Escolar Documentos
Profissional Documentos
Cultura Documentos
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
Table of Contents
Intro to RADIOSS for Impact Analysis
Examples using HyperMesh Desktop
Become familiar with the format of RADIOSS Engine and Starter files
Know how to use HyperWorks Solver Run Manager to execute a RADIOSS simulation
Analyze starter and engine output files
Learn to post-process key results from an explicit analysis with HyperView.
Model Description
The model simulates an impact of a thin-walled, closed hat section on a rigid wall.
Simulation time: 50 ms
Uniform thickness = 1.5 mm
Input files for this tutorial: FIRST_RUN_0000.rad and FIRST_RUN_0001.rad
Units: Length (mm), Time (ms), Mass (g), Force (N) and Stress (MPa)
Results requested:
o /ANIM/VECT/DISP Displacements
o /ANIM/VECT/VEL Velocities
o /ANIM/ELEM/EPSP Plastic Strain
o /ANIM/ELEM/VONM VonMises Stress
o /ANIM/ELEM/ENER Energies
o /ANIM/ELEM/HOURG Hourglass Energy
Johnson-Cook Elasto-Plastic Material /MAT/LAW2.
o = 7.8e-3 g/mm3 [Rho_I] Initial density
o E = 210,000 MPa [E] Youngs modulus
o = 0.29 [nu] Poissons ratio
o y = 180 MPa [a] Yield stress
o = 450.0 MPa [b] Plasticity hardening parameter
o n = 0.5 [nu] Plasticity hardening exponent
o MAX= 350 MPa [SIG_max0] Max Stress
Problem Setup
Copy the files: FIRST_RUN_0000.rad, FIRST_RUN_0001.rad into a working directory.
The length of the rail is about 1,000 mm. Using the speed of sound and the rail length, calculate the time
it takes for a shock wave to travel from one end of the rail to the other and enter this value as
SIMULATION_ TIME in the Engine file: FIRST_RUN_0001.rad file.
We would like to have 20 animation outputs of our analysis. Using the total simulation time, calculate the
time frequency animation output such that 20 animation steps are created. Enter this value as
ANIMATION_OUTPUT_FREQUENCY.
/VERS/140
# Simulation Time
/RUN/FIRST_RUN/1/
SIMULATION_ TIME
/ANIM/DT
# TSTART TFREQ
0.000000 ANIMATION_OUTPUT_FREQUENCY
/PRINT/-10
/RFILE
5000 0 0
/TFILE/0
Note that the time history file output request /TFILE is set to 0.1 ms for the restart later in this exercise.
Also note that the engine keyword /DT/NODA tells the solver to use nodal time step.
Step 3: Select the Starter file FIRST_RUN_0000.rad as the Input file and Run with
onestep option.
Tip: Option onestep tells RADIOSS to only run a single step of the simulation (in this case the
starter) and stop. If not specified, then the engine will run automatically if the starter finds no fatal
errors and if there is an engine file in the directory).
The time step calculated by the RADIOSS Starter is reported in the Starter output file. Search the file for
time step and review the value for Shell, Spring and Nodal time step. Which of these is controlling a
stable solution?
Step 5: Select the Engine file FIRST_RUN_0001.rad as the Input file and Run.
Tip: Option nt 4 tells RADIOSS to run the analysis on 4 threads (using 4 CPU) this should speed
up the analysis
Step 6: Review the Engine output file FIRST_RUN_0001.out with a text editor
Check the % Energy Error (ERROR) and Mass Addition (MAS.ERR) at the end of the
computation.
2. Or, open HyperView, load the file FIRST_RUN.h3d, and then Apply.
3. Click the Entity Attributes button to bring up the Entity Attributes panel. Click on
the Rigid Wall component in the graphics area and select the Transparent shading
4. Click the Contour button to enter the Contour panel. Select Velocity (v) for the
Result type: and click Apply.
6. Use the Next Time/Angle/Step button on the animation toolbar to incrementally step
through the animation of the Stress Contour.
This allows you to see the stress wave propagating from the front of the rail to the rear of the rail.
Check the time it takes the stress wave to reach the back end of the rail. How does it compare to
your calculation from Step 1?
1. Make a copy of the RADIOSS engine file FIRST_RUN_0001.rad and change the name
to FIRST_RUN_0002.rad.
2. Update the engine file with a new end time of 50 msec by updating the line beneath the
/RUN/FIRST_RUN/2/ command to 50.0 and updating the animation output to 2.5 msec
beneath /ANIM/DT as shown below.
# Simulation Time
/RUN/FIRST_RUN/2/
50.0
/ANIM/DT
# TSTART TFREQ
0.000000 2.5
/PRINT/-10
/RFILE
5000 0 0
/TFILE/0
Select the second Engine file (FIRST_RUN_0002.rad) as the Input file and Run
Notice that the animation files continue numbering from where the first run left off. Also notice that a
second time history file (T02) is created.
2. Split your page into 4 windows using the Page Window Layout:
3. In the first window, use the HyperView client to load the FIRST_RUN.h3d file with
a transparent rigid wall and contoured with Von Mises stress as in Step 7.
1. From the Build Plots panel , load the Time History File FIRST_RUN_T01 - note
that the second file FIRST_RUN_T02 will also be loaded.
2. Select the Y-type, Y-Request, and Y-Component as Rigid wall > RIGID WALL 1 >
Resultant and click Apply.
2. From the Build Plots panel, select the Y-Type: Global Variables, multiple Y Request:
Internal Energy, Kinetic Energy, Hourglass Energy, and TTE-Total Translational Energy
by control-clicking on each selection, and select the Y Component: MAG. Click Apply
to create the plot.
