Understanding Aircraft Structures

Understanding Aircraft Structures
John Cutler
Fourth Edition revised by Jeremy Liber
John Cutler 1981, 1992, 1999 John Cutler and Jeremy Liber 2005 Editorial ofces: Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK Tel: +44 (0)1865 776868 Blackwell Publishing Inc., 350 Main Street, Malden, MA 02148-5020, USA Tel: +1 781 388 8250 Blackwell Publishing Asia Pty Ltd, 550 Swanston Street, Carlton, Victoria 3053, Australia Tel: +61 (0)3 8359 1011 The right of the Author to be identied as the Author of this Work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher. First published in Great Britain by Granada Publishing 1981 Second edition published by Blackwell Scientic Publications 1992 Third edition published by Blackwell Science Ltd 1999 Fourth edition published by Blackwell Publishing Ltd 2005 ISBN-10: 1-4051-2032-0 ISBN-13: 978-1-4051-2032-6 Library of Congress Cataloging-in-Publication Data Cutler, John. Understanding aircraft structures / John Cutler.4th ed./rev. by Jeremy Liber. p. cm. Includes bibliographical references. ISBN-13: 978-1-4051-2032-6 (alk. paper) ISBN-10: 1-4051-2032-0 (alk. paper) 1. Airframes. I. Liber, Jeremy. II. Title. TL671.6.C88 2005 629.13431dc22 2005048077 A catalogue record for this title is available from the British Library Set in 10 on 12 pt Times by SNP Best-set Typesetter Ltd., Hong Kong Printed and bound in India by Replika Press Pvt Ltd, Kundli The publishers policy is to use permanent paper from mills that operate a sustainable forestry policy, and which has been manufactured from pulp processed using acid-free and elementary chlorine-free practices. Furthermore, the publisher ensures that the text paper and cover board used have met acceptable environmental accreditation standards. For further information on Blackwell Publishing, visit our website: www.blackwellpublishing.com
Preface
The fourth edition of Understanding Aircraft Structures builds, naturally enough, on the rm foundations of the earlier editions in aiming to keep up to date with an evolving industry. Whilst the fundamentals of how aircraft are constructed change only relatively slowly, the aerospace world as a whole tends to change at a more noticeable pace and many of these changes impinge upon the realm of aircraft structures. The major change since publication of the last edition has been in the regulatory environment: all of the national European aviation regulators have been subordinated beneath a new European Union body, the European Aviation Safety Agency (EASA). This change is reected in the chapter describing how to apply for approval of an aircraft modication and, to a lesser extent, in the section on quality and airworthiness. The broad range of subjects which support, and interact with, the work of the aircraft structures designer is covered in greater detail than before and the opportunity has been taken to provide a round up of the key points at the end of each chapter. In their own studies the authors have found that this kind of review at the end of a chapter greatly helps the learning process. Finally, in recent years we have seen the European Airbus organisation surpass the great Boeing Company in terms of numbers of aircraft sold. However, the American aircraft industry remains the dominant force in the world of aviation. Its inuence is such that, probably, the majority of the industry continues to work in imperial rather than SI units. For this reason, and after much soul-searching, it has been decided to retain both imperial and SI units in this book.
Contents
Preface Chapter 1 Chapter 2 Introduction History 2.1 Outline 2.2 Wire-braced structures 2.3 Semi-monocoque structures 2.4 Sandwich structures 2.5 Review of the key points Parts of the Aircraft 3.1 Terms connected with ight 3.2 Terms connected with control 3.3 Terms connected with high-lift devices 3.4 Terms associated with the shape and dimensions of the aircraft 3.5 Review of the key points Loads on the Aircraft 4.1 General ight forces 4.2 Acceleration loads 4.3 Further aerodynamic loads 4.4 Other loads 4.5 Further load factors 4.6 Loads acting on the whole aircraft 4.7 Review of the key points 4.8 References The Form of Structures 5.1 Structure relative to aircraft design 5.2 Historic form of structure 5.3 General form of structure 5.4 The basic load systems in structures 5.5 The forms of stress in materials 5.6 Bending and torsion 5.7 Compression
ix 1 3 3 4 9 14 15 29 29 31 32 32 37 38 38 44 48 50 51 53 55 56 57 57 57 59 59 63 72 85
Chapter 3
Chapter 4
Chapter 5
vi
Contents
5.8 The whole structure 5.9 Review of the key points 5.10 References Chapter 6 Materials 6.1 Choice of materials 6.2 Material properties 6.3 Smart structures (and materials) 6.4 Cost as a property of a material 6.5 Heat treatment 6.6 Reference number for materials 6.7 Review of the key points 6.8 References Processes 7.1 Introduction 7.2 Manufacturing 7.3 Jointing 7.4 Review of the key points 7.5 References Corrosion and Protective Treatments 8.1 Nature of corrosion 8.2 Causes of corrosion 8.3 Protection against corrosion 8.4 Review of the key points 8.5 Reference Detail Design 9.1 Sheet-metal components 9.2 Machined components and large forgings 9.3 Notching and stress raisers 9.4 Rivets and bolts 9.5 Joggling 9.6 Clips or cleats 9.7 Stringer/frame intersections 9.8 Lugs 9.9 The stiff path 9.10 Review of the key points Composite Materials in Aircraft Structures 10.1 What are composites? 10.2 The strength of composite materials 10.3 Types of structures 10.4 Joining composites 10.5 Fibres 10.6 Resins 10.7 Working safely with composites 10.8 Review of the key points
87 87 88 89 89 92 101 103 105 106 109 110 112 112 112 119 123 123 124 124 126 128 132 132 133 133 136 139 144 152 154 155 156 157 157 159 159 161 162 164 167 169 171 172
Chapter 7
Chapter 8
Chapter 9
Chapter 10
Contents
vii
Chapter 11
Quality and Airworthiness 11.1 Quality assurance and quality control 11.2 Control 11.3 Procedures and systems 11.4 Further notes on quality control functions 11.5 Airworthiness engineering 11.6 Continued airworthiness 11.7 Review of the key points 11.8 References Stressing 12.1 Introduction 12.2 The stressmans work 12.3 Stressing methods 12.4 Stress reports 12.5 Review of the key points 12.6 References Presentation of Modications and Repairs 13.1 Denitions 13.2 The essential paperwork associated with modications 13.3 Review of the key points 13.4 Conclusion 13.5 References
173 173 174 175 177 179 180 181 181 182 182 183 187 190 193 194 195 195 197 201 201 202 203 209
Chapter 12
Chapter 13
Appendices Index
Chapter 1
Introduction
The aim of this book is to present the principles of aircraft structures to the interested reader in a manner that is both clear and thorough whilst avoiding the necessity for complex mathematical formulae. No previous knowledge of the eld is assumed, only the desire to know more. Like most industries, aviation has various specialist elds which are often considered to be black arts by the uninitiated; the work of the aircraft structures analyst, or stressman, falls into this category. However, as with most disciplines, given a little curiosity and some application the basic principles may be easily understood, at which point much of the mystery disappears. There are many people involved in the engineering side of the aircraft industry, such as draughtsmen, tters or licensed engineers, who deal with aircraft structures in their daily lives, yet may have only an incomplete understanding of why aircraft are designed as they are. For those people, and others with enquiring minds, reading this book will not make them stressmen. It will, however, give them a qualitative understanding of the general principles of aircraft structures, allowing them to ask informed questions. The book is written in such a way that individual chapters may be read independently of each other if a topic is of particular interest to the reader. However, taken together they provide a logical progression and offer a thorough introduction to the subject. The book begins by giving an historical perspective, presenting a brief outline of the evolution of aircraft structures from the earliest ying machines to the present day. The one major change in philosophy made by aircraft structures designers is described. After this the basic shape and main structural elements of the aeroplane are dened, whilst introducing the fundamental reason why an aircraft requires a structure at all; that is to support the various loads applied to it. Subsequent chapters explain the form which that structure takes in order to efciently carry those loads, the materials that are employed to make it and the processes involved in its construction. There follows an introduction to the principles of corrosion protection and there is a distillation of much rule of thumb and good engineering practice so essential to the sound detail design of aircraft structures. The regulatory and quality environment within which the
aviation industry functions is outlined. Later chapters allow an insight into the work of the stressman and describe, in broad terms, the justication and documentation necessary to gain approval for aircraft modications. For anybody intending to become an aircraft structures specialist (stressman) Understanding Aircraft Structures will provide a rm foundation upon which to build and will provide explanations of phenomena which may not be easily found in more formal textbooks. Those who are simply looking for a broad appreciation of the subject will also nd all that they require in this book.
Chapter 2
History
2.1 Outline
In this chapter we will book at the general development of aircraft structures over the short period of their history. As with most subjects, knowledge of the steps which led to the present position is a great help in understanding current problems; later in the book there are more detailed comments concerning structures as they are now. Flying machines obviously changed enormously over the 70 years from the Wright Brothers Flyer at Kittyhawk to Apollo on the Moon, and a ghter ace of 1918 ew a very different aircraft from that own by his successor today, so a review of the whole development of ying would be a large task. However, there are many different branches of science and engineering which make up aeronautics as a whole and when these are looked at separately, the problem of dealing with them becomes more manageable. The main divisions of aeronautical engineering are (a) the science, which deals with the airow round the aircraft; (b) the power-plant engineering; (c) the avionics, that is the radios and navigation aids; (d) the airframe engineering, where the airframe includes hydraulic and electrical systems, ying and engine controls, interior furnishings and cargo systems; and (e) the section which concerns this book, which is the structure. All these divisions and subdivisions have developed at different rates. Power plants (engines) for instance have moved with two great strides, and many years of continuing short but rapid steps. Before the Wright Brothers could make their rst successful aeroplane, the power plant engineers had to make their rst stride and invent an engine which was light and powerful. The next stride was the invention of the jet engine but, in between, the power of piston engines increased nearly 200 times in just 40 years, from 12 hp (horse power) to over 2000 hp, with only a ten times increase of weight. As we shall see, structures have made only one major fundamental jump forward, but that was sufcient to change the whole character and appearance of aircraft.
2.2 Wire-braced structures

If we look at the aircraft in Fig. 2.1 we can have no doubt about the form of its construction. The wings and the fore and aft structures carrying the other covered surfaces were all made of rectangular frames which were prevented from collapsing (or parallelogramming) by wires stretched from corner to corner.Although the methods were not original, there were two imaginative pieces of structural thinking here. Firstly, the idea that two wings, one above the other, would make a lighter, stronger structure than the type of wing arrangement suggested by bird ight and, secondly, the idea that a rectangle could be held in shape with two light wires rather than with one much heavier diagonal member, like the structure of a farm gate. At this stage, and for the next 30 years, the major structural material was wood; at rst bamboo and later mainly spruce, a lightweight timber with very straight grain and medium strength. Strangely enough, balsa wood, which means so much to the model aircraft enthusiast, was not used during this period but has been used sometimes since then as a ller, or core material, in ooring panels for large aircraft. Wire bracing continued to be used as a major feature of aircraft construction for many years. Figures 2.2 and 2.3 illustrate its extensive use on early ghters and Fig. 2.4 shows it still in evidence into the era of metal aircraft. (Note: these particular aircraft are illustrated as they show a progression of designs for the same manufacturer: Sopwith Aviation Co., which became Hawker in 1920.)
2.2.1 Biplane structures

Biplane structures (Figs 2.12.3) dominated aircraft design for many years and, for certain particular requirements, such as aerobatic aircraft or agricultural crop sprayers, they still appear from time to time. The structural advantage of the arrangement is that the combination of upper and lower wing, the interplane vertical members (struts) and the wire bracing forms a deep, light member which is very rigid and resistant to bending and twisting.
2.2.2 The change to metal construction

The biplane era lasted until the middle of the 1930s, by which time wooden construction was being replaced by metal. The similarity of construction between the biplanes in Figs 2.2 and 2.3 is obvious but the main fuselage members of the later aircraft are steel tubes, as are the wing spars. Although wood was still being used very extensively at the time of the Hawker Fury, by the time that the Hurricane was produced (see Fig. 2.4(a)), the change to metal was almost complete. In spite of this change, the structures were still wire braced, and the principles of structural thinking behind the Sopwith Camel of 1917 and the Hurricane of 1935 showed some distinct similarities. Although the structures designer had
History
Fig. 2.1
Wright Flyer (1903).
Fig. 2.2
Sopwith Camel (1917).
been pushed into accepting the problems of designing monoplane wings as part of the progress towards better ying performance, the main fuselage structure of both aircraft was four longitudinal members called longerons and wire bracing was still being used extensively in the Hurricane. The side panels of the later fuselage were braced on the Warren girder principle, but this could not be called a great leap forward. This particular aircraft was designed at a period when some very rapid advances were being made in aircraft performance, and the airforces of the world were expanding in clear preparation for the coming Second World War, so some safe conservative thinking on the design of the structure was understandable. However, the difculty of combining metal construction with traditional fabric covering in a monoplane wing produced the strange design shown in Fig. 2.4(b). This was very short lived and all but the rst few Hurricanes had metal-covered wings, but it is worth seeing how the unfamiliar problems of preventing bending and twisting in a
History
Fig. 2.3
Hawker Fury (1931).
8 Understanding Aircraft Structures
Fig. 2.4(a) Hawker Hurricane (1935).
History
Fig. 2.4(b) Hawker Hurricane wing construction. (From Aircraft Production courtesy of IPC Transport Press.)
monoplane wing were tackled by one designer in the interim period between biplane construction and the type of structure we use today.
2.3 Semi monocoque structures

An essential feature of the fuselages discussed so far was the internal cross-bracing which, although acceptable for some types of aircraft, was denitely in the way when fuselages became large enough to carry passengers seated internally. A much more suitable type of construction had in fact been ying for some while.When we look back into aviation history, we refer to the ying boat without perhaps recognising just how well it was named. The fuselages of the early wooden ying boats were made to the highest standards of yacht construction with bent wood frames and double or triple ply skins, carvel built (smooth exterior) with a clear varnished nish. This method of construction presented a much more open and usable fuselage interior. The wooden frames and plywood skins passed from ying boats to land planes and the particular building technique produced some quite shapely aeroplanes compared with the very square slab-sided shapes which were the hallmark of the wire-braced fuselage. The main advantage of the boat type of construction from the engineering point of view was, and is, that the skin forms an integral working part of the structure, unlike the covering of the braced fuselage,
10
which is a complete framework and which would still be just as strong without the fabric. Because the skin is working, the boat construction came to be called stressed skin; it is also sometimes called semi monocoque. (Coque is the French word for shell and also the word meaning hull (of boats). Mono in French implies all in one piece or integral. Semi is an English addition because the construction uses an internal framework which is not a feature of pure monocoque.) The adoption of this type of construction, with its material changed from wood to metal, constitutes the one major fundamental stride forward that aircraft structures have taken since the earliest days of ying.The simplicity and elegance of this method of design and the inuence it brought to bear on the shape and appearance of aircraft is shown in Fig. 2.5. The Douglas DC.3 was not the rst all-metal stressed-skin aircraft, but it was a brilliant example of a new technology converted into successful engineering; it set standards and methods of structural design which have already lasted for decades and look as if they will go on for generations.
2.3.1 Stressed-skin wings

After the fuselage shells came the stressed-skin wings. The American aviation industry, by 1930, was already showing signs of dominance in the building of civil aircraft and was gaining stressed skin know how with large-capacity fuselages. European industry was more concerned with military ying and secrecy, and continued to design braced structures for perhaps two generations of aircraft (about ve to seven years at that time) longer than they should have done. To be fair, there was a problem of size. If a DC.3 wing was scaled down to a size suitable for a ghter aircraft, the sheet aluminium used for the construction tended to become unmanageably thin and imsy. However, some European design ofces, notably Messerschmitt and Supermarine, faced up to the problem and were making efcient stressed-skin wings for small aircraft by the mid-1930s.
2.3.2 Later developments

The later illustrations in this chapter, Figs 2.6 onwards, show the mainstream of structural development since the DC.3. There has been a number of different structural methods invented over the years but the arrangement of vertical members (called ribs in the wings and frames or rings in the fuselage) and longitudinal members (called stringers) has been the established convention for some years. A glance at these post-1955 designs shows very clearly that the structures of big aircraft are not just little aircraft structures scaled up (or the opposite way round). In fact, whatever the size of the aircraft, the fuselage frames are always about 500 mm (20 in.) apart and have between 75 mm (3 in.) and 150 mm (6 in.) deep cross-sections. This means that big aircraft have many structural members and small aircraft have few. This
History 11
Fig. 2.5
Douglas DC.2/3 (1933).
12
Fig. 2.6(a) Starship Corporation.)
fuselage
structure.
(Courtesy
of
Beech
Aircraft
Fig. 2.6(b)
Beech Starship. (Courtesy of Beech Aircraft Corporation.)
situation is not altogether what one would expect, and although scaling up a small structure to make a large one would give a strange result, scaling down from large to small may be seen to have more possibilities when the resultant small spaces between frames and stringers are compared with the centre cores of the composite materials mentioned in Section 2.4. Some other differences between the aircraft illustrated are discussed in Chapter 9, and we will restrict this chapter to saying that the developments which have taken place in recent years have been mainly towards (a) a reduction of the number of rivets in the aircraft, either by machining large pieces of structure from solid (see Fig. 9.12), or by the adhesive bonding assembly of components and (b) reducing the effects of minor
History 13
Fig. 2.7
British Aerospace HS 748. (Courtesy of British Aerospace.)
14
structural damage, either by providing sufcient members so that the failure of one is not disastrous (fail-safe structure), or by improving access for easy inspection of structures in service. The methods of design and manufacture for this stressed-skin type of construction are now so well understood by the manufacturers that the design of the structure for a new medium-size, medium-performance airliner could be regarded as a routine exercise, but the constant commercial pressure to reduce weight provides a strong incentive for improvement.Another area of major interest and innovation for the structures engineer is the search for methods of improving the reliability of structures; that is, reducing the time spent on the ground when minor structural defects are looked for and repaired. The British Aerospace 146/Avro RJ (shown in Fig. 2.10) is very interesting in this respect and the designers took a great deal of trouble to reduce the number of members and the number of places where structural deterioration, such as fatigue cracking or corrosion, may start and not be seen at inspections.
2.4 Sandwich structures

One of the main problems and limitations of the skin material in stressedskin construction is its lack of rigidity. As we shall see later in this book, skins often have to be made thicker than they might otherwise need to be because of a tendency to crumple under some types of load. A strip of paper illustrates this problem very well, one can pull it but not push it. A way of providing thin sheets with rigidity is to make a sandwich with one very thin sheet, a layer of very light but fairly rigid core material, and another very thin sheet, all bonded together with an appropriate adhesive. As with conventional semi monocoque structures, wooden construction led the way with sandwich structures. The famous and elegant de Havilland Mosquito of 1940 was built with plywood skins either side of a balsa-wood core. In todays major structures, a metal core of honeycomb-like cells is recognised as the most suitable core for metalfaced sandwich (see Figs 2.13 and 2.14 and Appendix 3).
2.4.1 Advances in sandwich construction

Although the advantages of honeycomb sandwich have been recognised for many years and at boards in metal and plastics are a standard product, shaped products such as those in Fig. 2.14(b) are difcult to make and therefore expensive. An appreciable percentage of Boeing 747 and 777 structure is sandwich construction (Fig 2.14(c)) but increasing use of composite material in other forms has mitigated against sandwich in the sense that glass- or carbon-reinforced plastic skin is thicker than metal skin and less prone to buckling (see Chapter 11). The Airbus A340 shown in Fig. 10.3 (page 168) has a high percentage of composite, including sandwich structure, but the rate of adoption of honeycomb sandwich is slower
History
15
than many structures designers would wish.The concept of shell structures that are as smooth inside as they are outside, and small fuselages with a ratio of structure thickness to outside diameter which is as good as that of large fuselages, was a dream rather than a reality when this book was rst published in 1981. Since then the Beech Starship (Fig. 2.6, page 12), which is clearly highly innovative in many ways, has been designed with a fuselage which exactly displays the advantages of a honeycomb shell, yet even this advanced aircraft has now ceased production and has been withdrawn from service. Compare Fig. 2.6 with Fig. 5.2 and the extent of the advance which has been made becomes apparent. Innovation on this scale clearly involves massive investment in research, design, tooling and courage and although the existing structural conventions will remain for some time, the Starship fuselage provides an elegant example of the benets of sandwich construction. A British innovation some years ago avoided one of the problems of straightforward honeycomb core material. By a modication of the hexagon pattern a core was made which could be formed around panels which were double curved, i.e. dome shaped. This facility is not possible with a conventional core, which has to be carved to shape from a large block.
2.5 Review of the key points

The structures of early aircraft were comprised of open wooden (or later metal) frames with diagonal wire bracing. The covering, if there was any, served no real structural purpose (its role being mainly aerodynamic). The one fundamental step forward made in the eld of aircraft structures was the replacement of this open form of structure by the stressed skin (or semi monocoque) form, the key difference being that in the latter the covering (or skin) is now a key structural member. Since the 1930s this type of construction has predominated and continues to do so to this day whatever the material used.
16
Fig. 2.8(a) De Havilland Canada Twin Otter. (Courtesy of The de Havilland Aircraft of Canada Ltd.)
History
17
18
Fig. 2.8(b)
De Havilland Canada Dash 7. (Courtesy of The de Havilland Aircraft of Canada Ltd.)
History
19
20
Fig. 2.9
Structural detail. Airbus A300B. (Courtesy of Airbus Industry S.A.)
History
21
22
No stringers no cleats
Fig. 2.10
The British Aerospace 146. (Courtesy of British Aerospace.)
History
23
3 ribs in C/S
C L
2 spars 146
C L
2 panels top
4 panels bottom 146
Fig. 2.11
Boeing 747. (Courtesy of The Boeing Commercial Airplane Company.)
History 25
Fig. 2.12
HS 125. (Courtesy of British Aerospace.)
26
Fig. 2.13 Typical at panel methods. (Courtesy of Ciba-Geigy, Bonded Structures Division, Cambridge.)
History
27
Fig. 2.14(a) Laying up a sandwich panel. (Courtesy of Ciba-Geigy, Bonded Structures Division, Cambridge.)
Fig. 2.14(b) Typical sandwich construction. (Courtesy of Ciba-Geigy, Bonded Structures Division, Cambridge.)
Fig. 2.14(c)
Typical use of sandwich structure. (Courtesy of the Boeing Commercial Airplane Company.)
Chapter 3
Parts of the Aircraft
3.1 Terms connected with ight

There are many words used to describe parts of the aircraft and its operation which, although not directly the concern of the structures engineer, will be used in the text and therefore must be clearly in mind before we start. Later chapters will examine and dene terms and expressions which are particularly relevant to aircraft structures. The study of how an aircraft ies is called the science of aerodynamics and the engineers who practise it are aerodynamicists. Immediately we have used a word which requires some explanation; the word aircraft. Specically aircraft covers all heavier-than-air ying machines, i.e. other than balloons or airships. Thus aircraft covers aeroplanes (or airplanes in America), be they powered or un-powered, which have wings that are xed relative to the rest of the structure, helicopters with wings that rotate above the main structure in order to produce lift and thrust and autogyros with wings that rotate above the main structure in order to produce lift but not thrust. Structurally they all follow the same principles and so, throughout this book, we generally use the term aircraft for the sake of simplicity. Aircraft engines or power plants are internal combustion engines, that is the basic fuel is burnt inside the engine and not externally as it would be in the case of a steam engine with an external boiler. The fuel is either gasolene (Avgas) or kerosene (Avtur). Power plants are either piston engines which drive propellers, or turbines which operate in one of the three following ways. A pure jet turbine is a producer of high-pressure gas which is expelled backwards to thrust the aircraft forwards. Prop-jets are turbine engines driving propellers of the same pattern as those driven by piston engines, and by-pass engines are turbines which combine a pure jet function with driving a multibladed propeller enclosed in a large circular duct. Aircraft are kept up in the air and can y because of lift produced by the wing or mainplane. We deal with lift in more detail in Chapter 4 but for the moment we will accept that a at strip of rigid material pushed edgeways through the air will generate lift if its forward or leading edge is slightly higher than its rearwards or trailing edge. Provided that the
30
angle of the plate to the airow is kept below about 1015 the greater the angle becomes, the higher the lift becomes. Also, if the angle is kept constant but the speed through the air is increased, then more lift is generated. (Some simple but quite informative experiments on lift and other properties of at plates or model wings can be made by a passenger in a motor vehicle holding the test piece out of a window, but the time and place for such activities should be selected with some care.) If we imagine an aircraft ying straight and level at a constant speed the lift produced by (or generated by) the wing must equal the weight of the aircraft. If the lift exceeds the weight, the aircraft will be pushed higher and if the weight exceeds the lift the aircraft will sink. As the speed is constant the lift is maintained at the correct amount by keeping the angle of the wing to the airow at a constant correct amount. In conventional aircraft this is achieved by the action of the tailplane. (Note here that mainplanes can be called wings but the horizontal tail surface, although very similar in shape and construction to a wing, is never referred to as a tail wing.) Other names for the tailplane are stabiliser, horizontal stabiliser or horizontal tail and it works in the following way. If the mainplane angle increases slightly the body of the aircraft is rotated in a tail downnose up direction and the tailplane is also given an increased angle to the airow. This increased angle produces a lift on the tail which levers the body and with it the mainplane back to their original angles. A nose down attitude is similarly corrected by the tail working the opposite way and pushing the rear of the aircraft down. This tailplane action is continuous and produces the quality called stability. The degree (or quantity) of stability is decided by the designer when he considers the specication of the whole aircraft. Clearly a single-seat aircraft intended for competition aerobatics and quick manoeuvres does not want to be too stable, and equally clearly a private aircraft intended for social air touring without the aid of expensive automatic pilot equipment needs to be stable so that the human pilot is not too busy just maintaining level ight. Achieving the correct amount of stability is a difcult problem for the aerodynamicist. We have already seen that some parts of the aircraft can have more than one acceptable name and Fig. 3.1 summarises the situation so far. This gure points out an interesting piece of aeronautical use of language. A bird has two wings but a conventional aeroplane is a monoplane, i.e. it has one wing. On the other hand, in the assembly shop of an aircraft manufacturing plant, when one side of a wing is being attached to a fuselage the action will be called bolting on a wing or something similar, it will not be bolting on half a wing. Some English names of aircraft parts have a French origin, reecting the heavy French inuence on early aeronautical research. In some books, although the use of the name has declined, the whole tail unit, that is the extreme rear fuselage, the vertical stabiliser and the horizontal stabiliser together with the control surfaces (see below), are referred to as the empennage. We shall notice other words with a French inuence later.
31
Fig. 3.1
Parts of the monoplane.
3.2 Terms connected with control

When the aircraft is ying it needs to be controlled by the pilot and we think of it being controlled about three axes of movement (Fig. 3.2). One axis is an imaginary pivot line stretching from wing tip to wing tip and rotation about this axis which causes nose up or nose down movement is called pitching. In engineering terms the pitching axis is also called the YY axis. The imaginary horizontal line extending from the extreme nose of the aircraft to the extreme tail is called the rolling, or XX, axis and movement about this axis is called rolling. The third axis is vertical and passes through the intersection of the other two. Rotation about this ZZ axis is called yaw and hence ZZ is the yaw axis. Figure 3.3 shows the controls which effect the movements about the three axes. Pitching is controlled by the elevators. Roll is controlled by the ailerons. Yaw is controlled by the rudder. These three are called control surfaces and are usually balanced as shown in the gure. Aerodynamic balance reduces the effort required to move the surface and the mass (or static) balance avoids utter, which is discussed in Chapter 4.
32
Fig. 3.2
Aeroplane axes.
A further small control surface is attached to the trailing edge of the main control surface. These small surfaces are called tabs. The operation and purpose of trim tabs and servotabs can be investigated in any book dealing specically with aerodynamics.
3.3 Terms connected with high-lift devices

The majority of aircraft are equipped with additional surfaces on their wings to improve lift at slow speed. Figure 3.3 shows the names of some typical high-lift devices which go under the general group name of aps. Again, it would be better for the reader who is interested in aerodynamics to investigate the function and working of aps for himself. So far as this book is concerned, it is sufcient to say that slots, which are the gaps between the aps and the main wing, smooth out airow and delay the stall (see also Section 4.6) and aps either increase camber (see Fig. 3.4), or increase wing area, or both.
3.4 Terms associated with the shape and dimensions of the aircraft
The names of the leading dimensions are shown in Fig. 3.4. The usual American practice (and as shown in the gures in this section) is to quote leading dimensions in feet and inches but other dimensions, even large
33
Fig. 3.3
Controls and aps.
distances such as door openings, in inches. Europe and other parts of the world tend to quote dimensions in metres and/or millimetres. Figure 3.5 shows the method of describing by drawing the complex shapes of an aircraft. The technique was inherited from the boat-building industry, which accounts for the name waterline for any horizontal section or cut. In the diagram the three sectional cuts are each shown as a single division, but in fact a large number of cuts would be required to describe
34
Fig. 3.4
Dimensions of aircraft.
a shape accurately. Any point on the surface of a solid shape will have a waterline, a section line and a buttock line passing through it. Point P on the diagram is an example. The problem for the engineer specifying the shape is that the points either side of P, that is points WW, BB and SS, must all make a smooth or fair curve with P and at the same time agree with another set of waterlines, butts and sections of their own. The process of achieving this agreement is usually computerised but in the past some very elegant aircraft have been shaped by having their lines drawn out full size by the mould loft department. Each of the cuts illustrated is crossed by the other two cuts which show as lines, and if the various divisions are made at regular intervals the shapes will each be covered by a grid pattern. In Fig. 3.5 the section illustrated is crossed by water and buttock lines which can then be dimensioned to dene the section outline. Sections are taken on station lines which are labelled as to their distance from the aircraft datum, so that on an EH101 helicopter for instance, station 2850 would be 2850 mm from the datum which is some distance ahead of the aircrafts nose. Waterlines are similarly labelled so that WL -505 would be 505 mm below (shown by the minus sign) the aircrafts horizontal datum. This datum is an arbitrarily positioned line; in the case of small piston-engined aircraft it is often on the centre of the propeller shaft, and on large air-
35
Fig. 3.5
Lines drawing.
36
Fig. 3.6
Alternative methods of describing wing geometry.
37
craft with circular section fuselages it is usually a line joining the centres of the circles in the main part of the fuselage; but note here that at nose and tail the centre of the fuselage sections would be below or above the horizontal datum. Buttock lines are labelled by their distance from the fuselage centre line, so that RBL 300 would be a butt taken at 300 mm to the right (starboard side) of the aircraft centre line. In general, fuselage frames (see Chapter 2) lie on stations, but there are no other fuselage structure parts that exactly follow the major loft lines. Wing ribs are correctly dened by butt lines, but unfortunately some companies confuse the situation by labelling wing rib positions as stations. Alternative methods of describing wing geometry are illustrated in Fig. 3.6; no one is more correct than any other but the differences are often confusing to the student trying to memorise and understand a mass of new words.

