LS-DYNA Implicit Workshop

Problem #1: Tensile Test

____________________________________________________________________________________

Objectives
• Learn how to activate LS-DYNA’s implicit mode.
• Learn how to select linear or nonlinear analysis.

Problem Description
A static tensile test is simulated using shell elements. One end of the specimen is
constrained, while concentrated nodal loads are applied at the other end. Uniform
stresses develop in the narrowed center section.

Input Filename: tensile1.k

Procedure
Copy the input file to your local directory. Using an editor, view the input file and
answer the following questions:

1. Which shell element formulation is used? default (Belytschko-Tsay type 2) _____

2. How many steps are used to apply the load? one ____________________________

Run the simulation, and post-process using LS-POST. Record below the total applied
load and total tip displacement (HINT: the ascii database NODFOR includes data for
each of the tip nodes).

3. Applied load 1000.0____ Tip displacement 2.23e-2___

Livermore Software Technology Corporation 1

Modify the input deck to increase the applied load 100x, and rerun the simulation.
Record the new data:

4. Applied load 100,000___ Tip displacement 2.27_____

5. Is the simulation linear or nonlinear ? (Does the tip displacement scale linearly with
applied load?)

6. What types of nonlinearity exist in this problem? ( ) material (X) geometric

( ) contact

Modify the *CONTROL_IMPLICIT_SOLUTION keyword to perform a linear

simulation, and record the results below for the large applied load:

7. What is the reaction force at the fixed end of the beam? 98,940 _______________

8. Why does the reaction force not equal the applied load? linear analysis with
nonlinear element (cross-sectional area has changed) ________________________

Modify the input deck to use the linear element formulation type 21, and repeat the
simulation.

9. Does the reaction force now match the applied load? (X) yes ( ) no

Livermore Software Technology Corporation 2

 LS-DYNA Implicit Workshop

Problem #2: Car Tire

____________________________________________________________________________________

Objective
• Learn to set up and run implicit dynamic and eigenvalue analyses.
• Learn the advantages of implicit dynamic compared with explicit dynamic analysis.

Problem Description
A coarsely meshed wheel and tire assembly is “kicked” by a brief transient load. The
dynamic response of a node at the top of the tire is monitored.

Input Filename: tire.k

Procedure
Copy the input file to your local directory. Using an editor, view the input file and
answer the following questions:

1. What loads and boundary conditions are applied to the tire? SPC constraints at
bottom, pressure spike on lower sidewall ___________________________________

Run the (explicit) simulation using the original input file. Note the times below.

method 2a. termination time 2b. CPU time

EXPLICIT 0.10 seconds 505 seconds

Livermore Software Technology Corporation 3

Copy the input deck to a new directory. Modify the input to run an implicit eigenvalue
analysis, computing the lowest 20 eigenvalues. Run the simulation and postprocess the
results.

3. What is the lowest frequency? 8.63 __ Hz What is it’s period? 0.116 _ sec.

Copy the input deck to a new directory. Modify the input to run a linear implicit dynamic
analysis. Increase the termination time to 1.0 seconds and run the simulation.

method 4a. termination time 4b. CPU time

linear IMPLICIT dynamic 1.0 seconds 79 seconds

Plot the Y-displacement of node ID 1000982 from the NODOUT file, and compare with
results from the explicit simulation.

5. Do the explicit and implicit results agree? Yes, until explicit TERM=0.10 ________
6. How much faster is the implicit simulation for this (tiny) model? 505/7.9 = 64x___
7. How could results of the eigenvalue analysis be used to select the timestep and output
interval? Need at least 10 steps per period for shortest period of interest _________

Copy the input deck to a new directory. Introduce some Newmark damping (try
GAMMA=0.60 and BETA=0.38). Run, plot the Y-displacement of node ID 1000982
from the NODOUT file, and compare with results from the previous simulations.

