Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Optimization in Consideration of Fatigue Results
Shown by the Example of an Aircraft Landing Gear
System
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
MAGNA´s Presence Worldwide *
Canada
56
8
USA
50
17
Mexico
23
S. America
2
W. Europe 74
24
E. Europe 11
2
S. Africa
238 Production
*As at March 2008
Asia Pacific 19
3
60 Engineering, R&D
Employees – 83,000
9
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Magna Powertrain Locations
Great Dunmow / GB
St Valentin / A Head Office
Engineering Centers (9) Roitzsch / D
Production sites (38)
Sales Office
Bièveres (Paris) / F
Joint Venture (Prod.)
Oberwaltersdorf / A Albersdorf / A Ilz/ A Lannach / A Benevento / I
North Sydney / CAN
Moscow / RUS 3x 2x Concord / CAN
Unionville / CAN
Woodbridge / CAN
Sterling Heights / USA
Lansing / USA
Muncie / USA
Howe / USA
Aurora / CAN Seoul / ROK Brampton / CAN Asan / ROK Rexdale / CAN Cheonan / ROK 3x Mississauga / CAN Tokyo / J Syracuse / USA 2x New Delhi / IND
Detroit (Troy) / USA Monterrey / MEX 2x Saltillo / MEX Pune / IND
Shanghai / RC Changzhou / RC Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
History
FEMFAT
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Engineering Center STEYR – Range of Services
System Integration
•• Product
Product Definition
Definition
Drivetrain
& Axle
•• Optimization
Optimization
•• Validation
Validation
Engineering
•• Functional
Functional Development
Development
••• Gearboxes,
Acoustics
Acoustics CVT’s
• Gearboxes,
CVT’s
••• Manual
Product
Cost
Optimization
Product
Cost
Optimization
Commercial
Truck
• Manual and
and automated
automated
•• transaxles
Production
Integration
Production Integration
transaxles
Engineering
•• Transfer
Transfer Cases
Cases
Axle
Drives
AxleDevelopment
Drives
•••• Cab
Cab
Development
••• Chassis
Planetary
Wheel
Planetary
Wheel Hubs
Hubs
Development
• Chassis
Development
Engineering
••• Engine
Beam
Axles
Beam Axles
Build
• Prototype
Prototype
Build
•• Vehicle
Vehicle Testing
Testing
•• Engine
Engine Development
Development
& Testing
•• Simulation
Engine
Engine Components
Components
Development
Development
Services
•• Electronics
Electronics
Injection
Injection Systems
Systems
•••• Structural
Analysis
Structural
Analysis
••• Vehicle
Engine
Integration
Engine
Integration
& Support
• Software
Vehicle Simulation
Simulation
••• Strength
Marine
Engines
Marine
Engines
// Fatigue
• StrengthFEMFAT
Fatigue Test
Test Lab
Lab
•• Measurement
Measurement Engineering
Engineering
Acoustics
Acoustics and
and Vibration
Vibration
•••• CAD/CAM/PDM/PLM
Technology
CAD/CAM/PDM/PLM
Technology
Diagnostics
Diagnostics
Production
In
Low
•• Electrics
Electrics
•• Electronics
Electronics
Volumes
•• ECS
ECS Software
Software Products
Products
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
FEMFAT – Local Stress Concept Crack Initiation
Stress Tensors
Material Properties
Stress Gradient
Mean Stress Influence
MultiAXial Load
Technological Influences
Size Influence
Temperature Influence
PLASTic Deformations
SPOT Joints s
Stress Amplitude
Application of specimen data to components
S/N1 modified
by FEMFAT
S/N material
from specimen tests
Load Cycles
Anisotropical Behaviour
of Arc WELDs
etc.
• Finally : Component S/N curve including all influences
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
FEMFAT - Influence Factors
Mean Stress
Notch Influence
(Stress Gradient)
Thermo Mechanical
Temperature
Isothermal
Temperature
Surface Treatment
Surface Roughness
Plastic Fiber
Orientation
Statistic
Boundary Layer
Tempering
(for Tempering
Steel only)
Cast Micro
Structure
Effective
Plastic Strain
Technological
Size
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Enhanced Optimization by using Fatigue Results
Classic Stress / Strain based
Optimization
FEM
Fatigue Based Optimization
Gray Cast Iron
Steel
optimal topology using only 30%
of the design space‘s volume
FLP Based
Optimization
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Optimization regarding fatigue
Start
FE-Model
Optistruct
Material
Life Solver
Damage,
Safety Factor
Adapted
FE-Model
Hyperstudy
Yes
New Design
Stop condition
fulfilled?
