Synthesis of Novel Aircraft Concepts for Future Air Travel

7/23/2019 Synthesis of Novel Aircraft Concepts for Future Air Travel

1/136

M S T

Synthesis of Novel Aircraft Concepts

for Future Air TravelDevelopment of a Conceptual Design Environment for

Conventional and Unconventional Aircraft Congurations

10-01-2014 R.J.M. Elmendorp B.Eng.

Faculty of Aerospace Engineering Delft University of Technology


2/136


3/136

Synthesis of Novel Aircraft Concepts

for Future Air TravelDevelopment of a Conceptual Design Environment for Conventional

and Unconventional Aircraft Congurations

M S T

For obtaining the degree of Master of Science in Aerospace Engineering atDel University of Technology

R.J.M. Elmendorp B.Eng.

10-01-2014

Faculty of Aerospace Engineering Del University of Technology


4/136

Copyright R.J.M. Elmendorp B.Eng.All rights reserved.


5/136

D U O TD O

F P P

e undersigned hereby certify that they have read and recommend to the Faculty of AerospaceEngineering for acceptance a thesis entitledSynthesis of Novel Aircra Concepts for Future AirTravelbyR.J.M. Elmendorp B.Eng. in partial fulllment of the requirements for the degree ofMaster of Science.

Dated: 10-01-2014

Head of department:prof.dr.ir. L.L.M. Veldhuis

Supervisor:dr.ir. R. Vos

Reader:dr. ir. R. de Breuker


6/136


7/136

v

Abstract

In the last 60 years many new technologies have entered the aerospace industry, but the overallaircra design remained virtually unchanged. If we compare an aircra built in the 1960s withthe latest generation, they look strikingly similar. Only small evolutionary changes entered thecommercial aircra market. A lot of these changes are driven by the ever-lasting quest to reducethe amount of burned fuel. However, the reduction in fuel usage which can be gained with thesesmall evolutionary changes decreases every aircra generation.

A revolutionary change in the aircra design is needed to make step-change in aircra fuel effi-ciency. Changing the conguration of the aircra could create opportunities for aerodynamic andstructural improvements, which result in higher fuel efficiency.

Current tools used in the eld of aircra design use a lot of empirical data obtained from the anal-ysis of existing aircra. ese tools are not capable of correctly analysing unconventional aircracongurations.

A design tool, called the Initiator, is created which is able to synthesise a conceptual aircra designbased on a given set of top level requirements. e Initiator is able to design and analyse thefollowing aircra congurations:

Conventional aircra

Canard (fore-plane) aircra

ree-surface aircra

Prandtl (box-wing) aircra

Blended-Wing-Body aircra

From top level requirements a rst estimation of the aircra characteristics is performed. eaircra geometry is sized based on the estimated aircra weight, wing loading and thrust-to-weightratio. is geometry is used in a chain of analysis modules which estimate the aircra weight andaerodynamic performance. e design process incorporates two convergence loops: e rst looprepeats the aircra sizing and analysis until the maximum take-off weight converges. e secondloop re-estimates the fuel weight until the harmonic range matches the requirements.

Physics based analysis methods are preferred over empirical methods since they are easier to adaptto unconventionalaircra. e weightestimation methods for themain wing and fuselage are Class

v


8/136

vi Abstract

II.V methods, which means that the main structure is sized using physical calculations. Secondarystructures are still estimated using empirical relations, which is sufficient unless the analysed partsdiffer greatly from the data used to create the empirical methods. All other parts (everythingexcept

the main wing and fuselage) are estimated used Class II methods, which are still highly empirical.e aerodynamics are calculated using a vortex lattice method. since this method is not capable ofcalculating the prole drag, an empirical method is used for the zero-li drag estimation.

e Initiator is veried by comparing the output of the Initiator with existing aircra. A selectionof thirteen reference aircra varying from small regional jets to wide-body long-range jet-poweredaircra is made to verify the tool. Top level requirements are dened which match the payload,harmonic range and runway performance specications of the reference aircra. e aircra gen-erated from these top level requirements are compared to the existing aircra.

e design process is proven to work, since it converges to a feasible aircra design which com-plies with the top level requirements. Since process contains design loops, inaccuracies in analyses

propagate easily through the whole design. By comparing the maximum take-off weights and op-erational empty weights, it can be shown that the generated aircra are similar to the referenceaircra. Nine out of the thirteen generated aircra are estimated to within 10% of the referenceaircra weights.

Visual inspection and comparison of external aircra dimensions show that the implemented de-sign rules are capable of generating an aircra which is similar to the reference aircra.

e design tool was used to compare the different aircra congurations. It is shown that a designprocess which iterates on the aircra maximum take-off weight and adjusts the fuel mass to matchrange requirements works for conventional, canard, three-surface and Prandtl aircra. Testing thedesign process for the Blended-Wing-Body was unfortunately not possible with the current state

of the sizing methods.

It is shown that the canard aircra provides a 12% reduction in fuel mass and a 28% reductionin operational empty mass in comparison with a conventional aircra designed for the same pay-load and harmonic range. However, since none of the analysis methods have been validated thecondence in the results gained from the conguration comparison in low.

Overall canbe concluded that it is possible to model a wide range of different aircra congurationsusing the implemented sizing rules and analysis methods.


9/136

vii

Acknowledgements

is thesis is the conclusion of the great time I had studying as an Aerospace Engineering studentat the Del University of Technology. Since no project can be completed without the support ofother people I would like to use the next few lines to thank the people who supported my duringthe last part of my Master.

First of all I would like to thank my supervisor Roelof Vos who guided me through the wholeproject and was always available for feedback on my work or just a good talk about aircra design.Second I would like to thank the members of mycommittee: Leo Veldhuis and Roeland de Breukerwho took the time to asses my work and share their expert insights.

I would also like to thank my friends for their support and my fellow students who made workingat the faculty a pleasure.

Finally, A special thanks to my parents and sister who have always supported me in everything Idid and encouraged my to pursuit my dreams.

vii


10/136

viii Acknowledgements


11/136

ix

Contents

Abstract v

Acknowledgements vii

List of Figures xv

List of Tables xviii

Nomenclature xix

I esis 1

1 Introduction 3

1.1 Research Question and esis Goal . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.2 Report Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 Background 7

2.1 Aircra Congurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.1.1 Conventional aircra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.1.2 Canard aircra. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2.1.3 ree-surface aircra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.1.4 Prandtl aircra. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.1.5 Blended-wing-body aircra . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.2 Aircra Design Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.2.1 Conventional Conceptual Design Process . . . . . . . . . . . . . . . . . . 10

2.2.2 Design and Engineering Engine . . . . . . . . . . . . . . . . . . . . . . . 11

ix


12/136

x Contents

3 Design Tool Description 13

3.1 Soware architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

3.2 Design process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

3.3 e Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

3.3.1 Sizing Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

3.3.2 Analysis Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

3.3.3 Design Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

3.3.4 Work-ow Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

4 Design Tool Verication 27

4.1 Reference aircra. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

4.2 Comparison based on Design point . . . . . . . . . . . . . . . . . . . . . . . . . . 294.3 Comparison based Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

4.4 Drag polar comparison. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

4.5 Comparison based on Weight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

4.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

5 Conguration Comparison 43

5.1 Top level requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

5.2 Design synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

5.3 Key performance indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

5.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

6 Conclusions and Recommendations 53

6.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

6.2 Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

II Code documentation 57

7 Introduction 59

7.1 Background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

8 Program Structure 61

8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

8.2 InitiatorController . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

8.2.1 Dependency handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

8.3 Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

8.4 Aircra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66


13/136

Contents xi

9 Geometry Denition 67

9.1 Geometry class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

9.1.1 Airfoil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 729.1.2 Lo. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

