3. API262
Background
3
Market forecast of aircraft
Middle(100-200 seats)* class
Requirement for future aircraft:
Low fuel burn and noise
Recent engine has higher BPR
Effective to reduce fuel burn and noise
Aircraft design in consideration of
engine/airframe interaction is required.
*http://www.jadc.or.jp/JADCF00.pdf
http://www.nasa.gov/topics/aeronauti
cs/features/future_airplanes.html
http://www.mtu.de/en/products_services/co
mmercial_mro/programs/ge90/index.html
Fan diameter gets bigger
4. API262
Background
4
Novel-Wing-Body(NWB)
150seats Blended-Wing-Body type aircraft
Concept has been proposed by Tokyo Metropolitan University and JAXA.
Smaller than the other BWBs aircraft
Upper mount engine integration
Design optimization for improvement aerodynamics
performance
Candidate of next generation aircraft
Engine integration considering Intake/Exhaust effect
Aerodynamic interaction between engine and airframe
Intake stagnation decreases airframe lift.
Complex flowfield around engine
Computational analysis based design
optimization is required.
http://air.mapping.jp/top.html
NWB
5. API262
Objective
5
Design optimization of NWB airframe considering
engine intake/exhaust flow interaction
Higher L/D at cruse condition using Efficient Global
Optimization(EGO)
Acquirement of design knowledge using data visualization
6. API262
Design Method
Airframe definition
Representation by modified PARSEC method
Thickness distribution and camber are defined respectively.
Design performance of leading edge is improved than original PARSEC
method.
Inboard wing design by camber control
6design variables×3cross sections
Total design variables are 18
6
Outboard Wing
Inboard Wing
Design target
PAX 150
Range[nm] 1800
Length[m] 28
Span[m] 42
Weight[kg] 7.17×104
7. API262
Design Method
7
Engine Shape
High bypass engine
The engine is based on the next generation
turbo fan engine*.
P/P0 T/T0
Fan face 1.21 -
Bypass nozzle 2.13 1.26
Core nozzle 1.48 2.93
Intake/Exhaust Effect
Give P/P0 as boundary condition to
boundary #1
Give P/P0 and T/T0 as boundary
condition to boundary #2 and #3
Intake Exhaust
* Data provided by JAXA engine research center
1
2
0 2
1
1
M
T
T
Mu t
x
1
2
0
0 2
1
1
1
M
T
Tp
p
t
t
BPR 13.8
Cruse thrust [kN] 29.2
8. API262
Design Method
8
Initial sampling is 42 samples
Additional sample is
added to improve the
model.
Objective Function
Maximize: L/D @ Cruse
Subject to CLCruise=0.13
9. API262
Design Method
9
Analysis of Variance (ANOVA)
One of multi-validate analysis tool for quantitative
information.
Visualizing the design variables contributions to
the objective function respectively.
Parallel Coordinate Plot (PCP)
One of visualization techniques from high-dimensional data into two
dimensional data.
Normalized design variables and
objective functions are set parallel
in the normalized axis.
)min(-)max(
)min(-)(
ii
ii
i
dvdv
dvdvx
P
10. API262
Design Method
10
Aerodynamics Evaluation
Tohoku university Aerodynamic Simulation (TAS) code
is employed.
Governing equation: compressible Navier-Stokes equation
Turbulent model: Spalart-Allmaras model
Time integration is carried out by LU-SGS implicit method.
Polar curve is drawn to acquire CD at target CL.
Reynolds number 6.63×107
Mach number 0.80
Cruse altitude [km] 10.0
weight [kg] 7.17×104
Angle of attack[deg.] 0, 1, 2
11. API262
Design Space
Design variable Lower Upper
dv1 rc at 0% 0.000 0.005
dv2 xc at 0% 0.150 0.550
dv3 zc at 0% -0.010 0.010
dv4 zxxc at 0% -0.100 0.100
dv5 zte at 0% -0.020 0.010
dv6 ate at 0% -10.00 15.00
dv7 rc at 16% 0.000 0.005
dv8 xc at 16% 0.150 0.550
dv9 zc at 16% -0.010 0.010
11
Design variable Lower Upper
dv10 zxxc at 16% -0.100 0.100
dv11 zte at 16% -0.020 0.010
dv12 ate at 16% -10.00 15.00
dv13 rc at 30% 0.000 0.005
dv14 xc at 30% 0.150 0.550
dv15 zc at 30% -0.010 0.010
dv16 zxxc at 30% -0.100 0.100
dv17 zte at 30% -0.020 0.010
dv18 ate at 30% -10.00 15.00
0%semi-span(Root)
16%semi-span(Cabin)
30%semi-span (Insertion)
13. API262
Results
13
Several better designs could be obtained.
The best L/D is more than twice that of baseline.
The AoA is smaller than the that of baseline.
L/D
14. API262
Results
14
Optimum Baseline
L/D=17.9 L/D=15.4
AoA=0.18[deg.] AoA=0.78[deg.]
CP distribution Isosurface CP distribution Isosurface
Flowfield visualization
Shock wave appears on the outboard wing
The optimum’s is weaker than the that of baseline.
Higher suction peak on the inboard wing of the optimum design
Due to low AoA
Outboard wing
(Upper surface)
Outboard wing
(Upper surface)
Upper Lower Upper Lower
15. API262
Results
15
Upper Lower Upper Lower
Span load distribution
Higher lift on the inboard wing of
optimum design
Due to get camber at the
inboard wing of the optimum
design
Optimum Baseline
L/D=17.9 L/D=15.4
AoA=0.18[deg.] AoA=0.78[deg.]
17. API262
Results
17
Visualization by ANOVA
dv12(T.E. angle @16%semi-span) shows the
largest effect to the aerodynamic performance.
dv14 and dv15(maximum camber position and
height @30%semi-span) also shows large
effect.
Flat camber from the LE to mid-chord
18. API262
Results
18
Visualization by PCP
dv12 is partial to around -2.5[deg.]
Cross-section@16%semi-span has steep
slope in aft.
Additional thrust is generated by positive
pressure.
19. API262
Conclusions
19
EGO for NWB airframe design considering effect of intake
and exhaust flow is carried out.
Optimum camber line of inboard wing for L/D maximization
Because the optimum design achieves lower AoA than
that of baseline, the shock wave on outboard wing is
reduced.
Design knowledge extraction by data mining
The negative camber around the trailing edge reduce
cruse AoA, then shock wave is waken on the outboard.
The additional thrust force is generated by S-shape
camber at the 16% semi-span.