Step 12: Plot Global Variables > Time Step vs Time in the remaining window
Once the post-processing is done, it can be saved by writing a session file or a report
template:
1) A session file (*.mvw) exactly reloads your post-processing on the same results files, or
results files of the same names (e.g. after a re-run)
2) A report template (*.tpl) allows post-processing of new simulation results (new run,
different filenames) based on the template format.
For this exercise we will create both a session file (*.mvw) and a template file (*.tpl) which
will be used in later exercises.
1. Click on File > Save > Session and save the file as Rail_Results.mvw.
2. Click on File > Save as > Report Template and save the file as Rail_Results.tpl.
The uniaxial tensile test uses a quarter size mesh with symmetric boundary conditions to reduce the
solver run time.
UNITS: Length (mm), Time (ms), Mass (kg), Force (kN) and Stress (GPa)
Simulation time: 10.0 ms
Boundary Conditions:
o The 3 upper right nodes (TX, and RY, RZ)
o The center node on right is totally fixed (TX, TY, TZ, RX, RY, RZ)
o A symmetry boundary condition on all bottom nodes (TY, RX, RZ)
At the left side a constant velocity is applied = 1 mm/ms on -X direction.
Tensile test specimen dimensions = 11 x 100 with a uniform thickness = 1.7 [mm]
Note the slight perturbation of the top right node so that necking will form at the center of
the specimen
Problem Setup
Copy the files: tensile_0000.rad to a working directory
Step 2: Create and assign the material Aluminum to the test specimen
1. Right-click in the Model Browser and select Create > Material.
3. In the Model Browser, select the component PSHELL1 and click on the Mat_ID and select
Aluminum to set the material.
2. In the Entity Editor, set the Name: to sheet_1.7 and enter Thick: of 1.7.
3. In the Model Browser, select the PSHELL1 component and click on the Prop_ID in the Entity
Editor and select sheet_1.7 to set the property.
2. For Name, enter constraint1, set Select type to Boundary Condition and set GRNOD
to Nodes.
3. Click on the nodes entity selector. Select the three nodes shown in the figure below and click
proceed.
6. Create a new boundary condition with the Name of constraint2 with the following node
selected and all six degrees of freedom constrained.
7. Create a third boundary condition with Name constraint3 with the following nodes
constrained in DOFs Ty, Rx, and Rz.
3. Click the Create/Select curve button to set the Curve ID. A curve editor appears.
4. Create a new curve with the Name Load with values as shown in table below.
5. Close the Curve editor to assign the created curve to this constraint.
Tip: Note that the graphics may not take the Y scale factor into account.
Step 9: Review the listing files for this run and verify the results
7. See if there are any warnings or errors in the .out files. If so, list them below.
8. Using HyperView, plot the displacement and strain contour and compare to the results on the
following page.
UNITS: Length (mm), Time (ms), Mass (kg), Force (kN) and Stress (GPa)
Simulation time: 1.5 ms in three steps of 0.05 ms for each load case
The outer circumference area is fixed on all translational DOFs (TX, TY, TZ) and the center node
is fixed in the X translation and the X and Y rotation (TX, RX, RY)
The bushing dimensions are: thickness = 100 mm, External Diameter = 200 mm and internal
diameter = 50 mm.
Hyper-Elastic Material /MAT/LAW42 (Rubber)
= 6.0 e-6 [kg/mm3] Initial density
= 0.495 [-] Poissons ratio
1 = 0.600 [GPa] Ground Shear mod. 1
1 = 2 [-] Alpha 1
2 = -2 [-] Alpha 2
Problem Setup
Copy the files: GASKET_0000.rad to a working directory
Step 2: Create a new M42_OGDEN material named Rubber and assign as the
Mat_Id for the GASKET component
Tip: When created, the component name should appear in bold, indicating it is the current
component and all new elements created will be assigned by default to this part.
2. In the 1D > rigids panel, switch the selector nodes 2-n to multiple nodes, and switch the
primary node to calculate node.
3. Click on the nodes entity selector and pick any node on the inner face of the gasket.
4. Click on the nodes entity selector to bring up the advanced selector menu and select by face.
HyperMesh Desktop will find and select all nodes on that inner face.
5. Click create to create the rigid body and click return to exit the panel.
10. For Name, enter OInneruter_BC, set Select type to Boundary condition and set the
GRNOD to Nodes.
11. Select the master node of rigid body created in the previous step. Click Proceed.
13. Click Create at the bottom of panel to create the inner fixed boundary condition.
16. Create a new curve under Curve ID using the XY curve editor named DISP_Y containing the
points {0, 0}, {0.5, 10}, and {1.0, 10}. Click update and close the XY curve editor
GUI.
19. Create a third Imposed Displacement boundary condition named ROT20DEG_Z on the master
node of rigid body with the Direction to ZZ, and Scale X and Scale Y to 1.0 which uses a new
curve named ROT20DEG_Z containing the points {0, 0}, {1, 0}, {1.5, 0.349}, and
{2, 0.349}.
21. For Name, enter Outer_BC, set Select type to Boundary condition and set the GRNOD
to Nodes.
22. Click on the nodes entity selector and pick any node on the outer face of the gasket.
24. Check the Tx, Ty, and Tz to fix the translational rotational degrees of freedom.
25. Click Create at the bottom of panel to create the outer fixed boundary condition.
Step 78: Set the control cards to a final run time of 1.51 ms with time frequency of
0.05 ms, a print to the *.out file and screen every 100 cycles, time history
information written in 0.0015 increments, and von Mises (VONM), density (DENS),
and pressure (P) outputs to the A* and h3d files
Tip: Use the HyperMesh Solver Browser from View > Browsers > HyperMesh > Solver to set up
the following control cards and parameters:
Step 110: Review the listing files for this run and verify the results
26. See if there are any warnings or errors in .out files.
27. Using HyperView plot the displacement and strain contour and vectors.
This exercise demonstrates how to set up a ball impact on regular glass versus safety glass
that has a thin layer of plastic. In all cases an FLD failure model is defined for shell rupture.