9.2 Part class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

9.2.1 Wing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

9.2.2 Fuselage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

9.2.3 BoxWing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

9.2.4 Engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

9.2.5 LandingGear. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

9.2.6 Cargo. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 809.2.7 ULD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

9.2.8 Spar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

10 User manual 87

10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

10.2 Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

10.3 Program Run . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

10.4 XML Layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

10.4.1 Aircra denition le . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

10.4.2 Settings le. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

10.4.3 Modules le . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

10.4.4 Materials le . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

10.4.5 Cargo le. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

10.5 Executable Build Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

References 93

A Aircra report generated by the Initiator 97

A.1 General Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

A.2 Specication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

A.3 Operational Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

A.4 Weight estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

A.5 Aerodynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

A.6 Propulsion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

A.7 Aircra Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104


14/136

xii Contents

B Code examples 107

B.1 Part implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

B.1.1 File creation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

B.1.2 Class denition le . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

B.1.3 Generate method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

B.2 Module implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

B.2.1 File creation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

B.2.2 Class denition le . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

B.2.3 Run method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

B.2.4 Adding module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

C Sample aircra denition le 111


15/136

xiii

List of Figures

1.1 History of the aircra fuel consumption (source: [1]) . . . . . . . . . . . . . . . . 3

1.2 Span efficiency of different non-planar wing congurations (based on [2]) . . . . . 4

1.3 Aircra conguration matrix (source: [3]) A: Flying wings B: Planar monoplane,single body C: Non(co)planar wings, single body D: Planar monoplane, multi-bodies E: Hybrid congurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2.1 (Modern) jet transport aircra . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2 ree-view of the Beechcra Starship [4] . . . . . . . . . . . . . . . . . . . . . . . 8

2.3 Top: trimmed conventional aircra; Bottom: Trimmed canard aircra (arrows nottrue to scale) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2.4 ree-view of the Piaggio P180 Avanti [5] . . . . . . . . . . . . . . . . . . . . . . 9

2.5 Impression of the Lockheed Greener Aircra [6] . . . . . . . . . . . . . . . . . . . 9

2.6 Boeing X-48B [6] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.7 Flowchartanaircra designprocess using Designand Engineering Engines (source:[7]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.8 Flowchart a Design and Engineering Engine (source: [8]) . . . . . . . . . . . . . . 12

3.1 Top-level UML of the Initiator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143.2 An abstract overview of the design process . . . . . . . . . . . . . . . . . . . . . . 15

3.3 N2 chart of the modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3.4 Drag polar comparison of AVL with ight test data from the Airbus A320-100 [9] . 19

3.5 Wake visualisation plot from Tornado. . . . . . . . . . . . . . . . . . . . . . . . . 20

3.6 Validation cases of Tornado with varying aspect ratios, source: [10] . . . . . . . . 21

3.7 Relation between wing box weight and total wing weight (source: [11]) . . . . . . 22

3.8 Verication of the fuselage weight estimation method (source: [12]) . . . . . . . . 23

3.9 CD0and e estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

xiii


16/136

xiv List of Figures

3.10 Activity diagram of the DesignConvergence module . . . . . . . . . . . . . . . . . 26

4.1 Payload - harmonic range combinations of the reference aircra . . . . . . . . . . 29

4.2 Design point of the Initiator generated aircra according to A320-200 specications 304.3 Comparison of the design point of the Initiator generated aircra and data from

Roux [13] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

4.4 A320-200 geometry comparison. . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

4.5 A320-200 comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

4.6 Comparison of top views. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

4.7 Drag polar output from the Initiator (A320-200) . . . . . . . . . . . . . . . . . . . 36

4.8 Cruise drag polar comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

4.9 Weight denitions (modied from: [14]) . . . . . . . . . . . . . . . . . . . . . . . 38

4.10 Comparison of the OEM/MTOM fraction of the generated aircra with referencedata from [13] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

4.11 Comparison of the maximum take-off mass (MTOM) and operational emptymass(OEM)calculatedfrom the Initiator with reference aircra weightdata from lodieRoux [13] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

4.12 Comparison of the weight breakdown. . . . . . . . . . . . . . . . . . . . . . . . . 41

4.13 Difference in weight fraction of the maximum take-off mass of the Initiator gen-erated aircra compared with reference data from [14] and [15]. . . . . . . . . . . 41

4.14 Comparison of the weight breakdown (continued) . . . . . . . . . . . . . . . . . . 41

5.1 Denition of the design space and design points . . . . . . . . . . . . . . . . . . . 455.2 3D-view of the design space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

5.3 Payload - range combinations of the design runs, aspect ratio written next to thepoint; Missing points did not converge . . . . . . . . . . . . . . . . . . . . . . . . 46

5.4 3D renders of the generated aircra . . . . . . . . . . . . . . . . . . . . . . . . . . 47

5.5 Relation between the aircra purchase price and the Operation Empty Weight(source: [16]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

5.6 Payload mass as a function of the range as used in Figure 5.8 . . . . . . . . . . . . 49

5.7 Contour plot of the KPIs for the conventional aircra with an aspect ratio of 9 . . . 51

5.8 Comparison of KPIs for the different aircra congurations, coloured bands show

inuence of the aspect ratio on the parameters . . . . . . . . . . . . . . . . . . . . 52

8.1 Top-level Initiator activity diagram . . . . . . . . . . . . . . . . . . . . . . . . . . 62

8.2 Top-level UML class of the Initiator . . . . . . . . . . . . . . . . . . . . . . . . . . 62

8.3 Class instantiation method of the Controller . . . . . . . . . . . . . . . . . . . . . 63

8.4 InitiatorController: runModulemethod . . . . . . . . . . . . . . . . . . . . . . . 64

9.1 UML of all Part and Geometry classes. . . . . . . . . . . . . . . . . . . . . . . . . 68

9.2 Activity diagram of thegetGeometrymethod . . . . . . . . . . . . . . . . . . . . . 70

9.3 Activity diagram of thegeneratemethod . . . . . . . . . . . . . . . . . . . . . . . 70


17/136

List of Figures xv

9.4 Class instantiation method of Geometry . . . . . . . . . . . . . . . . . . . . . . . 70

9.5 Activity diagram of thepropertyChangedmethod . . . . . . . . . . . . . . . . . . 71

9.6 Activity diagram of the propertyAccessedmethod. . . . . . . . . . . . . . . . . . . 71

9.7 Example of a property change in the Wing object. . . . . . . . . . . . . . . . . . . 72


9.9 Activity diagram of thegetDatFilemethod . . . . . . . . . . . . . . . . . . . . . . 73


9.11 Activity diagram of theresampleSectionmethod . . . . . . . . . . . . . . . . . . . 75

9.12 Wing geometry including spars . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76


9.14 Illustration of the algorithm to generate spars . . . . . . . . . . . . . . . . . . . . 78

9.15 Activity diagram of thegetSparLocationsmethod. . . . . . . . . . . . . . . . . . . 799.16 Geometry of an oval and a conventional fuselage . . . . . . . . . . . . . . . . . . . 81

9.17 BoxWinggeometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

9.18 Enginegeometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

9.19 Cargo part with ULD parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

A.1 Aircra geometry (all dimensions in meters) . . . . . . . . . . . . . . . . . . . . . 97

A.2 Loading Diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

A.3 Payload-Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

A.4 Manoeuvre diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99A.5 Mass distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

A.6 Loading diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

A.7 CG location. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

A.8 Drag Polars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

A.9 Aerodynamic efficiency of the aircra . . . . . . . . . . . . . . . . . . . . . . . . . 104

