2 integration points).
Fig 2: Pool game mesh.
Fig 3: Mesh for balls.

Fig 4: 16-node thick shell element.
The type 16 interface with the Lagrange Multipliers method is used to model the ball/ball and
balls/table contacts. An interface must be defined for each ball (that is: 16 interfaces in total). An
additional interface is used to define the contacts between the balls and the table (plate and frame).
Fig 5: Type 16 interface: slave SHEL16 for balls and master SHEL16 for the table.
Fig 6: Example of the type 16 interface defined for the contact between balls.
Slave nodes (red) are extracted from the external surfaces of the parts.

An initial velocity of 1.5 ms-1 in X direction is applied to all nodes of the white (cue) ball.
Fig 7: Initial translational velocities of the impacting ball.
All nodes of the lower face of the table are completely fixed (translations and rotations).
Gravity is considered for all the balls nodes. A function defines the gravity acceleration in the Z
direction compared with time. Gravity is activated using /GRAV.
Fig 8: Gravity function (-0.00981 mm.ms-2 )

Due to the faceting of the ball, contact between the impacting ball and the impacted balls is not
perfectly symmetrical and momentum is not homogeneously transmitted among the balls. An
apparent physical strike thus results.
Fig 9: Collision of the balls

Fig 10: History of the balls’ motions (contact control: type 16 interface).

9.2 - Collision between Two Balls
Study on Trajectories
Title
Collision between two
balls
Number
9.2
Input File
Collision study:
<install_directory>/demos/hwsolvers/radioss/09_Billiards/Collision_simulation/
COLLISION*

Overview
Two balls are now considered in order to study the behavior of impacting spherical balls.
The balls’ behavior is described using the parameters (angles and velocities) shown in Fig 11. The
numerical results are compared with the analytical solution, assuming a perfect elastic rebound
(coefficient of restitution is equal to 1).
Fig 11: Problem data.
Initial values: V1 = 0.7m.s-1; V2 = 1m.s-1; 1 = 40°; 2 = 30; massball = 44.514g.
The balls and the table have the same properties, previously defined for a pool game. The
dimensions of the table are 900 mm x 450 mm x 25 mm and the balls’ diameter is 50.8 mm. The
balls and the table are meshed with 16-node thick shell elements for using the type 16 Lagrangian
interface.

Fig 12: Mesh of the problem (16-node thick shells).
The initial translational velocities are applied to the balls in the /INIV Engine option. Velocities are
projected on the X and Y axes.
Fig 13: Initial velocities applied on the balls (initial position).
Gravity is considered for the balls (0.00981 mm.ms -2 ).

The ball/ball and balls/table contact is modeled using the type 16 interface (slave nodes/master 16-
node thick shells contact). The interface defining the ball/ball contact is shown in Fig 14.

Fig 14: Master and slave sides for the type 16 Lagrangian interface.
Analytical Solution
Take two balls, 1 and 2 from masses m1 and m2, moving in the same plane and approaching each
other on a collision course using velocities V 1 and V2, as shown in Fig 15.
Fig 15: General problem of collision between two balls.
Velocities are projected onto the local axes n and t. To obtain the velocities and their direction after
impact, the momentum conservation law is recorded for the two balls:
or
The shock is presumed elastic and without friction. Maintaining the translational kinetic energy is
respected as there is no rotational energy:
Such equality implies that the recovering capacity of the two balls corresponds to their tendency to
deform.

This condition equals one of the elastic impacts, with no energy loss. Maintaining the system’s
energy gives:
This relation means that the normal component of the relative velocity changes into its opposite
during the elastic shock (coefficient of restitution value e is equal to the unit).
The following equations must be checked for normal components:
The equations system using V’1 and V’2 as unknowns is easily solved:
It should be noted that these relations depend upon the masses ratio.
As the balls do not suffer from velocity change in the t-direction, maintaining the tangential
component of each sphere’s velocity provides:
The norms of velocities after shock result from the following relations.
In this example, balls have the same mass: m1 = m2.
Therefore:
and

The norms of the velocities are given using the following relations, depending on the initial
velocities and angles. Used to determine the analytical solutions (angles and velocities after
collision):
By recording the projection of the velocities, directions after shock can be evaluated using relation.
Used to determine the analytical solutions (angles and velocities after collision):

Simulation Results: Comparison of Numerical Results with the Analytical
Solution
The following diagram shows the trajectories of the balls’ center point obtained using numerical
simulation before and after collision.
Fig 16: Trajectories of balls (center of gravity).

Fig 17: Variation of velocities (collision at 40 ms).

Fig 18: Energy assessment.
For given initial values of V1, V2, 1 and 2, simulation results are reported in Table 1.
Table 1: Comparison of results for after collision
Numerical Results Analytical Solution
1’ 42.27° 1’ 44.72°
2’ 26.75° 2’ 26.48°
V 1’ 0.731 m/s V 1’ 0.731 m/s
V 2’ 0.969 m/s V 2’ 0.977 m/s

Conclusion
The simulation corroborates with the analytical solution. The 16-node thick shells are fully-
integrated elements without hourglass energy. This modeling provides a good transmission of
momentum. However, the type 16 interface does not take into account the quadratic surface on the
slave side (ball 2), due to the node to thick shell contact. Accurate results are obtained for a
collision without penetrating the quadratic surface of the slave side in order to confirm impact
between the spherical bodies.
A fine mesh could improve the results.

9.3 - Study on Interfaces
Comparison of Results Obtained using Different Interfaces
Title
Study on interfaces
Number
9.3
Input File
Inter_7_Penalty:
<install_directory>/demos/hwsolvers/radioss/09_Billiards/Contact_modelling/
Inter_7_Penalty/TEST7P*
Inter_7_Lagrangian:
Inter_7_Lagrangian/TEST7L*
Inter_16_tied:
Inter_16_tied/TEST16T*
Inter_16_sliding:
Inter_16_sliding/TEST16S*
Inter_17_tied:
Inter_17_tied/TEST17ST*
Inter_17_sliding:
Inter_17_sliding/TEST17S*
Overview
The balls and the table have the same properties as previously defined. The dimensions of the table
are 900 mm x 450 mm x 25 mm and the balls’ diameter is 50.8 mm.
Six interfaces are used to model the contacts (ball/ball and balls/table):

Table 2: Interfaces used in the problems.
Type 16 (Lagrange Multipliers) tied or sliding:

slave nodes/master solids contact
Type 17 (Lagrange Multipliers) tied or sliding:

slave 16-node shells/master 16-node shells contact
Type 7 (Lagrange Multipliers):

slave nodes/master surface contact
Type 7 (Penalty) sliding:

slave nodes/master surface contact
The type 16 interface defines contact between a group of nodes (slaves) and a curved surface of
quadratic elements (master part). The type 17 interface is used for modeling a surface-to-surface
contact. For both interfaces, the Lagrange Multipliers method is used to apply the contact
conditions; gaps are not required. Contact between the balls and the table is set as tied or sliding.
Contact between the balls themselves is always considered as sliding. The type 7 interface enables
the simulation of the most general contact types occurring between a master surface and a set of
slave nodes. The Coulomb friction between surfaces is not modeled here (sliding contact) and the
gap is fixed at 0.1 mm. The other parameters are set to default values.
The type 7 interface with the Penalty method is not available with 16-node thick shell elements.
Thus, brick elements replace the 16-nodes shells in this case (check in the input file).
Contact modeling between balls (always sliding).
Fig 19: Definition of slave and master sides for contact.

The symmetrical interface definition is not recommended when using the Lagrange Multipliers
method (types 16, 17 and 7-Lag). The problem using the interface with the Penalty method uses
two interfaces to model the symmetrical impact.
Fig 20: Symmetrical configuration of the type 7 interface using the Penalty method
Interface Slave (red) and Master (blue) Objects
Slave: nodes
Type 16 – tied
Master: solids (16-node shell)
Slave: nodes
Type 16 – sliding
Slave: 16-node shell

Type 17 – tied
Master: 16-node shell

Type 17 – sliding
Type 7 – Lagrange Slave: nodes

Multipliers Master: surface (segments)
Slave: nodes
Type 7 – Penalty method
Master: surface (segments)

Contact between the balls and the table (sliding or tied depending on the problem):
Fig 21: Definition of slave and master objects for balls/table contacts.

Interface Slave (red) and Master (blue) Objects
Slave: nodes
Type 16 – tied
Slave: nodes
Type 16 – sliding

Type 17 – tied

Type 17 – sliding
Type 7 – Lagrange Slave: nodes

Multipliers Master: surface (segments)
Slave: nodes
Type 7 – Penalty method
Master: surface (segments)
Pre-loading: quasi-static gravity loading to reach static equilibrium.

The explicit time integration scheme starts with nodal acceleration computation. It is efficient for
the simulation of dynamic loadings. However, a quasi-static simulation via a dynamic resolution
method needs to minimize the dynamic effects for converging towards static equilibrium and
describes the pre-loading case before the dynamic analysis. Thus, the quasi-static solution of
gravity loading on the model shows a steady state in the transient response.
To reduce the dynamic effect, dynamic relaxation can be used (/DYREL in the Engine file). A
diagonal damping matrix proportional to the mass matrix is introduced into the dynamic equation:
with, being the relaxation value by default, equal to 1, and T being the period to be damped (less
than or equal to the largest period of the system).
Thus, a viscous stress tensor is added to the stress tensor:
In an explicit code, the application of the dashpot force modifies the velocity equation:
without relaxation
with relaxation
with:

This option is activated in the Engine file (*_0001.rad) using /DYREL (inputs: = 1 and T = 0.2).
The dynamic problem (impact between balls) is considered in a second run managed by the second
Engine file (*_0002.rad) with a time running from 30 ms to 130 ms.
Simulation Results: Kinetic Energy Transmission between Balls during

Collision
Type 17 Interface
Contact between
quadratic surfaces
Balls/table contact: tied

Ball/ball contact: sliding
Type 17 Interface
Contact between
quadratic surfaces
Balls/table contact:
sliding

Type 16 Interface
Contact nodes/ quadratic

surface
Balls/table contact: tied

Type 16 Interface
Contact nodes/ quadratic
surface
sliding
Type 7 Interface
Lagrange Multipliers
method
Contact nodes/ linear

surface
(sliding contact)

Type 7 Interface
Penalty method
Contact nodes/ linear

surface
sliding

Conclusion
Interface 7
Interface Interface Interface 17 Interface Lagrange Interface 7
16 Tied 16 Sliding Tied 17 Sliding Multipliers Penalty
Cycles 241392 241385 241387 241385 241385 773099
Error on
-30.8% -1.4% -55.5% -10.8% -1.2% -46.1%
Energy
Rolling yes no yes no no no
Momentum quasi-
partial partial good good partial
Transmission perfect
master
Quadratic master master master and
and slave no no
surface side side slave sides
sides
A non-elastic collision appears using the type 7 interface Penalty method. After impact, each ball
has about half of the initial velocity. The momentum transmission is partial and can be improved by
increasing the stiffness of the interface despite the hourglass energy and degradation of the energy
assessment.
Error on energy is more noticeable for interfaces using the Tied option, due to taking into account
the rolling simulation.
This study shows the high sensitivity of the numerical algorithms for the modeling impact on elastic
balls. Regarding the interface type, the kinematics of the problem and the transmission of
momentum are more or less satisfactory. Type 16 interface allows good results to be obtained.

Example 10 - Bending
Summary
The bending of a straight cantilever beam is studied. The example used is a famous bending test for
shell elements. The analytical solution enables the comparison with the quality of the numerical
results. Carefully watch the influence from the shell formulation. In addition, the results for the
different time step scale factors are compared.

Title
Bending
Number
10.1
Brief Description
Pure bending test with different 3- and 4-nodes shell formulations.
Keywords
 Q4 and T3 meshes
 QEPH, Belytshcko & Tsay, BATOZ, and DKT shells
 Mesh, hourglass, imposed velocity, quasi-static analysis, and bending test
RADIOSS Options
 Rigid bodies (/RBODY)

Input File
BATOZ: <install_directory>/demos/hwsolvers/radioss/10_Bending/BATOZ/.../ROLLING*
QEPH: <install_directory>/demos/hwsolvers/radioss/10_Bending/QEPH/.../ROLLING*
BT (type1):
<install_directory>/demos/hwsolvers/radioss/10_Bending/BT/BT_type1/.../ROLLING*
BT (type3):
BT (type4):
DKT18: <install_directory>/demos/hwsolvers/radioss/10_Bending/DKT18/.../ROLLING*

Beginner benchmark

Overview

The purpose of this example is to study a pure bending problem. A cantilever beam with an end
moment is studied. The moment variation is modeled by introducing a constant imposed velocity on
the free end.
The following system is used: mm, ms, g, N, MPa
Several kinds of element formulation are used.
The material used follows a linear elastic law (/MAT/LAW1) and has the following characteristics:
 Initial density: 0.01 g/mm3
 Reference density: .01 g/mm3
 Young modulus: 1000 MPa
 Poisson ratio: 0
Fig 1: Geometry of the problem.
Three beams are modeled using quadrilateral shells and one beam with T3 shells. A rigid body is
defined at the end of each beam for applying the bending moment.
The four models are integrated into one input file. The shell element formulations are:
 Q4 mesh with the Belytshcko & Tsay formulation (I shell =1, hourglass control type 1, 2, and 3)
 Q4 mesh with the QEPH formulation (I shell =24)
 Q4 mesh with the QBAT formulation (I shell =12)
 T3 mesh with the DKT18 formulation (I shell =12)

At one extremity of the beam, all DOF are blocked. A rotational velocity is imposed on the master
node of the rigid body placed on the other side.
This velocity follows a linear function: Y=1
Fig 2: Beam meshes.
Numerical Results Compared to Analytical Solutions

As shown in Fig 1, rotation around X and displacement with regard to Y of the free end are studied.
The analytical solution of the Timoshenko beam subjected to a tip moment reads:
which yields the end moment for a complete loop rotation 2 :
The following tables summarize the results obtained for the different formulations. From an
analytical point of view, the beam deformed under pure bending must satisfy the conditions of the
constant curvature which implies that for = 2 , the beam should form a closed ring. However,

depending on the finite element used, a small error can be observed, as shown in the following
tables. This is mainly due to beam vibration during deformation as it is highly flexible. Good results
are obtained by the QBAT, QEPH and DKT18 elements, respectively. This is mainly due to the good
estimation of the curvature in the formulation of these elements. The BT family of under-integrated
shell elements is less accurate. With the type 3 hourglass formulation, the model remains stable
until = 6rad. However, the moment-rotation curves do not correspond to the expected response.
To reduce the overall computation error, smaller explicit time steps are used by reducing the scale
factor in /DT. The results reported in the end table show that a reduction in the time step enables to
reduce the error accumulation, even though the divergence problems for BT elements cannot be
avoided.
The following parameters are chosen for drawing curves and displaying animations:
BATOZ QEPH BT DKT
Scale factor 0.6 0.9 0.9 0.2
Imposed 0.005 0.005 0.005 0.005

velocity rot. rad/ms rad/ms rad/ms rad/ms

The following curves show the evolution previously shown (rotation and nodal displacement by
moment):
Fig 3: Moment versus rotation around X.
For
Fig 4: Moment versus displacement along Z.

Fig 5: Moment versus rotation around X.

BATOZ QEPH BT DKT
Type 1 Type 3 Type 4 Sf =0.3 Sf =0.2 Sf =0.1

Sf=0.9 Sf =0.8 Sf =0.6 Sf =0.9 Sf =0.8
Sf =0.9 Sf =0.1 Sf =0.9 Sf =0.1 Sf =0.9 Sf =0.1
CPU
(normalized)
2.18 2.43 3.14 1.23 1.34 42.64 7.07 2.62 108.60 1.03 7.17 5.44 8.21 16.21
# cycles
97600 109800 146400 95800 107800 59100 552600 182300 -- 59100 552600 364100 621600 1243200
Error
=2 0% 0% 0% 0% 0% 55.3% 99% 0% 0% 55.9% 99.9% 3.4% 28.88% 3.7%
(%)
err =20% 6.91 6.89 4.36 4.53 6.06 5.98 4.38 4.51 6.37
(rad)
-- -- -- -- --
degree 396° 395° 250° 260° 347° 343° 251° 258° 365°
Dz = 2п -500.5 -500.5 -500.5 -500.5 -500.5 -491.2 -525.8 -518.333 -506.0 -529.8 -433.8 -476.5 -496.5 -499.4
(mm)
Mx = 2п
+5 -4.04 -4.05 -4.06 -4.01 -4.01 -0.21 -0.11 -3.13 -2.38 -0.07 -0.02 -3.09 -3.02 -3.08
(x10 kN-
mm)

Conclusion
A description summary of the different tests is provided below:

 QBAT element:
This formulation gives a 2 -revolution of the beam with no energy error. However, a 20%
error is attained for = 384 degrees.
The decrease of the scale factor enables obtaining better results.
 QEPH element:
This formulation seems to be the best one to treat the problem. It enables a 2 -revolution
of the beam to be obtained. The error remains null until = 400 degrees.
 BT formulation:
This formulation does not provide satisfactory results and is not adapted to this simulation,
whatever the anti-hourglass formulation. This is mainly due to using a flat plate formulation
and the fact that the element is under-integrated. The type 3 hourglass formulation seems
to be better than others.
 For DKT formulation:
The bending is simulated correctly. However, the element is costly and the CPU time is
much longer.

Example 11 - Tensile Test
Summary
The material characterization of ductile aluminum alloy is studied. The RADIOSS material laws 2,
27 and 36 are used to reproduce the experimental data of a traction test by simulation. The
work-hardening, damage and rupture of the object are simulated by a finite element model. The
parameters of the material laws are determined to fit the experimental results. The influence of
the strain rate is also studied. A strain rate filtering method is used to reduce the effect of a
dynamic resolution on the simulation results.

11.1 - Law Characterization
Title
Law characterization
Number
11.1
Brief Description
Elasto-plastic material law characterization using a tensile test.
Keywords
 Shell element
 Johnson-Cook elasto-plastic model (/MAT/LAW2)
 Tabulated elasto-plastic (/MAT/LAW36)
 Elasto-plastic brittle (/MAT/LAW27)
 Necking point, damage model, maximum stress, and failure plastic strain
RADIOSS Options
 Material definition (/MAT)

Input File
Law_2_Johnson_Cook:
<install_directory>/demos/hwsolvers/radioss/11_Tensile_test/Law_2_Johnson-
Cook/.../TENSIL2*
Law 27_Damage:
<install_directory>/demos/hwsolvers/radioss/11_Tensile_test/Law_27_Damage/DAMAGE*
Law_36_Tabulated:
<install_directory>/demos/hwsolvers/radioss/11_Tensile_test/Law_36_Tabulated/
TENSI36*

Advanced

Overview
Aim of the Problem

It is not always easy to characterize a material law for transient analysis using the experimental
results of a tensile test. The purpose of this example is to introduce a method for characterizing
the most commonly used RADIOSS material laws for modeling elasto-plastic material. The use of
"engineering” or "true” stress-strain curves is pointed out. Damage and failure models are also
introduced to better fit the experimental response.
Apart from the experimental results, the modeling of the strain rate effect on stress will be
considered at the end of this example using a sensitivity study on a set of parameters for
Johnson-Cook’s model.
Traction is applied to an object. A quarter of the object is modeled using symmetrical conditions.
The material to be characterized is 6063 T7 Aluminum. A velocity is imposed at the left-end.
Units: mm, ms, g, N, MPa.
Fig 1: Geometry of the tensile object (One quarter of the object is modeled).
The material undergoes isotropic elasto-plastic behavior which can be reproduced by a Johnson-
Cook model with or without damage (/MAT/LAW27 and /MAT/LAW2, respectively). The tabulated
material law (/MAT/LAW36) is also studied.

Fig 2: Experimental results of the tensile test: engineering stress vs. engineering strain.
The average element size is about 2 mm in the mesh (Fig 3). There are 201 4-node shells and
one 3-node shell.
The shell properties are:
 5 integration points (progressive plastification).
 Belytschko elasto-plastic hourglass formulation (Ishell = 3).
 Iterative plasticity for plane stress (Newton-Raphson method; Iplas = 1).
 Thickness changes are taken into account in stress computation (Ithick = 1).
 Initial thickness is uniform, equal to 1.7 mm .

Fig 3: Mesh of the object.
 Node number 54 was renamed "Node 1" to be compliant with the Time History.
For node 54, only displacements in the x-direction (variable DX) are saved.
Fig 4: Sections saved for Time History.
For both sections, the variables FN and FTX, are saved; thus the following variables will be
available in /TH/SECTIO: FNX, FNY, FNZ (saved using "FN"), and FTX.
Engineering strains will be obtained by dividing the displacement of node 1 with the distance up
to the symmetry axis (75 mm). Engineering stresses will be obtained by dividing the force
through section 1 with its initial surface (10.5 mm 2). Therefore, the results shown correspond to
the engineering stress as a function of the engineering strain, equivalent to the force variation
compared to displacement (similar curve shape).
An imposed velocity of -1.0 m/s in the x-direction is applied to the nodes, shown below (abscissa
less than or equal to 25 mm). The displacement is proportional to time.
Fig 5: Imposed velocities

Fig 6: Variation of node 1 x-displacement in relation to time.
Only one quarter of the object is modeled to limit the model size and to eliminate the rigid body
motions. Symmetry planes are defined along axis x = 100 mm and axis y = 0. Two boundary
conditions cannot be applied to the same node 13 (corner).
Fig 7: Boundary conditions
The lower side is fixed in Y and Z translations and X, Y, and Z rotations.

The right side is fixed in X and Z translations and X, Y, and Z rotations; the node in the corner is
completely fixed.
Characterization of the Material Law
There are two steps to characterize the material law:
 Transform the engineering stress versus engineering strain curve into a true stress versus
true strain curve (this step applies to any material law).
 Extract the main parameters from the true stress versus true strain curve, to define the
material law (Johnson-Cook law and material coefficients for /MAT/LAW2 or the yield curve
definition for /MAT/LAW36).
- True stress/true strain curve
Engineering strains are computed using the following relationship:
And true strains are computed with the relationship:
Both strains, therefore, are linked together by:
Engineering stresses are measured by dividing the force through one section with the initial
section. True stresses are measured by dividing the force with the true deformed section:
Thus, to compute true stresses, the surface variation must be taken into account. Assuming that
Poisson’s coefficient is 0.5 during plastic deformation, the true surface in mono-axial traction is:
Thus, the relationship between true and engineering stresses is:
Characterization of the Material Law

The characterization will be made for /MAT/LAW2 (Johnson-Cook elasto-plastic), /MAT/LAW27
(elasto-plastic with damaged model) and /MAT/LAW36 (tabulated elasto-plastic). For each of the
material laws, the yield stress and Young’s modulus are determined from the curve.
The plastic strain can be defined as:
An important point to be characterized on the curve is the necking point, where the slope of the
force versus the displacement curve is equal to 0, and where the following relationships apply:

Fig 8: Guidelines for necking point.
Table 1: Equations used for analysis
Material
Generic Equation
Property
Engineering stress
Engineering strain
True stress
True strain
True strain rate

Experimental Results
An experiment designed by the "Norwegian Institute of Technology" as part of an EC-financed
program, "Calibration of Impact Rigs for Dynamic Crash Testing" is used. The following curve was
obtained from the experiment:
Fig 9: Engineering stress versus engineering strain curve (experimental data).
It is estimated that the necking point occurs between 6% and 8% (engineering strain). After
analyzing the experimental data, the first point satisfying the necking condition is at 6.68%.

Fig 10: Comparison between engineering and true curves (from experimental data).
Engineering formulation is converted into true formulation using the relationship:
The true stress curve is higher than the engineering stress curve, as it takes into account the
decrease in the objects cross-section.
Law 2: Elasto-plastic Material Law using the Johnson-Cook Model
Johnson-Cook Material Coefficients
The stress versus plastic strain law is: (Johnson-Cook model)

where, a is the yield stress and is read from the experimental curve and then converted into true
stress.
To compute b and n, two states are needed. This leads to the following formulas for b and n:
The first point is chosen at the necking point, then b and n are computed for each other point of
the curve and averaged out since the results tend to differ depending on the point chosen.
Characterization up to the Necking Point
The first stage when determining the material model is to obtain Johnson-Cook’s coefficients.
Neither the maximum stress, nor the failure plastic strain effects are taken into account here (set
at zero).
The values of coefficients are chosen so that the model adapts to the test data.

Fig 11: Variation of the engineering stress/strain according to Johnson-Cook’s model adapted to the test.

The material coefficients used for Law 2 are:
Initial density: 2.7x10-3 g/mm3 Yield stress: 90.27 MPa
Poisson’s ratio: 0.33 Hardening parameter: 223.14 MPa
Young’s modulus: 60400 MPa Hardening exponent: 0.375
Figure 12 compares the yield curve defined using the Johnson-Cook model with the one extracted
from experimental data.
Fig 12: Yield curves Johnson-Cook model 1
The true stress – true strain relationship can be described by:
The engineering stress deviations between experiment and simulation are described in the table
below:
Engineering strain 0.01 0.02 0.03 0.04 0.05 0.06 0.067
Deviation 7.9% 4.8% 1.8% 1.1% 1% 1.8% 2.9%
Comparison is performed up to the necking point (engineering strain = 6.68%) because after this
state, a rapid decrease in the engineering stresses occurs in the object. The rupture sequence is
simulated in the following paragraphs. Results using Law 2 remain within 8% of the experimental
curve.
The curve could be improved by slightly adjusting some of the values. The purpose of this test is
to propose a method for deducing material law parameters using a tensile test.

Beginning of the Necking Point
Necking Point Simulation

The Johnson-Cook model previously defined corresponds to the experimental results up to the
necking point. However, the slope of the numerical response does not enable the necking point to
start at the strain value observed experimentally.
The necking point is characterized by the slope value of the true stress versus the true strain
curve, which must be approximately equal to the true stress. The necking point numerically
appears by continuing simulation until the condition on the slope is observed.
The results are obtained using the Johnson-Cook model 1:
Fig 13: Beginning of the necking point using only the first coefficients of the Johnson-Cook model (a, b and n).

Fig 14: True stress versus true strain curve up to the beginning of the necking point.
The necking point can be simulated, either by adjusting the Johnson-Cook coefficients to obtain
an accurate slope, or by compelling curve with a maximum stress.
Simulation of the Slope near the Necking Point
By implementing an energy approach, the hardening curve can be modified to achieve an
engineering curve which resembles a horizontal asymptote near the necking point with the
purpose of simulating the behavior of the curve as observed in the test.
The Johnson-Cook coefficients used to describe the physical slope are:
Yield stress: 79 MPa
Hardening parameter: 133 MPa
Hardening exponent: 0.17
For this model, the new true stress/true strain relationship is:
(Johnson-Cook model 2)
The results obtained with those coefficients are provided below.

Figure 15 compares the Johnson-Cook model 3 with the experiment:
Fig 15: Adjusted engineering stress/strain curve to model the beginning of the necking point.
The shape of the yield curve versus the experimental data is depicted in Fig 16.

Fig 16: Yield curves.
The necking point is defined as .

This condition is characterized by the intersection of the true stress versus the true strain curve
with its derivate.

Fig 17: Superposition of engineering curve and true curve with its derivate.
Beginning of the Necking Point Using a Maximum Stress Limit, max
For this test, the Johnson-Cook coefficients input are those set in characterization up to the
necking point, the failure effect not being taken into account (the failure plastic strain is set to
zero). The beginning of the necking point is set using the choice of a maximum stress value. In
comparison to the experimental results (see Fig 10), the necking point is well defined for a
maximum stress set at 175 MPa. The limit in stress appears on the von Mises stress versus true
strain curve on elements where the necking point occurs.
The maximum true stress manages the beginning of the necking, as shown below:

Fig 18: Engineering stress versus engineering strain; necking point characterization

Fig 19: Variation of the von Mises stress with the true strain from shell 11.
Maximum stress max is reached for von Mises stress on shells where the necking begins. To
avoid overly-high stresses after the necking point, a maximum stress factor must be set
approximately equal to the true necking point stress.
The following curves show the evolution of the von Mises stress versus the true strain shell at
two characteristic locations of the object (3b and 3a in Fig 20):

Fig 20: von Mises stress curve with a maximum stress limit.
The beginning of the necking point is observed following the point where the stress is equal to
stress versus strain derivate .
Fig 21: Yield curve with maximum stress.
The yield curve is described by:

The derivate of the stress is very sensitive and strongly depends on the yield curve definition.
Thus, introducing the necking point into the simulation is very delicate (a small change can result
in many variations). The necking point should first begin on a given element for numerical
reasons. The preferred beginning of necking is addressed below.
Preferred Beginning of the Necking Point
Experimentally, the beginning of the necking point can appear anywhere on the object. The
beginning of the necking point should preferably be located on the right end elements in order to
propose a methodology for this quasi-static test. If the model only uses a quarter part of the
object, the necking point is found on elements 30, 125 and 78.
The beginning of the necking point is physically and numerically sensitive and can be initiated on
the right elements by changing a few of the coordinates along the Y-axis of the node in the right
corner (node 16) in order to decrease the cross-section and privilege the necking point in this
zone. Changing the node position by 0.01 mm is enough for achieving the preferential beginning
of the necking point.
Fig 22: Node 16 to be moved.
A second approach also enables the necking point to be triggered on the right end side by
defining an extra part, including shells 3, 11 and 4 by using a maximum stress slightly lower than
the remaining part, in order to initiate the necking point locally since the necking point stress is
first reached in the elements having the lowest maximum stress value, that is shells 3, 11 and 4.
This method, based on material properties, is quite appropriate for demonstrating the
characterization of a material law and will thus be used in the continuation of the example.
Fig 23: Localization of the beginning of the necking point according to the models using max.

The material is described as Johnson-Cook model 1:
max = 174 / 175 MPa
The following curves indicate the variation of the engineering stress versus the engineering strain
according to the beginning of the necking point zone and in comparison to the experiment.
Fig 24: Engineering stress/strain curve for each starting necking point location.
There is a fast decrease in the engineering stress after the right-end necking point. The necking
point, due to the boundary conditions of the y-symmetry plane (y-translation DOF released),
becomes more pronounced.
The variations in the section where the necking point is found are quite similar up to the necking
point. After such point, there is a sharp surface decrease for the right-end necking point, contrary
to the second case where the surface decrease is more moderate.

Fig 25: Variation of cross section (necking point zone).
Improvement of the Elements’ Contribution during the Necking Point

Sequence
In order to simulate physically the contribution of each element in the necking point, it is
advisable to adjust the curve by varying the Johnson-Cook coefficients in order to increase the
intensity of stress at the necking point. The main result is no longer the variation of the
stress/strain curve but rather the surface under the curve which characterizes the energy
dissipated during the test. This energy-based approach is relevant for crash tests since the final
assessment is often more significant than how it was achieved.

Fig 26: Engineering stress/strain curve obtained using adjusted Johnson-Cook coefficients.
The following graph compares the new yield curve with experimental data:

Fig 27: Yield curves.

Material is described in the the Johnson-Cook coefficients are:
Johnson-Cook Model 3:
(true stress/strain)
Yield stress = 50 MPa
Hardening parameter = 350 MPa
Hardening exponent = 0.38
Maximum stress is set to 189 or 190 MPa (according to the parts)
The results of adjustment to the Johnson-Cook coefficients are depicted below:
Fig 28: Shell contribution during the necking point sequence (von Mises stress).
As the necking point progresses, more physical results are obtained due to the new input data of
the material law coefficients having a better element contribution.

Fig 29: Variation of the von Mises stress on elements 110, 109, 108, 107, 11 and 106.
Damage Modeling with Plastic Strain Failure

The elasto-plastic model of Johnson-Cook is used until failure, which is simulated using a plastic
strain failure option. The element is deleted if the plastic strain reaches a user-defined value
max. This damage model shows good stability. A maximum plastic strain is defined for each
Johnson-Cook model:
Fig 30: max = 75% ; yield curve close to experimental data:

Fig 31: max = 47% ; yield curve adjusted with respect to lower stresses: .
Fig 32: max = 40% ; yield curve adjusted with respect to high stresses: .
Failure is reached for relatively high true strains.

Law 27: Elasto-plastic Material Law with Model Damage
Law 27 is used to simulate material damage following a Johnson-Cook plasticity law. Thus, model
damage is associated with the previous law in order to take account of failure.
The damage parameters are:
 Tensile rupture strain t1: damage starts if the highest principal strain reaches this tension
value.
 Maximum strain m1: the element is damaged if the highest principal strain is above the
tension value. The element is not deleted.
 Maximum damage factors max: this value should be kept at its default value (0.999).
 Failure strain f1: the element is deleted if the highest principal strain reaches the tension
value.
Fig 33: Stress/strain curve for damage affected material.
The following graphs display the results obtained using the material coefficients of two previous
Johnson-Cook models. Damage parameters complete those models.

Damage Model A
Damage model: t1 = 0.16 ; m1 = 0.72 ; dmax = 0.999 ; f =1; max = 16
Johnson-Cook model:

Damage Model B
Damage model: t1 = 0.16 ; m1 = 0.45 ; dmax = 0.999 ; f =1; max = 16
Johnson-Cook model:

Law 36: Tabulated Elasto-plastic Law
This is a tabulated law; therefore, the true stress versus plastic strain function can be directly
used. The rupture phase can be simulated by adding points to this hardening function.
Fig 34: Hardening function defined in law 36 to obtain the results below.

Fig 35: Results obtained with tabulated law 36.
The hardening curve has to be defined with precision around the necking point while the decrease
of the curve is very sensitive to its adjustment. In order to improve the modeling of the necking
point, two points can be interpolated, one "just before" the necking point, and one "just after"
with the slope between those two points equal to the necking point stress.

11.2 - Strain Rate Effect
Title
Strain rate effect
Number
11.2
Brief Description
The strain rate effect is taken into account, using filtering (cut-off frequency).
Keywords
 Shell element
 Johnson-Cook elasto-plastic model (/MAT/LAW2)
 Engineering strain/stress, strain rate effect, and filtering
RADIOSS Options
Input File
Time_History_files:
<install_directory>/demos/hwsolvers/radioss/11_Tensile_Test/TENSILET01

Advanced
Strain Rate Effect and Strain Rate Filtering (Cut-off Frequency)
In this additional study, the Johnson-Cook model is used to study the strain rate influence on
stress with or without filtering. There is no comparison with the experiment data in this section.
The study of sensitivity will be performed up to the beginning of the necking point.
 Stress-strain relationship:
The Johnson-Cook plasticity model will take into account the strain rate effect on the elasto-
plastic material behavior in order to improve the quality of simulation.

The law reads as follows:
where:
is the strain rate
0 is the reference strain
p is the plastic strain (true strain)

c is the strain rate coefficient
The two optional inputs, strain rate coefficient and reference strain rate must be defined for the
material. The purpose of the sensitivity study is to illustrate the influence of material parameters.
For further explanations about the Johnson-Cook model, refer to "Elasto-plasticity of Isotropic
Materials" in the RADIOSS Theory Manual.
Strain Rate Filtering
Because of the numerical application of dynamic loadings, the strain rates cause high frequency
vibrations, which are not physical; thus the stress/strain curves look "noisy". To obtain smooth
results, the strain rate filtering option will allow the reduction of those oscillations by removing
the high frequency vibrations. A cut-off frequency for strain rate filtering (Fcut) is used since its
value has to be smaller than half of the sampling frequency (1/ t).
In this example, t = 0.2163x10-3 ms.
The constants a, b and n which define the shape of the stress/strain curve are:
a = 90.27 MPa
b = 223.14 MPa
n = 0.375
The results are reported in the following tables.

Strain Rate Effect - Plasticity Model: Johnson-Cook
The influence of the strain rate and stress smoothing are shown below (with = 5x10-3 ms-1 and
c = 0.1):
Stress Comparison

Influence of the Cut-off Frequency for Smoothing

The following results show the effect of the reference strain rate, and strain rate coefficient, c:
Influence of the Reference Strain Rate

(c =0.1 and Fcut =10 kHz)
Influence of the Strain Rate Coefficient, c

(with = 10-2 ms-1)
Results are smoothed with correct cut-off frequencies.

Figure 36 compares the distribution of the first principal strain rate in the object, with and
without strain rate filtering.
Fig 36: First principal strain rate comparison at time t=4 ms.
A more physical strain rate distribution is achieved by filtering. Moreover, such results show
spatial oscillations when not damped by filtering. The explicit scheme is an element-by-element
method and the local treatment of temporal oscillations puts spatial oscillations into the model.
 Strain rate coefficient c influence:
If c is set to zero, the strain rate effect is not taken into account. This coefficient affects the
yield stress and it slightly translates curves in the plastic region. It must be adjusted in
accordance with the reference strain rate.
 Reference strain rate influence:
If the strain rate is lower than the reference strain rate, there is no strain rate effect.
Therefore, the lower the reference strain rate, the more the effect will be emphasized. The
effect appears as a translation of the curve towards higher stresses. An increase in the flow
stress using an increasing reference strain rate is observed.
 Cut-off frequency influence:
The cut-off frequency must not be set higher than half of the sampling frequency.
Smoothing is improved as the cut-off frequency comes closer to a particular value and the
convergence of the curve until a smoothing curve can be observed. A high-reference strain
rate requires low cut-off frequencies.
Conclusion
A tensile test is simulated using several material laws in RADIOSS. A method is set up to
correspond to the material parameters in the Johnson-Cook model. The rupture phase is very
sensitive and the simulation results strongly depend upon the starting point for necking. The
point-by-point definition of the hardening curve in law 36 enables to bypass the adaptation
difficulties when using the Johnson-Cook model. However, the results following the necking point
are very sensitive to the position of points defining the hardening curve.
A method to filter the strain rate is also demonstrated. The method can be generalized to the
industrial cases.

Example 12 - Jumping Bicycle
Summary
The purpose of this example is to illustrate how to use the RADIOSS description when resolving a
demonstration example. The particularities of the example can be summarized using dynamic
loading during a four-step scenario where a dummy is first put on a bike, then it rides on a plane
to subsequently jump back down onto the ground. The scenario described is created using
sensors.

Title
Jumping bike
Number
12.1
Brief Description
After a quasi-static pre-loading using gravity, a dummy cyclist rides along a plane, then jumps down
onto a lower plane. Sensors are used to simulate the scenario in terms of time.
Keywords
 Shell, brick, beam, truss, general spring, and beam
 Sensors on rigid bodies and monitored volumes (perfect gas)
 Quasi-static load treatment (gravity), kinetic relaxation, restart file, and MODIF file
 Dummy and hierarchy organization
 Type 7 interface self-impacting and rigid wall (infinite plane and parallelogram)
 Linear elastic law (/MAT/LAW1) and Johnson-Cook law (/MAT/LAW2)
RADIOSS Options
 Added mass (/ADMAS)
 Initial velocity (/INVEL)
 Kinetic relaxation (/KEREL)
 Monitored volume type gas perfect (/MONVOL/GAS)
 Sensor (/SENSOR)
Input File
Jumping_bicycle: <install_directory>/demos/hwsolvers/radioss/12_Bicycle/Bike/BIKERC*

Advanced

Overview
Aim of the Problem

The purpose of this example is to set up a demonstration in which sensors and restart files are
used to allow the change of a problem over time.
Subjected to the gravity field, a dummy cyclist rides on a higher plane, then jumps down onto a
lower horizontal plane. The problem can be divided into four phases:
 positioning the cyclist under the gravity effect
 running the bicycle on the high plane
 free fly
 the impact on the ground
The following system is used: mm, s, ton, N, MPa
Fig 1: Problem scenario.

The bike is meshed with 12103 4-node shells, 68 3-node shells, 62 trusses, 12 beams and six
brick elements. The dummy consists of 4779 4-node shells, 207 3-node shell and 27 springs (8).
Fig 2: Meshes of the main parts of the model.