These models are then updated to include XFEM (eXtended Finite Element Method) in which
failure can propagate across elements (rather than along mesh lines). Note: XFEM
requires version 14.0.220 or later.
UNITS: Length (mm), Time (s), Mass (Mg), Force (N) and Stress (MPa)
Simulation time: *_0001.rad [0 0.025 s]
Boundary Conditions:
o All degrees of freedom will be constrained on the master node of a rigid
body connected to top and bottom of rubber frame
o An initial velocity is given to a spherical rigid wall
Johnson-Cook Elastic Plastic with Brittle Failure Material /MAT/PLAS_BRIT
(Glass)
Problem Setup
You should copy the files: BallDrop-Start_0000.rad
Step 2: Define a new M27_PLAS_BRIT material named Glass with the following
properties and apply this material as the Mat_Id for the Glass plate component:
Tip: For the N: Data layer information within the Table Data , enter the following:
In the rigids panel on the 1D page, switch nodes 2-n to multiple nodes and switch
primary node to calculate node.
Click on the nodes entity selector and select one node apiece on the outer faces of both
Fixed upper and Fixed lower, respectively. Click again on the nodes entity
selector to bring up the advanced selector and choose by face.
Tip: HyperMesh Desktop will pick all of the nodes on the upper and lower faces.
Click create to create the rigid body and click return to exit the panel.
In the Model Browser window, right-click and select Create > Rigid Wall.
Complete the new Rigid Wall definition with the following parameters, selecting the
newly created node as the Base node for the sphere and selecting the Glass plate
component for gr_nod1 (S):
Step 10: Check the *.out files for warnings or errors and then review
the results in HyperView, plotting the displacement contour
Displacement Results
Edit the laminate property to replace the middle glass layer with the plastic film material
just created.
Step 13: Check the *.out files for warnings or errors and then review
the results in HyperView comparing the plot of the displacement
contour with that of the previous run
Displacement Results
Rigid body
t=3 mm
t=2 mm
UNITS: Length (mm), Time (ms), Mass (kg), Force (kN) and Stress (GPa)
Simulation time: 30.0 ms
Boundary Conditions:
o All degrees of freedom except in the Z-direction will be constrained on the
master node of a rigid body connected to the end of the tube
o An initial velocity is given to all nodes in the model
Problem Setup
Copy the file: TubeCrush-Start_0000.rad to a working directory.
Step 2: Review the material properties of the Steel to verify that the parameters
match those given in the exercise introduction
2. Right-click on the RBODY entry in the tab and select Edit Card to open the card
image editor for the element.
3. Scroll down in the top part of the editor to find the MASS entry and change the value to
500.
4. Click return to close the Card Image dialog. Close the RBODY Manager tab.
Step 6: Create a tied interface between the flanges of the C-sections making up
the tube. Recall that by default a Type 2 interface is a kinematic condition.
1. In the Model Browser window right click Create > Contact in the drop down menu to
open the Create Group dialog box. Change the Card Image TYPE2 and name the
contact Seamweld.
2. To set the slave nodes, click on the Grnod_id (S) Nodes entity selector. When the Add
entities panel appears, change the panels entity selector to nodes
3. Click return to exit the panel and return to the dialog box entries.
4. Select the Surf_id (M) Elements, entity selector and in the select entities dialog
change the entity selector to Elements. Select the inner area of the C-Channel elements
as shown in the next figure..
Step 7: Define a self contact Type 7 interface for the C-channels with the
following parameters shown below:
The default contact gap thickness method defined by Igap in TYPE7 contact is called
constant gap. This means that the actual thickness of the parts is not used for contact but
instead the thickness of the parts is considered the same for all parts and is defined by
GapMin. If GapMin is not defined then the constant gap thickness is calculated by
RADIOSS.
In this first example, constant gap thickness is used and the results will be compared to
another iteration that uses variable gap where the actual part thickness is used for contact.
1. In the Model Browser window right click Create > Contact in the drop down menu to open
the Create Group dialog box. Change the Card Image TYPE7 and name the contact
self contact.
2. Click on the Grnod_id (S) Nodes entity selector and change the entity type to Components
and click on Components and pick all the Components in the model.
3. Click on the GSurf_id (M) Element entity selector and change the entity type to
Components and click on Components and pick all the Components in the model.
4. Enter the best practices values of Istf = 4: K= min(Km,Ks) and Stmin= 1 kN/mm
5. Enter GapMin = 1.0 mm.
Note: The value of Gapmin was chosen to be less than thickness of the thinnest part
which should be small enough to avoid any initial penetrations but large enough to allow the
contact forces to prevent penetration.
6. Enter a steel against steel friction value as Fric = 0.2.
7. Enter additional best practice value of Iform = 2.
Step 8: Create infinite plane rigid wall named Wall at front of tube
1. In the Nodes panel, select the XYZ option , place the model in the XZ view ,
and click one of the nodes on the bottom of the c-channel three times to load its X, Y,
and Z coordinates, respectively, into the display fields.
Tip: The panel shows the Z-coordinate of the node is -304.5mm. This gives the user
enough information to determine the proper offset for the base node of the infinite wall.
3. Right-click in the Model Browser and select Create > Rigid Wall
Rename the rigid wall to Wall and set the geometry type to Infinite Plane with the
newly created node selected as the Base node, and a Normal vector of {0, 0, 1},
SLIDE set to 2 to enable friction, a fric parameter of 0.2, and a d value of 100 mm.
Tip: d represents the distance to search for slave nodes, which is the distance inside of
which nodes are seen by the rigid wall as potential contacts.
