Multi-Objective Design Exploration (MODE) Applied to a Regional Jet Design Shigeru Obayashi Institute of Fluid Science Tohoku University
Multi-Objective Design Exploration (MODE) Applied to a Regional Jet Design
Shigeru Obayashi Institute of Fluid Science
Tohoku University
Outline
n Introduction n Multi-Objective Design Exploration (MODE) n MRJ structure modeling
n Objectives n Structure design of MRJ Horizontal Tail
n Visualization of design space
n Conclusion
MRJ: Mitsubishi Regional Jet
MODE Applications to MRJ
MODE as aircraft MDO has been proposed and applied to MRJ wing design (MHI-Tohoku-university collaboration) ■ Previous Applications • Wing-body configuration MDO • Wing-nacelle-pylon-body configuration MDO • Winglet design MDO
MODE to Solve Multi-objective Optimization
Multi-objective Genetic Algorithm
Computational Fluid Dynamics
Design Database
Design Knowledge
Visualization and Data Mining
Multi-Objective Design Exploration (MODE)
Step 1Multi-objectiveShape Optimization
Step 2Knowledge Mining
Data mining:maps, patterns,models, rules
Reference vector for k-th unit Data vector
k-th unit High-dimensional Vector space
Map space
Self-Organizing Map (SOM)
" Visualization of design space " Neural network model proposed by Kohonen
n Unsupervised, competitive learning
" High-dimensional data → 2D map " Qualitative description of data
http://www.brain.kyutech.ac.jp/~furukawa/index.html
SOM provides design visualization: Seeing is understanding (Essential design tool)
MRJ Structural Features
ü Application of composite materials is about 12%
ü Need to reduce structural weight of main components • Aluminum wing box • Composite horizontal tail box (present application) • Composite vertical tail box • etc…
Aluminum 83%
CFRP 9%
GFRP 3%
Titanium 3%
Steel 2%
• Previous MODE applications - Aerodynamics : based on high-fidelity CFD model - Strength : based on low-fidelity structure model
• Need to introduce high-fidelity structure model into MODE
Previous MDO modeling for MRJ
High-Fidelity CFD model
Low-Fidelity Structure model
Aircraft model for MRJ MDO
skin
stringer
skin-stringer equivalent plate
Actual aircraft design
Strength Criteria
Plate compressive and tensile stress (rough evaluation)
Various buckling modes for skin, stringer and spar etc.
Fidelity Gap
Structure of Horizontal Tail Wing Key Points for H-Tail Structural Design
ü Various interactive criteria for buckling
• Euler buckling, Stringer crippling
• Skin bucking, Stiffened panel buckling
• Spar web shear buckling
à Hard to determine optimum combination between rib pitch (A) & stringer pitch (B)
Skin Stringer
Spar Rib A
B
Conventional Approach for Rib-Str. Pitch Optimization
ü Parameter study
à Limited search of design space with interactive buckling criteria
à Global optimum ? Local optimum ?
Rib P1 Str. P1
Rib P1 Str. P2
Rib P2 Str. P1
Rib P2 Str. P2
Weight Global Optimum ?
• Introduce a high-fidelity structure model into MODE based on detailed buckling modes using realistic aircraft structure model
• Apply the new MODE to MRJ Horizontal Tail
Objectives
Definition of Optimization Problem 1. Assumptions
• Fixed H-Tail OML & box config.
• Fixed design loads
• Fixed stacking sequence
• T-type stringer
à Fixed W & H à t = sizing parameter
• Sizing criteria
à Euler & Skin buckling à Stringer crippling à Stiffened panel buckling à Spar web shear buckling
2. Objective
• Minimize structural weight • Minimize number of main structure parts (rib & stringer)
H
W
t
t
T-type Stringer
H-Tail OML (fixed)
Str. Pitch
Rib Pitch
H-tail box (fixed)
3. Design variables
• Rib pitch (at regular intervals)
• Str. pitch (at regular intervals)
4. Criteria
• Str. thickness t > tmin (specified)
• Strain margin > Specified margin
Flowchart of Present Approach
Define design space
Choose sample points
Strength evaluation
Construct surrogate model
Check the model and front
Extract design space
Design of experiment: Latin hypercube
Response surface model: Kriging model
Find non-dominated front of EIs Optimization: ARMOGA
Data mining: SOM
NASTRAN-based structural sizing
NASTRAN-based Structural Sizing
Design parameters
Automated FEM generation
Static analysis by NASTRAN
Structural sizing
OML (fixed)
Design Loads (fixed) Rib & Str. Pitch, Initial thicknesses
Internal stress, strain
New thicknesses, Weight
No
Yes
Converged thicknesses & weight
Converged ?
FEM update
OML: Outer Mold line FEM: Finite Element Model
NASTRAN Static Analysis
Horizontal Tail FEM
Number of stringers and ribs
Results - Optimum Solution
• Kriging model was constructed based on 36 sample points
à Tradeoffs between two objective functions
à Improvements over the baseline design
Number of Structural Components
W eight
Initial points
1st update
2nd update
Baseline
Optimum Direction20kg
10
Constructed Weight RSM Comparison of objective functions among the baseline and sample points
Results - Design Knowledge
• SOMs were generated to extract design knowledge
Regular interval rib pitch (previous design concept)
Additional weight reduction by using larger outboard rib pitch than inboard
(new design concept)
Optimum pitch
à Optimum str. pitch can contribute to reduce both of the inboard & the outboard weights
Smaller pitch
à Smaller rib pitch is effective to reduce the inboard weight
Larger pitch
à Larger rib pitch is effective to reduce the outboard weight
Conclusion
MODE has been applied to the multi-objective structural design optimization for MRJ H-Tail n Including detailed buckling evaluation n Improved both of weight and number of main
structure parts compared with the baseline design
n Provided a better design concept utilized for MRJ H-Tail structural design
Stay tuned for the first flight in 2013!