8. Does the damping most affect high or low frequency response? Both, in this case _

Optional:
Copy the input deck to a new directory, and modify it to perform a nonlinear implicit
dynamic simulation. Reset the termination time to 0.1 seconds and run the simulation.
Compare results for node ID 1000982.

method 8a. termination time 8b. CPU time

n/l IMPLICIT dynamic 0.10 seconds 300 seconds

same, but ISTIF=9999 1.0 seconds 163 seconds

Livermore Software Technology Corporation 4

 LS-DYNA Implicit Workshop

Problem #3: Rigid Tip

____________________________________________________________________________________

Objective
• Learn the limitations of a linear approximation using a simple example.

Problem Description
A tip load is applied to a cantilevered beam made of shell elements. The tip of the beam
is constrained to be rigid.

Input Filename: rigid_tip.k

Procedure
Copy the input file to your local directory. Using an editor, view the input file and
answer the following questions:

1. How is the rigid tip modeled? *CONSTRAINED_NODAL_RIGID_BODY ______

2. How long is the beam? 5 inches _________________________________________

Run the simulation using the original input file.

3. What type of analysis was performed? Linear static _________________________

4. What is the tip deflection? 0.1635 inches __________________________________

Livermore Software Technology Corporation 5

5. What is the angle of rotation of the rigid body (Note: rotations in the RBDOUT file
are given in radians)? 4.8445e-2 rad = 2.776 ________________________degrees

6. What is the maximum X-stress at the supported end? 22,463 psi _______________
7. What is the maximum X-stress in the “rigid” elements? -38,677 psi_____________

Copy the input file to a new directory. Modify the input file to perform a nonlinear
analysis, and re-run the simulation.

8. What is the tip deflection? 0.163 in _____________________________________

9. What is the angle of rotation of the rigid body? 2.769 __________________degrees
10. What is the maximum X-stress at the supported end? 22,424 psi _______________
11. What is the maximum X-stress in the “rigid” elements? 7.0e-9 psi ______________

Copy the input file to a new directory. Modify the input file to again perform a linear
analysis, but with the constrained nodal rigid body removed. Run the simulation.

12. What is the tip deflection? 0.164 in ______________________________________

13. What is the maximum X-stress at the supported end? 22,463 psi _______________
14. What is the maximum X-stress in the tip elements? 775 psi ___________________

Why does stress develop in the “rigid” elements during the linear analysis? (Hint: plot
the (axial) X-displacement along the length of the beam for each simulation.)

(see rigid_tip_solution.ppt)

Livermore Software Technology Corporation 6

 LS-DYNA Implicit Workshop

Problem #4: Elastic-Plastic Tensile Test

____________________________________________________________________________________

Objectives
• Learn how to observe convergence behavior of nonlinear equilibrium iterations.
• Learn to use automatic time step control for nonlinear problems.
• Learn the benefits of force vs. displacement controlled simulations.

Problem Description
A static tensile test is simulated using shell elements and a nonlinear, elastic-plastic
material model. One end of the specimen is constrained, while concentrated nodal loads
are applied at the other end. Uniform stresses develop in the narrowed center section.

Input Filename: tensile2.k

Procedure
Copy the input file to your local directory. Using an editor, view the input file and
answer the following questions:

1. Which material model is used? What is the yield stress? Mat 37, sigy=200_______
2. How is load applied? Nodal forces _______________________________________
3. How many steps are used to apply the load? 100 ____________________________

Run the input deck. Does the job run to completion? (X) normal termination
( ) error termination

4. At what time does the solution begin to struggle? time = 0.67 ________________

Livermore Software Technology Corporation 7

Using the postprocessor, plot the X-displacement of an end node vs. time.

5. What is happening at the time shown above? plastic yielding through X-section _

6. max end displacement 7.07 _ max eff. stress 317 __ max eff. strain 9.62%

Activate the nonlinear print flag to get more information about the nonlinear solution
process, and repeat the simulation.

7. What two methods are available for this? “<ctrl-c> nlprint” or NLPRINT=1 ______

8. How many cycles are used in the simulation? 2366 _________________________

Switch from load control to displacement control, and repeat the simulation (Hint:
helpful keywords are commented out in the original input deck). Using the
postprocessor, again plot the X-displacement of an end node.

