Towards Aerodynamic Shape Optimization of …arrow.utias.utoronto.ca/ncss13/NCSA Presentations/Reist...Outline 1 Introduction 2 Blended-Wing-Body 3 Aerodynamic Shape Optimization 4

UTIAS

University of TorontoInstitute for Aerospace Studies

Towards Aerodynamic Shape Optimization ofRegional Class Blended-Wing-Body Aircraft

for Reduced Environmental Impact

Thomas A. Reist

David W. Zingg

May 16th, 2013

University of Toronto Institute for Aerospace Studies 2013 National Colloquium on Sustainable Aviation

Outline

1 Introduction

2 Blended-Wing-Body

3 Aerodynamic Shape Optimization

4 Optimization

5 Conclusions & Future Work


Introduction

1 IntroductionMotivation

2 Blended-Wing-BodyDesign BenefitsDesign ChallengesRegional BWB Design


4 OptimizationOptimization DefinitionOptimization Under Inviscid FlowOptimization Under Turbulent Flow



Airline Industry Factors

Increasing demand for air travel and air freight

Increasing and volatile fuel prices

Environmental pressures

Source: US Global Change Research Program


Environmental Goals

Source: ICAO Environmental Report 2010


Blended-Wing-Body







The Blended-Wing-Body (BWB)

The tube-and-wing design has served us well for over 60 years...

... But is a step change in configuration design required?


Design Benefits

AerodynamicHigh wetted aspect ratio gives high lift-to-drag ratioNatural ‘area-ruling’ improves high-speed performance

StructuralNatural spanloading reduces bending loads

PropulsiveBoundary-layer ingesting engines reduce fuel-burn

AcousticBody-mounted engines are acoustically shieldedLow landing speed reduces airframe noise

Liebeck, JoA, Vol. 41, No. 1, 2004


Design Benefits






Design Benefits







Design Benefits





Daggett, NASA/CR-2003-212670, 2003


Design Challenges

AerodynamicShock-free airfoils with sufficient thicknessMaintaining stability and control without a tail

StructuralDesign of non-cylindrical pressure vessel for the cabinMore complicated load-paths

PropulsiveRobust boundary-layer ingesting engine technology

Passenger comfortRide quality

Mukhopadhyay, 2012


Design Challenges





Mukhopadhyay, AIAA 2012-1999, 2012

Mukhopadhyay, 2012


Design Challenges





Daggett, NASA/CR-2003-212670, 2003

Mukhopadhyay, 2012


Design Challenges






Mukhopadhyay, 2012


The Regional Jet Segment

Comprises 30% of global aircraft fleet

Fastest growing segment over the past 30 years

RAA 2011 Annual Report 21

Reg

ion

al Aircraft Statistics

US Regional Aircraft Fleet

Scheduled Passenger Aircraft by Type (see airline breakdown page 23)

average seating capacity of regional aircraft

3741 42

4547

50 50 5153

55 56

0

10

20

30

40

50

60

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 20100

500

1000

1500

2000

2500

3000

3500

4000 Turboprop Aircraft

Regional Jets

Narrowbody*

Widebody

active in service aircraftREGIONAL MAINLINE

Notes:Includes most aircraft over 9 seats*Mainline narrowbody aircraft include Embraer 190 aircraft operated by US Airways and JetBlue Source: OAG Fleet iNet

“I see a day when the regional carrier is the face of a large and growing number of

small communities and not the mainline carrier as it is today...The regional sector will

increasingly become a large and larger face of the domestic market...but it will be done

with an increasing level of flying being done ‘at risk’. I can see a day when the regional

carriers of today are buyers of mainline domestic hubs. Consolidation inside the regional

sector is building companies with significant scope and scale on the operating side...”

Bill Swelbar, swelbar.org

at the RAA Fall Meeting December 2010

Source: Regional Airline Association 2011 Annual Report


Design Problem

Passengers 100†

Cargo volume 683 ft3

Payload 23,380 lbsMax range 2,000 nmiCruise speed 0.80 Mach† Single class at 31” pitch

Similar mission to the CRJ1000ER and E-190


Aerodynamic Shape Optimization (ASO)







Aerodynamic Shape Optimization

Range =aM

cT

L

Dln

(W0 + Wfuel

W0

) Flight condition effectsPropulsive effectsAerodynamic effectsStructural effects

Drag reduction under inviscid flow

Flow model: Euler equationsMinimize induced dragEliminate wave drag

Drag reduction under turbulent flow

Flow model: Reynolds-Averaged Navier-Stokes equationsMinimize induced dragEliminate wave dragMinimize profile drag



Geometry parameterization andmesh movement

B-spline geometryparameterization

Linear elastic mesh movementalgorithm applied to theB-spline grid

Robust for large shapechanges



Flow solver

Newton-Krylov-Schur parallelmultiblock implicit flow solver

Euler and Reynolds-AverageNavier-Stokes equations withthe one-equationSpalart-Allmaras turbulencemodel



Gradient evaluation

Via the discrete adjointmethod

Integrated with geometryparameterization, meshmovement, flow solver

Solution time independent ofnumber of design variables

76 CHAPTER 4. ANALYTIC SENSITIVITY ANALYSIS OF COUPLED SYSTEMS

0 400 800 1200 1600 2000

Number of design variables (Nx)

0

400

800

1200

Norm

aliz

ed tim

e

Complex step

2.1

+ 0

.92

× N

x

Finite difference

1.0 + 0.38 × N x

Coupled adjoint

3.4 + 0.01 × Nx

Figure 4-8: Computational time vs. number of design variables for finite differencing, com-plex step and coupled adjoint. The time is normalized with respect to the time required forone aero-structural solution.