A.10 Fuel tank layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

A.11 Fuselage geometry; (blue = cargo ULDs, purple = oors) . . . . . . . . . . . . . . 106

B.1 Squaregeometry;Length= 4,Width= 6,Position= (2,2,2), Orientation=(45,60,15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109


18/136

xvi List of Figures


19/136

xvii

List of Tables

4.1 Reference aircra requirements, source: [13] . . . . . . . . . . . . . . . . . . . . . 28

4.2 Initiator input parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

4.3 Initiator performance settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

4.4 Initiator geometry settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

4.5 Comparison of geometry parameters . . . . . . . . . . . . . . . . . . . . . . . . . 35

5.1 Requirements used for the conguration comparison runs . . . . . . . . . . . . . 44

5.2 Perceptual change of each conguration with respect to the conventional aircra . 51

8.1 InitiatorControllerclass properties and methods . . . . . . . . . . . . . . . . . . . 65

8.2 Aircraclass properties and methods . . . . . . . . . . . . . . . . . . . . . . . . . 66

9.1 Geometryclass properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

9.2 Geometryclass methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

9.3 Airfoilclass properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

9.4 Airfoilclass methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

9.5 Wingclass properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 779.6 Wingclass methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

9.7 Fuselageclass properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

9.8 BoxWingclass properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

9.9 Engineclass properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

9.10 LandingGearclass properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

9.11 Cargoclass properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

9.12 ULDclass properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

9.13 Sparclass properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

xvii


20/136

xviii List of Tables

10.1 Initiator command-line arguments . . . . . . . . . . . . . . . . . . . . . . . . . . 88

10.2 XML les used by the Initiator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

A.1 Max payload . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98A.2 Performance results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98A.3 Mass summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

A.4 Component masses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

A.5 Aerodynamic properties at cruise . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

A.6 Propulsion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

A.7 Main Wing dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

A.8 Horizontal Stabiliser dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

A.9 Vertical Stabiliser dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

A.10 Fuselage dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105


21/136

xix

Nomenclature

Latin Symbols

A Wing aspect ratio [-]

b Wing span [m]

CD Drag coefficient [-]

CD0 Zero-li drag coefficient [-]CL Li coefficient [-]

CL Li curve slope [-]

CLmax Maximum li coefficient [-]

Cm Pitching moment coefficient [-]

cr Root chord length [m]

ct Tip chord length [m]

e Span efficiency factor [-]

hcr Cruise altitude [m]cT Specic fuel consumption []1/s]

L/D Li-to-Drag ratio [-]

M Mach number [-]

Npax Number of passengers [-]

Rh Harmonic Range [km]

S Wing planform area [-]

Tstatic Static thrust [N]

Wp Payload mass [kg]

xix


22/136

xx Nomenclature

X Range parameter [km]

Greek Symbols

Wing taper ratio [-]

0.25 Quarter-chord sweep angle [-]

Abbreviations

AVL Athena Vortex Lattice

BPR By-pass ratioBWB Blended-Wing-Body

DEE Design and Engineering Engine

FM Aircra Fuel Mass

KPI Key Performance Indicator

MDO Multi-disciplinary Design Optimisation

MTOM Aircra Maximum Take-off Mass

OEM Aircra Operational Empty Mass

PLM Aircra Payload Mass

PRE Payload-Range EfficiencyTLRs Top-Level Requirements

TSA ree-Surface Aircra

ULD Unit Load Device

UML Unied Modeling Language

XML Extensible Markup Language


23/136

1

Part I

esis

1


24/136


25/136

3

Chapter 1

Introduction

In the last 60 years many new technologies have entered the aerospace industry, but the overallaircra design remained virtually unchanged. If we compare an aircra built in the 1960s withthe latest generation, they look strikingly similar. One can say that the changes are more of anevolutionary nature, real revolutionary changes have not entered the commercial aircra marketsince the introduction of the jet engine in the 50s1.

A lot of these changes are driven by the ever-lasting quest to reduce the amount of burned fuel.e design changes result in of course environmental and, with the ever-rising oil prices, econom-

ical benets. Figure1.1presents the total fuel consumption and the fuel consumption per seat ofaircra introduced in the last half century.

Figure 1.1:History of the aircraft fuel consumption (source: [1])

e introduction of the De Havilland DH.106 Comet brought the jet engine to the civil transportaircra market. From the Comets introduction onwards, the fuel efficiency of the engines and thetotal aircra started reducing every year and is levelling out at around 30% of the fuel burned perseat in comparison to the Comet. As can be seen in Figure1.1the curve seems to almost reached

1A notable exception is the short popularity of the supersonic transport aircra (Concorde & Tupolev Tu-144)

3


26/136

4 Introduction

its asymptote; to introduce another step-change in aircra fuel efficiency the evolutionary changepursued in the last half-century will probably not be sufficient. A revolutionary change in theaircra design is needed.

One big change would be the overall aircra conguration. As can be seen in Figure1.2a lot ofaerodynamic advantages can be gained from unconventional wing shapes. Also the general layoutof the aircra could create opportunities for aerodynamic and structural improvements. Figure1.3shows a wide variety of different congurations for subsonic transport aircra.

Bibplanee= 1.36

X-winge= 1.33

Branched tipse= 1.32

End plates

e= 1.38

Boxwinge= 1.46

Joined winge= 1.05

C-winge= 1.45

Tip-plated wingletse= 1.20

Winglets

e= 1.41

Dihedral (large)e= 1.03

e= induced drag efficiency factor

Figure 1.2:Span efficiency of different non-planar wing congurations (based on [2])

Figure 1.3:Aircraft conguration matrix (source: [3])

A: Flying wings

B: Planar monoplane, single body

C: Non(co)planar wings, single body

D: Planar monoplane, multi-bodies

E: Hybrid congurations


27/136

1.1 Research Question and esis Goal 5

1.1 Research Question and esis Goal

e possible improvements which could be gained by using different aircra congurations poses

the following research question:

Which aircra conguration has the potential to introduce a signicant increase in

fuel efficiency?

Current tools used in the eld of aircra design leverage the knowledge gained from aircra devel-oped in the last half-century. e aircra metrics and performance characteristics are captured indatabases and empirical design methods are derived from this information. Since these methodsare based on knowledge of existing aircra, they can only be used to design aircra with designfeatures that are similar to the aircra present in the database.

e aforementioned method is inadequate in the design of unconventional aircra. Since there isno real performance information available of unconventional congurations, the design of such anaircra requires more effort than designing a more conservative and traditional design. In orderto be able to get insight into the performance of unconventional aircra concepts a tool needs tobe conceived which enables the synthesis of such aircra in an efficient and fast manner very earlyin the design process.

Since the tool should be able to design and analyse unconventional aircra, favour should be givento rst principle physics-based methods. is will make the tool sensitive to design changes whichare not captured by empirical methods. An aircra design process should be implemented whichenables the synthesis of conventional and unconventional aircra based on a set of top-level re-

quirements. In order to be able to make a fair comparison of the different aircra all designs shouldbe analysed with the same methods, regardless of aircra conguration. In other words: all meth-ods should be able to analyse a wide range of aircra congurations.

At Del University of Technology previous research has been done on conceptual design methodsfor conventional and box-wing aircra [17], three-surface aircra [18] and Blended-Wing-Bodyaircra [19] and [20]. e effort of these projects should be combined in a single design toolto be able to compare the different aircra concepts. Besides the conceptual design tools whichcreate a rst aircra design based on top-level requirements more sophisticated design programsare developed as the Design and Engineering Engine. e design tool should be able to full therole of the Initiator, the program which creates the rst estimate of an aircra design and provides

the input for higher-delity analysis methods.Because of work done by the previously mentioned conceptual design tools the initial implemen-tation is limited to the design of:

Conventional aircra

Canard (fore-plane) aircra

ree-surface aircra

Prandtl (box-wing) aircra

Blended-Wing-Body aircra


28/136

6 Introduction

e design tool should be developed with modularity in mind. Adding more aircra design con-gurations in the future should be supported and extending the analysis capabilities should bepossible without the need of re-writing the whole program.