The major parts of an aircraft structure are:
The wing, or mainplane, which generates the lift required to keep the
aircraft up in the air. The wing usually has moveable ailerons let into the outboard (i.e. furthest from the fuselage) trailing edge to provide roll control. It may also have aps on the inboard trailing edge to help the aircraft y more slowly when landing. The fuselage, or body, which tends to link all of the other main parts and also provides space for the crew and payload. Empennage, or tail feathers, usually consisting of a vertical xed n and moveable rudder and a horizontal xed tailplane and moveable elevator.
Any part of an aircraft can be pin-pointed in space relative to a predetermined datum (or reference) point. This is achieved by dening its station (its distance aft, or forward, of the datum), its waterline (its distance above or below the datum) and its buttock line (its distance left or right of the datum).
Chapter 4
Loads on the Aircraft
Before considering the loads imposed on the structure of the aircraft, it will be as well if we remind ourselves of some of the engineering concepts and the associated terminology.
4.1 General ight forces

The method by which an aircraft ies is well known: air passing over the wing lifts by suction, which is a pressure reduction. This pressure reduction over the top of the wing is assisted by an upward pressure under the wing. However, although the aircraft ies by passing through the air, it is convenient for the engineer to consider that the aircraft is standing still and that the air is moving past it. These ideas are illustrated in Fig. 4.1. Figure 4.1(b) shows a piece of paper lifting when air is blown over it. Figure 4.1(c) is a diagram of a wing section with airow represented by lines. Note the direction of the airow directly behind the wing. Figure 4.1(d) shows diagramatically a slightly different concept of lift on a wing which is more convenient for the structures engineer. The lines are not now meant to represent airow, they represent direction of air pressure onto the wing. The line of pressure or force behind the wing is in the same direction as the deected airow in Fig. 4.1(c). Because the airow has been deected from its straight path, a force must have been applied to make it change direction. In the diagram this force is represented by the line R. From the principles shown in Fig. 4.1(d) we can divide the applied force R into two parts, a vertical part which we call W and a horizontal part which we call T. W is provided for us by the weight of the aircraft and T by the thrust of the engine. To further clarify the situation we dispense with the original lines of air forces and replace them with forces more directly opposing W and T. These forces we call lift and drag and we label them L and D. Lift opposed by weight, and drag opposed by thrust, represent the balanced forces which are established by the change in direction of the airow.
4.1.1 Resolution of forces

This concept of dividing forces into their component parts, as we divided R into the components W and T, is most important and basic for students
39
Fig. 4.1
Forces created by airow over surfaces.
who wish to go on to a serious understanding of the methods of engineering associated with the design of aircraft structures. Figure 4.2 illustrates some of these ways of dividing and joining forces; people not familiar with the principles should study this with some care. Before proceeding any further, we should also remind ourselves of some of the principles associated with turning moments, balance and the transfer of loads.
40
Fig. 4.2
Triangle of forces.
Start by looking at (a) and (b) in Fig. 4.3. Each of these diagrams is equally ridiculous. To stop the levers moving we would have to do something at the pivots. Joining the two bars together as shown in (c) makes the situation more possible; the lever does not swing, it is balanced, but
41
Fig. 4.3
Moments.
42
the picture is still not complete. It is only completed by a vertical reaction at the pivot, as shown in (d). Four units of upwards acting force are essential to oppose the four units acting downwards. In fact, the downward forces cannot exist without the opposition (or reaction). If we now redraw (a) in the way shown in (e) we can make it complete by adding a moment and a reaction: The moment (M ) = load distance =32 =6 The reaction = the load = 3 (but note the direction) In Fig. 4.3(c) the lever is balanced because the moments each side of the pivot are equal, and the reaction of four units is composed of the three units in Fig. 4.3 plus one unit from the lever to the left of the pivot. Looking at the pivot in (e) on its own, as drawn in (f), the system of forces which is acting at the pivot is as shown. The applied load and moment are the white arrows and the reacting force and moment are the black arrows. (Note here that between (e) and (f) the system has been labelled in lbf and lbf in. These read as pounds force and pounds force inches, or just pound inches. Students more used to other units may wish to relabel the diagrams so that lbf becomes N (newton) and lbf in becomes Nm (newton metre).) Note also that the two white arrows are the 3 lbf load which is away at the right-hand end of the lever, and this illustrates the important principle that a load can be transferred (relocated) but in its new position it must be accompanied by a moment. To further illustrate this last point, consider the beam in the two diagrams (g) and (h). The reactions at the wall have been omitted to simplify the drawing. Elements of the beam to the left of the plane of section XX will feel exactly the same load in each case. An engineer examining the strength of the beam at section XX would call 28 lbf in the bending moment (MXX) and 7 lbf the shear (SXX). The last diagram (i) illustrates one possible way of applying a moment at the end of a lever to balance a load. This method is used later in this chapter, see Fig. 4.4.
4.1.2 Balancing forces acting on the aircraft

Returning to our consideration of how the aircraft ies, we have a wing with lift and drag which we attach to a body as in Fig. 4.4. We said earlier that the lift balances the weight of the aircraft but it is not sensible to assume that these two forces will be directly opposing one another; without any doubt they will at times be offset from one another as shown in the diagram. This immediately produces an out of balance moment
43
Fig. 4.4
Flying loads.
which we attend to by the load on the tailplane. By the same argument, thrust and drag are not immediately opposite one another and we also balance those out by the tailplane. All the forces on the aircraft that we have considered so far come together and balance out in Fig. 4.4(c). This
44
Fig. 4.5
Distribution of lift.
concept of the whole aeroplane as a free stationary body surrounded by a system of balancing loads and reactions is a concept which the aircraft engineer also applies to every small part of the structure as he comes to examine it in detail. Now consider the wing as viewed from the front; it will have lift along its whole length. For reasons of airow, which we, as structural engineers and not aerodynamicists, need only accept without question, there is more lift towards the middle of the wing than at the tips, and the lift is in fact distributed as shown in Fig. 4.5. Of course this lift cannot exist on its own, it must be balanced by something, and it is balanced by the weight of the aircraft, labelled W in the diagram.
4.2 Acceleration loads

So far we have considered the aeroplane as ying straight and level. Obviously this is not always so, and since we have already said that we wish to consider it as a free body at rest under the inuence of the system of balanced loads and reactions, we need a system for recognising the forces which accompany a manoeuvre, which is the name for any disturbance of straight and level ight. Firstly, we must be convinced that a manoeuvre always involves acceleration. Imagine, for instance, an aircraft ying due north which then turns 90 and heads due east. Before the turn the aircraft had no speed towards the east but after making the turn it has some speed towards the east, therefore between the two conditions it must have received an acceleration. To achieve this acceleration a force was applied to the aircraft which
45
had to be resisted by the structure. A phenomenon which is familiar to all of us illustrates a method of introducing these accelerating forces into the free stationary body idea. Imagine being in a passenger lift or elevator; just by eye we cannot tell whether or not the elevator is moving. In fact, if it is moving at a steady speed, it is difcult for us to tell from the load in our legs whether the elevator is going up or down. It is only when the elevator stops or starts that the positive or negative acceleration appears to change our weight and and the load in our legs is altered. If the elevator is accelerating upwards the load in our legs is greater because the inertia forces in our body act in a downwards direction. (Inertia is the resistance of a body to any change in its motion. If the body is standing still, it requires some external effort to move it; if it is moving, it will continue to move in a straight line and will resist the force needed to stop it or change its direction.) This is just another instance of the action and opposite reaction rule. In this case, acceleration or accelerating force acts one way and inertia the opposite way. The analogy with the elevator is easy to understand when acceleration is up and down, and because we cannot see outside the elevator it is easy to accept the free stationary body idea. When the acceleration is horizontal, nding an analogy is more difcult. While travelling at a steady speed on a smooth road in a motor car, it may be difcult to admit that the loads acting on our body are exactly the same as if the car was stationary, but they are. The loads only change during acceleration away from a stop (positive or forward acceleration); during deceleration by braking (which is negative acceleration or the same as rearward acceleration); or during sideways acceleration when turning a corner. Again the acceleration one way, inertia the other way rule applies, so that, if the car turns to the left, some force (the action of the front-wheel tyres) has acted to the left and accelerated the car the same way, but the passenger has to restrain himself from being thrown to the right. The essential point in the above discussion is that when the elevator or the car either stops or starts or changes direction (that is, manoeuvres), the passengers weight appears to increase in the opposite direction to the force directing the acceleration. This is always so. Do not be confused by the apparent contradiction of this truth when the passenger lift starts to descend. The situation then is that the acceleration in the downwards direction is small and although the passengers weight does increase in the upwards direction the increase is not enough to completely overcome his normal weight due to gravity. If the acceleration downwards became great enough, the occupant of the lift would need to push upwards on the roof to keep his feet on the oor and the greater the acceleration the harder the push would have to be. In an aircraft the same situations apply, and structural engineers deal with them by saying that during a manoeuvre the apparent weight of everything in the aircraft is increased by a factor (n) of the weight due to gravity (g). In some cases the factor n is determined by calculation but usually a gure is specied by government
46
Fig. 4.6
Vn diagram (the manoeuvring envelope).
legislation working through its own Airworthiness Authority. (See the note in Fig. 4.6 for the various names given to n.)
4.2.1 Using Vn diagrams

The presentation of the Airworthiness Requirement for the inertia factor or load factor is usually by the Vn diagram, which is in a form shown in Fig. 4.6. Working from the European Aviation Safety Agency (EASA) Certication Specication 25 paragraph 337 (CS-25.337) for aeroplanes with a maximum take-off weight of 5700 kg (12 500 lb) or greater, for an aircraft weighing 22 690 kg (50 000 lb) or more n is 2.5 and for an aircraft weighing 6350 kg (14 000 lb) n is 3.1. Figure 4.6 describes a manoeuvring envelope and particular combinations of speed and load within the envelope are called Cases. A similar Vn diagram, which describes a gust envelope, is used to present Airworthiness Requirements of the effect of air currents likely to be met by the aircraft. The air currents, or gusts, have an accelerating effect and increase apparent weights in exactly the same way as ordinary manoeuvres; they may also operate in any direction and in general they become more and more severe with higher speeds.
47
4.2.2 Emergency alighting loads

Another group of load factors specied by Airworthiness Authorities are the crash loads or emergency alighting loads. In the current issue of the EASA CS-25 for large aeroplanes these are required to be from nine (9) times g forward to 1.5 g rearwards, 6 g downwards to 3 g upwards and 3 g sideways. Helicopters and light aeroplanes have somewhat different sets of inertia requirements quoted in the appropriate certication specication. This method of specifying the loads along three axes mutually at right angles raises a further interesting point. Inertias are not always so conveniently applied and the section of the now obsolete British Civil Airworthiness Requirements which dealt with emergency alighting insisted that the prescribed loads should be taken as acting together, although the combined load need not exceed 9 g. Bearing in mind the random nature of emergency landings (such as wheels up in a ploughed eld) and the consequent random directions of the imposed loads, the BCAR principle does not seem too ridiculous. It is sometimes asked why a particular gure (such as 9 g) is set, the arguments being that the normal passenger travelling in an aircraft would be unable to withstand the effects of a 9 g deceleration and therefore the requirement should be set at a lower gure; or alternatively that by increasing the requirement to 12 g and ensuring a stronger structure, more passengers would survive a crash. These arguments might well be true, but a gure must be set somewhere and the Airworthiness Authorities consider carefully the statistics available and apply their expertise before xing a factor at a certain level. Like most aspects of aircraft design, there is an element of compromise involved. It would be easy to set a gure which produced a very strong, safe aeroplane that was too heavy to y, and the gure chosen is a balance of careful thought with the bias, no doubt, on the side of passenger safety. Some of the arguments used in choosing a gure are set out in Tyes Handbook of Aeronautics. In case this question of load factors is not absolutely clear, we will consider the specic example of a piece of equipment bolted down to the oor of an aircraft and subjected to the emergency alighting conditions mentioned above. If the piece of equipment is a rectangular box of 10 lb weight, to satisfy the requirement of 4.5 g downwards the oor will have to be strong enough to support 45 lbf (read forty-ve pounds force) pushing down, i.e. 10 4.5 g. (Note that 4.5 g is just a factor with no dimensions but it does alter a weight to a force.) Similarly, the 9 g forwards requirement means that the attachments must resist 90 lbf parallel to the oor. If we are working in SI units the arithmetic is slightly more complicated. A weight of 1 kg at an acceleration of 1 g produces 1 kgf (read one kilogramme force) but since the unit of force in SI units is the newton, we must convert. (The conversion factor is 1 kgf = 9.81 N.) So 7 kg at 2.25 g = 15.75 kgf therefore
48
15.75 kgf 9.81 = 154.5 N For a quick and rough calculation use a factor of 10 and say: 15.75 kgf = 157 N(which is correct to less than 2%) (Another useful conversion factor is 1 lbf = 4.45 N.)
4.3 Further aerodynamic loads

4.3.1 Lift and drag
In Fig. 4.4 the load positions shown are exaggerated to illustrate the idea of the aircraft in balance. If this book was written about aerodynamics we should be a lot more careful about lift and drag and how and where they act, but as we are concerned with structures a few more ideas on air loads will sufce. Figure 4.1(c) showed a wing section (or airfoil or aerofoil) at an angle a (read alpha) to an air stream. This angle is called the angle of attack and there is a similar angle called the angle of incidence which is measured to the horizontal datum of the aircraft (see Fig. 3.4). As the angle of attack is increased from its usual working or cruising speed attitude of about 2, the lift and drag will increase. At about 15 the airow will no longer follow round the upper surface, the wing will stall which means that the lift signicantly reduces, the point at which it acts moves aft and drag rises signicantly: in essence the wing stops usefully ying. Conversely, if the angle of attack is reduced, lift will decrease to zero at a small negative angle and then start to produce the negative lift which allows aerobatic aircraft to y upside down. While the lift and drag are changing, the point at which they effectively act, which is called the centre of pressure (or CP), is also changing. At the angle of zero lift it is, for some wings, right back near the trailing edge and it moves forward as the angle is increased, usually to about a third of the wing chord distance. To clarify this situation, the aerodynamicists quote the lift and drag as acting through one imaginary point for all angles of attack and add a varying moment to take care of the CP shift. This imaginary point is usually at 25% of the wing chord and is called the quarter chord point. Lift also varies with wing area, air density and aircraft speed (in fact as speed squared, i.e. twice the speed, four times the lift). It is also affected by aspect ratio. For an absolutely rectangular wing, aspect ratio (AR) = span chord, so a short stubby wing has a low aspect ratio and a long, slender wing, like a sailplane, has a high aspect ratio. Now, if we think about the airow past a short wing with the air pressure reducing as it passes over the wing, we can imagine this reduced pressure pulling some air in from the ends. The effect of this is to change the line of the airow from a direct path between leading and trailing edges to a longer path slightly along the wing. With
49
the leading edge a little higher than the trailing edge to give an apparent angle of attack, the longer route taken by the air gives a reduced real angle of attack and hence a lower lift. If wings could have an innite span the air would travel square across the wing for greater efciency; they cannot have innite span but the higher the aspect ratio the better they are aerodynamically. Unfortunately, the structure is affected in just the opposite way: a short, stubby wing is lighter for the same area than a long, slender wing; so, as frequently happens in aircraft design, compromises have to be made in order that a particular requirement can be met.
4.3.2 Control surfaces and their loads

We will now consider the basic control surfaces, the ailerons, elevators and rudder, which are shown in Fig. 3.3. Figure 4.7 could be any of these surfaces, and when moved to the position shown it deects the airow and produces lift, drag and moment as we saw earlier in this chapter. These forces are of interest to us because the structure has to resist them, or, more accurately, has to transmit these forces so that they move the mass of the aircraft. The force needed to rotate a large weighty aircraft is considerable and if we say Fig. 4.7 shows the elevator of a Boeing 747, then the loads at the hinges will obviously be many thousands of pounds force. If, on the other hand, we could imagine anything so unlikely as a Boeing 747 that was as light as a balloon, or if we put our 747 size elevator on a small, light aircraft, then the loads needed to rotate the aircraft would be very much smaller. In fact we could have (if our imagination were strong enough) the same size elevator, the same angle of movement, the same air speed but much smaller loads at the hinges. What we are saying is that to exist at all, a load needs to be resisted. Without the bulky, weighty inertia of a big aeroplane to push against, our hinge loads fade away. It is obvious that the structural loads imposed by control movements relate to
Fig. 4.7
Airow deected by a control surface.
50
the size and weight of the aircraft they are mounted on; it is less obvious that, for the same size aircraft, control loads are altered by the weight being lifted. Control loads also vary with the rate of response of the aircraft; for instance, it takes more aileron load to roll an aircraft quickly than to roll it slowly. Methods of assessing all these loads are found in the Airworthiness Requirements publications (see Section 13.5). Ailerons cause special problems: in the 1920s they effectively started a whole new science which became the study of aeroelasticity. If we say that Fig. 4.7 shows an aileron out on the tip of a long, exible wing, by deecting the control down we are hoping to increase the lift at that wing tip and cause the aircraft to roll. However, because the additional lift is so far back and the wing is exible, the whole cross-section in the diagram will be twisted anticlockwise until the stiffness of the structure stops it. If the rotation is sufcient, the lift on the wing, far from being increased by the control movement, will actually be decreased, causing the aircraft to roll the opposite way to that expected. David B. Thurston in his very excellent book, Design for Flying, says of this control reversal effect . . . [it] could provide quite a bit of activity in the cockpit . . .. We took an aileron down case to illustrate the effect but aileron up shows a possible, even worse effect where forced rotation of the wing section causes a local stall, whereupon the unloaded wing springs back until the airow restores itself and the process starts over again. This is one form of a phenomenon called utter which can be induced in all sorts of ways. The study of utter is complex but well documented and modern aircraft do not suffer the catastrophic failures created by utter in early aircraft before its various causes were recognised. The general name, aeroelasticity, is given to the study of these problems of control reversal and utter, together with control surface balancing, resonance of vibrations and any other effect which follows from the fact that aircraft structures are springy and actually change their shape under load. Work in this area is probably the most difcult that the aircraft stressman and designer have to deal with. As we said of utter, a lot of study has gone into the subject, especially mathematical analysis which has been made possible by the computer power now available. Unfortunately, there are no rule of thumb methods for the certain elimination of aeroelastic problems. The Airworthiness Authorities lay down requirements and test procedures which eliminate the possibility of danger to the airline passenger or the private pilot, but it is in the nature of aircraft design that performance and technology must be pressed forward right to their limits and that is where aeroelastic problems abound.
4.4 Other loads

So far we have concentrated entirely on ying loads, apart from a mention of crash loads, but there are others which affect the structure. The loads due to pressurisation are important, especially if the fuselage is not circu-
51
lar. A fuselage with a non-circular cross-section will try to become circular under internal pressure and impose bending loads on the fuselage frames. The double bubble style of fuselage is a clever exception, with the shape maintained by tension in the oor beams (the term oor joists is not normally used in aircraft structure). A tragic but instructive accident illustrated an interesting pressurisation structural problem. In this accident the lower fuselage below the oor was suddenly and totally depressurised by the opening of an unlocked cargo door during ight. The oor beams, which were parallel-sided, constantsection bending members then collapsed; in fact, they tried to adopt a shape which was part of a circle. No structures engineer would criticise another for not foreseeing this illustration of the power of pressurisation loads, which led, interestingly, to the immediate adoption of automatic pressure-equalising valves between compartments. Landing loads also need to be considered. Airworthiness Requirements give advice and assistance with the assessment of landing loads which can have a large inuence on the structure. For instance on a normally congured twin-engined aircraft in ight, the weight of the engines, the fuel, the main undercarriage and the wing itself, are all nicely spread out along the wing and directly supported by the lift. On the ground, and worse still during a heavy landing, all those same weights are propped up on the top of the undercarriage leg which is a more difcult situation for the wing structure to deal with. High-speed ying produces another type of structural load which needs a classication of its own because it exists in addition to the ight and manoeuvring loads. This loading is an effect of heat produced by highspeed ight and the distortion which the heat causes. As the aircraft forces its way through the air, parts of the structure, especially the surface skinning, are warmed and at very high speeds become quite hot. At Mach 2 (read mark two), that is at twice the speed of sound, some areas reach 150C. This type of heating, called kinetic heating because of its association with movement, produces different temperatures on different parts of the structure. The temperature variations cause expansions which are different from member to member, and thus the internal loads that we are discussing are produced: members trying to expand because of the heat are being forcibly restrained by cooler members. (Note that both hot and cold members are loaded.) Although this idea is not easy to understand without a more general appreciation of the form of structures, it is an important source of load in the aircraft and will be referred to again in Chapter 12.
4.5 Further load factors

Assessment of the loads on each piece of structure under investigation constitutes more than half of the stressmans job. The notes in this chapter have dealt with major loads, details of which would be supplied to the
52
structures department by the aerodynamics and wind tunnel departments (and also, during a development programme, by the ight test department). Once the major loads have been established the designers and stressmen have to break each load down and distribute it through the structure; for instance, they might nd that they cannot make two hinges each strong enough to carry half an elevator load and therefore they will need to use three hinges. In a later chapter we shall discuss proof and ultimate stress (see Section 5.5), which are, respectively, working strength and breaking strength for a material. The majority of loads derived from specications and Airworthiness Requirements are called limit loads or working loads. To compare these with the abilities of the material they must rst be multiplied by a proof factor to give a proof load. This proof factor is also specied in Airworthiness Requirements (it is usually 1.0 so that proof load is numerically the same as limit load). The proof load is a load that the structure is expected to cope with repeatedly and without distress. Crash case loads (or emergency alighting loads) are usually ultimate loads, that is the structure must be strong enough to withstand the load once, but it is then understood to be too stretched or distorted for further service. The connection between proof and ultimate loads applied to a structure, and the working and breaking strengths for the material of which the structure is made, is clear and will be made clearer after reading Chapter 5. There is a further connection between proof and ultimate called the ultimate factor. This is specied in the Airworthiness Requirements and for most civil aircraft applications is required to be 1.5. The way it works is shown in the following simple example. A stressman is investigating a bracket supporting a piece of equipment weighing 10 lb. He checks with the Airworthiness Requirements and nds that (for one Case):
the limited load factor (which is quoted as a manoeuvring factor) the proof factor and the ultimate factor therefore the proof load on the bracket is to be taken as and the ultimate load on the bracket is to be taken as = 2.5 g = 1.0 = 1.5 = 10 lb 2.5 g = 25 lbf = 25lbf 1.5 = 37.5 lb
Strictly speaking the stressman should now make two sets of calculations, one set comparing 25 lbf with the proof stress that is the allowable proof stress of the material of the bracket, and the other set comparing 37.5 lbf with the ultimate stress and allowable ultimate stress of the material. In practice, he would be investigating a number of components all made of the same metal and he would know the relationship between its proof and
53
ultimate stresses. If, for instance, he knew that the proof stress was greater than two thirds of the ultimate stress he would know that by basing his calculations on the ultimate loads and the ultimate stress he would also be covering the proof conditions. All these factors are confusing on rst acquaintance and there are more to come. Each makes its contribution to the designers task of using structural material efciently, which means using it economically with safety. Reserve factors are dened on page 191 but before leaving the topic of strength requirements mention must be made of the tting factor. This is a requirement of CS-25.635 and means that the nal tting which joins two members together must have an additional strength reserve above its normal proof and ultimate factors (see above). Fittings such as control surface hinge blocks, wing attachments, equipment mountings, etc., are required to have proof and ultimate reserve factors of not less than 1.15.
4.6 Loads acting on the whole aircraft

With the aid of diagrams we will consider the arrangement of loads on the whole aircraft in four cases, cruising straight and level, turning, stalling and landing. The loading situation in the straight and level cruise condition is as shown in Fig. 4.8. In Section 4.2 we discussed inertia factors (load factors) which are written as ng. In the cruise the inertia factor is 1 g and applies to the weights of all the elements of structure and equipment which constitute the whole aircraft. Sometimes the inertia factor is referred to as an acceleration factor or acceleration coefcient.There is nothing wrong with such
Fig. 4.8
Cruising ight.
54
Fig. 4.9
Loads in a turn.
a descriptive name except that in the straight and level condition we are left to explain the statement . . . the acceleration factor is 1 g, when clearly the aircraft is not being accelerated in any direction. Remember that whether it is called inertia, load or acceleration it is always a factor and as such is a device used by engineers to bring into calculation the apparent increase in weight of objects which are being accelerated. As a multiplying factor it represents an unaltered situation when it is 1.0. When an aircraft is turning it suffers an apparent weight increase due to the centrifugal force which is trying to make it go straight on instead of round the curved path of the turn. In a correctly executed manoeuvre the loads balance as shown in Fig. 4.9. Stalling occurs when the lift surfaces of the aircraft are no longer able to provide enough lift to balance the weight, and the speed of the aircraft at which this occurs is called the stalling speed. Even this is not simple, because we can have one stalling speed with a clean wing and another lower speed when aps, slats and any other of a selection of liftimproving devices are added. We also have the situation described above, where during a turn the apparent weight of the aircraft is increased, so that the ability of the wing to provide enough lift will disappear at a rather higher speed in a turn than when the aircraft is ying straight and level. The aerodynamic action of stalling is a breakdown of the airow over the top surface of the wing, as shown in Fig. 4.1 and, except for some highspeed situations, occurs at an angle of attack of about 15. The structural load situation is that immediately after the stalling of the wing the aircraft starts to pitch nose down and, more or less, free fall; that
55
Fig. 4.10
Loads out of balance in a stall.
is all of the individual parts of the aircraft appear weightless. In Section 4.1.2 we spoke of the free body concept with all the loads in balance, so we might expect that when the lift disappears the opposing and balancing weight will also disappear. In fact the ideal situation is not quite realised as, although the wing has stalled, it is still producing a small amount of lift but at a point further aft than prior to the stall. The weight now signicantly exceeds the available lift and is also acting forward of this lift thus tending to pitch the aircraft nose downwards and, with the aid of the down force on the tailplane, producing an overall acceleration earthwards. In the stall case the unbalanced tailplane load causes a nose down pitching rotation which aerodynamically allows the aircraft speed to build up and stable ying conditions to be restored (see Fig. 4.10). Structurally the loads involved on the main structure in the stall are not usually of any signicance but the loads on control surfaces and aps may be important. Landing loads are a major consideration in the design of the structure and are involved with the designed cushioning characteristics of the undercarriage. The stiffer the springs of the undercarriage the more rapid the vertical deceleration of the landing aircraft and the higher the loads involved on the structure. The Airworthiness Requirements are very descriptive and demanding on this subject. Helicopters and, even more obviously, deck-landing aircraft generate high landing loads. If we look back at Figs. 4.74.9 we can see that the general loading on any major part of the aircraft structure is a bending or a twisting load or both. This general loading pattern is shown in Fig. 4.11 and will be discussed in the next chapter.

The principal forces acting upon an aircraft are lift, weight, thrust and drag. In straight and level ight lift equals weight and thrust equals drag. It can be useful to resolve a force into two other forces at 90 to each other. For example the force produced by a wing moving through the air is resolved into:
56
Fig. 4.11
Loads on the wing.
lift acting at 90 to the direction of travel; and drag acting parallel to the direction of travel.
A force acting at a point displaced from a support will produce a turning moment in addition to the direct force. Moment is force multiplied by the distance from the support (see Fig. 4.3). When an aircraft (or any other object for that matter) is in a steady state, i.e. not accelerating in any direction, although it might be moving at a steady speed, all forces and all moments acting on the aircraft are in balance. For an aircraft to change direction (manoeuvre) some of the loads acting upon the aircraft must be increased and this will cause the aircraft to accelerate. The increased loads are often quoted as a multiple of the normal force due to gravity, e.g. an aircraft might perform a 3 g turn where the loads acting upon the aircraft are three times those experienced in straight and level ight. Airworthiness codes usually specify what factors (multiples of nominal gravity) to apply when designing the structure.
4.8 References
Carpenter, C. Flightwise. Shrewsbury: Airlife Publishing. Thurston, D.B. Design for Flying. New York: McGraw-Hill. (See also references for Chapter 13.) Tye, W. Handbook of Aeronautics, No. 1, Structural Principles and Data, Part 1 Structural airworthiness, 4th edn. London: Pitman.
Chapter 5
The Form of Structures
5.1 Structure relative to aircraft design

In the last chapter we looked at the loads which the aircraft has to carry; in this chapter we consider how the structure is arranged to support the loads (or, more exactly, how the structure links the loads to the reactions). Perhaps the most signicant thing about a piece of aircraft structure is that it is a compromise. Some years ago an amusing cartoon was printed showing how different people in the aircraft business thought of an aeroplane. To the commercial operator it was a ying box for carrying people or freight; to the avionics engineer it was just a means of getting his radar aerials and navigation aids airborne; to the aerodynamicist it was a slim, shining and beautiful creation with no unsightly access panels or skin joints or rivet heads, not even room for passengers; to the structural engineer it was a few simple robust members reminiscent of a railway bridge or a tower crane. Like all good cartoons this was not totally ridiculous. An aeroplane is all of these things but not one thing alone, so compromises have to be made. The structures engineer is constantly compromising with his own requirements; for example, discarding methods or materials which would ideally satisfy one set of conditions but which would be unsuitable for another set, or correcting an apparently satisfactory design which proves to be impossible to manufacture. However, he mainly needs to adjust his ideas to satisfy both the aerodynamicist, who (broadly speaking) species the external shape of the aircraft, and the payload, that is the passengers, freight or armament, which, together with the fuel, determine the internal dimensions.
5.2 Historic form of structure

In the early days of ying the aircraft structure or airframe had almost only one requirement: it was the minimum which the designer (who was probably also the test pilot) thought adequate to keep the wings intact and support the control surfaces and the pilot (see Fig. 2.1). Later, the basic structure was covered with fairing which improved the streamline shape without contributing to the strength (see Fig. 2.3). Wings, of course,
58
Fig. 5.1
Structural boxes or tubes.
59
had from the beginning been covered with fabric or paper, but this covering did not form part of the structure (except that very locally it carried air loads to the ribs). The third phase, which is the method used almost universally on modern aircraft, incorporated the fairings into the loadcarrying structure or, looking at it another way, pushed the structure right out so that it lled the aerodynamic shape. This type of structure, which, as we said in Chapter 2, is called stressed skin or semi monocoque, is efcient in the engineering sense that it has material or components performing more than one function. In some aircraft the wing skins perform three functions: they transmit the air loads which provide lift, they make the tube or box which is the strength of the wing (see Section 5.6) and they are also the walls of the fuel tanks. Although it is very difcult to imagine a better type of structure for aircraft, it is always possible that one will be invented. There is even now a lower limit of size and weight of aircraft below which the stressed skin would become so thin that it would be unmanageable. Man-powered aircraft, for instance, are so light that the structure must belong to the second phase of structural development because the covering is too imsy to be made load carrying.
5.3 General form of structure

The aircraft structure in its various phases of development has always consisted of an assembly of large built-up tubes or boxes. These tubes are most clearly obvious in the stressed-skin type of structure, which from now will be considered as the conventional or normal type. Figure 5.1(a) illustrates the tubes in their simplest form, and Fig. 5.1(b) shows the renement which is the basis of the vast majority of aircraft, from single seaters to wide-bodied passenger airliners and supersonic bombers. The conventional method of construction is shown in Fig. 5.2(a) and (b), with an earlier type of construction illustrated in Fig. 5.3. This earlier construction does not so obviously form a tube but its construction allows it to carry the same types of load as the conventional structure. These are bending and torsion loads, and before proceeding further it would be as well to dene these and some of the other terms used by structures engineers.
5.4 The basic load systems in structures

There are three basic combinations of load and resistance to load:
shear tension compression

It is hard to nd aircraft examples of these load/load resistance systems where the load application can quite clearly be called one of the above
60
Fig. 5.2(a) Conventional stressed skin construction de Havilland Canada Dash 7. (Courtesy of The de Havilland Aircraft of Canada Ltd.)
without any element of another. The examples in Figs 5.4 and 5.5 (pages 62 and 63) are close to being pure. In Fig. 5.4(a) we can say that the bolt is in pure shear through faces AA and BB; that is, the bolt is resisting the pull by its reluctance to divide in the way shown in Fig. 5.4(b). In the example illustrated in Fig. 5.5 the bolt is in tension and will eventually divide as in Fig. 5.5(b). The act of division is called a failure. Compression is when the load application is the opposite of tension (see Fig. 5.6) and a piece of structure failing in compression is most difcult to show in a pure form because the start of a failure usually changes the form of loading. For instance, in Fig. 5.6(a) one side of the round block would probably collapse a little before the other and the load would no longer be pure. In Figs 5.4(b) and 5.6(b) load is indicated by two arrows. To nish the diagrams, we should add two more arrows to complete the whole load/load resistance system (see Fig. 5.7). It is most important to note that the arrows always come in pairs, each pair being in accordance with Newtons Third Law of Motion which states that there cannot be a load without a reaction. Aircraft engineers tend to be rather untidy in their use of the terms but in general they would call the black arrows loads and the white arrows reactions. Unfortunately for the student of structures, he will
61
Fig. 5.2(b) Conventional stressed skin construction British Aerospace 146. (Courtesy of British Aerospace.)
also nd the black arrows paired together as load and reaction. (The use of black and white arrows is not a convention in general use but is used just in these diagrams to illustrate the examples.) The next two most fundamental load/reaction systems are:
torsion bending
Torsion is very similar to shear but instead of the material being loaded as shown in Fig. 5.4(a) and eventually dividing as shown in Fig. 5.4(b), the torsion or twisting action is as shown in Fig. 5.8(a) (page 65) and the break in Fig. 5.8(b).
62
Fig. 5.3
Early type of construction.
Fig. 5.4
Shear.
Bending is the most commonly used load/reaction system in all types of structure. Fig. 5.9 (page 66) shows a piece of material being bent, but the way in which it eventually breaks (called the mode of failure) is likely to be complicated, as will be described in more detail in Section 5.6. below. In Figs 5.45.6 it cannot be emphasised too much that the simple pieces of structure shown are systems complete with load forces and reaction (or resistance) forces. Without a resistance there cannot be a load. For the
63
Fig. 5.5
Tension.
Fig. 5.6
Compression.
student of aircraft structures this rule is absolutely fundamental. Sometimes it is difcult to see the two opposing forces and to keep the rule in mind, but anyone wishing to make a serious analysis of a piece of structure (especially students becoming aircraft stress engineers) will nd that once they accept the truth of the rule their problems will be greatly eased.
5.5 The forms of stress in materials

