© ECS / Disclosure or duplication without consent is prohibited
Improving the assessment quality of a fatigue analysisALTAIR Motion solve, FEMFAT LAB virtual iteration and FEMFAT
© ECS / Disclosure or duplication without consent is prohibited
General Workflow
Author: FEMFAT SUPPORT 2Date: 05.05.2015
Application & calibration
Data acquisition on proving ground
Load data analysis
Fatigue testing
Fatigue analysis
Virtual iteration MBS model
Workflow of fatigue testing and analysis based on road load data
M U L T I P H Y S I C
W O R K F L O W
© ECS / Disclosure or duplication without consent is prohibited
General Workflow
Author: FEMFAT SUPPORT 3Date: 05.05.2015
Acceleration
Typical responses• Accelerations
– 1-axial– 3-axial
• Displacements – Draw wire displacement sensor
• Frame torsion• Strains (directly/calibrated to
forces)– Axle– Ball joint– Link– Rod– Spring– Stabilizer
• Load cells– Mount
• Wheel force transducers
Displacement
Strain – ball jointStrain – twist beam
Load cell Wheel force transducer
© ECS / Disclosure or duplication without consent is prohibited
General Workflow
Author: FEMFAT SUPPORT 4Date: 05.05.2015
Workflow of fatigue testing and analysis based on road load data
Application & calibration
Data acquisition on proving ground
Load data analysis
Fatigue testing
Fatigue analysis
Virtual iteration MBS model
M U L T I P H Y S I C
W O R K F L O W
© ECS / Disclosure or duplication without consent is prohibited
• FEMFAT LAB – Load data analysis software
• Module Virtual Iteration: – Load data generation for simulation models
based on measurement data (test track or test bench)
• General approach– Generate external load based on internal,
measured response
• Same approach as the iteration process in the laboratory (test bench)
• Excellent convergence between measurement and simulation
• Method is automated for – MotionSolveand additional 3 rd party MBD products
General Workflow
Author: FEMFAT SUPPORT 5Date: 05.05.2015
black…measurementred…...simulation
© ECS / Disclosure or duplication without consent is prohibited
General Workflow
Author: FEMFAT SUPPORT 6Date: 05.05.2015
Workflow of fatigue testing and analysis based on road load data
Application & calibration
Data acquisition on proving ground
Load data analysis
Fatigue testing
Fatigue analysis
Virtual iteration MBS model
M U L T I P H Y S I C
W O R K F L O W
© ECS / Disclosure or duplication without consent is prohibited
General Workflow
Author: FEMFAT SUPPORT 7Date: 05.05.2015
• MotionSolve model of front axle of a passenger car (half axle)
• Computing internal forces for fatigue analysis of the knuckle with
• Measurement signals from test track• Desired (measurement)
• Damper force• Ball joint force longitudinal• Ball joint force lateral• Tie rod force axial • Spring displacement (used for model-check)
• Goal: load at wheel (4 channels)• Vertical displacement (wheel center) • Longitudinal force (wheel center) • Steering torque (wheel center) • Lateral force (tire patch)
© ECS / Disclosure or duplication without consent is prohibited
General Workflow
Author: FEMFAT SUPPORT 8Date: 05.05.2015
Measurement signals (responses)
Damper force
Tie rod force
Ball joint forces Spring displacement
MBS: requests
© ECS / Disclosure or duplication without consent is prohibited
General Workflow
Author: FEMFAT SUPPORT 9Date: 05.05.2015
Workflow of fatigue testing and analysis based on road load data
Application & calibration
Data acquisition on proving ground
Load data analysis
Fatigue testing
Fatigue analysis
Virtual iteration MBS model
M U L T I P H Y S I C
W O R K F L O W
© ECS / Disclosure or duplication without consent is prohibited
FEMFAT LAB vi Introduction
Author: FEMFAT SUPPORT 10Date: 05.05.2015
Response
Often simply and cheaply measureable• Accelerations• Displacements (relative)• Strains• Forces (internal)
Drive
High effort or not measureable• Forces (external)
• WFT• Load cell• Strain gauges
• Displacements (absolute)
Inverse non-linear problemfind load for given response
MotionSolveInput un Output yn
Load(drive )
Desired(response)
© ECS / Disclosure or duplication without consent is prohibited
FEMFAT LAB vi Introduction
Author: FEMFAT SUPPORT 11Date: 05.05.2015
MotionSolveInput un Output yn
Load(drive )
Desired(response)
Calculation of the transfer function (MBS):
F(s) = y0(s) / u0(s) noise signal and its response
Calculation of first drive:
u1(s)= F-1(s) yDesired(s)
Calculation of further iterations:
un+1(s)=un(s)+ F-1(s) (yDesired(s) – yn(s))
© ECS / Disclosure or duplication without consent is prohibited
FEMFAT LAB vi Introduction
Author: FEMFAT SUPPORT 12Date: 05.05.2015
Iteration process
un+1 = un + F -1 ( yDesired - yn )
1. Noise
2. Response of noise
6. Response = desired 5. Response
3. Transfer function 4. Drive signal
© ECS / Disclosure or duplication without consent is prohibited
FEMFAT LAB vi Results
Author: FEMFAT SUPPORT 13Date: 05.05.2015