The cost of computing sensitivities using the coupled-adjoint procedure is in theory inde-

pendent of the number of variables. Using our implementation, however, some of the partial

derivatives in the total sensitivity equation (4.17) are calculated using finite differences and

therefore, there is a small dependence on the number of variables. The line representing

the cost of the coupled adjoint in Figure 4-8 has a slope of 0.01 which is between one and

two orders of magnitude less than the slope for the other two lines.

In short, the cost of computing sensitivities with respect to hundreds or even thousands

of variables is acceptable when using the coupled-adjoint approach, while it is impractical to

use finite-differences or the complex-step method for such a large number of design variables,

even with current state-of-the-art parallel computing systems.

4.5.2 Coupled-Adjoint Solution

The constant terms in the equations for the straight lines of Figure 4-8 represent the cost

of each procedure when no sensitivities are required. For the finite-difference case, this is

Martins, 2002


Optimization







Baseline Design

Car

go

Car

go

M M

C.G. location

M

Wing box

Monument

CapacityPassengers 98Crew 4

Cabin floor area 593 ft2

Cargo volume 683 ft3

Geometry

Planform area 2177 ft2

Total span 90 ftLength 74 ftMAC 44 ftAspect ratio 3.7Wetted aspect ratio 1.6

WeightMTOW 96,760 lbOEW 54,710 lbPayload 23,380 lb

Wing load at MTOW 44 lb/ft2

Cruise conditionsDesign range 500 nmiAltitude 40,000 ft

Reynolds number 69 ×106

Mach number 0.80xCG/ccenter-line 0.65


Optimization Problem

Objective:Minimize drag

Design variables:B-spline control points

Angle-of-attack

Geometric constraints:Cabin shape

Span and area

Geometric limits

Optimization under inviscid flow:Stability-constrained

Optimization under turbulent flow:Trim-constrained


Optimization Under Inviscid Flow

X/cC

p0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.20

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.40

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.80

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.00

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.60

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.98

Baselinepressure profiles

Y/b0.00.20.40.60.81.0

Y/b0.0 0.2 0.4 0.6 0.8 1.0

0.05

0.10

0.15

0.20

L/Ltotal

Cp: -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

Center of gravityNeutral point

OptimizedCenter of gravityNeutral point

Baseline

Baseline and optimized surface pressure andspanwise lift distributions

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.20

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.40

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.80

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.00

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.60

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.98

Optimizedpressure profiles


Inviscid Performance

AoA CM Kn L/D

Baseline 1.85 -0.021 -2.8 18.3Optimized 3.01 0.000 +4.9 39.6

Drag reduction of 55% while creating a

trimmed and stable design

Wave drag eliminatedInduced drag reduced

Stability constraint incurs a 2% dragpenalty

Mach number

ML/

D

0.70 0.75 0.80 0.85 0.900

5

10

15

20

25

30

35

40BaselineSP trim-constrainedSP stability-constrainedMP trim-constrained

Mach number

CD

0.70 0.75 0.80 0.85 0.900.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

0.045

0.050BaselineCmCm/KnCm-CGCm/Kn-CGCm MP

Mach number

CD M

2

0.70 0.75 0.80 0.85 0.900.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040BaselineCmCm/KnCm-CGCm/Kn-CGCm MP

Mach number

Ran

ge F

acto

r

0.70 0.75 0.80 0.85 0.900

5

10

15

20

25

30

35

40

45BaselineCmCm/KnCm-CGCm/Kn-CGCm MP

R =aM

cT

L

Dln

(W0 +Wfuel

W0

)


Optimization Under Turbulent Flow

X/cC

p0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.20

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.40

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.80

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.00

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.60

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.98

Baselinepressure profiles

Y/b0.00.20.40.60.81.0

Y/b0.0 0.2 0.4 0.6 0.8 1.0

0.05

0.10

0.15

0.20

L/Ltotal

Cp: -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

Center of gravityNeutral point

OptimizedCenter of gravityNeutral point

Baseline

Baseline and optimized surface pressure andspanwise lift distributions

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.20

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.40

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.80

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.00

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.60

X/c

Cp

0.0 0.2 0.4 0.6 0.8 1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Y/b = 0.98

Optimizedpressure profiles


Turbulent Performance

AoA CM Kn L/D

Baseline 3.96 0.007 -15.6 10.3Optimized 2.98 0.000 -4.0 16.7

Drag reduction of 40% while creating

trimmed design

Wave drag eliminatedInduced drag reducedProfile drag reduced

Baseline

Optimized


Conclusions







Conclusions & Future Work

Conclusions

Aerodynamic shape optimization is a powerful tool for dragreduction

High-fidelity aerodynamic shape optimization applicable tofull configuration optimization

Future Work

Aerodynamic shape optimization of an equivalenttube-and-wing design

Aerostructural optimization is required to demonstratefeasibility of regional jet BWB concept


Questions

Thank You

Questions?

Financial support provided by:


Documents

Towards Aerodynamic Shape Optimization of …arrow.utias.utoronto.ca/ncss13/NCSA Presentations/Reist...Outline 1 Introduction 2 Blended-Wing-Body 3 Aerodynamic Shape Optimization 4