Having talked about types of load systems, we will now dene some of the terms used to describe materials.
5.5.1 Stress and strain (in tension loads)

For the aircraft engineer the words stress and strain have particular meanings which are not exactly the same as those meanings which apply in normal conversational English. From The Concise Oxford Dictionary we
64
Fig. 5.7
(a) A tension member; (b) a shear member; (c) shear strain.
nd that strain means a pull or stretching force; an engineer would not use the word in this way. To the structures engineer strain is the change in dimensions of a body brought about by force. Again we have a rather careless usage because when engineers say strain they mean unit strain, which is a ratio, being the change in length associated with a particular load divided by the original length. For example, refer to Fig. 5.7(a): if the tie is loaded as shown it will stretch or extend a little. If we divide the extension by the original length we have the strain (or more exactly, the unit strain). This is clearer if we make a numerical example by giving the symbols in Fig. 5.7(a) some quantities:
65
Numerical Examples
(1) Imperial units Referring to Fig. 5.7(a) Length (l) = Load (P) = Extension (e) = Strain 5.0 in. 1000 lbf (pounds force) 0.013 in. e 0.013 in. (e ) = = l 5.0 in. = 0.0026
(2)
Metric units Referring to Fig. 5.7(a) Length (l) = Load (P) = Extension (e) = Strain (e ) = 125 mm 4500 N 0.326 mm e = 0.0026 l
Note: Strain is a ratio and has no dimensions. For e read epsilon. The stress in a member is a measure of the load compared with the amount of material available for carrying the load. Referring back to Fig. 5.5(a), it is clear that if the bolt was twice as big it could either carry double the load or carry the same load twice as easily. Now, twice as big does not mean twice the diameter, but twice the cross-sectional area, because two bolts in place of the original one would have the same effect. So, for the engineer, stress is load per unit area (= load/area). Again this is claried by putting some quantities into Fig. 5.7(a).
Fig. 5.8
Torsion.
66
Fig. 5.9
Bending.
Numerical Examples
(1) Imperical units Referring to Fig. 5.7(a) Width Thickness Load Area Stress (2) (w) = 1.0 in. (t) = 0.036 in. (P) = 1000 lbf (A) = w t = 0.036 in2 P (s ) = = 27 778 lbf in 2 A (read pounds force per square inch)
Metric units Referring to Fig. 5.7(a) Width Thickness Load Area Stress (w) = 25 mm (t) = 1 mm (P) = 4500 N (A) = 25 mm2 P (s ) = = 180 N mm 2 A (read newtons per square millimetre)
For s read sigma. Sometimes stress is given the symbol f . Note that we have not said what the material is but the stress is the same whether the member is made of steel or of aluminium or possibly of strong plastic, although in that case our calculations would lead us to expect that plastic would not be strong enough.
5.5.2 Shear stress and shear strain

In Fig. 5.7(b) we have a member in shear. The black arrows represent the load and the tinted area represents the material resisting the load, so again
67
stress (in this case shear stress) is load over area and is given the symbol t (read tau). The paragraphs above refer to tensile stress and strain. Compressive stress will (perhaps obviously) cause a strain which is a reduction of length. In most cases it can be accepted that the strain per unit of compressive load is very close to the equivalent tensile stress (but in the opposite direction). However, in general, compressive loads are not so easily dealt with as tensile loads. Shear strain is a slightly more difcult concept to embrace. Imagine the shearing faces in Fig. 5.7(b) being parted as in Fig. 5.7(c), then the shear strain is taken as the angle that the middle line has moved through. Unfortunately, the diagram looks very unconvincing and we know from experience that the material would not distort in such a tidy way. However, if the gap is made smaller and smaller the pattern of the distortion becomes more convincing without altering the angle taken by the median line. The symbol used for shear strain is g (read gamma). All the above notes are simple denitions and they form the basics from which the whole analytical study of aircraft structures (the process called stressing) is built up. The next group of ideas is slightly more complicated and included for the consideration of students who may want to go on to a serious study of structures analysis.
5.5.3 Modulus of Elasticity (Hookes Law)

We have said that if a force or load is applied to a structural member, the member will change its length. It was discovered many years ago that if the force was doubled, the change of length was doubled. In fact, the change of length is directly proportional to the change of force (up to a certain force value). This relationship is known as Hookes Law. If we convert the force to stress and the change of length to strain, the relationship becomes even more important. Looking back at the examples on pages 65 and 66, we worked out the stress in the member at a certain load, and the strain at the same load. If we apply half the load, we halve the stress and, according to Hookes Law, we halve the strain. We can plot this as a simple graph as in Fig. 5.10 and it becomes obvious that, arithmetically, stress divided by strain is a constant. In Section 5.5.1 we made the very important point that the stress in a member was not related to the material of that member. The same is not true of strain. Different materials extend by different amounts when subject to the same stress, as shown in Fig. 5.11. By relating stress and strain in the way suggested by Hookes Law we are given an important identifying characteristic of the material. This characteristic, or property of the material, is called Youngs Modulus or the Modulus of Elasticity and using the gures from pages 65 and 66 we can nd the modulus of the material used in the examples:
68
(1)
Imperial units Modulus of Elasticity (E ) = s e 27 778 lbf in 2 = 0.0026 = 10 683 846 lbf in 2 = 10.7 10 6 lbf in 2 (approx.)
(2)
Metric units Modulus of Elasticity (E ) = s e 180 N mm 2 = 0.0026 = 69 231 N mm 2 = 69 kN mm 2 (approx.)
In Chapter 6 there is a list of properties of various materials and from that we can see that the material used in our examples above is an aluminium alloy. These notes apply to both tension and compression and the Modulus of Elasticity for most aircraft structural materials is the same for both. Some materials, such as bre-reinforced plastics, may have different moduli relative to different directions of load application. Such materials are said to be not homogenous.
Fig. 5.10
Stress/strain relationship at small loads.
69
Fig. 5.11
Stress/strain relationships for two different materials.
5.5.4 Modulus of Rigidity

Shear stress and shear strain have a similar relationship but instead of referring to a modulus of elasticity in shear we call this the Modulus of Rigidity or the Shear Modulus (symbol G).
5.5.5 Application of Hookes Law

Although Hookes Law applies to virtually all structural materials it is only true up to certain stress levels. The diagrams in Fig. 5.12(a) and (b) show the type of curves obtained from tests on samples of (a) steel and (b) aluminium alloy. These tests are made in a materials laboratory or test house and are typical of those made to determine whether or not a sample of material is as strong as its specication says it should be. The main curves in the diagram have been stylised and distorted a little to illustrate their characteristics, but anyone looking at actual recordings of tensile tests will be able to pick out the points which are mentioned below. The rst point to notice on both diagrams is that Hookes Law applies from 0 to point P, called the limit of proportionality. After point P, as more stress is applied, the strain becomes disproportionately greater, so the curve starts to bend away from the straight line. In the case of curve (b), which is characteristic of aluminium alloy and most other structural materials apart from the softer or mild steels, the extension (strain) becomes greater and greater without much gain in applied load (stress) until the material breaks. The curve for steel is similar but has an irregularity at Y called the yield point, where the material suddenly stretches before starting to work properly again, which is interesting but of no great importance to the aircraft engineer, simply because it only applies to mild steel. However, it gives its name, especially in American technical books, to the start of the disproportionate increase of strain.
70
Fig. 5.12
Stress/strain relationships for (a) mild steel and (b) aluminium alloy.
In the region before the limit of proportionality (which is also called the elastic limit) if the material is off loaded the stress/strain curve returns to zero down the same line. The material is in fact elastic. When the stress has exceeded point P (Fig. 5.12(a) and (b)) the material is said to be in a plastic state (nothing to do with plastic materials) and if the load is then taken off (say from point p in Fig. 5.12(b)) the curve will not return to zero but to a new point S. At S there is no stress but there is a retained extension which is called a permanent set. This is an interesting and important feature for aluminium alloys because if we now reload the material we nd that the elastic limit is now at a rather higher stress level. In fact, the material has improved some of its characteristics slightly and has become work hardened (see also Section 6.2.2). There is a further signicance about point p on the curve. Structural designers in other branches of engineering usually make sure that the
71
stress levels in their structures are kept well below the elastic limit but the aircraft designer, who is ghting a constant battle to achieve the lowest weight of structure, does not want to waste any usable load-carrying capability of the material. So, to dene an upper working limit of stress value which he permits himself to use, he selects a proof stress (p) dened by the dimension 0S when Sp is parallel to 0P. When 0S is 0.001 (remember e is a dimensionless ratio) the stress at point p is called the 0.1% proof stress and written either as p1 or s1, or when 0S is 0.002 the associated stress is the 0.2% proof stress (p2 or s2).
5.5.6 Other properties of materials

Before nishing this section there are some other terms connected with materials which will eventually be of interest to anyone studying structures in greater detail. Ultimate tensile stress (UTS) is the stress level associated with breakage of the material and is usually quoted as a specication requirement. That is, a selected material quoted as having UTS of (say) 54 000 lbf/in.2 or 370 N/mm2 can be relied upon to be producing at least that stress when it breaks. Bearing stress is usually associated with bolted (or riveted) joints and is a value for the material being joined (not the bolt) which is obtained by dividing the load by the product of the diameter of the bolt hole and the thickness of the material. Permissible bearing stress or 0.1% bearing stress is quoted as a property in material specications. Some further comment will be found in Chapter 9. In Section 5.5.1 we pointed out that when a force is applied to a piece of material of a certain length, that length changes in the direction of the applied force. For metallic materials, and most others as well, there is also a change of dimension at right angles to the direction of the force. This effect is quite logical. Given that the volume of material remains the same, if we stretch it in one direction we would expect a contraction in the other two directions to compensate. Figure 5.7(a) shows this obvious effect. What is less obvious is that the ratio between the strain in one direction and the strain at right angles to it is a property of the material, and is known as Poissons ratio and given the symbol u (read nu). Hardness of a material is measured by indenting the surface with a ball (like a ball bearing) or a point, usually a very small square pyramid, and measuring the size of the indentation. There are several calibrating systems; Brinell, Vickers and Rockwell are used universally for conducting many thousands of tests each year. These hardness numbers are most useful for a quick mental comparison against known materials and there is a correlation with tensile strength (see Table 6.1). Brittleness of a material is measured by the Izod test. Of all the normal mechanical tests which are applied to materials this is the one most susceptible to small variations in the test piece. It gives an indication of notch sensitivity but otherwise its usefulness is minimal.
72
5.6 Bending and torsion

Now that we have listed some of the terms and expressions used we will look in more detail at the bending and torsion mentioned in Sections 5.3 and 5.4.
5.6.1 Structures subjected to bending

Figure 5.13(a) shows pictorially a simple bar loaded in the middle and Figure 5.13(b) shows the same bar diagramatically.The most obvious point is that the bar is being bent with the top face of the bar in compression and the bottom face in tension. It is less obvious that the bar is also being sheared but consider Fig. 5.13(c) and (d): it is clear that if the loaded bar was suddenly cut in either of the ways shown, there would be a movement
Fig. 5.13
Bending and shear due to bending.
73
across the cut face similar to the movement shown in Fig. 5.4(b), therefore the material of the bar must be capable of resisting that movement. In the case of a plain rectangular bar there is not much doubt that the necessary resistance is there, but in some other beams, particularly those of typical aircraft construction which are built up from sheet metal and have riveted joints, it is only too easy for the designer to lose sight of these shears. The beam in Fig. 5.14(a) is at rst glance a reasonable construction with the load W being resisted by the push and pull at the lugs A and B. However, because the shear has been ignored, the vertical web of the beam will collapse as shown in Fig. 5.14(b). To make the beam completely sound, an attachment at the end of the web is required. This would have been more obvious if the load W had been properly transferred according to the principle stated in Fig. 4.3, in which case Fig. 5.14(a) would have become like Fig. 5.14(c). Parts (d) and (e) of the diagram illustrate alternative solutions to the problem of coping with the shear in the web. Some other points should be noted about this piece of structure, which, as we said before, is typical of aircraft structural components. It is more complex than it appears at rst glance. Firstly, the rivets through the angles attaching the web to the anges are loaded in shear. We can imagine that if they were made of some very soft plastic instead of metal they would be cut through. Secondly, although in Fig. 5.14(d) we have divided the web into smaller sections, each section can still buckle into waves, as shown in the long unsupported web of (b). In his battle against weight, the aircraft structures engineer accepts this situation, knowing that even after the thin sheet has fallen into buckles (the post-buckling situation) the web will still carry load. This is a very important concept which was analysed by Professor Wagner in about 1930 and followed by S. Timoshenko in 1936, whose book The Theory of Elastic Stability, has been the inspiration of most of the mathematical work on analysis published since. Unfortunately, our second point above has an inuence on the rst point. The load on the ange rivets increases due to the action of the buckled web, and this fact gives a hint of the thoroughness with which structures of this type have to be investigated if they are to be safe and efcient. Consider now the effect of making the beam deeper. Refer again to Fig. 5.14(c). We will discuss the upper, tension-side ange, although the stages of the argument will apply equally to the compression side, but in the opposite sense. (a) As the beam is loaded it bends (or deects) and we can see that the ange length is increased because the length along the curve of the tension side is greater than the length at the centre of the beam (which has not altered). (b) The ange length is increased because it is loaded, therefore there is a stress (load/area), and a strain (increase in length/original length).
74
Fig. 5.14
Sheet metal beams.
75
(c)
If the deection of the beam is kept the same (i.e. the curve of the centre is kept the same) but the depth of the beam is increased, the strain in the ange will increase and, therefore, the stress will increase as we saw in Section 5.5. (d) Alternatively, if we increase the depth of the beam but keep the same strain in the ange, then the deection of the beam will have to be less in compensation. In fact, if we double the depth of beam the deection is halved. (e) Also, if we increase the depth of the beam, the load in the ange is reduced (see Fig. 4.2). Again, if we double the depth we halve the load. (f) We know that the ange load is the area of cross-section of ange multiplied by the stress, and we have already said in (d) that we are keeping the stress constant (actually, we said we would keep the same strain but that of course means that the stress will be the same), therefore, with the same stress but a smaller load, the area and hence the weight become less. (g) Summing up this important argument, we can say that by doubling the beam depth, we halve the deection and halve the weight (of the anges). It is usually said that the advantage of adding to the depth of a beam increases as the square of the depth, which is a rule of thumb which should be treated with some caution, but if the arguments above are clear and understood the effect can be considered in logical steps. We will make one more point before leaving the beam in Fig. 5.14. It is not a very important point because, although the type of construction is typical of aircraft structure, there are not many beams of this type which are entirely isolated; they are more usually part of a bigger structure. However, the principles involved are quite important. If the ange width is large compared with its thickness, the edges of the ange on the tension side will try to take a short cut across the curve of deection, as shown in Fig. 5.14(f). This means that the strain on the edges of the ange is less than the strain at the point where the ange is attached to the web. As the strain is less we know that the stress is less and therefore the load in the whole ange is not exactly the stress multiplied by the area, or more importantly the maximum stress is rather higher than the average found by dividing the load by the area. The phenomenon is called shear lag, although that term is also used in connection with the transfer of concentrated loads into aircraft skin. On the compression side of the beam, the situation is worse because the middle of the ange is being forced to reduce its length and the outer edges are unwilling to follow, so they tend to go into waves or buckles, which in turn push their way towards the middle of the ange and reduce its effectiveness. In fact, if we regard the tension and compression anges as separate members in their own right, we can say that while it is fairly easy to design a tension member, it is more difcult to design a compression member.
76
Fig. 5.15
Distortion due to twisting.
5.6.2 The effects of torsion

In Fig. 5.8 we showed a solid bar being twisted or subjected to torsion. Looking at a thin slice of the bar (Fig. 5.15) we can see that the metal distortion is greatest at the outside reducing to nothing at the middle. We already know that the distortion (the strain) is directly related to the stress in the material (Section 5.5.3), so that if we retain the hard-working metal and dispense with the under-worked metal by turning the bar into a tube we will save weight without losing much strength. Imagine now the situation if the tube is of a large diameter and the skin is very thin. If the tube is twisted (subjected to torsion) it will carry a small load but then fall into large buckles. Those who have the strength can prove this for themselves by twisting an empty soft-drink can. If we now divide the surface area with relatively stiff members, as shown in Fig. 5.16(a) or (b), the buckles are more contained and the load which can be carried is greatly increased. Combining Fig. 5.16(a) and (b) would give an illustration looking like a typical piece of aircraft structure. If we now look at just a strip of tube (which could be cut either along the tube, as shown in Fig. 5.16(c), or round the tube) then we have a piece of structure which is remarkably like the beam in Fig. 5.14(d), and which will be just as capable of carrying loads after buckling as the beam. Before the skin starts to buckle it is carrying the load in shear. In Fig. 5.13 we saw that bending a bar produced a shear load in the material. Conversely, in the current case, the shear due to torsion induces tension and compression loads along stiffening members, both the stringers and the frames. This is a difcult statement to justify but imagine the thin skin of the tube being jointed, edge to edge, with lines of rivets attaching each raw edge to a stiffening ring. If the large tube was now given a twisting load and then the two lines of rivets were instantaneously removed the effect would be very clear; it may be more difcult to picture a movement when the joint is along a stringer but the fact that a movement would take place illustrates that shear loads are present in the skin. This shear due to torsion induces, via rivets, tension and compression end loads in the stringers. Normally these loads balance themselves out because they act in opposite directions on either side of the
77
Fig. 5.16
Thin wall tubes with stiffeners.
stiffener member. However, they are important where the tubular member is cut out for cabin windows, access panels, aerials, etc.
5.6.3 More complicated beam analysis

In Sections 5.6.1 and 5.6.2 on bending and torsion we made particular reference to the buckling of thin skins or beam webs and especially to their ability to carry post-buckling load. Modern aircraft generally need to maintain a smooth outer skin and it would seem strange now to see wings with rippled skins which would upset the aerodynamics, but most early stressed-skin aircraft had wing skins which showed buckles when under load. In spite of the feeling that structures which are designed to buckle are slightly old fashioned, the concept is still important for two reasons. First, in some situations, structures exploiting post-buckling strength are still designed so that the buckling commences at, or just above, proof load (Section 4.5). Second, over the years since Professor Wagners work on the subject, a great deal of mathematical analysis has been recorded on the subject of buckling and much of this has been presented in the form of data sheets. For the student who wishes to become a stressman, some study of the buckling of thin sheets will give him a good insight into the mathematical methods used in structures analysis. It will also give him practice in the use of the data sheets. Data concerned with buckling are as complicated as any and by practising with them he will become familiar with the use of structures data sources in general. Most major aircraft
78
Fig. 5.17
Loads on a box structure.
companies have their own data sheets but for people in smaller organisations the ESDU (Engineering Sciences Data Unit) publish a very complete range.
5.6.4 Loads on box structures

In Section 5.3 we said that aircraft structures were basically boxes. If we discuss what happens to boxes when they are subjected to the types of load which they receive when they are part of an aircraft, that is, the loads we discussed in Chapter 4, we shall build up an understanding of why aircraft are built in the way that they are. The diagram Fig. 5.17 is similar to Fig. 4.11 and shows a representative piece of aircraft. We can call it a wing as we did in Fig. 4.11 (V is lift, H is drag and T is pitching moment), or we can call it a fuselage with the end ABCD built into the robust frames which carry the wings (V is then the tailplane load, H is the load on the n and T the torque due to load H having been transferred down from its true position above the level of PQ). To help us order our thoughts we are going to treat each of the three loads (V, H and T) separately; that is, we will discuss what happens when V is applied, when H and T are each applied, and then add together the separate effects. This method of dealing with complicated load systems is called the Principle of Superposition and can be applied to almost any piece of structure. In general, when a structure is loaded it bends or it stretches or in some way it deects. The load and deection, which are there for everyone to see, go hand in hand with stress and strain, which are the engineers concern and interest. The Principle of Superposition says that if two (or more) separate loads are applied to an elastic structure, each load will produce a deection and if the loads are applied together then the deections will be added together. (If the deections are in opposite directions, of course, one will be subtracted from the other.) When the loads and deections can be added then the stress and the strain produced by the loads and deections can also be added (or
79
subtracted if they are in opposite directions). This principle, which is used constantly by stressmen, can go awry if any part of the structure exceeds its elastic limit, nor does it work if the deection produced by one load upsets the ability of the structure to carry another load (see Section 9.2), but generally it is a useful tool which allows complex problems to be broken down into a series of simple problems. So, looking back to Fig. 5.17 and considering load V on its own, the rst obvious thing to notice is that if V was applied at a point on line RS, as shown in the drawing, the thin sheet would collapse. To prevent this happening we put in a at member PQRS. If the structure is a wing the member would be a rib, and if the structure is a fuselage the member would be a frame; for ease in this example we will call the structure a wing. With V pushing upwards, and thinking of the whole box as a beam, face PABQ is in compression and SDCR is in tension. However, we know (Section 5.6.1 above) that the loads tend to concentrate towards the web parts of sheet-metal beams, so for efciency we should have more metal in the corners, i.e. along lines QB, etc. Since in practice we have to make joints in the corners, the need of extra metal there is not inconvenient. (The joints do more than simply hold the bits together; they carry the loads from one component to another either by lines of rivets or by bonded, i.e. glued, attachment.) So far then, our structure has developed into that shown in Fig 5.18. We have elected to make this example a wing, which is an advantage because if we say that plane ABCD is on the centre line of the aircraft there is a
Fig. 5.18
Partially designed box structure.
80
mirror image of our piece of structure the other side. This means that because the loads are symmetrical the points A, B, C and D are as rigidly held in position as they would be if they were built into a wall, and we can think of them as being so. Remembering what was said in Fig. 4.2 about transferring loads, consider the effect of transferring load V to position YY which is a distance x from QR. (Actually we should only be transferring half of V because the other half would be carried by the other side of the box.) At YY we still have load V and we also have a moment Vx. Wherever YY is between QR and BC the load V is always the same, but the moment Vx changes from zero at QR to a maximum at BC. So if V, which is the shear, is constant, the web QBCR can be the same thickness from QR to BC, but if the moment Vx, which is called the bending moment, increases then the material along the corners QB and BC must be more highly loaded to resist the extra moment. For efciency we want to keep a constant high stress in the material in the corners, so the area of material has to be lower where the load is lower, and the angles shown in Fig. 5.18(a) should taper as shown in Fig. 5.18(b). Summarising the situation so far we have:
shear in webs ADSP and BCRQ compression in corners PA and QB increasing towards A and B tension in corners SD and RC increasing towards D and C
If we now consider load H on its own we follow exactly the same stages of thought and conclude that we have:
shear in webs PABQ and SDCR compression in corners PA and SD increasing towards A and D tension in corners QB and RC increasing towards B and C
Dealing with the torque T introduces some ideas which have not previously appeared in these notes. Earlier we discussed torsion in tubes of circular section and in Section 5.6.2 we introduced the concept of stiffeners. The present tube is just as capable of resisting or carrying torsion as a circular tube. We already have jointing members in the corners QB, etc., although the corner, even if it was only folded, could provide an anti-buckling stabiliser of the type which we needed to add to the circular tube. However, the large faces PABQ and SDCR will need to be divided into smaller areas to avoid large buckles or to delay the onset of buckling, since small areas of thin sheet begin to buckle later and at a higher load than large areas of the same thickness sheet. (This is an extension of the arguments in Section 5.6.1, but is fairly obvious since for the same sheet the smaller the area the greater the inuence of the thickness. A 1-mm thick sheet 10 ft square would buckle more easily than a 1-mm thick sheet only 6 in. square.) Redrawing the wing again gives a structure as shown in Fig. 5.19(a), which is now recognisable as part of an aircraft, even though in practice it would be shaped as shown in Fig. 5.19(b). This diagram is in advance of
81
Fig. 5.19
Completed box structure.
the discussion because we have not nished with the torque load. An important aspect of applying torque to boxes of this shape is that they distort under load in such a way as to warp ribs such as PQRS out of plane (Fig. 5.20). A rib ABCD would not warp because the other half of the wing, the other side of ABCD, produces a symmetry of loading. If there was no compensating structure, or if for some reason the load on the other half wing was different (as during a rolling manoeuvre), then rib ABCD would tend to warp as well. There are at least two important considerations here. First, if the rib is strong for some reason not connected with the particular case we are considering (for instance the undercarriage may be bolted to it) then it may object to being warped. In ghting against the warping loads it may produce other stresses back in the torsion box. Second, because it warps and is no longer a plane, at, easy to analyse piece of structure, it may upset the principle of superposition. Having introduced some of the side effects of torsion we can now say that the shear loads in the skin are not very different from those in the circular thin-walled tube. In fact they vary according to the area of crosssection of the whole tube or box, i.e. the bigger the box the lower the shear load, but the length of box makes no difference to the shear load. (Note, however, that it will make a difference to the deection and the warping.) The shear due to torsion acts all around the box but the shear due to V and the shear due to H each only act on two faces. Looking back to Fig. 5.17 we can see that the shear due to T increases the shear in PQBA, which
82
Fig. 5.20
Warping of rib or former in a thin-walled box subjected to torsion.
Table 5.1 Type of end load, i.e. tension (+) or compression (-) due to V Member PA QB RC SD Skin PQBA QRCB SRCD SPAD + + 0 + 0 due to H + + Type of shear load + 0 0 due to T 0 0 0 0 + + + + Type of load in total Large compressive load Smaller load Large tensile load Smaller load High-shear load High-shear load Lower-shear load Lower-shear load
was due to H, but reduces the shear in SRCD, and there is a similar effect in the shears due to V. We can now add together all the effects of V, H and T for the case we have considered and the results are shown in Table 5.1.The shear due to load T we will call + (read positive) for this example. Tension in a member we will call + and compression will be - (read negative). Without putting physical dimensions into the example we cannot say whether the members (also called edge members or booms) QB and SD are in tension or compression. This depends on the relative sizes of the loads due to V and H. Before proceeding further we must sound some warnings. The example above is not stressing which is more mathematical and much more searching. Nor is it really representative because a wing would not, for instance, have all its lift (load V) concentrated at a single point. However, as an example it introduces some of the concepts associated with box structures in torsion and bending and gives a hint of the way a stressman might begin to approach the problems of analysis. His later stages would involve detailed examination of the skins and members to see if each of
83
these could carry the loads imposed on it; examination of the loads in the rivets or other fasteners in the joints; reallocation of loads if he has been obliged to strengthen one part and alter the strain on another. After all this, remember, he has only examined the structure under one loading case with all other cases still to come.
5.6.5 Effect on beams of asymmetric loading

In an effort to present an uncluttered broad picture many important details have been glossed over. We will now discuss one which is very important. In Fig. 5.17 we had the point of application of the loads centrally placed and it was easy to accept that the inuence of the loads was not biased towards one or other piece of structure. Three situations could have upset this state of affairs: either the structure could have been asymmetrical, or the point of application of the load could have been elsewhere, or we could have had both of these situations simultaneously. If the point of application of the load is not convenient we can shift it to any other position we like, provided that we add a moment (see Fig. 4.2). In fact in Fig. 5.17 we had a moment T which could have resulted from a movement of V. The question of asymmetry is more complex, and we are now expanding the discussion away from box structures and on to beams of any sort. We can see that if the structure in Fig. 5.17 had been stiffer on the PS side then the QR side (say that members PA and SD were a heavier section than members QB and RC), application of load V at the point shown would have resulted in point Q deecting upwards more than point P. That is, the rib PQRS would have rotated anticlockwise as well as deecting upwards. In fact, we are saying that load V would have twisted the box, or applied a torsion to the box, as well as applying a bending. To benet from using the Principle of Superposition for the analysis of beams we need to nd a position to which we can move V, so that it bends the beam without twisting it. (Remember we can move V where we like by adding a moment which will change T.) Such a position is called the exural centre or shear centre and is a property of the cross-section of beam under consideration. When the section of the beam is symmetrical about the plane of bending, the shear centre is on the line of symmetry. When the section is not symmetrical the position of the shear centre has to be calculated. In Fig. 5.21(a) we have arranged the plane of bending to pass through the shear centre. If the bending load had been applied elsewhere we would have shifted it but of course the particular section shown is not a good shape to resist the torsion which such a shift would have produced. Figure 5.21 shows some sections with a rough indication of their shear centre. The position of the shear centre of the channel section is of some interest. Because it is easily made and easily attached the channel section is a popular structural member particularly for mounting pieces of equipment and for other jobs which at rst glance appear to be quite simple.
84
Fig. 5.21
Shear centre and its position.
Unfortunately, loading the member in what appears to be a straightforward manner, such as that shown in Fig. 5.22, produces a twisting action. A situation of this sort would not arise with a box section or a tube because the shear centre would be within the section and such sections are good torsion carriers anyway. Remember, the channel is a very poor torsion carrier and should only be used with caution. So far all of this section has been concerned with boxes made of thin sheet metal and of course this is the most familiar type of structure used on modern aircraft. Variants are also in general use where the webs and
85
Fig. 5.22
Effect of load on a channel section beam.
booms (see Fig. 9.12) are machined out of very large thick plates of metal. This method goes a long way to avoiding potential breakage (failure) points where rivet holes are drilled or other discontinuities of section occur in a built up structure. Further comment on this is made in Chapter 9. However, this type of structure is basically the same as the stressed-skin structure, but historically there have been other structures which occasionally reappear in new designs. Looking back at Fig. 5.19, the element could have looked like part of Fig. 2.2, or if it had been part of a fuselage instead of a wing it could have been like Fig. 5.2. One of the most interesting boxes, shown in Fig. 2.1, was used on the rst successful yer and subsequently on many other aircraft. When looking at a piece of aircraft structure, however important or insignicant, remember that it almost certainly will have to carry bending and torsion, so it should be, in some form, a box or tube. If it is not a box and is perhaps only carrying some local load then look for some other method of stabilisation, especially if it is a channel section or an angle section.
5.7 Compression
In Section 5.6 we mentioned compression load effects in the anges of beams. To have made the example structure in Section 5.6.4 more representative of a general case we could have added a compressive (or end load) to the box, but the larger elements of aircraft structure do not
86
usually suffer much from loads along their main axis. Fuselages are in tension from cabin pressurisation loads and, for some aircraft, in compression from engine loads, but usually end loads are a problem for parts of the boxes rather than the whole. In our example we had boom PA suffering compression which would produce a buckling problem. In the present case the buckling and stability problem is slightly different. Previously we discussed buckling in panels due to shear and buckling in compression anges of beams. In both of those cases the buckles were contained and forced to fade out by stiff edge members or by the web in a beam, so that under load the structure would sense the sudden formation of a buckle and give a little, but then the load would be redistributed into other paths. Buckling of compression members is usually more nal. In the case of the boom PA, for instance, if it chose to buckle (or collapse, or cripple) inwards, that would be nal. (The classic mathematical analysis of compression members or struts was the work of Euler (read oiler) in about 1750. The basis of his argument is that a slim straight springy strut (the usual modern given example is a bicycle wheel spoke) will support a compressive end load even though the strut is slightly bowed. Many students nd difculty in accepting this basic premise because practical experience suggests that if they press down on a spoke it will support a small load and then suddenly buckle and give way. If we think of it another way and start with a small sideways bow in the strut while pressing lightly down on the top we will feel the strut pushing back as it tries to straighten up. The down load on the strut is the Euler buckling load. Under this load the strut is in equilibrium and any increase of load causes a crippling failure.) The strength versus weight problems associated with the design of struts were a major worry in early aircraft structures. In a stressed-skin type of structure if a part is in compression and the calculations suggest that it may be in danger of buckling under an extreme loading case, it is usually possible to add some stabilising auxiliary structure to avoid the problem. In the example shown in Fig. 5.17, if PA looked as though it might fail we would probably put an extra rib, of lighter construction than rib PQRS, half way along the box. This would stabilise PA considerably. Solutions of this type were not open to designers of biplane interplane struts. However, stressmen and designers do still come up against strut problems on modern aircraft in the mounting of equipment and in non-stressed-skin areas such as undercarriages. Another area where there is a compression/buckling/stability problem is in riveted joints at the edge of shear-carrying panels. As we noted in Section 5.6.1, the rivets attaching beam anges to beam webs are loaded in shear and the same applies to all the rivets around the edge of a skin panel. If each rivet is loaded then there must be end loads in each piece of metal between the rivets. This end load can produce inter-rivet buckling in the skin, an undesirable thing in itself but worse because the deformed skin can pry (i.e. force up by leverage) under the rivet heads which are not designed to carry that sort of load.
87
5.8 The whole structure

We have discussed the general form of aircraft structure and the way it works and to complete this section we should record a few denitions of phrases used in connection with structure design:
Primary structure is structure which, if it failed, would endanger the

aircraft.
Class 1 parts are those components, the failure of which could cause
structure collapse, loss of aircraft control, injury to occupants or interference with essential systems. Class 2 parts are those not included in Class 1 which require more than visual examintion to ensure reliability. Class 3 parts are the remainder. Fail-safe structure is so designed that after failure of one part there is still sufcient strength in the remainder to prevent catastrophic collapse until the damaged part is found by routine inspection.
The above denitions were paraphrased from the old BCAR which are now replaced by CS (Certication Specications) (see Chapter 11). The current description for an important piece of structure is Critical Part. However, the earlier phrases create a clear picture for aircraft structures engineers whose major concern is with main structure, primary structure, Class 1 parts and critical parts by whatever name. Fail-safe concepts are usually linked with fatigue failures and modern thinking favours easier inspection of components likely to suffer from fatigue damage, thereby reducing the need for alternative load paths. The whole problem of evaluating the degree of success in a fail-safe design is very difcult and well beyond the scope of this book.