Step 10: Define time history output block for the rigid wall force
1. Right click in the model browser and select Create > Output Block.
2. The Output Block with name outputblock1 of solver keyword /TH/NODE
is created and appears in the Entity Editor (EE) in the bottom pane of the
model browser.
3. Change Entity IDs to Groups. Change Keyword to RWALL. Then select
Groups and select Wall.
4. Select NUM_VARIABLES and select 1. This will bring up a table in which
the Var DEF should be entered. (DEF is for default outputs, which for a
rigid wall are the forces: FNX, FNY, FNZ, FTX, FTY, FTZ). See EE image below:
1. Select File > Export > Solver Deck and click enter TubeCrush-Iter1_0000.rad
2. Check the Merger starter and engine file to try exporting the starter and engine files in one
file.
3. Click Export
Step 13: Review the output files for warnings and errors and view the animation
results in HyperView
1. While the model is running use a text editor to edit the TubeCrush-
Iter1_0000.out file and search for WARINING
EXPECTED RESULTS
TubeCrush-Iter1 (Constant Gap = 1.0), sectioned at Y = 40
Note: There are nodes that have penetrated the rigid wall. This is due to incompatible kinematic
conditions of the Type 2 interface and the rigid walls. The starter output file indicates that some
nodes in the Type 2 interface have been removed from the rigid wall slave set.
Step 14: Update the contact gap option to use variable gap (Igap = 2) in the Type 7
interface and rerun the model as TubeCrush-Iter2_0000.rad
1. Instead of using a constant contact gap thickness, it is more accurate to use the actual
thickness of the parts in contacts which is called Variable Gap in RADIOS.
2. Edit the self contact and enter the best practices values of Igap = 2: Variable gap
+ scale correction
Tip: For the most physical response, the contact gap for shells should use the shell
thickness.
5. Two more warnings message #343 and #477 are in the output file. The following
warning is most concerning and is the cause for the strange behavior of the
results.
WARNING ID : 343
** WARNING : INITIAL PENETRATIONS IN INTERFACE
DESCRIPTION :
-- INTERFACE ID : 2
-- INTERFACE TYPE : self contact
THERE ARE 860 INITIAL PENETRATIONS
( CONCERNING 550 NODES )
Step 15: Replot the results and compare to the TubeCrush-Iter1 and TubeCrush-
Iter2 runs in HyperView
*NOTE: This model blows up due to the initial penetration in the self contact of the 3 mm
shells at the top of the tube near the rigid body. The initial space between the flanges at the top
of the tube is 2.1mm but since the shells are 3 mm thick; the solver sees 0.9 mm of penetration
and applies a contact force as shown in the figure on the left to remove that penetration. The
forces from the initial penetration then cause the unphysical mesh deformation at Time=1.0 on
the right hand side. If there are no initial penetrations in a model, the contact force will be zero
at time=0.0.
Step 17: Use option to automatically reduce the contact gap thickness in the
initial penetration area and rerun RADIOSS.
1. Edit the self contact and enter the best practices values of Inacti = 6: Gap is
variable with time which will in areas of initial penetration cause the contact gap to be
reduced by the penetration value plus 5%.
2. Export the model as TubeCrush-Iter3_0000.rad and run in RADIOSS.
3. While the model is running use a text editor to edit the TubeCrush-
Iter2_0000.out file and search for WARINING.
*NOTE: Edit the TubeCrush-Iter3_0000.out file in a text editor and notice the WARNING
message about the initial penetrations still exists but now there is an additional message
concerning the automatic reduction of the initial gap.
REDUCE INITIAL GAP
REDUCE INITIAL GAP
WARNING ID : 343
** WARNING : INITIAL PENETRATIONS IN INTERFACE
DESCRIPTION :
-- INTERFACE ID : 2
-- INTERFACE TYPE : self contact
THERE ARE 860 INITIAL PENETRATIONS
( CONCERNING 550 NODES )
Plot the contact forces to make sure they are zero at time = 0.0
NOTE: A comparison of the rigid body master node Z-Displacement between iteration 1
constant gap and iteration 3 variable gap is shown below. The displacement of iteration 3 is
less because it does not crush as much.
NOTE: Plotting the rigid wall forces for the iteration 1 and iteration 3 below also shows the
effect of the gap thickness. Iteration 3 shows an earlier rebound as a result.
This message is similar to the FAQ in the RADIOSS help which deals with a self contact with
constant gap.
RADIOSS > Frequently Asked Questions:Contact Interfaces: What is the meaning of:
WARNING ID 94?
WARNING ID: 477 occurs when the maximum variable contact gap thickness > twice the
smallest side length of shell elements on the master side. When this happens there is a
possible over stiffening of the structure if the element is compressed by 50% as shown in the
following picture.
UNITS: Length (mm), Time (ms), Mass (kg), Force (kN) and Stress (GPa)
Simulation time: 25.0 ms
Boundary Conditions:
o All translation degrees of freedom are constrained on the edge nodes of the
block
o
A total force of 0.72 kN is apply in the negative X direction on the end of the
cantivlever beam
Johnson-Cook Elastic Plastic /MAT/PLAS_JOHNS (Steel plate) parameters
= 7.83e-6 kg/mm3 [Rho_I] Initial density
E = 210 GPa [E] Youngs modulus
= 0.3 [nu] Poissons ratio
a = 0.27 GPa [a] Yield Stress
b = 0.45 GPa [b] Hardening Parameter
n = 0.6 [n] Hardening Exponent
Problem Setup
Copy the file: bolted_cantilever-Start_0000.rad to a working directory.
Step 2: Review the material properties of the Steel plate and Steel bolt to
verify that the parameters match those given in the exercise introduction
Step 4: Review the already defined cross section which is used to calculate the
cross sectional force in the bolt.