9. max end displacement 7.0 __ max eff. stress 310 __ max eff. strain 8.61%

10. Why is this problem easier to solve? prescribing displacements eliminates unknowns

Return to the original input deck, and activate automatic time step control (IAUTO=1 on
the keyword *CONTROL_IMPLICIT_AUTO). Use 200 as the optimum iteration count,
and set the maximum stepsize to 0.050. Repeat the simulation.

11. What happens? Normal termination, 23 steps, 413 cycles ____________________

12. max end displacement 7.06 _ max eff. stress 317 __ max eff. strain 9.61%

Using the “ASCII” menu, load the GLSTAT database and plot the step size vs. time.

13. When does the step size change? Why? Quickly rises to DTMAX=0.05 because less
than 200 iterations are needed for each step. ________________________________

Livermore Software Technology Corporation 8

 LS-DYNA Implicit Workshop

Problem #5: Door Beam

____________________________________________________________________________________

Objective
• Learn to perform a springback simulation using an entirely static implicit analysis.
• Learn to specify key points during the solution which must be reached exactly.
• Learn about the available displacement convergence norm options.

Problem Description
A doorbeam subassembly is deformed by a rigid pole. Shell elements are used
throughout, and nodal rigid bodies are used to spotweld the components of the doorbeam.
The pole is displaced to deform the doorbeam, then retracted to evaluate springback.

Input Filename: doorbeam.k

Procedure
Copy the input file to your local directory. Using an editor, view the input file and
answer the following questions:

1. What type of contact interface is selected? Why? _NODES_TO_SURFACE because

doorbeam normals point everywhere ______________________________________

2. Why is the doorbeam chosen as the slave side? It’s normals are most erratic _____

Livermore Software Technology Corporation 9

Run the simulation and observe the convergence behavior.
Activate the flag IGAPF=2 on optional card “C” in the *CONTACT_… keyword.
Repeat the simulation.

3. Which gap flag produces the best convergence behavior? Sticky (400 vs 656 cycles)

Postprocess the results, and plot the slave interface force. Save the interface force curve
as file “curve1”.

4. What is the strategy for solving the springback problem? How is the applied load
removed? Simply retract the indentor, since analysis is static __________________
5. At what time is the full load applied? (HINT: Check the input file.) 100 ________
6. Why are the maximum reaction force and springback predictions from this simulation
misleading? Didn’t compute a solution at peak load _________________________

Define a key point at time=1.0 using the *CONTROL_IMPLICIT_AUTO keyword. Use

a maximum step size of 0.05 throughout the entire simulation. Repeat the simulation,
and plot the new interface force curve. Save this curve as file “curve2”. Compare these
results to “curve1”.

7. Is the peak load computed at the correct time? yes __________________________

8. Why is the force-deflection behavior not smooth? Poor equilibrium estimate _____

Select the alternate displacement tolerance scheme using DNORM=1 on

*CONTROL_IMPLICIT_SOLUTION, and repeat the simulation.

9. Does the alternative displacement tolerance become (X) more or ( ) less strict when
total displacements are large, as they are at the end of this problem?

10. Is the force-deflection curve more smooth? Why? More strict tolerance (d3hsp:
||u||=1.e-4 @t=1.0, compared with ||u||=1.e+0 in 7 above.)______________________

11. What is the springback deflection at the center of the doorbeam? 3.38 __________

Livermore Software Technology Corporation 10

 LS-DYNA Implicit Workshop

Problem #6: Truck Bumper

____________________________________________________________________________________

Objectives
• Learn the behavior contact interfaces in static implicit simulations.
• Learn how to set input parameters for implicit contact problems.

Problem Description
The rear bumper of a truck is modeled using shell elements. A solid, rigid bar is
displaced into the bumper, causing plastic buckling of the support. Post-buckling
response is determined.