Aircraft structures are almost invariably the result of a compromise between the requirements of, inter alia, the stress engineer, the aerodynamicist, the systems engineers, production department, the operators maintenance needs, etc. A piece of structure can be subject to three types of static stress, tension, compression and shear. Slightly less pure than these, we can also nd the structure subject to torsion and bending. When subject to a purely tensile load a member will experience tensile strain (e read epsilon). This can be dened as: the increase in length the original length The tensile load will also create a tensile stress (s, read sigma). This is dened as:
88
load area For a given material extension is proportional to load (within certain limits) which means, also, that strain is proportional to stress. The ratio of stress to strain is known as Youngs Modulus or the Modulus of Elasticity, E and; E= s e
An important property of typical aircraft aluminium alloys is that they can be cold worked. This happens when the material is stretched beyond its elastic limit (i.e. beyond the point where stress and strain are proportional) and this results in the material becoming work hardened, making it stronger. A beam in bending experiences tensile stress in its top surface, compressive stress in its bottom surface and shear stress through its thickness. The most efcient way of making a beam stronger in bending is to make it deeper. Closed tubes are the most efcient members for resisting torsion. When they fail it is usually due to buckling in shear and the onset of this buckling can be greatly delayed by dividing the tube up into smaller panels using frames and stringers. When trying to analyse a piece of structure subject to a complex set of loads it is helpful to use the Principle of Superposition. Using this very powerful principle it is possible to look at the effect of each load separately and then add the effects to obtain the whole picture. The classic compression member is the strut. All but the shortest struts will ultimately fail due to buckling rather than pure compressive stress. The classic work on buckling of struts was done by Euler.
5.10 References
Timoshenko, S., Theory of Elastic Stability. New York: McGraw-Hill. (See also references for Chapter 12.)
Chapter 6
Materials
6.1 Choice of materials

With the exception of reinforced concrete, in aircraft construction you can nd almost all the engineering materials used by man. The original aeroplanes were made mainly of wood with fabric covering and steel wire bracing. Today the main material used is aluminium alloy, which is pure aluminium mixed with other metals to improve its strength. In other parts of the aircraft, for instance in the avionics, valuable metals such as gold and platinum can be found. Traditionally plastics have been used in the furnishings, both in sheet form and as upholstery foam, but in recent years plastics have become more common in primary structures in the form of resin-reinforced bres. Even paper has been employed, for instance as the centre of sandwich structures used in the oor of wide-bodied aeroplanes. With so many choices open to him the aircraft engineer needs to be very selective when considering which material to use for the job that he has in hand. Selection of some materials is very obvious and scarcely needs recording; for instance, the gases coming from a jet engine are extremely hot and therefore it is not possible to use any material that is likely to be damaged by heat in any position where it can be touched by jet efux. It is also ridiculous to suggest that anything other than a transparent material could be used for windscreens or cabin windows, but nevertheless it is not particularly easy to nd a transparent material which will stand the loads and extremes of temperature to which such components are subjected. At other times, choice of materials may be determined or restricted by the manufacturing processes likely to be employed on them; for instance, not all steels are suitable for welding. As a process, welding is not much used in aircraft construction now, but in cases where it is, the particular alloys used must be selected with considerable care. In some cases aluminium alloys have to be softened before they can be bent or formed into shape. This softening process may be accompanied by a later heat treatment process to regain the former strength of the alloy. If this is the case it is often impossible to perform the heat treatment after the material has been bent or formed due to the effects of distortion, and it will then be necessary to choose another alloy.
90
Amongst the different types of materials available to the engineer, there is an enormous number of sub-variants. The properties and characteristics of these subvariants are carefully controlled, and the materials manufacturers produce their products to written specication which may be issued by the standards institutions of the various countries in which they are made, or by the aircraft industry itself. By far the most common material used in aeroplane structures is aluminium alloy, also called light alloy. A typical light alloy used in aircraft structure may be 4% copper, 0.5% silicon, 0.5% magnesium, a little manganese, iron and chromium and the remainder, more than 90%, will be aluminium itself. Mixing the aluminium with these very small quantities of other metals has an enormous effect on its strength; the strength of the structural alloys is six to eight times the strength of pure aluminium. When it was rst discovered, aluminium was very difcult to rene and for some years in the middle of the 19th century it was virtually a semi-precious metal. Even when production methods were improved and cheapened it was of very little use as a structural material because of its weakness; it was not until about the time of the First World War that the discovery of methods of producing alloys presented the opportunity for the design of the type of structures which have changed very little up to the present time. By far the most important design objective for the aircraft structural engineer is lightness, but of course it must be coupled with adequate strength. We discussed in the last chapter the meaning of the word strength. Strangely enough, in terms of tensile strength, there is very little benet to be gained from using light alloy: although it is only one third of the weight of steel, it also has only one third of the strength. However, for other types of structural members, such as bending members or struts, it is useful to have a bulk of material and it is then that aluminium alloy comes into its own. This need for bulk is very well illustrated by the skin sheeting used on fuselages and wings. The thickness of the skin on quite a large aeroplane may be as little as 0.030 to 0.040 of an inch (1 mm). Much the same strength of skin could be achieved with steel a third of the thickness without any loss of weight, but clearly skins which are only 0.010 of an inch (0.3 mm) thick would not be a very practical proposition. Aluminium alloy starts life as a cast billet. There are many uses for cast aluminium but in general it is not good structural material. Cast aluminium is widely used on reciprocating engines (crank cases and pistons) but it is not a ductile material. This lack of ductility, which is another way of saying it is rather brittle, makes it unsuitable for many structural applications and this unsuitability is recognised by the Airworthiness Authorities who ask for considerable reserves of strength to be applied to any structural members using castings. More usually, the cast billet is worked on and submitted to various processes which change its character into what is termed a wrought alloy. Wrought alloys are used in several forms:
sheet and plate forgings
Materials
91
extrusions tubes
In the process of changing from cast to wrought material the cast billet, which is probably a cylinder perhaps 10 in. diameter and 25 in. long, is rst subjected to a repeated squeezing process not unlike kneading bread. The object of this is to change the random crumbly structure of the casting into a material with a grain or a recognisably continuous bound together type of internal texture.To make sheet or plate, the kneaded block or billet is passed through a succession of heavy rollers which reduce the thickness of the material to usable and carefully controlled dimensions. For forging, the large billet is cut into smaller pieces which are then, quite literally, hammered and squeezed into shape, in most cases between carefully prepared dies, to produce the nished part. Components which have been forged are sometimes difcult to distinguish externally from components which have been cast, but the manufacturing processes of forging produce a much better quality of material than the casting process. The process of making extrusions is often referred to as being like squeezing toothpaste out of a tube and this is a good analogy, with the nozzle of the toothpaste tube being shaped to produce long strips of material of complex sections. It may be difcult to believe that this is literally what happens with a metal, but given that the machines are very robust, this is indeed the process. Tubes start life as hollow extrusions which are subsequently drawn (i.e. pulled) through circular dies. This drawing process again improves the quality of the material, as well as thinning it. Accepting that light alloy is the main material used in the construction of aeroplane structures, we can list its major advantages as follows:
It is a light material, especially when used in bending or compression

members which form the major part of any aircraft structure.
It is easily produced in usable forms. It is also a material which is comparatively easy to work. (Chapter 7
discusses some of the manufacturing processes used.) However, materials which have advantages also have disadvantages, and light alloy is no exception. One of the major problems is a process known as ageing. Further on in this chapter there are some notes on the heat treatment of aluminium alloys, and it will be seen that part of the process of improving the strength of copper-bearing alloys is a process known as precipitation, or articial ageing. Unfortunately, this process continues throughout the life of the material and although modern materials are less susceptible to this change than older alloys were, a material carefully selected today may be found to have different properties in ten years time. Another disadvantage with aluminium alloy which has become well publicised in recent years is the phenomenon known as fatigue. Fatigue occurs eventually in metal members which are subjected to varying stresses, especially where the stress varies from tension to compression in a cyclic way. A large number of cycles would be involved before failure.
92
Unfortunately, the nature of aircraft structures being as it is, with constant pressurisation and de-pressurisation of cabins, bending stresses produced by gusts on wings and, perhaps worst of all, stresses produced by vibration in the region of engines or helicopter blades, fatigue becomes a major problem for the aircraft structure designer. Basically, the problem can be dealt with by limiting the stress on a member to a gure which is considerably less than the maximum stress permitted by the specication of the material. This fatigue limit stress lies in the region of 10 ton/in.2 (150 N/mm2) for almost any light alloy, and designers using wrought alloys at higher stress levels have to consider very carefully the number of fatigue cycles to which the member is likely to be subjected. A great deal of research has been put into the phenomenon of fatigue, and published literature is widely available to the serious student. Possibly an unfair impression has been given that aluminium alloys are more susceptible to fatigue than other metals. This is not necessarily so but aircraft design is always so much closer to the limits of technology, research and performance, that aluminium alloys are always in the forefront when fatigue is being discussed. A third disadvantage of aluminium is its susceptibility to the effects of notching, or small cut-outs, and this is discussed in more detail in Chapter 9.
6.2 Material properties

In specialist books dealing with metallurgy, it is usual to list metals in a certain order which is followed here.
6.2.1 Steel
Steel is iron with other elements, especially carbon, added to make an alloy. The alloying elements are present in a vast variety of different combinations and quantities to make a range of materials tted for many different tasks. The scope is so wide and we are so familiar with the material that it requires some mental effort to realise that the rusty reinforcing bars sunk into the concrete foundations of a building are made of the same basic material as the surgeons scalpel. When the aircraft industry developed beyond its initial problem of achieving sustained ight it quickly became aware that wooden constructions deteriorated due to moisture absorption. The obvious step was to turn to the universal material of engineering construction and the earliest metal aircraft used steel members where wood had been used before, but retained the fabric covering. Figure 2.3 shows a structure of this type with a tubular steel framework quite clearly similar to the wooden frame of the earliest ying machines. (The rst all-metal aeroplane was the Tubavion, a French design of 1912, but the more usually accepted rst successful all-metal aircraft was the J1 built by Professor Junkers in 1915.)
Materials
93
Subsequently, the amount of steel used in aircraft structures has been steadily reduced to the point where the designer uses it only with reluctance when he cannot nd another material to suit his purpose. This is partly due to improvements in the properties of aluminium alloys and the use of titanium but more importantly because structures made of mixed materials are generally best avoided. With aircraft at their present size it is not sensible to consider an aircraft with an all-steel outer covering (see Section 6.1) so if the skin is aluminium alloy, that tends to dictate the material of the whole structure.There are two reasons for avoiding mixed materials: one is a corrosion problem which is mentioned in Chapter 8, but a more important reason is that steel is stiffer than aluminium (higher Youngs Modulus) and so, as mentioned in Section 9.9, it will try to carry all the load, leaving the aluminium underworked, and therefore inefcient. Nevertheless, steel is still an important aircraft material and is used in undercarriages, engine mountings, joint plates (at wing roots, etc.), door latches, for bolts and fasteners, and in numerous other places where its valuable, well-documented properties, its reliability and its comparative cheapness outweigh its disadvantages of high weight, corrosive tendencies and high fabrication costs. (The high weight comment above requires some qualication. In general, steel is three times the weight of aluminium but is also three times as strong so there is no intrinsic weight penalty in the use of steel provided that all the material used is fully worked. Unfortunately, this ideal situation is virtually impossible to achieve. For example, the head of a steel bolt carries material that is only there for the benet of the spanner which tightens the bolt perhaps only once in its life. It is really that sort of waste and even larger volumes of low-stressed material in more complex ttings which account for the high weight of steel.) The principal alloying element added to iron to make steel is carbon. This addition of carbon to pure iron is virtually a denition of steel, although some strong steels have been developed which have only very small quantities of carbon (see maraging steels below). In general, the greater the carbon content the stronger the steel. The student reading textbooks on metallurgy will be confused to nd that cast iron, which is a relatively weak material, contains between 2.5% and 3.5% of carbon, while the carbon content of a low-strength mild steel is about 0.2%, and in the case of a high-strength alloy steel it is about 0.5%. The situation is that in the two materials the carbon is in different forms. In cast iron it exists as graphite suspended in the pure iron like particles in muddy water, while in the steel it is truly in solution with the iron. The other major alloying elements are:
manganese (used between 0.5% and 2%) chromium (may be up to 25% in some heat-resistant stainless or
corrosion-resistant steels) nickel (can be as high as 20%) molybdenum (0.1% to 0.8%)
94
silicon vanadium sulphur phosphorus
There is a vast range of steel specications (internationally there are several hundred), each having properties making it suitable for a certain task. These various specications are very loosely grouped into:
carbon steels

low carbon high carbon alloy steels stainless or corrosion-resistant steels heat-resistant steels and super alloys
The stainless and heat-resistant steels speak for themselves and are not of much interest now to the structures departments of the aircraft industry. Although some years ago the Bristol Aeroplane Company built an all stainless steel ghter-size supersonic research aircraft called the Type 188, a similar project now would probably use titanium. Stainless steels may be specied for some new projects where its comparatively low cost relative to titanium may still be attractive but this is becoming less likely. Low-carbon steels are the group which go to make bridges and horseshoes and almost every other engineering construction in the world but they are not of use to the aircraft engineer. Although they have a much higher Youngs Modulus than light alloy (207 103 N/mm2 [30 106 lbf/in.2] as against 72 103 N/mm2 [10.5 106 lbf/in.2]) their ultimate tensile stress, about 21 ton/in.2 or 325 N/mm2, compares unfavourably with fairly uncomplicated light alloys at about 25 ton/in.2 (385 N/mm2). The remaining carbon and alloy steels are used in aircraft structures mainly for ttings and fasteners. The usual group names are very vague and really only jargon; for the aircraft engineer it is more convenient to group the specications into those which are weldable (or not weldable), those which are difcult to machine, etc., and to grade them all in order of strength. Nevertheless, we should attempt a denition and say that alloy steels can be considered as those with total alloying constituents above 5%, and high-carbon steels those with carbon above 0.2% and with manganese as the only other constituent. It can be seen from Table 6.1 that the range of strengths available is wide and that even when requiring a property as particular as welding suitability, the designer is still faced with a choice and therefore the need to make a decision. A total list of available steel specications would increase the problem so that the short list shown should only be taken as a general indication of the values of mechanical properties. A designer or stressman would know (or would nd out) which steels were in use or in stock within his company.
Table 6.1 Steels Permissible tensile stress2 Country of origin UK UK Ref. No. BS.S.1 S.92 S.514 T.45 S.154 S.98 S.28 S.80 S.521 at yield 5165 000 69 000 90 000 90 000 96 000 145 000 152 000 98 000 29 000 ultimate (UTS) 78 000 90 000 112 000 100 000 127156 000 168 000 224 000 127 000 78 000 Elongation at break (%) 10 15 10 12 10 8 12 30 Hardness Brinell (B) or Rockwell (R) B or C (B) 160 (B) 180250 (B) 230 (B) 250300 (B) 360 (B) 445 (B) 288
Izod 15 35 40 25 15 25
Remarks 0.20% carbon. Lowest grade steel for aircraft use Bar Sheet Tube Suitable for welding and compatible
UK UK UK UK
General-purpose bar High-strength bar Very high-strength alloy bar, but note Izod trend Bar Sheet General-purpose CRES. Sheet is not structural but good forming properties
USA USA
SAE.1015 4130 4130
45 000 70 000 103 000
55 000 90 000 125 000 17 10
(R) B64 (R) B90 (R) C26
0.15% carbon. Lowest grade suitable for aircraft General purpose steel in various forms and HT conditions. The major a/c steel Compare with BS.S.98
Materials
USA (1) (2) (3) (4) (5)
4340
145 000
180 000
14
(R) C40
The properties shown are only approximate and for design purposes reference must be made to the published specication. Stress is quoted in lbf/in.2. To convert to N/mm2 divide the gure given by 145.0 (see also note 5). Youngs Modulus for steel is 2830 106 lbf/in.2 (read twenty eight to thirty by ten to the sixth pounds force per square inch). Poissons ratio for steel is 0.27. Stress may be quoted in units called kips, 1 kilo pound = 1000 lb.
95
96
In the 1960s a range of steels evolved called maraging steels, from which carbon was virtually eliminated. The alloy steels which achieve their strength with a high carbon content are usually very difcult to machine (note, the Brinell hardness number in Table 6.1 for S.28 and S.98) and are often rather brittle or notch sensitive (note the Izod number in Table 6.1). The maraging steels are strong, tough and comparatively easy to work. Their heat treatment process is different from carbon steel and simpler (see Section 6.5) but their initial cost is high. This may be offset against lower machining costs but the total cost of any particular component in maraging steel should be compared with the cost in titanium, also bearing in mind the considerations of Section 6.4.
6.2.2 Non-ferrous metals

Non-ferrous metals are usually taken to mean copper and the copperbearing alloys: brass, bronze and gunmetal. None of these are of specic interest to the aircraft structures engineer, but copper has an interesting physical property known to metal workers for many years which also exists in aluminium and is exploited in some aircraft manufacturing processes. Copper that is stretched or is deformed by hammering becomes tougher in a process known as work hardening. The phenomenon is only mentioned here to introduce the term; its usefulness in working aluminium will be discussed in Chapter 7.
6.2.3 Aluminium and aluminium alloys

Aluminium and its alloys have already been discussed in Section 6.1. It seems certain that these will remain as the major materials of aircraft construction for many years to come, although, in percentage terms, titanium may, and reinforced plastics will, make progressively larger inroads. There is also, at present, a limitation on the use of light alloy on supersonic aircraft because at high speed the skin of the aircraft can approach temperatures which are detrimental to the strength of the material. Concorde was designed as a Mach 2.2 aeroplane because higher speeds would have dictated the use of a material other than aluminium alloy. Even then a special alloy was developed particularly for the aircraft with the object of obtaining an improvement in the strength and fatigue characteristics over a wide range of temperatures. In the notes on steel above we said that steel was graded or denominated by its carbon content, and then by the quantity of other elements in its composition. A similar system applies to light alloy, which falls into three groups. The original formulation was patented and marketed under the trade name of Duralumin. For many years the name Dural was freely applied to all varieties of light alloy (at least in Europe) in much the same way that the name Hoover was applied to all makes of vacuum cleaner. Dural is a copper-bearing alloy which was reputedly discovered by accident and possessed the rather strange characteristic that if it was heated
Materials
97
(to about 480C) and suddenly quenched in water it became soft and easily bent or formed, but then the formed shapes, over a period of a few hours, regained their earlier strength and hardness naturally and without any further attention. This remarkable property, which is entirely different from the work-hardening characteristic of copper and brass (although most aluminium alloys also work harden), has inuenced the techniques, and hence the design philosophy, of aircraft manufacture for 80 years. It was later found that a further heat treatment, which was called articial ageing (at about 170C), made an additional improvement in the strength. For comparison, the 0.1% proof stress of naturally aged Dural was about 33 000 lbf/in2 (230 N/mm2) and modern derivatives still have much the same strength. The same alloy articially aged would have a 0.1% proof stress of about 51 000 lbf/in2 (350 N/mm2). The articially aged copper-bearing alloy was well established by the early 1930s, with the second group of alloys already making an appearance. This group differed from the Duralumin group by the addition of nickel and a higher magnesium content. Not widely used for structures, this group was mainly employed in aero engines and as forgings. The third group, which came into use in the middle 1940s, used zinc and magnesium as alloying elements. The 0.1% proof stress of a typical early member of this group was 74 000 lbf/in2 (510 N/mm2), which represented an enormous increase over the previous families of alloys. Unfortunately, possibly because wartime pressures had inuenced the development programme, the desirable properties were accompanied by some highly undesirable tendencies. Fatigue cracking (see Section 6.1) accompanied by higher notch sensitivity (Section 9.3) led to some unpleasant failures. A new phenomenon called stress corrosion, initiated by unrelieved internal stresses caused by heat treatment, led to cracking which was very difcult to detect. Many designers considered that the disadvantages outweighed the benets, but recently improved versions have brought this group of alloys into more general use. Table 6.2 lists the mechanical properties of some of the widely used alloys. The designation used is nominally international but most familiar in America. The current equivalent British specications can easily be checked in the brochures published by the materials producers. Some notes on the various methods of numbering specications are given in Section 6.6.
6.2.4 Titanium
Titanium is effectively the last on the list of metals used in aircraft structures, although magnesium was used in the past and beryllium may be used in the future. One of the most plentiful metals on Earth, it is difcult to rene and therefore expensive. Its weight is between that of steel and light alloy, but it is strong and corrosion resistant even at high temperatures. These characteristics have led to its increasing use for specialised pieces of structure, such as fairings near jet engine efux or intake
98
Table 6.2 Light alloys Permissible tensile stress
Country of origin USA UK UK USA UK
Ref. No. 2014T4 L.164 L.102 2014T6 L.165
0.2% proof
ultimate (UTS) 56 000 61 000
Elongation at break (%) 14 8
Brinell hardness No. 105 135
Remarks L.164 is sheet material { L.102 is a bar Note the effect of heat treatment on properties. UTS increases 9%, proof stress increases 43%
}
} }
35 000 50 000
USA UK USA UK UK UK USA UK
2024T4 L.109 QQA250/11 L.111 L.113 L.114 7075T6 L.88
39 000
59 000
15
120
37 000
43 000
L.111 is a bar L.113 is a sheet L.114 is a tube 150
This is weldable alloy. Properties are before welding
62 000
71 000
Aluminiumzinc alloy
(1) The properties are approximated to illustrate the grouping. Figures must not be used for design and reference must be made to the specications. (2) Stress is quoted in lbf/in.2. To convert to N/mm2 divide the gure given by 145.0 (see also note 5). (3) Youngs Modulus for light alloy is 9.510.7 106 lbf/in.2. (4) Poissons ratio for light alloy is 0.3. (5) Stress may be quoted in units called kips. 1 kilo pound = 1000 lb. (6) The sufxes T4 and T6 signify that the material can receive two different heat treatments. The effect is that UTS increases 9% and proof stress increases 43%. L.164 and L.165 are the same basic alloy but in the UK system the different heat treatments give the two specications separate reference numbers.
Materials Table 6.3 Titanium and maraging steel Permissible tensile stress Country of origin UK UK US UK US Typical ref. No. T.A.10 T.A.6 AMS.4935 DTD.5212 AMS.6521 Description Titanium sheet Titanium sheet Titanium sheet Bar maraging Sheet steel 0.2% proof 130 000 67 200 130 000 245 000 270 000 ultimate (UTS) 140 000 82 900 140 000 260 000290 000 280 000
99
Elongation at break (%) 8 15 8 8 8
(1) The properties shown are approximated to illustrate the comparison with the materials shown in Tables 6.1 and 6.2. Figures must not be used for design and reference must be made to the specications. (2) Stress is quoted in lbf/in.2. To convert to N/mm2 divide the gure given by 145.0 (see also note 4). (3) Youngs Modulus for maraging steel is 27.5 106 lbf/in.2. Youngs Modulus for titanium is 15.516 106 lbf/in.2. (4) In some lists of material properties stress is quoted in the units called kips. 1 kilo pound = 1000 lb.
edges where stainless steel may have been used previously. On the Boeing 727 designed in the early 1960s, less than 2% of the structure weight was titanium, but by the late 1960s the Boeing 747 structure was almost 10% titanium by weight and this trend was almost certainly exceeded by military aircraft. In recent years the move towards composite structures has rather reversed the trend towards titanium. Bearing in mind the high initial tooling cost of composites and the comparatively low repair cost of metal structures, the pendulum may yet swing back towards titanium. Returning to the Boeing 747, it is interesting to speculate on how much the designers were driven, against their will, to use titanium because weight in the nal design exceeded that of the project calculations. If they were forced to use more of the expensive titanium than they would have wished, then the price of the aircraft must have been higher than expected; but in the event it is a magnicent money-making machine and supports the argument that the cost of raw material should not necessarily be a deterrent to its use. (See also Section 6.4.) As with other metals, titanium is used as an alloy, in fact as two families of alloys, one with aluminium and one with manganese. These give a good range of available properties (see Table 6.3) which should be compared with those in Tables 6.1 and 6.2.
6.2.5 Reinforced plastics and bonded structures

Reinforced plastics (or composites) and bonded structures are now widely used in aircraft structures, especially in oors and ancillary structures such
100
as overhead luggage bins, but increasingly now in primary structures. Gliders and some light aircraft have effectively a 100% composite airframe and the structure of the Airbus A400M is 30% composite and the A380 of the order of 40% composite by weight. The most familiar of these materials is Fibreglass which is another of those names like Dural and Hoover where a trade name has become a part of the language. More correctly Fibreglass is one companys name for a loose mat of random bres of glass. The glass is spun, in the sense that spiders spin the threads of their webs, by passing molten glass through very small holes. These very thin bres are also spun in the other sense, like cotton or wool bres, into threads (or rovings) which can be woven into cloth. Because the glass is spun so ne, the threads can be bent to a radius of the same order as the thickness of the cloth so that a brittle material has been changed into a exible material without losing its strength or resistance to corrosion. What is usually known as breglass is, in fact, glass-reinforced plastic (GRP), manufactured by laying glass cloth, or glass random mat, onto a prepared surface and then impregnating the cloth with liquid plastic which sets hard into a stiff composite sheet. Two groups of plastics are in common use for making GRP. Polyester synthetic resin is the more common and epoxy resin is the more expensive but has rather superior properties. The highest strength glass-reinforced layups use a special cloth called unidirectional which is not woven but laid as carefully prepared papers of lightly bonded threads. Each layer of paper has its threads running parallel to one another and successive layers are placed at right angles to each other like the grain in plywood. The resulting sheet is often used as the facing for a sandwich panel board, the interior of the board being a honeycomb of plastic impregnated paper (see also Section 6.1). Used in this way reinforced plastics are very efcient in terms of strength and lightness, and components grouped under the general heading of composite and bonded structures are becoming more and more accepted. Helicopters have been built using GRP almost exclusively in their structure, including rotor blades, but xed-wing civil aircraft designers have avoided making wings or fuselages out of composites. Until fairly recently military aircraft designers have been more adventurous and the AV-8B may have as much as 60% of its structure in composites. (It is not the intention to give the impression that all bonded structures are necessarily either plastics or sandwich structures. Metal-faced, metalhoneycomb-cored panels and metal panels with adhesives are all bonded structures and, in the early days of ying, wooden laminated structures bonded with casein glue were used for some excellent aircraft, among them the beautiful Lockheed Vega. A later version using plywood skins on balsa wood core bonded with synthetic resin glue was used for the structure of the even more elegant de Havilland Mosquito.) The essential point of composites is that they combine two materials, particular properties of one material being used to compensate for perceived shortcomings in the other. A non-aviation, but easily digested,
Materials
101
example is the steel reinforced concrete lintel. Concrete is strong in compression and cheap but is not particularly good in tension. Steel is very strong in tension but expensive relative to concrete. Constructing a concrete lintel, or beam, with steel rods laid into it where the tensile stresses are expected, allows for a still relatively cheap solution to the builders problem.
6.2.6 Plastic
Plastic is also used without reinforcement but other than for windows it is not normally used as part of the structure. In pressurised aircraft the windows are very denitely part of the structure and have been developed into a standard form. In the great majority of designs there are three separate layers: a load-carrying layer, a back-up load carrier in case the rst one cracks, and an inside plastic shield to prevent scratching by passengers. The load carriers will probably be glass, either heat-treated toughened glass like ovenware, or laminated three-ply glass. (Again we have trade names for these products which have been adopted into the language, in this case Pyrex and Triplex.) The most usual transparent materials are acrylics (Perspex, Plexiglass, etc.) in sheet form. Although there is a ammability problem with acrylic, it is not so serious as to exclude the use of the material. However, there is now a newer range of transparent plastics called polycarbonates which are more re retardant and much tougher than acrylics although with similar strength.
6.2.7 Strength to weight ratio

All the materials mentioned in the above paragraphs have their own virtues and values. It is sometimes thought necessary to make a comparison of materials by what is known as their strength/weight ratio. At rst glance this appears simple and indeed for a straightforward tension member it probably is, but for any more complex piece of design there will be so many factors to consider that the material will almost certainly be selected on the basis of some other criterion. If we discuss the selection of material on a component to component basis then probably the only sure method is to make complete designs for each component in the alternatives and only then select the lightest design. As there is seldom time available for such an exercise then perhaps the strength to weight ratio may have some signicance as a guide, but it can only be used with caution.
6.3 Smart structures (and materials)

The term smart structures now encompasses a wide-ranging set of concepts in engineering, nding applications, at least in a developmental
102
sense, in aerospace, civil engineering, sport, medicine and the process control industry. Some of the main concepts are set out below.
6.3.1 Sensory structures

These are structures made of any material to which have been added sensor devices of some form. An example is an aircraft wing skin fabricated in composite materials such as glass or carbon-reinforced plastic that contains, embedded within the resin matrix, a series of strain gauges. The technology for measuring strain, which is dened in engineering terms on page 63, could be based on electrical/electronic gauges or on new optical bre techniques. (A strain gauge is a short, thin wire with electrical connections at each end. This is bonded to the structure in such a way that as the material of the structure stretches, the wire stretches and changes its electrical resistance which can then be measured and converted into a measure of length.) Either way, the gauges or sensors provide a form of articial nervous system for the skin which then feels the effects of applied loading. Strain measurement is only one example. Sensors could also be used to detect temperature distributions, measure vibration, listen for cracks forming or bres breaking, or simply measure the shape of the skin as it distorts during different phases of ight. For a structure to be termed smart in this way, it is generally required to have multiple measurement points (sensors) and is all the smarter if the sensors are amalgamated with the structural material (as in the example above), rather than just bonded to the surface. The addition of sensors, while giving extra functional capabilities, does not otherwise directly improve mechanical performance. Sensory structures are essentially passive but the information gathered from them can be used imaginatively to create smart active structures.
6.3.2 Smart active structures