All the above results in the following thesis goal:

e development of a exible automated conceptual design environment for the syn-

thesis and analysis of conventional and unconventional aircra designs.

1.2 Report Structure

is report consists of two parts. PartIdescribes the work done in the context of the thesis, PartIIcontains the implementation details of the design tool.

First, background information about the different aircra congurations, the aircra design pro-cess and an introduction to the Design and Engineering Engine developed at Del University ofTechnology is given in Chapter2.

earchitectureofthedesigntoolisanimportantchoicesinceiteventuallyinuencestheexibilityof the design tool. is is presented in Chapter3. Also the implemented design process and theused design and analysis methods are elaborated in this chapter.

To be able to use the design tool it needs to veried. is is done by comparing the design tooloutput to data from existing aircra. e method and results of the tool verication can be foundin Chapter4.

Chapter 5 will present an application of the design tool by using it to compare a set of conventionaland unconventional aircra designed for a wide set of payload and range requirements.

e thesis is concluded in Chapter6and recommendations for future research are presented.

Chapter7gives an introduction to the second part of this report: the code documentation.

Chapter8draws an outline of the structure of the program. Here the composition of the differentcomponents and their implementation is shown.

Chapter9presents the aircra parts also known as theHigh Level Primitives. e different pa-rameters which are required to dene the parts and the methods of generating the geometry are

elaborated.

In Chapter 10 the installation and ways of operating the program are explained. Also the structureof the les which are needed to operate the application is presented.


29/136

7

Chapter 2

Background

2.1 Aircra Congurations

2.1.1 Conventional aircra

(a) De Havilland DH.106 [21] (b) Boeing 367-80 [22] (c) Airbus A350XWB [23]

(d) Sud Aviation SE 210 Car-

avelle [24]

(e) Bombardier CRJ900 [25]

Figure 2.1:(Modern) jet transport aircraft

e majority of all currenly operated aircra can be called Conventional aircra. As mentionedpreviously, the De Havilland DH.106 Comet (Figure 2.1a) brought the jet engine to the commercialtransport aircra market. An unique design feature were the in the wing root integrated engines.

e Boeing Dash 80 (later developed into the Boeing 707 in 1958) placed the engines in nacellesunderneath the wing, as can be seen in Figure2.1b. is solved a lot of structural problems, with

7


30/136

8 Background

the added advantage that bigger engines could be installed without a complete redesign of theaircra. is tube-and-wing conguration with podded engines has remained unchanged duringthe last six decades, with the latest example the Airbus A350 (Figure2.1c).

A slight variation on this conguration comes in the form of fuselage mounted engines. is con-guration is particularly popular with regional jet aircra. An 1959 example of this congurationis the Sud Aviation Caravelle (gure2.1d) and a modern example is the Bombardier CRJ900 inFigure2.1e.

2.1.2 Canard aircra

Figure 2.2:Three-view of the Beechcraft Starship [4]

e canard aircra is a variation on the conventional aircra. Instead maintaining longitudinal sta-bility and control with the horizontal stabiliser attached to the tail, the canard aircra sports a hor-izontal stabiliser in front of the main wing. A ying example of a canard aircra is the BeechcraStarship (Figure2.2).

Figure 2.3:Top: trimmed conventional aircraft; Bottom: Trimmed canard aircraft (arrows not

true to scale)


31/136

2.1 Aircra Congurations 9

Assuming the centre of gravity positioned between the canard and main wing, the aircra canbe trimmed with both surfaces providing li. is decreases the induced drag of the aircra incomparison to the conventional aircra, since the later always has a down force on horizontal

stabiliser, which the main wing needs to compensate for with extra li. is is illustrated in Figure2.3.

2.1.3 ree-surface aircra

e three-surface aircra is based on the same principle as the canard aircra with the addedexibility of an additional control surface. e basic conguration is a fuselage, main wing, canardand a horizontal tail. e three surfaces allows the designers to have more design freedom whilemaintaining the aircra stability and the lack of negative li on the tail surface. An example of athree surface aircra is the Piaggio P180 Avanti, which can be seen in Figure2.4.

Figure 2.4:Three-view of the Piaggio P180 Avanti [5]

2.1.4 Prandtl aircra

Figure 2.5:Impression of the Lockheed Greener Aircraft [6]

e Prandtl-plane design is based on research by Ludwig Prandtl in 1924 [ 26]. e Box-wingdesign is a derivative of Prandtls Best wing system and a great induced drag reduction is expectedfrom this design.


32/136

10 Background

Currently thereareno commercial available aircra which featurethe Prandtl-plane conguration.e Lockheed submission on the NASA Greener Aircra project in Figure2.5is an example of aconceptual Prandtl-plane design.

2.1.5 Blended-wing-body aircra

Figure 2.6:Boeing X-48B [6]

e Blended-wing-body aircra concept is an aircra where the body and wing are integrated intoone blended shape. e fuselage has an airfoil shaped cross-section and is designed to contributesignicantly to the li. Since the aircra weight is better distributed span-wise over the aircra,less bending moments are introduced into the aircra structures. is should result into a morefuel efficient aircra.

Currently there are no commerciallyying bended-wing-bodyaircra. Boeing and NASAare con-ducting experiments with the remote-controlled X-48 aircra (Figure2.6) which should provideinsight into Blended-wing-body performance in the near future.

2.2 Aircra Design Process

e aircra design process is a complex task involving a lot of different disciplines. Because thereare is a lot of mutual inuence between the different disciplines the design synthesis is an iterativeprocess.

e section discusses two different approaches to conceptual aircra design. First the traditionalconventional aircra design process is discussed. Secondly the Design and Engineering Engine asunder development at Del University of Technology is presented.

2.2.1 Conventional Conceptual Design Process

e design methodologies currently in widespread use in the aerospace industry are based onmethods developed in the 1960s and 1970s. e conceptual design phase of these methods arecharacterised by the extensive usage of empirical methods to create a rst estimate of the aircra.ese empirical relations are based on previously designed aircra, which are sometimes tuned to


33/136

2.2 Aircra Design Process 11

account for increases in technological capabilities. Only later in the detailed design computationalmethods and wind-tunnel tests are used to analyse the aircra.

is approach enables designers to create aircra concepts and estimate their performance withgood accuracy, but restricts the design to aircra which are similar to the aircra on which theempirical methods are based.

2.2.2 Design and Engineering Engine

e development of unconventional aircra concepts needs a different approach to the designprocess. Instead of relying on empirical relations to design and analyse the aircra components asemi-empirical or fully physics based methods are preferred. e interactions between disciplinesare modelled directly, which moves the aircra design process in the realm of multi-disciplinarydesign optimisation (MDO).

e design starts by creating a rst estimation of the aircra based on top-level requirements. isdesign is analysed and the result is used to create a higher-delity design of the aircra in ques-tion. Every step of increasing delity needs a starting point, this task is preformed by the so-calledInitiator. In fact, every design loop can be regarded as the Initiator of the next design loop; as isillustrated in Figure 2.7. e rst Initiator which generates a rst preliminary design of the aircrais the program developed in the context of this Masters thesis project. Every step in the designprocess is performed by a Design and Engineering Engine (DEE).