The material of the metallic parts use the Johnson-Cook law (/MAT/LAW2) with the following
properties:
 Density: 7.9x10-9 GKg/l
 Yield stress: 185.4 MPa
A QEPH formulation (Ishell = 24) is used for tires in order to prevent hourglass deformations. A
Belytschko & Tsay element with a type 4 hourglass formulation is used for the other shell parts. A
global plasticity model is used.
Materials and proprieties are provided in the table below:
Table 1: Proprieties and materials of main parts
Parts Properties Materials
Frame Shell Q4 – 3 mm Steel – Law 2
Spokes Truss – 2 mm2 Steel – Law 2
Rim Shell Q4 – 3 mm Steel – Law 2
Tires Shell QEPH – 3 mm Rubber – Law 1

Bike
Hubs Beam – 900 mm2 Steel – Law 2
Saddle Brick Foam – Law 1
Pedals Beam – 900 mm2 Steel – Law 2
Tube of saddle Shell Q4 – 3 mm Steel – Law 2
Body (limbs) Shell Q4 – 3 mm Law 1

Dummy
Joints Spring (8) -
Hierarchy organization:
 Bike model: 6 subsets comprising 23 parts
 Dummy model: 11 subsets comprising 38 parts

Monitored Volumes / Perfect Gas
A perfect gas monitored volume is defined to model the pressure in the tires. For further details
about monitored volumes, refer to the RADIOSS Theory Manual.
The main properties are:
 Initial internal pressure: 0.75 MPa
 Gas constant: 1.4
All other properties are set to default values.
Fig 3: Visualization of a monitored volume (yellow part).
 Quasi-static loading: gravity effect on initial static equilibrium

The quasi-static solution of gravity loading on structure deformation corresponds to the
steady state part of the transient response. It describes the pre-loading case before the
dynamic analysis. Therefore, the simulation is divided into two phases: quasi-static
response (structure subjected to the gravity) and dynamic behavior (run, jump and
landing). The solution is obtained from kinetic relaxation (see /KEREL). Gravity is defined
by /GRAV.
 Contacts modeling
The type 7 interface using the penalty method serves to model contacts between the
dummy cyclist and the bike. A self-impacting interface (symmetrical) is required to treat
the landing of the bike. It is modeled by a type 7 interface having default values. Figure 4
below illustrates the description of the interface.

Fig 4: Contacts modeling with type 7 interface (Penalty method).
A type 11 interface models contact between the pedals (beams) and the feet (shells).
 Links between man and bicycle
The spring type 8 (/PROP/SPR_GENE) general spring property model the links between the
feet/pedals and the hands/handlebar.
 Stiffness (TX, TY and TZ): 100 kN/m
 Mass: 1 g
 Inertia: 0.1 kg/mm2
A rupture criteria based on displacements is activated by the beams connecting the hands and
handlebar in order to simulate the fall of the cyclist after landing.
 Left hand: Z = 20 mm
 Right hand: Z = 20 mm

Fig 5: Link right hand/handlebar (Type 8 springs)
 Dummy joints
Fig 6: Type 8 Springs
The general type 8 springs, characterize a spherical hinge with a stiffness given for each DOF.
Directions are local and attached to a moving skew frame. Two coinciding nodes define a spring.
Limbs are linked to the springs via the slave nodes of the rigid bodies, as shown in Fig 7.

Fig 7: Example of connection rigid body – spring 8 – rigid body.
 Wheel rotation
Beam elements are used to attach the wheel to the forks. The rotational DOF is released around
the beam axis.
Fig 8: Wheel/forks junction

Two types of rigid walls are set up:
 A fixed infinite plane (floor)
 A fixed parallelogram (springboard)

Fig 9: Position of the rigid walls
The characteristics of the parallelogram plane are: 2013 mm x 1200 mm. Both rigid walls are tied
to allow the wheels to turn.
The infinite plane is defined by the normal vector ( ) and the parallelogram by the
coordinates of three corners (M, M1, and M2). For both rigid walls, the slave nodes are obtained
from the tire and rim parts (displayed in green in Fig 10).
Fig 10: Slave nodes definition (green) and profile view of rigid walls
Several rigid bodies are created (/RBODY) and activated by sensors for use at the appropriate
time and in a chronological manner (sens_ID not equal to 0). Thus, every rigid body is not active
at the same time. The activation order is described in the paragraph dedicated to /SENSOR.
According to their activation time, the rigid bodies are classified in groups which are indicated in
following table.

Fig 11: Classification of rigid bodies (group).
The inertias of rigid bodies are set in local skew frames for groups A, C and D.
Rigid body activation – deactivation:
Groups A and B: The rigid bodies are activated during pre-loading up to equilibrium then
applied to the initial velocity start. They are activated again just before the impact of the bike on
the inferior plane.
During the free fly phase, both the cyclist and the bike undergo a rigid body motion. In order to
save the computation time, the motion can be simulated by putting the whole structure into a
global rigid body (Group D). The rigid body is deactivated just before landing.

Group C: Three rigid bodies include the dummy, the frame and both wheels (not including the
tires). This configuration allows just the wheels to turn, taking into account the active tires action
on the plane. This rigid body is activated while the bike is running on the springboard.
Group D: This global rigid body, including all nodes of model is activated as long as the bike is in
the free fly phase and is deactivated just before impact on the floor.
Group E: This rigid body is activated before impact ensures the stiffness level of the lower fork.
A 8333 mms-1 (30 km/h) initial velocity (/INIVEL) is applied to all nodes of the model (bicycle
and cyclist) in a parallel direction to the high plane at time t = 0.004 s. This initial condition is
defined in the Engine file “*_0002.rad" (start time: 0.004 s) which is run after the quasi-static
equilibrium with gravity loading.
Options in Engine file (*_0002.rad):
/INIV/TRA/X/1 initial translational velocities in direction x
8333 of 8333 mm/s
1 338000 on node 1 to 338000
Fig 12: Initial translational velocities of the model bike – man (30 km/h) at t = 0.004 s.
Gravity is applied to all nodes of the model. A constant function defines the gravity acceleration
in the Z direction versus time. Gravity is activated by /GRAV.

Fig 13: Gravity function (-9810 mm.s-2 ).
The explicit time integration scheme starts with the nodal acceleration computation. It is efficient
for the simulation of dynamic loadings. Nevertheless, quasi-static simulations via a dynamic
resolution method need to minimize the dynamic effects to converge towards the static
equilibrium. Among the methods usually employed, the kinetic relaxation method is quite
effective and is activated in the Engine file (*_0001.rad) with /KEREL. All velocities are set to
zero each time the kinetic energy reaches a maximum value.
Fig 14: Kinetic relaxation method with /KEREL.
Rigid bodies are activated and deactivated with sensors (/SENSOR). A sens_ID flag characterizes
the sensors and it is required in the rigid bodies’ definition. The five types of sensors used are:
 TIME (activated with time)
 DIST (activated with nodal distance)
 INTER (activated after impact on rigid wall)
 SENSOR (activated with sensor IS1 and deactivated with sensor IS2)
 NOT (ON as long as sensor IS1 is OFF)

Fig 15: Events definition for the activations and deactivations of sensors.
At the beginning of the simulation (time=0), the rigid bodies are automatically set to ON, as long
as the sensors are not active. Thus, in order to deactivate the rigid bodies at the first cycle,
active sensors at time t=0 should be used. Consequently, the rigid bodies are active when the
sensors are not active.
Added masses and inertia, as well as the flag for the gravity center, are ignored when a rigid
body is managed by sensors. By default, the gravity center is only computed by taking into
account the slave nodes mass (ICoG set at 2). The master node is moved to the computed center
of gravity where added mass and inertia are placed. In order to distribute the mass to the
dummy over the rigid bodies, option /ADMAS is used.
Sensors used are:
Table 2: Sensors used for simulation

Rigid body’s
Name Type Definition group using
senor
S1 TIME Time 0s. -
Distance between rear hubs and extremity

S2 DIST -
of springboard equal to 1810 mm.
Distance between rear hubs and extremity

S3 DIST -
of springboard equal to 345 mm.
S4 RWALL When the infinite rigid wall is impacted. -
SEN(S2,S3) SEN Activated with S2 and deactivated with S3 -
SEN(S3,S4) SEN Activated with S3 and deactivated with S4 -
SEN(S2,S4) SEN Activated with S2 and deactivated with S4 Group A/B
Deactivated with S2 and

NOT(SEN(S2,S3)) NOT Group C
activated with S3
Deactivated with S3 and

NOT(SEN(S3,S4)) NOT Group D
activated with S4

Sensor (S4) is also used for deactivating both the beam type springs modeling links between the
feet and pedals (Isflag set to 1). A case could be considered without this sensor to study the risks
of automatic pedals.
The following graphs show the active and deactivated zones of sensors and rigid bodies.
Fig 16: Activation and deactivation of sensors and rigid bodies.

The elements included in a rigid body are deactivated. Therefore, the element flags saved in
/TH/RBODY provide information on the activation and deactivation of rigid bodies during
simulation.
Fig 17: Activation and deactivation of main model parts (elements flag ON/OFF).

Fig 18: Distribution of the von Mises stress on the frame after quasi-static loading.

Fig 19: Kinetic relaxation effect on kinetic energy with /KEREL.

Fig 20: Simulation phases (impact at t = 4.6 s).

Fig 21: Configuration of a dummy cyclist during impact on the ground (shoes not attached).

Fig 22: Variation of von Mises Stress for a shell element of the frame.

Example 13 - Shock Tube
Summary
This famous experiment is interesting for observing the shock-wave propagation. Moreover, this
case uses the representation of perfect gas and compares the different formulations: The ALE
uses Lagrangian or Eulerian and Smooth Particle Hydrodynamics (SPH).
The first part of the study deals with the modeling description of perfect gas with the
hydrodynamic viscous fluid law 6. The purpose is to test the different formulations:
 Lagrangian (mesh points coincident to material points)
 Eulerian (mesh points fixed)
For the Eulerian formulation, different scale factors on time step are also tested. Furthermore,
the SPH formulation is also tested; which does not use mesh, but rather particles distributed
uniformly over the volume.
The propagation of the gas in the tube can be studied in an analytical manner. The gas is
separated into different parts characterizing the expansion wave, the shock front and the contact
surface. The simulation results are compared with the analytical solution for velocity, density and
pressure.

Title
Shock tube
Number
13.1
Brief Description
The transitory response of a perfect gas in a long tube separated into two parts using a
diaphragm is studied. The problem is well-known as the Riemann problem. The numerical results
based on the SPH method and the finite element method with the Lagrangian and Eulerian
formulations, are compared to the analytical solution.
Keywords
 Lagrangian and Eulerian formulations
 SPH modeling and hexagonal net
 Scale factor for time step
 Hydrodynamic viscous fluid law (/MAT/LAW6) and perfect gas modeling
RADIOSS Options
 SPH symmetry conditions (/SPHBCS)

Input File
Eulerian_formulation:
<install_directory>/demos/hwsolvers/radioss/13_Shock_tube/Eulerian_formulation/
TACEUL*
Lagrangian_formulation:
<install_directory>/demos/hwsolvers/radioss/13_Shock_tube/Lagrangian_formulation/
TACLAG*
SPH_hexagonal-net:
<install_directory>/demos/hwsolvers/radioss/13_Shock_tube/SPH_formulation/TUBSPH*

Advanced

Overview
Aim of the Problem

The shock tube problem is one of the standard problems in gas dynamics. It is a very interesting
test since the exact solution is known and can be compared with the simulation results. The
Smooth Particle Hydrodynamics (SPH) method, as well as the Finite Element method using the
Eulerian and Lagrangian formulations served in the numerical models.
A shock tube consists of a long tube filled with the same gas in two different physical states. The
tube is divided into two parts, separated by a diaphragm. The initial state is defined by the values
for density, pressure and velocity, as shown in Figures 1 and 2. All the viscous effects are
negligible along the tube sides; it is also assumed that there is no motion in the beginning.
Fig 1: Sketch of the shock tube.
Fig 2: Initial states with discontinuities.
The initial state at time t = 0 consists of two constant states 1 and 4 with p4 > p1, 4 > 1, and V1
= V4 = 0 (table).

Table 1: Initial conditions in the shock tube.
High pressure side (4) Low pressure side (1)
Pressure p 500000 Pa 20000 Pa
Velocity u 0 m/s 0 m/s
Density 5.7487 kg/mm3 0.22995 kg/mm3
Temperature T 303 K 303 K
Just after the membrane is removed, a compression shock runs into the low pressure region,
while a rarefaction (decompression) wave moves into the high pressure part of the tube.
Furthermore, a contact discontinuity usually occurs.
Perfect Gas Modeling with RADIOSS

The hydrodynamic viscous fluid law 6 is used to describe compressed gas.
The general equation describing pressure is:
with
where, p is the pressure, Ci are the hydrodynamic constants, En is the internal energy per initial
volume, and is the density.
Perfect gas is modeled by setting all coefficients C0, C1, C2 and C3 to zero.
Also:
C4 = C5 = -1
Where, is the gas constant.

Then the initial internal energy, per initial volume is calculated from initial pressure:
Under the assumption = CST = 1.4 (valid for low temperature range), the hydrodynamic
constants C4 and C5 are equal to 0.4.
Gas pressure is described by:

Parameters of material law 6 are provided in Table 2.
Table 2: Material properties of gas in law 6.
Initial volumetric energy

density (E0) 1.25x106 J/m3 5x104 J/m3
C4 and C5 0.4 0.4
Density 5.7487 kg/mm3 0.22995 kg/mm3
Analytical Approach
The shock tube problem has an analytical solution of time before the shock hits the extremity of
the tube [1].
Fig 3: Schematic shock tube problem with pressure distribution for pre- and post-diaphragm removal.
Evolution of the flow pattern is illustrated in Fig 3. When the diaphragm bursts, discontinuity
between the two initial states breaks into leftward and rightward moving waves, separated by a
contact surface.

Each wave pattern is composed of a contact discontinuity in the middle and a shock or a
rarefaction wave on the left and the right sides separating the uniform state solution. The shock
wave moves at a supersonic speed into the low pressure side. A one-dimensional problem is
considered.
Fig 4: Diagram of the shock, expansion waves and contact surface.
There are four distinct zones marked 1, 2, 3 and 4 in Fig 4. Zone 1 is the low pressure gas which
is not disturbed by the shock wave. Zone 2 (divided in 2 and 2' by the contact surface) contains
the gas immediately behind the shock traveling at a constant speed. The contact surface across
which the density and the temperature are discontinuous lies within this zone. The zone between
the head and the tail of the expansion fan is noted as Zone 3. In this zone, the flow properties
gradually change since the expansion process is isentropic. Zone 4 denotes the undisturbed high
pressure gas.
Equations in Zone 2 are obtained using the normal shock relations. Pressure and the velocity are
constant in Zones 2 and 2’.
The ratio of the specific heat constant of gas is fixed at 1.4. It is assumed that the value does
not change under the temperature effect, which is valid for the low temperature range.
The analytical solution to the Riemann problem is indicated at t=0.4 ms. A solution is given
according to the distinct zones and continuity must be checked. Evolution in Zones 2 and 3 is
dependent on the constant conditions of Zone 1 and 4. The analytical equations use pressure,
velocity, density, temperature, speed of sound through gas and a specific gas constant.
Equations in Zone 2 are obtained using normal shock relations and the gas velocity in Zone 2 is
constant throughout. The shock wave and the surface contact speeds make it possible to define
the position of the zone limits.

Zone 1 – Zone 4
Zone 4 Zone 1
Pressure p p4 = 500000 Pa p1 = 20000 Pa
Velocity u u4 = 0 m/s u1 = 0 m/s
Density 4 = 5.7487 kg/mm3 1 = 0.22995 kg/mm3
Temperature T T4 = 303 K T1 = 303 K
Speed of sound through gas:
Specific gas constant:
a a4 = 348.95 m/s a1 = 348.95 m/s
R 287.049 J/(kg.K)

Zone 2
Analytical solution Results at t = 0.4 ms
Pressure p p2 = 80941.1 Pa
Velocity u u2 = 399.628 m/s
Density 2 = 2RT2 2 = 0.5786 kg/mm3
Temperature T T2 = 487.308 K
Shock wave speed:
Therefore, x2/1 = Vs * 0.4 + 500 = 765.266 mm
Zone 2'
Pressure p p2 = p2' p2' = 80941.1 Pa
Velocity u u2 = u2' u2' = 399.628 m/s
Density 2' = 3(x4/3) 2' = 1.5657 kg/mm3
Temperature T p2' = r2'RT2' T2' = 180.096 K
Surface contact speed: Vc - u2
Therefore, x2/2' = 2 * 0.4 + 500 = 559.85 mm

Zone 3
Zone 3 is defined as:
where, x = 500 + X
Pressure p
Velocity u u3 = 290.792 + 2.0833 X
Density
Temperature T

Continuity verifications:
Finite Element Modeling with Lagrangian and Eulerian Formulations

Gas is modeled by 200 ALE bricks with solid property type 14 (general solid).
The model consists of regular mesh and elements, the size of which is 5 mm x 5 mm x 5 mm.
Fig 5: Mesh used for Lagrangian and Eulerian approaches.
In the Lagrangian formulation, the mesh points remain coincident with the material points and
the elements deform with the material. Since element accuracy and time step degrade with
element distortion, the quality of the results decreases in large deformations.

In the Eulerian formulation, the coordinates of the element nodes are fixed. The nodes remain
coincident with special points. Since elements are not changed by the deformation material, no
degradation in accuracy occurs in large deformations.
The Lagrangian approach provides more accurate results than the Eulerian approach, due to
taking into account the solved equations number.
For the ALE boundary conditions (/ALE/BCS), constraints are applied on:
The nodes on extremities have material velocities fixed in X and Z directions. The other nodes
have material and velocities fixed in X, Y and Z directions.
The ALE materials have to be declared Eulerian or Lagrangian with /ALE/MAT.
Smooth Particle Hydrodynamics Modeling (SPH)

The 12798 particles are distributed though a hexagonal compact net. No mesh is used.
Fig 6: Smooth Particle Hydrodynamics modeling with hexagonal compact net.

The nominal value h0 is the distance between each particle and its closest neighbor. According to
the assigned property of the part, the mass of the particles should be calculated. The mass is
related to the density and the size of the net, in accordance with the following equation:
Where:
 Particle mass of low pressure part: mp = 1.25265x10-5 g
 Particle mass of high pressure part: mp = 3.13166x10-4 g
Particle mass is specified in the SPH property set.

The scale factor of the time step is set to 0.3 in order to ensure cell stability computation.
Boundary conditions are used to introduce SPH symmetry conditions (/SPHBCS). This option is
specific to the SPH modeling and consists of creating ghost particles, symmetrical to the real
particles with respect to the symmetry plane.
Fig 7: SPH symmetry planes definition.
Each symmetry condition is defined according to the plane passing through the frame origin
attached to the plane and is normal in relation to the local direction of this frame.

Selected nodes and SPH symmetry condition frame along (-x) axis:
Six symmetry planes are used:

 x and (-x) symmetry conditions: SLIDE without rebound (Ilev =0)
 y and (-y) symmetry conditions: SLIDE without rebound (Ilev =0)
 z and (-z) symmetry conditions: TIED with elastic rebound (Ilev =1)
For the SLIDE-type condition, the material is perfectly sliding along the plane
The particles must lie on the symmetry planes at t = 0.
Fig 8: Local direction of frame

Particles should move into the positive semi-space defined as:
Where, O is the origin of the frame, P is a point of the plane, and is the local direction of the
frame.
Comparison of the Finite Element Results with the Analytical Solution

Simulation results along the tube axis at 0.4 ms are shown in the following diagrams.
Pressure

Density
Velocity
 Lagrangian formulation: Scale factor = 0.1

 Eulerian formulation: Scale factor = 0.5

Scale Factor Effect on Eulerian Results
Case 1: Scale factor = 0.5
Case 2: Scale factor = 0.9
Pressure
Density

Comparison of SPH Results and Analytical Solution
Simulation results along the tube axis at 0.4 ms. Scale factor: 0.3 and 0.67.
Pressure – Hexagonal Net and SPH Symmetry Conditions
Density – Hexagonal Net and SPH Symmetry Conditions

Velocity – Hexagonal Net and SPH Symmetry Conditions
Indications on computation for each formulation are given in the following table (the scale factor
is set to 0.5):
Finite Element approach SPH approach
Formulation Lagrangian Eulerian SPH
Normalized CPU 1.08 1 1809
Number of cycles
(normalized) up to 0.4 1.42 1 3.46
ms
(DTsca=0.5)

Pressure Distribution along Tube at 0.4 ms
Fig 9: Pressure wave produced in the shock-tube at t = 4 ms for different approaches and animations regarding pressure,
density and velocity

Reference
[1] J. D. Anderson Jr., Modern Compressible Flow with Historical Perspective, McGraw Hill
Professional Publishing, 2nd ed., Oct. 1989.

Example 14 - Truck with Flexible Body
Summary
This example compares different studies with flexible or rigid bodies. The method for using the
flexible bodies in an explicit analysis is also studied.
At first, the truck is modeled using a classical finite element model for explicit analysis. All parts
of the truck are modeled using different kinds of finite elements, such as shells, bricks, springs
and beams. The volumes monitored with perfect gas characterize the tires.
The problem is divided into two loading phases. First, gravity is applied as a quasi-static load.
Then, the truck’s Virtual Proving Ground (VPG) is studied to observe the truck driving over an
obstacle (bump).
For the gravity loading phase, the explicit approach using relaxation techniques or not is
employed. For the VPG analysis, three approaches are compared: (i) classical finite element
model; (ii) simplified finite element model with a global rigid body; and (iii) finite element model
involving a flexible body. The last approach requires the first run to compute the Eigen and static
modes. A flexible body input file is then generated for use in a second time-history run. The main
interest of this method is to economize the CPU time.

14.1 - VPG with a Complete Finite Element Model
Title
VPG with a complete
finite element model
Number
14.1
Brief Description
After applying gravity, a truck runs on a horizontal plane and passes over a bump.
Keywords
 Shell, brick, beam, beam type spring, and monitored volume (perfect gas)
 Quasi-static load treatment and kinetic relaxation
 Type 7 and 2 interfaces, self-impacting, and rigid wall (infinite plane and cylinder)
 Linear elastic law (/MAT/LAW1), elasto-plastic law (/MAT/LAW2), and void material law
(/MAT/LAW0)
RADIOSS Options
 Initial velocity (/INIVEL)
 Monitored volume type gas perfect (/MONVOL/GAS)
 Skew frame (/SKEW)
Input File
VPG_complete_model:
<install_directory>/demos/hwsolvers/radioss/14_Truck_with_FXB/VPG_complete_model/
TRUCK*

Advanced

Overview

The truck model is placed on the ground under the gravity field until static equilibrium is
obtained. Then, under the impulse of 15.6 m/s (56 km/h) initial speed, the truck runs in a
straight line and passes over a speed bump. The shock is expected to cause major deformation in
some highly solicited parts.
Units: mm, s, ton, N, MPa.
Fig 1: Problem studied.
In order to simplify modeling, most of the parts undergo the linear elastic material law
(/MAT/LAW1).
 Density: 7.85x10-9 Kg/l
The elasto-plastic Johnson-Cook model (/MAT/LAW2) mainly describes the joint and
strengthening elements, such as the beams and spring.
 Density: 7.85x10-9 Kg/l
The truck represents a simplified model having the essential parts. The weight of the truck is
approximately 8 tons.


Finite Element mesh:
The truck model is meshed with 21430 elements - 148356 degree of freedom, as follows:
 1D elements: 173
Details of the elements used are provided in Table 1 below:
Table 1: Composition of the EF mesh.
Number
Node 24726
4-node shell 18471
3-node shell 1638
Brick 1148
Beam 47
Spring 126
Part 159
The improved Belytschko hourglass formulation (type 4 hourglass, I shell =4) is used for shell
elements in the explicit computation. The Eigen analysis requires fully-integrated elements since
the computation mode needs an implicit option. Compatible element formulations are set by
default.
Fig 3: Overall mesh of truck.

The main parts of the model are shown in the table below:
Fig 4: Definition of the part.

Monitored Volumes / Perfect Gas
Monitored volumes are used to model the pressure in the tires. They are defined with one or
more shell property sets and the surface must be closed. The monitored volume used is the
perfect gas type.
The main properties for this type are:
 Initial internal pressure: 0.3 MPa
 Constant gas: 1.4
All other properties are set to the default values. The parts modeled with the monitored volumes
are highlighted in Fig 5:
Fig 5: Visualizing the monitored volumes (yellow parts).
Connections between Parts

In order to assemble the parts, four link types are used in the model:
 Beam type spring (type 13)
 Rigid body (kinematic condition)
 Tied interface type 2 (kinematic condition)
 Merged nodes
The beam type spring elements are useful for modeling the welding points. The modeling
techniques are described in the RADIOSS User's Guide.

In this example the beam type spring properties are:
 Inertia: 2x10-4 kg.m2
 Mass: 2x10-6 ton
Force and moment are read from the input curves:
Table 2: Input force versus displacement curve.
Displacement -1 0 1
Fx, Fy, Fz -105 0 105
Table 3: Input moment versus rotation curve.

Rotation -1 0 1
Mx, My, Mz -106 0 106
Fig 6: Beam type springs (13) used in the model.
The type 2 tied interface rigidly connects a set of slave nodes to a master surface. The kinematic
constraint is set on the slave nodes which remain in the same position on their master segments.
This interface is a kinematic condition. The Spotflag spotweld formulation is set to zero in order to
connect two meshes without coincident nodes. The master surface should be the coarser mesh.

Fig 7: Tied interface (type 2).
Fig 8: Tied interfaces (type 2) used in model
Fig 9: Example of the tied interface modeling connections between the fuel tank and its support

Rigid bodies are created to join two or more parts together. For these rigid bodies no added mass
is required and the master node can be located anywhere.
Slave nodes may not accept the other kinematic conditions (such as tied interface).
Fig 10: Visualization of rigid bodies in model.
A spherical inertia must be used for the rigid bodies having only two slave nodes for ensuring the
stability of the connected elements (set Ispher = 1). Thus, inertia is spherical and not computed
from data.
Contact Modeling – Self-impacting
Taking into account self-impacting parts, a type 7 self-impacting interface must be used. The
Block Format definition of this interface is to define master surface (/SURF/PART), then define
slave nodes as all nodes on this surface (/GRNOD/SURF).
Gap is equal to 0.5 mm.

Fig 11: Type 7 interface – self-impacting (slave side in red and master side in blue).
Wheel Rotation Modeling

Wheels are linked to a frame using an axle attached to the brake systems. A beam element (in
red in the opposite figure) models the axle causing released rotations at the node linked to the
wheel rim.

Fig 12: Wheel model (without brake disk)

Two types of rigid walls are set up:
 A fixed infinite plane (ground);
 A fixed infinite cylinder having a diameter, D = 1500 mm (bump).

Fig 13: Infinite plane and cylindrical wall for modeling the ground and bump (slave nodes displayed in green).
The cylindrical wall is defined by point M (500, 0, -600), M1 (500, 100, -600) and the diameter.
Fig 14: Cylindrical wall definition.
Both rigid walls are tied to allow the wheels to turn. The tire parts define the slave nodes for the
infinite plane (contact of ground and tires) and only the nodes of the front right tire are set as
slave for the speed bump in order to model a local bump. The obstacle is not infinite.
A kinematic condition is applied on each impacted slave node. Therefore, a slave node cannot
have another kinematic condition; unless such condition is applied in an orthogonal direction. In
such a manner, incompatible kinematic conditions can be detected, due to the coincident normal
orientations along the Z-axis of the cylindrical and plane walls. However, the common slave
nodes are not affected simultaneously by both kinematic conditions.
Fig 15: Incompatible kinematic conditions (no orthogonal directions of normals).

A 15600 mm.s-1 (56 km/h) initial velocity (/INIVEL) is applied to all nodes of the structure in the
X direction at t = 0.3 s. This initial condition is defined in the D02 restart file (*_0002.rad, start
time: 0.3 s), which is run after achieving the quasi-static equilibrium with gravity loading.
Option in restart file (*_0002.rad):
/INIV/TRA/X/1 initial translational velocities in the X direction
15600 of 15600 mm/s
1 265130 on node 1 to 265130 (/INIV/TRA/X/1)
Fig 16: Selected nodes for the initial translational velocity of the truck (56 km/h) at t = 0.3 s.
Quasi-static Loading: Gravity Effect on Initial Static Equilibrium

The quasi-static solution of gravity loading on structure deformation is the steady state part of
the dynamic response and describes the pre-loading case before the transient analysis. Thus,
simulation is divided into two phases: quasi-static response (structure deformation under the
gravity effect) and dynamic behavior (run and impact on the bump). The solution is obtained
using the kinetic relaxation method.
Gravity is applied to all nodes of the model. A constant function defines the gravity acceleration
in the Z direction versus time and is equal to -9810 mms-2. Gravity is activated with the /GRAV
option.
The explicit time integration scheme assumes starting with nodal acceleration computation. It is
very efficient for simulating dynamic loadings. Nevertheless, quasi-static simulations via a
dynamic resolution method need to minimize the dynamic effects in order to converge towards
static equilibrium. Among the usual methods employed, the kinetic relaxation method is quite
effective and is activated in the Engine file (*_0001.rad) using /KEREL. All velocities are set to
zero each time the kinetic energy reaches a maximum value.

Fig 17: Kinetic relaxation method using the /KEREL option.
Animation of the passing over the speed bump:
Fig 18: Distribution of von Mises stress on the model during bump passage.

Fig 19: Cab deformation (initial state and after bump passage).

14.2 - VPG with Flexible and Rigid Bodies
Title
VPG with flexible and
rigid bodies
Number
14.2
Brief Description
After applying gravity, a truck runs on a horizontal plane and passes over a bump. The major
part of the truck is described using a flexible body.
Keywords
 Eigen and static analysis
 Eigen modes
 Flexible body
RADIOSS Options
 Eigen modes computation (/EIG)
 Flexible body input file (/FXINP)
 Flexible body (/FXBODY)
Input File
VPG_Rigid_body:
<install_directory>/demos/hwsolvers/radioss/14_Truck_with_FXB/VPG_Rigid_body/TRUCK
*
VPG_Flexible_body:
<install_directory>/demos/hwsolvers/radioss/14_Truck_with_FXB/VPG_Flexible_body/
Model_EIG/TRUCK_EIG_*
<install_directory>/demos/hwsolvers/radioss/14_Truck_with_FXB/VPG_Flexible_body/
Model_FXB/TRUCK_FXB_*

Advanced

Overview
Aim of the Problem

The purpose of this example is to perform an Eigen analysis on a complete truck model with the
purpose of creating a flexible body which will be used to model the truck’s main part, excluding
transmission (wheels, left-springs, differential, shaft, brakes and axles). In order to appreciate
the quality of the modeling, the results will be compared with those obtained using two other
models: one without a flexible body (previous analysis) and the other substituting the flexible
body with a rigid body.
The study deals with:
 an Eigen analysis to create a file containing the dynamic response of the structure
 a quasi-static analysis (explicit pre-loading by gravity)
 an explicit dynamic analysis with a global flexible body
 an explicit dynamic analysis with a global rigid body
The original model and two alternative models are compared:
1 model including a 1 model including a

1 complete model
global flexible body global rigid body
In the previous section where a complete finite element model is used, it is noted that the stress
and strain levels are low for most parts of the global model. Thus, the CPU time can be
considerably reduced if the elements working in the linear elastic field are replaced with a flexible
body. The purpose of this example is to provide an overall view of using flexible bodies in
RADIOSS.
The top part of the truck, where no damage and no plastic strain occurs, is first successively
modeled with a rigid body (non-deformable) and then with a flexible body (deformable), as
shown in Fig 20.
Parts of the truck covered by rigid or flexible body is shown in the following diagram:

Fig 20: Top part of truck included in a flexible/rigid body depending on the model.

Eigen and Static Modes Computation – Flexible Body Creation
A flexible body is similar to a rigid body where displacement is computed on nodes corresponding
to vibration modes. The input file for a flexible body uses the RADIOSS Eigen modes and static
modes computation. Modes can derive from experimental analysis, as well as from vibratory
software.
The total displacement field for every point of a flexible body is obtained by displacing the local
frame defining the rigid body modes and from an additional local displacement field
corresponding to the body’s small vibrations.
Fig 21: A flexible body is deformable according to its Eigen modes (from vibratory analysis).
A preliminary study with RADIOSS extracts Eigen or static modes for creating the flexible body
input file used in a second run. This computation phase requires the /EIG and /FXINP options.

The /EIG option is set up in the Starter input file and defines the part to be included in the
flexible body, as well as the type and number of modes to be computed.
In this example, the main data is:

 Number of modes = 25
 Maximum Eigen frequency = no
 Minimum Eigen frequency = 13 Hz
 Number of Eigen modes per block = 10
Two types of modes can be obtained:
 Free Eigen modes
 Static modes
Eigen modes (or dynamic modes) are computed for the entire structure without any specific
boundary condition. The equation solved is:
Ku = Mu
In this approach, rigid body modes in the structure are possible and give null Eigen frequencies.
If Kur = 0, K is not singular and ur 0, therefore, Mur = 0 and =0
In addition, static modes can be computed if boundary conditions are added to a node group in
the flexible body frontier. They correspond to the static response of the structure. All degrees of
freedom in the set of interface nodes concerned by the additional boundary conditions are fixed
and one static mode is computed for each constrained degree of freedom. The equation solved is:
Ku = F
Static modes are displayed with null frequencies in animations.
Rigid modes are not permitted and generate null pivots during inversion of the stiffness matrix.
It should be noted that modes computation requires the implicit options in the Engine file
(/IMPL/LINEAR and /IMPL/SOLVER/1).
 Eigen frequencies are provided in the Engine output file. One animation exists per
computed mode.

 The /FXINP option is used in the Engine file for creating a flexible body input file .fxb. The
flexible body has the same support as that defined in /EIG. You should enter:
- Identification number of the Eigen mode or static mode problem defined in /EIG;
- The critical structural damping coefficient used for computing the Rayleigh damping
coefficient to be introduced in the flexible body (it is recommended to use default value
0.03);
- Type of flexible body (1 = free flexible body, 2 = fixed flexible body).
 The flexible body input file can be used in a second run using /FXBODY in the Starter Input
file to generate a flexible body. The flexible body input file name ending in .fxb for the
RADIOSS format and master node coordinates are required (possible coordinates are given
at the top of the .fxb file).
Eigen Analysis
Run using Flexible Body
(writing FXB input file)
 Master nodes of rigid

 Rigid bodies + master bodies.
FXB domain
nodes.  Master node of the
can contain
 Boundary conditions. flexible body.
 Interfaces.
 Free parts.
 Slave nodes on the  Rigid bodies (slave
flexible body frontier. nodes).
FXB domain  Rigid body overlapping on  Slave nodes on the
must not flexible body and the rest flexible body frontier.
contain of structure.  Rigid body overlapping on
 Truss elements. flexible body and the rest
 Void material. of structure.
 Monitored volumes.
Options incompatible with the implicit solver must be avoided.
Fig 22: Flexible body creation from RADIOSS options.

The inputs files used with the specific options are:
For the truck model, the global flexible body includes 14344 nodes, 120 of which are the master
nodes of the inside rigid bodies. Thus, the flexible body takes into account constraints of the rigid
bodies.
Eigen Run
In addition, you can define nine interface nodes linking the flexible body and the rest of the truck
with the translation fixed along the X-, Y- and Z-axis. Thus, 27 static modes will be computed.
Only the translation degrees are retained in order to minimize the input file size of the flexible
body, given that preliminary studies have shown that additional static modes computed by fixing
rotational degrees have not substantially improved flexible body behavior.

Fig 23: Nine interface nodes with blocked translations for computing static modes.
A static mode is computed for each fixed degree of freedom, in addition to the Eigen modes.
Thus, the number of modes is equal to the number of Eigen modes, plus the number of blocked
degrees of freedom.
Flexible Body Run
The rigid bodies and tied interfaces included in the flexible body domain should be removed for
the second run. Those kinematic conditions are only considered in Eigen modes computation.
The coordinates of the center of mass (possible master node) indicated in the flexible body input
file are:
X: 3.267252E+03 Y: -1.71759E+01 Z: 1.407584E+03 (node 265200)
The master node should be included in the nodes groups for gravity loading and initial velocity. It
should be defined in the Starter file (/NODE).
Connections between the parts covered by the flexible body and other parts of the model are
modeled with beams and the rigid body, as shown in Fig 24. Connection is set at the beam
extremity.
Fig 24: Example of the connection point for flexible body.

Fig 25: Characteristic Eigen modes (arbitrary displacement).

Comparison of Animation Results
Deformed configurations are compared with global bodies according to the modeling used:
Fig 26: Face view of the different models’ behavior during bump passage (displayed with multi-models option)
Animation Results
 Animations multi models: cab deformation face view
 Animation flexible body model: cab deformation
 Animation original model: cab deformation
Conclusion
This example introduced a method for creating and employing a flexible body using an Eigen
analysis performed by RADIOSS. The number of retained modes and the frequency range set for
the Eigen analysis are according to the parameters which influenced the results.
Simulation using the flexible body provided accurate distribution of deformations in the model,
compared with the modeling not having a substitute body. However, the amplitudes obtained are
very low. The flexible body behavior could be enhanced by improving connections between the
flexible body and the rest of the structure to ensure transmission of the shock wave up to the
flexible body.
The flexible body input file required the IMPLICIT module for the Eigen modes computation.

Example 15 - Gears
Summary
The main purpose of this example is to study how to represent a quadratic contact. Two different
interfaces are compared:
 Type 16 interface (node to surface)
 Type 17 interface (surface to surface)
Moreover, 20-node brick and 16-node shell elements are used for the mesh to represent the
curvature of the teeth. Constant acceleration is also applied to the gear using constant initial
rotation and an imposed velocity.
Finally the results of the contact force show that the type 16 interface is more adaptable than a
type 17 interface.

Title
Simple gears
Number
15.1
Brief Description
The problem studied is a twin gear having an identical pitch diameter and straight teeth.
Keywords
 Type 16 and 17 interfaces (sliding)
 20-node brick and 16-node thick shell
 Quadratic surface contact
RADIOSS Options
 Interface type 16 (/INTER/LAGMUL/TYPE16) and type 17 (/INTER/LAGMUL/TYPE17)
Input File
Interface_type16:
<install_directory>/demos/hwsolvers/radioss/15_Gears/Inter16/DIF24416*
Interface_type17:
<install_directory>/demos/hwsolvers/radioss/15_Gears/Inter17/DIF24416*

Advanced

Overview
Aim of the Problem

The purpose of this example is to illustrate the use of quadratic contact via node-surface and
surface-to-surface interfaces of types 16 and 17.
The following study shows a gear example using interface type 16 and type 17. The finite
elements used to model this gear are the thick SHELL16 elements and the quadratic BRICK20
elements. In the first stage the interface type 16 is used to model contact between the teeth.
Then, interface type 17 is used to manage a surface-to-surface contact.
The gear system is turning with a constant acceleration ( = 0.002 rad/ms2). The acceleration is
applied to both of the gears. It is assumed that contact between the teeth does not generate any
friction.
Steel characteristic (elastic /MAT/LAW2) are:
 Young Modulus: 210000 MPa
 Poisson’s coefficient: 0.29
 Number of teeth: Z =19
 Diametric pitch: P = 1/mo = 1/40
 Pressure angle: ao = 20 degrees
The following parameters are calculated as:

Pitch diameter: Dp = mo * Z, then Dp = 760 mm
Root diameter: Db = mo * cos(ao), then Db = 714.17 mm
Addendum: ha = mo, then ha = 40 mm
Dedendum: hf = 1.25 * mo, then hf = 50 mm
Circular pitch: p = PI * mo, then p = 125.66 mm
Fig 1: Global view of the mesh.

The finite elements used to model this gear are the thick SHELL16 type elements and the
quadratic BRICK20 elements.
A quadratic solid mesh is used to take into account the teeth’s curvature. The external BRICK20
elements are then converted to solid SHELL16 shells using pre-processing. The interface types 16
and 17, manage contact between the quadratic surfaces of the SHELL16 elements.
Fig 2: View of the teeth.
Interface types 16 and 17 use the Lagrange multipliers. Type 16 interface is built with a slave
node group impacting a quadratic master surface. Type 17 interface is built with two quadratic
surfaces.

Fig 3: Contact modeling between quadratic surfaces (type 16/17 interfaces).