1. Select the Model tab, under Components turn off the plate and block parts.
2. Under Cross Section right click on bolt_section and select Review.
Note: The RADIOSS section is defined with a group of element and group of nodes.
Step 6: Define a Type 24 interface named contact which includes all the parts in
the model.
In the first simulation, the default Inacti option will be used.
1. In the Entity Editor right click and select Create > Contact and enter self contact for
the name.
2. Select Card Image = TYPE24 and select Yes to the pop-up message about changing the
card image.
3. Define the best practices value of Isft=4, which uses the minimum stiffness between the
slave and master entities. Enter Fric= 0.2, and Inacti = 1000, which means only small initial
penetrations will be taken into account and larger penetrations ignored.
4. To create a TYPE24 self-contact interface only Surface 1 needs defined. In the Surf_ID1 (S)
right click and change the entity type to Set.
5. To the left of the entity selector right click and pick Create to create a new set.
Step 7: Export the model and run in RADIOSS and run in RADIOSS
1. Using File > Export > Solver Deck export the model making sure to check the
Auto export engine file bolted_cantilever_inacti1000_0000.rad.
2. Start HyperWorks Solver Run Manager program by selecting RADIOSS from the
HyperWorks menu.
3. Submit an initial run using 4 cores by specifying np 4. Make sure the Use solver control
option is checked.
4. Note that a larger than normal amount of mass scaling is used make the
simulation run faster.
By default TYPE24 contact ignores all initial penetrations except for very numerical small
penetrations. The initial penetrations between the top of the plate and bottom of the bolt head
cause the plate to slip through the bolt and contact the next row of bolt elements.
2 3
6. Use the display mode icon, , to change to wireframe mode and rotate the model to
see the intersecting elements.
7. In the lower right core of the Penetration tab Click the close icon.
Step 9: TYPE24 option INACTI=-1, initial penetrations will be taken into account.
1. Although we could manually translate the nodes to remove this initial penetration, we can
also use the INACTI option in to cause RADIOSS to apply forces based on the amount of
initial penetration.
2. Under Groups select self contact and change to Inacti= -1 which will cause RADIOSS
to apply contact forces at t=0.0 if there are intersections.
2. Click on the vector plot icon and change the Result type: Contact / Pressure
/Normal (v) and click on Apply.
Due to Inacti=-1 being used, RADIOSS applies contact forces at t=0.0 proportional to the
amount of initial penetration. These forces can sometimes remove the initial penetration
between the parts.
3. Change back to shaded elements mode, , and then use the contour panel to plot
Result type: Von Mises(s) with Averaging method: Simple and notice the stress on in the
parts before any load is applied due to the initial penetration. If desired use the section cut,
5. Animate the model to the end of the simulation and review, Von Mises(s), Displacement
(v) and Plastic Strain (s)
Step 12: TYPE24 option INACTI=5, the master segment is shifted by the initial
penetration value.
1. Under Groups select self contact and change to Inacti= 5 which will cause RADIOSS
to apply contact forces at t=0.0 if there are intersections. This will cause RADIOSS to offset
the master segment by the penetration value. The contact will now behave as though the
elements with initial penetrations at t=0.0 are touching but without any contact force.
2. Repeat steps described as before to plot the cross section Von Mises(s), at t=5.0
3. Repeat steps described as before and review Von Mises(s), Displacement (v) and Plastic
Strain (s).
4. Last, use a text editor to edit the *0001.out engine output files for the 3
simulations and compare the %Energy Error. The simulation that used Inacti=
- 1 had a higher %Energy Error caused by the application of the initial contact
forces at t=0.0.
Step 15: Optional: plot and compare the bolt cross forces of the three models.
Step 16: Optional: plot and compare the contact and total energy of the three
models.
Learn about kinematic conditions: boundary conditions, /BCS and rigid bodies, /RBODY
Define time history output for interfaces
This exercise demonstrates how to set up 3-point bending model with symmetric boundary in Y.
UNITS: Length (mm), Time (s), Mass (ton), Force (N) and Stress (MPa)
Simulation time: in Engine file 7.0e-002 s
Only one half of the model is modeled because it is symmetric.
The supports are fixed & an imposed velocity of 1000 mm/s Z- applied on the Impactor
Model size = 370 mm x 46.5 mm x 159 mm
Honeycomb Material /MAT/LAW28: HONEYCOMB
3
= 3.0e-10 ton/mm [Rho_I] Initial density
Eij = 200 MPa [E11], [E22] and [E33] Youngs modulus
Gij = 150 MPa [G11], [G22] and [G33] Shear modulus
Elasto-Plastic Material /MAT/LAW36: Inner, Outer and Flat
3
= 7.85-09 ton/mm [Rho_I] Initial density
E = 210000 MPa [E] Youngs modulus
= 0.29 [nu] Poisson's ratio
Elastic Material /MAT/PLAS_JOHNS: Impactor
3
= 8e-09 ton/mm [Rho_I] Initial density
E = 208000 MPa [E] Youngs modulus
= 0.29 [nu] Poisson's ratio
Stress-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
STRESS 300 310 320 330 340 350 360 370 380 400
Problem Setup
Copy the file: BENDING_0000.rad to a working directory.
Step 2: Create a law 1 elastic material named Rigid Material and assign to the
Impactor and Support parts
5. Select New, enter StressStrain for the name of the new curve, and click proceed
to return to the Curve Editor.
Tip: Note that functions 5-10 used in the HCFoam definition are already defined in the model.
Step 6: Right-click the Shell Property in the Property section of the Model
Browser and select Duplicate. Rename it to Rigid Shell and assign this
property to Impactor and Support.
Note: This modeling practice, where shells are used to define the surface of a
rigid structure along with a rigid body, is useful for describing arbitrarily
shaped rigid surfaces. In this case, the shell property can be a simple 1-
integration point. The property is duplicated and N set to 1.