Results: 10. iteration, rough road
© ECS / Disclosure or duplication without consent is prohibited
FEMFAT LAB vi Results
Author: FEMFAT SUPPORT 14Date: 05.05.2015
Damper force – time domain 15 th to 20 th second Ball joint force lateral – time domain 15 th to 20 th second
Ball joint force longitudinal – time domain 15 th to 20 th second Tie rod force axial – time domain 15 th to 20 th second
black…measurementred…...simulation
Results: 10. iteration, rough road
© ECS / Disclosure or duplication without consent is prohibited
FEMFAT LAB vi Results
Author: FEMFAT SUPPORT 15Date: 05.05.2015
Damper force - frequency domain Ball joint force lateral - frequency domain
Ball joint force longitudinal - frequency domain Tie rod force axial - frequency domain
Results: 10. iteration, rough road black…measurementred…...simulation
© ECS / Disclosure or duplication without consent is prohibited
FEMFAT LAB vi Results
Author: FEMFAT SUPPORT 16Date: 05.05.2015
Spring displacement(model check)
Spring displacement – time domain
Spring displacement – time domain 15 th to 20 th second Spring displacement – frequency domain
Results: 10. iteration, rough road
black…measurementred…...simulation
© ECS / Disclosure or duplication without consent is prohibited
FEMFAT LAB vi Results
Author: FEMFAT SUPPORT 17Date: 05.05.2015
Results: 10. Iteration – relative damage values (signal based)
Relative damage comparisonsimulation to measurement
1,021,03
0,99
0,970,97
0,90
0,95
1,00
1,05
1,10
Damper force Bolt force lateral Bolt forcelongitudinal
Tie rod force Spring deflection
rela
tive
dam
age
valu
e
© ECS / Disclosure or duplication without consent is prohibited
General Workflow
Author: FEMFAT SUPPORT 18Date: 05.05.2015
Workflow of fatigue testing and analysis based on road load data
Application & calibration
Data acquisition on proving ground
Load data analysis
Fatigue testing
Fatigue analysis
Virtual iteration MBS model
M U L T I P H Y S I C
W O R K F L O W
© ECS / Disclosure or duplication without consent is prohibited
Fatigue analysis
Author: FEMFAT SUPPORT 19Date: 05.05.2015
Internal forces of MOTION SOLVE
using VI:
Rough road
Specimen material data
FE structure & FE stresses for each load case
Multi axial fatigue assessment
RESULT:
Damage Values
© ECS / Disclosure or duplication without consent is prohibited
Fatigue analysis
Author: FEMFAT SUPPORT 20Date: 05.05.2015
� Stress Tensors
� Material Properties
� Stress Gradient
� Mean Stress Influence
� MultiAXial Load
� Technological Influences
� Size Influence
� Temperature Influence
� PLASTic Deformations
� SPOT Joints s
� Anisotropical Behaviourof Arc WELDs
� etc.
S/N1 modifiedby FEMFAT
Load CyclesS
tres
s A
mpl
itude
S/N materialfrom specimen tests
© ECS / Disclosure or duplication without consent is prohibited
Fatigue analysis
Author: FEMFAT SUPPORT 21Date: 05.05.2015
� Stress Tensors
� Stress Gradient
� Mean Stress Influence
� MultiAXial Load
� Temperature Influence
� PLASTic Deformations
� etc.
Influences in FEMFAT
Finally : Component S/N curve including all influences
FOR EACH NODE
Stress Amplitude
Load cycles
Mean Stress
Stress Amplitude
UTSUCS
Specimen Material Data
Mean Stress
Stress Amplitude
UTSUCS Load cycles
© ECS / Disclosure or duplication without consent is prohibited
Fatigue analysis
Author: FEMFAT SUPPORT 22Date: 05.05.2015
• Transformation of all stress tensors into several planes (2D,3D)
• Filtering of interesting planes
• Generation of the load historiesof the stress components
• Rainflow counting in all selected planes
• Damage analysis(Influence Parameter Method)
• The cutting plane with maximum damage isassumed to be the critical plane for fatigue failure
)(
)(
)(
zzzzyyzxxz
yzzyyyyxxy
xzzxyyxxxx
nnne
nnne
nnne
σσσ
σσσ
σσσσ
++×+
++×+
++×=
γ
σa n
σm n
τ a
τ m
ϕ
© ECS / Disclosure or duplication without consent is prohibited
Result comparison
Author: FEMFAT SUPPORT 23Date: 05.05.2015
Crack observed after 97 h
Test bench
Crack predicted after 28 h
FEMFAT with WFT
Crack predicted after 60 h
FEMFAT with Virtual Iteration
© ECS / Disclosure or duplication without consent is prohibited
Conclusion
• Fully automated co-simulation FEMFAT LAB and MotionSolve• Easier and cheaper measurements can be used to get a better
correlation• Including FE-structures in a MotionSolve model can account also
for dynamic effects in the fatigue analysis• Including a MotionSolve simulation can also account for nonlinear
effects (nonlinear dampers, springs…)• No limitations for the used flex body co-simulation (attachment
parts to full vehicle)• Major increase of the accuracy in fatigue with FEMFAT with a
minor increase of the FEA simulation effort • FEMFAT is available in the partner program• FEMFAT LAB will follow soon !
Author: FEMFAT SUPPORT 24Date: 05.05.2015
© ECS / Disclosure or duplication without consent is prohibited
The future is ours to make.
Author: FEMFAT SUPPORT 25Date: 05.05.2015