No
Process
controlled by the
optimzation tool
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Multi axial fatigue analysis based on
modal approach
FE-Model
time
domain
frequency
domain
Dynamic Loads
Optistruct
Static Behavior
(Mean Stress)
Optistruct
(linear)
Mode Stresses
(real)
Mode Participation
Factors (complex)
Inverse Fourier
Transformation
FEMFATMAX
Endurance
Safety Factors
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Optimization regarding fatigue Landing Gear
FEA Model
Design Variables (Shape Optimization)
Stress Based Optimization
Fatigue Based Optimization
Conclusion
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
FEA Model
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
FEA Model
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Design Variables
Baseline
Design / Shape Variable
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Enhanced Optimization Loop By
Using HyperStudy (Stress Based)
Hyper Study
FEM
Hyper Mesh / Optistruct
Shape optimization
Minimize maximum stress value in critical area
Optimization Study
(Stress Results)
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Enhanced Optimization By
Using HyperStudy (Stress Based)
Objective definition:
Minimize stress values
(max. stress value from the critical node group)
Adaptive Response Surface method (HyperOpt)
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Enhanced Optimization By
Using HyperStudy (Stress Based)
Result design variables:
Adaptive Response Surface method (HyperOpt)
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Stress Based Optimization (LC Braking)
Node 102153
Safety Factor:0.44
Baseline
Node 102153
v. Mises Stress: 762 MPA
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Stress Based Optimization (LC Braking)
Node 102153
Safety Factor:0.48
Run 17
Node 102153
v. Mises Stress: 653 MPa
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Enhanced Optimization By
Using HyperStudy (Fatigue Based)
FEM
Hyper Mesh / Optistruct
Hyper Study
Optimization Study
(Fatigue Results – Safety Factor)
Shape optimization
Minimize maximum stress value in critical area
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Enhanced Optimization By
Using HyperStudy (Fatigue Based)
Objective definition:
Maximize safety factor
(min. safety factor value from the critical node group)
Adaptive Response Surface method (HyperOpt)
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Enhanced Optimization By
Using HyperStudy (Fatigue Based)
Result design variables:
Adaptive Response Surface method (HyperOpt)
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Fatigue Based Optimization (LC Braking)
v. Mises Stress: 762 MPa
Baseline
Safety Factor: 0.44
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Fatigue Based Optimization (LC Braking)
v. Mises Stress: 640 MPa
Run 16
Safety Factor: 0.5
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Summary
Baseline
Safety factor : 0.44
v. Mises stress : 762 MPa
Mass
: 133.00 kg
Stress based
Safety Factor : 0.48
+ 9%
v. Mises stress : 653 MPa
Mass
: 133.67 kg
Fatigue based
Safety factor : 0.5
+ 13%
v. Mises stress : 640 MPa
Mass
: 133.45 kg
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Outlook
Including dynamic effects:
Basic FEM-model
Optistruct
Basic FEM-model
Optistruct
Life-Solver
Adapted
FEM-model
Motionsolve
Adapted
FEM-model
Controller
Life-Solver
New design
yes
Convergence
criteria reached
Enhanced
Optimization Loop
no
Controller
New design
yes
Convergence
criteria reached
no
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Manual Approach
Attachment part
Damping ≤ 2%
Damping ≥ 6%
Eigenfrequency
Analysis
Impact Analysis
Transient Response
Frequency Response
6
5
4
acceleration [g]
3.0
2.5
2.0
1.5
acceleration [g]
acceleratio [g]
acceleration [g]
4.0
3.5
1.0
0.5
0.0
0.0
5.0
10.0 15.0 20.0 25.0 30.0 35.0 40.0
Representative
Collective
2
1
0
-1
0
0.1
0.2
0.3
0.4
0.5
0.6
-2
-3
frequency
freqency[Hz]
[Hz]
Stressdistribution
3
time
[s]
time [s]
(Quasi) Static Stress distribution
(Load at CG, component depending)
Fatigue
(FEMFAT)
Fatigue
(FEMFAT)
Total life time
Representative
Collective
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Manual Approach: Spare wheel carrier (ξ
ξ=1%)
Beschl. x [g]
Beschl. y [g]
Beschll. z [g]
4,5
Acceleration [g]
4,0
3,5
3,0
2,5
2,0
1,5
1,0
0,5
350
Maximum
Stress
324
300
250
[N/mm2]
5,0
Stress component for each Eigenfrequency:
Von
Mises
Stress
Mises
Vergleichspannung
acceleration at frame: 0,18mm (harmonic, vertical)
response:
3,2g
(bracket outside)
200
150
100
50
30
27
9
0,0
0
5
10
15
20
Frequency [Hz]
25
30
35
21
0
1
2
3
Mode
1st Mode is dominant
Up to 2% modal damping stress combination
of different modes at the Eigenfrequencies is not necessary
4
6
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Summary / Conclusion
Topology optimization leads to
global design
Shape optimization leads to
local design improvement
Integration of FEMFAT leads to improved optimization
results:
- adequate interpretation of static and dynamic loads
- consideration of load histories
- consideration of material properties
- many other influence factors can be considered
Copyright © 2008 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.
Summary / Conclusion
Last-Zeitverlauf
8000
6000
4000
Kraft [N]
+ enables consideration of
complex loads
2000
0
-2000
-4000
-6000
199
188
177
166
155
144
133
122
111
89
100
78
67
56
45
34
23
1
12
-8000
Zeitschritt [0,2s]
Seitenkraft F y
Aufstandskraft F z
σUlt
σa
R
=
+ enables consideration of
material properties
Längskraft F x
-
+ enables proper cosideration of
static and dynamic load portion
σEndu
8
8
R
0
5
4
2, 3
9
NEndu
+ allows consideration of durability, endurance and over loads
+ provides consideration of
many other influences e.g. welds, spot welds
- needs additional CPU-time
=
6, 7
1
Rp 0,2 Rm
σm