To apply constant acceleration, start at time, t, with a constant initial rotation phase to reduce
the initial value effects. Then, an imposed velocity is applied to the gears to manage a rotational
acceleration of 0.002 rad/ms2.
Fig 4: Initial velocities on the temporary rigid bodies’ master nodes.
To apply the initial rotational velocity to the gears, two rigid bodies are created, as shown in Fig
4. Then both the rigid bodies are set to OFF to ensure a realistic deformation of parts after the
first loading phase.

The out-of-plane rotation of the rigid bodies is set free. A YZ symmetry plan is used to stabilize
the model.
Fig 5: Free out-of-plane rotation of rigid bodies.
Number of options:
Options Quantity
BCS 3
BRIC20 950
FUNCT 1
GRBRIC 2
GRNOD 20
IMPVEL 1
INIVEL 2
INTER 1
MAT 1
NODE 10757
PART 3
PROP 3
RBODY 4
SENSOR 1
SHEL16 380
TH 4
Minimum time step: 0.4E-03 ms

Time History Results

Figure 6 shows the application of velocities on the gears.
Fig 6: Imposed rotational velocity curves.

Fig 7: Contact force obtained with type 16 interface.

Fig 8: Contact force obtained with type 17 interface.
Figures 7 and 8 compare the contact forces obtained for two different models; one using the
interface type 16 and the other using the type 17 interface. The comparison shows that some
numerical problems may appear when using the interface type 17, due to the complexity of the
algorithms; especially when two surfaces with nonlinear curvatures are used.
On the other hand, interface type 16 obtains an overall physical response.
Animation with Multi-model Mode
Conclusion
The type 16 interface provides overall satisfactory results for this kind of application, where the
contact surfaces are complex and there is no gap.

Example 16 - Dummy Positioning
Summary
The problem of a dummy positioning on the seat before a crash analysis is the quasi-static
loading which can be resolved by either RADIOSS explicit or RADIOSS implicit solvers. If
deformation remains small, a linear analysis may be used as a simple approach to determine the
position after applying gravity force. However, this method is not valid if the contact surface
between the dummy and the seat is not correctly estimated before analysis. When comparing the
implicit and explicit solvers, it's shown that the implicit computation enables saving time in the
computation. However, the rigid body modes of the dummy must be controlled. This is not the
case if the explicit solver is used.

16.1 - EXPLICIT Solver
Title
EXPLICIT solver
Number
16.1
Brief Description
A dummy is sat down via gravity using the quasi-static load treatment.
Keywords
 Shell, brick, beam, and dummy
 Quasi-static analysis by explicit, kinetic and dynamic relaxation, and Rayleigh damping
 Type 7 interface (symmetric)
 Kelvin-Voigt visco-elastic model (/MAT/LAW35) and linear elastic law (/MAT/LAW1)
RADIOSS Options
 Added mass (/ADMAS)
 Material definition (/MAT)
 Rayleigh damping (/DAMP)
Input File
Rayleigh_damping:
<install_directory>/demos/hwsolvers/radioss/16_Dummy_Positioning/EXPLICIT_solver/
RAYLEIGH/.../SEAT_RAYLEIGH*
Dynamic_relaxation:
DYREL/SEAT_DYREL*
Kinetic_relaxation:
KEREL/SEAT_KEREL*

Without_damping:
Without_damping/SEAT*

Advanced
Overview
Aim of the Problem

The topic of this study concerns quasi-static load treatment using kinetic relaxation, dynamic
relaxation and Rayleigh damping. The explicit solutions provided by the three different
approaches will be compared and analyzed.
The purpose is to position a dummy on a foam seat under the gravity field using a quasi-static
approach prior to a possible dynamic crash simulation.
Units: mm, s, ton, N, MPa
The dummy weighs 80 kg (173.4 lbs.). The material introduced does not represent the physical
case; however, the global weight of the dummy is respected. As the dummy deformation is
neglected in this loading phase, simplifying the material characterizations has no incidence on the
simulation.

Material for seat brace - both the columns and the floor are made of steel with the following
properties (/MAT/LAW1):
 Density: 7.8 x 10-9 Gkg/l
The seat columns have the following characteristics:
 Area: 2580 mm2
 Inertia: IXX = 554975 mm4; IYY = 554975 mm4; IZZ = 937908 mm4
The thickness for the seat back and the floor:

 Brace thickness = 2 mm
 Floor thickness = 1 mm
The seat cushion is made of foam which can be described using the generalized Kelvin-Voigt
model. The material properties of the foam are:
 Young’s modulus: 0.2 MPa
 Poisson’s ratio: 0
 Density: 4.3 x 10-11 Gkg/l
 E1 and E2: 0 MPa
 Tangent modulus: 0.25 MPa

 Viscosity in pure shear: 10000 MPa/s
 C1 = C2 = C3 = 1 (visco-elastic bulk viscosity)
RADIOSS material law 35 is used. The open cell foam option is not active (I Flag = 0) and the
pressure is read using the following input curve:
Table 1: Pressure versus compression curve.
Compression -100000 -10 0 3000 209000 210000
Pressure -1000 -1000 0 7.633 7.633 18.5
Visco-elastic Foam Material Law (/MAT/LAW35):

Based on the Navier equation, law 35 describes materials using visco-elastic behavior. The effect
of the air enclosed is taken into account via a separate pressure versus compression function.
Relaxation and creep can be modeled.
The schematic model in Fig 2 describes the generalized Kelvin-Voigt model where a time-
dependent spring working in parallel with a Navier dashpot is put in a series with a nonlinear
rate-dependent spring.

Fig 2: Generalized Kelvin-Voigt model – Law 35.
Two pressure computations are available in RADIOSS for foam having no open cells. The
expression used by default is:
See the RADIOSS Theory Manual for explanation of coefficients.

Pressure may also be computed using the pressure versus compression curve defined by you.
The compression, u is defined as:
Where, is the density at a time t, and 0 is the initial density.
The model consists of two subsets:
 a dummy composed of 38 parts (limbs and joints).
 a seat comprised of six parts (foam seat back, foam seat cushion, seat back brace, seat
bottom brace, seat columns and the floor).

Fig 3: Model mesh. Fig 4: Model mesh.
(Perspective view – Shaded display) (Profile view – Edges display)
The seat cushion is meshed with 70 brick elements defined by general type 14 solid property.
 Quadratic bulk viscosity: 1.1
 Linear bulk viscosity: 0.05
 Hourglass viscosity coefficient: 0.1
The dummy and seat brace are modeled with shell elements, divided into 4871 4-node shells and
203 3-node shells (Dummy: 5004 shells and seat: 70 shells).
Using a dummy in the model, the /DEL/SHELL/1 option should be activated in the Engine file to
avoid a small time step, due to the low density of material defining the dummy envelope.
The shell properties are:
 Belytschko hourglass formulation (Hourglass type 4, Ishell = 4).
 Membrane hourglass coefficients: 0.01 (default value)

 Out-of-plane hourglass: 0.01 (default value)
 Rotation hourglass coefficient: 0.01 (default value)
Contacts between the dummy and the seat cushion, as well as between the foot and the floor,
use type 7 interface models with the penalty method. Additionally, symmetrical contact between
the body and seat is achieved by creating two complementary interfaces, as shown below:
First interface: Dummy parts: slave nodes
Seat: master surface
Second interface: Dummy parts: master surface
Seat: slave nodes

Fig 5: Contacts modeling with type 7 symmetrical interface
The gap between the symmetrical interfaces is equal to 5 mm, while a gap of 0.5 mm is set for
the other interface.
The type 7 interface allows sliding to occur between surfaces. A Coulomb friction can be
introduced; in addition, a critical viscous damping coefficient can be defined to damp sliding.
The symmetric interfaces properties are:
 Coulomb friction (Fric flag) = 0.3
 Critical damping coefficient (Visc flag) = 0.05
 Scale factor for stiffness (Stfac flag) = 1
 Sorting factor (Bumult flag) = 0.20
See the RADIOSS Theory Manual and Starter Input for further information about the definition of
the type 7 interface.
The goal is to set the body on the seat using a quasi-static approach in order to obtain static
equilibrium. The positioning phase is not included in this study. Thus, all nodes of the dummy are
placed in a global rigid body in order to maintain the dummy’s initial configuration.
In order to save the CPU, a second global rigid body includes parts of the seat and the floor;
except for the seat cushion parts, which will only have active elements during simulation.

Fig 6: Set up of both rigid bodies.
When the ICoG flag is set to 1 for the rigid body of the seat, the center of gravity is computed
using the master and slave node coordinates, and the master node is moved to the center of
gravity, where mass and inertia are placed.
When the ICoG flag is set to 3 for the rigid body of the dummy, the center of gravity is set at the
master node coordinates defined by you. The added masses and added inertia are transmitted to
the master node coordinates.
The master node coordinates and skew are extracted from the pelvis part of the original rigid
body.
Gravity is applied to all nodes of the model. A function defines gravity acceleration in the z
direction versus time. Gravity is activated by /GRAV in the Starter file (*_0000.rad).
Fig 7: Input gravity function (-9810 mm.s-2) and nodes selection (yellow).

The six rigid body modes of the seat are removed by completely fixing the rigid body master
node attached to the seat. In order to limit the out-of-plane vibrations, the master node of the
dummy's rigid body is fixed in translation along the Y axis.
Fig 8: Boundary conditions on the rigid bodies’ master nodes.
Static analysis: quasi-static treatment of gravity loading up to static equilibrium.

simulating dynamic loading. However, a quasi-static simulation via a dynamic resolution method
needs to minimize the dynamic effects in order to converge towards static equilibrium. This
usually describes the pre-loading case prior to dynamic analysis. Thus, the quasi-static solution
of gravity loading on the model is the steady state part of the transient response.
To reduce the dynamic effect, three options are available in the Engine file:
Kinetic Relaxation Method
All velocities are set to zero each time the kinetic energy reaches a maximum value. This option
is activated in the Engine file using /KEREL (input is not required).
Fig 9: Kinetic relaxation method with /KEREL (also named energy discrete relaxation).

Dynamic Relaxation Method
Dynamic loading is damped by introducing a diagonal damping matrix, proportional to mass
matrix, in the dynamic equation:
with, being the relaxation value (recommended default value 1), and T being the period to be
damped (less than or equal to the highest period of the system).
Thus, a viscous stress tensor is added to the stress tensor:
Using an explicit code, application of the dashpot force reduces the velocity equation
modification:
This option is activated in the Engine file (*_0001.rad) using /DYREL (inputs: and T ).
Rayleigh Damping Method
Dynamic loading is damped by introducing a damping matrix, proportional to the mass and
stiffness matrix, in the dynamic equation. This simplified approach will allow you to reduce the
global equilibrium equation to n-uncoupled equations by using an orthogonal transformation. This
damping is said to be proportionally uncoupled.
Where, a and are the pre-defined constants.

The orthogonal transformation using this proportional damping assumption leads to:
with, i being the ith being the damping ratio of the system, and i being the ith being the
natural frequency of the system.

Fig 10: Rayleigh type damping.
If you have some experimental results, the proportionality factors, a and are found by
evaluating the damping for a pair of the most significant frequencies used. Thus, two equations
with two unknown variables are obtained:
If several frequencies are available, an average of computed values, a and may be used.
This model of proportional damping is not recommended for complex structures and does not
enable good experimental retiming.
This option is activated in the Engine file (*_0001.rad) using /DAMP (inputs data: a and ).
Parameters Used
In this example, a and are set to the following values:
 First case: a = 10 and = 10
 Second case: a = 0 and = 10
 Third case: a = 10 and =0
 Fourth case: a = 20 and =0
The resulting assumptions are:
 First case: [C] = 10[M] + 10[K]
 Second case: [C] = 10[K]
 Third case: [C] = 10[M]
 Fourth case: [C] = 20[M]


Results Obtained using Kinetic Relaxation: /KEREL
Fig 11: Z-displacement of the rigid body’s master node on dummy (node 14199).

Fig 12: Kinetic energy of global model.

Results Obtained using Dynamic Relaxation: /DYREL
Fig 13: Z-displacement of the rigid body’s master node on dummy (node 14199).

Fig 14: Z-velocity of the rigid body’s master node on dummy (node 14199).
The period T to be damped is estimated from the velocity curves (highest period).

Results Obtained using Rayleigh Damping: /DAMP
Fig 15: Z-displacement of the rigid body’s master node on dummy (node 14199)

Fig 16: Z-velocity of rigid body’s master node on dummy (node 14199)

Comparison of the Different Approaches
Fig 17: Comparison of the nodal displacements’ display on the seat at time t = 1.48 s

Fig 18: Comparison of damping on displacement obtained using the three static approaches
(Z-displacement of the rigid body’s master node on dummy: node 14199)
Conclusion
It is undeniable that the damping methods used to converge towards static equilibrium provide
accurate results, especially in the case of this problem where the low rigidity of the seat caused
very little quenched oscillations.
The kinetic relaxation introduced in /KEREL, was relatively effective having a swift convergence of
the solution towards a static solution, in addition to being easy to use since no input is required.
Stability was obtained at 0.137 s.
The /DYREL and /DAMP options are based on viscous damping conducted for the same response,
with convergence in three oscillations. Stability was obtained at 0.75 s. Furthermore, dynamic
relaxation and the Rayleigh damping methods are basically equivalent in this problem, due to the
low stiffness of the seat cushion (Young’s modulus is equal to 0.2 MPa), which breaks the balance
between the mass and the weight stiffness in the Rayleigh assumption. Moreover, the boundary
conditions and the loading applied on the model lead to a problem described using a predominant
natural frequency. Thus, only one parameter, a is needed to describe this physical behavior,
which reverts back to the dynamic relaxation assumption.

Using =1 and T =0.18s for dynamic relaxation and =10 for Rayleigh damping, you achieve:
 Dynamic relaxation:
 Rayleigh damping:
[C] = [M] + [K]  [M]
 10[M]
In conclusion, the approaches available in RADIOSS provided after convergence a single solution,
namely displacement of the dummy by -12.66 mm along the Z-axis and an identical deformation
of the seat cushion.

16.2 - IMPLICIT Solver
Title
IMPLICIT solver
Number
16.2
Brief Description
A dummy is sat down via gravity using the implicit approach (static).
Keywords
 Shell, brick, beam, spring, and dummy
 Linear and nonlinear static solution by implicit solver
 Type 7 interface (symmetric) and tied interface (type 2)
 Kelvin-Voigt visco-elastic model (/MAT/LAW35) and linear elastic law (/MAT/LAW1)
RADIOSS Options
 Imposed displacement (/IMPDISP)
 Time step control method for implicit (/IMPL/DT)
 Initial time step for implicit (/IMPL/DTINI)
 Static linear implicit solution (/IMPL/LINEAR)
 Static nonlinear implicit solution (/IMPL/NONLIN)
 Print frequency for implicit (/IMPL/PRINT)
 Implicit solver method (/IMPL/SOLVER)
Input File
Linear_implicit_model:
<install_directory>/demos/hwsolvers/radioss/16_Dummy_Positioning/IMPLICIT_solver/
Linear/SEAT_IMPL_LIN*
Nonlinear_implicit_model:
<install_directory>/demos/hwsolvers/radioss/16_Dummy_Positioning/IMPLICIT_solver/
Nonlinear/
 Imposed_displacement: //.../Imposed_displacement/SEAT_IMPL_DISP*
 Concentrated_load: //.../Concentrated_load/SEAT_IMPL_CLOAD*
 Gravity_loading: //.../Gravity/SEAT_IMPL_GRAV*

Advanced
Linear and Nonlinear Analysis by Implicit Solver
The main advantages of implicit resolution are:

 Unconditional stable scheme
 Large time step
 Treatment of the static problem
However, the implicit algorithm uses a global resolution which requires convergence for each
time step and has low robustness in comparison to the explicit (null pivots, divergence for high
nonlinearities, etc.).
The implicit methods result in solving a linear system for each time step, which is relatively
expensive but enables a large time step: few expensive calculations. The explicit method treats
linear or nonlinear systems depending on the problem. It is less expensive and faster for each
step, but requires short time steps to ensure stability of the scheme that has many inexpensive
cycles.
 Implicit integration scheme: Newmark
This scheme is unconditionally stable, the stability condition being independent of the time step
choice. See the RADIOSS Theory Manual for further information about the Newmark scheme.
RADIOSS has a linear and a nonlinear solver. Only static computations are available and loading
should be defined as a monotonous increasing time function for nonlinear analysis.
The main computational methods available in RADIOSS:
 Cholesky (direct method, linear solver)
 Preconditioned Conjugate Gradient (linear solver)
 Modified Newton-Raphson method (nonlinear solver)
The precondition methods for linear solver available in RADIOSS:
 No preconditioned
 Diagonal Jacobi
 Incomplete Cholesky
 Stabilized incomplete Cholesky
 Factored Approximate Inverse (by default)
You should define the tolerance and stop criterion for the linear and nonlinear solver (residual).
Strategies of resolution for nonlinear static computation/time step control:
 Iterations number limit for updating stiffness matrix
 Convergence iterations number for increasing time step
 Convergence iterations number for decreasing time step
 Increase time step factor
 Decrease time step factor

 Minimum time step
 Maximum time step
 Initial time step
The nonlinear solver uses the modified Newton-Raphson method and the resolution is based on
sparse iterative techniques.
Fig 19: Newton-Raphson resolution in the case of load control technique.
The modified Newton-Raphson method is based on maintaining the tangent matrix for all
iterations and can be combined with the line search acceleration technique for accelerating
convergence.
Piloting techniques available in RADIOSS:
 Displacement norm control
 Arc-length control
An automatic time step control is used.

Static Analysis and Implicit Options
This example deals with two implicit analysis:
 A static linear computation (loading by gravity),
 A static nonlinear computation (three computations are performed: dummy positioning
using an imposed displacement, followed by a concentrated load and a gravity loading).
An adapted modeling methodology is set up for each analysis. Contact with the different
interfaces depends on the computations taken into account and then the material can be
updated.
The goal for this analysis is to propose a modeling method for different loading cases, with
specific input data used in the implicit strategies. The studies by linear implicit and nonlinear
implicit using imposed displacement are no longer comparable with results obtained by explicit
due to the different physical approaches. Comparisons are only valid for the positioning by
gravity loading.
Linear Static Analysis

Type 7 interface uses nonlinear algorithms to check contact. Thus, in order to be used in a linear
solver, it is replaced by a type 2 tied interface which creates kinematic conditions between slave
nodes and master surfaces. Gravity loading is studied.
Fig 20: Type 2 tied interface linear contact for dummy / seat cushion modeling.
The visco-elastic law 35 (generalized Kelvin-Voigt model) describing the foam of a seat is
converted into a linear plastic law 1 (properties are maintained):
 Young’s modulus: 0.2 MPa
 Poisson’s ratio: 0
 Density: 4.3 x 10-11 k g/l

You can select BATOZ formulations for the shell elements and HA8 formulations using 2x2x2
integration points for the brick elements.
The linear implicit methods used are:
Implicit type: Static linear
Linear solver: Direct Cholesky
Precondition method: Factored Approximate Inverse
Stop criteria: Relative residual of preconditioned matrix
Tolerance: 10-6
The implicit options used in the Engine file are:

/IMPL/PRINT/LIN/-1 Printout frequency for linear
iteration
/IMPL/SOLVER/3 Solver method
5 0 3 0.0
/IMPL/LINEAR Static linear computation
Results
Only one animation corresponds to the static solution.
Fig 21: Linear static implicit solution of gravity loading (type 2 interface is used).

It should be noted that this modeling contact slightly modifies the problem which is no longer
comparable with the previous explicit models.
Table 1: Indication of time computation.
Explicit Solver - Implicit Solver –

/DYREL Linear
Normalized 170 1
CPU
Nonlinear Static Analysis
Positioning Using an Imposed Displacement

The modeling methodology defined in the explicit studies is maintained (visco-elastic material
law, type 7 interface, etc.). Brick elements are modeled by default element formulation.
Fig 22: Nonlinear contact modeling with self-impacting type 7 interface.
In addition to the constant gravity load, an imposed displacement along the Z-axis is applied on
the master node of the global rigid body covering the dummy. This approach allows computation
to converge and the rigid body modes to be removed (no null pivot). An input curve for the
imposed displacement is required. The boundary conditions on master node 14199 are: 110 111.

Fig 23: Imposed displacement along the Z-axis as a monotonous increasing time function.

Stop criteria: Relative residual in force
Tolerance: 0.01
Time step control method: Arc-length method and "line-search"
Initial time step: 5s
Minimum time step: 0.01 s
Maximum time step: no
Desired convergence iteration number: 6
Maximum convergence iteration number: 20
Maximum increasing time step scale factor: 1.1
Spring-back option: no
Implicit parameters are set in the Engine file with the options beginning with /IMPL/.
The implicit options used are:
/IMPL/PRINT/NONLIN/-1 Printout frequency for nonlinear
iteration
5 0 3 0.0
/IMPL/NONLIN Static nonlinear computation
5 2 0.01
/IMPL/DTINI Initial time step determines the
1 initial loading increment

/IMPL/DT/STOP Min-Max values for time step
1e-3 0
/IMPL/DT/2 Time step control metho 2 - Arc-
6.0 20 0.8 1.1 length+Line-search will be used
with this method to accelerate and
control convergence
Due to the contact problem, the tolerance value (Tol) is set to 10 -2 (default = 10-3).
Some options are not compatible with the implicit solver. Refer to RADIOSS Starter Input for
more details about implicit options.
Results
The last animation corresponds to the static solution.
Fig 24: Nonlinear static implicit solution of the imposed displacement.
The Z-displacement of the dummy should not be considered as a result but as an input data
(imposed displacement on the master node 14199).

Table 2: Time computation comparison between explicit and implicit computations:

/DYREL Nonlinear
Normalized 1.26 1
CPU
Number of 56704 (1718) 33 (1)

cycles
(normalized)
Positioning Using a Concentrated Load

The modeling methodology defined in the explicit studies is maintained. The gravity loading is
taken into account by applying a constant concentrated load of 813.05N (dummy weight + added
masses) on the master node of the rigid body, including the dummy. In order to remove the rigid
body modes, the dummy is connected to fixed nodes via type 8 spring elements.
Fig 25: Concentrated load along the Z-axis as a monotonous increasing time function.
Fig 26: Springs type 8 defined for removing rigid body modes during implicit computation.

The properties of the general type 8 springs are:
 Linear elastic behavior
 Mass = 1g
 Inertia = 0.001
 Translational stiffness: TX = 1 N/mm
TY = 1 N/mm
TZ = 1 N/mm
 Rotational stiffness: RX = 100 Mg.mm 2/(s2.rad)
RY = 100 Mg.mm2/(s2.rad)
RZ = 100 Mg.mm2/(s2.rad)
Implicit options are the same as the previous implicit problem; except for the initial time step is
set to: 2s.
Results

/DYREL Nonlinear
Normalized CPU 3.07 1
Number of 56704 (1090) 52 (1)

cycles
(normalized)
Z– -12.75 mm -12.49 mm
displacement
(master node
dummy)
Positioning Using Gravity Loading

The modeling methodology defined in the implicit model is maintained using a concentrated load.
Gravity loading is applied on the slave nodes and the master node of the rigid body, including the
dummy. In order to remove the rigid body modes, the dummy is connected to fixed nodes via
type 8 spring elements.

Fig 27: Gravity loading as a monotonous increasing time function.
Implicit options are the same as the previous implicit problem (initial time step is set to: 2s).
Results
Explicit solver - Implicit solver –

/DYREL Nonlinear
Normalized CPU 2.53 1
Number of cycles 56704 (1090) 52 (1)

(normalized)
Z – displacement -12.75 mm -12.42 mm

(master node
dummy)

Fig 28: Convergence results of the X- and Z-displacement of master node 14199 (rigid body dummy) for the implicit
models using gravity loading and concentrated load.

Fig 29: Final dummy position obtained using IMPLICIT (model using gravity loading) and EXPLICIT (model with gravity
loading and kinetic relaxation).
Conclusion
This example brings awareness to the use of the RADIOSS implicit solver in resolving quasi-static
problems. On the other hand, it illustrates different convergence acceleration techniques when an
explicit solver is applied to the quasi-static problems. The advantages and drawbacks of the
methods are compared.

Example 17 - Box Beam
Summary
The crashing of a box beam against a rigid wall is a typical and famous example of simulation in
dynamic transient problems. The purpose for this example is to study the mesh influence on
simulation results when several kinds of shell elements are used.
At first, the quality of the results obtained for the different mesh densities is studied using
several element formulations. Then the mesh transition influence is highlighted. Finally, the
meshes are disturbed and the simulation results are compared.
This example illustrates element sensitivity for various kinds of mesh, in the case of a crash
analysis.

17.1 - Densities
Title
Box Beam - Densities
Number
17.1
Brief Description
A steel box beam, fixed at one end and impacted at the other end by an infinite mass.
Results for mesh with different densities are compared.
Keywords
 Shells Q4
 Type 7 and 11 interface
 Global plasticity, iterative plasticity, and variable thickness
 BT_TYPE1, 3, 4, QEPH, BATOZ, DKT18 and C0 formulation
RADIOSS Options
Input File
Mesh 0:
<install_directory>/demos/hwsolvers/radioss/17_BoxBeam/Densities_mesh/mesh0/.../
BOXBEAM*
Mesh 1:
BOXBEAM*
Mesh 2:
BOXBEAM*
Mesh 3:
BOXBEAM*

Advanced

Overview

A steel box beam fixed at one end, is impacted at the other end by an infinite mass. The
dimensions of the box beam are 203 mm x 50.8 mm x 38.1 mm, and its thickness is 0.914 mm.
As symmetry is taken into account, only one quarter of the structure is modeled.
The material used follows an isotropic elasto-plastic material (/MAT/LAW2) using the Johnson-
Cook plasticity model, with the following characteristics:
 Initial density: 7.8 x 10-3 g/mm3
 Maximum stress: 340 MPa

Four kinds of meshes are used to model the beam. The initial mesh is uniform using a total of 60
x 8 elements. For the three other meshes, the element length is multiplied by 2, 3 and 4, as
shown in the following diagram.
For each model, several element formulations are tested:
 BT_TYPE1
 BT_TYPE3
 BT_TYPE4
 QEPH
 BATOZ
 C0 (T3 element)
 DKT18 (T3 element)
Fig 2: Meshes.
The 3-node shell mesh is obtained by dividing the 4-node shell elements.
 Boundary conditions:
Take into account the symmetry, all nodes in the Y-Z plan are fixed in a Y translation and
an X and Z rotation. One quarter of the structure is modeled.

 Rigid body:
The lower (fixed) end is modeled using a rigid body connecting all lower nodes (Z = 0.0).
The rigid body is completely fixed using translations and rotations.
 Wall:
The impactor is modeled using a sliding rigid wall having a fixed velocity (13.3 m/s) in a Z
direction and is fixed for other translations and rotations.
 Interfaces:
The structure’s self-impact is modeled using a type 7 interface on the full structure. The
interface master surface is defined using the complete model. The slave nodes group is
defined using the master surface.
On top of the beam, possible edge-to-edge impacts are dealt with using a type 11 self-impacting
interface. The edges use the master surface of the type 7 interface as the input surface.
Fig 3: Boundary conditions.
The results are compared using two different views:

 The role and influence of the mesh for a given type of element formulation
 The shell element formulations for a given mesh
Three criteria are used to compare the quality of results obtained:
 Crushing force versus displacement
The crushing force corresponds to normal force in the Z-direction of the impactor (rigid
wall), multiplied by 4 due to the symmetry.
In comparison, the displacement corresponds to the Z-direction motion of the rigid wall’s
master node.
 Total energy
Total energy is the sum of all energies.

Mesh Influence of a Given Shell
Fig 4: Total energy for a BATOZ formulation.
Fig 5: Force for a BATOZ formulation.

Fig 6: Total energy for a QEPH formulation.
Fig 7: Force for a QEPH formulation.

Fig 8: Total energy for a BT_TYPE1 formulation.
Fig 9: Hourglass energy for a BT_TYPE1 formulation.

Fig 10: Force for a BT_TYPE1 formulation.



Fig 17: Total energy for a CO formulation.

Fig 18: Force for a CO formulation.
Fig 19: Total energy for a DKT formulation.

Fig 20: Force for a DKT formulation.
Influence of Element Formulation using Mesh 3
Fig 21: Total energy for different element formulations.

Fig 23: Hourglass energy for different element formulations.

Fig 24: Displacement for different element formulations.
Fig 25: Displacement for different element formulations

MESH 0
MESH 1

MESH 2
MESH 3

MESH 0 MESH 1 MESH 2 MESH 3
EI
3.25 x 105 3.82 x 105 4.88 x 105 7.23 x 105
t = 8 ms
Ehr
- - - -
t = 8 ms
EK
1.32 x 104 1.23 x 104 1.26 x 104 1.10 x 104
t = 8 ms
Total
3.38 x 105 3.94 x 105 5.00 x 105 7.34 x 105
Energy
Error 0.3% 1.1% 1.6% 2.9%

t = 8 ms
Maximum
normal
force on 10350 10491 10953 11555
the wall
(N)
Formulation: QEPH

EI
3.38 x 105 4.55 x 105 5.49 x 105 8.13 x 105
t = 8 ms
Ehr
- - - -
t = 8 ms
EK
1.32 x 104 1.36 x 104 1.35 x 104 0.93 x 104
t = 8 ms
Total
3.51 x 105 4.68 x 105 5.63 x 105 8.23 x 105
Energy
Error 2.0% 2.9% 3.2% 8.0%

t = 8 ms
Maximum
normal
force on 10345 10574 11335 11865
the wall
(N)
Formulation: BT_TYPE1

EI
3.19 x 105 3.60 x 105 4.68 x 105 5.19 x 105
t = 8 ms
Ehr
2.42 x 104 4.17 x 104 3.87 x 104 8.80 x 104
t = 8 ms
EK
1.29 x 104 1.23 x 104 1.16 x 104 1.35 x 104
t = 8 ms
Total
3.32 x 105 3.72 x 105 4.79 x 105 5.32 x 105
Energy
Error -6.4% -9.3% -5.8% -11.5%

t = 8 ms
Maximum
normal
force on 10344 10505 10971 11569
the wall
(N)

EI
3.14 x 105 3.73 x 105 4.46 x 105 4.94 x 105
t = 8 ms
Ehr 11.90 x
2.02 x 104 3.80 x 104 6.56 x 104
t = 8 ms 104
EK
1.31 x 104 1.24 x 104 1.32 x 104 1.29 x 104
t = 8 ms
Total
3.27 x 105 3.85 x 105 4.60 x 105 5.07 x 105
Energy
Error -5.5% -8.2% -11.0% -16.7%

t = 8 ms
Maximum
normal
force on 10353 10526 11000 11670
the wall
(N)

EI
3.23 x 105 3.52 x 105 4.60 x 105 5.26 x 105
t = 8 ms
Ehr
1.26 x 104 1.94 x 104 3.74 x 104 5.02 x 104
t = 8 ms
EK
1.30 x 104 1.24 x 104 1.21 x 104 1.31 x 104
t = 8 ms
Total
3.36 x 105 3.64 x 105 4.72 x 105 5.39 x 105
Energy
Error -3.3% -4.0% -5.8% -6.5%

t = 8 ms
Maximum
normal
force on 10344 10538 11011 11568
the wall
(N)
Formulation: C0

EI
3.45 x 105 4.56 x 105 4.79 x 105 8.64 x 105
t = 8 ms
Ehr
- - - -
t = 8 ms
EK
1.29 x 104 1.30 x 104 1.10 x 104 1.12 x 104
t = 8 ms
Total
3.58 x 105 4.69 x 105 4.90 x 105 8.75 x 105
Energy
Error 0.2% 0.8% 1.7% 2.5%

t = 8 ms
Maximum
normal
force on 10355 10344 10875 11435
the wall
(N)
Formulation: DKT18

EI
3.21 x 105 3.75 x 105 3.97 x 105 4.32 x 105
t = 8 ms
Ehr
- - - -
t = 8 ms
EK
1.29 x 104 1.34 x 104 1.13 x 104 1.45 x 104
t = 8 ms
Total
3.34 x 105 3.88 x 105 4.08 x 105 4.47 x 105
Energy
Error 0.5% 0.8% 1.6% 1.9%

t = 8 ms
Maximum
normal
force on 10348 10367 10800 11139
the wall
(N)


17.2 - Transitions
Title
Box Beam -
Transitions
Number
17.2
Brief Description
A steel box beam, fixed at one end, impacted at the other end by an infinite mass.
Results for meshes with different transitions are compared.
Keywords
 Q4 shells
 BT_TYPE1-3-4, QEPH, BATOZ, DKT18 and C0 formulation
RADIOSS Options
Input File
Mesh 0:
<install_directory>/demos/hwsolvers/radioss/17_BoxBeam/Transition_mesh/mesh0/...//
BOXBEAM*
Mesh 1:
BOXBEAM*
Mesh 2:
BOXBEAM*
Mesh 3:
BOXBEAM*

Advanced

Overview

As symmetry is taken into account, only one quarter of the structure is modeled. Four kinds of
mesh and three plasticity formulations are compared (global plasticity, five integration points and
iterative plasticity).
The material used follows an isotropic elasto-plastic material (/MAT/LAW2) with the Johnson-
Cook plasticity model, having the following characteristics:
 Initial density: 7.8 x10-3 g/mm3
Fig 26: Problem studied

Four types of mesh are used to model the beam. The beam is divided into two parts: a fine mesh
for one half (8 x 26 elements) and a coarse mesh for the other half. Transition between the two
meshes of a single beam is carried out using a mesh transition element of the same length as
each particular case. You will compare results using a reference case which has a uniform mesh.
The layout of the elements is shown in the following diagram. The following are tested for each
model:
1. Element formulation:
 BT_TYPE1
 BT_TYPE3
 QEPH
 BATOZ
 C0
 DKT18
2. Plasticity:
 Global plasticity
 Progressive plasticity with five integration points
 Iterative plasticity with five integration points and variable thickness
Fig 27: Meshes

 Rigid body:
The rigid body is completely fixed in translations and rotations.
 Wall:
The impactor is modeled by a sliding rigid wall using a fixed velocity (13.3 m/s) in the Z-
direction and fixed for other translations and rotations.
 Interfaces:
On top of the beam, the possible edge-to-edge impacts are dealt with using a type 11 self-
impacting interface. The edges use the master surface of the type 7 interface as the input
surface.

The results are compared using three different views:

 Role and influence of the mesh for a given type of element formulation
 Shell element formulations for a given mesh
 Plasticity options for a given mesh and element formulation
Three criteria are used to compare the quality of the results obtained:
The crushing force corresponds to the normal force in the Z-direction of the impactor (rigid
wall), multiplied by 4 due to symmetry.
For comparison, displacement corresponds to the Z-direction motion of the rigid wall’s
master node.
 Total energy
Total energy is the sum of all energies.

Mesh Influence for a Given Shell Using Global Plasticity
BATOZ
Fig 29: Total energy for a BATOZ formulation.
Fig 30: Force for a BATOZ formulation.

QEPH
Fig 31: Total energy for a QEPH formulation.
Fig 32: Force for a QEPH formulation.

BT_TYPE1

BT_TYPE3


BT_TYPE4

C0
Fig 42: Total energy for a CO formulation.

Fig 43: Force for a CO formulation.
DKT18
Fig 44: Total energy for a DKT formulation.

Fig 45: Force for a DKT formulation.

Influence of Element Formulation Using Mesh 3 and Global Plasticity

Fig 48: Hourglass energy for different element formulations.
Fig 49: Force for different element formulations.

Fig 50: Displacements for different element formulations.
Influence of Plasticity Options Using Mesh 1 and BT_TYPE3 Formulation
Fig 51: Total energy for different plasticity computations.

Fig 52: Hourglass energy for different plasticity computations.
Fig 53: Force for different plasticity computations.

MESH 0
MESH 1

MESH 2
MESH 3

Formulation: QEPH


Formulation: C0
Formulation: DKT18



17.3 - Distorted
Title
Box Beam - Distorted
Number
17.3
Brief Description
A steel box beam, fixed at one end and impacted at the other by an infinite mass.
Results for distorted meshes are compared.
Keywords
 Q4 shells
 BT_TYPE1-3-4, QEPH, BATOZ, DKT18 and C0 formulation
RADIOSS Options
Input File
Mesh 0:
<install_directory>/demos/hwsolvers/radioss/17_BoxBeam/Distorted_mesh/mesh0/...//
BOXBEAM*
Mesh 1:
BOXBEAM*
Mesh 2:
BOXBEAM*
Mesh 3:
BOXBEAM*

Advanced

Overview

As symmetry is taken into account, only one quarter of the structure is modeled. Four kinds of
mesh and three plasticity formulations are compared (global plasticity, five integration points and
iterative plasticity).
The material used follows an isotropic elasto-plastic (/MAT/LAW2) with the Johnson-Cook
plasticity model, with the following characteristics:
 Initial density: 7.8 x10-3 g/mm3

Four beams are modeled with different kinds of mesh, all having 56 elements in length and 8 in
height. The layout of the elements is shown in the following diagram.
The following are tested for each model:
1. Element formulation:
 BT_TYPE1
 BT_TYPE3
 QEPH
 BATOZ
 C0
 DKT18
2. Plasticity:
 Global plasticity
 Progressive plasticity with five integration points
 Iterative plasticity with five integration points and variable thickness
Fig 55: Meshes.

 Rigid body:
The rigid body is completely fixed in translations and rotations.
 Wall:
The impactor is modeled using a sliding rigid wall with a fixed velocity (13.3 m/s) in the Z-
direction and fixed for other translations and rotations.
 Interfaces:
On top of the beam, the possible edge-to-edge impacts are dealt with using a type 11 self-
impacting interface. The edges use the master surface of the type 7 interface as the input
surface.
The results are compared using three different views:

 The role and influence of the mesh for a given type of element formulation.
 Shell element formulations for a given mesh.
 Plasticity options for a given mesh and element formulation.

Three criteria are used to compare the quality of the results obtained:
The crushing force corresponds to the normal force in the Z-direction of the impactor (rigid
wall), multiplied by 4 due to symmetry.
For comparison, displacement corresponds to the Z-direction motion of the rigid wall’s
master node.
 Total energy
Total absorption energy is the sum of internal energy and hourglass energy.
Mesh Influence of a Given Shell Using Global Plasticity and BT_TYPE3 Formulation


Influence of Element Formulation Using Mesh 1 and Global Plasticity
Fig 60: Total energy for different element formulation.
Fig 61: Total energy for different element formulation.

Fig 62: Hourglass energy for different BT element formulation.
Fig 63: Force for different element formulation.

Fig 64: Force for different element formulation.
Influence of Plasticity Options Using Mesh 0 and BT_TYPE3 Formulation
Fig 65: Total energy.

Fig 66: Hourglass energy.
Fig 67: Crushing force.

MESH 0
MESH 1

MESH 2
MESH 3



Formulation: QEPH
Formulation: BATOZ

Formulation: DKT18
Formulation: C0
Conclusion
The crash of a box beam using several meshes and finite element formulations was studied in
detail. The simulation results for uniform, mapped and transit meshes are classified and
compared for each different shell formulation. The results obtained illustrate the sensitivity of the
shell elements with respect to the quality of the mesh for a typical crash problem.

Example 18 - Square Plate
Summary
A square plane subjected to in-plane and out-of-plane static loading is a simple element test. It
allows you to highlight element formulation for elastic and elasto-plastic cases. The under-
integrated quadrilateral shells are compared with the fully-integrated BATOZ shells. The triangles
are also studied.