Figure 2.7:Flowchart an aircraft design process using Design and Engineering Engines

(source: [7])

In Figure 2.8 the layout of a Design and Engineering Engine is shown in a more detailed ow chart.Here can be seen that for the different disciplines the same aircra geometry generated by the multimodel generator is used to create their input. e output is combined to calculate the performance


34/136

12 Background

of the analysed aircra. is information is used to determine the feasibility of the design and canbe used as an objective function of the optimiser.

Figure 2.8:Flowchart a Design and Engineering Engine (source: [8])


35/136

13

Chapter 3

Design Tool Description

isChapter describes the design tool:the Initiatordeveloped in this thesis project. First the globalarchitecture of the tool is presented. In Section3.2the aircra design process implemented in thetool is elaborated. e different modules of the Initiator are presented in Section3.3.

3.1 Soware architecture

e Initiator is developed in a modular manner. To enable this modularity object-oriented pro-

gramming is used. All the different program features are dened as classes. Once a class is instan-tiated it will be represented as an object in the program. Basic classes are created which can beextended to represent more sophisticated object.

e Initiator consists of module and geometry objects which are controlled by a controller object.All program ow is directed by the controller object. e top-level layout of the Initiator can befound in an UML in Figure3.1. Modules can specify modules they depend on for which the con-troller will make sure they are completed before the module is started. ere are four differenttypes of modules dened:

Sizing modules

Analysis modules

Design modules

Workow modules

Sizing modules perform the preliminary sizing of the aircra based on the top-level requirementsand conguration settings that are given as an input. e resulting design can be used by analysisand design modules to respectively analyse the aircra (aerodynamics, weights, etc.) or design aspecic part (cabin, control surfaces, etc.). e workow modules facilitate the Initiator by pro-

viding XML read/write methods and implement design workows (to converge to a consistentdesign).

13


36/136

14 Design Tool Description

e aircra is built up of separate parts called high-level primitives. Each primitive is able to gen-erate its own geometry. Besides this, the primitives implement methods which are able to calculatefor example surface areas, mean aerodynamic chords, etc, whatever is applicable to the primitive

in question. e main primitives are:

Wing: Used for wings, tail surfaces, etc.

Fuselage: Used for conventional as well as oval fuselages

Engine: Used to model engine pods

Besides the main primitives, some simple primitives are dened to model cargo containers, landinggear, spars and combine wings into a box-wing. For more information about the implementationof the primitives, please read Chapter 9. Geometry is generated on-the-y when it is needed and isautomatically agged for re-generation when parameters are changed. is means that at all times

an up-to-date model of the aircra is used, without the need to re-generate all geometry aer everychange.

+runModule()+resetModule()

+getModuleResults()

XMLFileSettings

...

InitiatorController

+findPart()

+addPart()+checkAllParts()

Name

Description

Type

Requirements...

Aircraft

Name

Location

Part

+run

...

Module

Figure 3.1:Top-level UML of the Initiator

3.2 Design processis section describes the design process. e process ow leverages two different mechanisms inthe Initiator. First, the sequence of modules is controlled by module dependencies. is providesthe basic forward-feeding module ow. e feedback loops are generated by the DesignConver-gencemodule, which resets a set of modules, changes input parameters and calls the controller tore-run the modules.

A high-level view of the design process can be found in Figure3.2. e whole process starts withthe denition of the top-level requirements (mission requirements), conguration parameters anda set of performance parameters (initial guess of the drag polars, specic fuel consumption andCLmaxvalues). is is represented in the rst row in Figure3.2.


37/136

3.3 e Modules 15

From the specied top-level requirements a Class I weight estimation is performed. e wingloading and thrust-to-weight ratio of the aircra is determined by evaluating performance andregulatory constraints. A Class II weight estimation (Raymer [27]) is performed to get a more

design-sensitive weight and centre-of-gravity estimation.e Class II weight results are used to perform an aerodynamic analysis by using a vortex-latticemethod (AVL [28]). e aerodynamic forces and moments on the aircra surfaces are used inthe Class II.V methods for a more rened and physics based estimation of the wing and fuselageweight.

Since the vortex-lattice method is an inviscid analysis method, the prole drag is calculated sepa-rately. is prole drag is calculated empirically using Torenbeeks method [ 29]. All aerodynamicand weight information is gathered and the overall aircra performance is calculated.

Aer every design loop the calculated (Class II.V) weight is compared to the weight previouslycalculated. If the deviation is larger than 1% of the maximum take-off mass, the aircra weight

and performance data is fed back to the Class I methods, the wing loading and thrust-to-weightratio are recalculated and the whole design process is repeated until the weight converges.

When the aircra weight is converged, the range is tested against the payload-range requirements.To meet the required range, the sensitivity of the fuel mass on the aircra range is determined usinga linear regression and the fuel mass for the next iteration is calculated. Because this changed fuelweight also alters the gross aircra weight, the whole weight convergence loop is repeated. isprocess is repeated until the calculated range is within 1% of the required range. One aircra takesaround 300 seconds on an Intel i7-3610QM CPU to converge for both the range and weight.

Database

RangeConverged?

Configuration

Mission RangePassengers

CargoAirports

Top-Level Requirements

Class II WeightEstimation

Preliminary Sizing(Class I methods)

AerodynamicAnalysisi

Class II.V Weight

Estimation

PerformanceEstimation

WeightConverged?

KPIs and Geometry

Figure 3.2:An abstract overview of the design process

3.3 e Modules

is section gives an overview of all modules currently implemented in the Initiator. In Figure3.3the N2 chart of the module can be found. Inside the rst black square all sizing modules canbe found. ese modules generate a rst estimation of the aircra which is subsequently used bythe design and analysis modules, which can be seen in the second big square in the N2 chart. e


38/136


Initiator contains two loops, which are marked with red squares in the N2 chart. e inner loopmakes sure the weightand aerodynamic loads converge. To converge the wing loading and weightsthe outer loop is used.


39/136

3.3 e Modules 17

Database

Weights

Fuselage&

E

ngine

dim

ensions

Engine

parameters

ClassIWeight

Estimation

Weights

W

eights

Weights

Wing-T

hrust

Loading

Win

gloading

Ge

ometry

Estimation

Aircraft

geometry

Aircraft

geometry

Aircraft

geometry

Aircraft

geometry

Aircraft

geometry

Aircraft

geometry

A

ircraft

geometry

Aircraft

geometry

Aircraft

geometry

CabinDesign

Cabin&Cargo

dimensions

Cabin&Cargo

dimensions

ClassIIWeight

Estimation

Weights&CG

s

Weights&CGs

Weights&CGs

Weights&CGs

Weights&CGs

Weights&CGs

AVLVLM

Aerodynamic

loads

Aerodynamic

loads

CL

,CD

,spanE

ff

CLmax

Emprirical

CLmax

CLmax

EMWETWing

Weight

W

eights

FuselageWeight

Estimation

W

eights

Weights&CG

s

Aerodynamic

loads

Aerodynamic

loads

Class

II.V

Weight

Estimation

Weights&CG

s

EngineModel

SFC

EmpiricalDrag

Estimation

Profiledrag

Weights&CGs

Performance

Estimation

Range,C

Lmin

,

CDmin

,e

Dragpolars,

SFC

,C

Lmax

Fu

elmass

Design

Convergence

W

eights

Figure

3.3:

N2

chartofthemodules


40/136


3.3.1 Sizing Modules

e sizing modules are used to create an initial preliminary sizing of an aircra based on a set of

top-level requirements. ey are skipped when a pre-dened aircra is detected in the XML-le.e end result is a rst estimation of the aircra geometry, weights, propulsion and performance.