Input Filename: bumper.k

Procedure
Copy the input file to your local directory. Using an editor, view the input file and
answer the following questions:

1. What type of contact interface is used? *CONTACT_SURFACE_TO_SURFACE _

2. Is the thickness of the shell elements considered? Yes (SHLTHK=1) ___________

Run the simulation and postprocess the results.

Livermore Software Technology Corporation 11

3. Why does the simulation stop? Convergence failure at step #1 ________________

Temporarily increase the default displacement convergence tolerance to DCTOL=0.005

using keyword *CONTROL_IMPLICIT_SOLUTION, and repeat the simulation. Stop
the job after about 10 steps using the interactive command “<ctrl-c> stop”. Postprocess
the results, and magnify the displacements 10x.

4. What is wrong with the contact interface? Sticking along lower edge___________

Activate the flag IGAPF=2 on optional card “C” in the *CONTACT_… keyword.
Repeat the simulation.

5. Does the job run? Why? No, energy explosion at first contact _________________

To reduce the severity of the initial contact, decrease the contact penalty scale factor
using SLSFAC=0.01 on *CONTROL_CONTACT. Rerun the simulation.

6. Does the interface separate cleanly? yes ___________________________________

Plot the resultant interface force for the slave side of the interface. Save the curve data.

7. Why is the curve not smooth? Poor estimate of equilibrium (sloppy tolerance)_____

Restore the convergence tolerance DCTOL to it’s default value of 0.001, and repeat the
simulation. Compare the new interface force with the saved data.

8. Is the new curve smoother? Why? Yes, more accurate equilibrium______________

Livermore Software Technology Corporation 12

 LS-DYNA Implicit Workshop

Problem #7: Implicit – Explicit Switching

____________________________________________________________________________________

Objective
• Learn to use the implicit method for static initialization of an explicit simulation.
• Learn to switch formulations during a simulation.
• Learn to perform a simple eigenvalue analysis.

Problem Description
A cantilevered strip of shell elements is loaded using the static implicit method. The
analysis type is then switched to explicit, the load is removed, and the dynamic response
is simulated. The first fundamental response frequency is verified by eigenvalue
analysis.

Input Filename: ie_shell.k

Procedure
Copy the input file to your local directory. Using an editor, view the input file and
answer the following questions:

1. How is the implicit – explicit switching activated manual method, curve ID=9 ____
2. When does the simulation run implicit? Explicit? Implicit until t=0.01, then explicit
3. How many implicit steps will be taken? 20 _________________________________
4. How is load applied? When is load removed? End force, max @ t=0.01, then zero _

Livermore Software Technology Corporation 13

Run the simulation, and postprocess the results. Plot the Y-displacement of a tip node vs.
time.

5. Estimate the period of response for the first fundamental mode: 0.0075 ___ seconds.

Modify the input deck to conduct an eigenvalue analysis. Solve for the lowest five
eigenvalues. Repeat the simulation. Observe the names of the two new output files
which are created for eigenvalue analyses.

6. Which two new output files are created? d3eigv, eigout _______________________

View the file eigout using a text editor.

7. What is the period of the first fundamental mode? 0.00744 sec _________________

Postprocess the binary file d3eigv using LS-POST. View the mode shapes.

8. What is the meaning of the “time” associated with each mode shape? Freq (cycles) _

Livermore Software Technology Corporation 14

 LS-DYNA Implicit Workshop

Problem #8: Springback using DYNAIN file

____________________________________________________________________________________

Objective
• Learn to generate a DYNAIN output file at the end of a simulation.
• Learn to apply artificial stabilization in a multi-step springback simulation.

Problem Description
This exercise involves two simulations. First, a cantilevered strip of shell elements is
loaded using a dynamic explicit simulation. An output file named “dynain” is created at
the end of this simulation. A second, implicit simulation is then performed which reads
the dynain file and computes the springback deformation.