A form of articial nervous system was described above but a smart active structure has the muscles to go with it. The exibility of todays aircraft wings (that is, their ability to operate in differing conditions) derives from their complex arrangement of compound panels able to translate under actuator control. The wing shape is controlled and changed to the optimal geometry for the ying conditions or particular stage in the ight prole. Optimum in practice engineering terms, always means compromise. Smart active structures may allow aircraft designs to more closely approach the optimum. The concept of an aircraft with an active smart structure may include, for example, a deformable wing with a continuous surface able to distort itself into the optimum shape for its ying condition. This might be achieved by embedding within the fabric of the wing a distributed network of micro-actuators made of piezo-electric ceramic material, i.e. material
Materials
103
which changes shape when an electric current is passed through it. If activated in the appropriate combinations and with programmed voltages, these actuators could conceivably be made to ex the surrounding structure to the desired shape. Smart active structures can also embody the sensor networks of sensory structures. Linked with a computer processor, a closed-loop system can be envisaged in which a structure senses its changing condition and automatically recongures itself in accordance with a prescribed logic. An aerospace example may be automatic compensation for a loss of lift due to damage or power reduction, or reconguration of a ying surface during transitions to and from supersonic ight. A more immediate and simple application is the forced damping of vibration. Sensors at different locations can detect local vibrational modes. Their signals, when amplied, inverted and fed back to local actuators then cancel out the vibration.
6.3.3 Use of smart structures in aerospace

The foregoing descriptions of smart structures have hinted at examples from aerospace that could possibly benet from these concepts. The introduction of smart engineering into aerospace will follow a gradual, step by step path. The rst advantages to aircraft operators, whether they are military or civil, will result from better quality information on airframe status, leading to more operationally effective or cost-effective maintenance. As items in the aircraft designers toolbox, active structures open up a world of possibilities. In the long term, continuously deformable ight surfaces could lead to revolutionary aircraft designs in terms of operational capabilities, fuel efciency and safety.
6.4 Cost as a property of a material

Materials are selected to perform particular functions on the evidence of their properties and hence their suitability. When the function is structural then the major material property to be considered is strength. However, if strength considerations permit the possible use of several different materials, then other factors begin to affect the selection. Quite clearly one of these factors is cost. Unfortunately, while it is possible to list the cost of materials in terms of price per unit weight, this is virtually no help in making a decision. Cost is more properly associated with the nished component; that is, bringing into account manufacturing costs which are fairly easy to assess, although even this assessment may be affected by quantity as well as by the difculty of working the material. If a component is to be produced in very small quantities the tooling and design/drawing costs will create a different situation from that associated with large quantities. In these circumstances it may, for example, be cheaper to make a single shelf out of
104
sandwich board with ready-made rigidity, rather than make a sheet-metal shelf with stiffeners and formed edges. The sheet-metal material would be cheaper than the sandwich board but the design/drawing time for the sheet-metal construction would be much greater. Major companies now run special departments to assist the designer with decision making in these areas but in most design organisations, especially when dealing with the mass of small detail parts, very little conscious consideration is given to these factors. Two other cost problems may inuence materials selection. One is the replacement cost of life expired (worn out) components, bearing in mind that this must include labour and inspection costs. The other is operating costs in terms of fuel expended in carrying weight during the life of the aircraft. The following calculation was made for the second edition of this book (in 1992) and since then large trans-Atlantic aircraft have improved so much in efciency that reworking the example at todays values yields a much less impressive result. However, the principle remains true so the argument is repeated here:
the price of aluminium alloy is the price of titanium alloy is the price of turbine fuel is the average all-up weight of a Boeing 747 half-way between London and New York
= $ 6.00/kg = $44.00/kg = $ 0.20/kg
= 280 000 kg = 11 000 kg/h = 80 000 h = (80 000 11 000)kg = 880 000 000 kg = (880 106 0.20) = $176 000 000 = (176 106) (280 103) = $628
the average cruising fuel consumption aircraft operating life assuming 20 years
at 4000 hours per year
therefore total fuel consumed during

the life of the aircraft
total cost of fuel during the life of

the aircraft
therefore the fuel cost of carrying 1 kg

during the life of the aircraft
Beside this large gure, the additional cost of specifying, for example, titanium instead of aluminium (where all other factors are equal) is insignicant. (We should be aware that the calculation above is very crude and only intended to indicate the direction of an argument. One opposing argument is that with an investment interest rate running at about 20%, a dollar spent today would be worth $15.40 in 15 years.)
Materials
105
In spite of this arithmetic, aircraft have to be sold and airline buyers still tend to look at initial price, so that although material cost is important, the relative costs of different materials may still prevent the lightest solution to a design problem being adopted.
6.5 Heat treatment

The heat treatment of steel is not now normally carried out by the aircraft manufacturers because of the complex heat treatment requirements of these high-strength steels. Nevertheless, it is useful to have some knowledge of the terms and expressions used. Heat treatment changes the characteristics of a steel, changing one desirable feature for another; for instance, a tough springy alloy may be softened so that it can be bent and then rehardened to its former properties. In the carbon steels, the carbon is held in the alloy either as pearlite or cementite and it is the difference between these two states that determines whether the steel is soft or hard. When the material is red hot the carbon is in cementite state and if allowed to cool slowly the carbon changes to pearlite and the material is soft. (Used here, soft is a comparative term because, even in this condition, the high-carbon and alloy steels are difcult to work.) If the material is cooled rapidly from red hot by quenching in oil or in water, then the carbon is trapped in its cementite form and the steel is strong, hard and brittle. This brittleness is subsequently reduced in exchange for some reduction of strength and hardness by heating to a lower temperature and allowing some of the carbon to change its form. The three conditions described above are entitled:
annealed (or sometimes normalised) hardened tempered

but for the structures engineer the names, with some idea of their meanings, are more important than the processes because specications often use these words in their titles; for example, a steel might be described as annealed steel wire for oil-hardened and tempered springs. Also, the terms have been carried over into the description of light alloy heat treatments. Maraging steels are heat treated in rather a different way, in that some part of one of the alloying elements is persuaded to break its bond with the other constituents, leaving exactly the correct amount to give the material properties desired. When the element in question is fully dispersed through the material it is said to be in solution, and while it is changing it is said to be precipitating. The heat treatments which induce the change from one condition to another are thus called solution treatment and precipitation treatment.
106
The heat treatment of aluminium alloys follows roughly the same pattern as that for steels. That is, an alloy can be in the annealed or soft state and then be made progressively stronger by solution treatment and precipitation treatment which are roughly equivalent to the tempering process for steel. Most aluminium alloys work harden like copper, and this property is exploited in some workshop processes (see Chapter 7). Because of work hardening, light-alloy sheet comes out of the rolling mill in a condition which might be described as unnatural and before being supplied to the aircraft manufacturer it is heated and cooled to change its characteristics by dispersing the crystals of the alloying elements (like copper) evenly through the material. This particular solution treatment or normalising process is carried out by soaking the material in a bath of melted salt, or by baking it in a hot-air oven at approximately 450C for about half an hour and then quenching it in water. (More exact gures are shown in Table 6.4.) An interesting property of most light alloys is that immediately after normalising, and for about half an hour longer, the metal is soft. This property is useful for the manufacturing processes of bending and forming. After about an hour the solution-treated material starts to age harden. This is an important process because it continues to some extent throughout the life of the metal and components made from it. Precipitation treatment is articial or accelerated ageing. The various stages of heat treatment are called the conditions of the alloy and are listed and coded by the Standards Bureaux of the major manufacturing countries. British and American codes are shown in Table 6.4 with space for noting other equivalents. In itself, this information is not very important to the engineer, but it is a useful reference when considering the specications of common aircraft materials, because variations in treatment conditions may result in considerable changes of strength.
6.6 Reference numbers for materials

The quantity of different variants in the materials used in aircraft construction is enormous and totally bewildering. All of the major industrial countries have their own alloys, and to compound the confusion each has their own unique reference numbering system. Indeed some countries have more than one system. The United Kingdom has at least two systems for aerospace quality materials and the USA has at least three for aluminium alone. One of the American systems has, at least in theory, been adopted as an international standard and some aluminium suppliers refer to it in their brochures. Within this system which was designed by the Aluminium Association, alloys are allocated a four digit code number, followed by a code which identies the heat treatment condition. The pattern of the whole number then looks like this: 2024T3
Materials Table 6.4 Heat treatment designations for light alloy Ref. No. (see also Table 6.2) 2014 L.164 L.102 2014 L.165 2024 L.109 QQA250/11 L.111 L.113 L.114 7075 L.88 Solution treatment temperature (C) 502 Precipitation temperature and time None None 160C for 18 h 175C for 10 h TF T4 None TB TB T6
107
British designation
USA designation T4
Other designations
505 5 502 505 5 493 495 5
510540
180C for 8 h
TF
482 460 10
121C for 24 h 135C for 12 h TF
T6
Note: This table illustrates some of the difculty in being certain of exact equivalence in similar specications; for example, 2014T4 and L.164 are similar materials (see Table 6.2) but the difference in solution treatment temperature (502C to a possible 510C) might make a difference.
The rst digit indicates the principal alloying element and there are a total of seven groups, of which only the 1000, 2000, 6000 and 7000 series are of much interest to the aircraft structures engineer. The code is as follows: 1000 series pure aluminium (99% pure) 2000 series copper alloys 6000 series magnesium alloys 7000 series zinc alloys (In Section 6.2.3 we discussed three groups of alloys. The rst group (Duralumin) is in the 2000 series and the third group is in the 7000 series. The middle group (known in the UK as Y-alloy derivatives) is mainly in the 2000 series but variants appear in the 6000 series.) The second digit of the code denotes a revision to the original alloy (thus 2618 is the sixth modication of an alloy originally called 2018). The third and fourth digits are index numbers. Following the dash, the principal codes associated with aircraft materials are:
108
0 T3 T4 T6
annealed solution treated and then cold worked (i.e. attened or straightened) where this improves mechanical properties solution treated and naturally aged solution treated and articially aged
There are a good many of these heat-treatment codes to cover complex conditions; for instance, T6511 is a subvariant of the T6 state, indicating that the product has been stress relieved by stretching and subsequently straightened. A full list of heat treatments, alloys, designations, etc., can be found in Aluminium Standards and Data but of the very many alloys and variants available, those shown in Table 6.2 constitute the major interest of aircraft structures engineers. A major snag with this otherwise worthy system is that the principal American reference numbering system for steels also uses a four-digit code which is totally unrelated to the aluminium code. Fortunately, the main aircraft steels fall into groups 3000 and 4000, which do not conict with the aircraft aluminiums. Also, the number is usually preceded by the letters SAE (Society of Automotive Engineers) but caution is necessary; for example, a screw made of SAE.2330 is steel, not aluminiumcopper alloy. Other numbering systems used in the USA are Federal Specications, Military Specications and Aerospace Material Specications and they look like this (in order): QQ-A-367 MIL-A-22771 AMS.4127 These systems also cover materials other than metals. In fact there are MIL specications covering every imaginable aircraft material, including sewing thread, and, for example, an adhesive product may meet Federal Spec. MM-A-132, Type 1, Class 3, and MIL-A-5090-D, Type 1. A further word of caution is necessary here, the Aluminium Associations numbers identify the material but the Federal and Military Specications list the requirements for the material.This is really a quality control matter and the British system, which is discussed below, recognised this from the beginning (see also Chapter 10). The ofcial specications, by listing minimum performance standards and by requiring that those standards should be demonstrated by test, provide a clear level of condence for the designer and eventual user and an unequivocal basis against which the material manufacturer can quote prices. The British industry uses two systems which do not overlap. A material which in the rst place has been developed by private industry has its characteristics listed by a government department (Directorate of Technical Development) in the form of a specication. This specication carries a number which looks like this:
Materials
109
DTD.610 The material, which is now called by its specication number, is used and developed and if it becomes well established it is adopted into the British Standards Institutions list of aircraft materials and given a number which looks like this: L.72 In fact, the complete number would be BS.3.L.72, where the 3 signies the third issue of the specication. (This incidentally brings up another minor difference between British and American practice. The British number the issues of a document and the Americans number the revisions. Thus when a document suffers its rst change, if it is British it becomes issue B or issue 2 and if it is American it becomes revision A or revision 1.) The strength and weakness of the British specication numbering system is that the number is just a number with no hint of indexing about it. This is illustrated by L.102, which is a bar material and its published specication is complete; from it we can nd the heat-treatment condition, the strength and almost anything else we could want to know about this particular bar stock. The same applies to a sheet material called L.164. What we cannot quickly recognise just from the numbers is that they are related and are both 2014 alloy. (The random nature of these numbers is also illustrated by the history of the specications for Dural bar. First issued in 1924 as DTD.18, the specication reached its third issue as DTD.18C and then became successively L.1, through to 6L.1, L.39, L.64 and L.102.) DTD and British Standard numbers also cover other materials. The DTD numbers do not change their pattern for other materials but in the British Standards steels (for instance) come into an S group, e.g. S.80 is a stainless steel. Tubes can be confusing, L.114 is a light alloy but so was T4. As we said at the beginning of this section, the numbering situation is very confusing but structures designers and stressmen quickly become familiar with the materials favoured by the company for whom they are working.

It is important to remember that there is an enormous range of materials available depending upon, inter alia, the application, the manufacturing processes to be used and the designers imagination. However, a given in
110
aircraft structures is that the lighter they are the better. Every kilogram (or pound) of extra structure is a kilogram (or pound) less of revenue earning cargo and extra fuel is required to carry it about the sky. Aluminium is used in its alloyed forms due to the much higher strengths available compared to pure aluminium. A variety of alloying elements is used with a similar variety of results. Aluminiumcopper alloys are heattreatable to give a good compromise between strength, toughness and corrosion resistance. Aluminiumzinc alloys may be heat treated to give higher ultimate strength but may be more notch sensitive and prone to corrosion. The specic strength (strength to weight ratio) of alloy steels is broadly similar to that of aluminium alloys. However steel is much less widely used in aircraft due to what might be called the incidental weight penalty. This is manifest in, for example, wing skins needing to be of a particular thickness to resist buckling and in-service handling rather than pure strength. The use of alloy steels is mostly limited to where very high local strength is required, such as for wing lugs, engine mounts, bolts, etc. There is a continuing trend toward more composites in aircraft structures. Their use allows the number of parts to be reduced and very clean aerodynamic external surfaces can be achieved. The costs of composites tooling need consideration, especially for low volume production, and composite structures can be more difcult to repair than aluminium. Smart structures are a combination of structure, sensors, actuators and processing power. They change shape in response to loads or other parameters. It could be argued that gust alleviation systems on large airliners are a form of smart structure but, beyond that, their use on aircraft is still essentially in the developmental stage. Aluminium alloys are designated by a four digit numbering system followed by their temper, e.g. 2024T4. The aerospace industry, though, references material by much more tightly dened specications. These specications give the desired level of assurance that the key properties, such as tensile strength, are achieved.
6.8 References
Aluminium Standards and Data 900 19th Street NW Washington, D.C. 20006 Tel: 202 8625161 The Properties of Aluminium and its Alloys The Aluminium Federation Broadway House Calthorpe Road Birmingham B15 1TN Tel: 0121 4561103 Fax: 0121 4562274
Materials
111
Metallic Materials and Elements for Aerospace Vehicle Structures (MILHDBK-5*) Department of Defense Naval Publications and Forms Centre 5801 Tabor Avenue Philadelphia, PA 19120
Chapter 7
Processes
7.1 Introduction
The range of manufacturing processes used by the aviation industry is so wide that to give only an indication of its scope in the same way that this book gives an introduction to structures would need a companion volume. We will therefore mention only a few of the basic processes, and show some sketches and diagrams of the machinery used. The workshop processes of joining components by riveting, etc., are important in the construction of aircraft and we take this opportunity of showing how bolts and rivets are identied by stores reference numbers, as well as showing something of the range of fastenings available.
7.2 Manufacturing
When the aviation industry wants to specify a single piece of an engineering assembly it will refer to a piece part, a part, an item, a detail, a component or sometimes even to a dash number. An assembly will have an associated parts list, bill of materials or schedule of parts (or some permutation of these words) which itemises the components of the assembly, notes the quantity of each that is required, names each item (or states its nomenclature), and gives the part number or stores reference of each physical part. This number is also (usually) the reference number of the engineering drawing of the item. Often a drawing will give information on a range of similar parts, especially in the type of situation commonly found in the aircraft industry where parts are handed (i.e. parts installed on opposite sides of the aircraft are mirror images of each other). In this case, the identifying part number will usually be the drawing reference number plus a sufx or dash number; for example, a port or left-hand wing may be shown on drawing number 28503 and have a reference assembly number 28503-1 while the starboard or right-hand wing will be 28503-2. The number of variants of this numbering system is the same as the number of companies in the aerospace industry, but the principles are fairly general.
Processes
113
Fig. 7.1
Drilling machine.
7.2.1 Types of component

The majority of aircraft components are either sheet-metal parts (selfexplanatory) or machined parts which may be made from bar stock, extrusion (see Section 6.1), forgings, castings or occasionally tube or plate. The exceptions are forgings, castings or lengths of extrusion which are used as supplied or which receive some minor work, such as drilling (see Fig. 7.1). Machined parts are those which have been worked on a machine tool. The most common machine tool after the drilling machine is the lathe (see Fig. 7.2). Lathes produce parts which have circular sections, such as bolts, wheels, bearing bushes, pistons, etc. In aircraft production, however, the milling machine has probably overtaken the lathe as the most important machine tool (see Fig. 7.3). The milling process makes smooth at faces, grooves, shaped recesses or slots. It does this by cutting chips of metal from a surface by rotating a tool with one or more cutting teeth and driving it through the metal to be removed. By variations on this process some very complicated shapes can be produced (see Fig. 9.12). Although sheet-metal parts are also worked on machine tools, these are in a separate range from those mentioned above. The tools for working sheet metal are enlarged and power-operated versions of the hand tools used by sheet metal workers for hundreds of years. With the exception of hole drilling, which is a very basic metal-working operation, there are only two operations which can be performed on sheet metal: it can be cut and it can be formed. (Note here that the difference between sheet and plate
114
Turning
Facing
Parting
Fig. 7.2
Lathe.
is usually taken to be that material over 0.25 in. or 6 mm is plate, but the sheet referred to in this paragraph is less than 0.15 in. or 3 mm.)
7.2.2 Cutting and forming

Cutting can be done in one of three ways, by guillotine, blanking or routing. The guillotine or shear (Fig. 7.4) operated either by foot pedal or power, makes long, straight cuts (410 ft dependent on machine size). Blanking is a punch and die process. Any irregular shape can be cut out as a hole in the thick (comparatively) steel plate to make the die. A matching piece of similar material (the punch) made so that it will just pass through the hole, is mounted above the die in a press. When a piece of thin sheet is placed between punch and die the punch is pressed down and cuts one piece of matching shaped sheet. The process is very versatile and can reproduce accurate components, at a very high speed if necessary, from subminiature size to dimensions limited only by the power of the press.
Processes
115
Horizontal mill
Vertical mill
Fig. 7.3
Milling machines.
Fig. 7.4
Guillotine.
116
Fig. 7.5
Fly press.
The tools (that is, the punches and dies) are expensive but the labour cost of the nished component is low, therefore this method is economical for long production runs with the runs needing to be longer as the component size (and therefore tool size) increases. Presses vary enormously in appearance and even in their exact function. A simple hand-operated machine, called a y press, is shown in Fig. 7.5, but hydraulic power presses of 500 tonne capacity are quite common equipment. Routing also reproduces irregular sheet-metal shapes. In this method, a special milling cutter run at very high speed (18 00024000 rpm) cuts through a pack of ve to ten sheets of material. The slot, which may be about 0.25 in. wide, can be of any length and follow any path guided by a template or model of the shape required. The tools are much cheaper than blanking tools and there is virtually no upper size limit, but the accuracy of prole is rather dependent on the skill of the operator and the labour cost per part is high. The name routing is sometimes given to the milling process of making shaped recesses in plate material.
Processes
117
Folder bar Beam Machine bed Pivot
Fig. 7.6
Folding machine.
Forming sheet metal is probably the major activity of aircraft production. Straight bending of anges or single-curvature skin panels, such as those on the parallel central part of a fuselage or the surface of a conventional wing, is very easy. Forming parts with double curvature, such as the skin panels at the nose of an aircraft or the lips around the anged lightening holes (see Fig. 9.2), is considerably more difcult, but the processes are well understood in the industry and seldom cause major problems. There are some further comments on the design aspects of forming in Chapter 9. Bending or folding of straight anges is carried out . in a folding machine (Fig. 7.6) or a press brake (Fig. 7.7). The latter machine can also be set up for blanking or piercing (i.e. blanking small holes). Double curvature forming may be carried out in a press with punch and die tools similar to those described above, but arranged with sufcient clearance so that the sheet is not cut between the two parts of the tool but is formed to the shape of the punch. Again such tools are very expensive and only justied by a long production run or the absence of an alternative method. A similar but cheaper, and in some ways more adaptable, version is the rubber press method. The punch is still required but is mounted on the bottom plate of the press and instead of a die the top plate carries a rectangular block of rubber contained inside a strong box or fence with the
118
Fig. 7.7
Brake press.
underside open.As the top and bottom press plates are squeezed together, the rubber moulds itself and wraps the sheet metal around the punch. Punches (or form blocks) can be made of materials which are more easily worked than the steels used for matched punch and die sets, and the rubber (or sometimes now a plastic called polyurethane) is long lasting. Purpose-built rubber presses are very large and may work at several thousands of tonnes load. The process of forming sheet metal depends to some extent on the property of material called work hardening (see Section 6.2.2) and luckily aluminium, the major material of aircraft structures, has good workhardening characteristics. To illustrate this point, imagine a component made of sheet aluminium and shaped like a deep dish or saucer. The manufacturing tools required will be a punch and die which respectively t the inside and the outside of the nished component. Also required is a clamping ring, or pressure plate, to hold the unpressed metal blank in position over the die ready to receive the punch. As the punch is pressed down it starts to stretch the sheet into the hollow of the die, because the clamping ring (intentionally) prevents material being drawn in from the sides. Now as the material is being stretched it must get thinner and wherever it is thinnest we would expect it to be weakest and therefore to stretch more, so that it progressively becomes thinner in one small area. However, because of work hardening this is not the case. The initial small stretch works the material, improving the strength of the sheet locally, transferring the deformation to a new point and so by progressive action ensures that stretching takes place evenly over the whole area of the dish.
Processes
119
7.3 Jointing
7.3.1 Welded joints
Welding is the process whereby components made of metal or some plastics are joined by being made so hot in an area local to the joint, that the material of both components melts and runs together so that when the heat is removed the material is continuous across the joint. Usually, to compensate for gaps between the joint faces, a ller of similar material is melted into the weld. There are two other similar processes, soldering and brazing, in which the components being joined are not melted but are made hot enough to fuse a softer material such as tin or brass (note brazing uses brass) which runs between them and acts like an adhesive. Welding is a very useful solution to rather too many design problems in the sense that the designer may take the easy road out of his own difculties and specify welding in place of some other process, thereby passing the problem to his colleagues in the production departments.Their difculties lie in the fact that welding requires highly skilled labour and is extremely difcult to inspect for deciencies. More advanced welding techniques using lasers or electron beams are interesting for some specialised work, but from the structures viewpoint welding is not a process to be used too freely. Welding can signicantly change the material properties in the heat-affected area, not to mention introduce residual stresses and distortion. These problems can be minimised by the use of MIG or TIG welding techniques and further overcome by re-heat treating and/or stress relieving after welding but this all adds to the cost and complexity of the exercise.
7.3.2 Bonded joints

Adhesive bonding of metal structures is widely used by some aircraft manufacturers. All designers of airframe structures use adhesives for glass-reinforced plastic components and sandwich structures in various combinations of wood, paper, plastic and metal, but in the bonding of stringers to skins and other metal to metal joints in primary structure it is not so widely used. The de Havilland Comet and the Fokker F.27, products of the early 1950s, both used Redux bonding in their main structures. In both aircraft it was entirely successful and reliable and the two manufacturers continued to use the techniques in their later designs. One structural disadvantage of riveted construction is that every rivet hole is potentially the start of a tear or a crack, so the apparently simple drilled hole in fact requires care in its preparation, and an additional reserve of strength in its surrounding structure, with the consequent weight disadvantage. Bonded structures go a long way towards removing this particular problem and in many ways they are effectively similar to structures machined from solid stock material. Possibly even more effective is a mixture of machining and bonding to make the type of structure
120
shown in Figs. 9.10 and 9.11.A further major advantage with bonded structure is the excellent exterior nish produced. There are some manufacturing problems. Cleanliness before bonding, application of pressure and heat during bonding and inspection of the joint after curing of the adhesive are all difcult. The provision of the necessary facilities should therefore be a company decision rather than a design ofce dictate.
7.3.3 Bolted and riveted joints

Some of the design aspects of bolted and riveted joints are discussed in Chapter 9. In this paragraph the patterns and types of rivets and bolts and some special fasteners are shown, together with notes on the methods of identifying various standard hardware items for design and drawing ofce purposes. Various shapes of rivet and bolt heads are shown in Fig. 7.8 and the name description for a particular item would be in the following pattern:
5 Rivet, 32 diam., csk head
(read: rivet ve thirty two diameter, countersunk)
Bolts are often named by their thread form: Bolt 1032, hex. head (read: bolt ten thirty two hexagon head)
These names are only a general indication of the part, and the positive identication for specication or call-out purposes is by a unique number. These numbers are in an even more bewildering profusion than the material specication numbers referred to in Chapter 6. Table 7.1 lists some of the major prexes, and although we shall be referring to bolts and rivets the systems also embrace washers, hinges, clevis pins, etc., and many other items of hardware.
Table 7.1 Prexes used for component identity numbers Name of standards group or bureau American: Air ForceNavy Aeronautical Standards National Aircraft Standards Military Aeronautical Standards British: British Standard Aircraft Part Society of British Aerospace Companies Standard Aircraft General Standard Prexes for item numbers AN NAS MS SP A AS AGS
Processes
121
Fig. 7.8
Some general information on fasteners (Courtesy of Kaynar (UK) Ltd.).
122
The prex is followed by a number which identies the style of bolt or rivet, that is its head shape or its particular function. In some cases the number also refers to size but this particular practice is being dropped in the more modern systems. Examples: (Amer.) (Amer.) (Brit.) (Brit.) (Amer.) AN4 AN365 A102 SP85 MS27039 refers to a 0.25 in. diam. hexagonal head bolt refers to hexagonal nuts (but no particular size) refers to hexagonal head bolts refers to mushroom head rivets refers to pan head screws
The third part of the number completes the unique reference, so that dependent upon the amount of information given by the number so far, the third part must indicate at least:
diameter (and/or thread form) material length

Examples:
(Amer.) AN4CH10
1 is a 28 UNF threaded 4 hexagonal head bolt made of corrosion-resistant steel with a head and shank drilled for locking wire 9 with a grip length 16 in. (0.563 in.)
(AN4) (C) (H) (10) (A102) (5) (E)
(Brit.)
A102.5E
is a hexagonal head bolt made of high-tensile steel with a grip length of 0.5 in. 1 and 28 UNF thread diam. 4 is a mushroom head rivet made of aluminium alloy to specication BS.L86 1 with a diameter of in. (0.125 in.) 8 5 and a length of in. (0.625 in.) 8 is a countersunk head rivet made of aluminium alloy 2024T4 3 of 32 in. diameter (0.094 in.) 3 and in. long (0.75 in.) 4
(Brit.)
SP85.410
(SP85) (4) (10) (MS20426) (DD) (3) (12)
(Amer.)
MS20426DD312
It can be seen that the standards organisations have devised some complex number codes, but with practice the systems become more readily understandable. For British readers, B.A.S. (aircraft components)
Processes
123
Ltd publish a concise catalogue, and for American readers, information from Litton Fastening Systems Inc. is equally helpful.