18.1 - Square Plate Torsion
Title
Square plate torsion
Number
18.1
Brief Description
Torsion test on a cantilever plate submitted to two opposing forces on the same side.
Keywords
 Q4 shells
 T3 shells
 Hourglass, mesh, and concentrated loads
RADIOSS Options
 Concentrated loads (/CLOAD)
 Element formulation (/PROP)
Input File
4Q4: <install_directory>/demos/hwsolvers/radioss/18_Square_plate/Torsion/4Q4/.../
TORSION*
8T3: <install_directory>/demos/hwsolvers/radioss/18_Square_plate/Torsion/8T3/.../
TORSION*
8T3 inv:
<install_directory>/demos/hwsolvers/radioss/18_Square_plate/Torsion/8T3_inv/.../
TORSION*
2Q4-4T3: <install_directory>/demos/hwsolvers/radioss/18_Square_plate/Torsion/
2Q4-4T3/.../TORSION*

Beginner

Overview

This example concerns a torsion problem of an embedded plate subjected to two concentrated
loads, as shown in the following diagram. This example illustrates the role of the different shell
element formulations with regard to the mesh.
The material used follows a linear elastic behavior with the following characteristics:
Four different types of mesh are used:
 Mesh 1: two quadrilateral shells and four triangular shells (2Q4-4T3)
 Mesh 2: four quadrilateral shells (4Q4)
 Mesh 3: eight triangular shells (8T3)
 Mesh 4: eight triangular shells (8T3 inverse)

For each model, the following shell formulations are tested:
 QBAT formulation (Ishell =12)
 QEPH formulation (Ishell =24)
 Belytshcko & Tsay formulation (I shell =1 or 3, hourglass control type 1, 3)
 C0 and DKT18 formulations
Fig 2: Square plate meshes.

The boundary conditions are such that the three nodes of a single side and the two middle ones
are blocked, while the others are free with respect to the Y axis.
Two concentrated loads are applied on the corner points of the opposite side. They increase over
time as defined by the following function:
F(t) 0 10 10
t 0 200 400

Fig 3: Boundary conditions and loads.

This example compares several models concerning:
 the use of different element formulations for each mesh
 the different types of mesh for a given element formulation
Two criteria used to compare the results are:
 absorbed energy (internal and hourglass)
 vertical displacement of the node under the loading point
The following diagrams summarize the results obtained.

Energy Curves / Comparison for Element Formulations
Mesh 1: 2Q4-4T3
Fig 4: Internal energy for 2 x Q4 and 4 x T3 elements.

Mesh 2: 4Q4
Fig 5: Internal energy for 4 x Q4 elements

Meshes 3 and 4: 8T3 and 8T3_INV
Fig 6: Internal energy for 8 x T3 elements.

Energy Curves / Comparison for Mesh Definitions
Fig 7: Internal energy for different meshes.
Fig 8: Hourglass energy for different meshes.

Displacement and Maximum Energy Comparison
2 Q4- 4 T3 4 Q4 8 T3 8 T3 Inverse
BT_TYPE
QEPH BT_TYPE4 BATOZ QEPH BT_TYPE1 BT_TYPE4 BATOZ DKT C0 DKT C0
1
IEmax 2.74x10-2 2.35x10-2 2.37x10-2 7.21x10-2 3.64x10-2 2.93x10-2 2.97x10-2 2.30x10-2 1.37 x10-1 1.69x10-2 1.37x10-1 1.69x10-2
HEmax --- 1.01x10-4 1.03x10-4 --- --- 1.94x10-4 1.98x10-6 --- --- --- --- ---
DZmax 1.75x10-3 1.78x10-3 1.78x10-3 1.21x10-2 2.42x10-3 2.95x10-3 2.97x10-3 2.30x10-3 1.44x10-2 1.69x10-3 1.44x10-2 1.69x10-3
Conclusion
A square plate under torsion is a severe test to study the behavior of shell elements in torsion-bending. A general overview of the results obtained
highlight the following key points:
 For the 4Q4 mesh, the results obtained using QBATOZ and QEPH are similar. BT elements are too flexible and are not significantly
influenced by the hourglass formulation, due to the in-plane mesh.
 For triangular meshes, the DKT element is able to bend much better, the co-element being too stiff.
 The mesh with both Q4 and T3 elements may not comment like the other two, as one part uses the triangle elements employed in
RADIOSS.

18.2 - Square Membrane Elastic
Title
Square membrane
elastic
Number
18.2
Brief Description
Square plate submitted to two opposing in-plane end forces.
Keywords
 Q4 shells
 T3 shells
RADIOSS Options
Input File
4Q4:
<install_directory>/demos/hwsolvers/radioss/18_Square_plate/Membrane_elastic/4Q4/.
../TRACTION*
8T3:
<install_directory>/demos/hwsolvers/radioss/18_Square_plate/Membrane_elastic/8T3/.
../TRACTION*
8T3 inv:
<install_directory>/demos/hwsolvers/radioss/18_Square_plate/Membrane_elastic/
8T3_inv/.../TRACTION*
2Q4-4T3:
<install_directory>/demos/hwsolvers/radioss/18_Square_plate/Membrane_elastic/
2Q4-4T3/.../TRACTION*

Beginner

Overview

This example concerns the in-plane traction-comparison problem of an embedded plate subjected
to two concentrated loads, as shown in the following diagram. This example illustrates the role of
the different shell element formulations with regard to the mesh.
The material used follows a linear elastic behavior and has the following characteristics:
 Initial density: 7.8 x 10-3 g/mm3
Four different types of mesh are used:

For each model, the following shell formulations are tested:
 Belytshcko & Tsay formulation (I shell =1 or 3, hourglass control type 1, 3)
 C0 and DKT18 formulations

are blocked, whereas the others are free with respect to the Y axis.
Two concentrated loads are applied on the corner points on opposing sides. They increase over
time, as defined by the following function:
F(t) 0 10 10
t 0 200 400

Simulation Results


Mesh 1: 2Q4-4T3
Fig 13: Y Displacement for 2 x Q4 and 4 x T3 elements

Mesh 2: 4Q4
Fig 14: Internal energy for 4 x Q4 elements.

Energy Curves / Comparison for Mesh Definitions

2Q4-4T3 4Q4 8T3 8T3_INV

Elastic
Plate BT_TYPE 1 and BT_TYPE 1 and
QEPH BATOZ QEPH BATOZ DKT CO DKT CO
3 3
IEmax 1.07 x 10-2 1.19 x 10-2 1.07 x 10-2 1.24 x 10-2 1.44 x 10-2 1.24 x 10-2 6.42 x 10-3 6.42 x 10-3 6.42 x 10-3 6.42 x 10-3
HEmax --- 2.10 x 10-5 --- --- 3.49 x 10-6 --- --- --- --- ---
Dymax 1.18 x 10-3 1.38 x 10-3 1.18 x 10-3 1.24 x 10-3 1.44 x 10-3 1.24 x 10-3 6.42 x 10-3 6.42 x 10-3 6.42 x 10-3 6.42 x 10-3
(Traction) (Traction) (Traction)
Conclusion
In the case of elastic flat plate modeling, when the loading is in-plane, the shell elements are reduced to become a membrane if the loads applied
do not cause buckling.
A general overview of the results obtained highlight the following key points:
1. The quadrilateral shell elements QEPH and QBAT have the same in-plane behavior.
2. The different types of hourglass formulations in the BT shell elements lead to the same results, as there is no out-of-plane deformation and
the material is supposed to be elastic.
3. The three in-plane behaviors of the DKT18 and T3C0 RADIOSS triangles are exactly the same, as both of the elements are used for the same
membrane formulation.
4. The triangles are stiffer than the quadrilateral elements and do not provide good results, especially when the mesh is coarse.
Refer to the RADIOSS Theory Manual for more details.

18.3 - Square Membrane Elasto-Plastic
Title
Square membrane
elasto-plastic
Number
18.3
Brief Description
Square plate submitted to two opposing in-plane end forces.
Keywords
 Q4 shells
 T3 shells
RADIOSS Options
Input File
4Q4: <install_directory>/demos/hwsolvers/radioss/18_Square_plate/Membrane_elasto-
plastic/4Q4/.../TRACTION*
8T3: <install_directory>/demos/hwsolvers/radioss/18_Square_plate/Membrane_elasto-
plastic/8T3/.../TRACTION*
8T3 inv:
<install_directory>/demos/hwsolvers/radioss/18_Square_plate/Membrane_elasto-
plastic/8T3_inv/.../TRACTION*
2Q4-4T3:
<install_directory>/demos/hwsolvers/radioss/18_Square_plate/Membrane_elasto-
plastic/2Q4-4T3/.../TRACTION*

Beginner

Overview

This example concerns the torsion problem of an embedded plate subjected to two concentrated
loads, as shown in the following diagram. This example illustrates the role of different shell
element formulations with regard to the mesh.
The material used follows an isotropic elasto-plastic behavior with the Johnson-Cook plasticity
model (/MAT/LAW2), with the following characteristics:
 Initial density: 0.0078 g/mm3

Four shells are modeled with different types of mesh:
For each model, the following element formulation is tested:
 Belytshcko & Tsay formulation (Ishell =1 or 3, hourglass control type 1, 3)
 C0 and DKT18 formulation

are blocked, while the others are free with respect to the Y axis.

Two concentrated loads are applied on the corner points of the opposite side. They increase over
time, as defined by the following function:
F(t) 0 10 10
t 0 200 400
Simulation Results


Mesh 1: 2Q4-4T3
Fig 21: Hourglass energy for 2 x Q4 and 4 x T3 elements.

Fig 22: Force for 4 x Q4 elements.

Mesh 2: 4Q4
Fig 23: Internal energy for 4 x Q4 elements.
Fig 24: Hourglass energy for 4 x Q4 elements.

Fig 25: Total energy for 4 x Q4 elements.
Fig 26: Force for 4 x Q4 elements.

Fig 27: Internal energy for 8 x T3 elements.
Fig 28: Force for 8 x T3 elements.

Mesh 4: 8T3_INV
Fig 29: Internal energy for 8 x T3 elements (inversed mesh).
Fig 30: Force for 8 x T3 elements (inversed mesh).

Comparison for Different Meshes

Fig 33: Force for different meshes.

Conclusion
The purpose of this example was to highlight the role of the elasto-plastic treatment when
formulating RADIOSS shells. The in-plane plasticity was considered here. Regarding the applied
boundary conditions and the Poisson effect on the plate, the test may be very severe with
respect to the behavior of plastic in under-integrated elements.
In the case of a mesh with four quadrilaterals, the QBAT element always provides the best results
as it allows four integration points to be put over the element. The plasticity computation over
the integration points is thus more accurate. The under-integrated elements, having just one
integration point at the center, allows only two integration points to be put through the width of
the mesh. Another point concerns the role of Poisson’s ratio in the plasticity computation. In
fact, the QEPH element uses an analytical expression of the hourglass energy which takes into
account the accurate expression in terms of the Poisson ratio (refer to the RADIOSS Theory
Manual for further information). However, some approximations are induced in its elasto-plastic
formulation, possibly influencing the results, especially for low levels of work-hardening. In the
BT element formulation with a type 3 hourglass control, the Poisson ratio effect on the plastic
part of the hourglass deformation is computed by a simplified expression which minimizes its
role. In fact, the results obtained using BT_TYPE3 are slightly affected by the change in (use
=0 for the example studied and compare the results obtained). The BT elements are generally
more flexible and provide better results for a very coarse mesh.
For triangular meshes, the in-plane behavior of DKT18 should be noted as being the same as the
T3C0 element. In fact, the elements are essentially different with respect to their bending
behaviors.
When combining the T3 and Q4 elements, the results generally come between a uniform
triangular mesh and a quadrangular mesh.

Example 19 - Wave Propagation
Summary
Elastic shock wave propagation on a half-space is studied using two different approaches:
 Lagrangian formulation
 ALE (Arbitrary Lagrangian Eulerian) formulation
The simulation results are compared with an analytical solution. A bi-dimensional problem is
considered.
The domain subjected to the vertical impulse load undergoes an elastic material law process. The
generated shock wave is composed of a longitudinal wave and a shear wave. Results are
indicated in 0.77 ms, for which the longitudinal wave is predicted to reach the lower boundary of
the domain. In order to ensure an accurate wave expansion, an infinite domain is modeled using
a non-reflective frontiers (NRF) material law available in the ALE formulation.

Title
Lagrangian
Number
19.1
Brief Description
Elastic wave propagation on a half-space subjected to a vertically-distributed load.
Keywords
 Bi-dimensional analysis, quad and general solid
 Impulse load, shock wave propagation, longitudinal and shear waves
 ALE and Lagrangian modeling
 Non-reflective frontiers (NRF) material and infinite domain
RADIOSS Options
 Bi-dimensional analysis (/ANALY)
 Function (/FUNCT)
 Non-reflective frontiers (NRF) material law 11 (/MAT/BOUND)

 Lagrangian vs. ALE modeling/Analytical solution
Input File
Lagrangian modeling:
<install_directory>/demos/hwsolvers/radioss/19_Wave_propagation/
Lagrangian_formulation/WAVE*
ALE modeling:
<install_directory>/demos/hwsolvers/radioss/19_Wave_propagation/ALE_formulation/
WAVE*

Medium

Overview
Aim of the Problem

This example studies wave propagation through a bi-dimensional domain. Two analysis' are
performed using:
 a Lagrangian formulation
 a ALE (Arbitrary Langrangian Eulerian) formulation
The simulation results are compared to an analytical solution.
A half-space is subjected to a vertical load distributed over a varied time span and creating wave
propagation in the domain. The dimensions of the model are 8 m x 4.76 m and the impulse load
is applied over a 1 m-width zone.
Units: m, s, Kg, N, Pa.
The material used follows a linear elastic law (/MAT/LAW1) and has the following characteristics:
 Initial density: 2842 kg.m-3
 Young’s modulus: 73 GPa
The expansion process of the shock wave is comprised of the longitudinal and shear waves.
Based on these material properties, the propagation speed of longitudinal waves in the material
correspond to 6169.1 m.s-1 and 3107.5 m.s-1 for shear waves. Thus, the longitudinal waves
should reach the lower boundary of the domain in about 0.77 ms.

The wave pattern caused by a distributed load is shown in Fig 2.
Fig 2: Longitudinal and shear waves comprising the wave pattern.
The impulse load is described by the sinusoidal function: F(t) = sin(2 * 105t) GPa
The part is modeled using a regular mesh with 19080 QUAD elements (44.9 mm x 44.4 mm with
lc =63.15 mm).
Fig 3: Mesh of the bi-dimensional domain.

A bi-dimensional problem is considered. The flag N2D3D defined in /ANALY is set to 2. The 2D
analysis defines the X-axis as the plane strain direction.
The applied vertical pulse is a concentrated load (/CLOAD) in the form of a sinusoidal function
having an amplitude F = 1 GPa and a time period of T = 2 x E-5 s.

Fig 4: Variation of the impulse load over time.
Specific Options for the Lagrangian Modeling

Boundary conditions: Three sides of the model are fixed in terms of translation.
Fig 5: Fixed sides.
The limitation of this approach is the reflection on the domain’s boundaries. Simulation results
are shown for the point in time prior to the shock hitting the low side (< 0.77 ms).
Specific Options for the ALE Modeling
Non-reflective frontiers (NRF): The mesh includes quiet boundary elements to model the infinite
domain. These elements minimize the reflection of the propagating waves. The material used for
these elements follow a non-reflective frontiers (NRF) material law 11 (type 3) as a non-reflective
frontiers (NRF), and has the following characteristics:
 Initial density: 2842 kg.m3
 Characteristic length: 0.0632m

Fig 6: Infinite domain modeled by the non-reflective frontiers (NRF) material law 11 (type 3).
The materials have to be declared ALE using /ALE/MAT in the input desk.
Comparison of Lagrangian and ALE Results with the Analytical Solution

Figures 7 and 8 represent the von Mises stress wave propagation and the velocities at t=0.77
ms.
Fig 7: von Mises isovalues at t=0.77 ms.

Fig 8: Velocity isovalues at t=0.77 ms.
The shock wave propagation is well predicted. Simulation results obtained at t=0.77ms
corroborate the analytical solution: Longitudinal and shear waves.
Lagrangian Results
Wave Pattern
The wave pattern produced by the distributed load shown previously can be identified in the
deformed configuration when the longitudinal wave reaches the lower boundary of the mesh.
Fig 9: Wave pattern in domain at t=0.77 ms.

Vertical Displacement
The graphs below shows the vertical displacement (DZ) of three nodes respectively positioned at
0 m, 3.2 m and 4.75 m under the edge of the distributed load.
Fig 10: Nodes saved in Time History.
Figure 11 shows the vertical displacement of Node 0. The beginning of the wave propagation can
be seen during the time [0; 1.35e-04]. The response after the end of the application force
[1.35e-04; 4e-04] is due to the shear wave.

Fig 11: Z-displacement of "Node 0".
The vertical response of Node 1 shows that the longitudinal wave reaches it in 0.47 ms (Fig 12).
The reflection can be seen after 0.97 ms. The shear wave does not appear because its motion is
in the horizontal direction.

Fig 12: Z-displacement of "Node 1".
The displacement of Node 2 placed at the other extremity of the pattern, shows that the
longitudinal wave crosses the model in 0.7 ms, in accordance with the analytical results.

Fig 13: Z-displacement of "node 2".

Horizontal Displacement
Figure 14 shows the horizontal displacement of Node 1 (placed 3.2 m below the load surface).
The horizontal component of the longitudinal wave reaches the node in 0.49 ms, while the shear
wave arrives at 1.1 ms. Any response after this time results from the different reflections of the
longitudinal and shear waves.
Fig 14: Y-displacement of "node 0".

Fig 15: Global energy assessment.
ALE Results
The wave pattern produced by a distributed load can be identified in the deformed configuration
by displaying the pressure. The grid is fixed and nodal displacements are equal to zero. The
following figure shows propagation when the longitudinal wave reaches the lower boundary of the
mesh.

Fig 16: Pressure isovalues at time t=0.77 ms.
Conclusion
The wave propagation in a finite domain is studied using Lagrangian and ALE approaches. The
Lagrangian formulation does not allow an infinite domain to be defined. Reflections of the
longitudinal and shear waves against boundaries restrict simulation in terms of time (t < 0.77
ms). The ALE approach allows you to model an infinite domain by defining the non-reflective
frontiers (NRF) material (Law 11 - type 3) on the limits. Such specific modeling minimizes the
reflection of the expansion wave.
The bi-dimensional analysis illustrates a planar propagation. An accurate representation of the
wave pattern is obtained and the simulation results are in a closed agreement with the analytical
solution.

Example 20 - Cube
Summary
The fall of an ice cube dropping on two sloped beams is studied to illustrate the use of an explicit
time integration scheme in resolving a transient dynamic analysis with free deformable flying
objects. The impact and the rebound are modeled easily using various types of RADIOSS contact
algorithms. Due to the rotary motion of the ice cube, a co-rotational solid formulation is required.

Title
Cube
Number
20.1
Brief Description
Ice cube dropping on two sliding channels.
Keywords
 Brick elements and 16-node shell elements
 Co-rotational formulation
RADIOSS Options
 Type 7 interface (/INTER/TYPE7) and type 16 (/INTER/LAGMUL/TYPE16)
Input File
Model: <install_directory>/demos/hwsolvers/radioss/20_Cube/CUBE*

Beginner
Overview
Aim of the Problem

This problem demonstrates comparing two interfaces which will allow a sliding contact between
an ice cube and the steel beams to be modeled.
The cube is submitted to gravity and slides on inclined fixed beams and is collected in a cup. The
width of the cube is 30 mm and the dimensions of the beams are 40 x 30 x 500 mm.
The material used for the cube is ice and has a linear elastic behavior (/MAT/LAW1), with the
following characteristics:
 Initial density: 916 Kg/m3

The material used for the beams and the cup is steel and follows an isotropic elasto-plastic
material (/MAT/LAW2) with the Johnson-Cook plasticity model, having the following
characteristics:
 Initial density: 7800 kg/m3
Units: m, s, Kg, N, Pa
Fig 1: Overview of the problem.
Contact between the ice cube and the first beam is modeled using a type 16 interface. Contact
between the ice cube and the second beam is modeled using a type 7 interface. A type 7
interface defines contact between the ice cube and the cup.
 The first beam is modeled using twelve 16-node thick shell elements.
 The second beam is modeled using twelve 8-node brick elements.
 The ice cube is modeled using 8-node brick elements having a co-rotational solid
formulation.
 The cup is modeled with twelve standard shell elements.

Boundary conditions:
 The ice cube nodes are constrained in the Y translation and rotation is around the X-Z axis.
 The lower nodes of the beams are constrained in all directions.
 The cup is constrained in all directions.
Load:
 A gravity load (g = 9.81 m/s2) in the Z-direction is applied on the ice cube’s nodes.
Interface:
 The type 16 interface is used by deactivating the "tied" option, which enables a sliding
contact to be modeled. Ice cube nodes are slave and the upper surface of the beam defines
the master surface.
 The type 7 interface between the ice cube and the second beam uses the Penalty method,
with an initial gap of 1.5 mm. Ice cube nodes are slave and the master surface is defined
using the upper surface of the beams. Friction is not taken into account.
 The type 7 interface between the ice cube and the cup uses the same parameters as those
defined above. Ice cube nodes are slave and the cup defines the master surface.
Fig 2: Boundary conditions. Interfaces.

The results below represent the trajectory of the ice cube and the cube’s reaction forces on the
channels. The ice cube trajectory is obtained using a post-processing option, which enables to
draw the trajectory of a picked node (here the center ice cube node) throughout simulation.
Fig 3: Trajectory of the ice cube.

Fig 4: Reaction forces of the ice cube on channels.
Conclusion
This demonstrative example illustrated the capacity of RADIOSS to simulate sliding contacts,
either using a Lagrangian (type 16 interface) or a Penalty method (type 7 interface).
The co-rotational solid formulation is essential in this case, taking into account the ice cube’s
rotary motion.

Example 21 - Cam
Summary
A cam can be considered as a device that translates motion from circular to linear. The camshaft
of a car takes the rotary motion of the engine and translates it into the linear motion required for
operating the intake and exhaust valves.
The purpose of this example is to illustrate the capacity of RADIOSS to simulate the dynamic
behavior and the kinematic motion of a cam-valve system. The smooth motion of the valve can
be simulated using an adequate and accurate contact model having contact algorithms with
quadratic surfaces without any gaps.

Title
Cam
Number
21.1
Brief Description
The modeling of a camshaft, which takes the engine’s rotary motion and translates it into linear
motion for operating the intake and exhaust valves, is studied.
Keywords
 Penalty/Lagrangian contact, type 7 interface, and type 16 interface
 Linear/quadratic elements and quadratic surface contact
RADIOSS Options
 BRIC20 elements (/BRIC20)
 SHEL16 elements (/SHEL16)
 Initial velocities around axis (/INIVEL/AXIS)
 Spring element (/PROP/SPRING)
 Type 16 interface (/INTER/LAGMUL/TYPE16) and type 7 (/INTER/LAGMUL/TYPE7)
Input File
Interface 16:
Fine mesh:
<install_directory>/demos/hwsolvers/radioss/21_Cam/interface16/fine_mesh/I16S16FM*
Coarse mesh:
<install_directory>/demos/hwsolvers/radioss/21_Cam/interface16/coarse_mesh/I16S16CM*
Interface 7:
Penalty method:
<install_directory>/demos/hwsolvers/radioss/21_Cam/interface7/penalty/slave_cam/
I7PMCAM*
<install_directory>/demos/hwsolvers/radioss/21_Cam/interface7/penalty/
slave_valve/I7PMVALVE*
Lagrange multipliers:
<install_directory>/demos/hwsolvers/radioss/21_Cam/interface7/lagrange/slave_cam/
I7LMCAM*
<install_directory>/demos/hwsolvers/radioss/21_Cam/interface7/lagrange/
slave_valve/I7LMVALVE*
Friction:
<install_directory>/demos/hwsolvers/radioss/21_Cam/interface7/friction/I7PFMCAM*

Medium

Overview
Aim of the Problem

Modeling a contact between a plane and a curved surface uses a faceted curved surface.
Interfaces 7 and 16 are compatible with the geometry of the problem and the faceting, are
described and compared.
This problem demonstrates a cam rotating with an angular velocity of 314 rad/s, which interacts
and triggers off the translation of a valve tied to two springs. The superposed springs have
varying stiffness (spring 1: 30000 N/m and spring 2: 15000 N/m). The springs control the higher
and lower rotary frequencies.
 The cam is 36 mm in length, with a maximum width of 14 mm and a thickness of 18 mm.
 The valve is 44 mm in diameter, with a thickness of 3 mm (Fig 1).
 The spring is 40 mm in length.
The following system is used: mm, s, kg, mN , KPa.
The material used for the cam and the valve is steel. It is characterized by the isotropic elasto-
plastic material (/MAT/LAW2) and the Johnson-Cook plasticity model, with the following
properties:
 Initial density: 7.8 x 10-06 Mkg/l
 Young modulus: 2.1 x 10+08 KPa
 Yield stress: 20000 KPa
 Hardening parameter: 40000 KPa

The problem raised in this example is the modeling of an interface between a plane and a curved
surface. In this case, using quadratic elements is the most appropriate.
A type 16 interface with 16-node shell elements for both the valve and the cam should be chosen
first for the modeling. 20-node brick elements are used for the mesh inside the cam (Fig 2).
Fig 2: BRIC20 and SHEL16 mesh.
Another model using linear elements is studied. Contact between the cam and valve is defined by
a type 7 interface. In order to satisfy the closed contact between parts, the Lagrange Multipliers
method is selected.

Fig 3: BRICK elements mesh.
The springs are modeled using RADIOSS type 4 springs. The stiffness is linear and defined by the
following functions. Damping is neglected.
l-l0 (mm) -40 0 50
Fspring 1 (mN) -1.5 e+06 -0.3 e+06 1.2 e+06
Fspring 2 (mN) -0.75 e+06 -0.15 e+06 0.6 e+06

 Rigid bodies:
In order to apply a constant angular velocity to the cam, a rigid body is created over the
internal nodes, as shown in Fig 4. The master node is moved to the camshaft axis.
To attach the valve head to the spring, another rigid body is created to distribute the
internal spring force over several nodes.

Fig 4: Rigid body cam
- Master node of the cam is blocked, except when rotating around Y.
- Master node of the valve is blocked, except when translating around Z.
- One extremity of the spring is fixed to the valve, while the other is blocked.
Fig 5: Boundary condition on valve.

 Imposed velocity:
A rotational velocity of 314 rad/s is imposed on the master node of the rigid body. This
velocity is activated by a temporal sensor, with a short activation delay (Tdelay =0.0002s).
This sensor is necessary to avoid applying the initial and imposed velocities at the same
time.
Fig 6: Imposed velocity.
 Initial velocity:
An initial rotational velocity is applied to all the cam’s nodes, including the master node of
the rigid body. You must define the origin (center of rotation) and the orientation vector.
Fig 7: Initial velocity.

 Interfaces:
The type 16 interface simulates a contact between a quadratic master surface and a group
of nodes. In the case of contact between a curved and a plane surface, the curved surface
is defined as the master surface and the nodes of the plane part are slave.
Fig 8: Interface 16.
The type 7 interface works either with Penalty or with Lagrange multipliers. In its basic
formulation, the interface simulates contact between two facetisated surfaces. The use of the
Lagrange Multipliers method enables to precisely satisfy the kinematic contact without
introducing a gap.
Fig 9: Interface 7.

At first, you are interested by the kinematics of the problem. The results obtained for velocity
and acceleration at the master node of the rigid body’s valve are thus compared.
The graphs in Fig 10 provide the velocity of the valve’s master node when a type 7 interface is
used with the Penalty method. The raw results obtained are noisy due to the fact that the Penalty
method applies discontinued forces. The smooth velocity curve is obtained by using a low pass
CFC 180 (3db) filter. Figure 11 shows the acceleration of the same master point.
Fig 10: Vertical velocity of the master node valve for a type 7 interface, using the Penalty method.
Attention should be paid to the use of filters. The filtered curves are in fact generally affected by
a border effect. Filtering induces an error in the beginning and end parts of the curves (for this
example, take the intervals of 0 < t < 0.002 and 0.038 < t < 0.04).

Fig 11: Vertical acceleration of the master node valve for a type 7 interface, using the Penalty method.
The filtering quality depends on the number of samples which, in this case is the number of
points computed by RADIOSS for each curve. Therefore, a low value for the /TFILE parameter in
the Engine file (*_0001.rad) is used to obtain good results, especially for the acceleration curve.
In the following sections, only the filtered curves are represented in order to the compare
different models.
Comparison of Interfaces
Figures 12 and 13 represent velocity and acceleration curves for a model using a type 7 interface
with the Penalty method. As for the master and slave part definition, the results are slightly
different.
Figures 14 and 15 give velocity and acceleration curves for type 7 interface using the Lagrange
Multipliers method.

Fig 12: Vertical velocity of the valve’s master node for a type 7 interface, using the Penalty method.
Fig 13: Vertical acceleration of the valve’s master node for a type 7 interface, using the Penalty method.

For both Figures 12 and 13, the model using slave nodes on the cam and a master surface on the
valve seems to be the most realistic.
Fig 14: Vertical velocity of the valve’s master node for a type 7 interface, using the Lagrange Multipliers method.
Fig 15: Vertical acceleration of the valve’s master node for a type 7 interface, using the Lagrange Multipliers method.

Even if using a type 7 interface with the Penalty or the Lagrange Multipliers method good results
can be achieved, a quadratic mesh with the type 16 interface will enable the reduction of
oscillations, due to facetisation.
Figures 16 and 17 compare the results for models using type 7 and 16 interfaces.
Fig 16: Vertical velocity of the valve’s master node.

Fig 17: Vertical acceleration of the valve’s master node.
Comparison of Meshes
Considering a contact modeled with a type 16 interface, the influence of the mesh density is
studied using two relatively coarse and fine meshes
Fine mesh: Cam: 200 external SHEL16 elements, 250 internal BRIC20 elements
Valve: 88 SHEL16 elements
Coarse mesh: Cam: 40 SHEL16 elements
Valve: 12 SHEL16 elements

Fig 18: Vertical velocity of the valve’s master node.
Fig 19: Vertical acceleration of the valve’s master node.

Although the coarser mesh amplifies the facetisation of the curved surfaced, the mesh density
does not influence the results for velocity after filtering. However, the fine mesh provides better
results for acceleration, having limited parasite oscillations for each node/surface contact.
Friction
An option in interface 7 using the Penalty method allows you to add friction to the model. Several
friction models are available. The Coulomb friction model is used here. A comparison is made
between models with and without friction.
Fig 20: Type 7 interface using Penalty and friction.

Fig 21: Vertical velocity of the valve’s master node for a type 7 interface, using the Penalty method.
Fig 22: Vertical acceleration of the valve’s master node for a type 7 interface, using the Penalty method.

Comparison of the CPU and Time Step
CPU
Simulation Time Step
(normalized)
Type 16 interface with fine 22,50 0.8365x10-7

mesh
Type 16 interface with 1 0.207x10-6

coarse mesh
Type 7 interface with 1.65 0.2133x10-6

Penalty method
(slave nodes on cam and
master surface on valve)

penalty method
(slave nodes on valve and
master surface on cam)

Lagrange multipliers
method

Lagrange multipliers
method

Penalty method and friction

Penalty method and friction
Conclusion
This example illustrated the ability of RADIOSS to model mechanisms, particularly in the case of
this contact mechanism. Interface types 16 and 7 can be used to model contact between plane
and curved surfaces. The type 16 interface enables you to simulate contact between quadratic
surfaces without using a gap and provides accurate results within a reasonable computation time.
The type 7 interface allows a frictional modeling of the contact, needing little computation time
and provides good simulation results.

Example 22 - Ditching using SPH and ALE (Mono-Domain and Multi-
Domain)
Summary
The ditching of an object into a pool of water is studied using SPH and ALE approaches. The
simulation results are compared to the experimental data and to the analytical results.
Furthermore, the study is performed using different impact velocities. The object is modeled
using a triangular section. In the first approach, a SPH model is used for water. This example
deals with the problem of an interface definition between the two parts. First, the SPH boundary
and type 7 interface are used. Moreover, the object undergoes a linear elastic law; the water
being defined by the hydrodynamic viscous fluid law 6. The results are compared with regard to
the pressure and acceleration outputs. The OUTLET boundary conditions provide appropriate
results. In the second approach, the water is modeled with an ALE mesh while the structure is
Lagrangian. The interface type 18 is used to treat the fluid-structure interactions. The results
compared to Von Karman theory, illustrate the robustness and stability of the ALE method.
Here RADIOSS Multi-Domain applied to Fluid Structure Interaction (FSI) is demonstrated, which
is easy to define Multi-Domain with Single Input Format. Highly effective computation speedup
and results are equivalence whatever H-MPP setup.

22.1 - Ditching using SPH (Mono-Domain)
Title
Ditching using SPH
(Mono-Domain)
Number
22.1
Brief Description
Impact of a simple object on water.
Keywords
 SPH modeling and hexagonal net
 Hydrodynamic viscous fluid law (/MAT/LAW6) and impact on water modeling
 Type 7 interface
RADIOSS Options
 Accelerometer (/ACCEL)
 Interface (/INTER)
 SPH outlet (/SPH/INOUT)

 Experimental data provided by Politecnico di Milano (Polytechnic University of Milan) [1]
 Analytic solution proposed by Von Karman [2]
Input File
Impact_velocity=3.5m/s:
<install_directory>/demos/hwsolvers/radioss/22_Ditching/Ditching_Mono_Domain_SPH/
v_35/*
Impact_velocity=6.8m/s:
v_68/*
Impact_velocity=11m/s:
v_110/*

Skilled
Overview

The problem consists of a simple object falling into water simulating the ditching of a helicopter.
Units: mm, ms, KN, GPa, kg.
Impact of a triangular section object on water is performed and the results are compared
qualitatively [2], also using the experimental data obtained from the Politecnico di Milano [1].
The computation is performed using several impact velocities: 3.5 m/s, 6.8 m/s and 11 m/s.
The material used for the object follows a linear elastic law (/MAT/LAW1) with the following
characteristics:
 Initial density: 7.8 x 10-6 kg.mm-3
 Young modulus: 206.9 GPa
The material law for water is a hydrodynamic viscous fluid law (/MAT/LAW6) with the following
characteristics:
Initial density: 1 x 10-6 kg.mm-3
Kinematic viscosity: 0 Pressure cutoff: -0.0001 GPa
C0: 0 GPa Pressure shift: 0 GPa
C1: 2199 GPa C4: 0 GPa
C2: 5351 GPa C5: 0 GPa
C3: 7324 GPa Initial energy/unit of volume: 0 mJ/mm3

The object is modeled using shell elements with an average mesh size of 15 x 15 mm 2.
The water is modeled using SPH particles having a hexagonal compact net with a smoothing
length "ho" equal to 28.2843 mm. Each particle of the net represents a volume equal to 16 mm 3
and weighs 16 g. This part uses 36075 SPH cells.
The size of the water block is adapted to the shape of the object for the purpose of reducing the
model’s size and the simulation’s CPU time.
 Rigid body:
The object is modeled using a rigid body, and a mass of 23.0422 kg is added to the rigid
body’s master node (ID: 287002).
An initial velocity, in accordance with the Z-axis, is set on the rigid body’s master node and
its value is set successively at 3.5 m/s, 6.8 m/s and 11 m/s.
 Gravity:
A gravity load gz = -9.81 m.s-2 is applied to the object.

 Accelerometer:
An accelerometer is set on the rigid body’s master node.
 SPH outlet:
The parallelepiped water mesh is surrounded on five faces by Outlet SPH boundary frontier
absorbing conditions. A control surface is placed at a distance equal to 2 x h o inside the
water. This surface shown in green on Fig 2, is oriented so that its normal vector points
face the interior of the domain. On this outlet surface, specific non-reflective frontiers (NRF)
are applied to the SPH cells.
 Interfaces:
One type of contact occurs in the simulation. Contact between the skin structure (shell
finite elements) and the water (SPH cells) is modeled using a sliding interface (type 7). The
gap between the surface skin and the SPH cells is equal to 3 mm. After optimization, a
scale factor on the Penalty stiffness interface equal to 0.1 is used for controlling the
interface forces between the rigid object and the water.
Fig 2: SPH outlet boundary surface non-reflective frontiers (NRF) (green).


Output Pressure
The SPH module does not enable output pressure time history values; only the animation values
may be compared to the experimental data. Figures 3 and 4, show a variation in the pressure
when the wedge ditches into the SPH water, corresponding to an impact velocity of 6.8 m/s. In
order to attain a representative pressure gradient, the pressure range has been fixed at -0.01
GPa for the minimum value and 0.08 GPa for the maximum value. In comparison to the
experimental data, the pressure values are basically the same, the wave propagation being well
described.
Fig 3: Ditching simulation (from the beginning to 3 ms).

Fig 4: Ditching simulation (from 4 ms to 9 ms).
Output Acceleration
For the object, an accelerometer is set on the master node of the corresponding rigid body. The
acceleration values expressed in g units are compared to both the experimental values [1] and
the analytic solution proposed by Von Karman [2]. The signal is filtered using a CFC 60 (-3db)
filter frequency after calculation. The filtering reduces discrepancy between the peaks.
The following diagrams indicate the time history acceleration results at the wedge object’s rigid
body’s master node for the three cases of impact speed: 3.5, 6.8 and 11 m/s.

Fig 5: Deceleration of the wedge for an impact velocity of 3.5 m/s.

Fig 6: Deceleration of the wedge for an impact velocity of 6.8 m/s.

Fig 7: Deceleration of the wedge for an impact velocity of 11 m/s.
For these three cases, the SPH approach using the OUTLET SPH boundary conditions indicates a
good deceleration. For an impact velocity nearing the 8 m/s of the Helicopter ditching
configuration, the deceleration is in correlation with the experimental data [1] and also with the
analytic solution proposed by Von Karman [2].
Conclusion
The simulations show that the SPH approach using the OUTLET option, allows the ditching of
simple objects to be modeled without any numerical problems.
The SPH and OUTLET results are very close to the experimental test results and also to the
analytical solution. In conclusion, to achieve ditching simulations with the correct results, it is
necessary to model the water block using the SPH method with the OUTLET boundary conditions.
References
[1] CAST Deliverable 5.5.1 Generic Water Impact Tests performed at Politecnico di Milano
(Polytechnic University of Milan)
[2] Olivier Pastore Study and modelization of rigid bodies impact during sea landing phase;
Annex 1 Von Karman's Theoretical Models, T. Miloh et al. May.

22.2 - Ditching using ALE (Mono-Domain)
Title
Ditching using ALE (Mono-
Doman)
Number
22.2
Brief Description
Impact of a simple object on water simulated by ALE approach.
Keywords
 ALE modeling
 Type 18 interface
RADIOSS Options
 /MAT/ELAST
 /ALE/GRID/DONEA
 /UPWIND
 /MAT/BIPHAS
 /ALE/MAT
 /INTER/TYPE18

 Experimental data provided by Politecnico di Milano (Polytechnic University of Milan) [1]
 Analytic solution proposed by Von Karman [2]
Input File
<install_directory>/demos/hwsolvers/radioss/22_Ditching/Ditching_Mono_Domain_ALE/*

Skilled

Overview

Units: mm, ms, KN, GPa, kg.
Impact of a triangular object on water is performed and the results are compared qualitatively
[2], also using the experimental data obtained from the Politechnico di Milano [1].
The computation is performed using several impact velocities of 3.5 and 11 m/s.
The material used for the object follows a linear elastic law (/MAT/LAW1) with the following
characteristics:
 Initial density: 7.8 x 10-6 kg.mm-3
 Young modulus: 206.9 GPa
The material law for water is BIPHAS law (/MAT/LAW37) with the following characteristics:

The object is modeled using shell elements with an average mesh size of 15 x 15 mm 2. In order
to simplify the computation, it is put in a rigid body with an accelerometer on the master node.
The water is modeled using 15x15x15 mm mesh with a total of 166023 elements. The material
law BIPHAS (/MAT/LAW37) is used.
The air is modeled using a BIPHAS material with the following characteristics:
#blank
# RHO_L CL ALPHA_L NU_L LAMBDA_ON_RHO_L

1e-6 2.089 0 .00089 0
# RHO_G GAMMA P0 NU_G LAMBDA_ON_RHO_G
1.22E-9 1.4 .1e-3 .014607 0
The boundary conditions are applied to the pool as following:

 Z displacement component fixed for lower and upper faces
 Y displacement component fixed for lateral faces normal to Y
 X displacement component fixed for lateral faces normal to X
An interface type 18 is defined to manage the contact between the solid in Lagrangian mesh
(Prism) and the fluid (pool). The diedra is defined as master and the nodes in the pool (air and
water) as slave.
The interface type 18 forces are computed by Penalty method. The interface stiffness is
proportional to impact velocity. The results obtained by the ALE approach can be highly
dependent to the interface stiffness factor Stfac, which should be adjusted in function of size of
element and fluid properties.
The interface type 18 forces are computed by Penalty method. The forces are proportional to the
stiffness factor Stfac which should be calibrated. The following graph shows the dependence of
results (acceleration at the accelerometer) to the mesh and Stfac.