Input Filename: cant_dynain_load.k, cant_dynain_springback.k

Procedure
Copy the input files to your local directory. Using an editor, view the first input file and
answer the following questions:

1. What keyword creates the DYNAIN file?

*INTERFACE_SPRINGBACK_LSDYNA _________________________________
2. What material model is used? Elastic _____________________________________

Run the first simulation, and postprocess the results.

Livermore Software Technology Corporation 15

3. What is the initial Y-coordinate for all nodes? 0.0 ___________________________
4. What is the maximum effective stress at the end of the loading? 146,475 _________

View the contents of the DYNAIN file which was created at the end of the first run.

5. Which keywords are used in the DYNAIN file? *NODE,

*ELEMENT_SHELL_THICKNESS, *INITIAL_STRESS_SHELL _____________

View the contents of the second input file, and answer the following:

6. Could this file be easily created from the first input file? How? Subset of initial
input, plus *CONTROL_IMPLICIT section ________________________________
7. How many steps will be used in this simulation? 4 __________________________
8. How is load applied in this simulation? Internal initial stress ___________________

Run the second input file, making sure the DYNAIN file produced by the first run is
available for *INCLUDEing, and postprocess the results.

9. Is the termination time reached? Why? No, convergence failure step 1, 80 iter. ____

Activate automatic time step control and artificial stabilization, and repeat the simulation.
Postprocess the results, and plot the Y-coordinate of a tip node vs. time.

10. Does the springback deflection occur uniformly, or abruptly? Abrupt at end of sim.
11. After springback: max. effective stress = 2081 __ max. Y-coordinate = 6.67e-3 _

Modify the artificial stabilization scale factor to SCALE = 0.050, and repeat the
springback simulation. Postprocess, and plot the Y-coordinate of the tip node again.

12. How has the evolution of springback deflection changed? More gradual _________
13. After springback: max. effective stress = 5765 __ max. Y-coordinate = 7.3e-2 __

Modify the nonlinear convergence test to compare the displacement increment to the total
displacement over the current step (DNORM=1), repeat the simulation, and postprocess.

14. After springback: max. effective stress = 1997 __ max. Y-coordinate = 7.43e-4 _

Livermore Software Technology Corporation 16

 LS-DYNA Implicit Workshop

Problem #9: Ellipsoidal Dome

____________________________________________________________________________________

Objectives
• Learn the behavior of different element formulations in static implicit simulations.
• Learn how to minimize hourglass problems.

Problem Description
A static load is applied to the center of an ellipsoidal dome. Shell elements are used.
Nodes at the base of the dome are constrained, and included in a NODFOR output
database.

Input Filename: ellipse.k

Procedure
Copy the input file to your local directory. Using an editor, view the input file and
answer the following questions:

1. What is the maximum value of the applied load? 500,000 _____________________

2. How many load steps are used? 100 + auto _______________________________

3. What shell element formulation is used? default (B/T) ______________________

Livermore Software Technology Corporation 17

4. What hourglass control type is used? type 4 (stiffness) ______________________

Run the simulation, and postprocess the results (Note: NODFOR file contains reaction
forces at constrained base nodes.)

5. What happens? Normal termination, 19 steps, 280 cycles _____________________

6. total reaction force 499,951__ center displacement -11.78 _

Repeat the simulation using the S/R Hughes-Liu shell (ELFORM=6), and postprocess the
results.

7. total reaction force 499,936__ center displacement -11.37 _

Repeat the simulation using the Fast Fully Integrated shell (ELFORM=16), and
postprocess the results.

8. total reaction force 500,005__ center displacement -10.85 _

9. What conclusion can you draw about the effect of hourglass deformation on overall
structural stiffness for this problem? Hourglassing produces “softer” response ___

Concentrated applied loads are often responsible for initiating hourglass deformation.
Edit the input deck and replace the single concentrated nodal load with pressure acting on
the adjacent segments (Hint: Helpful keywords are included in the comments inside the
input deck). Repeat the simulation using the default shell.