All aircraft parts are identied by a unique part number. If the design of the part changes to the extent that the t, form or function are affected then the new version must be given a new part number. Sheet metal parts are made by any combination of cutting, folding, pressing and drilling. Thicker metal components made from large blocks (or billets) of material are gradually formed to shape by using machines such as lathes or mills which remove a small quantity of material at a time. Certain aluminium or steel alloys can be joined by welding. Whilst welding is a most useful technique it has some drawbacks in that it can change the material properties and introduce residual stresses and distortion. Bonding does away with stress-raising rivet and bolt holes and allows defect-free external surfaces. Process control should be tightly controlled as it is almost impossible to effectively inspect a bonded joint in the traditional manner. The range of fastener types available is huge and each has its particular set of qualities and features.
7.5 References
B.A.S. (aircraft components) Ltd Cramptons Road Sevenoaks Kent TN14 5DS Litton Fastening Systems, Inc. 3969, Paramount Blvd Lakewood, CA 90712
Chapter 8
Corrosion and Protective Treatments
8.1 Nature of corrosion

Almost all construction materials seem to deteriorate. Buildings made by the Romans two thousand years ago have almost all disappeared, worn away by a natural deterioration which nobody troubled to prevent. Most materials try to go back to the form they were in before they became building material. There are exceptions like gold, glass and some plastics but aircraft materials in general either rot or corrode. The earliest aircraft only needed to have short lives; either they crashed or they became obsolete before they had time to wear out. The rst deterioration troubles became evident in the 1920s in wooden aircraft when water found its way into glued joints and lay trapped in undrained corners. Indeed a major cause of the aircraft industrys swing to all-metal aircraft was an expectation of longer life. In general metal aircraft are free from deterioration by corrosion but only because their designers and makers are aware of the problems involved and have taken a great deal of care to avoid them. All metals can be chemically attacked and changed into salts of one type or another. The most familiar change of this type is the change of steel into rust, which is iron oxide. Aluminium corrodes into a white powder. Some metals are recognised as being resistant to attack but by that we usually mean atmospheric attack and, consequently, some metals, like stainless steel, can catch us out and give unexpected problems. Although there are many different ways in which metal can be attacked, for all practical purposes aircraft metals are damaged by an electrochemical process which takes place between the metal and another substance in the presence of an electrolyte. This happens so easily and the other substances are so common that really all that is needed to set up some corrosion is the electrolyte water in small amounts is an excellent conductor. Having said that, we must make it clear that the problem varies in intensity. There is always trouble when two different metals are close to one another and we shall consider this situation in more detail below. Aircraft operating close to the sea have very severe problems. Naval aircraft and oil-rig helicopters require special protection, and the owners of the few remaining ying boats carry on a running battle against the effects
125
of salt water which assist the metal in combining with atmospheric oxygen very quickly. Electrochemical, electrolytic or galvanic corrosion is basically an action similar to that which takes place in an electric battery. Metal, electrolyte and another substance (which may be another metal) form an electric cell; current ows and chemical changes take place, especially on the anode material, which is the one with the higher electropotential. In an assembly, if dampness penetrated between an unprotected steel bolt and an aluminium alloy tting, an electrochemical action would be set up and some of the aluminium would be changed to aluminium oxide. This cell action can take place over very small areas of the same piece of metal, as in the case of direct chemical attack (by acid or strong alkali), and it can even take place between the constituent metals of an alloy. Generally, corrosion needs a supply of oxygen and surface oxidisation is the common type of deterioration which we know well and know how to combat. Other forms of corrosion are fairly obvious, like the deterioration of aluminium into aluminium sulphate under the action of battery acid. A form of corrosion which is extremely difcult to counter is that form called intergranular corrosion and its associated form, stress corrosion. As its name implies, intergranular corrosion attacks the interior of the metal, effectively treating each grain as though it were a separate component and spoiling its surface. This happens in spite of lack of oxygen, in an action which does not seem to be very well understood. It certainly is understood that the action is accelerated when components are under load, including the types of load induced by the manufacturing processes or heat treatment and known as residual stress. Unfortunately, the corrosion has a knock-on type of effect in that the internal damage in the material is very similar to cracking, with the attendant sharp ends which are stress raisers, propagating in turn more cracks. As with so many problems in the aircraft industry, the causes are no worse than they are for other industries, but the effects are worse because we always have to push our materials, structures and indeed all our engineering right to the limit of their capabilities. Consequently, when corrosion attacks a component it is not eating into fat, it is eating into the bone. Apart from intergranular corrosion, which requires specialist equipment and expertise for its detection, most other corrosion is fairly easy to see (provided that it is not attacking a piece of structure which is hidden away from sight). Some corrosion, like rust or copper oxide, is coloured and makes telltale marks or stains. Surface corrosion has the advantage (if such a word can be used in connection with corrosion) that the oxide formed has a much greater volume than the metal from which it was formed. This phenomenon is well known to the owners of elderly motor cars when the product of corrosion under paintwork pushes up a paint blister out of all proportion to the size of the pitting in the metal. Similarly, when aluminium corrodes, it produces a lot of powder which looks much worse than the damage justies. None the less, the damage is there and advice must be sought from the aircraft manufacturer (via the main-
126
tenance manual) about how much damage is acceptable in a particular area before a component is scrapped. Strangely, although aluminium oxide is a pale grey to white powder quite unlike new aluminium, as a structure gets older and any internal paint starts to ake a little, the oxide is harder to see, or perhaps easier to overlook, than one would imagine.
8.2 Causes of corrosion

In a sense the primary cause of corrosion is failure to provide protection from atmospheric attack. However, this is a fairly obvious source of trouble and can be avoided most of the time, but there can be catches. For instance in a company where the usual practice is to use Alclad material (see Section 8.3), a sudden requirement which calls for the use of unclad sheet can lead people to leave the interiors of closed structures without protection. Some years ago manufacturers of crew and passenger seats were very fond of making their structures of welded tube, which is virtually impossible to protect on the inside (see Section 8.3.6). Unfortunately, where a structure which needs paint is left bare it is usually because the painter cannot reach it and this invariably also means that it cannot be seen, so any eventual corrosion is likely to remain undetected. Another prime cause of corrosion must be failure, during routine maintenance, to carry out the manufacturers instructions on protection; but this is a quality control matter (see Chapter 11) and just outside the scope of this book. The corrosion giving the most trouble to structure designers arises from contact between dissimilar metals. The situation is that one metal corrodes the other, but within certain very tight limits this can be acceptable, or at any rate tolerated. The existence of a degree of tolerance does sometimes allow a situation in which it is possible to interpose a metal separator between two non-compatible materials such that the middle material avoids seriously damaging either of the others. It is by this method that we can accept cadmium-plated steel fasteners in light-alloy structures, but in most cases it is better to separate dissimilar metals with an insulator (plastic sheet, Tufnol or Paxalin preferably, or even jointing compound if the problem is only marginal). One of the traps for the unwary is the stainless-steel (CRES) fastener which may need plating to avoid an unacceptable level of potential corrosion against some light alloys. Even different aluminium alloys or the same alloy at different tempers can cause trouble, although seldom so much that it cannot be avoided by the use of jointing compound. On the other hand, ying boats and oil-rig helicopters sometimes suffer corrosion between the light-alloy skins and stringers, to the point where the growth of the aluminium oxide is sufcient to pull rivet heads through countersinks in the skin. Stress corrosion was mentioned in Section 8.1 as being a form of intergranular corrosion and this seems to affect the high-strength zinc-bearing
127
light alloys more than some of the others. Following a remark earlier in this book (see Section 6.2.3) the reader will sense little enthusiasm for high-zinc alloys; however, it is necessary to use them at times to save weight. In at sheets or gently curved panels they are satisfactory, but in any situation where bending or heat treatment could possibly produce built-in unrelieved stresses they are prime candidates for stress corrosion. Another situation in which stressed metal can be attacked by intergranular corrosion is created under bolt heads which have been overtightened. Designers should make sure that the workshops know the permissible torque loadings for fasteners in sensitive areas. All riveted and bolted joints are liable to fretting corrosion (faying) which results from constant small relative movements between clamped faces. If the joint is in a structure there is almost certainly a load across the joint. If there is a load there is a movement (no stress without strain), so solutions to the fretting problem which involve clamping bolts up tighter are only wishful thinking. The solution is to separate the metal faces with jointing compound on permanent joints or with anti-faying lubricant on other joints. The last type of corrosion which is particularly relevant to conventional aircraft structure is shielding corrosion. It appears where wood, furnishing materials or sound-proong hold a wet patch in contact with metal, and may at rst glance seem similar to the type of corrosive situation in which areas of structure are compelled to lie in accumulated water. (It is not possible for instance to put drainage holes in the bottom of a pressurised fuselage, and toilet and galley spillage, plus condensation, can make a very corrosive combination.) This latter situation can easily be recognised and guarded against, unlike shielding corrosion in which the rate of corrosive attack is in fact greater than would be expected if the metal was totally immersed in water. Again there is a possible trap when using stainless steel in this situation. The particular property which makes stainless steel resistant to corrosion is an ability to instantly form for itself a thin protective skin. Unfortunately, this skin only forms in a well-ventilated situation. So in a situation of the type described above, the wet material breaks the surface skin of the stainless steel which then nds itself unable to immediately repair itself, therefore the steel is no longer stainless. Some CRES alloys in this situation can completely corrode away faster than unprotected mild steel. The only satisfactory solution is to insulate the metal from the water carrier in just the same way as one would insulate it from a dissimilar metal. Even then there is still a snare which has deluded some eminent designers (especially in the automotive industry). In an effort to reduce maintenance down time, the dry lubricated bearing is a tempting proposition, but there are problems. Stainless shafts rotating in turned nylon bushes (moulded bushes with a very smooth bore are better) can suffer from shielding corrosion with the result that the shaft swells by corrosion and starts to squeak and bind.
128
8.3 Protection against corrosion

The classic and only satisfactory way of stopping a metal from returning to its earth is to cover it; either paint it, plate it or allow it to produce its own oxide covering without disturbance. The last of these methods is the most debatable so we will discuss it rst.
8.3.1 Oxide lms

Most constructional metals and alloys start to corrode on their surface, and the initial oxide lms so formed cover the parent metal and tend to inhibit it from further attack. With some alloys this action can provide sufcient protection on its own, as in the case of stainless steel (CRES) and titanium. Most oxide lms are hard and tough but some are very thin and can be broken. Usually this is no cause for worry because a new protective skin immediately forms. If, however, a condition exists which prevents this, more serious corrosion may result. Designers who are relying on this form of protection should also ensure that components are well nished (polished), because small surface imperfections can enlarge into pits, which in turn become stress raisers. The major material of aircraft construction, aluminium, has, at least in its unalloyed state, a (comparatively) thick and stable or passive natural skin. It is so patient and gives so little trouble that we have to be careful not to take it too much for granted. In general the higher the strength of the alloy the more it is susceptible to corrosion, but even this problem is effectively defeated by the Alclad process. This elegantly simple but highly inventive idea is to coat the vulnerable high-strength sheet with a very thin surface of non-vulnerable low-strength pure aluminium. The coating is hot rolled on in the manufacturing process and forms an absolutely unpeelable bond. Some major aircraft constructors are satised that structures made from Alclad sheet are sufciently protected, at least internally. Since other manufacturers paint everything (see below), there is obviously room for sensible debate. Unexpectedly, Alclad sheet is less vulnerable at cut edges and drilled holes than unclad sheet. This bonus is due to sacricial protection by the unstressed aluminium in favour of the stressed aluminium alloy. This takes place as galvanic corrosion, acceptable because the loss of the soft pure aluminium takes place slowly and with no loss of strength to the structure. So although the Alclad process is in a very real sense a covering or coating protection, most of its efciency derives from the self-healing passive oxide lm on the aluminium.
8.3.2 Cadmium plating

One primary protection for steel parts is cadmium plating. This is a metallic electrodeposition process like decorative chrome plating or domestic silver plating on cutlery. As protection, cadmium plating works in at least three ways. Firstly, cadmium is a passive metal and offers good resistance
129
to atmospheric attack. For this property alone it was considered a suitable protection when some vital steel structure components were exposed to the elements. Biplane tie wires were often left cadmium plated but unpainted. Secondly, cadmium plating acts as a sacricial anode in a similar way to that described for aluminium cladding (Section 8.3.2). This means that cadmium-plated components are not sensitive to minor surface damage even when the plating is scratched right through. Thirdly, the cadmium acts as an intermediary between steel and aluminium as it has a level of electropotenital acceptable to both. Cadmium plating is a very thin lm. Depending upon the exact application the lm thickness usually varies between 0.0002 and 0.0006 in. (i.e. 0.005 to 0.015 mm) but as this is of the order of 0.001 in. on the diameter of bolts, etc., draughtsmen need to be careful when specifying manufacturing tolerances to state whether their dimensions are before or after plating. The British specication for the whole process in DEF STAN 03-19 and the American specication is QQP416; it is very much the responsibility of the designer to know what these specications are giving him. In terms of the British specication it is, strictly speaking, necessary to stipulate a pre-treatment DTD 901 (degreasing), plus any subsequent treatments such as DTD 934 (which applies to high-tensile and stainless steels), de-embrittlement which applies to springs, and passivation which applies fairly generally, and certainly when there may be a painting operation. To assist draughtsmen, most design organisations provide a shorthand note, or some term which avoids repetitive delineation of exact processes. This does not, however, relieve the draughtsman of the responsibility of ensuring that he is asking for the correct processes. There is a principle here which applies over the whole eld of aircraft and aircraft equipment design, and we can use the shorthand note Cad and Pass to illustrate its importance. By putting such a note on his drawing, the draughtsman is stipulating a process: Cadmium plate to DEFSTAN 03-19 preceded by . . . etc., through to . . . and passivate, and provided that he is prepared to accept the decision of the process department on the standard required, then his shorthand note is sufcient. If, however, he is specifying a Class 1 component (see Section 5.8) then he must ensure that the process department understands that his note means nothing less than the full treatment. On the other hand, if the component involved is relatively insignicant, it would be wrong to waste ofce time, workshop cost and inspection effort on providing more than the minimum standard of nish. The information that the draughtsman is communicating should make this quite clear. This general philosophy should apply to all the engineering departments work. Where quality requirements demand the best, the designer must be assured that he is going to get it. However, if there is no justication for expensive materials or processes or labour, then he should be equally certain that the draughtsman is not guilty of over-engineering the design.
130
8.3.3 Anodising
The standard nish for aluminium alloy components made of anything other than Alclad sheet is anodising, or more exactly, anodic oxidation. The current British specication reference for anodising is DEFSTAN 0324 and the American specication is MIL-A-8625. Anodising articially produces the tough oxide lm on aluminium alloy which, as we already know from Section 8.3.1, forms naturally on pure aluminium. As far as Alclad is concerned, there is no reason why it should not be anodised, but the process is pointless unless the material is to be subsequently bonded or subjected to a similar operation, for which anodising provides an excellent preparation. There are various processes under the general heading of anodising, including decorative nishes, and an abrasion-resistant hard anodising process, but for structural parts the chromic acid method is most widely used. This process, however, produces only a thin oxide lm which is porous and easily damaged, and must therefore be sealed either by rinsing in boiling water or by being immediately painted (or at least primed, see below). Neither action is necessary if the part is to be bonded. The remarks made above, concerning the need to specify exactly what is required by the design, apply equally here. The draughtsman is in the business of communicating the designers wishes to the workshop and he must make sure here, as much as in the more glamorous areas of aircraft design, that what he puts on paper carries the right message. Some authorities suggest that it is possible to anodise welded or even riveted assemblies. The chromic acid process has the advantage that it shows up cracks in welds remarkably well, but even so the danger of trapping the uids used in the process is too great. The answer with riveted assemblies is to anodise or paint the components before riveting. The answer with welded parts is simply to avoid welding altogether; there is usually an alternative method that will give less trouble in manufacture, quality control and service. (There are exceptions to this heavy dogma. Some automatic or machinecontrolled welding and some electron-beam or laser-beam welding may justify its cost for some jobs, otherwise welding belongs with steel tube structures back in history.)
8.3.4 Protective pastes

An alternative to the full electrochemical immersion anodising process is the application of one of a number of proprietary pastes, such as Alocrom, which produce a protective lm. These are especially useful on assemblies (or repairs) when the original anodising has been damaged, perhaps by some machining process which could not be completed before assembly. However, these treatments must be used with caution as some, which at rst sight appear entirely satisfactory, have to be painted over immediately. (The golden rule is to read the instructions on the packet.)
131
8.3.5 Painting
There are a multitude of protective treatments which may apply to aircraft in general but the only remaining treatment which must be mentioned in relation to aircraft structure is paint. Even here there is a great variety of alternative nishes: cellulose, stove enamel, epoxy, acrylic, polyurethane with, no doubt, new ones appearing regularly. Paints are the universal protectors of structures and with the possible exception of parts made of Alclad sheet, structures need it. Paints usually contain inhibitors intended to ward off atmospheric corrosion chemically but essentially they provide a decorative covering less susceptible to deterioration than the structure, easier to replace when damaged and not contributing to strength. When thinking of paint it is necessary to think of a whole paint scheme with three or two compatible elements: (a) The primer, which is the protector for the material and is made to adhere to the material surface. (b) The undercoat, which is the opaque colour. (c) The top coat, a transparent decorative nish which also is hard or tough enough to protect the primer. Some aircraft schemes have been designed to save weight by dispensing with element (b). When selecting a scheme, check that if the paint is heat cured the heat can be applied safely and that the temperature does not approach the heat-treatment temperature of the material being painted. Primers for metal surfaces are usually etch primers. The most universal contain zinc chromate, which gives treated parts a distinctive yellow or greenyellow nish, which for some reason always looks stable and reliable. For interior parts which do not require a decorative nish, a good primer may be sufcient in itself. Although some etch primers (especially the two-part synthetic resin types) are very tolerant of poor pre-priming surface preparation, degreasing by vapour bath for light alloy or mechanical abrading (sand blasting) for steel is almost essential. Decorative exterior paint schemes usually specify their own particular pre-treatment. Some exterior paints are specically applied to protect the strength of the structure from erosion by stone chippings thrown up from the runway or ice particles in the air. As with all other protective treatments, specication is the responsibility of the design ofce and must be complete.
8.3.6 Unprotected components

There remain the components or parts of components which cannot be protected. The inside surfaces of tubular structures and the inside of holes reamed on assembly are typical examples. Unfortunately, most manuals and handbooks are remarkably unhelpful on this subject and regrettably
132
this book is no exception. Perhaps the most practical advice is to check that the part really cannot be reached. Maybe it can be reached by being dismantled after a trial tting. In another case a coating of lanolin applied by any available method may be better than nothing, but the real answer is that if a part cannot be protected then the design is wrong.

Due to the very short expected service lives of early aircraft, corrosion protection was not an important issue for the early manufacturers. Virtually all materials used in aircraft construction, composites included, are subject to some form of degradation associated with time in service. It is the designers job to understand the degradation process for the material being used and so specify the protection method which best retards that degradation. Metal corrosion results from a galvanic process similar to that which occurs within a battery. Contact between dissimilar metals will greatly aggravate the problem, especially if water is present. Composite materials generally suffer degradation when exposed to ultra-violet light (as found in sunlight) and can also suffer in the presence of heat and moisture. Fretting corrosion arises when there is relative movement between adjacent parts, usually in joints. Applying a jointing compound during assembly will help greatly by lling any small voids within the joint and thus eliminating moisture. It will come as no surprise to nd that the principal techniques used to protect structures are aimed at keeping dissimilar metals from touching whilst also eliminating moisture. Some materials, such as aluminium and stainless steel, helpfully tend to produce their own, non-porous protective lms without any assistance from us. Anodising is the method used for achieving this protective lm on aluminium. The most widespread form of corrosion protection remains painting with a good quality epoxy primer and nishing paint. Lanolin (a waxy coating) is sometimes applied to internal surfaces, such as the insides of metal tubes, after painting.
8.5 Reference
CAP 562, Civil Aircraft Airworthiness Information and Procedures Westward Digital Ltd 37 Windsor Street Cheltenham Glos. GL52 2DG Tel: 01242 235151
Chapter 9
Detail Design
9.1 Sheet-metal components

The essence of detail design is understanding the limitations of the materials being used. Although an increasingly large number of aircraft structural components are now machined from solid billets or thick plate, sheet aluminium alloy is still the major material of aircraft construction. Figures 9.19.5 show some typical components and it will be seen that apart from the skin panels, almost all the items are bent or formed in some way. When bending or folding the structural alloys it is necessary to make a radius in the corner as shown in Fig. 9.6. The actual radius required depends on the material and its heat-treatment condition (see Chapter 6). Within half an hour of normalising, most alloy can be bent to an inside bend radius equal to 2t (that is, twice the material thickness) but higher strength alloys may require a 5t internal radius or even more if the bend angle exceeds 90. The penalty for trying to reduce the bend radius recommended by the material supplier is cracking or splitting of the metal on the outside of the bend. Unfortunately, the requirement for a bend radius brings problems in its train which we shall discuss below and illustrate later in Fig. 9.15. Bending sheet metal along a straight edge (also called folding) is a comparatively straightforward process but any variation from this simple bending causes manufacturing problems which are disproportionately severe. Most of the diagrams 9.1 to 9.5 show bends which have joggles (see Fig. 9.4) and it is almost impossible to avoid their use in sheet-metal structures but, in addition to some reservations about their strength, they are very difcult to make. The two small brackets in Fig. 9.2 are a case in point; the smaller of the two is a simple job for the folding machine but the larger, being joggled, would require a special press tool and longer production labour time. These differences tend to disappear with larger quantities, but for small production runs joggling should be avoided where at all possible. Bending anges on curved members is also difcult. Figure 9.7 shows a curved channel member. A component of this type would be made by pressing or hand beating a at development over a block or tool specially made to the correct shape. Edge AA on the diagram is stretched during
134
Fig. 9.1
Examples of detail design, BAe 748. (Courtesy of British Aerospace.)
Detail Design
135
Fig. 9.2
Typical sheet-metal construction.
the forming process and can be made smooth, but the difculty is that edge BB has to shrink and tends to fall into waves which are difcult to eliminate. Again, bends of this type cannot be completely avoided but designers try to nd alternatives where they can. The need for a substantial bend radius makes for difculty in the corners of a certain type of sheet-metal component or tting. Fig. 9.8(a) illustrates the point and (b) and (c) show the effect on the development of the corner. (Developing a sheet metal part simply means laying it out at.) For reasons which are discussed in Section 9.3, the shaded area in (b) cannot be cut out to leave a sharp corner at P; nevertheless some companies accept that development (b) is satisfactory while other companies require solution (c). Figure 9.8 shows at (d) and (e) two design situations which crop up from time to time. Consideration of the developed shape will show that (d) is an impossible problem. Shape (e) is not impossible, but it is very difcult to make and should be avoided. The more complex shapes in sheet metal are produced by pressing, which was mentioned in Chapter 7. The anged holes in Fig. 9.2 and the components with curved edges in Fig. 9.1 would all be press formed. The methods of production are very versatile but designers have to consider how their components will be made and, as with all components, simplicity in the design of pressings is a virtue.
136
9.2 Machined components and large forgings

Some aircraft structural components, such as undercarriages, are made almost exclusively of components machined out of solid bar material or forged blocks. This book is more concerned with the larger components, such as fuselages and wings, and in recent years the percentage of such structures machined from solid stock has increased substantially. One of the major advantages of machined structure is the reduction of notching effects, which will be discussed in their relationship to riveted structure in Section 9.3. (Notching is the introduction of an abrupt change of section due to drilled holes, grooves, sharp-edge undercuts, etc.) Figure 9.5 and Figs. 9.99.12 illustrate some typical machined structural parts. It can be seen in Fig. 9.5. that there can be an immediate weight saving compared with riveted structures because of the elimination of overlaps at joints and stringer attachments. More careful and thoughtful design using adhesive bonding (as shown in Figs. 9.9 and 9.11) produces even more advantages. Savings on cost of materials, reduced machining costs by the use of extrusions, and the option of using very high strength alloys in selected locations are added to the more obvious advantages mentioned above. As with almost all design options in aircraft engineering, there are advantages and disadvantages. Repairs to machined structural components are much more difcult than repairs to sheet-metal structures; initial tooling costs are very high; capital investment in machinery is very high; the cost of representative prototype test pieces is high and mitigates against risking early test failures, which in turn suggests heavier than necessary structure. It is probable that for the next few years the larger companies dealing with advanced designs will increase their use of machined structures, while smaller aircraft companies facing commercial pressure for a lower initial selling price will continue to use the more traditional methods of construction.
Fig. 9.3 (a) Typical constructional details (Courtesy of The de Havilland Aircraft of Canada Ltd); (b) Boeing 747 (Courtesy of The Boeing Commercial Airplane Company); (c) BAe 146 (Courtesy of British Aerospace).
Detail Design
137
Fig. 9.3
Cont.
138
Fig. 9.4
Joggles.
Detail Design
139
Fig. 9.5
Structural parts.
9.3 Notching and stress raisers

Figure 9.13 shows some examples of what is meant by notching, which is perhaps the most dangerous trap of all for the aircraft structural designer. Apart from situations where ground staff have not fastened doors correctly, etc., and at some risk of sounding melodramatic, it is difcult to think of a structural failure leading to an aircraft accident in recent years
140
Fig. 9.6
Sheet metal bending.
Fig. 9.7
Sheet metal forming.
that was not initiated by a notch effect which might have been avoided by better design. That is not to suggest undue criticism of the particular designers; all designers get caught by the ease with which unintentional stress raisers can be built in, and those who recognise and retrieve the situation before an expensive failure occurs know that they have been very lucky indeed. The essential feature of notching is that because it is notched, a component is disproportionately more liable to breakage. The most common example of the phenomenon is its exploitation for glass cutting. Glass sheet which has been scored with a very light groove will break at a small fraction of the load which would be required to break unmarked sheet.
Detail Design
141
Fig. 9.8
Sheet-metal corners.
142
Fig. 9.9
Machined structural details. (Courtesy of British Aerospace.)
Fig. 9.10 Constructional details with machined components. (Courtesy of British Aerospace.)
Detail Design
143
Fig. 9.11 Machined and bonded structure. (Courtesy of Ciba-Geigy, Bonded Structures Division, Cambridge.)
In Fig. 9.13(b) the maximum stress in the material near the hole is much higher than the average stress which would be calculated by dividing the load P by the area shown shaded. Figure 9.13(d) indicates the situation by showing the lines which represent an even stress distribution suddenly crowded together at the groove. A very common aircraft structure situation is shown in Fig. 9.13(c). In 1 this case, if the rivet holes are 3.2 mm ( in.) in diameter and the rivets are 8 spaced 25 mm (1.0 in.) apart, the actual maximum stress in the sheet will be approximately 2.6 times the nominal stress which would be expected from dividing the load by the sheet thickness and length. Unfortunately, neither structures nor machines can be designed entirely without notches. The best we can do is to recognise them, appreciate the type of effect they have and make allowances for them by lowering the nominal stress level. In the two examples shown in Fig. 9.13(b) and (c) we would probably have to make the bar and the skin thicker. (It should be noted here that using stronger light-alloy material may not provide an answer because strong materials are often more susceptible to notching effects than softer, more ductile materials.)
144
The examples shown in Fig. 9.13(f) illustrate simple precautions which must always be taken. Example (e) is also easily avoided (see Fig. 9.8). Other problems which may crop up in a design will need their own solutions.
9.4 Rivets and bolts

Rivets and bolts are part of a general group called fasteners. Under the same heading there is a large range of proprietary fasteners which offer particular advantages for special purposes, such as shear strength above the normal to be expected from a rivet, but with a lower installed weight than a bolt. Another special fastener acts in a similar way to a lightly loaded bolt, but is designed to be released by only a 90 twist with a screwdriver. This last example belongs to a subgroup called quarter-turn fasteners and, although they do not have any relevance to major structure, they are used to retain light fairings, engine cowlings and access panels. Some indication of the range of fasteners is shown in Fig. 9.14 and notes on the methods of reference by number are given in Chapter 7. In Chapter 7 reference was made to the grip length of bolts. When specifying bolt length it is important that the smooth shank of the bolt extends right through the hole so that no thread is against a bearing surface, so
Fig. 9.12(ac) Structural components machined from forgings or heavy plate. (Courtesy of British Aerospace.)
Detail Design
145
Fig. 9.12(b)
146
Fig. 9.12(c)
the grip length is the plain length and is approximately equal to the thickness of the materials being fastened together plus the thickness of any washer under the bolt head. For countersunk or at-head bolts the length is measured from the top of the head, and for other bolts from under the head (see Fig. 7.8). In almost all cases, specifying the correct length of bolt shank will result in having a small amount of plain shank exposed past the hole so that a washer or washers will be needed under the nut to allow tightening. When specifying the length of a rivet, allowance must be made for forming a head. This allowance is about one and a half times the nominal rivet diameter but the design ofces of the aircraft manufacturers have their own standards for this important dimension. In a normal structural use of bolts the fastening is completed with a nut. All so-called self-locking nuts (see Fig. 7.8) are better referred to as stiff nuts and (for structures) must not be used singly; that is, the minimum number of bolts and stiff nuts in a group must not be less than three. For some applications nuts are wire locked as shown in Fig. 7.8. If bolts are used through ttings which are unlikely to be removed for checking during the service life of the aircraft then serious consideration should be given to the use of lighter fasteners of the type produced by the HiShear Corporation (see Fig. 9.14). In addition to the solid rivets, there are various types of rivets which can be placed without access to both sides of the structure. These are referred to under the general heading of blind rivets. An indication of the general principles is shown in Fig. 9.14. The method for using almost all the blind rivets involves having a hole right down the length of the shank
Detail Design
147
Fig. 9.13
Notches.
148
Chobert (a) A steel mandrel which has an opposite taper on the head is drawn through from the tail end of the rivet, expanding the rivet tail around the rear side of the hole forming a shoulder. (b and c) The mandrel continues to be pulled through the rivet, symmetrically expanding the rivet shank to fill the hole. (d) This ensures the rivet has good bearing in the hole and a parallel bore is left in the rivet. The cycle is completed and the next rivet is ready to be placed into the prepared hole.
How the Avdel MBC Rivet system works 1. The MBC Rivet is loaded into the nose of the placing tool, and the tool applied to a prepared hole in the workpiece. 2. When the tool is actuated, jaws in the nose of the tool grip the rivet stem and exert an axial pull, drawing the stem through the rivet shell to give a high-clench joint and complete hole-fill. 3. The placing tool automatically shears the stem flush with the rivet head, and is mechanically locked into the rivet shell.
Fig. 9.14 Some proprietary fasteners. (Courtesy of Avdel Ltd (a and b), HiShear Corporation (c), Dzus Fastener, Europe Ltd (d), Cherry Rivet (e), Huck Fasteners (f).)
Detail Design
149
Fig. 9.14
Cont.
150
and with some types, such as the Chobert, this hole remains after the rivet is installed. In other types the hole is plugged automatically as part of the rivet-setting process. All the various patterns have their own advantages and there is a great deal of commercial competition in this market. For discussion purposes, and since we have referred above to Chobert rivets, the companion self-plugging rivet from the same manufacturer is the Avdel MBC. Chobert are for light duties, comparatively cheap to buy, very cheap (in labour terms) to place and are produced in several different materials for different purposes. Avdel MBC are much more expensive to buy and to place than Chobert but are very strong (at least as strong as solid rivets) and apart from the facility of being able to place them from one side of a structure, the fact of mechanical installation means that vulnerable surfaces such as wing skins are not damaged by the hammering which is required to set up the strong types of solid rivet. Wherever possible, aircraft structures are designed so that bolts and rivets are used in shear. Some companies allow rivets to be used where they may be subjected to a small amount of tension but, in general, fasteners in tension must be bolts. The exception to this rule is the use of rivets for attaching the skin. On the top surface of wings, where there is negative pressure, and on the skin of pressurised fuselages, where there is positive internal pressure, the skin rivets are in tension. The justication for this bending of the rules is that there are a lot of rivets required to hold the skin to the structure members and carry the shears mentioned before so the tension in each rivet is small. Figure 9.15 and other illustrations in this book show the way fasteners are set out in aircraft structures. Some particular points in Fig. 9.15 should be noted. The distance A in Fig. 9.15(b) must be sufcient to allow a riveting tool to have access above and below unless a blind rivet is being installed from below. As a guide, for 3.2 mm (1/8 in.) diameter solid rivets in 1.0 mm (0.04 in.) thick sheet, dimension A should be at least 7.6 mm (0.3 in.) and the edge distance should be 6.4 mm (0.25 in.). For blind rivets installed from below, A could be reduced to 6.4 mm (0.25 in.). The above dimensions should all be increased by 1.3 mm (0.05 in.) for 4.0 mm and 4.8 mm (5/32 and 3/16 in.) diameter rivets but these dimensions are sufciently important that draughtsmen must check the standard practice in their particular company before quoting gures. Unsuitable edge distance allowances by the designer can cause quality control problems, e.g. a rivet hole positional tolerance of +/-0.5 mm (0.02 in.) could result in a 6.5% strength loss in a 7.6 mm (0.3 in.) edge and a 10% loss in a 5.0 mm (0.2 in.) edge, but a 7.6 mm edge is, by denition, 50% heavier than a 5 mm edge. Some of these arguments apply to Fig. 9.15(c). In this case dimension A must be determined by the clearance needed for wrench access. However, because of the load conditions indicated by the arrows, the edge distance must not be less than dimension A. Even if the two dimensions are equal, the tension load in the bolt will be twice the load P in the angle.
Detail Design
151
Fig. 9.15
Dimensions of riveted and bolted joints. ((a) Courtesy of British Aerospace.)
152
The use of the prole washer is necessary to avoid bending the ange of the angle, but it also helps keep A to a minimum. The basic loading of bolts (or rivets) in a shear joint is shown in Fig. 9.16(a). Ignoring any design criticism of the whole joint (such as the possibility of distortion putting the bolt in tension as shown in (b)) we can see that the shear load (see Section 5.5.2) in the bolt is P. Similarly, in Fig. 9.16(c) we have doubled the load but shared it between two bolts so the shear in each bolt is again P. Now consider (d). In this case we have increased the load three times and increased the number of bolts to three, so that we are tempted to say that the load in each bolt is still P. Unfortunately, this is not necessarily true for the following reason: If we consider (c), the material between the two bolts is carrying a load and is stressed. Because it is stressed it is strained (i.e. it is extended) and because the stress is constant over the whole length between the bolts the extension is even and uniform over the length. Now, if we imagine (c) being kept in a loaded condition while a hole is drilled through both parts to receive the third bolt shown in (d) we can see (because the extension is even) that the holes in both parts will still line up and still accept the bolt after the load is removed. The situation is that the middle bolt can be placed in its hole when the joint is either loaded or unloaded, so it is quite clear that when the load is applied the middle bolt carries no load; i.e. the two outer bolts in (d) each carry 1.5 P. In practice, because of minor distortions, the centre bolt does receive a load (but not a third of the total). Also we can taper the joint plates as shown in Fig. 9.16(e), which has the effect of making the strain between the outer bolts uneven over the whole distance, and so by careful design we can make the loads on the bolts very nearly equal. Although the example above is somewhat over-simplied, the underlying principle is that in a line of bolts, such as shown in Fig. 9.16(f), the loads on the bolts are not equal. Also, in a joint where the plates are bonded together with an adhesive (such as epoxy resin), the shear stress in the adhesive is not constant along the joint but is concentrated at the ends. Students who have become practised in consideration of stress and associated extension will be able to satisfy themselves that the shear distribution diagram shown in Fig. 9.16(g) is representative. (It should be noticed that the diagram must be symmetrical because the joint is symmetrical.)
9.5 Joggling
Joggles are features of the type shown in Fig. 9.4. In sheet-metal structures they are virtually unavoidable. (Figure 9.4(b) shows a method of avoid-
Detail Design
153
Fig. 9.16
Joints.
ance but it is heavy and expensive and only justied where strength or tooling considerations demand drastic solutions.) As far as the strength of joggles is concerned, their problem is illustrated in Fig. 9.4(c), but the difculty is made more severe because, in spite of knowing that this failing exists, the designer can still be persuaded otherwise by the robust appearance of joggled components. If an attempt is
154
made to react the unbalanced load P by hoping that a rivet head may take some tension (Fig. 9.4(d)), and if the angle of the joggle is 45 then the load R is, by triangle of forces (see Fig. 4.3), the same as P. Load R can be reduced by increasing the length of the joggle and hence reducing the angle in the triangle of forces, and this partial solution, carefully used, is satisfactory. The illusion of strength in joggles is most marked when angles are joggled and it disguises the fact that the joint in Fig. 9.4(e) is as poor as that in (f). A better solution is shown in (g).
9.6 Clips or cleats

The terms clips and cleats are used, respectively, in the American and British industries. Typical cleats are shown in Fig. 9.17(a) and (b). All sheet-metal structures use these or similar connections and it must be remembered that
Fig. 9.17
Cleats or clips.
Detail Design
155
they are structural components requiring the same stress analysis as more glamorous components. Figure 9.17(c) illustrates the fault which makes the use of single rivet or bolt connections highly suspect. The reason for not being more denite and saying that such joints are impossible is the slight reservation that if the only load involved was in the direction T then the distortion shown might not occur. These clips are sometimes called shear connections, and if this is true then rivets are adequate. If, however, the load is as shown in Fig. 9.17(c) at NP, a typical pressure normal (or perpendicular) to the skin, then it must be remembered that, as shown in (b), the rivets can be in tension. In (a) the rivet tension is limited by the lack of torsional stiffness of the cleat between the bend and the rst rivet. It should be clear from the above remarks that the design of cleats is not a task to be undertaken lightly.
9.7 Stringer/frame intersections

A primary difculty in the detail design of stressed-skin structures made from sheet metal occurs at the point where the horizontal members (the stringers) cross the vertical members (the frames). Obviously at the point of crossover it is not possible for both members to be continuous. Some solutions to the problem are shown in Fig. 9.3. Figure 9.3(b) shows a particular arrangement from the Boeing 747, and we can see from the very careful shaping of the cut-out in the frame, which allows the stringer to pass through, that the design has warranted, and been given, a great deal of thought. The cut-out is clearly an unavoidable notch in the frame, and the design effort has been directed to minimising the stressraising effects. The stringer clip (or cleat) and the joggling of the stringer have also been carefully designed to avoid the criticisms of Sections 9.5 and 9.6. On an aircraft of the size of the Boeing 747 there are so many of these stringer/frame intersections that a company can justify an expensive design effort aided by computers and physical test programmes devoted to obtaining maximum efciency from the engineering of this joint. Efciency in this context means weight reduction without loss of structural integrity in the frame, and integrity means that the frame will perform all its specication functions without fatigue failure problems reducing the life of the structure. All aircraft manufacturing companies have their own arrangements for this type of crossover joint, and students of design will need to follow the standards set within their company. With machined structures the problem either does not arise or is not so pressing, and just the avoidance of this one factor alone in the design would make the integral machined skin and stringer combination extremely attractive.
156
As structural components such as fuselages or wings are usually tapered, it is necessary to reduce the number of stringers towards the narrow end. Stringers should end at a frame and not in the middle of a panel. If they are stopped without being cleated (clipped) to a frame, movement between the very stiff frame and the stiff stringers, caused either by the ight loads or by vibration, will cause cracking in the relatively exible skin. Although this particular point is a fairly obvious and easily avoided problem, designers have to beware of similar problems in other areas, such as galley structures, freight containers, etc., where a framework is covered with a skin which carries load.
9.8 Lugs
The present normal type of stressed-skin structure does not use tension lugs of the type shown in Fig. 9.18, but they are mentioned here because they are not uncommon in machined components and the ways in which they can break if overloaded are interesting and relevant to some related components, such as the cleats shown in Fig. 9.17.
Fig. 9.18
Lugs.
Detail Design
157
9.9 The stiff path

We shall mention in Chapter 12 that some structures are unavoidably designed so that the applied loads can be carried in more than one way. The example used is the oor panel, which is attached on four sides but which may be quite capable of carrying sufcient load if it was only attached along two opposite sides. In Chapter 12 this is presented as an advantage for the stressman because he may nd by analysing one load path and nding it strong enough on its own, that he does not need to go to the effort of nding the share of the load carried by each of the possible paths. However, in detail design the situation is slightly different. Here the designer must be careful to avoid the situation where he has (for instance) two load paths, neither of which is strong enough to carry the applied load on its own (an apparently satisfactory design) but he also has one path more rigid than the other. He may then nd that the exible structure does not start to resist its share of the load until after the rigid structure (the stiff path) has exceeded its load-carrying capability and broken. An easy trap to fall into is that set by the use of different materials in one structure; for instance, if a designer attempts to reinforce a light-alloy structure with a steel addition he may nd that the steel tries to carry all of the load, because the Youngs Modulus of steel is much higher than that of light alloy; i.e. the steel presents a stiffer load path. When calculating the stresses in a situation of this sort it is not only necessary to nd the simple load over area stress, but also to check the stress due to deection (i.e. stress = Youngs Modulus (E) strain).