Module: Class1WeightEstimation

is module developed by Slingerland as part of the MSc. thesis performs a Class I weight esti-mation [30]. e module is able to size an aircra based on multiple payload-range combinationsand calculates an harmonic range for which all missions are feasible. With this information therequired fuel mass is calculated with the lost range method [3]. e relation between the OEMand MTOM is estimated using a linear regression. By using this relation in combination with thefuel mass and payload mass the MTOM and OEM of the aircra in question can be calculated [27].

Module:WingustLoading

is module chooses the wing loading and thrust-to-weight ratio of the aircra. A large numberof constraints are used to limit the design space:

Take-off distance

Landing distance

Cruise speed

Climb gradients Climb rate

Stall speed

Turn rates

e evaluation of the constraints are based on Roskam [31]. For a full description of the moduleplease read Slingerlands Master thesis [30].

Module:GeometryEstimation

e GeometryEstimationmodule uses the wing loading and weight information to create a rstestimate of the aircra geometry. e wings are sized to meet the wing loading estimated by theWingrustLoadingmodule where canard planform is included to get the required loading. Sincethe canard is estimated with volume coefficientmethods [27], the canard panform area is dened asa function of the wing planformarea and mean aerodynamic chord. erefore the wing and canardsizing is solved simultaneously and iterated until the desired planform area is achieved. All othertail surfaces are also estimated using the volume coefficient method. e sweep angle and taperratio are calculated using empirical relations used by Raymer [27]. e fuselage is sized to meet therequired cabin oor to hold the payload by using a pre-set passenger density. e slenderness isestimated using database values of aircra with similar payload-range requirements. e engineslength and diamter is calculated by using the thrust to interpolate database information.


41/136

3.3 e Modules 19

e main part (which is the fuselage for the congurations currently implemented in the Initiator)is place at (0,0,0). All other parts are positioned with respect to this part by using length fractionsdened in settings. ese settings for the different congurations are fractions of the main part

dimensions.Design rules for the Blended-Wing-Body aircra are created, but only create feasible aircra ge-ometries for very high payload requirements.

3.3.2 Analysis Modules

e Analysis modules are used to analyse the aircra generated by the sizing modules or a pre-dened aircra. All analysis modules require a fully dened aircra; in other words: the sizingmodules need to be completed, or an aircra needs to be present in the XML le.


e Class2WeightEstimation moduleperforms a Class II weight estimation for all the differentpartsof the aircra. It uses the method as written by Raymer [27]. Note that the connectors of theprandtl-wing and winglets are currently not included in the weight estimation.

Module:AVLVLM

AVLisavortex-latticemethoddevelopedbyMarkDrelaofMIT[ 28]. eAVLVLMmodule createsan AVL model of the aircra and performs several AVL runs on a trimmed aircra to get:

Stability and Control derivatives

CLand CDat several angles of attack to calculate the drag polar

Forces and moments at n= 1and n= nmax

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0.01 0.02 0.03 0.04 0.05 0.06

Flight test

AVL

CD

CL

7.5 cts

Figure 3.4:Drag polar comparison of AVL with ight test data from the Airbus A320-100 [9]

Figure3.4shows the comparison of an AVL model of the Airbus A320-100 and ight test data.Because vortex-lattice is an inviscid method, the CD0from the ight test data is also used to con-struct the AVL drag polar. Note that the AVL model has symmetric airfoils (which is not the case


42/136


on the real A320) and the twist angles could have a measurement inaccuracy since these were notclearly visible in the technical drawings. As can be seen the drag polar is estimated quite well, withan error of 7.5 drag counts at CL= 0.5, providing the estimation of the zero-li-drag is accurate.

Since the vortex-lattice method is incapable of modelling thickness effects, inaccuracies may occurin the case of very thick liing surface (as is the case with the Blended-Wing-Body).

Module:TornadoVLM

is module is capable of modelling the aerodynamics of the aircra liing surfaces with theTor-nadovortex-lattice method [32] which is extended by Vaessen [10] with a relaxed wake model andan improved compressibility correction. An example of the output of the Tornado module can beseen in Figure3.5. CurrentlyAVLis used in favour ofTornadosince the Trez-plane analysis ofTornadois very inaccurate. Another fact in favour ofAVLis the runtime; AVLis multiple timesfaster thanTornado. Tornado is validated by Vaessen during his Masters thesis work. e resultsof a few validation cases can be found in Figure3.6.

Figure 3.5:Wake visualisation plot from Tornado

Module: CLmaxEmpirical

is module, developed by Jan Mariens, calculates the CLmax of a clean wing and is based on theESDU 89034 method [33]. Since it does not give reliable results, the method is currently disabled

and CLmaxvalues provided by the user and used. is needs to be xed in a future version.

Module: EmpiricalDragEstimation

eEmpiricalDragEstimationis developed by M. Warmenhoven [34] and is a soware implemen-tation of the method described in Appendices F and G of prof.dr.ir TorenbeeksSynthesis of Sub-sonic Aircra Design[29]. e method can estimate the drag of an aircra in cruise conditionand with different low-speed congurations (including ap deections). Currently only the cruisedrag is calculated, from which only the prole drag is used in other modules of the Initiator. einduced drag is already acquired from theAVLVLMmodule. Note that the method is an empiricalmethod based on data from conventional aircra congurations. Please be suspicious to results


43/136

3.3 e Modules 21

Figure 3.6:Validation cases of Tornado with varying aspect ratios, source: [10]

calculated with this module in the case of unconventional congurations. In Warmenhovens re-port the tool is compared with the Fokker 100 aircra. e tool gives aCD0of0.0215, where thereal aircra has aCD0 of0.0188, an overestimation of 14%. is is compensated in the tool byunderestimating the CLin cruise condition. Since the CD0is the only result used by the Initiator,this could result in inaccurate drag estimation results.

Module:EMWETWeight

is module is a wrapper around the student version of theEMWETClass II.V wing weight esti-mation method developed by A. Elham as part of his PhD thesis [35].EMWETis a quasi-analyticalwing weight estimation method which uses load data from an aerodynamic analysis to calculate thematerial distribution in the wing box which is needed to withstand the loads. [11] Semi-empiricalrelations are used to estimate the secondary weights. Figure3.7shows the correlation between thecalculated wing box weight and the actual wing weight. is t is used inside the program to derivethe wing weight aer the wing box sizing.

Module: FuselageWeightEstimation

ismoduleperformsa Class II.V weight estimation of the fuselage and is developed by K. Schmidtas part of his Masters thesis work [12]. Is is able to perform a weight estimation on conventional aswell as oval-cross-subsection fuselages. It uses the moment and forces calculated byAVL as well ascomponent weights calculated by the Class II or Class II.V weight estimation methods. In Figure3.8the verication of the fuselage weight estimation can be found. e reference aircra used areall quite small, therefore it would be advisable to validate the method with a more varied set ofreference aircra.


44/136


103

104

103

104

F50

B737

B727

A300

B777 A330

Calculated wingbox weight (kg)

Actualwingweight(kg)

Figure 3.7:Relation between wing box weight and total wing weight (source: [ 11])


e Class25WeightEstimation extends the Class II weight estimation module by iterating AVL,EMWETand the Class II.V fuselage weight estimation method until the maximum take-off mass,wing mass and fuselage mass stabilisewithin a pre-set margin. All other parts (eveythingexcept themain wing and fuselage) are calculated with the same methods as the Class II weight estimation.

Module:HighLiDevices

is module, developed by Jan Mariens, uses the ESDU 99031 method [36] to estimate the li-curves of the wing-fuselage combination with high-li devices. It evaluates all possible ap/slatcombinations and chooses the most simple solution which meets the runway requirements. Since

this module had a very long runtime (around 1 minute on an Intel i7-3610QM CPU) it is notincluded as a dependency for other modules. Also the accuracy of the method regarding uncon-

ventional aircra congurations is unknown.