10. What happens? Normal termination, 15 steps, 247 cycles ____________________

11. total reaction force 519,437__ center displacement -10.20 _

Livermore Software Technology Corporation 18

 LS-DYNA Implicit Workshop

Problem #10: Adaptive Ellipsoidal Dome

____________________________________________________________________________________

Objectives
• Learn to activate mesh adaptivity in an implicit simulation.
• Learn how to minimize hourglass problems.

Problem Description
A static load is applied to the center of an ellipsoidal dome. Shell elements are used.
Nodes at the base of the dome are constrained, and included in a NODFOR output
database. Adaptivity is used to automatically refine the mesh in areas of high curvature.

Input Filename: aellipse.k

Procedure
Copy the input file to your local directory. Using an editor, view the input file and
answer the following questions:

1. How frequently will the mesh be evaluated for refinement? 0.10 sec_____________

2. How many times can each element be subdivided? 2 _______________________

3. How do you indicate which parts will be adapted? ADPOPT=1 on *PART ______

Livermore Software Technology Corporation 19

4. Which element formulation is used? type 2 (B/T)___________________________

5. How is load applied? Point load ________________________________________

Execute the simulation, and view the results with the postprocessor.

6. How many time steps and cycles were used? 16 __ steps 203 ___________ cycles

7. Applied load 519,089___ Center displacement -11.29 _

8. Does the adaptive mesh improve the hourglassing problem better, still bad locally_

Switch to pressure driven load application, and repeat the simulation. Postprocess the
results. Using the NODFOR database, verify that the load is applied correctly as the
mesh is refined.

9. How many time steps and cycles were used? 17 __ steps 372 ___________ cycles

10. Does the pressure load improve nonlinear convergence? no___________________

11. Applied load 519,089___ Center displacement -11.29 _

Experiment with shell element formulation #16.

12. Does shell type #16 improve hourglassing? Hourglassing not bad with pressure load

13. Does shell type #16 improve convergence behavior (number of steps/cycles)? Yes
(14 steps, 149 cycles) __________________________________________________

Livermore Software Technology Corporation 20

 LS-DYNA Implicit Workshop

Problem #11: Ball Pressed Into Plate

____________________________________________________________________________________

Objectives
• Learn when implicit methods can run faster than explicit methods.
• Learn to troubleshoot contact and convergence problems.

Problem Description
A hemispherical ball of brick elements is displaced into a plate of brick elements. The
plate is supported around its edges.

Input Filename: hemi.k

Procedure
Copy the input file to your local directory. The simulation is initially set up to run
explicitly. Start the job and estimate how long it will take to finish. Stop the job using
“<ctrl-c> stop” when your patience runs thin.

1. How long will the explicit job take to run? About 8 minutes ___________________

Edit the input deck and activate the implicit method. Run the simulation and postprocess
the results.

2. Does the problem run to completion? No, negative eigenvalue error message, then
divergence ___________________________________________________________

Livermore Software Technology Corporation 21

3. What is the meaning of the “negative eigenvalue” warning message? Singular
stiffness matrix _______________________________________________________

Modify the input deck to perform an eigenvalue analysis. Extract the lowest 30 modes.

4. Are any “zero” or “rigid body” modes detected? Yes, three ___________________

5. What was the problem with this model? Forgot to add constraints dx=dy=rz=0 ___

6. Why did this model run as an explicit analysis? Rigid body modes OK for dynamic

Use *BOUNDARY_SPC_SET to add appropriate constraints to the ball (use the existing
node set ID #1.) Repeat the simulation.

7. Examine the contact condition (HINT: cut a section plane normal to the Y-direction
using SPLANE buttons). How deeply does the ball penetrate the plate? 0.29_____

To reduce penetration, use the penalty scale factor to increase the contact interface
penalty stiffness by 10x (SLSFAC=1.0 on *CONTROL_CONTACT). Repeat the
simulation.

8. How long does the simulation take to run using implicit method? 4 cpu seconds ___

9. Is the penetration problem improved? Yes, g=-0.061 _________________________

Livermore Software Technology Corporation 22