Detail design begins with an understanding of how the material can be shaped into the desired part so that account can be taken of the necessary bend radii, the available manufacturing processes, etc. Be especially wary of joggles! Avoiding double curvature in sheet metal parts will keep manufacturing costs down. Machinings can often give more optimised solutions than sheet metal designs. Machined parts tend to be used in areas of high load concentration, e.g. wing to fuselage joints, undercarriage legs. The cost of machinings can be minimised by starting with extrusions (rather than bar-stock) which more approximately match the desired nished shape, thus reducing the amount of material to be removed. Even so machined parts might well still be more expensive than sheet metal fabrications and more difcult to repair. Beware of notches. Sharp corners to holes act as stress raisers and the local stress will be much higher than the simple load/area calculation would lead you to think. Always give internal corners as large a radius as
158
possible. Note that even machining marks will act as stress raisers so the smoother the nish, the better. The choice of fastener will be the result of a compromise between, amongst others, ease of installation, maintainability, bulk, cost and the type of loads being carried by the joint. Details such as joggles have their own peculiarities (such as poor load path stiffness) but, in general, all riveted joints should have a minimum of two rivets per side so they are able to react a torque load. The relative stiffness of different load paths will have a signicant effect upon the load taken by each of those load paths and so stiffness is a major consideration when trying to work out how strong a part needs to be.
Chapter 10
Composite Materials in Aircraft Structures
10.1 What are composites?

When any two or more substances are used together in a structure it can generally be termed a composite. In aircraft design the term composite is used when the main material of construction consists of strands of strong bres stuck together with an adhesive (resin or matrix). Composite materials are normally used in the form of layers of thin cloth or at tapes, although small-diameter bundles of bres, called rovings, are also sometimes used. These materials are easily formed into complex curved shapes of almost any size, giving very clean, aerodynamically smooth surfaces. Careful design and selection of materials can result in structures that are stronger, stiffer and lighter than their metal counterparts. During the 1960s, sailing boat manufacturers began to exploit composites to make smooth-curved hull shapes of low cost and needing little maintenance. The material used was chopped-strand glass bres with polyester resin. This gave smooth and cheap hull mouldings but for its weight was not very strong.Aircraft made from this kind of material would be too heavy. Glider manufacturers saw the advantage of the beautiful, smooth surfaces possible with composites but needed something a little stronger and lighter. Glass bres woven into a cloth (and used as long strands of rovings for wing spars) were combined with stronger epoxy resin to give a better strength to weight ratio but at a greater cost. This combination of materials is now used for almost all gliders, as well as many home-built or kit planes and even some light aircraft such as the Slingsby T-67 Firey and Grob 115 trainer. Further improvements in strength to weight ratio were required before composites could be more widely adopted in the aircraft industry. This came from the introduction of carbon bre (often called graphite ber in the USA) which offered increased strength and stiffness and a reduction in weight (see Fig. 10.1). Improvements in manufacturing techniques have reduced the amount of resin needed to bond the bres together, yielding further weight reductions.
160
Fig. 10.1
Stress : strain diagram comparing composites with aluminium.
For military and commercial aircraft in the 21st century, composites are mostly used in the form of uni-directional tapes, where the straight bre strands are all laid side by side and run in the same direction like a ribbon. These are supplied with a measured quantity of resin already squeezed around the bres and are called pre-pregs (for pre-impregnated). They can be laid into moulds by hand or by a programmed, robotic tape-laying machine. These high-strength pre-preg systems need to be cured in an autoclave (a large pressure chamber that applies heat and pressure to the component) and the total cost of the materials, labour and capital equipment can be very high. Examples of the use of pre-preg materials can be seen in the wings and forward fuselage of the AV-8B Harrier II and the tailplanes of the Airbus A-320. Apart from a limited number of Beech Starship aircraft, no allcomposite commercial aircraft has yet gone into production but composites are used extensively in combat aircraft like the Euroghter Typhoon and SAAB Grippen, in helicopter structures and rotor blades and in fairings and control surfaces of airliners.
161
10.2 The strength of composite materials

A piece of composite material, such as a tape, with all of its bres running along its length will be much stronger when pulled in the direction of the bres (longitudinal strength) than when pulled to the side (transverse strength). This is unlike metals, which have virtually the same strength in all directions. In many structures the loads (stresses) are predominantly in one direction, so the designer can place most of his bres in that same direction to maximise strength and minimise weight. Careful selection of the bre directions in a structure can cause it to twist or warp in a controlled manner and it is this effect that has made possible forward-swept wings for some experimental aircraft. The starting point for considering the strength of the various types of composite material and comparing it to metal can be taken as a single strand of bre or bundle of bres like a string, wetted out with the appropriate amount of resin and cured. In this case all the bres run in the same direction and share the load equally. Typical design values of tensile stress (taken from actual aircraft design manuals) are listed below, the aluminium alloys are included for comparison: Glass bre, wet lay-up Carbon bre, wet lay-up Carbon bre, pre-preg Aluminium alloy sheet 2024-T3 Aluminium alloy plate 2014-T651 310 MPa (44 950 psi) 292 MPa (42 360 psi) 585 MPa (85 500 psi) 414 MPa (60 000 psi) 460 MPa (66 700 psi)
(Note: refer to Section 5.5 for an explanation of the different types of stresses.) This list does not tell the whole story. The reason composites are often considered for aircraft design is because of their strength to weight ratio. A high strength to weight ratio will result in a lighter aircraft structure. If we take the tensile strengths listed above and divide them by their respective material weight (density), we can get a comparison of strength to weight ratio, starting with aluminium alloy sheet as 100%: Aluminium alloy sheet Aluminium alloy plate Glass bre, wet lay-up Carbon bre, wet lay-up Carbon bre, pre-preg 100% 111% 126% 182% 235%
In a simple design case like the anges of an I beam or the spar caps of a glider wing these strength to weight ratios can actually be achieved but in many other cases they cannot. The reason is that while the composites are very strong along the direction of the bres, they are very weak across them. To make a composite panel equally strong both along its length and
162
across its width, half of the bres would have to be turned through 90 to run across the panel. This would result in the composite having only half of the strength in each direction and, consequently, its strength to weight ratio would also have halved. The metal panel, of course, always had equal strength in each direction. Typical strengths for composite panels with equal numbers of bres running both along it and across it (0/90 cloth) are listed below and can be compared with the values above: Glass bre, wet lay-up Carbon bre, wet lay-up Carbon bre, pre-preg 109 MPa (15 801 psi) 148 MPa (21 462 psi) 315 MPa (45 900 psi)
The strength to weight ratios are much lower, compared to the aluminium sheet: Glass bre, wet lay-up Carbon bre, wet lay-up Carbon bre, pre-preg 44% 72% 126%
In a panel subjected to shear stresses, such as the vertical web of an I beam, the best strength is achieved when the bres in the composite are aligned at 45 to the web so that they run diagonally up and down it. Special 45 double-bias cloths are available to serve this purpose. The 45 cloths are also used in wing skins and other parts of the aircraft structure to resist torsion loads or twisting (Table 10.1). It can be seen from the strength to weight comparisons above that the use of composite materials does not automatically result in large weight savings in an aircraft design. It is necessary to calculate the exact amount of material to place in each direction and to use the minimum amount needed to carry the design loads. It is by this optimisation of the design that an efcient lightweight structure is achieved and the nature of composites allows it to be more easily achieved than with metals.
10.3 Types of structures

The simplest type of composite construction is a solid laminate skin supported by ribs and frames, just like a conventional metal or wooden structure. This is common in military aircraft, where the structure can be subjected to very large loads and the resulting components become quite thick. A combat aircraft wing skin made from pre-preg tapes could be 25 mm thick. A wing skin this thick has the advantage that it can resist buckling but it can become quite difcult to get the loads transferred from the skin into the spar webs. This is because of the poor through-the-thickness strength of the laminates, which might allow the skin to tear away
Table 10.1 Composite design values Youngs Modulus, GPa (p.s.i.) 13.8 (2.0 106) 4.55 (0.66 106) 37.6 (5.45 106) 3.45 (0.50 106) 41.9 (6.08 106) 41.9 (6.08 106) 13.9 (2.02 106) 82.7 (12 106) 82.7 (12 106) 70 (10.2 106) 70 (10.2 106) 17 (2.47 106) 130 (19 106) 130 (19 106) Shear modulus, GPa (p.s.i.) 1.65 7.03 1.65 1.65 4.07 4.07 19.5 4.62 4.62 (0.24 (1.02 (0.24 (0.24 (0.59 (0.59 (2.83 (0.67 (0.67 106) 106) 106) 106) 106) 106) 106) 106) 106) Tensile/compressive stress, MPa (p.s.i.) 109 (15 810) 36 (5264) 310 (44 950) 14 (2027) 148 (21 462) 119 (17 206) 63 (9090)(T) 292 (42 360) 234 (33 960) 315 (45 900) 273 (39 780) 77 (11 115)(T) 585 (85 500) 507 (74 100) Shear stress, MPa (p.s.i.) 30 (4307) 89 (12 905) 30 (4307) 30 (4307) 23 (3339) 23 (3339) 110 (16 018) 26 (3792) 26 (3792) 27 (3990) 27 (3990) 194 (28 101) 27 (3990) 27 (3990)
Type and orientation 0/90 Glass cloth, wet lay-up 45 Glass cloth, wet lay-up Undirectional glass tape, wet lay-up, axial Undirectional glass tape, wet lay-up, transverse 0/90 Carbon cloth, wet lay-up, tension 0/90 Carbon cloth, wet lay-up, compression 45 Carbon cloth, wet lay-up U/D Carbon tape, wet lay-up, axial, tension U/D Carbon tape, wet lay-up, axial, compression 0/90 Carbon tapes, pre-preg, axial, tension 0/90 Carbon tapes, pre-preg, axial, compression 45 Carbon tape, pre-preg U/D Carbon tape, pre-preg, axial, tension U/D Carbon tape, pre-preg, axial, compression
4.8 (0.70 106) 4.8 (0.70 106) 34 (4.93 106) 4.8 (0.70 106) 4.8 (0.70 106)
163
164
from the spars. Many spars, ribs and shear-webs may be needed to spread the load transfer over a large area of skin. Ten or more wing spars might be used in a combat aircraft. In less severely loaded aircraft structures, the skins will become thinner and the number of ribs, frames and spars reduced. In light aircraft and gliders a single main-spar and just a few ribs are used. The skin and ribs can become so light that they are too thin to resist buckling and some additional form of stiffening is needed. A common solution to this problem is to build the skins as a sandwich by inserting a lightweight ller or core layer into the laminate. The skin will now consist of this sandwich core layer with thin composite face layers glued to either side of it (Fig. 10.2). This construction greatly increases the laminates bending stiffness and therefore its resistance to buckling, with only a very small increase in weight. (See Section 2.4 for more information on sandwich construction.) Another application for sandwich construction is in the oor panels of airliners, where a large at panel must carry bending loads. Glass-bre facing skins are normally used over a honeycomb core layer. The glassbre facing skins resist the bending loads, rather like the anges of an I beam and the sandwich core carries the shear force. The thicker the core layer that is used, the thinner and lighter become the skins and the less will be the bending deection of the oor under load. The weight of the core layer will remain the same because the thicker core can be made from a lighter material to carry the same shear force. Many home-built light aircraft use a different form of construction, usually referred to as mouldless composite construction. The basic idea is that a huge block of polystyrene foam is carved (usually with a hot wire) into the shape of the wing and glass-bre skins are simply built up over the outside. The foam is left in place and serves as the shear-carrying ribs and spar webs. Wing main spars can be built in too. Sometimes the spar booms and webs are made separately and tted into the foam block before the skins are applied. Other things must also be built in, such as hinges and controls for ailerons and aps and wiring conduits for lights and aerials. Although very simple and requiring no major moulds, the structures produced are a little heavy and a great number of man hours can be spent achieving a smooth nish.
10.4 Joining composites

Since composites consist mostly of bres stuck together with epoxy resin, it is quite straightforward to glue components together with epoxy adhesives. Most glass-bre gliders and light aircraft are almost entirely glued together, bolts being used only where particularly heavy loads are applied or at disassembly points. Strong glued joints can only be achieved by the correct selection of the adhesive and very careful preparation of the surfaces to be joined.
Fig. 10.2
Composite wing spar construction (one of several possible methods).
165
166
In larger aircraft structures it is quite common to eliminate many of the glueing operations by forming several of the components at the same time, this is called co-curing. For example, a wing skin can be laid uncured into its mould and then the ribs and spar webs, also uncured, are added, with tooling to support and position them. The whole lot is then cured together in an autoclave. This leaves only the closing skin to be attached by other methods. The tooling needed for co-curing large aircraft parts is extremely complex and very expensive. It is quite possible to use mechanical fasteners, such as bolts and rivets, to join composite components together or to attach them to metal components. Specially designed rivets and high-strength fasteners are available for use in composites. Using rivets designed for metal may damage the composite. Ordinary bolts may be used and often some local reinforcing feature may be added, such as a sleeve bushing in a hole or a gluedon metal face plate. Galvanic corrosion of the bolts or rivets must be prevented when carbon bre is present (see Section 8.1). The main precaution is to prevent the ingress of water by applying jointing compounds or adhesive and wet assembling the joint. Specially coated or plated fasteners are available and it is common to use titanium rather than steel bolts. Drilling holes in the composite for fasteners will greatly reduce its strength. To get around this it is necessary to reinforce it. Where a tting is bolted to the composite structure it is often adequate to just glue metal facing plates either side of the composite and then bolt through the whole lot. Another way is to build up the thickness of the composite around the joint by adding more layers and placing them with the bres lying in the direction that gives the best joint strength. Bolting things to a foam or honeycomb sandwich structure requires something to stop the lightweight core being crushed when the bolts are tightened. A reinforcing block of dense foam, plastic or even wood can be imbedded in the core during manufacture. Alternatively, the cells of a honeycomb could be lled with a strong adhesive ller paste. Metal inserts can be bought or made, then glued into large holes cut into one side of the sandwich. The face skins of a sandwich panel are usually very thin and need to be reinforced to carry the bearing loads around a bolt.The metal inserts often used have large diameter heads or anges that spread the load over a larger area of skin. Large-diameter washers or metal face plates can be glued on to serve the same purpose. The idea is to spread the load over the face skin, rather than take it all on the thin edge of the hole. A large number of specialist items are available for making attachments to composite. Adhesively bonded screw studs, anchor nuts and cable tie bases can be used to deal with many attachment problems without having to drill holes in the composite.
167
10.5 Fibres
10.5.1 Glass bre
There are two main types of glass bre: E-glass, which is the most common type, and S-glass, which is available mainly in the USA. Both types are similar but S-glass is slightly stronger. Woven glass cloths and tapes are available in a wide variety of styles and thicknesses. Because the bare glass bres are very delicate, they are coated with a substance called size to protect them during manufacture and weaving. There are several different types of size and epoxy resin does not stick properly to all of them. A glass cloth with a nish that is suitable for use with the chosen resin system must be used. Silane is the most popular for use with epoxy resin but there are others and some universal nishes; a check should be made with the technical department of the supplier before using them.
10.5.2 DuPonts Kevlar aramid bre

Aramid bres look like thin yellow string and are very soft to the touch. DuPonts Kevlar is the best known but there is also Twaron by Akzo. It is very strong in tension and very light. Stiffness is much better than glass bre but not as good as carbon. A good range of cloths is available from many different suppliers. It is very easy to wet-out with resin and bonds well with epoxy to give a tough, lightweight composite. Aramids are excellent for making stiff, lightweight shell structures but there are a few drawbacks: the material is weak in compression and is therefore not very good for carrying bending loads or for use in compression struts; it does not cut very easily, so special drills, saws and shears are needed for both the dry cloth and nished composite; the bres tend to go uffy around cut edges and holes and sanding down the surface can be a disaster. The material is also prone to soaking up moisture through exposed bre ends (called wicking), which can damage the structure and therefore the components must be well sealed with paint. One big advantage of Kevlar is that laminates made from it are very resistant to impact damage. Kevlar is sometimes used in hybrid composites, mixed with carbon to get the strength and stiffness of carbon combined with the toughness and lightness of Kevlar.
10.5.3 Carbon bre (or graphite bre)

Carbon bres are very strong, stiff and light. There are two main varieties, called high-modulus and high-strength, caused by variation of the manufacturing technique. Data on the performance of specic types of bre is normally readily available from the manufacturers and suppliers. The actual strength of carbon bre is not much greater than that of glass bre but it is lighter and three times as stiff (i.e. its Youngs Modulus is much greater, see Section 5.5.3). Unfortunately, it is also much more
168
expensive than glass bre. It is available in a wide range of woven cloths, stitched fabrics and uni-directional tapes (narrow tapes with all the bres running parallel in the same direction), both as a dry cloth or as a prepreg with epoxy. Carbon bre is easy to work into complex curved shapes, wets-out well as a wet lay-up and forms a strong bond to epoxy resin. It can be cut and drilled quite easily, although special drills may be needed with pre-pregs to prevent damage around the holes. One problem with carbon bre is that it is a high-resistance electrical conductor, which might explode when struck by lightning. Aeroplanes and yacht masts made of carbon bre must be tted with lightning conductors unless the cross-sections of the components are so big that they can absorb a lightning strike without overheating. For aircraft this usually means building copper conductors into the wing around carbon spars or adding a conductive layer, such as aluminium mesh, over the outside of the whole aircraft. The use of bres in aircraft structures is illustrated in Fig. 10.3. Figure 10.4 shows an aircraft with an entirely composite main structure.
Fig. 10.3
Application of composite materials to the A340.
169
Fig. 10.4
CMC Leopard 002 an aircraft with an all composite main structure.
10.6 Resins
10.6.1 Polyester resin
Polyester resin is widely used in commercial, industrial and marine applications. It is not normally used in aircraft because epoxies offer better strength and durability. The resin is easy to work with, being supplied as a clear liquid resin to which a few drops of liquid catalyst (hardener) are added and stirred-in just before use. It is sometimes used for low-cost tooling and moulds when accuracy and durability are not vital.
10.6.2 Epoxy resin

Epoxy is the resin system used in most aircraft structural applications. There is a wide range of different formulations available to suit different
170
purposes; the manufacturers or suppliers technical departments will give advice on the most suitable system for any specic application. The main differences are between the different temperatures used to cure the resin, the working temperature it is expected to see in service and the manufacturing method used to make the composite components. The greatest strength is achieved with a resin system that is cured at high temperature, typically 175C under pressure in an autoclave. To get the lightest weight components for aerospace use, all excess resin must be eliminated. This is normally done by running the bre tapes or cloth through a bath of mixed epoxy resin and then between rollers that force resin into the cloth and squeeze away the excess. This is done by the material suppliers and the resulting material is called a pre-preg. To stop the resin from hardening in the pre-preg cloth before it can be cut and laidup in the mould it is kept at cold temperatures, this is typically -10C. Even when stored in a freezer, the epoxy will eventually go off, so these materials have a limited shelf life of 6 or 12 months. The high-temperature curing epoxy systems are expensive to use because not only is an autoclave needed but the mould tools must be strong enough to withstand the repeated heating and cooling cycles of component production; this means that they must be made from expensive materials. Epoxy resin systems are being continuously developed to improve strength and reduce the cost and difculty of component manufacture. Low-temperature curing pre-preg systems are now available that give excellent strength when cured at temperatures of 75100C and that need only the pressure of a portable vacuum bag system, rather than the higher pressures of an autoclave. For light aircraft and gliders the epoxy resin is bought in a two-part pack and mixed just before use, where it is brushed, rolled and squeezed into the dry cloths laid in the mould. Unlike polyester resin, it is very important to get exactly the right quantity of resin and hardener into the mixture. This is done by using accurate scales to weigh it out or by using a metering pump that delivers the correct ratio. Different resin systems require different mixing ratios and this is explained on the data sheet supplied with the resin. The component is left in the mould to cure at room temperature for 24 hours but it can take up to 14 days before it has reached its full strength. To speed up the curing and to make the component more resistant to high temperatures, it can be cooked in an oven (post-cured) at 4580C for several hours.
10.6.3 Vinyl ester resin

Some kit planes use vinyl ester resin systems which can be considered to fall between polyester resin and wet lay-up epoxy resin in terms of both strength and cost. These resin systems are mixed and handled in a very similar way to polyester and result in good chemical resistance and moderate strength. In most aircraft applications, the greater strength of epoxy is considered to outweigh the cost saving of vinyl ester.
171
10.6.4 Phenolic resin

Interior wall panels and cabin furnishings are made using phenolic resin systems.This is because phenolic is the only resin system that is adequately re resistant and can meet the smoke and toxic fume emission requirements when burnt. It is not strong enough to be used for the structure of the aircraft and is generally more difcult to process than epoxy.
10.7 Working safely with composites

While it is obvious that working with composites can be sticky, messy and dusty, there are some real health hazards that mean precautions must be taken to protect the user. Resin systems by law must be supplied with safety information either on the container or in a data sheet and this is worth reading. Polyester resins are quite benign when mixed, except for the styrene fumes, so working in a well ventilated place or even outdoors is the main precaution against headaches and nausea. The hardener used with polyester is another story. It is often a strong peroxide called MEKP, which can burn the skin and cause severe permanent eye damage if spilled or splashed. Overalls and latex gloves should be worn to keep it off bare skin and goggles prevent it being rubbed or splashed into the eyes. With epoxy resins, the problem most often encountered is a skin reaction. This can accumulate with exposure so that the skin becomes sensitised after a period of trouble-free working. Some resin systems are much worse than others but they all have the potential to cause problems. Keep skin covered when working with epoxy. Use disposable latex gloves and preferably a barrier cream as well, try never to touch wet resin with bare hands. Epoxies also give off harmful chemical fumes, again some worse than others. A good-quality respiratory lter mask will be required and goggles too, to keep the substance out of the eyes. The workshop needs to be well ventilated, preferably with fume extraction fans. Working outdoors is not normally an option because airworthiness requirements demand a controlled environment and in any case the cure time may be more than 12 hours and good temperature and humidity cannot be guaranteed overnight. The problems do not stop when the resin has cured. Cutting and sanding the nished components brings its own problems. A respiratory mask and goggles are required to keep the irritating dust out of the eyes and lungs. A barrier cream or gloves should be used too, to stop tiny bres burying themselves in the skin. Dust extractors in the workshop are needed, along with good ventilation. Care should be taken not to spread too much dust around the building, so dusty clothes should be removed on leaving the work area. All these precautions mean that undertaking composites manufacture on a commercial basis requires some signicant expenditure on workshop
172
equipment and preparation and on disposable protective clothing for the workers. Heating, air-conditioning and a de-humidier will also be needed to meet the requirements for a controlled environment. A calibrated thermo-hygrograph is also necessary to record the data for quality.

In aerospace terms composite usually means long strands of strong bres in a matrix of epoxy resin. Glider manufacturers were the rst to embrace the technology, followed by military aircraft and light aircraft builders. Today large commercial aircraft have signicant elements of primary structure made in composite materials. Composites are strong in the direction of the bres but comparatively weak across them.The strength of composite materials is of a similar order to that of good aluminium alloys but their real advantage is in their high strength to weight ratio. The strength to weight ratio of carbon bre prepregs is over double that of good aluminium alloys. However, in practice, this advantage is only fully realised in relatively few applications, i.e. when the loading is mainly in a single direction, such as in wing-spar booms (or caps). The need to have strength in all directions can result in the strength to weight ratio for a nished composite component being similar to, or worse than, that for an aluminium part. Where composites really score is in the ability to place material in exactly the place where it is needed and nowhere else. This is the key to minimising structural weight using composites and is more easily achieved than with conventional materials. Designing in composite bring its own set of considerations. On larger aircraft, simply copying an aluminium alloy structure can result in very thick lay-ups with the attendant problem of through-thickness load transfer. Conversely, on lighter aircraft, items such as wing skins can end up being very thin and thus require stabilising against buckling. This is often achieved by creating a sandwich structure. Classically, composite parts are joined by being glued together. However, by using co-curing (where several parts are cured at the same time, as an assembly) the number of joints can be signicantly reduced. Mechanical fasteners (bolts, rivets, etc.) can be used but the resultant stress concentration usually means that local reinforcement is necessary. Load transfer between composite parts is best achieved by way of largearea shear connections. The variation in cost, strength, stiffness, workability, etc., of materials such as glass, aramid and carbon bres is so great that a mix of several different bres is often used in the same structure. Of the various resins available, however, only epoxy is really used for aircraft as its strength and dimensional stability characteristics outweigh the lower cost of other resins.
Chapter 11
Quality and Airworthiness
11.1 Quality assurance and quality control

Quality assurance may be dened as the organisational systems and procedures put in place to ensure that the companys product or service consistently meets the required standard. Quality control is, typically, the inspection of nished parts to check that they conform with the drawing. Both activities are fundamental to ensuring aviation safety. Quality is complementary to aircraft structure and therefore only complementary to this book. However, the airframe engineer is so constantly involved with quality that it will be useful to give an indication of the methods used to ensure that the structure meets the specications which it is intended to meet. Everybody associated with an aircraft is concerned with its quality (especially those who y in it) and a very large percentage of those associated bring a personal inuence to bear. A cabin cleaner on a remote outstation and a radiographer examining a wing joint for cracks are both inuencing quality. In general industry outside aviation, product quality awareness starts at the top and is controlled by using the two-way management communication system. The board of directors of a company (or the company president) set the quality standard for their product according to their assessment of the market. Middle management and the engineering department specify the quality as part of the design and instruct the manufacturing staff on how to achieve the correct standard. The manufacturing staff then make the product to the standard required and offer it to middle management for acceptance. Middle management monitor the work and report back to the directors. In the aircraft industry the process is somewhat modied by the inuence of agencies outside the company. In the case of both military and civil aircraft, these agencies set standards for design and manufacture which must be adhered to. In the civil aircraft market minimum quality standards are set by the Airworthiness Authorities. Table 11.1 shows the titles of the authorities in various countries. These authorities are backed by the law of their respective countries and publish very detailed requirements of design quality which must be achieved before the aircraft is per-
174
Understanding Aircraft Structures Table 11.1 Airworthiness Authorities Europe European Aviation Safety Agency (EASA) Postfach 10 12 53 D-50452 Koln Germany www.easa.eu.int Civil Aviation Authority (CAA) Aviation House South Area Gatwick West Sussex RH6 0YR United Kingdom www.caa.co.uk Federal Aviation Administration (FAA) 800 Independence Ave SW Washington DC 20591 USA www.faa.gov
United Kingdom
USA
mitted to y. They also require the quality on production aircraft to be kept at a continuing level of excellence which they endorse and monitor. Because of the status of the Airworthiness Authority it would not be possible for a company making or operating aircraft to set low standards of quality for commercial advantages. In fact, since the earliest days of ying there has been remarkably little conict between the Airworthiness Authorities and commercial interest, and most aircraft manufacturers set their quality standards above the already high legal minimum, as do the airline operators.
11.2 Control
Quality control starts with education to ensure that the newest recruit to the industry is aware of the importance of quality to the safe and efcient operation of aircraft. The second aspect is checking. Everyone checks his own work until he is satised with its correctness. Then the inspection department acts as a screen or lter to pick out the inevitable errors which we all make. Finally, the technical director (or chief engineer) and the Airworthiness Authority monitor the inspection department to ensure that necessary systems and procedures do not fall into disuse. The third aspect is the setting up of systems and procedures designed to make quality control more foolproof. Typical of the many systems required for quality control are:
a system for issuing and distributing the latest issues (corrections and
updates) of engineering drawings
175
a procedure for notifying the drawing ofce of errors found on

drawings a system for ensuring that workshop measuring instruments are regularly checked for accuracy systems to ensure that parts of an aircraft found to be defective on one ight are repaired before the next.
The nal aspect is specication (usually by engineering drawing or maintenance manual) linking the customers and the Airworthiness Authoritys quality requirements with the companys manufacturing or maintenance capabilities. A major function of the Airworthiness Authorities particularly relevant to this book is the approval of the design of the aircraft and its equipment down to the smallest detail. The methods employed to examine the design and assess its quality vary in detail from country to country, but in general they are based on a system of acceptable reporting. This means that the authority recognises the competence of a company (or in some cases an individual), and when that company reports that a design meets the authoritys requirements then the design is accepted as being airworthy. This system operates within specied limits for each company; that is a company approved to issue reports on the airworthiness of engines would not expect to be considered competent to report on airframe structure. Also, the approved design organisation is subject to regular monitoring, in that the Airworthiness Authority reserves, and exercises, the right to review design data presented by approved design organisations. The standards which a company must achieve before they can be considered as competent to report to the authorities are noted in each countrys Airworthiness Requirements.
11.3 Procedures and systems