Fig 2: Higher peak force with coarse mesh and interface stiffness dependence
Higher peak forces are obtained with the coarse mesh. That can be partially corrected by
filtering, as shown in Fig 3.

Fig 3: Filtering of results for coarse and fine meshes
Using a filter CFC 60, -3 dB, the simulation results of the ALE and SPH approaches are compared
to Von Karman theoretical solution and experimental measures in Fig 4.
Fig 4: Comparison between simulation results, theoretical solution and experimental measures (acceleration)

SPH and ALE approaches respectively give the maximum acceleration of 83g and 84g. However,
the Von Karman theory delivers 82g; with the maximum value by experience is between 83g and
73g.
On the other hand, the duration for acceleration beyond 40g is respectively 7.9ms and 8.2ms for
SPH and ALE simulation methods, where the experience provides values between 7.5ms and
8.5ms and the Von Karman theory provides 8 ms.
Note that:
 The computation time is similar for both SPH and ALE approaches.
 Using a fine enough mesh, both RADIOSS methods SPH and ALE provide satisfactory
results, compared to experience and analytical solution.
 The ALE approach is more robust and stable when the SPH signal is highly vibrated.
 The use of a coarse mesh in the ALE approach requires the interfaced stiffness calibration.
References
[1] CAST Deliverable 5.5.1 Generic Water Impact Tests performed at Politecnico di Milano
(Polytechnic University of Milan).
[2] Olivier Pastore Study and modelization of rigid bodies impact during sea landing phase;
Annex 1 Von Karman's Theoretical Models, T. Miloh et al. May.

22.3 - Ditching using Multi-Domain for SPH and ALE
Title
Ditching using Multi-Domain
for SPH and ALE
Number
22.3
Brief Description
Impact of a simple object on water using Multi-Domain to approach SPH and ALE.
RADIOSS Options
 Multi-Domain (/SUBDOMAIN)
 Rad2rad (/RAD2RAD/ON) in Engine
Compared
 Results of Mono-Domain and Multi-Domain
Input File
SPH (Multi-Domain):
<install_directory>/demos/hwsolvers/radioss/22_Ditching/Ditching_Multi_Domain_SPH/*
ALE (Multi-Domain):
<install_directory>/demos/hwsolvers/radioss/22_Ditching/Ditching_Multi_Domain_ALE/*

Skilled
Overview

Unit: mm, ms, KN, GPa, kg.
Use Example 22.1 as SPH ditching and Example 22.2 as ALE ditching, then add Multi-Domain
approach in SPH ditching and ALE ditching

Modeling Methodology of Multi-Domain (Single Input Format) approach

For Multi-Domain approach, take the model of Mono-Domain and:
1. Setup /SUBDOMAIN in Starter:
Use /SUBDOMAIN to define sub-domain. In this example part ID 19 is defined as sub-domain.
For SPH ditching:
For ALE ditching:
Note: The subdomain_title “SPH_SubDomain” (or “ALE_SubDomain”) must be the same as sub-
domain Engine name SPH_SubDomain_0001.rad (or ALE_SubDomain_0001.rad).
Fig 1: Domain description in Multi-Domain approach
2. Setup in two Engine files:

First Engine file (main-domain):
 Take the Engine of Mono-Domain and insert /RAD2RAD/ON keyword.
 Free element time step control for ALE or free nodal time step control for SPH in the first
Engine file.

Second Engine file (sub-domain)
 Take the Engine of Mono-Domain and insert /RAD2RAD/ON keyword.
 Update the Engine name and also the run name in /RUN. They must match the
subdomain_title in Starter /SUBDOAMIN.
 Set the Lagrange parts (part ID 19) specific time step control
Fig 2: Multi-Domain setup in Starter and Engine (SPH ditching as example)
Contact between main-domain and sub-domain

As of RADIOSS version 14.0, Starter automatically creates coupling links between domains with
Multi-Domain Single Input Format. In SPH ditching using /INTER/TYPE7 coupling links between
Lagrange sub-domain and SPH main-domain are automatically created. In ALE, ditching using
/INTER/TYPE18 coupling links between Lagrange sub-domain and ALE main-domain are
automatically created.
Results between Mono-Domain and Multi-Domain

To compare the results between Mono-Domain and Multi-Domain we take the model with 11m/s
in initial velocity in SPH ditching and in ALE ditching.

It shows same results for Mono-Domain and Multi-Domain.
 Energy
SPH ditching ALE ditching
Fig 3: Energy comparison between Mono-Domain and Multi-Domain
 Acceleration
Fig 4: Acceleration comparison between Mono-Domain and Multi-Domain

The time step in Multi-Domain is much bigger than in Mono-Domain.
 Timestep
Fig 5: Time step comparison between Mono-Domain and Multi-Domain
Theoretical speedup a formula:
Where, Ne is the number of elements, Nc is the number of cycles, and C is the the element
average time cost/cycle.
Speedup a could also be formulated with following factors:
 Relative element number:
 Relative time step (relative cycle numbers):
with dt - time step
 Relative cost per cycle:
Lagrange shell= 1
Lagrange sold ~ 3
ALE solid ~ 6
SPH cell ~ 15

Then speedup a will be:
Elapsed times:
SPH ditching
ALE ditching
After comparing the elapsed time, observe Multi-Domain FSI speedups ranging from 6 times to 9
times of the Mono-Domain.
Conclusion
 The Multi-Domain is particularly adapted to FSI simulations involving a large amount of

either ALE elements or SPH cells which the time steps are significantly higher than the one
in the much smaller in terms of elements of the Lagrange structure
 The easiness of setup makes the Multi-Domain applied to FSI much simpler than when
applied to purely Lagrange models where the cross-domain contacts need to be minimized
as long as this aspect is not automated
 Due to this setup easiness the Multi-Domain results in FSI give less possibility to deviations
when compared to classical mono-domain results
 FSI simulations often involving large ALE or SPH domains with high time steps the Multi-
Domain speedups are significantly higher than in pure Lagrange simulations like in Crash

Comments and Hints
 For Multi-Domain FSI there is no need to minimize the contacts in TYPE7 interface (SPH) or
in TYPE18 or TYPE22 interfaces (ALE) between Fluid and Structure whereas it is needed in
pure Lagrange Multi-Domain.
 For Multi-Domain FSI it is advised to use for the main-domain a free element time step with
a scaling factor of 0.5 for ALE and a free nodal time step with a scaling factor of 0.6.
 When Lagrange parts are assembled using tied contact interface TYPE2. Then poor
performances may occur with SPOTflag = 0 or 1.
o Set SPOTflag = 26 to switch from kinematic to penalty formulation to solve the

problem. In such cases (SPOTflag = 26) contacts become permissive depending on
assigned contact stiffness. It is also recommended to check the contacts quality in
animation results to make sure no significant penetrations occur.

Example 23 - Brake
Summary
A frictional mechanism is studied, which consists of a brake system, defined by a disk pinched
between two pads. The main aspects of the model are the initial rotary motion of the disk and
the interface definition, between the disk and the pads. Carefully watch the accuracy of the
simulation results compared to the analytical solution.

Title
Brake
Number
23.1
Brief Description
A brake system is simulated using a finite Lagrangian mesh element.
Keywords
 Brick elements and HEPH formulation
 Type 7 interface and friction
RADIOSS Options
 Type 7 interface (/INTER/TYPE7)
 Skew frame (/SKEW)
Input File
Lagrangian formulation:
<install_directory>/demos/hwsolvers/radioss/23_Brake/Lagrangian_formulation/
BRAKE2*

Advanced

Overview
Aim of the Problem

The purpose of this example is to highlight the capacity of RADIOSS to simulate frictional
mechanisms. The braking system retained consists of a disk pinched in between two pads.
A disk with a hole in the center rotates at 0 = 120 rad/s around its axis. It is subjected to
frictional contact using two small brake pads, placed on two faces.
Description of the disk: Radius = 100 mm
Width = 50 mm
Thickness = 5 mm
Mass = 1 kg
Inertia = 0.57x10-2 kg/m2 (about its free rotation axis).
Description of the pads: Length = 65 mm
Width = 28 mm
Thickness = 5 mm
A constant P = 300N pressure is applied on the back of each pad to push them against the disk.
A Coulomb friction coefficient is assumed as being 0.15.
Units: m, s, kg

The material used for the disk follows an isotropic elasto-plastic law (/MAT/LAW2) using the
Johnson-Cook plasticity model, with the following characteristics:
The material used for the pads follows a linear elastic law, with the following characteristics:
The two parts are modeled using a regular mesh having 720 BRICK elements for the disk and 80
such elements for the pads. The HEPH formulation is used to describe the BRICK elements.
Two steps are necessary to compute the model: First, an initial velocity 0.03ms is applied to the
disk. In the second step, pressure is applied to the pads to push them onto the disk.
 Rigid bodies:
Two rigid bodies are created to put the disk into motion: the first (called RBODY1) contains
all the nodes of the disk, except those in the disk’s internal periphery, which are contained
in the second rigid body (called RBODY2). Both rigid bodies are activated in the first step of
computation; however, RBODY1 is deactivated in the D02 file.
Two other rigid bodies are created to model the pads’ faces where concentrated loads are
applied.

Fig 2: Rigid bodies on the disk (RBODY1 on the left and RBODY2 on the right).
Fig 3: Rigid body on a pad.
For the disk’s rigid bodies, all DOF, except the rotation around Y are fixed. For the pads’
rigid bodies, all DOF; except translation around Y are fixed.
 Load:
Two concentrated opposite forces are applied to the rigid bodies’ master nodes for the
pads.
An initial rotational velocity 0 = 120 rad/s is applied to the disk’s master nodes during the
first computation phase.

Angular Velocity of the Disk

The normal contact force between the pads and the disk is FN = 600N. Then the tangential
friction force on the surface of the disk is obtained at FT = 0.15 x FN = 90N. The torque around
the axis of the disk is T = r x FT = 7.1 Nm, with r = 0.0789 m, which corresponds to the
orthogonal projection on a radial axis with regard to the distance between the center of the disk
and the point of the pad where the load is applied. This leads to an angular deceleration of a = T /
IR = 1246 rad/s2.
The necessary time to stop the disk can be computed as: t = 0 / a = 0.096 second.
The simulation by RADIOSS using the explicit solver allows similar results to be obtained, as
shown in the following diagrams. The following graph shows the time history for angular velocity.
The disk stops at t = 0.095 s, which corresponds to the analytical solution.
Fig 4: Angular velocity of the disk.

Rotation of the Disk
The following graph shows the total rotation of the disk, which rotates 5.65 rad before stopping.
Fig 5: Total rotation of the disk.

Reaction Forces
Fig 6: Reaction forces.
The reaction forces value in Fig 6 is about 90 N, which corresponds to the analytical value.
Energies
The total energy remains constant during computation. After braking, the kinetics energy
decreases smoothly while the contact energy increases. There is no hourglass energy as a HEPH
solid element is used.
Fig 7: Energies.

Contact Forces
The following diagram presents the tangential contact forces for three consecutive moments.
Fig 8: Variation of X-component of tangential contact forces.
Conclusion
The accuracy of the results obtained, using the simulation and corresponding to the analytical
solution, proves that RADIOSS is able to simulate mechanisms, such as braking systems.

Example 24 - Laminating
Summary
The lamination of a metal strip under two rolling cylinders is studied. Several formulations are
compared. Large and small strain assumptions are respectively used. The influence of the
number of elements concerning the thickness of the metal strip, as well as element formulation is
discussed.

Title
Laminating
Number
24.1
Brief Description
Two rolling rigid cylinders squeeze a plate to laminate it.
Keywords
 Brick element, solid formulation, co-rotational formulation, and fully-integrated element
 Constant pressure formulation and plasticity options
 Large deformation/Small strain
RADIOSS Options
 Constant time step (/DT/BRICK/CST)
 Elasto-plastic material law (/MAT/PLAS_JOHNS)
 General solid property (/PROP/SOLID)
Input File
Thickness: 2 elements:
<install_directory>/demos/hwsolvers/radioss/24_Laminating/Thickness/2_elements/
ROLLING*
5 elements: //.../radioss/24_Laminating/Thickness/5_elements/ROLLING*
Formulation: Isolid=12:
<install_directory>/demos/hwsolvers/radioss/24_Laminating/Formulation/Isolid12/
ROLLING*
Icpre=0: //.../radioss/24_Laminating/Formulation/Icpre0/ROLLING*
Icpre=1: //.../radioss/24_Laminating/Formulation/Icpre1/ROLLING*
Temperature: T=800°C:
<install_directory>/demos/hwsolvers/radioss/24_Laminating/temperature/T=800/
ROLLING*
T=1200°C: //.../radioss/24_Laminating/temperature/T=1200/ROLLING*

Advanced

Overview

This analysis shows a phase of a rail rolling. A metal strip is successively passed through two
rollers aiming at reducing its thickness. Both rollers have a constant angular velocity of 6.85
rad/s, and the metal strip is dragged along a moving machine bed. This process may be
considered quasi-static and involves high deformation (mainly compression).
Units: mm, s, Mg, N, MPa
Both rollers and the metal strip are made of mild steel. They exhibit an isotropic elasto-plastic
behavior which can be modeled using the Johnson-Cook law (/MAT/LAW2):
 Initial density: 7.8 x 109 Mg/mm3
Note: Any temperature influence will not be taken into account. The problem is studied using the
following parameters:
 Temperature exponent: 1
 Melting temperature: 2073 K (around 1800°C)
 Specific heat at constant pressure Cp: 460 J/Kg.K
 Geometry: The metal strip has a cross-section of 80 x 20 mm and the rollers have a radius
of 100 mm. After the passage of the first roller, the thickness is reduced by 7 mm, then by
another 5 mm after the second roller.

It is not necessary to pass many elements over the width of the metal strip, but rather to obtain
an accurate stress distribution over its thickness by passing a minimum of five elements over the
thickness of the metal strip. Depending on what is being looked for, passing five elements over
the thickness may seem like a lot. This issue should be discussed in the early part of the analysis.
Concerning the rollers as the elements is of first order, as it is not easy to perfectly model the
curvature. The mesh must be fine enough to estimate the curvature with as much accuracy as
possible, one element over the width being sufficient.
Some details are made: The moving machine bed is not modeled and all lower nodes of the
metal strip are constrained in the Z-direction. Moreover, an initial velocity is applied to the metal
strip to initiate contact with the first roller. Assuming there is a Coulomb friction between the
metal strip and the roller using a friction coefficient of 0.3, the metal strip is then dragged by the
roller.
Assuming the rollers are rigid, a constant angular velocity to the master nodes is applied.
As this process may be considered as a quasi-static one, the density is artificially increased by a
factor of 10000, that the density used is now 7.8 x 10 -5. Consequently, the time step will be
higher, and assuming the kinetic energy is negligible compared to the energy of deformation,
results will be correct.
Note: Density can be increased as long as dynamic effects are negligible.

The lower nodes of the metal strip are constrained in Z (to represent the moving machine
bed).
The master nodes of the rollers are constrained in all directions, except rotation around the
X-axis.
 Imposed velocity:
A constant angular velocity around the X-axis is applied to both of the rollers’ master
nodes.
An initial velocity of 500 mm/s in the X-direction is applied to every node of the metal strip.
 Interfaces:
Contact between the metal strip and the rollers are modeled by a type 7 interface. The
master surface is defined by the external surface of the rollers, and the slave nodes by the
metal strip (/GRNOD/PART). As there is only one element over the width of the rollers. The
previous interface does not need to be symmetrically arranged. The gap is chosen
arbitrarily at 1 mm.

Simulation Results
Number of Elements over the Thickness

It was mentioned earlier that passing five elements over the thickness is a minimum; but how
was that number reached? It is not an easy question, as this number depends on what is being
looked for. If interested in the stress distribution over the thickness, the higher the better; thus
the choice would be a trade-off between accuracy and the CPU time. Compare the von Mises
stress and the plastic strain for two elements over the thickness using one integration point
(Isolid =1) and five elements with 8 integration points (I solid =12).
Table 1: Comparison between two and five elements.
Passing two elements over the thickness is not enough to see the stress (or strain) distribution;
five elements is enough though. If the deformed shape is not smooth and/or the gradient
between the two elements is too high, consider refining the mesh; however, this can be
somewhat costly! Additionally, it takes 12 times longer to run the model with five elements over
the width.
What if the only interest is in the reaction force acting on the cylinder?

Fig 2: Reaction forces acting on the cylinder.
The above graph indicates the reaction force on the first cylinder using two or five elements over
the thickness. Both curves are almost identical and it takes much longer to use five elements.
Thus, to save CPU time, there is no need to use more than two elements.
Influence of the Small Strain Formulation
Usually for problems involving large deformations, a large strain formulation would be used. In
RADIOSS this is the default setting, but it is also possible to use a small strain formulation. This
formulation is not very accurate for large deformations, but it is more robust and enables the
time step to not decrease too much. Indeed, large deformation/rotation problems may lead to
mesh distortion which causes the time step to drop drastically; computation may even stop due
to a negative volume. The small strain formulation overcomes all this by assuming a constant
volume, consequently the time step becomes constant, and even if the mesh is completely
distorted, computation will not be stopped due to the negative volume.
This formulation can be applied from t=0 by setting the flag I smstr to 1, directly in the type of a
specific part. It is also possible to switch from a large strain formulation to a small strain
formulation during the simulation in order to prevent a negative volume and/or to maintain a
decent time step using the /DT/BRICK/CST option in the Engine file (*_0001.rad) having a
critical time step.

In this example the results between a full large strain formulation, a full small strain formulation
and a "mixed formulation” using /DT/BRICK/CST are compared. The average time step is 20%
higher for a full small strain formulation and 3.5% higher for a "mixed formulation” in comparison
to the default large strain formulation. Thus, there is a significant gain in terms of CPU time using
the small strain formulation. The deformed shape is not good (Table 2).
Table 2: Deformed shape.
Moreover, looking at the plastic strain, using a small strain formulation from t=0 leads to major
errors (Table 3). First of all, the strain distribution is not well determined and most importantly,
the maximum is far too low, which means permanent deformation was under-estimated.

Table 3: Deformed shape.
In such a case, it may be of interest to use the small strain formulation but only for a few
elements reaching a critical time step (using /DT/NODA/CST); as the time step will not stop, due
to a distorted element. However, for accuracy reasons, the number of elements switching to a
small strain formulation should be checked, the lower the better.

The HA8 Formulation
The improved solid formulation HA8 overcomes the drawbacks of the standard 8 integration
points’ formulation (Isolid=12 or 14). In particular, in the case of a pure bend, "shear-locking”,
which makes the standard formulation rather stiffer, does not exist. It is also possible to use the
small strain formulation, which contrary to the 8 integration points’ formulation is not compatible.
It is now possible to use up to 9 integration points for each direction.
Depending on the value given for the I cpre flag, the HA8 formulation may use a reduced pressure
integration. In this part the influence of this flag on simulation will be investigated in order to find
out the most suitable value. Table 4 shows both the deformed shape and the equivalent plastic
strain for the different solid formulations. The new HA8 formulation with reduced pressure
integration behaves the same way as the standard 8 integration points formulation. If the
reduced pressure integration is not activated, the results will not be correct, with the metal strip
deforming badly and the plastic strain being over-estimated (Table 4). The reaction force acting
on the cylinder will also be over-estimated (Table 3), which means that the structure will be
stiffer without the reduced pressure integration.
Fig 3: Rolling Force for different formulations

Table 4: Plastic Strain for different formulations.
The HA8 formulation must always be used with reduced pressure integration, the only time when
this option must be deactivated is in the case of emulating a thick shell formulation with 8-nodes
bricks.
Temperature Influence
When metal forming, one of the main parameters engineers' study whether the force is needed
during the process. In this particular case, it concerns the rolling force applied by the cylinders.

Engineers try to minimize this force in order to use less power and to reduce maintenance on the
cylinders. One way to do so is to form metal at a higher temperature. RADIOSS takes into
account the temperature dependency using the same law already used (Johnson-Cook law). The
following diagram indicates the rolling force in accordance with the temperature and, as expected
the higher the temperature, the lower the force. Consequently, metal forming is a trade-off
between the power saved using a lower force and the power required to raise the temperature.
Fig 4: Rolling Force.
Conclusion
The squeezing of the metal strip below two rolling cylinders is simulated by RADIOSS. The large
deformation formulation, when a sufficient number of elements are used, obtaining physically-
acceptable results is allowed. The small strain option leads to bad results, but with low cost. The
element formulation and the number of integration points through thickness are other
parameters influencing results; the higher the precision, the higher the cost. On the other hand,
as the problem is considered to be quasi-static, resolution using the RADIOSS implicit solver can
be envisaged.

Example 25 - Spring-back
Summary
The spring-back simulation of sheet metal bent into a hat-shape is studied. The problem is one of
the famous tests from the Numisheet’93. As spring-back is generally a quasi-static unloading, the
use of the RADIOSS implicit solver is justified. The RADIOSS explicit solver is also used to
compare the methods’ efficiency. However, for the stamping phase only the explicit solver is
used, as the forming process is highly dynamic.
The example illustrates how to link up the explicit computations. It highlights the efficiency of the
implicit solver for the spring-back simulation.

Title
Spring-back
Number
25.1
Brief Description
An explicit stamping simulation is followed by a spring-back analysis using implicit or explicit
solvers for stress relaxation. Results are compared with a reference.
Keywords
 Explicit stamping simulation, implicit/explicit spring-back simulation, and stress relaxation
 Implicit strategy and time step control by arc-length method
 Anisotropic elasto-plastic material law (/MAT/LAW43) and Hill model
 Orthotropic shell formulation, QEPH, progressive plastification, and iterative plasticity
 Type 7 Interface, Penalty method, and friction
RADIOSS Options
 Implicit parameters (/IMPL)
 Implicit spring-back (/IMPL/SPRBACK)

 Experimental data
Input File
Explicit spring-back: <install_directory>/demos/hwsolvers/radioss/25_Spring-back/
Explicit_spring-back/DBEND_44*
Implicit spring-back: <install_directory>/demos/hwsolvers/radioss/25_Spring-back/
Implicit_spring-back/DBEND_44*

Advanced

Overview
Aim of the Problem

This example deals with the numerical simulation of a stamping process, including the spring-
back. This refers to one of the sheet metal stamping tests "2D Draw Bending" indicated in
Numisheet’93. The final shape of the formed sheet metal, after releasing all constraints on the
blank sheet is studied. During the spring-back simulation, an explicit-to-implicit sequential
solution method is used, where a dynamic forming process using the explicit solver is used first,
followed by an implicit modeling of the spring-back deformations by statically removing the
stamping stress.
Thus, two analyses are performed in order to compare the performances and the accuracy of
results:
 Explicit stamping and implicit spring-back simulations
 Explicit stamping and explicit spring-back simulations (using dynamic relaxation)
The spring-back simulation of the forming sheet metal uses an elasto-plastic nonlinear approach.
The implicit input options and the incremental strategy used are described in the modeling
section.
A numerical simulation of stamping is performed up to 960 ms. Spring-back computation is
carried out from 960 ms to 1000 ms for implicit (static approach) and to 6000 ms for explicit
(quasi-static approach).
A standard stamping operation is studied. The stamping tools include a punch, a die and a blank
holder.
A load F of 1225 N is vertically applied on the blank holder in order to flatten the sheet metal
against the die. The load is removed before spring-back simulation.
The sheet metal stamping operation is managed using a variable imposed velocity applied on the
punch with a maximum set to 0.1 ms-1. The tools are withdrawn after the stamping phase in
order to enable the spring-back to be observed.
Fig 1: Description of the problem.

The main geometrical dimensions of parts are:
 Radius of die’s corners: 5 mm
 Radius of punch’s corners: 5 mm
 Width of punch: 50.4 mm
 Sheet metal dimensions: 35 mm x 175 mm
The thickness of the sheet metal is defined at 0.74 mm. The Coulomb friction coefficient between
the sheet metal and the die is defined at 0.129.
The stamping tools’ material undergoes a linear law using the following properties:
 Initial density: 8x10-3 g/mm3
The material of the sheet metal under the roller has distinct characteristics of anisotropy. Its
anisotropic elasto-plastic behavior can be reproduced by a Hill model (/MAT/LAW43). This law
can be considered as a generalization of the von Mises yield criteria for anisotropic yield behavior.
The yield stress is defined according to a user function and the yield stress is compared to
equivalent stress:
The Ai coefficients are determined using Lankford’s anisotropy parameters range. Angles for
Lankford parameters are defined according to orthotropic direction 1.
A hardening coefficient is used to describe the hardening model as full isotropic (value set to 0)
or based on the Prager-Ziegler kinematic model (value set to 1). Hardening can be interpolated
between the two models, if the coefficient value is between 0 and 1.
The material parameters are:
 Initial density: 8x10-3 g/mm3
 Lankford 0 degrees: r00= 1.73
 Lankford 45 degrees: r45 = 1.34
 Lankford 90 degrees: r90= 2.24
The yield curve used is shown in the diagram below. Failure is not taken into account.

Fig 2: User’s yield function.
Taking symmetry into account, only a quarter of the structure is modeled. The symmetry plane is
along axis y = 17.5 mm and x = 0 mm.
Fig 3: Finite mesh elements of the problem studied.

The punch is shown in purple, the blank holder in green and the die in red. The sheet metal
(blue) is modeled using 4-node shell elements.
The sheet metal is discretized by a non-regular mesh and a fine mesh is used for parts to be
plastically deformed. The smallest size of the shell element is 1.5 mm.
Fig 4: Progressive mesh of the sheet metal.
In order to achieve accurate simulation results, the QEPH shell element formulation is used in
explicit and implicit analyses. A Lagrangian formulation is adopted.
In accordance with the elasto-plastic Hill model for the material law, the sheet metal is described
by the shell elements using the orthotropic property (Type 9). The shell characteristics are:
 Five integration points (progressive plastification)
 Interactive plasticity with three Newton iterations (I plas = 1)
 Thickness changes are taken into account in stress computation (I thick = 1)
 Initial thickness is uniform, equal to 0.74 mm

 Orthotropy angle: 0 degree
 Reference vector: (1 0 0)
The input components of the reference vector are used to define direction 1 of the local
coordinate system of orthotropy. The orthotropy angle, in degrees defines the angle between
direction 1 of the orthotropy and the projection of the vector on the shell.
Three type 7 interfaces using the Penalty method are employed to model contacts between the
stamping tools and the sheet metal. The parameters defining the contact are:
 Coulomb friction: 0.129
 Gap: 0.37
 Critical damping coefficient on interface stiffness: 1
 Critical damping coefficient on interface friction: 1 (default)

Fig 5: Contact modeling using a type 7 interface considered with the Penalty method (master/slave sides).
In the implicit approach, the contact using the Penalty method with fictional springs is stored in a
separate stiffness matrix to the main one. Therefore, supplementary memory is needed and
information of the second contact stiffness will be printed when contact is active.
Critical damping coefficients (inputs) description:
 The normal force computation is indicated by:
Where,
K0 is the initial interface spring stiffness
VISCS is the critical damping coefficient on interface stiffness (default value: 0.05)
 The tangential force computation is indicated by:

Ft = min(Fric * Fn, Fad )

where:
Fad = Ct Vt is the adhesion force
VISf is the critical damping coefficient on interface friction (default value: 1)
For spring-back computation by implicit, the removing of the stamping tools is taken into account
by deleting all interfaces using the input option in the second *_0002.rad Engine file as follows:
/DEL/INTER
1 2 3  Interfaces ID 1, 2 and 3 are deleted.

Simulation deals with:
1. Stamping simulation by explicit: from the beginning up to 960 ms.
2. Spring-back simulation:
 using explicit (dynamic approach): from 960 ms to 6000 ms:
- From 960 ms to 2000 ms: Stamping tools are slowly withdrawn because the quasi-static
analysis requires dynamic effects to be minimized during spring-back. Thus, the
interfaces are not deleted. Options are defined in the *_0002.rad Engine file.
- From 2000 ms to 6000 ms: A dynamic relaxation (/DYREL) is activated in the
*_0003.rad Engine file in order to converge towards quasi-static equilibrium.
 using implicit (static approach): from 960 ms to 1000 ms:
- The input implicit options are added in the *_0002.rad Engine file. Stress relaxation is
activated using the /IMPL/SPRBACK keyword. All interfaces are deleted and specific
boundary conditions are added on the stamping tools. Tools are not withdrawn.
In the simulation, the tools are modeled using rigid bodies (/RBODY) as shown in Fig 6.

Fig 6: Modeling of the stamping tools as rigid elements.
An automatic master node is chosen. The center of gravity is computed using the master and
slave node coordinates and the master node is moved to the center of gravity where is placed
mass and inertia (ICoG is set to 1). No mass or inertia are added to the rigid bodies.
A quarter of the structure is modeled in order to limit the model size and to eliminate rigid body
modes for implicit computation. Symmetry planes are defined along the y axis = 0.
Fig 7: Boundary conditions (/BCS) on the sheet metal according to the symmetries.
The nodes on the longitudinal plane are fixed in the Y translation and X, Z rotations.
For the other symmetry plane, the nodes are fixed in the X translation and Y, Z rotations.

Stamping tools are restricted to moving only along the Z-axis. The boundary conditions are
applied on the master nodes of the rigid bodies, including the parts (Fig 7).
For the numerical simulation of the implicit spring-back, additional conditions must be added in
the *_0002.rad Engine file in order to remove the rigid body modes that are not permitted in the
implicit approach. The stamping tools are fully fixed (X, Y, Z translations and X, Y, Z rotations).
The translation of the ID 427 node is fixed along the Z-axis allowing the sheet metal to move
towards the final shape without rigid body mode.
Fig 8: Added boundary conditions on the 427 node for implicit spring-back.
Imposed velocities are applied on the stamping tools via the master nodes of the rigid bodies.
The velocity of the punch is controlled by a specific input curve, as shown in Figures 9 and 10.
During implicit spring-back, all velocities are set to zero. Explicit spring-back computation up to
6000 ms necessitates imposed velocities on tools in order to withdraw them as of 1000 ms.

Fig 9: Imposed velocity on punch via the rigid body’s master node.
Fig 10: Imposed velocity on die and blank holder via the rigid bodies’ master node.
" Punch part …………… Curve 1, scale factor set to -1.

" Die part ………………. Curve 2, scale factor set to 1.
" Blank holder part ……. Curve 2, scale factor set to -1.
The stamping is performed by explicit simulation up to 960 ms using Curve 1. The implicit
simulation is carried out only for the spring-back stage from 960 ms to 1000 ms. Curve 2,
therefore, is only defined for explicit spring-back simulation.

Fig 11: Imposed velocities on tools in two phases: stamping then tools removing.
Considering the symmetries, a constant concentrated load of 612.5 N is vertically applied on the
blank holder via the master node of the rigid body. The load is set to zero from 960 ms before
studying the spring-back.
Implicit spring-back analysis is launched using /IMPL/SPRBACK.
Tolerance: 0.025
Time step control method: Norm displacement (arc-length)
Initial time step: 0.08 ms
Minimum time step: 10-5 ms
Maximum time step: no

Desired convergence iteration 6
number:
Maximum convergence iteration 20
number:

Maximum increasing time step
0.0
scale factor:
Spring-back option: Activated
A solver method is required to resolve Ax=b in each iteration of the nonlinear cycle. It is defined
using /IMPL/SOLVER.
Linear solver: Direct solver MUMPS
Precondition methods: Factored approximate inverse
Maximum iterations number: System dimension (NDOF)
Stop criteria: Relative residual on force
The input implicit options added in the *_0002.rad Engine file are:
/IMPL/PRINT/NONL/-1 Printout frequency for nonlinear iteration
/IMPL/NONLIN/1 Static nonlinear computation

2 0 0.25e-1

5 0 3 0.0
/IMPL/DTINI Initial time step determines initial loading increment

0.08
/IMPL/DT/STOP Min-Max values for time step

0.1e-4 0.0
/IMPL/DT/2 Time step control method 2 - Arc-length+Line-search will be used

6 .0 20 0.67 0.0 to accelerate and control convergence
/IMPL/SPRBACK Spring-back computation (stress relaxation)
Refer to RADIOSS Starter Input for more details about implicit options.
Explicit spring-back analysis uses the dynamic relaxation in the *_0003.rad Engine file from
2000 ms.
the simulation of dynamic loading. However, a quasi-static simulation via a dynamic resolution
method is needed to minimize the dynamic effects for converging towards static equilibrium, the
final shape achieved after spring-back.

The dynamic effect is damped by introducing a diagonal damping matrix proportional to mass
matrix in the dynamic equation.
where, is the relaxation value which has a recommended default value 1, and T is the period
to be damped (less than or equal to the highest period of the system).
The inputs of the relaxation dynamics are:
 Relaxation factor: 1
 Period to be damped: 1000 ms
This option is activated using the /DYREL keyword (inputs: and T).
In the metal stamping operation, the highly nonlinear deformation processes tend to generate a
large amount of elastic strain energy in the metal material in addition to some of the plastic
deformed areas. The internal energy, which is stored in the sheet metal during stamping, is
subsequently released once the stamping pressure has been removed. This energy released is
the driving force of the spring-back in the sheet metal forming process. Therefore, the spring-
back deformation for sheet metal forming is mainly due to the amount of elastic energy stored in
the part while it is being plastically deformed.
The material density has been multiplied by 10,000 to obtain a reasonable computation time
using explicit simulations. An additional time period is also required for slowly withdrawing the
tools, prior to the explicit spring-back simulation in order to achieve a good result. Thus, explicit
stamping takes longer than stamping followed by implicit spring-back computation.
Figure 12 shows the deformed configurations using implicit simulation. The symmetrical part is
added.

Fig 12: Deformed sheet metal before and after spring-back (implicit spring-back).
Stamping is performed from the beginning up to 960 ms. The final shape after the spring-back
process is achieved after 1000 ms using the implicit solver and after 6000 ms using the explicit
solver.

Fig 13: Deformed mesh of the sheet metal before and after the spring-back (multi-models mode).
The animations in Fig 14 include the results of the spring-back during simulation. There is an
increasing number of stresses in the sheet metal from the start up to 960 ms, after which, the
stresses begin to decrease as a result of the spring-back (stress relaxation).
Fig 14: Stamping results on the sheet metal before and after spring-back.

Figure 15 shows the internal energy stored in the sheet metal during the stamping.
Fig 15: Internal energy in the sheet metal part (explicit spring-back simulation).
The dynamic relaxation used in the explicit spring-back computation enables to improve
convergence towards quasi-static solution. The variation of the kinetics energy on the sheet
metal in the explicit spring-back simulation is depicted in Fig 16 (from 960 ms up to 6000 ms):

Fig 16: Convergence towards quasi-static equilibrium (explicit spring-back simulation).
Comparison with experimental data on geometry after spring-back is shown in Table 1.

Table 1: Simulation results compared to experiment
Experiment
105.7 77.7
(means values)
Implicit 100.8 78.6
Explicit 106.1 78

The performance results are presented in Table 2.
Table 2: Implicit/explicit computation time.
Stamping Spring-back Spring-back
CPU cycles CPU Total CPU
(cycles) (iter. Num.) (CPU per cycle)
Explicit 1160 229379 (-) 2698 (0.01) 3858
(92326)
Implicit - 120 (354) 1589 (13.2) 2749
The implicit simulation for spring-back is performed from 960 ms to 1000 ms. Explicit spring-
back simulation is performed until the kinetics energy on the sheet metal reaches a minimum
value (quasi-static equilibrium). The final computation time is set to 6000 ms.
Explicit and implicit analysis' both obtain good results in this test, with implicit computation being
40% faster than the explicit computation. The implicit approach is; however, 1320 times more
expensive per step than the explicit solver. The use of the implicit approach allows you to
economize on the overall computation time.

Example 26 - Ruptured Plate
Summary
Failure of a circular plate subjected to the impact of an infinite rigid sphere is studied. Material
models, with or without a dedicated failure criteria, are compared. The new failure criteria
available adds to the simple rupture models existing in such material laws as Law 2 and Law 27.
The study is divided into three parts:
 Rupture using a damage model in Law 27
 Failure using the Johnson-Cook model
 Advantage of Forming Limit Diagram as a failure model
The sensitivity of the results for the different failure models is demonstrated.

Title
Ruptured plate
Number
26.1
Brief Description
A metallic thick plate is perforated by a rigid sphere. Simulation of the rupture uses different
failure models.
Keywords
 Rupture, elements deletion, maximum stress, and failure plastic strain
 Johnson-Cook failure model, failure model using Forming Limit Diagram
 Johnson-Cook elasto-plastic material law and damage integrated in the material law
 General shell formulation, progressive plastification, and layers
RADIOSS Options
 Johnson-Cook failure model (/FAIL/JOHNSON)
 Forming Limit Diagram failure model (/FAIL/FLD)
 Material law 2 (/MAT/PLAS_JOHNS) and law 27 (/MAT/PLAS_BRIT)
 Rigid Sphere (/RWALL)
Input File
Law 2 without failure:
<install_directory>/demos/hwsolvers/radioss/26_Ruptured_plate/Law2/Without_FAIL/
LAW2*
Johnson failure:
<install_directory>/demos/hwsolvers/radioss/26_Ruptured_plate/Law2/JOHNSON_model/.
../FAILURE_JOHNSON*
FLD failure:
<install_directory>/demos/hwsolvers/radioss/26_Ruptured_plate/Law2/FLD_model/
Ishell=1_without_epsmax/.../FAILURE_FLD*
Law 27: <install_directory>/demos/hwsolvers/radioss/26_Ruptured_plate/Law27/LAW27*

Advanced

Overview
Aim of the Problem

The purpose of this example is to model the perforation of a thick plate using a rigid sphere. The
simulations are performed using different failure models:
 Damage model integrated in the elasto-plastic material law (/MAT/LAW27)
 Johnson-Cook failure model, in addition to the elasto-plastic material law (/MAT/LAW2)
 Failure model using the generic Forming Limit Diagram (FLD), in addition to the elasto-
plastic material law 2
Numerical results are not compared with experimental data. However, this example proposes
different approaches to take account of failure.
A 3 mm thick plate is impacted at its midpoint by a 12.7 mm diameter sphere with an imposed
velocity of 1 ms-1.
The material undergoes an isotropic elasto-plastic behavior which can be reproduced by a
Johnson-Cook model, independently of the failure model: .

Material properties are:
 Density: 2.8 x 10-3 g/mm3
 Hardening parameter: 562.3 MPa
The maximum stress and the failure plastic strain are considered in the failure modeling section.
The strain rate effect is not taken into account in this example.

The plate is meshed with 4-node shell elements.
The shell properties (Type 1) are:
 5 integration points (progressive plastification)
 Belytschko elasto-plastic hourglass formulation (Ishell = 3)
 Iterative plasticity for plane stress (I plas = 1)
 Thickness is constant (Ithick = 0)
 Initial thickness is uniform, equal to 3 mm
Fig 2: Mesh of the metallic plate with the initial rigid sphere position.

A sphere with non-zero mass is considered as a rigid body, modeled by a rigid sphere of 12.7
mm diameter. Slave nodes include the plate part.
Here in this example, non-zero mass (5g) in /RWALL will keep energy balance. If mass =0, then
99% energy error is received. This is because you have external work done by the rigid wall and
this external work is proportional to the mass of the rigid wall. But if the mass is zero, then the
external works is also zero. The loss of external work will lead to bad energy balance.
A constant imposed velocity of -1.0 ms-1 in the Z-direction is applied on the rigid sphere via the
ID 4067 master node. Its displacement is proportionally linked to time.
Boundaries of the plate are clamped, as shown in Fig 3.