Step 8: Create a rigid spider for the Impactor with a calculated primary node
1. From the Model Browser right-click and select Create > Component.
2. For name, enter Rigid Bodies.
3. In the entity editor set card image as none.
4. Go to 1D page, then select the rigids panel.
5. Verify that you are in the create subpanel.
6. For dependent nodes 2-n switch to comps.
7. For primary node switch to calculate node.
8. Click comps.
9. Select Impactor, then click select
Note that the imposed velocity arrow in the graphics window may not correspond to the
actual direction defined.
Step 11: Define a Type 7 contact between the Flat (Slave) and Support
(Master) parts named Support
Tip: In this case we choose GAPmin of 0.45 because that is half of the thinnest shell thickness. The
parameter Igap of 2 gives us variable gap , INACTI of 6 will adjust the gap if any initial penetrations exist.
Also, when selecting friction in contact, set to 0.1 here, the flag Iform should be set to 2 for Stiffness
formulation.
Hint: All of the slave nodes can be selected by component. Similarly the master segments can be chosen
by selecting the corresponding component. This will generate /GRNODE/PART and /SURF/PART/EXT,
respectively.
Step 15: Select the Delete Unused dialog box by going to Tools > Unused and
review the model to determine that all properties and materials are assigned.
Tip: This dialog only displays entities within the model which are unassigned through
reference in other entities and can be a quick way to determine if there are curves,
groups, properties, or materials which are unassigned.
Step 18: Review the output files for this run and view the results in HyperView
von Mises (Max) Stress Contour (MPa) Only steel parts shown
Step 19: Plot a graph of the Z-normal force resultant over time from the
Imp_Outer interface from the file 3PBENDINGT01 using HyperGraph
Model Description
UNITS: Length (mm), Time (ms), Mass (Mg), Force (N) and Stress (MPa)
Simulation time *_0001.rad [0 0.03 seconds]
Boundary Conditions:
o The rigid plate applying the load is fixed except for the imposed displacement in the
vertical direction
o The bottom of the bottle is supported by a rigid wall
o A 0.25 inch imposed displacement is applied to the rigid plate
Objective is to calculate the max top load for bottle.
Materials
Elastic Plastic Piecewise Linear Material /MAT/LAW36 (Bottle)
-10 3
= 9.75e Mg/mm [Rho_I] Initial density
E = 800.0 MPa [E] Youngs modulus
= 0.4 [nu] Poissons ratio
fct_ID1 Yield stress function
Exercise 9a: Evaluate the Time Step and review the Model
Purpose: Use various methods to determine the time step of a simulation and
improve the time step by modifying elements
3. Click the icon, navigate to the directory with the starting file and select the initial model
bottle_timestep_0000.rad
4. Using the Model Browser, review the material properties, shell properties, contact, rigid
wall, and boundary conditions. The top load is applied using an imposed displacement
applied to a rigid meshed plate. The bottom of the bottle is supported by a rigid wall,
/RWALL.
3. Submit an initial run using 4 cores by specifying np 4. Make sure the Use solver control
option is checked.
4. Notice the low time step, DT = 4.41E-08, and the long estimated remaining solution time,
REMAINING TIME= 3566.03 s. Note: this number will vary depending on the speed of
the computer and the number of CPUs used for the solution.
5. Stop the run by checking the Stop checkbox and selecting Send Command then Close
6. Using a text editor open the bottle_timstep_0001.out and find the section that shows
the run cycles. Notice that it says that SH_3N (a tria element) element number 9279 is
controlling the time step.
Next we will use a few different methods to find the time step of the whole model
1. Start HyperView then File > Open > Model and click the and select
the bottle_timstepA001.
3. Select the contour icon, and select Result type: Time Steps (s) click Apply
4. Press the M key to turn on the mesh.
5. Rotate the model to find the location with the low time step which will be dark blue by
default. Zoom in to see that some very small elements are causing the small time step
6. In HyperMesh, press F10 for the Check Elems panel, toggle the Time
7. Set the time step for all types: 1.0e-7, and click check elems, then save failed
8. To isolate the elements that failed, press F5, for the Mask panel click elems, retrieve;
elems, by adjacent; elems, reverse; then the green mask button
9. Zoom in and find the 5 small thin elements that are controlling the timestep.
10. To delete the small elements, press F2 for the delete panel.
11. Select the 5 small elements and click delete entity, then return
12. Access the Edges panel by Pressing SHIFT+F3 keys
13. With comps entity selector highlighted click once on an element on the
screen
14. Enter tolerance = 0.1 and click preview equiv to see the nodes that
will be equivalence
15. Click equivalence to equivalence the nodes
16.Select File > Export > Solver Deck and click enter
bottle_element_0000.rad click Export
Purpose: Run updated model using the default element time step calculated by
RADIOSS.
1. Start HyperWorks Solver Run Manager program and submit an initial run using 4
cores by specifying np 4 . Make sure the Use solver control option is checked.
1. Using a text editor open the bottle_element_0000.out
1. Use Find in the text editor to search for time step in the file and notice that the time step is
now much higher listed for each element type and the nodal time step.
2. Compare the smallest element time step listed in the bottle_element_0000.out file to
the time step listed in the output while the model is running or in the
bottle_element_0001.out file. Why are they not the same?
3. After the solution has finished, look at the end of bottle_element_0001.out for
ELAPSED TIME = ________________ .
4. Start HyperView and use File > Open > Model to load the bottle_element.h3d file.
6. Animate the bottle crush by clicking on the animate button on the toolbar,
Next click File > Open > Report Template, and select the report template file
force_displacement.tpl Click on PLOT_FILE_1 and select the file open icon and
select the bottle_elementT01. Last click Apply and a second page will be added to
HyperView with the force displacement plot of the load applicator as shown below.