Module:PerformanceEstimation

e PerformanceEstimationmodule combines all data calculated by the analysis modules to createthe aircra drag polar, manoeuvre loading diagram and payload-range diagram. e drag polar isconstructed by combining the induced drag from the Trez plane analysis ofAVLand the proledrag calculated with empirical methods. e drag polar is commonly approximated with a polar


45/136

3.3 e Modules 23

100

101

102

100

101

102

Werkelijk gewicht [x10 3 k g]

Geschatgewicht[x10

3

kg]

Class 2.5Tor enb ee kRaymerHoweNicolai0%15 %35% fout

foutfout

A300B2

A320-200

ATR42

B707-121

B707-320

B707-321

B727-100

B737-100

DC8-55DC8-72

DC9-15

F100

L1011-1

Figure 3.8:Verication of the fuselage weight estimation method (source: [12])

which has the form:

CD = CD0+ C2L A e

(3.1)

To calculate the values ofCD0and e the method presented in Section 5.3 of [29] is used, which isillustrated in Figure3.9. A linear t is made around the cruise li coefficient forCD = f(C2L).e intersection of this line with the C2L = 0is the value ofCD0of the polar approximation.


46/136


0 0.02 0.04 0.06 0.08 0.1

0

0.2

0.4

0.6

0.8

1

1.2

CD

CL2

k =1

A e

CD 0

drag polar

approximation

Figure 3.9:CD0 and e estimation

3.3.3 Design Modules

Design modules add or change the aircra design. In implementation they are not different fromthe analysismodules. However, to separate analysis from design, a different module typeis created.

Note that including design modules in an optimiser workow need to be done with caution. Onehas to be sure that parameters which are present in the design vector are not changed by the mod-

ules.

Module:CabinDesign

is module creates the fuselage interior by creating the cabin oor (and wall and roof in an ovalfuselage) and the cargo bay. It creates aCargopart into the fuselage and ts ULDs container intothe bay. e module is created as part of Schmidts Master thesis [12].

Module:ControlAllocation

eControlAllocationmodule allocates control surfaces to various parts of the aircra by usingsome simple design rules and writes the results to the wing parts. is module will be extended by[37].

Module: PositionLandingGear

ePositionLandingGearmodule estimates the placement of the landing gear by using a large setof constraints which result in the best placement of the landing gear. e module is developed byHeerens [38].


47/136

3.3 e Modules 25

3.3.4 Work-ow Modules

e work-ow modules dont analyse or change the aircra directly. ey implement design work-

ows and in- and output modules.

Module:XMLReader

eXMLReaderreads the aircra denition le and stores the data into the objects foruse by the modules. Existing data previously loaded are overwritten by the values in the XML le.

Module:XMLWriter

eXMLWritermodule writes all data from the Initiator to the aircra denition le. Existing datain the le is overwritten.

Module:DesignConvergence

is module implements the design loop which feeds back the Class II.V weight estimation tore-evaluate the wing loading and thrust-to-weight ratio. e weight converge when the changebetween two iterations fall within a pre-set tolerance. e module also has a loop which changesthe fuel mass until the desired range is met. e work-ow can be found in the activity diagram inFigure3.10. is is the implementation of the design process outlined in Figure3.2.

Module: InteractiveMode

e interactive mode implements a text interface for the user. It has the possibility to run and resetmodules and print results in the console.

Module:BatchMode

e BatchMode modulereadsthe tagintheaircradenitionleandrunsthemodulesin the order they are specied in the le.

Module:PlotTool

e PlotToolmodule has two different run cases. When invoked from the InteractiveMode itpresents the user an input prompt from where the user can plot the aircra geometry and g-ures created by the modules. When called fromBatchMode, it reads its own module input andplots the specied gures.

Module: ReportWriter

e ReportWritermodule collects results from the analysis modules and writes this information toa nicely formatted PDF le. is module required a working LATEX installation. An example of areport generated with the module can be found in AppendixA.


48/136


Get initial va lues

(Call getModuleR es ults method)

Get c onvergence criteria from s ettings

Class 2.5 converged?

R ange converged?

R anges of all requirements met?

R e-run all SizingModules

With C lass 2.5 results

R un Clas s 2. 5 weight estimation

R un performance estimation

Calculate fuel mass required

to achieve range

Figure 3.10:Activity diagram of the DesignConvergence module


49/136

27

Chapter 4

Design Tool Verication

In this chapter the Initiator design tool developed in this thesis project is compared against existingaircra. e Initiator is a tool designed to synthesise a preliminary aircra design from a set oftop-level requirements. An aircra design is in essence a non-unique solution to a problem posedby a set of TLRs. Since the Initiator is a design tool there is, in contrast to an analysis tool, noright answer which can be used to validate the tool. However, assuming commercially designedaircra represent an optimally designed aircra, the design tool output can be veried againstexisting aircra designed for similar top-level requirements. Each analysis tool should be validatedseparately, this is however outside the scope of this thesis. e current state of the different modules

can be read in Section3.3.

A selection of aircra varying from small regional jets to wide-body long-range, jet-powered air-cra is made with different wing and engine congurations. Aircra are chosen for which suffi-cient reference data could be acquired. e aircra are generated by the Initiator utilising a de-sign routine which designs an aircra for a specied payload versus harmonic range combination.e requirements are set to match the performance characteristics of the reference aircra. egenerated aircra are compared with the reference aircra designed for the same payload-rangerequirements.

4.1 Reference aircrae full list of reference aircra requirements can be found in Table4.1. Of all aircra the per-formance gross weights and empty weights are available. Six aircra (A300 B2-100, A320-200,B737-200, DC-10-30, MD-80 and F-100) can be compared using a more detailed weight break-down acquired from [14] and [15]. Drag polars from ight-test data is available for the A320-200,B737-200, B737-800, DC-10-30 and the MD-80. Figure4.1shows the harmonic range and maxi-mum payload mass of different aircra used for the tool verication.

From the aircra performance specications top level requirements are created and Initiator inputles are created. An example of an input le (the Airbus A320-200) can be found in AppendixC. Besides the requirements, the conguration (low or high wing, standard or T-tail, number of

27


50/136

28 Design Tool Verication

Table 4.1:Reference aircraft requirements, source: [13]

Aircra Npax[-] Wp[kg] Mcruise[-] hcr[m] R h[km] Ltakeoff[m] Llanding[m]

A300-B2-100 269 31750 0.78 10058 1720 1850 1768A320-200 150 20536 0.76 11278 2870 2180 1440A340-300 295 50800 0.82 11887 9167 3000 1964A340-600 380 67400 0.82 11887 10556 3100 2240A380-800 555 83900 0.85 11887 12149 2990 2160B737-200 97 11385 0.73 10668 1774 1829 1350B737-800 162 21319 0.79 11887 1363 2101 1440B777-300 394 64000 0.84 10668 3142 2574 1860BAe 146-200 88 10250 0.65 9144 2000 1646 1192F70 70 9302 0.77 10668 1085 1296 1210F100 107 11300 0.72 10668 2556 1856 1321

DC-10-30 285 46180 0.82 9449 6995 2996 1820MD80 155 18236 0.76 10668 1453 2195 1481

engines) and the wing aspect ratio is dened. Also the CLmaxof the aircra in clean, take-off andlanding conditions need to be specied. e choice of the landing CLmaxhas great inuence on theaircra performance since this is currently the only constraint for the maximum wing loading.1

eCLmax values are chosen such that the wing loading of the generated aircra lies within theloading of its reference aircra. e used parameters can be found in Table4.2, general settingused for all congurations can be found in Table 4.3. All other aircra parameters are estimatedby the Initiator. All aircra are designed by running the DesignConvergencemodule. e global

working of this design convergence module can be found in Chapter3.