Fig 3: Side fixed in X, Y, Z translations and X, Y, Z rotations.
Failure Modeling
Law 27: Elasto-plastic Material Law using a Damage Model

Law 27 is used to simulate material damage following a Johnson-Cook plasticity law. Thus, a
damage model is incorporated into the material law to take into account the damage evolution
with stress decreasing up to element rupture.
The damage parameters are:
 Tensile rupture strain t1: damage starts if the highest principal strain reaches this tension
value.
 Maximum strain m1: the element is damaged if the highest principal strain is above the
tension value. The element is not deleted.
 Maximum damage factors dmax1: this value should be kept at its default value (0.999).
 Failure strain f1: the element is deleted if the highest principal strain reaches the tension
value.
Fig 4: Stress/strain curve for damage affected material (with i=1,2)

For the first principal direction: For the second principal direction:
= 0.14 t2 = 0.14
t1
m1 = 0.15 m2 = 0.15
f1 = 0.151 f2 = 0.151
dmax1 = 0.999 dmax2 = 0.999
The maximum stress and the failure plastic strain are activated:
 max = 0.151
 max = 425 MPa
The element is removed if one layer (one integration point) of the element reaches the failure
tensile strain.
For further information about this law, refer to the RADIOSS Theory Manual and RADIOSS User's
Guide.
Johnson-Cook Failure Model

The elasto-plastic behavior of the material is defined using the Johnson-Cook law (/MAT/LAW2),
with or without damage ( max and max). The failure model is independent from the material law
and the hardening model.
The Johnson-Cook failure model is defined using /FAIL/JOHNSON in the input deck.
The model uses cumulative damage to compute failure.

 f  D1  D2 exp D3 * 
where,
D refers to the current damage (failure if D = 1)
is the normalized mean stress
is the increment of plastic strain during the loading increment

D1, D2 and D3 are the first three parameters
The strain rate and thermo-plastic effects are not taken into account in this example. Therefore,
only three parameters are required (D1, D2 and D3).
Two cases are considered:
 The maximum stress and the failure plastic strain are not taken into account.
 In addition to the Johnson-Cook failure model, the maximum stress and the failure plastic
strain are activated.

Two failure approaches are also investigated:
 Shell element is deleted if damage D > 1, for one layer (Ifail_sh set to 1)
 The layer stress tensor is set to zero and the shell element is deleted if damage D > 1, for
all layers (Ifail_sh set to 2).
Therefore, the four simulations performed are shown in the following table:
Ifail_sh = 1 Ifail_sh = 2
/FAIL /FAIL
only /FAIL only /FAIL
max , max max , max
D1 = 0.11 D1 = 0.09 D1 = 0.11 D1 = 0.09

D2 = 0.08 D2 = 0.08 D2 = 0.08 D2 = 0.08
Johnson-
Cook failure D3 = -1.5 D3 = -1.5 D3 = -1.5 D3 = -1.5
model max = 0.151 max = 0.151
max = 425 MPa max = 425 MPa
For further information about this failure model, refer to the RADIOSS Theory Manual and the
RADIOSS User's Guide.
FLD Failure Model (Forming Limit Diagram)

This failure model uses the generic forming limit diagram, defined for the given material. The
curve is expressed in the area of principal strains (max and mini strains) and defines the failure
zone.
An input curve and the flag Ifail_sh (same as Johnson-Cook model) are required. However, the
results obtained using Ifail_sh = 1 and Ifail_sh = 2 are very similar and only if Ifail_sh = 1 is
presented.

Two failure modes can be simulated by adjusting the diagram. Shells elements are deleted if one
layer is in the failure zone.
Explosive Perforation (Hole Creation) Perforation by Shell Tearing
Curve 1 Curve 2
For further information about this failure model, see the RADIOSS User's Guide.

The failure simulations considered in the example are:

Law 27 Incorporated damage model + max , max
Material law 27
max , max (without /FAIL)

Law 2
Material law 2
Ifail_sh = 1 Ifail_sh = 2
Law 2 + /FAIL
/FAIL only /FAIL /FAIL only /FAIL
max , max max , max
Johnson failure
model
FLD FLD 1
failure
model FLD 2
Engine file:

In the /DEF_SHELL option defined in the input desk, the I strain flag must be set to 1 for
computing strains in view of post-processing.
During simulation, failure of the elements can be checked in the output file runname_0001.out.
 Example of output file (extract):
[…]
3869 0.5145 0.1330E-03 SHELL 159 0.0% 2451. 0.1692 2.336 0.0000E+00 0.0000E+00
3870 0.5147 0.1330E-03 SHELL 159 0.0% 2452. 0.1697 2.335 0.0000E+00 0.0000E+00
-- RUPTURE OF SHELL ELEMENT NUMBER 151
3871 0.5148 0.1330E-03 SHELL 159 0.0% 2453. 0.1694 2.361 0.0000E+00 0.0000E+00
3872 0.5149 0.1330E-03 SHELL 159 0.0% 2452. 0.7397 3.424 0.0000E+00 0.0000E+00
3873 0.5151 0.1330E-03 SHELL 159 0.0% 2446. 3.288 6.740 0.0000E+00 0.0000E+00
3874 0.5152 0.1330E-03 SHELL 159 0.0% 2443. 5.818 8.481 0.0000E+00 0.0000E+00
3875 0.5153 0.1330E-03 SHELL 159 0.0% 2443. 6.888 8.419 0.0000E+00 0.0000E+00
3876 0.5155 0.1330E-03 SHELL 159 0.0% 2443. 7.988 8.214 0.0000E+00 0.0000E+00
3877 0.5156 0.1330E-03 SHELL 159 0.0% 2437. 12.35 9.924 0.0000E+00 0.0000E+00
3878 0.5157 0.1330E-03 SHELL 159 0.0% 2430. 17.44 11.89 0.0000E+00 0.0000E+00
3879 0.5159 0.1329E-03 SHELL 159 0.0% 2425. 21.28 13.40 0.0000E+00 0.0000E+00
3880 0.5160 0.1329E-03 SHELL 159 0.0% 2421. 23.82 14.56 0.0000E+00 0.0000E+00
3881 0.5161 0.1329E-03 SHELL 159 0.0% 2419. 25.85 15.13 0.0000E+00 0.0000E+00
[…]
Fig 6: Perforation of the plate by the rigid sphere at 5 ms (case: Johnson-Cook failure model without failure plastic strain,
Ifail_sh=2).
The following table compares the results provided by simulations in terms of plate deformation,
hole dimension, residual shells, etc.

Conclusion
The rupture of a circular plate, due to the impact of a rigid sphere was studied and several failure
models with different simulation parameters were compared. The results obtained highlight the
sensitivity of the numerical models to simulate the failure.
Laws 2 and 27, with or without the failure models were compared. The comparison shows that
the results are quite similar when coherent simulation parameters are used.

Example 27 - Football (Soccer) Shots
Summary
This is mainly a demonstration example. An unusual application is used: In 1976, the European
Football (Soccer) Cup final was between the Bayern of Munich and Saint Etienne. During the
match, two shots from Saint Etienne rebound on the opposite bars, shaped as a square cross-
section. The fact that England is the only European country not having replaced its square bars
by round bars always makes French supporters believe that Saint Etienne could have won the
final if the bars had been round. This example provides an answer through simulation; however,
the controversy will no doubt continue.

Title
Football shot
Number
27.1
Brief Description
Simulation of a football (soccer) shooting impact on bars.
Keywords
 Q4 and T3 meshes, and Orthotropic shell
 Airbag modeling and sensor
 Rigid cylinder
 Initial velocity
RADIOSS Options
 Initial velocities around axis (/INIVEL/AXIS)
 Monitored volume type airbag (/MONVOL/AIRBAG1)
 Rigid bodies (/RBODY)
 Rigid cylinder (/RWALL)
 Sensor (/SENSOR)

 Video films for a square cross-section case.
Input File
Bathenay’s shot:
<install_directory>/demos/hwsolvers/radioss/27_Football_shots/Bathenay_circular/
BAT_CIR*
<install_directory>/demos/hwsolvers/radioss/27_Football_shots/Bathenay_square/
BAT_SQR*
Santini’s header:
<install_directory>/demos/hwsolvers/radioss/27_Football_shots/Santini_circular/
SANT_CIR*
<install_directory>/demos/hwsolvers/radioss/27_Football_shots/Santini_square/
SANT_SQR*

Medium

Overview
Aim of the Problem

During the European football (Soccer) Cup final in 1976 (Bayern of Munich versus Saint Etienne),
a shot from Bathenay (Fig 1) and a header from Santini (Fig 2) rebounded off the square cross-
section frame of the German team’s goal. The purpose of this demonstration is to determine the
influence of a square or a round cross-section bar for both cases.
Fig 1: Bathenay’s shot.
Fig 2: Santini’s header

The main differences between both shots are the incidence, the velocity and the impact point of
the ball on the bar (its vertical value).
The material used for the ball follows a linear elastic orthotropic law (/MAT/LAW19) with the
 Young modulus (dir. 1 and 2): 20000 MPa
 Shear modulus (12, 23 and 31): 10000 MPa

The ball is modeled using 60 3-node shells and 1420 4-node shells. The shell element
formulations are set by default. The bar and the ground are also modeled with 4-node shell
elements, but their mesh will not be used for the computation.
Fig 4: Ball mesh.

 A rigid body is created, containing all the nodes of the ball. It is deactivated just before
impact on the bar.
 Two initial velocities are applied to the rigid body’s master node. They are defined using
/INIVEL/AXIS for the rotary motion and /INIVEL/TRA for the translating motion.
 The bars are modeled with a cylindrical rigid wall for the round bar and two rigid
parallelograms for the square bar.
 Gravity is taken into account using a gravity load.
 The ball is considered as an airbag, which is activated when the rigid body is deactivated.

Fig 5: Impact of the ball on a square and a round cross-section.
Fig 6: Trajectory of the ball for Bathenay’s shot (impact on a square and a round bar respectively).

Fig 7: Trajectory of the ball for Santini’s head (impact on a square and a round bar respectively).
Conclusion
Even using a simple modeling of the impact (bars modeled with rigid walls instead of parts), the
simulation provides quite accurate results in the case of a square cross-section when simulations
are compared to reality. The results obtained for the bars with a round cross-section show that
the ball enters to the goal for both shots. However, several impact parameters, such as friction
and rotational velocity are estimated as calibrating the case of a square cross-section.
Conclusions on the results of the match in case of cylindrical bars should be moderated.

Example 37 - Analytical Beam
Summary
This example deals with the use of RADIOSS linear and nonlinear solvers. A beam submitted to a
concentrated load on one extremity and fixed on the other hand is studied. This problem is well
known and results can be compared with analytical solutions.
Different meshes are tested: beam, shell, thick-shell, and brick elements. For thick-shell
representation, different formulations are also tested: HA8, HSEPH, and 16-node thick-shell.
Moreover, this study tests with solvers: linear and nonlinear implicit solvers, as well as nonlinear
explicit solver. The linear solver is used for small displacements, whereas nonlinear can solve
more problems.
The main propose of this example is to illustrate how to prepare a RADIOSS deck for linear
analysis. It also demonstrates a high quality of RADIOSS finite elements to resolve linear and
nonlinear problems.

Title
Analytical beam
Number
37.1
Brief Description
A cantilever beam submitted to a shear-bending is tested by RADIOSS linear and nonlinear
solvers. Different kinds of RADIOSS finite elements provide results close to the analytical one.
Keywords
 Linear solver
 Beam, shell, thick-shell and brick elements
 BATOZ, HA8 and HSEPH formulations
 Dynamic relaxation and implicit solver
RADIOSS Options
 Implicit options (/IMPL)

 Analytical results
Input File
2 Bricks:
<install_directory>/demos/hwsolvers/radioss/37_Analytical_Beam/Analytical_beam/
2_bricks/explicit_nonlinear/POUTRE*
 Linear: //.../2_bricks/implicit/Linear/.../POUTRE*
 Nonlinear: //.../2_bricks/implicit/Nonlinear/.../POUTRE*
4 Bricks:
<install_directory>/demos/hwsolvers/radioss/37_Analytical_Beam/Analytical_beam/
4_bricks/explicit_nonlinear/POUTRE*
 Linear: //.../4_bricks/implicit/linear/.../POUTRE*
 Nonlinear: //.../4_bricks/implicit/nonlinear/.../POUTRE*

Beam:
<install_directory>/demos/hwsolvers/radioss/37_Analytical_Beam/Analytical_beam/bea
m/explicit_nonlinear/POUTRE*
 Linear: //.../beam/implicit/linear/.../POUTRE*
 Nonlinear: //.../beam/implicit/nonlinear/.../POUTRE*
Shell:
<install_directory>/demos/hwsolvers/radioss/37_Analytical_Beam/Analytical_beam/she
ll/explicit_nonlinear/POUTRE*
 Linear: //.../shell/implicit/linear/.../POUTRE*
 Nonlinear: //.../shell/implicit/nonlinear/.../POUTRE*
Thick Shell:
<install_directory>/demos/hwsolvers/radioss/37_Analytical_Beam/Analytical_beam/thi
ck_shell/explicit_nonlinear/POUTRE*
 Linear: //.../thick_shell/implicit/linear/.../POUTRE*
 Nonlinear: //.../thick_shell/implicit/nonlinear/.../POUTRE*

Beginner
Overview
Aim of the Problem

This study deals with two choices: finite element approach and resolution techniques.
A simple beam is fixed on one extremity, and loaded on the other hand by concentrated load:

The material behavior is linear elastic law (/MAT/LAW1) with the following properties:
 Density: = 0.0078 g/mm3
 Young modulus: E = 210 000 MPa
 Poisson coefficient: = 0.29
Analysis Assumptions and Modeling Description
The beam is modeled with four different kinds of mesh: beams (/BEAM), shells (/SHELL), thick-
shells (/PROP/TSHELL), and bricks (/BRICK).
Fig 2: Different meshes.
Each formulation has particular properties (/PROP). Beam elements use the default formulation
(Ismstr = 0). Furthermore, in order to satisfy Timoshenko’s beam assumptions, use the following
properties:
 Cross section: 100 mm2
 Moment of inertia: 833.33 mm4
This conducts to a shell thickness of 10 mm, where BATOZ shell formulation is used (I shell = 12).
For the solid mesh, the HA8 formulation (I solid = 14) is applied. No reduced pressure integration
is necessary for implicit computation, as the behavior is elastic (I cpre = 0).
For thick-shell elements (/PROP/TSHELL), several formulations are tested: HA8 (Isolid = 14),
HSEPH (Isolid = 15), and 16-node thick-shell (Isolid = 16), which require a specific nodal
connectivity, as shown below:
Fig 3: Thick-shell element numbering

The boundary conditions option (/BCS) is used to fix the beam on one extremity.
To apply the load, a rigid body is created (/RBODY), and then the beam is submitted to an
increasing force (/CLOAD and /FUNCT).
When the loading is small enough, the beam behavior can be considered to be linear. The easiest
way to resolve the problem is to use RADIOSS linear solver.
The implicit RADIOSS solver considers a static loading. It needs to define some parameters with
the /IMPL option. Under the linear behavior assumption, RADIOSS linear solver can be used by
putting /IMPL/LINEAR in the Engine file. The nonlinear implicit solver can be activated by
/IMPL/NONLIN in the Engine file. To solve the equation, both methods need a numerical linear
solver as direct solver MUMPS (/IMPL/SOLVER/2).
The nonlinear explicit solver can also be used. In this case, the load is considered to be applied
dynamically. The dynamic relaxation technique (/DYREL) allows speed converge to the static
solution by introducing a numerical damping. That implies the knowledge of the frequency to
damp. To obtain the dominating period of the structure to damp, look at the kinetic energy
variation in time during the first phase of run. The modal analysis solver by eigen values in
RADIOSS can also be used.
Analytical solution
Fig 4: Notation.
For the linear problem, the analytical solution gives:

For nonlinear case, the Timoshenko and Gere study provides the following results [1]:
Fig 5: Timoshenko and Gere results.
In the case of linear behavior, the numerical results are quite close to the analytical solution
(error lower than 0.01% for all meshes).
Fig 6: Linear solutions.

For the nonlinear case, the results obtained by RADIOSS with explicit and implicit solvers are in
good concordance with the analytical solution of Timeshenko and Gere.
Fig 7: Comparison of results.
The numerical error or the difference between the numerical results and the analytical solution
can be evaluated by:
(4)

Fig 8: Error on w/L
Here it appears that errors for the slender case ( =2) are high for the explicit solver. This is
mainly due to difficult convergence of the dynamic solution to the static response. High flexibility
means low frequency vibrations and thus low efficiency of dynamic relaxation method. On the
other hand, the implicit solver converges to the good solution. The error is remains lower than
0.5%.
Regarding the thick-shell elements, RADIOSS results are quite close to the analytical solution.
The error remains always less than 0.5%.

Fig 9: Thick-shell solutions.
Fig 10: Error for thick-shell formulations.

A schematic deformation mode for each case of mesh is shown in the following figures.
Fig 11: Beam animation
Fig 12: Shell animation.

Fig 13: Shell 16 animation.
Fig 14: Brick animation

Reference
[1] S.P. Timoshenko, J.M. Gere, "Mechanics of materials", D. Van Nostrand Co, 1972

Example 39 - Biomedical Valve
Summary
This example aims at demonstrating how to perform an FSI run using RADIOSS on a relatively
simple case. The maximum deflection of a flap in an interaction with a transient fluid is computed
once the stationary state is reached.
In this example, the two following points are emphasized:
 How to set up an FSI case study
 Fast description of the various options used in an ALE/CFD run (refer to the RADIOSS
Theory Manual for more information)

Title
Biomedical Valve
Number
39.1
Brief Description
A Fluid-Structure-Interaction (FSI) problem is studied. The RADIOSS ALE/CFD solver is used to
resolve the problem.
Keywords
 FSI, CFD, and Fluid
 INLET and OUTLET
 Turbulent Fluid
RADIOSS Options
 /MAT/BOUND
 /ALE/MAT
 /ALE/BCS
 /PROP/FLUID
 /VEL/ALE
Input File
<install_directory>/demos/hwsolvers/radioss/39_Bio_Valve/BIO_VALVE/VALVE*

Skilled

Overview

A simplified heart valve is modeled. The valve opens under the pressure of the incoming blood
flow. As the opening process of the valve is taken into account, the problem is transient.
Additionally, fluid-structure interaction must be taken into account, as the flap deforms under the
pressure of the blood.
Units: Kg, m, s, N, Pa
Fig 1: Definition of the problem.
RADIOSS ALE/CFD Terminology
Euler Formulation
The Eulerian formulation is classical in fluid mechanics. The mesh is fixed and material flows
through the mesh. Equations are modified with respect to the Lagrangian formulation in order to
take into account the convective terms.
It can be activated for a specific part by a flag in material data:
/EULER/MAT/mat_ID
Where, mat_ID is the identification number of the material to be set Eulerian.
In this case, the Eulerian formulation cannot be used because the boundaries of the domain (and
mainly the flap) move.
ALE (Arbitrary Lagrangian Eulerian) Formulation
The material flows through an arbitrary moving mesh and it can degenerate either in a
Lagrangian or an Eulerian formulation.
This option can be activated for a specific part by a flag in material data:
/ALE/MAT/mat_ID
Where, mat_ID is the identification number of the material to be set ALE.

Grid velocities and displacements are arbitrary.
In practice, built-in algorithms determine smooth grid deformation according to displacements of
the ALE domain boundaries. Several algorithms are available (DONEA, SPRINGS, DISP, and
ZERO), in this case, the DISP option is used: the velocity of a node is computed using the
average velocities of the connected nodes.
Boundary nodes between ALE and Lagrangian materials must be set Lagrangian: grid and
material velocities are equal. Boundary nodes between ALE and Eulerian materials with must
have a fixed grid velocity.
Both conditions are set using the /ALE/BCS option.
We can also specify extended boundary conditions for ALE nodes (grid velocity components can
be set to 0 or to the material velocity), or impose grid velocities or ALE links to any nodes in a
similar manner to classical kinematic conditions.
Nodal Boundary Conditions
Kinematic constraints act on material velocities and accelerations. In RADIOSS CFD, a wide
variety of such constraints can be defined. For fluid applications, options of interest are:
 fixed and full slip boundary conditions
 imposed velocities (for example: imposed flux at inlet)
 rigid links (temporarily adds during restarts)
 rigid bodies to model rigid structures and connections and also to compute drag and lift
forces (that is: fluid impulse on rigid body is stored in time history database)
Grid constraints act only on grid velocities. You can specify:
 fixed and full slip grid conditions
 Lagrangian conditions, that is: grid and material velocity are set equal.
 ALE links to maintain regular distribution of nodes.
 imposed grid velocities (for example: moving inlet and outlet)
Elementary Boundary Conditions
Boundary elements allow prescription of element values at domain boundaries. They can be
specified by assigning material law type 11 (or type 18 in purely thermal cases) to boundary
elements. Those are quads in 2D and solids in 3D. For each variable P, rho, T, k, epsilon, internal
energy, you can recommend:
 imposed varying conditions according to user function
 continuity
 smoothly varying predefined function
Non-reflective frontiers (NRF) (material type 11, option 3) ensures free field impedance to
pressure and velocity fields.
With RADIOSS ALE/CFD, any combination of the above options can be specified. On the
counterpart, the closure of the various convection and diffusion equations has to be verified
carefully by you.

Generally the following elementary boundary conditions are used:
 Inlet, flux is imposed using imposed velocities; density, energy, and turbulent energy (that
is, k) are imposed as constants. Continuity is imposed for pressure (display purposes only)
and for epsilon. Turbulent energy, rho k is set to zero for external flows and to
1.5*rho*(0.06 Vin)2 for internal flows.
 Outlet, continuity for all variables except pressure, which is imposed. When using the NRF
option, you need to provide a value for sound speed and a typical relaxation length, which
must be greater than the biggest wave length of interest.
 Sides, continuity for all variables with NRF option or slip conditions without boundary
elements.
If no element exists at boundary, continuity is assumed but kinematic conditions are necessary to
disallow fluxes; otherwise, the convection equation is not closed and the program might diverge.
A quasi-uniform solid mesh is used for domain discretization. One element through the thickness
with a fine enough mesh along the axis is used.
Shell elements are used to model the flap. The flap is clamped on one side and its nodes are
attached by the springs to the clamp. One row of meshes are created at each extremity to define
inlet and outlet.
The problem is incompressible; therefore, in order to increase the time step, the speed of the
sound in the fluid has been arbitrarily reduced to 50 m/s.
When doing such an approximate, it must be verified that the velocity of the fluid is much lower
than the modified speed of the sound.
Fig 2: Mesh of the model

Four material laws are defined:
 A linear elastic material for flap (/MAT/ELAST),
 A turbulent fluid material for the main parts of model (/MAT/LES_FLUID),
/MAT/LES_FLUID
Rho = 960.0 Kg/m3
Sound speed = 50.0 m/s
Molecular kinematic viscosity = 5.45E-05 N.s/m
Sub-grid scale model flag = 0
Cs = 0.1
Csp = 0.1
 A fluid material for the inlet to define density, energy and pressure of fluid (/MAT/BOUND),
/MAT/BOUND
Rho = 960.0 Kg/m3
Ityp = 2 (General case)
 A fluid material for the outlet to define pressure of fluid outside the domain (/MAT/BOUND).
/MAT/BOUND
Rho = 960.0 Kg/m3
Ityp = 3 Non-reflective frontiers (NRF)
Characteristic length = 1.0E-03 m
The format /ALE/MAT is assigned to each of fluid materials.
Two imposed velocity are applied to the inlet nodes:
 Upper inlet: Vx= 1.253 m/s
 Lower inlet: Vx= 0.849 m/s
The boundary conditions are defined in the following table:
Type Position Boundary Condition
1 /BCS Lateral nodes Translation Vz = 0

2 /ALE/BCS Lateral nodes Grid velocity Wz = 0
3 /BCS Nodes domain on the Translation Vz = 0
lateral edge of flap
4 /ALE/BCS Nodes domain on the Wx = Vx
lateral edge of flap Wy = Vy
Wz = Vz
5 /BCS Nodes on flap Translation Vz = 0
Rotation Wx = 0
Rotation Wy = 0
6 /ALE/BCS Nodes on flap Wx = Vx
Wy = Vy
Wz = Vz

An interface type 2 is created to connect the nodes of fluid domain on the lateral edge of flap, to
the Lagrangian mesh of flap. Thus, the fluid domain is connected to the structural part.
With the use of this method, it is possible to have different meshes and mesh densities between
the fluid and the structure.
Vorticity distribution in the transient period gives a good overview of the problem evolution in
time before stabilization.
Fig 3: Vorticity distribution in time and in space

The main purpose of this study is to obtain the maximum deflection of the flap in time. Plotting
the vertical displacement of the node 23360 given in the following graph in which the flap
position is stabilized at time t=1 s.
The pressure stabilization in time is shown in Fig 5 for elements 3370 and 3992.
Fig 4: Vertical displacement of the free extremity (node 23360) of the flap in meter
Fig 5: Pressure stabilization for elements 3370 and 3992 in Pa

Fig 6: Position of elements 3992 and 3370
This example demonstrates RADIOSS capabilities to simulate transient Fluid-Structure-

Interactions. The use of the ALE formulation attached to a Lagrangian mesh is described. Some
elementary explanations to RADIOSS ALE/CFD terminology are mentioned.

Example 42 - Rubber Ring: Crush and Slide
Summary
The model consists of a deformed rubber ring resting on a flat, rigid surface. Another circular
rigid roller rests at the top of the ring, and is in contact with the ring at just a point. Contact is
defined between the rigid surfaces and the outside surface of the ring and self-contact is defined
in the inside surface of the ring. The loading is applied in two steps – in the first step, the circular
roller is pushed down enough to produce self-contact of the inside surface of the ring. In the
second step, the roller is simultaneously translated and rotated such that the crushed ring rolls
along the flat rigid surface producing a constantly changing region of contact.
This example is considered a static problem and the nonlinear implicit solver is used.

Title
Rubber-ring
Number
42.1
Brief Description
A rubber ring resting on a flat rigid surface is pushed down by a circular roller to produce self-
contact on the inside surface of the ring. Then the roller is simultaneously rolled and translated
so that crushed ring rolls along the flat surface.
Keywords
 Nonlinear implicit large displacement analysis
 Self-contact
 Hyper-elastic material
RADIOSS Options
 Hyper-elastic rubber material (/MAT/LAW42)
 Releasing of DOF (/BCSR)
 Incompressible solid element (/PROP/SOLID)
 Contact definition (/INTER/TYPE7)
 Implicit analysis (/IMPL)
Input File
Rubber_ring:
<install_directory>/demos/hwsolvers/radioss/42_Rubber_ring/rubber_ring*

Beginner

Overview
Aim of the Problem

The purpose of this example is to demonstrate a nonlinear large displacement implicit analysis
involving hyper-elastic material and contacts using RADIOSS.
The deformed ring has an inner diameter of six units and an outer diameter of 8 units and the
thickness of the ring is 0.67 units. The diameter of the circular rigid roller is 2 units. The
thickness of the rigid flat surface and the circular roller is 0.01units. A question that might come
to mind is: "Why a half-symmetric model is not be used in this example?" Now, even though the
loading in the first step is symmetric, the loading in the second step is not, so the full model had
to be used for the example to simulate the loading of both the steps.
Fig 1: Geometry of the rubber ring and roller model.
The hyper-elastic rubber ring has been modeled using the Ogden, Mooney-Rivlin material
(/MAT/LAW42) with the following properties:
 1: 0.7
 2: -0.5
 1: 2.0
 2: -2.0
The rigid barriers (roller and flat surface) have been modeled using elastic material, but a 1D
rigid link has been connected to all the nodes of each of the barriers, making them essentially
rigid. The material used for the rigid barriers has the following properties:
 Density: 7.9e-9

Geometric Linear (NLSTAT) or Geometric Nonlinear (NLGEOM) Analysis

In geometric linear analysis all deformations and rotations are small – displacements of 5% of
the model dimension are considered small.
For this rubber ring example, the final deformations and strains after crushing the rubber ring are
much larger than the above mentioned limit. So, the geometrically linear static NLSTAT analysis
could not be considered for this example.
1) Mesh and properties
The ring mesh is a regular solid mesh modeled with four elements (dimension of 0.25 units)
along the width and two elements (dimension of 0.335 units) through the thickness. The flat rigid
surface and the circular roller are both modeled as a regular shell mesh of 0.01mm thickness
with the flat surface around 0.75 units in dimension and the roller being around 0.467 units in
dimension.
Fig 2: Geometry of the rubber ring and roller model.
The ring has been modeled using first order fully-integrated solid elements.
/PROP/SOLID/5
WHEEL
14 10 1 222
The flat surface and roller have been modeled using the first order reduced integration shell
elements with three integration points through the thickness. Full integration elements were not
considered as you are not interested in any detailed post-processing of the barrier.
/PROP/SHELL/6
BARRIER
1 2
3 0.01

2) Load and boundary conditions
The boundary conditions applied to the rubber ring in step 1 are shown in Figure 3.
Fig 3: Boundary conditions applied to ring in first step
The boundary conditions applied to the flat rigid surface and circular roller in step 1, are shown in
Figure 4.
The flat surface is constrained in all DOF's, while the roller is pushed down by 6.22 units in Y-axis
so much that self-contact is established within the inner surface of the ring.
Fig 4: Boundary conditions applied to flat surface and roller in 1st step

In the second step, the top roller is to be simultaneously translated and rotated such that the
wheel in the crushed configuration rolls along the flat rigid surface in –X direction. So, the X
translation and ZZ rotations of the circular roller have to be released from the primary node of
/RBODY. Additionally, the center nodes of the ring that were constrained in X DOF (Fig 3) need
to be released for the ring to roll along the flat surface. So, the Engine file for the second step
has the following cards representing release of the above-mentioned degrees of freedom.
/BCSR/TRA/X/
5 6 8 9 15 16 17 18
87 88 89 93 94 95 241 242
243 244 245 246 5717 5699 5681 5663
5662 5537 5517 5497 5477 5476 2269
/BCSR/ROT/Z
2269
3) Contact definition
Several contacts have been defined: i) contact between the circular roller and rubber ring, (ii)
contact between the flat rigid surface and rubber ring, and (iii) self-contact within the inner
surface of the rubber ring.
A small physical gap (0.05 units) has been introduced between the circular roller and the rubber
ring and also between rubber ring and the flat rigid surface. The minimum gap specified for the
contact is slightly higher than the physical gap for contact to take effect. Static Coulomb friction
of 0.5 is defined for all the interfaces. The definition of one such interface is shown below:
/INTER/TYPE7/14
TOP_Rubber
25 30 4 0
0.5 0.055
000 0
0 2
Also, since the contact involved is between a rigid part and a very soft hyper-elastic material, it is
advisable that the E*h (Young’s modulus * thickness) of the rigid part be approximately the
same order as the bulk modulus of the rubber material.
The hyper-elasticity and contact causes major nonlinearities. Therefore, a static nonlinear
analysis is performed using the arc-length displacement strategy. The time step is determined by
a displacement norm control.

Nonlinear solver: BFGS Quasi-Newton
method
Termination criteria: Relative residual in
energy
Tolerance: 0.001
Time step control method: Arc-length
Initial time step: 0.001
Minimum time step: 1e-6
Maximum time step: 0.001
Line search method: AUTO
Special residual force computation
with 5
contact interfaces present:
Desired convergence iteration 6
number:
Maximum convergence iteration 15
number:
Maximum increasing time step scale 1.1
factor:
Spring-back option: No
A solver method is required to resolve Ax=b in each iteration of a nonlinear cycle. It is defined in
the option /IMPL/SOLVER. The linear implicit options used are:
Linear solver: Direct
Precondition methods: Factored approximate

Inverse
Maximum iterations System dimension (NDOF)

number:
Stop criteria: Relative residual of

preconditioned matrix

A restart analysis is performed for the second load step.

The input implicit options set in both the Engine files are:
/IMPL/PRINT/NONLIN/-1 --- Printout frequency for nonlinear iteration
/IMPL/NONLIN/2 ---- Static nonlinear computation
5 1 0.001
/IMPL/SOLVER/3 ----- Solver method (solve Ax=b)
5 0 3 0.0
/IMPL/DTINI ----- Initial time step determines initial loading increment
0.001
/IMPL/DT/STOP ------- Min Max values for time step
1e-6 0.001
/IMPL/DT/2 ------ Time step control method 2 – Arc-length + Line-search will be

used with this method to accelerate and control convergence.
6 0 15 0.8 1.1
/IMPL/AUTOSPC/ALL ---- Constraining automatically zero stiffness dof
/IMPL/LSEARCH/3 ---- Line search method for nonlinear analysis
/IMPL/RREF/INTER/5 ------ Special Reference residual computation with contact
Refer to the RADIOSS manual for more details about implicit options.

Animations
The deformed shape of the rubber ring after the circular roller is pushed down enough is shown
in Figure 5.
Fig 5: Deformed shape of the rubber ring after 1st step
Figure 6 shows the slide of the crushed rubber ring along the flat rigid surface after the roller has
been simultaneously translated and rotated.

Fig 6: Deformed shape of the rubber ring after 2nd step
The stresses in the rubber ring after it has been crushed and sliding along the flat rigid surface
are shown in Figure 7.
Fig 7: Stress in the rubber ring

Example 43 - Perfect Gas Modeling with Polynomial EOS
Summary
Polynomial EOS is often used by RADIOSS to compute hydrodynamic pressure. It is cubic in

compression and linear in expansion.
where,
(1) and (2)
Material law 6 (/MAT/HYDRO) uses this equation to compute hydrostatic pressure. It is possible
to consider absolute values or relative variation (Table 1). This example shows how to build
material control cards for each of the following cases:
Mathematical
Case Pressure Energy
model
1 P  , E  absolute absolute
2 P   , E  relative absolute
3 P   , E  relative relative
4 P   , E  absolute relative
Table 1: Modeling formulation for perfect gas with /MAT/HYDRO
A simple test of compression/expansion is made to compare these formulation outputs with

theoretical results.

Title
Perfect Gas
Modeling with
Polynomial
EOS
Number
43.1
Brief Description
Polynomial EOS is used to model perfect gas. Pressure or energy can be absolute
values or relative. Material law 6 (/MAT/HYDRO) is used to build material cards for each
of these cases.
Keywords
 Perfect gas
 Polynomial EOS
 Absolute/Relative formulations
 Pressure shift
RADIOSS Options
 Hydrodynamic fluid material (/MAT/LAW6 (HYDRO))
 Boundary conditions (/ALE/BCS)
Compare to / Validation method

Input File
Model 1:
<install_directory>/demos/hwsolvers/radioss/43_perfect_gas_polynomial_eos/
01-Pabsolute_Eabsolute/*
Model 2:
02-Prelative_Eabsolute/*
Model 3:
03-Prelative_Erelative/*

Model 4:
04-Pabsolute_Erelative/*

Beginner
Overview
Aim of the Problem

The purpose of this example is to plot numerical pressure, internal energy, and sound speed for a
perfect gas material law. Comparison to theoretical results is made.
This test consists with an elementary volume of perfect gas undergoing spherical expansion and
compression.
Initial conditions are listed below:

P0 = 1e5 Pa
V0 = 1000 m3
0 = 1.204 kg/m3
0 = 0
The fluid will be assumed to be a perfect gas. Volume is changed in the three directions to
consider a pure compression (-1 < < 0) followed by an expansion of matter (0 < ). See
Figure 1.
This test will be modeled with a single ALE element (8 node brick) and polynomial EOS.
Evolutions of pressure, internal energy and sound speed will be compared between numerical
output and theoretical results.

Fig 1: Elementary volume change. Length is modified with /IMPDISP card; its influences on V and are plotted.

Nodes on each of the faces are moved with imposed displacement (/IMPDISP).
Boundary nodes are defined as Lagrangian with the /ALE/BCS card.
Element pressure, density and internal energy density are saved in the Time History file.
Polynomial EOS
Polynomial EOS is used in material law 6 (/MAT/HYDRO) to compute hydrodynamic pressure. It is
cubic in compression and linear in expansion.
Where, P is the hydrodynamic pressure.
(1)

and
(2)
are called hydrodynamic coefficients and they are input flags. Hypothesis on the
material behavior allows determining of these coefficients:
 General case corresponds to Mie-Guneisen EOS (see Appendix C of the Theory Manual)
 Incompressible gas
 Linear elastic material
 Perfect gas
This example is focused only on Perfect Gas modeling.
Theoretical Results
The purpose of this section is to plot pressure, internal energy, and sound speed in function of
the single parameter V or .
1. Pressure
Perfect gas pressure is given by:
PV   1 Eint
(3)
Then,
RADIOSS assumes the hypothesis of an isentropic process to compute the change in internal
energy:
dEint = -PdV
This theory gives the following differential equation:
dP P

dV V
y '  x  0
This has the form and the general solution is:
y  Cst.x 

Pressure is also polytropic:
PV   PoVo

 V0 
P(V )  P 0  
V  (4)
Here, is the material constant (ratio of heat capacity). For diatomic gas =1.4. Air is made
mainly of diatomic gas, so set gamma to 1.4 for air.
2. Internal Energy
Equations (3) and (4) lead to the immediate result:
3. Sound Speed
Perfect gas sound speed is:
(5)
Equation (4) gives its expression in term of volume:
The theoretical results are listed in the table below. Pressure, internal energy, and sound speed
are expressed both in function of V and .
Pressure (Pa) Internal Energy Density (J) Sound Speed (m/s)
PREF(V) PREF( ) eREF(V) eREF( ) cREF(V) cREF( )

Corresponding plots are shown below:
Fig 2: Perfect Gas Pressure
Fig 3: Perfect Gas Internal Energy
Fig 4: Perfect Gas Sound Speed

A single ALE brick element is used. Material is confined inside the element by defining brick nodes
as Lagrangian. For each face, displacement is imposed on the four nodes along the normal.
Material law 6 (/MAT/HYDRO) is used and describes the hydrodynamic viscous fluid material.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
/MAT/LAW6/mat_ID/unit_ID or /MAT/HYDRO/mat_ID/unit_ID
mat_title
i
C0 C1 C2 C3
Pmin Psh
C4 C5 E0
Pressure Shift
Material law 6 introduces flag Psh which allows shifting computed pressure in the polynomial
equation of state:
RADIOSS Engine shifts C0 flag and computed pressure P( ,E) with an offset of -Psh.
Minimum Pressure
The theoretical value is Pmin = 0 Pa (absolute pressure) with a default value of -1030, to accept a
negative value in relative pressure formulation.
This flag has to be manually offset with -Psh.