7. Save a HyperView session file of these results, on File >Save > Session enter the name
bottle_results.mvw and click save.
Purpose: Re-run the model using a constant nodal time step to reduce total
runtime RADIOSS.
1. Copy the bottle_element_0000.rad and bottle_element_0001.rad files to
bottle_nodal1_0000.rad and bottle_nodal1_0001.rad into the new directory.
2. Using a text editor and open bottle_nodal1_0001.rad to add the following Engine
Control Card to enable nodal time step calculation.
/DT/NODA/CST
0.67 2.0e-6
/ANIM/NODA/DMAS
NOTE: The above command will cause RADIOSS to use the nodal method of calculating the
time step. The 0.67 is the time step scale factor while 2.0e-6 is the minimum constant time
step (CST) that will be used by the RADIOSS simulation. This /DT/NODA/CST option will cause
RADIOSS to add mass to any node whose time step is less than 2.0e-6 / 0.67 3.0e-6.
When mass is added to a node its time step increases. /ANIM/NODA/DMAS causes RADIOSS
to create a contour output of the, change in mass / original mass.
3. As before, start the HyperWorks Solver Run Manager and start running the
bottle_nodal1_0000.rad file using 4 cores.
4. Observe the amount of mass added in the last column of the bottle_nodal_0001.out
simulation by text editing the bottle_nodal_0001.out file. This MAS.ERR is the
(change in mass) / (original mass). To get % mass added multiply this number by 100. So
in this simulation 117.4% mass has been added which is too much.
NOTE: Good engineering judgement must be used to determine how much mass is an
acceptable amount to be added to a model to reduce the runtime. Adding too much mass can
affect the physics of a drop or impact simulation. This is because the object being simulated
weighs more than the real part. In general it is recommended to keep the amount of mass
added to less than 5% but more may be acceptable depending on a particular simulation.
8. In HyperView click on the add page icon and add another HyperView page.
9. If needed, change the page type from HyperGraph 2D to HyperView using the icon
10. Click on the new empty window and use File > Open > Model and the select the
bottle_nodal1.h3d file and click Apply.
11. Select the contour icon, and contour plot Mass Change (s) which is the (change in
nodal mass) / (original nodal mass).
It is always good to understand where the mass is being added in your model. It can be
important to minimize the amount of mass is added to critical areas of the design.
7. Did the amount of mass added by using a time step of 2.0e-6 cause a large difference in
the results compared to the results from the first simulation with elemental time step?
8. Copy the files to bottle_nodal1_0000.rad and bottle_nodal1_0001.rad files to
bottle_nodal2_0000.rad and bottle_nodal2_0001.rad
12. Determine a time step that adds less mass by editing the bottle_noda11_0000.out
RADIOSS starter output file and searching for the NODAL TIME STEP. Starting with the
page that contained the element time step contour plot by using .
16. Change page to bottle_element.h3d results and change to two windows using this icon,
to
17. Click on the new empty window and use File > Open > Model and the select the
bottle_nodal2.h3d file and click Apply.
18. Select the contour icon, and select Result type: Stress (t): vonMises to create a
contour plot of the stress of the crushing bottle in HyperView.
19. How does the stress results compare to the first simulation with element time step?
20. Click on the Open Report Panel Icon to plot the force
displacement curve for the nodal time step. Click beside the PLOT_FILE_1 and pick the
bottle_nodal2T01 file. Uncheck Use report colors and select Overlay under
Mode and click Apply.
21. How does the top load force results compare to the other two simulations?
22. After the solution has finished, look at the end of bottle_nodal2_0001.out for
ELAPSED TIME = ________________ .
23. How much faster does this nodal time step with mass scaling compared first run with
element time step ______________?
Purpose: Run the model using the Advanced Mass Scaling (AMS) method to
further reduce run time. The load needs to be applied slower to get a converged
quasi-static maximum top load force and remove vibrations in the force
displacement curve. By using AMS with a higher time step and apply the load
over a longer time, a converged quasi-static maximum top load force can be
calculated.
1. Launch HyperMesh, then select Preferences > User Profile > BRADIOSS Block140
2. From the pull down menu bar, File > Import > Solver Deck
Due to the way AMS changes the mass matrix, high frequencies are damped out of a system.
This makes AMS best used for quasi static or low speed simulation where the high frequency
spectrum is not important to the results.
Experience has shown that a good starting time step for an AMS simulation is 10x the minimum
NODAL TIME STEP from the starter output file or 10x a valid /DT/NODA/CST simulations time
step.
To eliminate the vibrations in the force deflection curve, the load needs applied 10X slower.
9. In Model Browser > Load Collector, click on 3disp and enter scale_x=10. This apply a
scale factor of 10x to the abscissa so the load will be applied from 0 0.3 seconds.
10. In Model Browser > Card click on ENG_ANIM_DT and enter, Tfreq=0.03 to that 10
animation files are created in 0.3 seconds of run time.
11. In Model Browser > Card click on ENG_RUN T_STOP=0.3 so that the simulation ends at
0.3 seconds.
12. Select File > Export > Solver Deck and click enter bottle_ams_0000.rad click
Export
13. As before, start the HyperWorks Solver Run Manager and start running the AMS model
bottle_ams_0000.rad using 4 cores by specifying the option np 4.
14. Compare the total run time to the other simulations.
15. Switch to the first page that contained the element time step contour plot and nodal time
step contour plot change the page to three windows using this icon,
16. Click on the new empty window and use File > Open > Model and the select the
bottle_ams.h3d file and click Apply.
17. Select the contour icon, and select Result type: Stress (t): vonMises to create a
contour plot of the stress of the crushing bottle in HyperView.
18. How does the stress results compare to the first simulation with element time step?
Using AMS to increase the time step, allows the load to be applied slower and results still be
obtained in a reasonable amount of time.