Table 4.2:Initiator input parameters

Aircra A CLmax,landing CLmax,take-off CLmax,clean

A300-B2-100 7.73 2.6 2.2 1.2A320-200 9.39 3.2 2.2 1.2A340-300 10.02 2.5 2.2 1.2A340-600 9.16 2.5 2.2 1.2A380-800 7.52 2.6 2.2 1.2B737-200 8.83 3.2 2.2 1.2

B737-800 9.45 3.4 2.2 1.2B777-300 8.68 2.8 2.2 1.2BAE146-200 8.98 2.8 2.2 1.2F70 8.43 3.0 2.2 1.2F100 8.43 3.0 2.2 1.2DC-10-30 6.91 2.7 2.2 1.2MD80 9.62 3.2 2.2 1.2

1In future Initiator versions the wing loading will also be constrained by buffet onset in cruise conditions


51/136

4.2 Comparison based on Design point 29

A300 B2-100

A320-200

A340-300

A340-600

A380-800

B737-200

B737-800

B777-300

BAE146-200

DC-10-30

F-100

F-70

MD-80

0

10

20

30

40

50

60

70

80

90

0 2000 4000 6000 8000 10000 12000 14000

Payloadmass[metrictons]

Range [km]

Figure 4.1:Payload - harmonic range combinations of the reference aircraft

Table 4.3:Initiator performance settings

Parameter Value Description

CD0,take-off 0.0350 Zero-li drag increment, take-off conguration, aps & landing gearCD0,landing 0.0850 Zero-li drag increment, landing conguration, aps & landing gearetake-off 0.05 Span efficiency increment, take-off congurationelanding 0.10 Span efficiency increment, landing congurationBPR 6.0 Engine bypass ratio

4.2 Comparison based on Design point

e design point is the wing loading - thrust-to-weight-ratio combination chosen for the designed

aircra. e design space is the space set up by all possible wing loading versus thrust-to-weightratio combinations. e design space is constrained by a multitude of requirements. e choice ofdesign point hasa biginuences on the aircra design. But because limited information is availableat the conceptual design phase a lot of assumptions need to be made. A majority of constraints onthewing loading are a resultof the high-li performanceof theaircra. Becausethere is no high-lior clean CLmaxprediction available in the tool at the time of writing, all CLmaxvalues are set by theuser. e design point which results in the highest wing loading and still meet all the constraintsis selected by the design tool. For more information please read Slingerlands Masters thesis [30].

As can be seen in Figure4.2, the wing loading of the aircra with the A320-200 requirementsmatches the wing loading of the aircra present in the Initiator database. Note that the referenceaircra presented in the Figure are not the reference aircra used in this Chapter. e refer-


52/136

30 Design Tool Verication

ence aircra used in the sizing are gathered from a database inside the Initiator. Based on thepayload-range requirements a selection of similar aircra is made. For more information pleaseread Slingerlands Masters thesis [30].

0 1000 2000 3000 4000 5000 6000 7000 80000

0.1

0.2

0.3

0.4

0.5

Takeoff Wing Loading (W/S) [N/m2

]

Ta

keoffThrusttoWeightratio(T/W)[]

sL

= 1440 m

bmax

= 80 m

(c/V)FAR 25.111c

= 1.2 %

(c/V)FAR 25.119

= 3.2 %

(c/V)FAR 25.121a

= 0 %

(c/V)FAR 25.121b

= 2.4 %

(c/V)FAR 25.121c

= 1.2 %

(c/V)FAR 25.121d

= 2.1 %

sTO

= 2180 m

Mcr

= 0.76

tclimb

= 10 min to h = 4000 m

Design PointReference Aircraft

design

space

Figure 4.2:Design point of the Initiator generated aircraft according to A320-200 specica-

tions

e activeconstraints are the landing distance, cruisespeed and take-off length. In other words, thedesign space is bound by these constraints. Note that the landing and take-off distances are highlydependentonthe CLmaxvalues suppliedby the user, therefore it cannot be stressed more that a goodCLmax estimation method is required to improve the design tool and make it less dependent on

values given by the user. e cruise speed constraint (and also the climb requirements) are highlysensitive to changes in the aircra drag polar. Since the prole drag is currently estimated with anempirical method, its validity cannot be ensured for unconventional aircra congurations.

All design points of the generated aircra are compared to the reference aircra in Figure4.3. ewing loading estimation is quite good, although this is highly inuenced by the selected CLmax

values. e thrust-to-weight ratio is not estimated correctly for all aircra. Further inspectionreveals that the thrust-to-weight ratio is driven up by the cruise speed requirement in the caseswhere the thrust-to-weight ratios have a large difference.

For example the generated aircra with AirbusA380-800 requirements hasa thrust-to-weight ratioof0.34, where the real A380-800 has a thrust-to-weight ratio of0.24. e cause of this high

value of the cruise speed constraint is the high drag coefficient, which is a direct result of the highcruiseCLof0.75. eCD0 is 203 drag counts, where the totalCD is 491 cts; from this can beconcluded that the li-induced drag is very high, which is a result of the high cruise li coefficient.In reality this high CLcould result in buffet onset in cruise conditions. Since the aircra is alwaysdesigned such that the li equals the weight, this means that the wing loading is too high for theaircra. erefore a buffet onset constraint is needed to limit the wing loading, which would resultin a lower cruise CL. is lowerCL will result in a lower CD during cruise, which decreasesthe required thrust (or trust-to-weight-ratio). is results into smaller engines, but larger wings.e other aircra which have a large thrust-to-weight ratio show the same high cruise CL andcorresponding high CDvalues.


53/136

4.3 Comparison based Geometry 31

F

-70

BAe-146

-200

F-100

B737-200

MD-80

A300-B2-100

B737-800

B777-300

A32

0-20

0

DC-10-30

A380-800

A340-300

A340-600

0.2

0.22

0.24

0.26

0.28

0.3

0.32

0.34

0.36

3000 4000 5000 6000 7000 8000

Roux

Initiator

Wing loading [N/m2]

Thrust-to-weightratio[-]

Figure 4.3:Comparison of the design point of the Initiator generated aircraft and data from

Roux [13]

4.3 Comparison based Geometry

e Initiator design tool generates 3-dimensional geometry of the designed aircra. e congu-ration (high or low wing, tail type) and the aspect ratio are dened by the input. All other variablesare calculated by the design tool. e Initiator uses a lot of settings to generate the aircra. emost relevant settings are listed in Table4.4. is section will compare the geometry generated bythe Initiator with the actual aircra.

Figure4.4compares the A320-200 with an aircra generated by the Initiator with the same re-quirements. As can be seen the geometry is estimated quite well, note that since the aspect ratio isa user-input, this is an exact match. e vertical tail size (and to a lesser degree, the horizontal tailtoo) is a little underestimated. However, the vertical tail is positioned further a than the reference

aircra, which results in a longer moment arm.

e fuselage length is clearly overestimated. Its planform area is sized by using a pre-set passengerdensity. is setting is currently an average density found in literature for a three-class congura-tion. (For more information please see [12]). is could be improved by making this parametersensitive to range, since longer hauls require more room per passenger than short ights. Alsothere is no distinction made between single and twin-aisle fuselages.

Other little differences canbe seenbetwee

Synthesis of Novel Aircraft Concepts for Future Air Travel

Documents

Transcript of Synthesis of Novel Aircraft Concepts for Future Air Travel