Material Control Cards

Material is supposed to be a perfect gas. The following cases have been investigated:
P  , E 
 Case 1: Both Pressure and Energy are absolute values:
P   , E 
 Case 2: Pressure is relative and Energy is absolute:
P   , E 
 Case 3: Both Pressure and Energy are relative:
P   , E 
 Case 4: Pressure is absolute and Energy is relative:
Case 1: Both Pressure and Energy are absolute values
1. Equation of State
Equation of state can be written:
with
Expanding this expression and identifying the polynomial coefficients leads to:
P   , E    C4  C5   E
where,
2. Corresponding Input
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
AbsolutePRESSURE_AbsoluteENERGY
i
0 0 0 0
0 0
C4 = -1 C5 = -1

3. Output Results
Time Initial
Measure Unit
History Value
/TH/BRICK (P) P P0 Pressure
/TH (IE) Eint (= E x V0) E0V0 Energy
/TH/BRICK (IE) Eint / V E0 Pressure
4. Comparison with Theoretical Result

Numerical result for perfect gas pressure is given by time history. Element time history (RADIOSS
/TH/BRICK) allows displaying it. This result is compared to a theoretical one. Curves are
superimposed.
P  , E 
Fig 5: Numerical pressure, model 1:
Internal energy can be obtained through two different ways. The first one is internal energy density
(Eint / V) recorded by element time history (RADIOSS /TH/BRICK). The second one is the internal
energy from the global time history because the model is composed of a single
element.
P  , E 
Fig 6: Numerical internal energy, model 1:

Case 2: Pressure is relative and Energy is absolute
Equation of state for a perfect gas is:
Calculating Pressure from a reference one provides relative pressure:
Expanding this expression and identifying with polynomial coefficients leads to:
P( ,E) = P( ,E) = Psh = -Psh + (C4 + C5 )E
where,
2. Minimum Pressure
Then, the minimum pressure must be set to a non-zero value Pmin = -P0.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
RelativePRESSURE_AbsoluteENERGY
i
0 0 0 0
-P0 P0
C4 = -1 C5 = -1

4. Output Result
Initial
Time History Measure Unit
Value
/TH/BRICK (P) P 0 Pressure
/TH (IE) Eint (= E x V0) E0V0 Energy
/TH/BRICK (IE) Eint / V E0 Pressure

Element time history (/TH/BRICK) is the pressure relative to Psh. The resulting curve is then shifted
with Psh value and starts from 0.
P   , E 
Internal energy can be obtained through two different ways. The first one is internal energy density
(Eint / V) recorded by element time history (/TH/BRICK). The second one is the internal energy from
the global time history because the model is composed of a single element.
P   , E 

Case 3: Both Pressure and Energy are relative
Initial internal energy can be introduced:
Calculating pressure from a reference one provides:

P( ,E) - P0 = P = ( - 1)(1 + )( E + E0) - P0
Where,
P( , E) = P( ,E) - Psh = C0 - Psh + C1 + (C4 + C5 ) E
where,
C0 = C1 = E0( - 1)
C4 = C5 = -1
E0 = 0
Psh = P0
2. Minimum Pressure
The minimum pressure must be set to a non-zero value Pmin = -P0
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
RelativePRESSURE_RelativeENERGY
i
E0( - 1) E0( - 1) 0 0

-P0 P0
C4 = -1 C5 = -1 0
4. Output Results
Initial
Time History Measure Unit
Value
/TH/BRICK (P) P 0 Pressure
/TH (IE) Eint (= E x V0) 0 Energy
/TH/BRICK (IE) Eint / V 0 Pressure

Element time history (/TH/BRICK) is the pressure relative to Psh. The resulting curve is then
shifted with Psh value and starts also from 0.
P   , E 
Internal energy can be obtained through two different ways. The first one is internal energy
density (Eint / V) recorded by element time history (/TH/BRICK). The second one is the
internal energy from the global time history because the model is composed
of a single element. This numerical internal energy is relative to its initial value; it is shifted
with the E0V0 value from the absolute theoretical one and also starts from 0.

P   , E 
Case 4: Pressure is absolute and Energy is relative

Initial internal energy can be introduced:
Which leads to:

P( ,E) = ( -1)(1 + )(E0 + E)
P( ,E) = C0 + C1 + (C4 + C5 ) E
Where,
C0 = C1 = E0 ( - 1)
C4 = C5 = -1
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
AbsolutePRESSURE_RelativeENERGY
i
E0( - 1) E0( - 1) 0 0
0 0
C4 = -1 C5 = -1 0

3. Output Results
Time Initial
Measure Unit
History Value
/TH/BRICK (P) P P0 Pressure
/TH (IE) Eint (= E x V0) 0 Energy
/TH/BRICK (IE) Eint / V) 0 Pressure

Element time history (/TH/BRICK) gives absolute pressure. This result is compared to a
theoretical one. Curves are superimposed.
P   , E 
Internal energy can be obtained through two different ways. The first one is internal energy
density (ΔEint / V) recorded by element time history (/TH/BRICK). The second one is the internal
energy from the global time history because the model is composed of a single
element. This numerical internal energy is relative to its initial value; it is shifted with the E0V0
value from the absolute theoretical one and also starts from 0.

P   , E 
Sound Speed and Time Step

Material law 6 computes sound speed through the usual expression for fluids:
It can be written in function of :
Then,
The total differential of P in terms of internal energy E and is:
In case of an isentropic transformation (reversible and adiabatic), the change of internal energy
Eint with volume V and pressure P is given by:
dEint = -PdV
Using relation which links Eint and E leads to:
can be expressed in terms of volume ratio:
its variation in function of the volume change is also:

Change in internal energy per unit volume E is then:

Finally, the sound speed is given by:
(5)
This expression computes the sound speed for a given equation of state P( ,E). In the case of
perfect gas, it was shown that for each type of formulation (absolute or relative), EOS can be
written:
P( ,E) = C0 + C1 + (C4 + C5 )E
Equation (5) is used to compute sound speed:
(6)
This calculation is then applied for each of the four cases.
Numerical Sound Speed vs. Theoretical Expression
C0 C1 C4 C5 Comparison with
Case c2 from Eq (5)
theoretical value
1 0 0 -1 -1 c = cREF
2 0 0 -1 -1 c = cREF
3 E0( - 1) E0( - 1) -1 -1 c = cREF
4 E0( - 1) E0( - 1) -1 -1 c = cREF
For each of the four formulations, the computed sound speed by RADIOSS is the same as the
theoretical one. Time step and cycle number are also not affected.

Example 44 - Blow Molding with AMS
Summary
The aim of this example is to introduce high quality time step control Advanced Mass Scaling
(AMS). Time step will be computed by RADIOSS. Small element sizes may lead to small time step
and; therefore, occupy many CPU sources. Increase time step could use time step control, but
using old option of time step control will for example increase the mass or kinematic energy. If
the increase is not small enough, it will affect the solution, but with this high quality time step
control AMS, there is no change in inertia effects on translational global acceleration, non-
diagonal mass added. With AMS similar results are received, like the old one, but with much less
computation time.

Title
Blow Molding with
AMS
Number
44.1
Brief Description
Blow molding with Advanced Mass Scaling (AMS).
Keywords
 Advanced Mass Scaling (/AMS)
 Time Step for Advanced Mass Scaling (/DT/AMS/Iflag)
 Type 7 interface (/INTER/TYPE7)
 Visco Elastic Plastic Piecewise Linear Material law (/MAT/LAW66)
 Shell property (/PROP/SHELL)
RADIOSS Options
 Boundary condition (/BCS)
 Impose displacement (/IMPDISP)
 Pressure Load (/PLOAD)
Input File
Example44:
<install_directory>/demos/hwsolvers/radioss/44_blow_molding_ams/E4_66_AMS/*
<install_directory>/demos/hwsolvers/radioss/44_blow_molding_ams/E4_66_no_dt_
control/*
<install_directory>/demos/hwsolvers/radioss/44_blow_molding_ams/E4_66_Noda_CST/*

Advanced

Overview
Description of the Physical Problem

A hollow plastic parison (tube-like) has been formed. Then the parison is clamped into a mold
and air is pumped into it. Here pressure load is used to model air pressure. Let it push the plastic
out and then match the mold. The dimension of the parison is cylinder with 30mm and its
thickness 2mm. The dimension of the mold is 207mm x 120mm and its thickness is 1.0 mm.
Fig 1: Problem description for blow molding
Units: mm, s, Mg , N , MPa

The mold material using the Elastic model (/MAT/ELAST), with the following characteristics:
 Initial density = 7.8e-9 Mg/mm3
 Young modulus = 200000 MPa
 Poisson ratio = 0.3
The plastic parison using Visco Elastic Plastic Piecewise Linear material (/MAT/LAW66), with the
 Initial density = 1e-9 Mg/mm3
 Young modulus = 4 MPa

Modeling Methodology of Blow Model with AMS

Blow molding using AMS will be modeled as follows:
1. Define /AMS in Starter. Select the part group which will use AMS. If the part group has not
been specified, then the whole model will use AMS.
2. Use /DT/AMS in Engine. For example:
/DT/AMS
0.67 1.15e-4

Rigid body and Boundary conditions:
 Using rigid body, two molds have been fixed in all direction of rotation and translations of
y-direction and x-direction. They are only free in z-direction (translation).
Pressure Load:
 The air pressure on the plastic parison is modeled using pressure load /PLOAD from inside
towards outside
Fig. 2: Pressure load on plastic parison
Imposed displacement:
 Two molds are moved in opposite directions with imposed displacement.
Interface:
 Type 7 interface has been defined between mold and plastic parison with friction 0.7.

The following figures show the plastic strain, von Mises stress on plastic parison. (See below).
Fig. 3: Plastic strain and von Mises stress on plastic parison
Performance
Using the AMS technique, CPU time is reduced by a factor of approximately 3, in this case.
Below shows results comparison of tests:
 Without time step control (no mass scaling)
 With standard mass scaling /DT/NODE/CST
 With AMS

Table 1: Results of model computation with and without AMS
With standard mass

Without time step
scaling With AMS
control
/DT/NODA/CST
Time step(s) 1.15e-4 0.34e-04 1.15e-4
Total Number of 78200 24280 6966

cycle
CPU time(s) 2027.82 723.02 522.83
Speed-up
- 2.80 3.88
Results quality
- Bad Good
Fig. 4: Plastic strain for tests without time step control (no mass scaling). With /DT/NODA/CST and with AMS at time
0.4s.

Fig. 5: Internal energy on plastic parison with and without AMS
It shows at time 0.4s for the same speed up factor with AMS you get more accurate results
compare with no mass scaling test than with node mass scaling.
Conclusion
To obtain a CPU saving factor of about 3, the target time step should be about 10 times higher
than the one without AMS; AMS treatment itself is taking some CPU cost.
Standard mass scaling technique can also speed up the calculation by a factor of about 3, but the
results quality will be affected.
In general, AMS technique for a given speed up, gives more accurate results than standard mass
scaling.
The AMS technique does not change the total mass; the mass is added only on non-diagonal
terms of the mass matrix.
It is applicable to the entire model.
There is no change in inertia effects on translational global acceleration
Note:
 Result accuracy, in terms of stress and strains, is normally not affected; by the way AMS is
affecting Eigen modes of the structure(s) to which it is applied. Higher frequencies are
lowered.
 AMS technique is highly scalable; large models could show even more significant speed up
factors.

Example 45 - Multi-Domain
Summary
The Multi-Domain technique aims at optimizing performance of large scale RADIOSS models
containing one sub-domain with significant time step discrepancy, often related to mesh
refinement differences. It makes it possible to split a whole model into master domain and
several sub-domains. Each domain is computed as a separate RADIOSS model, using its own
timestep. The force and momentum transfers between them are managed by a separate program
insuring stability constraints. The aim of this example is to show how to use the new Multi-
Domain Single Input Format and how to prepare a model.

Title
Multi-Domain
Number
45.1
Brief Description
Separate the whole model into master domain and sub-domain and solve each one with its own
timestep. The new Multi-Domain Single Input Format makes the sub-domain part definition with
the /SUBDOMAIN keyword.
Keywords
 Multi-Domain Decomposition (/SUBDOMAIN)
 Multi-Domain Coupling (/RAD2RAD/ON)
 /INTER/TYPE2
Input File
FRAME_MODIFIED: <install_directory>/demos/hwsolvers/radioss/45-
multidomain_tied/monodomain/*
<install_directory>/demos/hwsolvers/radioss/45-multidomain_tied/multidomain/*

Advanced

Overview
Description of the Physical Problem

The bumper beam impacts the rigid cylinder (254 mm diameter) with an initial velocity of 5m/s.
The crash box behind the bumper is fixed in all directions, except translation in x direction - the
pillar is fixed in all directions.
The pillar is fixed in all directions.
Fig 1: Problem description
Modeling Description
In order to get more accurate results, use a fine mesh in the area of interest (high deformation
location). In this example, the part near the impact area (highlighted in red) where the bumper
highly deforms was chosen. The fine mesh size is about 2 mm, whereas the mesh size of the
remaining part is about 7 mm. The finely meshed part has 18370 shell elements that represent
nearly 50% of the whole model. In explicit analysis the timestep will be controlled by this fine
mesh. In order to improve performances, this computation is usually treated with classical mass
scaling (/DT/NODA/CST). It is possible to use a small timestep for the fine meshed part and a
large timestep for the coarse meshed part using the Multi-Domain technique.
There are two different methods to use the Multi-Domain technique. The first (old) method
requests users to build separate Starter and Engine files for each domain and to define a /LINK
keyword for the connections between domains.
A new feature was introduced in RADIOSS v11.0.220 called Multi-Domain Single Input Format.
Using the keyword /SUBDOMAIN, you can insert the parts, which use a different timestep into
the new sub-domain and create an Engine file with its specific timestep. RADIOSS will
automatically create the Starter information for the sub-domain and its link for connection
between the two domains.

In this example one Starter file with two Engines files are used:
bumper_LL4_0000. Starter file
rad
bumper_LL4_0001. Engine file for master domain
rad
FINE_MESH_0001.r Engine file for sub-domain
ad
Only the following two keywords need to be added:

1. /SUBDOMAIN in bumper_LL4_0000.rad  Starter
With /SUBDOMAIN defining our sub-domain in Starter file, as shown below:
Fig. 2: Parts of the sub-domain
Remark:
Keep the “subdomain_title” in /SUBDOMAIN the same as the sub-domain Engine file
root_name.
In this example, the “subdomain_title” in /SUBDOMAIN is “FINE_MESH”. Therefore, the
Engine sub-domain is “FINE_MESH_0001.rad”.
2. /RAD2RAD/ON must be present in both Engine files:
In order to use the Multi-Domain technique, /RAD2RAD/ON in both Engine files needs to be
defined, so that the Multi-Domain coupling can be established.

Definition of the Contact between Domains
Tied contact (/INTER/TYPE2) between sub-domain and master-domain need to be defined. This
connection is fully compatible with Multi-Domain single input format and automatically creates
the coupling links between domains during RADIOSS Starter run.
In this example, sub-domain will not impact the master domain during the simulation. So, define
two self-contacts for each domain. In case the sub-domain impacts the master domain during the
simulation, it is recommended to define the following four contact interfaces type 7.
 Two internal contact interfaces (/INTER/TYPE7) each one treated in its own domain:
Contact interface 1: self-contact interface for the master domain
Contact interface 2: self-contact interface for the sub-domain
 Two contact interfaces (/INTER/TYPE7) for the interaction between domains treated in the
sub-domain:
Contact interface 3: contact subframe/car – subframe on the slave side
Contact interface 4: contact car/subframe – subframe on the master side
In this case, if all contacts are treated in one single contact (/INTER/TYPE7), every element of
the model would impact the sub-domain and all the elements of the model would be duplicated in
the sub-domain. If the model is large, the Multi-Domain interface would be huge, the CPU cost of
RAD2RAD would be very high; therefore, the performance of the computation is very poor. The
warning message “Multi-Domains interface is too big” will be printed by Starter, in this instance.
Units: mm, ms, kg, N, GPa

In order to show the performance of a Multi-Domain, the following two tests were conducted:
 Test 1: Mono-domain – free DT control
 Test 2: Multi-Domain – free DT control
Table 1: Total calculation time comparison of two tests
Multi-Domain (2)
Test Mono-Domain (1)
free DT control Master domain Sub-domain
free DT control free DT control
CPU 1 1 1
Timestep [ms] 2E-04 8E-04 2E-04
CPU time [s] 5.93E+03 0.67E+03 2.64E+03
Elapsed time [s] 5.9E+03 3.3E+03
For Test 1, the timestep will be controlled by the fine mesh part. In order to avoid small
timesteps, use the Multi-Domain approach (Test 2). The master domain (coarse mesh part) has a
free timestep of about 8e-4 ms and the sub-domain (fine mesh part) has a timestep of about 2e-
4 ms. The total calculation time is only 3.3e-3 s (almost 2 times faster than Test 1). Fig 3 shows
the same failure behavior between Test 1 and Test 2. Fig 4 and Fig 5 show exactly the same
results between Test 1 and 2. Test 2 is faster and provides the same results quality as in Test 1.

Fig. 3: Failure of crash box in the two tests
Fig. 4: Impactor force of the two tests

Fig. 5: Internal energy of the two tests
When using Multi-Domain, the performance will be more significant for large models or for cases
where the fine mesh part represents less than 30% of the whole model (50% here is a lot).
The RADIOSS domains are treated sequentially, which means that only one RADIOSS process is
running at a time. The full CPU resource is automatically allocated to the running process and the
other is put into a no CPU consuming idle mode.
Fig. 6: RAD2RAD multi-timestep method

Conclusion
By using the Multi-Domain approach, the computation time can be reduced significantly, with
preserving high quality results.
The new single input file format makes the Multi-Domain more user friendly, compared to the
set-up process in earlier versions.

Example 46 - TNT Cylinder Expansion Test
Summary
The Cylinder Expansion Test is an experimental test used to characterize the adiabatic expansion
of detonation products. It allows determining JWL EOS parameters.
It consists in a copper cylinder filled with an explosive (TNT). Detonation is initiated at the
bottom of the explosive with a planar detonation wave. It propagates along cylinder axis and
radial expansion of the copper cylinder is measured at a given point of external surface.
In order to simulate this experience, a model is created with the following details:
 3D mesh of a quarter-cylinder with eight node brick elements
 Jones Wilkins Lee Equation-of-State for TNT detonation products (/MAT/JWL)
 Hydrodynamic Johnson-Cook material law for the copper cylinder (/MAT/HYD_JCOOK)
 Multi-Material, Fluid and Explosive material law (/MAT/LAW51)
The simulation results are then compared to the experiment data.

46.1 - Lagrange Formulation
Title
Cylinder
Expansion Test
with Lagrange
formulation
Number
46.1
Brief Description
Detonation is initiated at the bottom of the explosive. Radial expansion of the cylinder is
measured and compared to experimental data.
Keywords
 Lagrange formulation
 Jones Wilkins Lee EOS (/MAT/JWL)
 Hydrodynamic Johnson-Cook Material (/MAT/LAW4)
 Gruneisen equation of state (/EOS/GRUNEISEN)
RADIOSS Options
 Solid property (/PROP/SOLID)
 Detonation plan (/DFS/DETPLAN)
 Time history on node (/TH/NODE)
Input File
Cylinder Test:
<install_directory>/demos/hwsolvers/radioss/46_TNT_Cylinder_Expansion_Test/
Lagrange/*

Advanced

Overview

The purpose of this example is to show how to simulate the cylinder expansion test and compare
the simulation result to experimental data.
A OFHC copper cylinder (1.53cm diameter, 0.26cm thickness, 30.5cm height) is filled with an
explosive (TNT). Detonation is initiated at the bottom of the explosive. Radial expansion is
measured at a length of 8*D cm.
Since this problem is axisymmetric, only a quarter of the cylinder is modeled.
Fig 1: Problem description for cylinder test
Units: cm, µs, g, Mbar

The TNT material uses Jones Wilkins-Lee Material (/MAT/JWL) and Lagrange formulation with the
 Initial density = 1.63
 A = 3.7121
 B = 0.0323
 R1 = 4.15
 R2 = 0.95
 = 0.3
Chapman Jouget parameters enable detonation time to compute and burn fraction evolution:
 Detonation velocity D = 0.693
 Chapman Jouguet pressure PCJ = 0.21
 Detonation energy E0 = 0.07

RADIOSS Card (TNT)
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/MAT/JWL/2
TNT
# RHO_I
1.63 0
# A B R1 R2 OMEGA
3.7121 .0323 4.15 .95 .3
# D P_CJ E0
.693 .21 .07
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
Using Hydrodynamic Johnson-Cook material law (/MAT/LAW4), the copper cylinder material has
the following characteristics:
 E-Module = 1.24
 Poisson = 0.35
 A = 0.9e-3
 B = 0.292e-2
 N = 0.31
 max = 0.0066
 C = 0.025
 0 = 1e-5
 M = 1.09
 0Cp = 3.461e-3
 Tmelt = 1656
The Gruneisen equation of state (/EOS/GRUNEISEN) is used for copper with the following
characteristics:
 C = 0.394
 S1 = 1.489
 0 = 1.97
 a = 0.47
 E0 = 8.96

RADIOSS Card (Copper)
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/MAT/HYD_JCOOK/1
Copper
# RHO_I
8.96
# E0 nu
1.24 .35
# A B n epsmax sigmax
.9E-03 .292E-02 .31 0 0.0066
# Pmin
-1.E30
# C EPS_DOT)0 M Tmelt Tmax
.25E-01 .1E-05 1.09 1656.0 1e30
# RHOCP
.3461E-04
/EOS/GRUNEISEN/1
Copper
# C S1 S2 S3
.394 1.489 0 0
# GAMMA0 ALPHA E0 RHO_0
1.97 .47 0 8.96
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|

Modeling methodology
A 3D mesh is made of brick elements. The element size is approximately of 0.035 cm x 0.035 cm
x 0.035 cm.
The mesh is dragged along the z direction (z = 30.5 cm). It is important to have no discontinuity
in element volume in order to ensure a good propagation of detonation wave and shock wave.
Fig 2: Model mesh

Due to the symmetries of the model, a quarter of the cylinder is modeled. Boundary conditions
are set on the yOz plan at x = 0 (Tx = 0) and on the xOz plan at y = 0 (Ty = 0) to simulate the
symmetry.
A planar detonation wave is defined at the bottom of the cylinder.
In order to plot the curve of radial expansion, displacements of node n 201 520 at z = 24.48 cm
on the outer wall of the copper cylinder are saved in time history. It corresponds to L/D=8 in
agreement with experimental protocol.
A scale factor of 0.5 (on time step for all elements) is used for this type of application.

In solid properties, qa and qb default values are used. These values have to be changed
depending of the formulation (ALE, Euler).
 Isolid is set to 14 for copper solid properties.
RADIOSS Card (TNT)

#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/PROP/SOLID/2
TNT
# Isolid Ismstr Icpre Inpts Itetra Iframe dn
0 0 0 0 0 0 0
# q_a q_b h LAMBDA_V MU_V
0 0 0 0 0
# dt_min istrain IHKT
0 0 0
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|

#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/PROP/SOLID/1
Copper
0 0 0 0 0 0 0
0 0 0 0 0
0 0 0
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|


The two following diagrams display the pressure and density in the cylinder and the explosive.
Fig. 3: Pressure distributed in copper and TNT at time = 13µs.

Fig. 4: Density distributed in copper and TNT at time = 13 µs.
The following diagram shows the comparison between the experimental and simulation
measurement of radial expansion.
Fig. 5: Comparison between experimental results and simulation results

Conclusion
Good correlation between experimental and simulation results. A thinner meshing could improve
the correlation between simulation and experimental curves.
Elapsed time for simulation: t = 11 441 s, 8514 cycles, (4 cpu intel core i7 Q 840 @ 1.87 GHz).
As the model is Lagrangian, the mesh becomes very distorted at the end of the simulation to
obtain a proper mesh, it is possible to use the Euler method.
References
[1] Adiabatic Expansion of high explosive detonation products, LANL, Wilkins (1969)
[2] A Constitutive model and data for metals subjected to large strains, high strain rates and high
temperatures, Gordon R. Johnson, William H. Cook

46.2 - Euler Formulation
Title
Cylinder Expansion
Test with Euler
formulation
Number
46.2
Brief Description
Detonation is initiated at the bottom of the explosive. Radial expansion of the cylinder is
measured and compared to experimental data.
Keywords
 Euler formulation
 Multi-Material, Fluid and Explosive material law (/MAT/LAW51)
RADIOSS Options
 Detonation plan (/DFS/DETPLAN)
Input File
Cylinder Test:
<install_directory>/demos/hwsolvers/radioss/46_TNT_Cylinder_Expansion_Test/Euler/*

Advanced

Overview

The purpose of this example is to show how to simulate the cylinder expansion test and compare
the simulation result to experimental data.
A OFHC copper cylinder (1.53cm diameter, 0.26cm thickness, 30.5cm height) is filled with an
explosive (TNT). Detonation is initiated at the bottom of the explosive. Radial expansion is
measured at a length of 8*D cm. With an Euler formulation, the air has to be modeled to
measure radial expansion.
Since this problem is axisymmetric, only a quarter of the cylinder is modeled.
Fig 1: Problem description for cylinder test

Using the Multi-Material Solid, Liquid, and Gas material law (/MAT/LAW51), the copper cylinder
material has the following characteristics:
 Initial pressure C0mat1 = 1e-6

 Hydrodynamic coefficients:
 C mat1 = 1.38
1
 C2mat1 = 1.372
 C4mat1 = 0.87
 C5mat1 = 0.87
 Elastic shear modulus G1mat1 = 0.519
 Yield stress amat1 = 0.9e-3
 Plastic yield factor bmat1 = 0.292e-2

 Plastic yield exponent nmat1 = 0.31
 Plastic strain rate factor cmat1 = 0.025
 Plastic reference strain rate = 1e-6
 Thermal exponent mmat1 = 1.09

 Specific heat Rhocvmat1 = 3.461e-5
 Tmelt = 1656

#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/MAT/LAW51/1
Copper
# Iform
10
# P_ext NU Nu_Vol
0 0 0
# ALPHA0_mat1 RHO0_mat1 E0_mat1 Pmin_mat1 C0_mat1
1 8.96 0 0 1E-6
# C1_mat1 C2_mat1 C3_mat1 C4_mat1 C5_mat1
1.38 1.372 0 .87 .87
# G1_mat1 a_mat1 b_mat1 n_mat1
.519 9E-4 .00292 .31
# c_mat1 EPSILON_DOT0_mat1
.025 1E-6
# m_mat1 T0_mat1 Tmelt_mat1 Tlim_mat1 Rhocv_mat1
1.09 0 1656 0 3.461E -5
# Epspmax_mat1 sigma_max_mat1 KA_mat1 KB_mat1
0 0 0 0
0 .0012 2.5E-6 -1E-20 0
0 0 0 .4 .4
0 0 0 0
0 0
0 0 0 0 0
0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0
0 0
0 0 0 0 0
0 0 0 0

0 1.63 .07 -1E-20 1E-6
# A B R1 R2 W
3.712 .0323 4.15 .95 .3
# D PCJ C1_mat4
.693 .21 .036
/EULER/MAT/1
# Modif. factor.
0
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
Using Multi-Material Solid, Liquid, and Gas material law (/MAT/LAW51), the TNT material has the
 Explosive cavitation pressure Pminmat4 = -1e-20
 Initial explosive pressure C0mat4 = 1e-6
 Explosive coefficient B1 = 3.712
 Explosive coefficient B2 = 0.0323
 Explosive coefficient R1 = 4.15
 Explosive coefficient R2 = 0.95
 Explosive coefficient = 0.3
 Explosive coefficient C1mat4 = 0.036
Chapman Jouget parameters compute detonation time and burn fraction evolution:
 Detonation velocity D = 0.693
 Chapman Jouguet pressure PCJ = 0.21
 Initial explosive energy per unit initial volume E0mat4 = 0.07
RADIOSS Card (TNT)

#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/MAT/LAW51/2
TNT
# Iform
10
# P_ext NU Nu_Vol
0 0 0
0 8.96 0 0 1E-6
1.38 1.372 0 .87 .87

.519 9E-4 .00292 .31
.025 1E-6
1.09 0 1656 0 3 .461E-5
0 0 0 0
0 .0012 2.5E-6 -1E-20 0
0 0 0 .4 .4
0 0 0 0
0 0
0 0 0 0 0
0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0
0 0
0 0 0 0 0
0 0 0 0
1 1.63 .07 -1E-20 1E-6
# A B R1 R2 W
3.712 .0323 4.15 .95 .3
# D PCJ C1_mat4
.693 .21 .036
/EULER/MAT/2
# Modif. factor.
0
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|

Using the Multi-Material Solid, Liquid, and Gas material law (/MAT/LAW51), the Air material has
the following characteristics:
 Initial energy per unit initial volume E 0mat2 = 2.5e-6
 Hydrodynamic cavitation pressure Pminmat2 = -1e-20
 Hydrodynamic coefficient C4mat2 = 0.4
 Hydrodynamic coefficient C5mat2 = 0.4
RADIOSS Card (Air)

#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/MAT/LAW51/3
Air
# Iform
10
# P_ext NU Nu_Vol
0 0 0
0 8.96 0 0 1E-6
1.38 1.372 0 .87 .87
.519 9E-4 .00292 .31
.025 1E-6
1.09 0 1656 0 3.461E -5
0 0 0 0
1 .0012 2.5E-6 -1E-20 0
0 0 0 .4 .4
0 0 0 0
0 0
0 0 0 0 0
0 0 0 0
0 0 0 0 0

0 0 0 0 0
0 0 0 0
0 0
0 0 0 0 0
0 0 0 0
0 1.63 .07 -1E-20 1E-6
# A B R1 R2 W
3.712 .0323 4.15 .95 .3
# D PCJ C1_mat4
.693 .21 .036
/EULER/MAT/3
# Modif. factor.
0
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
Using the Multi-Material Solid, Liquid, and Gas material law (/MAT/LAW51), the Boundary
material has the following characteristics:
RADIOSS Card (Boundary)

#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/MAT/LAW51/4
Boundary
# Iform
3
# ALPHA_1 RHO_01 E_01 P_min1 C_01
0 8.96 0 0 1E-6
1 .0012 2.5E-6 -1E-20 1E-6
0 0 0 0 0
/EULER/MAT/4
# Modif. factor.
0
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|

A 3D mesh is made of brick elements. The element size for the copper cylinder is approximately
of 0.035 cm x 0.035 cm x 0.035 cm.
The mesh is dragged along the z direction (z = 30.5 cm). It is important to have no discontinuity
in element volume in order to ensure a good propagation of detonation wave and shock wave.
Fig 2: Model mesh

Due to the symmetries of the model, a quarter of the cylinder is modeled. Boundary conditions
are set on the yOz plan at x = 0 (Tx = 0) and on the xOz plan at y = 0 (Ty = 0) to simulate the
symmetry.
A planar detonation wave is defined at the bottom of the cylinder. A scale factor of 0.5 (on time
step for all elements) is used for this type of application.
In solid properties, qa =1.1 and qb =0.05. These values must be set to 0 for the Lagrange
formulation.
 Isolid is set to 0 for TNT and copper solid properties.

RADIOSS Card (TNT)
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/PROP/SOLID/2
copper
0 0 0 0 0 0 0
1.1 .05 0 0 0
0 0 0
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|

#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/PROP/SOLID/1
TNT
0 0 0 0 0 0 0
1.1 .05 0 0 0
0 0 0
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|


The following figure displays the variation of density in the cylinder, explosive and air.
Fig. 3: Density distributed in copper and TNT at time = 33µs.
The following diagram shows the comparison between the experimental and simulation
measurement of radial expansion. The displacement values are estimated on the animations
using the density contour.

Fig. 4: Comparison between experimental results and simulation results
Conclusion
Good correlation between experimental and simulation results. A thinner meshing could improve
the correlation between simulation and experimental curves.
References
[1] Adiabatic Expansion of high explosive detonation products, LANL, Wilkins (1969)
[2] A Constitutive model and data for metals subjected to large strains, high strain rates and high
temperatures, Gordon R. Johnson, William H. Cook

Example 47 - Concrete Validation
Summary
RADIOSS includes the material model CONC to model concrete failure modeling under
compression and tension.
Three kinds of tests are performed in this example:
 Uniaxial tests (uniaxial compression and uniaxial tension) where experimental results have
been used to calibrate the model reference
 Multi-axial tests to evaluate the simulation/experiment correlation
 Cyclic tests to illustrate the right behavior of the model used
In order to simulate this experience, a model is created with the following details:
 A one element cube with eight node brick elements
 Concrete material law (/MAT/LAW24)
The simulation results are then compared to the experiment data.

Title
Concrete Validation
Number
47.1
Brief Description
Three kinds of tests are performed in order to evaluate the simulation/experiment correlation and
to illustrate the good behavior of the model used.
Keywords
 Concrete material law (/MAT/LAW24)
RADIOSS Options
 Pressure load (/PLOAD)
Input File
Concrete Failure: <install_directory>/demos/hwsolvers/radioss/47_concrete_test/*

Advanced

Overview

The purpose of this example is to compare the simulation results to experimental data.
A concrete cube is subjected to various tests:
Kupfer Tests [2]
1.1: C000 - Uniaxial compression principle stress 1 = -1; 2 = 0; 3 =0

1.2: T000 - Uniaxial tension principle stress 1 = 0; 2 = 0; 3 =1
1.3: CC00 - Biaxial compression principle stress 1 = -1; 2 = -1; 3 =0
2.1: CC01 - Compression/Compression principle stress 1 = 0.52; 2 = 0; 3 = -1
2.2: TC01 - Compression/Tension principle stress 1 = -.052; 2 = 0; 3 = -1
2.3: TC02 - Compression/Tension principle stress 1 = 0.102; 2 = 0; 3 = -1
2.4: TC03 - Compression/Tension principle stress 1 = 0.204; 2 = 0; 3 = -1
Units: mm, ms, mg, MPa

The Concrete material law (/MAT/LAW24) has the following characteristics:
 Initial density = 0.0022 mg/mm3
 Concrete elasticity Young’s modulus Ec = 31700 MPa
 Poisson’s ratio = 0.22

 Concrete uniaxial compression strength fc = 32.22 MPa
 Concrete biaxial strength fb/fc = 1.15
 Concrete confined strength f2/fc = 4.2
 Concrete confining stress so/fc = 0.8
 Concrete plasticity initial value of hardening parameter ky = 0.35
 Concrete plasticity dilatancy factor at yield α y = -0.6
 Concrete plasticity dilatancy factor at failure α f = -0.2

RADIOSS Card (Concrete)
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/MAT/CONC/1
Concrete
# RHO_I
.0022
# E_c NU
31700 .22
# fc ft_on_fc fb_on_fc f2_on_fc s0_on_fc
32.22 0 1.15 4.2 .8
# H_t D_sup EPS_max
0 0 0
# k_y r_t r_c H_bp
.35 0 0 0
# ALPHA_y ALPHA_f V_max
-.6 .2 0
# f_k f_0 H_v0
0 0 0
# E sigma_y E_t
0 0 0
# ALPHA1 ALPHA2 ALPHA3
0 0 0
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
Schickert and Winkler Tests

3.1: TRX1 - Meridian Compression m= 51 MPa; 1 = 0.5; 2 = 0.5; 3 = -1
3.2: TRX0 - Meridian Shear m = 51 MPa; 1 = 1; 2 = 0; 3 = -1
3.3: TRX2 - Meridian Tension m = 51 MPa; 1 = 2; 2 = -1; 3 = -1

 Concrete elasticity Young’s modulus E c = 23000 MPa
 Poisson’s ratio = 0.19

 Concrete uniaxial compression strength fc = 30.6 MPa
 Concrete data tensile tangent modulus Ht = -31700
 Concrete plasticity initial value of hardening parameter ky = 0.35
 Concrete plasticity base plastic modulus H bp = 29170
 Concrete plasticity dilatancy factor at yield α y = -0.6
 Concrete plasticity dilatancy factor at failure αf = -0.2

#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/MAT/CONC/1
Concrete
# RHO_I
.0022
# E_c NU
23000 .19
30.6 0 0 0 0
# H_t D_sup EPS_max
-31700 0 0
# k_y r_t r_c H_bp
.35 0 0 29710
-.6 .2 0
# f_k f_0 H_v0
0 0 0
# E sigma_y E_t
0 0 0
0 0 0
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
Cyclic Tests
4.1: BBX0 Tension-Compression-Tension Cycle without Reinforcement

 Young modulus for concrete: E c =57600 MPa
 Poisson’s ratio: = 0.25

 Compressive strength of concrete: fc = 35.60 MPa

#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/MAT/CONC/1
Concrete
# RHO_I
.0022
# E_c NU
57600 .25
35.60 0 0 0 0
# H_t D_sup EPS_max
0 0 0
# k_y r_t r_c H_bp
0 0 0 0
0 0 0
# f_k f_0 H_v0
0 0 0
# E sigma_y E_t
0 0 0
0 0 0
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
4.2: BBX1 Tension-Compression-Tension Cycle with Reinforcement

 Young modulus for concrete: E c =57600 MPa
 Poisson’s ratio: = 0.25

 Compressive strength of concrete: fc = 35.60 MPa
 Reinforcement percentage: α3 = 1%
 Young modulus for steel: Ec = 210000 MPa
 Yield stress for steel: y = 500 MPa

#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/MAT/CONC/1
Concrete
# RHO_I
.0022
# E_c NU
57600 .25
35.6 0 0 0 0
# H_t D_sup EPS_max
0 0 0
# k_y r_t r_c H_bp
0 0 0 0
0 0 0
# f_k f_0 H_v0
0 0 0
# E sigma_y E_t
210000 500 0
0 0 .01
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
The results for each test are:

 Stress in various directions as a function of deformations
 Von Mises stress as function of Pressure

A 10 mm cube is modeled with a one element brick.
Fig 1: Geometry of the cube

Boundary conditions depend on the test.
A scale factor of 0.1 (on time step for all elements) is used for “single elements models”.
Solid properties are:
 qa =1.1 and qb =0.05 (default values)
 Isolid = 1
 Iframe = 2 (co-rotational formulation)
 Istrain is set to 1 (to post-treat strains).
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|
/PROP/SOLID/1
Concrete
# Isolid Ismstr Icpre Inpts Irot Iframe dn
1 0 0 0 0 2 0
# q_a q_b h
1.1 0.05 0
# dt_min Istrain
0 1
#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----

Curves
Test C000: Uniaxial Compression
The X displacement is fixed for nodes 2, 3, 6 and 7. A negative displacement is applied on the
face defined by nodes 1, 4, 5 and 8.
Fig 2: Uniaxial compression with RADIOSS (blue curves) and experiment (red curves)
Comments
1. The stress/strain curve is made of three line segments.
2. After failure, the behavior obtained with RADIOSS curves (left) is perfectly plastic whereas
there is experimentally a softening phenomenon (right).
3. The yield stress is obtained at σ = 0.337 f c for theoretical, numerical and experimental
curves.

Test T000: Uniaxial Tension
The X displacement is fixed for nodes 2, 3, 6 and 7. A positive displacement is applied on the
face defined by nodes 1, 4, 5 and 8.
Fig 3: Uniaxial Tension with RADIOSS (blue curves) and experiment (red curves)
Comments
1. Failure is modeled by stress and elastic modulus softening.
2. On the RADIOSS curve there is a residual stiffness in concrete after the softening: (1-Dsup)E
3. Dsup is set to 0.9 (default value = 0.99999).

Test CC00: Biaxial Compression
The X displacement is fixed on nodes 2, 3, 6 and 7. The Y displacement is fixed on nodes 3, 4, 7
and 8.
A displacement is applied in X direction on the face described by nodes 1, 4, 5 and 8 and in Y
direction on the face described by nodes 1, 2, 5 and 6.
Fig 4: Biaxial Compression with RADIOSS (blue curves) and experiment (red curves)
Comments
1. The yield stress is obtained at σ=0.197 fc for theoretical, numerical and experimental curves.
2. Failure mode is similar to uniaxial compression.

Test CC01 Compression/Compression
The X displacement is fixed on nodes 2, 3, 6 and 7. The Z displacement is fixed on nodes 5, 6, 7
and 8.
A pressure load is applied in X direction on the face described by nodes 1, 4, 5 and 8 and in Z
Fig 5: Compression/Compression with RADIOSS (blue curves) and experiment (red curves)
Comments
1. Theoretical yield strength: 0.288 fc
2. Theoretical failure: 1.926 fc
3. Experimental failure: 1.22 fc
4. Theoretical and numerical results are the same, but different from experimental results; linear
interpolation between the traction meridian and the compression meridian is too coarse for
small confinement.

TC01 Compression/Tension
X displacement is fixed on nodes 2, 3, 6 and 7.
Z displacement is fixed on nodes 5, 6, 7 and 8.
A pressure load is applied in X direction on the face described by nodes 1, 4, 5 and 8, and in Z
Fig 6: Compression/Tension with RADIOSS (blue curves) and experiment (red curves)
Comments
4. Theoretical, numerical and experimental results are the same.

and 8.
Comments
4. Theoretical and numerical results are the same, but slightly different from experimental
results.

and 8.
Comments
4. Theoretical and numerical results are the same, but different from experimental results.