20. The AMS force deflection curves are smoother then element and nodal results because the
load was applied slower resulting in less dynamic affects. The AMS maximum top load
force is very close to the converged quasi-static force of 205 N.
Your task is to review the model for completeness and finish the model setup using the information below
to get a valid simulation.
Problem Setup
Copy the files: phonedrop_start_0000.rad to a working directory.
Assume the parts interact with each other with a friction value of 0.2.
Loading Conditions
The phone drops from a height of 1.5 meters in the Z-direction to a flat floor on the Z plane.
The forces between the phone and the floor need to be recovered.
The total simulation time is 1 millisecond. Animation output requests should include plastic strain for
solids and shells, vonMises stress, tensor of shell membrane stress, and nodal added mass.
Your task is to review the model for completeness and finish the model setup using the information below
to get a valid simulation.
Front Rail
Crush Box
Bumper Beam
Problem Setup
Copy the files: bumper-start_0000.rad to a working directory.
Connections
The Crush Box is connected to the Bumper Beam with a MIG Weld along the top and bottom
flanges (4 places). Green box shows one location.
The Crush Box is attached to the Front Rail with bolts of diameter 10 mm (8 places). Red circle
shows one location.
Impact Conditions
The bumper impacts a finite sized wall centered at (-503.5, -425, 180) that is 240 mm high
by 560 mm wide. Assume friction of 0.2.
The initial velocity of the frame is 4.5 mm/msec.
Assume all the parts have a friction coefficient of 0.2
The total simulation time is 30 milliseconds
Bumper Results
HyperStudy Interface
HyperStudy Overview
For response extraction, HyperStudy uses HyperGraph readers and hence any result that
can be read by HyperGraph can also be read by HyperStudy. HyperStudy can also extract any
value from an ASCII output file.
The material to be characterized is a 6063 T7 Aluminum: it has an isotropic elasto-plastic behavior which
can be reproduced by a Johnson-Cook model without damage (RADIOSS Block Law2), defined as
follows:
For the simulation results, engineering strains will be obtained by dividing the
displacement of node 1 by the reference length (75 mm), and engineering stresses will be
obtained by dividing the force in section 1 by its initial surface (12.012 mm2).
Problem Setup
You will need to copy the files: TENSILE_TEST_0000.rad, TENSILE_TEST_0001.rad, and
exper.xy
5. Select E, the Youngs Modulus value, by starting at the beginning of row 51 and highlighting the
first 20 fields.
Tip: To assist you in selecting 20-character fields, press CTRL to activate the Selector (set to 20
characters) and then click the value.
6. Right-click on the highlighted fields and select Create Parameter from the context sensitive
menu.
7. In the Parameter Varname_1 dialog, enter E_Young in the Label field.
8. Set the Lower Bound to 50000, the Initial Bound to 60400, and the Upper Bound to 70000.
9. In the Format field, enter %20.5f.
1. To start a new study, click File > New from the menu bar, or click on the toolbar.
2. In the HyperStudy Add dialog, enter a study name, select a location for the study, and click
OK.
3. Go to the Define models step.
5. Click Import Variables. Five design variables are imported from the
TENSILE_TEST_0000.tpl resource file.
6. Go to the Define design variables step.
7. Check the design variable's lower and upper bound ranges.
8. Go to the Specifications step.
Difference between experimental stress and RADIOSS at Strain equal 0.02 (1)
9. Click OK.
10. Repeat steps 4.5 through 4.8 to add a second vector labeled Force_sim.
Problem Setup
You will want to copy the files: boxbeam_0000.rad, boxbeam_0001.rad
7. Right-click on the highlighted fields and select Create Parameter from the context menu.
8. In the Parameter - varname_1 dialog, Label field, enter Upper part.
9. Set the Lower bound to 0.5, the Initial value to 1.0, and the Upper bound to 2.0.
10. Set the Format to %20.5f.
16. In the Parameter - varname_2 dialog, Label field, enter Lower part.
17. Set the Lower bound to 0.5, the Initial value to 1.0, and the Upper bound to 2.0.
18. Set the Format to %20.5f.
19. Click OK.
1. Start HyperGraph.
3. From the menu bar, click File > Open > Document.
4. In the Open Document dialog, open the boxbeam1.tpl file. The text editor displays the following
design variables that are defined by Templex parameter statements:
{parameter(t1,"Upper part",1.0,0.5,2.0)}
{parameter(t2,"Lower part",1.0,0.5,2.0)}
7. Click . The parameterized /PROP/SHELL cards, which reference the design variables,
highlight.
9. Repeat steps 2.5 through 2.7, and search for /PROP/SHELL again. You will find:
2. To start a new study, click File > New from the menu bar, or click on the toolbar.
3. In the HyperStudy Add dialog, enter a study name, select a location for the study, and click
OK.
4. Go to the Define models step.
5. Add a Parameterized File model.
a. From the Directory, drag-and-drop the boxbeam1.tpl file into the work area.
6. Click Import Variables. Two design variables are imported from the boxbeam1.tpl resource
file.
7. Go to the Define design variables step.
8. Review the design variable's lower and upper bound ranges.
9. Go to the Specifications step.
1. Click the Iteration History tab to display data in a tabluar view. The optimal design is highlighted
green, the infeasible designs are shown with red text, and the violated constraints are indicated
in bold text.
2. Click the Iteration Plot 2D tab to plot the iteration history of the study's objectives, constraints,
and design variables.
Tip: In the initial design, the design was infeasible as indicated by the large circular marker for
the first iteration. A view of the constraint plots shows that the second constraint was violated
in the initial design. Initially, the optimizer added some weight in order to satisfy the design
constraints. Notice that both constraints are near their bounds in the optimal design.