Test TRX00: Triaxial Meridian Shear
and 8. The Z displacement is fixed on nodes 5, 6, 7 and 8.
A hydrostatic pressure of 51 MPa is applied on the sample.
Fig 9: Triaxial Meridian Shear with RADIOSS (blue curves) and experiment (red curves)
Comments
1. Theoretical yield strength: 64.3 MPa
2. Theoretical failure: 88.9 MPa
3. Experimental failure: 93 MPa

Test TRX01: Triaxial Meridian Compression
A hydrostatic pressure of 51 MPa is applied on the sample.
A pressure load is applied in X direction on the face described by nodes 1, 4, 5 and 8, in Y
direction on the face described by nodes 1, 2, 5 and 6, and in Z direction on the face described
by nodes 1, 2, 3 and 4.
Fig 10: Triaxial Meridian Compression with RADIOSS (blue curves) and experiment (red curves)
Comments
4. The behavior of the model under hydrostatic loading is elastic, whereas there are non-
linearities experimentally.

Test TRX02: Triaxial Meridian Compression
An hydrostatic pressure of 51 MPa is applied on the sample.
A pressure load is applied in X direction on the face described by nodes 1, 4, 5 and 8, in Y
direction on the face described by nodes 1, 2, 5 and 6, and in Z direction on the face described
by nodes 1, 2, 3 and 4.
Fig 11: Triaxial Meridian Compression with RADIOSS (blue curves) and experiment (red curves)
Comments

Test BBX0: Tension-Compression-Tension Cycle without Reinforcement
A displacement is applied in the Z direction on the face defined by nodes 1, 2, 3 and 4 with a
tension-compression-tension cycle as shown below:
Fig 12: Velocity imposed
Fig 13: Cycle without reinforcement

This test illustrates the behavior of the model LAW24:
 Failure, damage and cracks reopening
 Cracks reclosing
 Plastification and hardening
 Compressive strength
 Residual plastic deformation
Test BBX1: Tension-Compression-Tension Cycle with Reinforcement

A velocity is set among the Z direction on the face defined by nodes 1, 2, 3 and 4 with a tension-
compression-tension cycle, as shown below:
Fig 14: Cycle with reinforcement
Comments
1. Steel reinforcement improves compressive and tensile strength by 5 MPa compared to the
same model without steel reinforcement.

Von Mises / Pressure Curves
The following von Mises/Pressure curves show the different areas described by tensile and
compressive tests.
Each orange point corresponds to the failure point of the model considered.
fc is the uniaxial compressive strength.
Fig 15: von Mises/Pressure curves
Conclusion
Under complex loading, concrete mechanic behavior between RADIOSS simulation, theory and
experiments are demonstrated. With three kinds of tests, the mechanic behavior of concrete can
be well characterized using LAW24.
References
[1] A non-uniform hardening plasticity model for concrete materials, Mechanics of Materials, D.J.
Han and W.F. Chen, 1984.
[2] Behavior of Concrete under Biaxial Stresses, Journal of the Engineering Mechanics Division,
ASCE, V. 99, No. 4, pp. 853-866u, LKupfer, B., and Gerstle, K., 1973.

Example 48 - Solid Spotweld
Summary
The aim of this example is to introduce solid element modeling for spotweld connection.

Title
Solid Spotweld
Number
48.1
Brief Description
Solid spotweld defined with /MAT/LAW59 and /FAIL/CONNECT connect two metal sheets with tied
contact.
Keywords
 Shell element (for sheet metal)
 Tied contact interface /INTER/TYPE2 (between solid spotweld and sheet metal)
 Material law /MAT/LAW59 and failure model /FAIL/CONNECT for solid spotweld
 Solid element property for connect material /PROP/TYPE43 for solid spotweld
 Johnson-Cook elasto-plastic material law /MAT/LAW2 for sheet metal
RADIOSS Options
 Boundary condition /BCS
 Rigid wall /RWALL
Input File
Frame Modified: <install_directory>/demos/hwsolvers/radioss/48_solid_spotweld/*

Advanced

Overview

A steel crash box with solid spotweld, fixed at one end, is impacted at the other end by a rigid
wall. The dimensions of the crash box are 333.6 mm x 122.7 mm x 68.28 mm, and its thickness
is 1.8 mm.
Fig 1: Problem description and beam cross section
Units: mm, s, Mg , N , MPa

Using the Johnson-Cook plasticity model (/MAT/LAW2), the sheet metal material has the
 Initial density = 7.8e-9 [Mg/mm3]
 Young modulus = 210000 [MPa]
 Yield stress = 250 [MPa]
 Hardening parameter = 350 [MPa]
 Hardening exponent = 0.3
In this connection material (/MAT/LAW59) use stress - plastic strain curve to describe the
material characters of spot-weld. With failure model (/FAIL/CONNECT) it is possible to use two
different failure criteria to describe the failure of solid spot. In this example, use the following
characteristics:
 Initial density = 7.8e-9 [Mg/mm3]
 Young modulus = 210000 [MPa]
 Failure relative displacement in normal direction = 1.0
 Failure relative displacement (elongation) in tangential plane = 1.8

Fig 2: Material curve of sheet metal and solid spotweld.
In this simple example a linear behavior of spotweld has just been assumed. For more accurate
results, you can also put nonlinear behavior of spotweld through real physic test and CAE
validation.

Modeling Methodology of Solid Spotweld
Solid spotweld will be modeled as follows:

 /MAT/LAW59 to define the material
 /FAIL/CONNECT to define the rupture criteria
 /PROP/TYPE43 to define the spotweld property, which should be only used with
/MAT/LAW59
 /INTER/TYPE2 to define the connection between spotweld and sheet metal
Comments
1. In this spotweld modeling surface 1 (Node 1, 2, 3, and 4) and surface 2 (Node 5, 6, 7, and 8)
are defined. The normal direction has been defined from surface 1 to surface 2.
2. Surface 1 and 2 for each solid spotweld should always connect with a shell element. If one
surface of one solid spotweld is not connected with a shell element, then the stiffness will not
be computed correctly and an error may occur.
Rigid body and Boundary conditions:

 One end of the crash box is modeled using a rigid body and this rigid body is completely
fixed using translations and rotations.
Rigid wall:
 The impactor is modeled using a moving rigid wall having a fixed velocity (5 m/s) in a X
direction and is fixed for other translations and rotations.
Interface:
 The structure interaction is modeled using a type 7 interface on the full structure with self-
impact, where the master surface is defined using the complete model and the slave nodes
group is defined using the master surface.

With /ANIM/BRICK/TENS/STRAIN and /ANIM/BRICK/TENS/STRESS you get the strain and stress
results of the solid spotweld (see below).
Fig. 3: Strain of solid spotweld
Fig. 4: Stress of solid spotweld
Here in the model you define max. relative displacement in normal direction EPS_N_MAX=1.0 in
option /FAIL/CONNECT. Therefore, two solid spotwelds (circled in red) failed after reaching this
criterion.

Fig. 5: Strain of solid spotweld (display erode element in HyperView)
Fig. 6: Force vs displacement of crashbox
Performance
Compared with spring beam spotweld molding, the following performance could be observed:
If you use this solid spotweld modeling, it shows less sheet metal mesh size dependence
compared with spring beam element.
 With coarse sheet metal mesh size, you got similar deformation and similar Force vs
Displacement curve for solid spotweld modeling and spring beam spotweld modeling.

 Once you remesh the sheet metal shell element with fine mesh size (for example with
2.5mm mesh size here), then you still get similar deformation and Force vs Displacement
curve in solid spotweld modeling but not in spring beam spotweld modeling. This means
less sheet metal mesh size dependence for this new solid spotweld modeling.
 The solid spotweld element is time step is free. The element stability is assured by its nodal
connection. The node of the solid element must be connected (tied interface, common
nodes, and rigid connection).
 The solid spotweld element height (length in local Z direction) can be null and the spotweld
results are independent from its height.
 Strain rate of solid spotweld can be taken into account.

Example 49 - Bird Strike on Windshield
Summary
Birds striking an aircraft windshield is a very important topic, for safety reasons and also
financially. The aim of this example is to introduce how to simulate a bird hitting a windshield.

Title
Bird Strike on
Windshield
Number
49.1
Brief Description
Using SPHCEL to simulate a bird hitting a windshield.
Keywords
 /SPHGLO
 /SPHCEL
 /MAT/HYD_VISC
 /MAT/PLAS_TAB
 /MAT/PLAS_BRIT
 /PROP/SPH
 /PROP/SH_SANDW
 /INIVEL/TRA
RADIOSS Options
 Rigid Body (/RBODY)
 Interface (/INTER/TYPE7 and /INTER/TYPE2)
Input File
Bird Strike on Windshield:
<install_directory>/demos/hwsolvers/radioss/49_bird_strike_on_windshield/*

Advanced

Overview
Description of the physical problem

A bird weighing about 4 pounds (1.93kg) with an initial speed of 80m/s impacts a windshield of a
plane. The bird model is extremely simplified with one cylinder (diameter: 106mm; height:
195mm) but allows getting familiar with the major options used in a bird strike analysis. Fix the
edge of the plane model.
Units: mm, ms, kg, KN, GPa

A bird with the following characteristics:
The bird uses material law 6 which is a simple hydro-viscous law based on water properties
/MAT/HYDRO.
 Initial density = 9.82242e-7 [kg/mm3]
 C1 = 2.24
with SPH cells property /PROP/SPH
 mp = 1.725149e-4 [kg]
 qa = 2e-30

 qb = 1e-30
 order = 0
 h = 6.286 [mm]
Notes:
1. In property, mp is the mass of each particle. There are 11190 particles in the bird model, so
the bird weighs 11190*1.725149e-4 = 1.93kg (4 pounds).
2. Set qa to 2e-30 and qb to 1e-30 for no damping.
To properly simulate a bird strike, the compression phase of the bird is the most important;
therefore, the initial compression wave should properly propagate through the bird. Energy
conservation is also important, so that the bird communicates the most realistic amount of
energy to the target, thus ensuring that the impact is reasonably conservative.
A low value of qa and qb should be chosen. The behavior of the bird for very low values of the
damping is very similar, and there is no evidence of instability; therefore, negligible values of
qa and qb (that is 1-30 and 2-30) are considered.
3. In property, h is smoothing length:
1 1
 m p 2  3  1.725149 104 2  3
h    6.286 [mm]
    9.82242 107 
   
The smoothing length, h is highly recommended to use the minimal distance between any
particle and its closest neighbor.
4. SPH correction order is used to satisfy the consistency conditions. It is generally to insure a
better representation when the particles are not so well organized than into a hexagonal
compact net. High order value is very expensive. In this example, set order to 0.
For the windshield in this example, use composite material with matrix glass and plastic fabric.
Both use /MAT/PLAS_BRIT (composite with isotropic layers).
Glass matrix:
 Initial density = 1.74e-6 [kg/mm3]
 E-Modules = 61 [GPa]
 Yield stress a = 0.045 [GPa]
 Plasticity hardening b = 0.05
 t1
 = 0.01
 m1
 = 0.03
 f1
 element deleted at =0.06

'
Fabric plastic:
 Initial density = 8e-7 [kg/mm3]
 E-Modules = 3 [Gpa]
 Yield stress a = 1e30 [Gpa]
with /PROP/SH_SANDW
 Thickness of windshield = 9 [mm]
 Number of lays N = 9
 Ithick=1
For metal parts (aluminum) in the airplane, use /MAT/PLAS_TAB.

 Initial density = 2.7E-6 [kg/mm3]
 E-Modules = 73.1 Gpa
For metal parts connection (bolt) in the airplane, use /MAT/PLAS_JOHNS and /PROP/BEAM.
In /MAT/PLAS_JOHNS:
 Initial density = 7.8E-6 [kg/mm3]
 E-Modules = 210 [Gpa]
 Yield stress a = 0.58 [Gpa]

 hardening parameter b =0.6
 hardening exponent n =1
 Strain rate coefficient =0.017
In /PROP/BEAM:
 Area = 50.265 [mm2]
 Iyy = Izz = 201.061 [mm4]
 Ixx = 403.12 [mm4]
 Bird modeling
A 4 pound bird model with 11190 particles, the distributions follow a hexagonal compact
net. The cell distribution must be cubic centered. The diameter for each particle is about
5.0 mm.
Fig 2: Perspective view of the cubic centered faces net

 SPH simulation setup
In order to define the SPH bird, the following keywords are needed:
o Use /PROP/SPH to define the bird SPH cells property.
o Use /SPHCEL to define SPH cells for bird. Use bird part ID in this keyword.
o Use /SPHGLO to defined global parameter for this simulation.
 
o Set sort = 0.25 (default value). sort is a security coefficient which is used when
searching for neighbors, so that for each particle more than the actual neighbors are
found. The computational time can be reduced.
 Contact between SPH bird and structure and impact setup
o Use interface type 7 to simulate the Contact between SPH bird and structure. Set the
structure as master and the bird as slave.
o Set Igap=1 (use variable gap). The gap is 1/2(particle diameter)+1/2(thickness of the
target). This is the physical value for contact.
o The stiffness between bird and structure are quite different. In order to get results more
close to reality. Normally set Istf=0 and Stfac=0.1. This means interface stiffness equals
to 0.1 times stiffness of master side (structure).
o Use /INIVEL/TRA to set bird part with initial velocity 80m/s strike on the structure.
 Connection between the windshield and fuselage
o Use spring beam to simulate the weld connection between windshield and fuselage, and
use tied contact to connect them.
 Other remarks
In order to decrease the size of animation files and get the best display of SPH particles,
use /ANIM/VERS/44 in Engine file.

Figure 3 shows von Mises stress of the windshield at time 3ms ~ 8ms.
Fig. 3: von Mises Stress at time 3ms~8ms

The Interface Force between the bird and the windshield is shown in Fig. 4.
Fig. 4: Interface force between bird and windshield
Conclusion
 The size of the mesh and the number of particle seems to be a good compromise between
the accuracy of the problem and the CPU considerations (especially for larger models such
as 4 and 8 pound birds).
 For interface parameters default values could be used.
 The simplified viscous law is good.
 Impact velocity remains small compared to the sound speed in the material (that is less
than 300 m/s).
 The h value should be set to a value larger than the inter-cell distance (to avoid excessive
de-cohesion), but not too large to avoid excessive energy absorption by the bird.
 Using this model, the discrepancies between the real tests and the simulations are usually
due to parameters independent from bird modeling: boundary conditions, target material
laws and test sensitivity and so on.
 The final shape of the bird is more “physical”, so more experimental data would be
required, for instance multiple penetration cases.

Example 50 - INIVOL and Fluid Structure Interaction (Drop Container)
Summary
The aim of this example is to introduce /INIVOL for initial volume fractions of different materials
in multi-material ALE elements, /SURF/PLANE for infinite plane, and fluid structure interaction
(FSI) with a Lagrangian container.
Title
FSI and ALE container drop
Number
50
Brief Description
A container that is partially filled with liquid is dropped and the interaction between the liquid and
structure can be studied. The liquid fill is defined using /INIVOL and infinite plane /SURF/PLANE.
The contact between the structure and liquid is defined using /INTER/TYPE18.
Keywords
 Solid element /BRICK
 Material law /MAT/LAW51 and /ALE/MAT
 ALE grid velocity /ALE/GRID/DONEA
 Upwind /UPWIND
 Property /PROP/SOLID
 Load /INIVEL
 Define initial volume /INIVOL
 Infinite surface plane /SURF/PLANE
 Fluid structure contact /INTER/TYPE18
Input File
<install_directory>/demos/hwsolvers/radioss/50_invol_and_fluid_structure_interaction
/fsi_drop_container*

Advanced

Overview
Description of the physical problem

A container partially filled with water is simulated being dropped from a height of 1 meter. The
container is partially filled with water with the remainder filled with air.
Modeling description
A hex mesh is created that fully encloses the structural container. The mesh size of the hex mesh
should be ½ the size of the structural mesh. Ideally the hex mesh should also be ¼ of the
structural mesh size in the direction of impact. To simplify this example, the hex mesh in this
model does not adhere to the ¼ mesh size guideline.
Boundary conditions
Each outer side of the hex mesh is constrained to prevent displacement in the direction normal to
the side. For example, the top and bottom of the hex mesh is constrained in the z translation
DOF (Fig. 2). The same is done for the other four sides. The velocity at impact of a drop from 1
meter would be 4429 mm/s. Since the simulation is started right before impact, an initial velocity
of 4429 mm/s is applied to the container and the fluid hex mesh (Fig. 2).

Fig 2: Boundary condition of container in z-direction

In one /MAT/LAW51 card, three different phases can be defined. The two phases are: Air and
Water
Air is defined with the following characteristics:
 Reference density used in E.O.S (equation of state) = 1.2e-12
 Initial density of air = 1.2e-12
 Initial energy per unit volume = 0.25
 Hydrodynamic cavitation pressure = -1e-20
 Hydrodynamic coefficient C41 = 0.4
Water is defined with the following characteristics:
 Initial density of water = 1e-9
 Hydrodynamic cavitation pressure = -1e-20
 Hydrodynamic coefficient C13 = 2250 (Liquid bulk modulus)
/ALE/MAT should also be defined for the /LAW/MAT51 material, to indicate that is an ALE
model.
Coupled Euler-Lagrange (CEL) interface
To define the contact between the fluid and the structure a visco-elastic penalty formulation
/INTER/TYPE18 interface is defined as follows:
Master is the Lagrange structure
Slave is the ALE fluid nodes

Gap is the Interface gap. The recommended value is 1.5 * fluid element size along the normal
direction to contact
 * 2 * Sel
Stfval 
Gap

is the (highest) fluid density
 is the velocity. as defined below:
 For incompressible models (ditching, sloshing, etc.), use the velocity of the event.
 For compressible but not supersonic, use the speed of the sound in the material.
 Compressible and transonic (Mach 0.8 to 1.0), replace the term v 2 with v*c
where,
 is the speed of the sound in the material and c is the speed of sound in air
 Compressible and supersonic, use the velocity of the event
 For an explosion, use the Chapman Jouguet velocity
Sel is the surface area of the Lagrangian elements.
Gap is the interface gap, as defined above.
For this example:
Gap=1.5 * fluid element size = 1.5*2.5=3.75[mm]
 * 2 * Sel 1e  9 * 44292 * (5 * 5)
Stfval  =  0.131
Gap 3.75
Filling the container with /INIVOL

With /INIVOL, the water line can be defined in this part.
part_ID = Part ID of ALE hex mesh

surf_IDn = 3-nodes or 4-nodes surfaces only or /SURF/PLANE
ALE_PHASE = Phase of the multi-material ALE material

FILL_OPT = 0 or 1
=0: filling the side which along normal direction
=1: filling the side which against normal direction
/INIVOL uses successive filling actions of the initial background multi-material ALE mesh, to get
the final configuration of the initial volume fractions (three containers and three ALE phases).
Initially the volume is filled by the first material defined in the /MAT/LAW51 field. In this case,
the first material is air, so the entire hex mesh is first filled with air. Next, a surface is defined
from the container part ID.
/SURF/PART/998
Vessel_Surf_Part
85
Since the surface normal of container part point outside, use FILL_OPT = 1 to fill the water
(phase 3) inside the container (filling the side which against surface normal direction).
/INIVOL/86/10003507
INIVOL
# Surf_ID ALE_PHASE FILL_OPT
998 3 1
Now, ALE mesh is filled with ALE material 1 (air) from /MAT/LAW51 on the outside of the
container and material 3 (water) inside the container. Lastly, define a surface plane,
/SURF/PLANE to define the fill height. The normal of this plane points upward, use FILL_OPT = 0
to fill the air (phase 2) above the plane (filling the side along normal direction).
/INIVOL/86/10003507
# Surf_ID ALE_PHASE FILL_OPT
9999 2 0

To check the initial fill, the following animation options can be used in the Engine file.
/ANIM/ELEM/DENSITY
/ANIM/ELEM/VFRAC
You can contour the model and use section cut to see inside, or use iso-surface, as shown in the
image below.
/UPWIND - Anti-diffusive Technique

The amount of diffusion or mixing between the /MAT/LAW51 materials can be reduced by
changing the third parameter, η3 in /UPWIND.
η3 = 1: Full Upwind (default value, recommended)
η3 = 1e-20: Less Upwind (less diffusive) => used in this example model.
η3 = 1e-30: Zero Upwind (less diffusive, but potentially unstable)
η3 = -1: Full Downwind (Anti-Diffusive Technique, potentially unstable)
/ALE/GRID/DONEA - ALE Grid Velocity

This activates the J. Donea Grid Formulation, where the velocity of a given grid node depends on
velocity and displacements of neighboring grid nodes.

Engine Control
It is recommended to use time step scale factor 0.5 for ALE in /DT/BRICK in order to keep
computation stable, and use fac=1.0 in /UPWM/SUPG. This option provides better velocity field in
Cartesian grids when ALE material velocity is not normal to brick faces.
To see the movement of the water in the container, and isosurface plot of results type, User Var
40 can be done. If the simple averaging method is used in HyperView, the results will look
smoother, as shown below.
Also notice that water is starting to splash up the sides of the container at the end of the
simulation.

Example 51 - Optimization in RADIOSS for B-Pillar (Thickness
optimization)
Summary
The new feature, Optimization in RADIOSS, makes it easy to set up an optimization problem in
RADIOSS Block Format. A typical optimization problem has three elements:
 Objective (minimize mass of the structure)
 Constraints (stress less than 200 MPa and maximum displacement less than 10mm)
 Design variables (parameters that can be adjusted to improve the design)
The setup of optimization in RADIOSS requires an extra input file apart from the usual Starter
and Engine input files. The required file is an optimization input file named <name>.radopt (the
Starter and Engine files are named <name>_0000.rad and <name>_0001.rad, respectively). The
<name>.radopt file defines optimization entities such as the optimization objective, optimization
constraints, design variables, optimization responses and so on.
For more RADIOSS optimization details, refer to Design Optimization in the User’s Guide.

Title
Optimization in RADIOSS for B-

Pillar (Thickness optimization)
Number
51
Brief Description
This example defines a crash test on B-Pillar. The optimization objective is to minimize the mass of the
B-Pillar by changing the shell thickness. The intrusion, which is defined by the optimization constraint
is required to not be larger than the original model, to keep the passenger safe.
RADIOSS Optimization Keywords

 Optimization objective (/DESOBJ)
 Optimization design variable (/DESVAR)
 Relate design variables to analysis model properties (/DVPREL1)
 Optimization design response (/DRESP1)
 Optimization design constraint (/DCONSTR)
RADIOSS Options
 Material law (/MAT/LAW36)
 Shell property (/PROP/SHELL)
 Interface type 7 (/INTER/TYPE7)
Input File
Optimization in RADIOSS for B-Pillar (Thickness Optimization):
<install_directory>/demos/hwsolvers/radioss/51_optimization_bpillar

Advanced

Overview
Aim of the Problem

The purpose of this example is to show how to set up an optimization, based on a crash test
(RADIOSS Block model).
Cut the B-Pillar from the full car model and set the initial velocity in the rigid cylinder.
Units: mm, s, ton, N, and MPa
Optimization Problem
 Objective: Minimize mass
 Constraint: Maximum displacement of node 2021524 in Y direction (inside reinforcement) <
19.7 [mm]
 Design variables:
- Shell thickness of the middle reinforcement - allowable range [0.5mm, 3.0mm]
- Shell thickness of the inside reinforcement - allowable range [0.5mm, 3.0mm]

Detailed Optimization Setup
Fig 4: RADIOSS optimization setup
Optimization Objective
/DESOBJ is used to define optimization objective. In this example, it defines the minimal response
#1.
Response #1 defines an optimization response that is the combination of mass in part group
#2000329 (defined in the RADIOSS Starter file), which includes both parts of the middle
reinforcement and the inside reinforcement.
Fig 5: Optimization objective setup in RADIOSS optimization

Optimization Constraint
In this example, you want to constrain the intrusion to not be larger than the original model
(before optimization). In the original model, max. y-displacement in node 2021524 is 19.7mm.
Define this value as cmax in /DCONSTR and in response 2, define 19.7 as the max. y-displacement
(with RTYPE=5, ATTA=7) in node 2021524 (with PTYPE=1, ATTI=2021524).
Depending on your own optimization criteria, the value in cmax may differ.
Fig 6: Optimization constraint setup in RADIOSS optimization
Design Variable
/DESVAR is used to define design variables and /DVPREL1 to relate design variables to analysis
model property.
Two different design variables are defined: One for part 2000327 and one for part 2000329
For example, the design variable for inside reinforcement (part 2000327) is defined:
 In /DESVAR with the range [0.5,3.0] - this will be used in /DVPREL1.
In /DVPREL1 with prop_typ and prop_fid the above variable can be used for shell
thickness, and with prop_ID the thickness in shell property 2000327 will be used in the
optimization run.
 COEF_1 in /DVPREL1. In each iteration, Pi (the thickness value) will be equal to C0+(1.0*X),
where
X  [0.5,3.0] which is defined in /DESVAR.

Pi  C 0   (COEFi * DVIDi )
i
Fig 7: Design variable setup in RADIOSS optimization

/INIVEL and /INTER/TYPE7 are also used.
Use /INIVEL to optimize the B-Pillar under same initial kinetic energy.
In this example, you want to change the shell thickness to have minimal mass of B-Pillar. There
is initial penetration, due to thickness changes between two contact parts. Defining initial
thickness in /PART can avoid this issue and the thickness is only used to calculate the gap in
interfaces.
Fig 8: /PART card in RADIOSS Starter file

In the results check the following:

1. The latest design provides the best results. Verify whether the result is feasible or not. In the
*.out file (Neon-b_pillar.out) this data is at end of each iteration.
“FEASIBLE DESIGN (ALL CONSTRAINTS SATISFIED).“ - indicates that the design is good.
“INFEASIBLE DESIGN (AT LEAST ONE CONSTRAINT VIOLATED).” - indicates that the design
variable definition needs to be checked/improved.
This information can also be found in the output file hwsolver.mesg.
2. In the *.out file (Neon-b_pillar.out), check/verify the optimization definition. “Objective
Function: Minimize Combined Mass”, “Run Type: Sizing Optimization” and so on.
3. By running the Optimization in RADIOSS, an equivalent OptiStruct model will also be
automatically created and named *.fem (Neon-b_pillar.fem).
Check the results in the *.eslout (Neon-b_pillar.eslout), in the *h3d file, or in each ANIM,
T01 file. The *.eslout file contains the value of optimized objective in each iterations.
Fig 9: Optimized results in each iterations in *selout file
In the original model, the mass of the two parts is 3.0011e-3[Ton] and the optimized mass for
these two parts is 2.4158e-3[Ton]. The mass is reduced by approximately 19.5%.
The thickness of part #2000327 is 1.328mm and the optimized thickness is 0.5539mm. The
thickness is reduced.
The thickness of part #2000329 is 0.7060mm and the optimized thickness is 0.8933mm. The
thickness in this model increased in order to get better performance of these two parts (in
response 1, ATTB =1 (for COMB) is defined).
Fig 10: Optimized results of shell thickness of two reinforcement parts

Fig 11: Optimized results of total mass of two reinforcement parts
In the last iteration, the mass was reduced to 2.4158e-3[Ton]. This new design still meets the
constraint (< 19.7[mm]), defined in /DCONSTR.
In node 2021524, the max. y-displacement:
19.57[mm] (last iteration) < 19.7 [mm] (in constraint). Meets the constraint.
Fig 12: y-displacement on node 2021524 in original model and optimized model

Example 52 - Creep and Stress Relaxation
Summary
The aim of this example is to introduce how to use typical visco-elastic material to simulate creep
and stress relaxation tests. Stress relaxation is the phenomena of how polymers relieve stress
under constant strain, and creep is the phenomena of how polymers or metal move slowly or
deform permanently under constant stresses. This simulates the creep and relaxation processes
over a short period of time in quasi-static.
Title
Creep and Stress Relaxation
Number
52
Brief Description
Use visco-elastic material law /MAT/LAW40 to simulate the creep and stress relaxation.
Keywords
 /MAT/LAW40
RADIOSS Options
 Concentrated force load (/CLOAD)
Input file
Creep and Stress Relaxation:
<install_directory>/demos/hwsolvers/radioss/52_creep_and_stress_relaxation/*

Advanced

Overview

A foam sample with dimension: Radius 10 mm and high 15 mm.
 For stress relaxation test: The foam sample has been compressed until a given strain and
kept in this state.
 For creep test: The foam sample has been tensile under constant force.

To describe the phenomenon stress relaxation and creep, use viscous material law /MAT/LAW40
with the following characteristics of foam:
 Initial density = 2e-9 [Mg/mm3]
 Bulk modulus = 66.67 [MPa]
 Long time shear modulus Ginf = 10 [MPa]
 Shear modulus G1 = 90 [MPa]

 Decay constant 1 = 0.01 [1/ms], 2 = 0.05 [1/ms] and 1 = 1 [1/ms] for compare

Fig 2: Stress relaxation test under constant displacement and creep test under constant force
For stress relaxation test: The foam sample has been compressed under constant displacement
(/IMPDISP).
For creep test: The foam sample has been tensile under constant force (/CLOAD).
The stress relaxation test shows stress relieve under constant displacement with different
relaxation parameters (Decay constant, defined as the inverse of relaxation time ) and
shows a different stress relive tendency.

Fig. 3: Stress relieved with different Decay constant β in stress relaxation test under constant displacement
The creep test shows deformation increased under constant force and with different relaxation
parameter it shows a different deformation increase tendency.

Fig. 4: Sample deformed with Decay constant β in creep test under constant force
In LAW40 shear modulus is reduced with time and tends to G∞ after an infinite period of time.
The softening speed is determined by relaxation parameter . Higher relaxation parameter
means quick softening.
5
G  t  = G  +  G ie
βi t

with
The general case of viscous materials represents time-dependent in elastic behavior. Creep is
time-depended deformation and stress relaxation is a time-depended decrease in stress. Viscous
material can describe these two phenomenons. In RADIOSS, the following material laws describe
viscous:
Visco-elastic law
 /MAT/LAW34: visco-elastic generalized Maxwell model, Boltzmann (solids)
 /MAT/LAW35: visco-elastic generalized Maxwell-Kelvin-Voigt (shells + solids)
 /MAT/LAW38: visco-elastic tabulated (solids)
 /MAT/LAW40: visco-elastic generalized Maxwell-Kelvin (solids)
 /MAT/LAW42: Ogden/Mooney-Rivlin with Prony viscosity (Hyperelastic materials)
 /MAT/LAW62: Ogden (Hyperelastic materials)
 /MAT/LAW70: visco-elastic tabulated (solids)
 /MAT/LAW77: visco-elastic tabulated with porosity and air flow
Visco-elastic plastic law
 /MAT/LAW33: visco-elastic plastic (solids) and user-defined yield function
 /MAT/LAW52: Gurson, visco-elasto-plastic porous metals, and strain rate dependent
 /MAT/LAW66: semi-analytical plastic model. Yield surface built from curves in tension,
compression and shear + /VISC/PRONY
The creep compliance and the relaxation modulus are often modeled by combinations of springs
and dashpots. The two typical simple schematic model of visco-elastic material are Maxwell
model and Kelvin-Voigt model. The Maxwell model represents the material relaxation, but it is
only accurate for secondary creep (creep with slow decrease in creep strain rate) as the viscous
strains after unloading are not taken into account. The plasticity can be introduced in the models
by using a plastic spring. Base on the Maxwell and Kelvin-Voigt models adding other springs
could get a generalized model. The Maxwell and Kelvin-Voigt models are appropriate for ideal
stress relaxation and creep behaviors. Although, they are not adequate for most of physical
materials. A generalization of these laws, like LAW34, LAW35 and LAW40 are a better choice,
which can describe deviatory behavior of material.
Maxwell model Kelvin-Voigt model

Example 53 - Thermal Analysis
Summary
Thermal analysis, like heat exchange (between two contact surfaces, between heat object and
surrounding atmosphere though convection or radiation, inside the object through conduction),
deformation is due to thermal expansion or heat generated, due to friction can be simulated in
RADIOSS. In this example heat exchange is discussed between a moving heat source and one
plate, due to contact and also between plate and atmosphere (water) through convective flux.

Title
Thermal Analysis –
Heat Exchange
Number
53.1
Brief Description
A heat source moved on one plate. Heat exchanged between a heat source and a plate through
contact, also between a plate and the atmosphere (water) through convective flux.
Keywords
 /HEAT/MAT
 /CONVEC
 /IMPTEMP
 /INTER/TYPE7
 /MAT/LAW2
RADIOSS Options
Input File
<install_directory>/demos/hwsolvers/radioss/53_thermal_analysis/Heat_exchange/*

Advanced

Overview

A heat source with a constant temperature of 800K is moved under imposed displacement on one
plate with an initial temperature of 298K. The dimension of the heat source is 5mm x 5mm and
the plate is 100mm x 100mm.
Units: mm, ms, g, N, and MPa

/MAT/LAW2 and /HEAT/MAT are used to describe the aluminum heat source and plate, with the
 g 
 mm3 
 Initial density: 2.8 x 10 -3
 Young modulus: 70000 [MPa]

 Yield stress: 206 [MPa]
 Hardening parameter: 450 [MPa]
 Room temperature: 298 [K]

 N  mm 
ρC p  
: 2.51  mm  K 
3
 Specific heat
 Initial temperature for heat source: 800 [K] and for plate: 398 [K]
 N  mm 
 ms  mm  K 
 Thermal conductivity coefficient AS: 0.23
/HEAT/MAT is an additional material law card used to describe the material thermal character. So
the material ID in the material law in /MAT and in /HEAT/MAT must be the same. The thermal
parameter defined in /HEAT/MAT will recover the same parameters which are defined in the
material law.
Heat capacity provides heat and mass the ability to change the temperature. In engineering and
science, it is recommended to use specific heat capacity, which is heat capacity divided by mass,
 J   J   N  mm 
 kg  K  C p  897    897  gK 
 in SI unit. Heat capacity is  kg  K   
for aluminum.
Refer to Material Constants in the Theory Manual Appendices for more information on heat
capacity of ordinary material.
 N  mm 
0.23 
For the thermal conductivity coefficient AS,  ms  mm  K  . Thermal conductivity
 W   N  mm 
k  230    0.23 
mK  ms  mm  K  for aluminum, and constant thermal conductivity. Set
BS=0. Since thermal conductivity k=AS+BS*T, then k=AS, in this case.
With /IMPTEMP, imposed temperature will be set on a group of nodes. The source constant
temperature is defined for heat source.

Use /CONVEC to describe the heat exchange between a structural component and its surrounding
atmosphere (infinite room).
The surrounding atmosphere is water with a constant temperature of 298K, which is described in
function, fct_IDT (Figure 3).
Fig 3: Temperature in water.
Where, H is the heat transfer coefficient between structural component and its surrounding
 J 
 s  m2  K 
infinite room with unit . In general, the convective heat transfer coefficient for water
 J 
 s  m2  K 
(free convection) is about 20 - 100 and water (forced convection) is about 50 -
 J 
 s  m2  K 
10000 . Forced convection in water is
 J   N  mm 
H  500   5e  4
 s  m  K   s  mm2  K 
2
.
In /INTER/TYPE7, heat exchange between the heat source and plate during the contact, is
defined Ithe=1 to activated heat transfer between master and slave.

Ithe_form set to 1 for heat exchange between all pieces in contact.
There are two ways to define heat exchange between contact parts.
 W 
 m 2  K 
1. Define constant heat exchange coefficient using Kthe ( in SI unit). In this case,
fct_IDK = 0.
2. If fct_IDK ≠ 0, the heat exchange coefficient is the function of contact pressure using this
curve and Kthe is the scale factor.
K  Kthe  fct_IDK (Ascale K , P)

Interfacial heat transfer coefficient, K described conductive heat flux through a unit area of a
plate with particular thickness. The range of this heat transfer coefficient can be very large, which
will affect the accuracy of simulation. To get a more accurate result, experimental test are
required.
 W   N  mm 
Kthe  15000  2   0.0.15 
m K  ms  mm  K 
2
Set Fheats and Fheatm to zero, to not consider heat friction.

The following figure shows nodal temperature at time 10[ms], 20[ms] and 30[ms]. Part of heat
transferred to plate through contact. Therefore, the temperature under the trace increased. The
temperature on the plate decreased during the time, due to the convection with water.
Fig 4: Nodal temperature in plate at time=10[ms], 20[ms] and 30[ms].

Below the nodal temperature on Nodal N641, N1034, N958 and N1708 are illustrated.
1. Nodal N641 is not under trace. The temperature changed, only due to convection with water.
2. Nodal N1034, N958 and N1708 are under trace. At first the temperature decreased before the
heat source began, due to convection with water, and then increased, due to the heat
exchange from the heat source through contact. Once the heat source is removed, the
temperature decreased again, due to the heat conduction inside the material and convection
with water. So the slope of the temperature decrease is much larger than N641 (only
convection).
Fig 5: Temperature on Nodal N641, N1034, N958 and N1708

RADIOSS 2017 Tutorials and Examples

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

RADIOSS 2017 Tutorials and Examples

Uploaded by

Copyright:

Available Formats

HyperWorks 2017

RADIOSS Tutorials and Examples

Learn more at altairhyperworks.com

HyperWorks® 2017 Products

Additional Altair Products:

Altair Packaged Solution Offerings (PSOs) Copyright© 2008-2016

Altair Simulation Cloud Suite:

Altair PBS Works™:

Software products of solidThinking, Inc., a wholly owned subsidiary of Altair Engineering:

ALTAIR ENGINEERING INC. Proprietary and Confidential. Contains Trade Secret

Third party software licenses

Software Security Measures:

Altair Support on the World Wide Web

Altair Training Classes

Telephone and e-mail

Location Telephone e-mail

Australia 64.9.413.7981 anzsupport@altair.com

Brazil 55.11.3884.0414 br_support@altair.com

Canada 416.447.6463 support@altairengineering.ca

China 86.400.619.6186 support@altair.com.cn

France 33.1.4133.0992 francesupport@altair.com

Germany 49.7031.6208.22 hwsupport@altair.de

India 91.80.6629.4500 support@india.altair.com

Italy 39.800.905.595 support@altairengineering.it

Japan 81.3.5396.2881 support@altairjp.co.jp

Korea 82.70.4050.9200 support@altair.co.kr

Mexico 55.56.58.68.08 mx-support@altair.com

North America 248.614.2425 hwsupport@altair.com

Scandinavia 46.46.460.2828 support@altair.se

Spain 34.910.810.080 support-spain@altair.com

South Africa 27.21.8311500 support@altair.co.za

United Kingdom 01926.468.600 support@uk.altair.com

Accessing the Model Files.............................................................................................................................. 4

Introductory Tutorials ................................................................................................................................... 5

List of Examples......................................................................................................................................... 273

Example 1 - Twisted Beam ......................................................................................................................... 281

Example 2 - Snap-thru Roof ....................................................................................................................... 287

Example 3 - S-beam Crash ......................................................................................................................... 301

Example 4 - Airbag .................................................................................................................................... 317

Example 5 - Beam Frame ........................................................................................................................... 328

Example 6 - Fuel Tank ................................................................................................................................ 335

Example 7 - Pendulums ............................................................................................................................. 353

Example 8 - Hopkinson Bar ........................................................................................................................ 371

Example 9 - Billiards (pool) ........................................................................................................................ 390

Example 10 - Bending ................................................................................................................................ 417

Example 11 - Tensile Test .......................................................................................................................... 426

Example 12 - Jumping Bicycle .................................................................................................................... 465

Example 13 - Shock Tube ........................................................................................................................... 486

Example 14 - Truck with Flexible Body ....................................................................................................... 506

Example 15 - Gears.................................................................................................................................... 531

Example 16 - Dummy Positioning .............................................................................................................. 540

Example 17 - Box Beam ............................................................................................................................. 573

Example 18 - Square Plate ......................................................................................................................... 640

Example 19 - Wave Propagation ................................................................................................................ 671

Example 20 - Cube..................................................................................................................................... 685

Example 21 - Cam...................................................................................................................................... 691

1 RADIOSS 2017 Tutorials Altair Engineering

Proprietary Information of Altair Engineering

Example 24 - Laminating ........................................................................................................................... 741

Example 25 - Spring-back .......................................................................................................................... 752

Example 26 - Ruptured Plate ..................................................................................................................... 771

Example 27 - Football (Soccer) Shots ......................................................................................................... 782

Example 37 - Analytical Beam .................................................................................................................... 789

Example 39 - Biomedical Valve .................................................................................................................. 801