Two examples are given to illustrate the working of quality assurance; both relate to aircraft structures in the manufacturing stage. The rst example concerns the procedure for dealing with defects in manufactured components found by the inspector. Because of the outside agencys inuence on the industry, aircraft inspection is rather different from that in, say, the automobile industry. The aircraft inspector is employed and paid by the company he works with, but in the event of a conict between commercial interest and quality, then his loyalty lies not with his employer but with the Airworthiness Authority. This sounds dramatic but in the event of a major snag or defect being found no company would give priority to any consideration of the commercial problems created by grounding an aircraft or scrapping a component. A minor defect, on the other hand, is dealt with by a procedure as follows. If the inspector nds that some items in a production batch of bolts (for example) have a thickness of head which is less than is allowed by the tol-
176
erance on the drawings, then he raises a reject note which he sends to the production supervisor with the faulty parts. (Tolerance is the sum of the variation above and below the nominal dimension. Nominal dimension and upper and lower variations (called limits) are shown on the drawing.) However, he does not think that head thickness is a feature of such importance that the bolt should be scrapped. He therefore arranges with the production supervisor for concession procedure to be started. This procedure is, in effect, a request that the technical department authorise an alternative design standard for the bolts, in this case with a thinner head than that specied on the drawing. Initially a concession request is addressed to the technical department requesting that a concession be granted. The request form will identify the bolt by description and part number, say how many items are involved and probably include a sketch to clarify the defect. In their turn the various sections of the technical department (drawing, stress, weights) would examine the request and comment on the effect of the fault on interchangeability (for spares), strength or any other aspect within their interest. Provided that there is no adverse comment the chief engineer (or chief designer) signs the note to indicate his approval of the request which is returned to the inspector as an authority to accept the bolts as usable parts. This effectively cancels the reject note and the two documents (reject note and concession note) are led together as part of the history of that order for bolts. A similar document to the concession note is the production permit. This is used before the event rather than after. It is probably on the same printed form and follows the same route. A typical use is to request permission to produce a part in an alternative, more readily available, material. The second example traces the route of a quantity of aluminium alloy from manufacture to conversion into a nished component. (1) The alloy is cast and is given a cast number. (2) The cast billet is converted into extrusion and the extruded bars are marked with the cast number and a batch number. Test pieces are taken from the batch and checked to ensure that the material meets its specication. (3) The extrusion is sold to a number of stockists (metal dealers) and in each case is accompanied by a document signed by the manufacturer certifying that the material meets its specication. This document, called an Approved Certicate (or a release note or a statement of conformance), also notes the batch number and test report number. (4) The stockist gives the extrusion his own batch number and holds it in a store which has been approved by the Airworthiness Authority until he sells the material to various manufacturers. (5) Part of the original batch is delivered to a company who receive several feet of extrusion and a delivery advice note. The goods inwards inspector of the receiving company immediately places the material into quarantine store. Material in quarantine cannot be used
177
until the inspector authorises its movement into bonded store, and he will not do this until he receives the stockists Approved Certicate. (Note that the original material manufacturer has issued one Certicate but the stockist may issue several relating to the original batch of material as it disperses to various users.) The stockists Approved Certicate passes on the assurance that the material is to specication and notes the batch number. The inspector again rebatches the material into his companys own system and notes the details (i.e. material specication, dimensions, supplier, date received, etc.) into his batch book. The material passes into stock and awaits further conversion. (6) Eventually, part of the stock is withdrawn from store for conversion into a structural component. At this point it acquires its last (or probably its last) batch number. This number appears on a route card (or job card or work sheet) which accompanies the material from process to process until it is converted into a nished component. The component is nally stamped or marked (according to drawing instructions) with the last batch number and goes into service. This seemingly laborious, but simple, task, which applies to all sorts of aircraft materials and components is a very old, established cornerstone of aircraft quality control procedure. It has two fundamental uses. Firstly, anyone receiving any piece of material or equipment together with an Approved Certicate can use that material with condence that it will perform its function according to its specication. Parts for which an Approved Certicate cannot be obtained can be considered as bogus parts of doubtful manufacture. Secondly, in the event of a component proving defective in service, others with the same batch number can be traced. In the extreme case, by tracing backwards through the various companies batch books to the original cast number and then forward again, all the components which were eventually made from the cast can be traced.
11.4 Further notes on quality control functions

Some of the functions of quality control and some of the terms and expressions used tend to be loosely applied, especially in smaller companies, and warrant some further comments. For example: (1) A statement of conformance (or a letter of conformity) from a supplier referring to a piece of material is useless unless it states clearly to what specication or standard the material conforms. It is equally useless if the standard is unacceptable to the customer. (2) A concession note starts life as a request for somebody to change his requirement or lower his standards and therefore must be addressed to that person. Typically in (say) a subcontracting situation the note
178
(3)
(4)
(5)
(6)
would initially be from the chief inspector of the subcontractor addressed to the chief designer of the prime contractor. Batch numbers can easily be allocated to any groups of material or parts received into a company but if a nut from batch A is tted to a bolt from batch B the assembly must be given its own number or the chain of the system is broken. Finished parts should be marked with their batch number but, as a quick examination of any random selection of equipment will show, this is not always strictly enforced. However, if a complete piece of equipment is given a serial number, the manufacturers records, i.e. route cards, history cards or similar systems, will show all the components which went to make that particular article and their batch numbers. Companies purchasing and receiving equipment for use on aircraft should ensure through their goods inwards inspection that these procedures have been adhered to. The function of the inspector is that of a lter. If a faulty component eludes him he has made a mistake in his job; but the faulty component was nevertheless made by the production department and it is not possible for them to wash their hands of the responsibility. Material or parts cannot be withdrawn directly from a quarantine store. By denition a material in quarantine has not yet been proved to be faultless. If it has already been proved to be faultless it should have been moved (by authority of the goods inwards inspector) into the bonded store. The second example set out above illustrates an example of traceability. This synthetic word tries to sum up the features and systems which must exist and surround each component concerned with the aircraft. It should, for instance, be possible to pick any part and by looking at the numbers printed on it, trace back to the raw materials and also to trace the applicable manufacturing drawing, applicable stress calculations and test reports, and eventually an authoritative signature by which the designer acknowledges his opinion that the part will perform properly. The object and benet is mainly to ensure that if something does not perform as it should, the reason can be tracked down and corrected.
For design ofces, a quality assurance system which ensures control of the content and amendment of all drawings, reports and other documents is essential. In addition, amongst other things, design ofce quality systems must also concern themselves with interchangeability, i.e. all parts which have the same part number are identical in terms of t, form and function. A good rule of thumb here is to ask yourself, when a drawing is updated, whether you care if you get a part made to the old or new issues. If you do care then the old and new design standards are sufciently different to require different part numbers.
179
11.5 Airworthiness engineering

With the ever increasing codication of the laws and rules which control aircraft design and operation, there is a need for an engineer who can take an overview of the whole design with the object of ensuring that the published requirements are met. In a sense, the work is the opposite of innovative design in much the same way that careful accountancy is the opposite of entrepreneurial enthusiasm. The suggestion is that the airworthiness engineer should not attempt to control design, but make it his job to ensure that designers are aware of the connes within which they must work, and the requirements with which they must demonstrate compliance, also bearing in mind that the efforts of one specialist designer may unknowingly be affecting some other specialists area. This book is about structures, and the requirements dening their desired strength have been stable over a long period. This situation produces exactly the atmosphere in which newly published requirements may go unnoticed by designers, who may then waste expensive time through working to outdated regulations. The timely advice of an airworthiness engineer who continually keeps himself aware of revisions to the rules governing safety should help prevent wasted effort. Because of the element of cross-specialisation involved, the airworthiness engineer must have a wide general experience of the aircraft industry (although by natural growth the airworthiness engineer has become the airworthiness department, which has people specialising in specic areas). He must also have a sufciently open mind to accept that large sections of his carefully acquired knowledge may have to be drastically changed by some new legislation. More difcult to accept, he may also nd that standards he has worked to for many years are not acceptable standards in another country. The quantity of written material published by the authorities which is required reading for the airworthiness engineer is so great that he cannot be expected to have every facet of the requirements at his ngertips. What he must know is where to start looking for information. For civil aircraft in Europe the basic source of information is contained in the Certication Specications (CS) published by EASA. In America the requirements are dened in Federal Aviation Regulations (FAR). It is no coincidence that CSs and FARs are almost identical. There is a long standing harmonisation process in place, the aim of which is to arrive at a unied set of requirements between Europe and America. For military aircraft, the source document for UK requirements is DEF STAN 00-970. US requirements are contained in MIL Specs. A selection of other publications that an airworthiness engineer must be aware of are listed below, with comments on the areas of particular interest to structures engineers. (Note here that, in common with many sources of information, no-one should attempt to learn word by word what is written down. Examination assessors and quiz masters still seem to put
180
some store by a persons ability to recite facts, but it is far better if you can say that the information you require is in book A. Accident Reports, from a variety of sources, help by highlighting errors which might have been avoided by better engineering. Mandatory Occurrence Reports (MOR) allow, amongst other things, people involved in the operation of aircraft to bring to the attention of the airworthiness engineer any incidents which may have a detrimental effect on safety. CAP 562, Civil Aircraft Airworthiness Information & Procedures, are a mine of information on aircraft workshop practices. Repair manuals are published by the aircraft manufacturer and either describe specic repairs for specic damage or set out general rules for dealing with damage met in service. An American FAA publication, AC43.13-1B, Acceptable Methods, Techniques & Practices, is another immensely useful and practical source of generalised information that the airworthiness engineers should be aware of. Service Bulletins and Alert Service Bulletins are issued by manufacturers to operators to inform them of problems which require attention, or of difculties which have been met by other operators.
11.6 Continued airworthiness

The above notes mainly concern manufacturing and design quality, but quality control has to continue into the operating life of an aircraft structure. This is achieved through an approved maintenance schedule which denes a series of checks and inspections to be carried out after so many ying hours and/or elapsed calendar time. For large aircraft these are listed and dened by the aircraft manufacturers, but to give an indication of how a typical schedule works the following would be reasonable for a light aircraft: Check A a pre-ight check at least before the rst ight of the day Check 1 50 ying hours or two months, whichever occurs sooner Check 2 100 ying hours Check 3 once a year or 500 hours Check 4 once every two years or 1000 hours The following structure items would be among the details inspected at the various checks: Check A skins, moving control surfaces, windscreens and windows, door locks, all visually inspected for any obvious defects
181
Check 1 all xed and moving structures, doors, windows, access panels, fuel tanks inspect for visible corrosion, operation and damage Check 2 internal structures, especially near potential moisture accumulation points inspect for corrosion Check 3 structural joints inspect for slackness or corrosion Check 4 all structural parts inspect for corrosion, slackness, cracking wear or any form of deterioration. Items found to be worn or damaged beyond the limits dened in the aircrafts maintenance manual at any inspection would be rectied before the aircraft continued in service. To avoid taking aircraft out of service too frequently or for long periods it is often the practice to agree a system of running checks with the regulator. Under such a system all the items required to be checked yearly (say) would be listed, and instead of all being checked at one grounding of the aircraft they would be attended to in weekly or monthly batches.

Quality standards for the aviation industry are set by the relevant airworthiness authority, e.g. EASA in Europe and FAA in the USA. Before any safety critical work is released it is rst checked by the person doing the work and then independently checked by another appropriately qualied and approved person. This applies to drawings, reports, tests, manufactured parts and maintenance. Regulators approve organisations as being competent to carry out and certify particular functions. They then regularly audit these organisations to verify that standards are being maintained. An example of a quality assurance process is that employed to ensure traceability. For example, any metal component can be traced back to the raw material from which it was made. In design the airworthiness engineer guides the designers with regard to the airworthiness requirements with which compliance must be shown before the product can be certied. The continued airworthiness of a product is ensured by way of a carefully dened set of in-service inspections.
11.8 References
Although it is the responsibility of companies to set up their own quality assurance and control systems, advice and help is available from the publications of the Airworthiness Authorities and inspection procedures in particular are well documented. Lists of EASA and FAA publications are available on the appropriate web site.
Chapter 12
Stressing
12.1 Introduction
Stressing, or stress analysis, is primarily that process which estimates whether or not the proposed structure is strong enough to carry the loads which will be imposed on it by the operation of the aircraft. The designer (or, with the complexity of modern aircraft, the design team) is required to design a structure which will allow the aircraft to perform in its role efciently. There are many constraints which inuence the designs, such as maintenance accessibility, production costs, aesthetic appeal and so on, but this chapter is concerned only with the structure and its problems. The overall loads to be applied to the structure are determined by the airworthiness requirements. These required loads, the specication of the role of the aircraft and the nature of the constructional materials and manufacturing processes available, are the raw data with which the designer begins. It is in the nature of aircraft that the lighter they are, the better they perform in their role. Also, commercially, the cost of carrying superuous weight in terms of total fuel consumed during the life of the aircraft is very high. The designers task then includes achieving maximum lightness and to do this he will employ the most advanced materials and constructional techniques which are within the production capability of his company. (The last sentence is not intended to suggest that complexity is at any time desirable. Simple solutions are almost always the best for any design problem. One of the great aircraft designers, Ed Heinemann, said Simplicate and add lightness.) Going hand in hand with the advance of materials and techniques is more and more detailed and accurate analysis by the stressman. This has a twofold aim: to reduce weight by ensuring that every piece of structure is fully employed and also to eliminate expensive test failures. Airworthiness approval only occasionally relies on the stressmans calculations; there is always sufcient testing on major structural items to conrm compliance with requirements, although tests do not necessarily show up components which are over strength. Also, with larger aircraft and more expensive manufacture, companies are very keen to eliminate any early
Stressing
183
failures on test. In spite of the fail-safe type of design philosophy, premature failure of one component can scrap a whole test-piece structure. So, if the primary stressing task is to ensure adequate strength, the secondary task is to assist the designer with the aim of producing structures which are exactly strong enough and no more. Stressing is now a highly complex subject relying heavily on mathematical analysis. In this chapter the techniques used are discussed in their most elementary form, with the idea, as stated in Chapter 1, of giving a foundation vocabulary on which readers can build a more detailed study. Stressing is a mathematically biased subject but it is not necessary for stressmen to be more than competent mathematicians. The majority of the work done in the examination of structure relies on the use of data and formulae which only require practice and sensible application for their satisfactory use. Writing the data and evolving the formulae do require considerable mathematical skill but the stress ofce always has a place for the type of engineer who has a feel for structures and who can recognise the weak points in two minutes, as well as the man with several degrees who can identify the same weaknesses after two weeks of calculating. Both types of stressman are useful and even essential to a balanced team.
12.2 The stressmans work

When a stressman examines a piece of structure he must (a) Determine the load distribution throughout the structure. (b) Determine the extent to which the elements of the structure are capable of supporting the loads imposed on them. Statement (a) is simple enough when we consider a simple structure surrounded, as we said in Chapters 4 and 5, by a balanced system of loads and reactions. If a thin chain is suspended from a hook in the ceiling and a weight put at the bottom of the chain, then the weight is the applied load, and what happens at the hook is the reaction (Fig. 12.1). At every link in the chain we have the same situation: an applied load at the bottom and a reaction at the top. (If we ignore any weight in the chain itself, the loads and reactions are numerically the same all the way down.) If we now look at half a link we realise that what we were calling a reaction is in fact a load when we think of the link in isolation, and the reactions are provided by the other half of the link. (One bits load is another bits reaction.) By progressively thinking our way through this simple structure we have determined the load distribution. Many structures can be examined and the internal distribution of loads analysed by statics alone; that is, by considering in careful stages how each load is reacted and by progressively dividing the structure into sections (as we did with the chain), balancing loads and reactions as we go. Figure 4.3 from earlier in this book is relevant here.
184
Fig. 12.1
Loads and reactions in a thin chain.
Unfortunately aircraft structures usually fall within a group known as statically indeterminate structures. These are structures which have more than one load path and Fig. 12.2 shows two classic examples. Frame (a) is a structure which is sound enough but it is also still a satisfactory structure at (b) with one member removed. Beam (d) is a good structure but it would still hold up the two loads if it was divided at the middle support. The continuity at the centre support in (d) and the removable member in (a) are called redundancies (or redundant members) and the load distributions in statistically indeterminate or redundant structures have to be found by consideration of the deformation. (We can use Fig. 12.2 for consideration of two other ideas. If we remove the second diagonal (BD) from (b) the structure would cease to be a structure and would become a mechanism, that is, the bar (CD) would swing into some position of repose which would not be where we wanted it to be. The other thought is that members (AD) and (BC) are in tension. It happens that the load in (BC) is greater than the load in (AD) and also (BC) is longer. If we are not careful with their sizes the two members could stretch as shown in (c) so that the line of action of the centroid (i.e. the centre of gravity, CG) of the two weights could come outside point B, which would put member (AD) into compression. If (AD) was not capable of carrying compression, the structure would have changed into a mechanism and what appeared at rst to be satisfactory would have turned out to be quite the reverse.) Determining load distribution can be difcult but with practice and thought most problems can be solved. Chapter 5 emphasised the importance of the relationship between stress and strain. The vital key to a clear analysis of some very complex structures is the very simple idea that there is no stress without strain and, equally, if a part of a structure is strained it is also stressed. This concept is applied in an interesting way in the consideration of the effects of kinetic heating (see Section 4.4). Because these
Stressing
185
Fig. 12.2
Statically indeterminate structures, etc.
effects and ideas are not easily understood we will digress to look at the example of a simple structure which is not an aircraft part, but will illustrate principles that do apply to aircraft structures. Imagine a camera tripod with its feet rmly anchored to the ground (set in concrete). Now imagine a fourth leg added to the tripod. Provided that the extra leg is exactly the correct length, its inuence is zero, but the structure which was three legged and needed all its members is now redundant, or statically indeterminate; that is, any one of the four legs could be sawn through without the camera falling. Obviously, we cannot affect our
186
camera support by kinetic heating, but we could heat one leg with a blow lamp. If this one structural element is heated it will expand, or try to expand, by an amount which can be calculated by multiplying the temperature rise by the coefcient of expansion (a known physical property of the material). Because of the heating, this extension takes place without any stress being produced, but if the extension is prevented or reduced by the reactive efforts of the other members (as it will be in this case, since they are anchored to the ground), then the amount by which the extension is forcibly reduced is effectively a compressive strain with consequent stress and hence load in the member. We now have the heated member in compression, being loaded by the other members, which in our example, are in tension; all these loads have been produced by heating just part of the whole structure. Structures of this type with more members than they need but with a different form (stressed skin) are normal in aircraft and are affected by kinetic heating in a similar way to the way in which the camera mount was affected by the blow lamp. The situation of one members load pushing or pulling on another member is typical of statically indeterminate structures under any externally applied load (apart from or in addition to the heating condition). Usually in aircraft structures the problem is complicated because as well as straightforward changes to the lengths of the members, there are also angular deections and rotations at the joints between members to be considered. However, the principles are simple and if we compare, step by step, the extensions, the strains, the stresses and, consequently, the loads, we can arrive at the condition where all the loads are balanced, with one members load reacting the next members load, and the problem is solved. Once the distribution of loads around the structure is quite clear (and do not forget that, as we pointed out in Section 4.2, there is more than one case to be considered), each element of the structure has to be examined to see whether or not it is capable of carrying the load applied to it. Apart from the very common possibility that the stress level is too high, the stressman must also be sure that members are not fundamentally incapable of carrying the type of load imposed. A typical example of this sort of problem is one that frequently causes dissent between draughtsmen and stressmen. The draughtsman will fasten two components together with rivets (which are good shear connectors but are poor in tension) and the stressman will say that they must be bolts. If at this point the draughtsman shows good reasons for not having bolts then there will inevitably be disagreement. Arguments of this type can be avoided if the draughtsman and the stressman work together from the beginning of a design task. If such a partnership can be made to operate smoothly it can be of great benet to the job. Inevitably it is not an easy relationship since the draughtsman is inclined to think of the stressman as a head-in-the-clouds academic, while the stressman is certain that he could make a better job of the design if only he had more time. Neither of them is right. The designer/draughts-
Stressing
187
man is bound to know more about manufacturing processes and design aesthetics than the stressman because that is the eld in which he has gained his experience. On the other hand, if he listens to the stressmans advice on load distribution and size of members, he can possibly pre-empt the inevitable day of reckoning when all structure design work still on the drawing-board is declared by the chief designer to be both too weak and too heavy! The remaining part of the stressmans work is to continue his academic studies, either formally through further education, or informally through reading and discussion with others in the profession.
12.3 Stressing methods

There are no short cuts to learning stress analysis. Study and work under the instruction of experienced stressmen are the only routes. Some approaches to study are, however, easier than others and design draughtsmen or licensed engineers who would like the satisfaction of designing a repair scheme, or even a piece of structure which the stressmen can approve without argument, should be able to teach themselves enough theory by reading. Some knowledge of mathematics is essential and some further knowledge gives added interest. The essential knowledge is basic algebra (for the manipulation of formulae) and trigonometry (for the solution of triangles). A knowledge of calculus and how to read the mathematical shorthand which goes with it opens a whole eld of interest, because without this it is almost impossible to read the textbooks and understand how the formulae are evolved. Schools and colleges sometimes fail to explain the connection between pure mathematics and the study of practical subjects such as the theory of structures. An ability to use maths with ease enables the engineer to set up mathematical models. These are close approximations to the true and accurate situation, which can be thoroughly analysed by mathematical methods. A typical example in the stressing of beams would be a beam with a loading pattern so close to being evenly distributed that the stressman could fairly assume that it was exactly uniform. This assumption of exactness means that the next stage of calculation examines a truly parabolic bending moment distribution and the parabola is one of many geometric shapes which can be integrated or differentiated or otherwise mathematically dissected with ease by someone familiar with the processes. Extra practice with an increasing knowledge of mathematics enables the analyst to set up models which more closely and accurately represent more complex problems. The remainder of this chapter notes some things which are not usually mentioned in textbooks and also continues the policy of explaining some of the terms and expressions used in stressing.
188
12.3.1 Analysing the structure

The best way to start stressing a piece of structure (perhaps the only way) is to stand back and look at it. At the stressing stage it will only be a drawing so looking at it means understanding the drawing and visualising it in your mind. It is a piece of structure so it is being loaded. (If there is no load we are not looking at a piece of structure but at a fairing or a piece of in-ll.) As it is being loaded the loads are being reacted, so we look for the places where the loads are coming in and where the reactions are being applied. The structure is a link between the loads and the reactions and between the two it will be pushed or pulled or sheared or bent or twisted, or it may be subjected to all of these, some of them simultaneously. We must think about it and be absolutely sure that the structure is arranged to cater for all the loads being applied to it. It is ridiculously easy for a designer to cope admirably with a load in one direction and entirely forget another load in another direction. Once we are satised that the structure is complete we can begin the arithmetic. The rst move in the calculation is to balance the external loads; that is, to make sure that every load is properly reacted. Checking the balance of the external loads is a continuous process right through the analysis of the structure. Every component must be in balance in every loading case. Being in balance means that the sum of all the loads (or the forces) and all the moments is zero. To check this easily, forces are divided (see Fig. 4.3) into parts along three axes which are mutually at right angles. The usual axes taken are the x axis fore and aft along the aircraft, the y axis left and right across the aircraft and the z axis up and down. One of the most difcult things at this stage is keeping track of whether the load is positive or negative, but if we keep checking for zero balance any sign error will soon become apparent. At this point the going is almost certainly becoming harder. Being an aircraft structure, the piece being examined stands a very good chance of being statically indeterminate and, as mentioned in Section 12.2, such structures need special methods of analysis. With the help of computers it is becoming quite easy to use nite element methods. The practice is to divide the complex structure into elements or parts which can be examined separately, and their deections under load matched with their neighbours deections. The structure of Fig. 12.2(d) can be treated as two elements and solved without the help of even a hand-held calculator. However, structure 12.2(a), which is still not very complex, has six elements and if we had chosen to put joint plates at all the intersections instead of having pin joints (i.e. joints with clevis pins) the analysis would have already become too tedious for hand calculations. Powerful computer programs have been written to deal with this sort of problem and are now accessible to most stressmen, so that after a short
Stressing
189
study course on how to divide the structure and how to present information to the computer, structures which a few years ago could only have been analysed approximately, can now be examined very accurately. This approach is now available to the smallest of companies and even humble equipment manufactures are using these methods to rene the structures of their products. There are already several very usable, cheap programs for desk-top computers, which are readily available and capable of analysing simple redundant structures. Approximate solutions are often sufciently accurate for smaller or ancillary structures, especially where the weight penalty of excess structure is being carried for other reasons. (A typical example here is a piece of oor being analysed for a load in the middle of the panel. When held all around its edge the structure is statically indeterminate, but it might prove strong enough if only two edges were held.) A stressman faced with alternative load paths might quite convincingly and reasonably imagine all the load going on one path and if he found that strong enough he might not look further, satised that if one path was strong enough, the two paths would be stronger. Finding and applying approximations to load distribution problems requires the knowledge which comes with experience and discussion, but the majority of senior stress engineers are always ready to talk over problems and pass on a little know how. So far we have glossed over the problems of load path analysis. These can be very difcult and it is part of the stressmans work which requires more imagination and insight than other parts which can be operated with data sheets. Having made that admission, it is also true that licensed engineers and draughtsmen who attempt design stressing soon develop a pattern of thinking which proposes simpler structures with clearer load paths, which is to their advantage and to the benet of the aircraft.
12.3.2 Calculating stresses

Once the loads in the members of the structure are determined, calculating the actual stress level is usually straightforward, given that an adequate library of data sheets is available. One of the books listed in the References (Handbook of Aeronautics No. 1) was written by the compilers of many of the initial Royal Aeronautical Society data sheets and it reproduces all those which are of fundamental importance. The talent that the stressman brings to the job at this point is the ability to pick the correct data sheet (or the correct formula). To do this he must visualise how the member or component he is considering will fail. (For readers who may be faint-hearted potential air travellers, the phrase . . . will fail . . . is another piece of jargon meaning . . . how the component would eventually break or collapse if the application of load was to be continued past the required level.) Once he has reached a decision on the mode of failure (i.e. the way it is going to break) he refers to the appropriate data sheet which will present him with one of the following:
190
(a) A permissible load on the component (e.g. the permissible shear on a specic rivet and sheet-metal combination). (b) A stress level which he must match against the specied capabilities of the material used. (c) A factor by which a calculated stress or load must be corrected when the original calculation was not sufciently rened; for example, a rectangular section bar which has a hole drilled through it is loaded in tension. The apparent tensile stress is the load divided by that area of section of bar which is left either side of the hole. We know, however, from previous reading (Chapter 9) that the hole is a stress raiser. So we look for information on such a condition (the author found it in Formulas for Stress and Strain), and nd that the rst calculation of stress will need to be increased by a substantial amount (depending on the proportions of the bar the factor will be of the order of 2.0 to 3.0). The factor does not always indicate a stress increase. For instance some beams which appear to be overstressed on rst investigation may be satisfactory after the application of a stress-reducing factor relying on the proportions of depth to length, called a short beam bending factor. The majority of stressing data is well presented and easy to understand. Often the compiler includes a worked example. For people who are not actually working in a stress ofce the most difcult type of data sheet to obtain will probably be that dealing with the strength of riveted joints. All the major aircraft manufacturers publish their own tests but only for their own people. A very short table of rivet strengths is shown in Appendix 1 at the end of this book. In a large proportion of investigations, unless there has been very close co-operation between draughtsman and stressman from the beginning of the design, the rst run through the stress calculations will reveal faults in the design. As these are corrected it can happen that the load distribution is affected and so the whole process must be re-run, but unless exceptional accuracy of design is being sought, two passes at the problem are usually enough.
12.4 Stress reports

The object of the stress analysis is to determine whether or not a structure is strong enough and to achieve this the analyst compares the stress in each component of the structure with the allowable or permissible stress for the material of which the component is made. Sometimes, as in the case of bolts and rivets, the comparison is made between load and permissible load but the effect is the same. At the end of what may have been a massive calculation the stressman notes the actual stress level which he estimates will be in the material, and the calculation nishes on one line
Stressing
191
which says in essence: the permissible stress is greater than the actual stress, therefore the component is strong enough. There are two common ways of quantifying this. The British method usually quotes a reserve factor, and the American method quotes a margin of safety. It will be clearer to use gures to explain the difference. (The examples use imperial units, but followers of SI units will understand the points made.) We assume that the stressman has investigated the specication and the requirements, and has found the loads on the whole structure. He has multiplied by proof and ultimate factors, found the loads on the particular item in question, and calculated the stresses in the item. He then notes down (for example): max. tensile stress found permissible UTS for . . . material Therefore (in the British system) reserve factor (RF) permissible actual 51 000 = 45 530 = 1.12 = = 45 530 lbf/in.2 = 51 000 lbf/in.2
or therefore (in the American system) margin of safety (MS) permissible - actual actual 51 000 - 45 530 = 45 530 = 0.12 =
(Note that MS = RF - 1) Reserve factors are usually noted in a right-hand margin and so the calculation ends: max. tensible stress permissible UTS = 45 530 lbf/in.2 = 51 000 lbf/in.2 RF = 1.12
(Note here that RF < 1.0 read a reserve factor less than one means that the strength of that component is inadequate and RF > 1.0 read RF greater than one means that the component is unnecessarily heavy.)
192
Note also that the RF (or MS) is assessed against the maximum stress found. In an earlier example (Section 12.3) we saw the effect of a stress raiser, but in fact that maximum high stress would be localised in a very small volume of the material. So if the RF is found for the place of maximum stress, as it has to be, then in our example, and also in most other components, there is a great volume of metal where the RF > 1.0 and extra weight is being carried, which takes us right back to Section 6.2.1, where we discussed why steel is a heavy material to use compared with aluminium alloy. As aircraft design and manufacturing techniques improve we will progress towards the situation in which every ounce of material is fully worked and we have RF = 1.0 throughout the structure. At that point, the arguments in favour of using titanium may be overwhelming unless an even better material has been found. The justication for discussing the last line of the stress calculation before the rest is that we need to see the ultimate objective. (Incidentally, in any technical report it is a good idea to make the rst paragraph, after a description of the objectives of the report, a summary of the conclusions. The rest of the report is then a justication of the conclusions and a detailed report of how the end result was reached.) We said at the beginning of this section that the stress in every component is investigated. This is certainly true for all the components which are newly designed for the structure being investigated, and a reserve factor (or margin of safety) is found and noted for each part identied by its unique part number. It is also generally true for load-carrying items of standard hardware such as bolts and rivets; the only type of exception being where, for instance, a group or line of rivets is so arranged that either the loading is evenly spread or it is clear that one rivet is more highly loaded than the rest. So far we have the report listing reserve factors; it should also make a statement noting the minimum reserve factor found throughout the whole structure being reported. This would be part of a summary statement which effectively says that the stressman has examined the structure in question and nds it strong enough, within certain limits. Having made this statement, the report must also list those limits. These are the specications and requirements referring to the aircraft in its role, any exceptions to specication or limitations of role, and a note of any assumptions made during the calculations leading up to the report. (Assumptions may concern conditions external to the structure about which no exact information was available at the time of the calculations, or they may concern approximations of the calculation method by the simplication or idealisation of a component which is difcult to analyse. The oor panel example in Section 12.3.1 would be analysed on the assumption that only two edges were held.) The whole report will then list:
Document reference number Issue or revision letter
Stressing
193
Date Title: Stress Report Subject: the structure identied by part number and name
(If it is a long report there may follow an index and a list of effective pages with their own current issue status. Some reports have an amendment record wherein is noted a prcis of the changes that have been incorporated in the text at each update.)
Summary of results and conclusions Associated specications and requirements Load cases: a list of load cases and combinations of load cases
considered
Assumptions: a summary of major or signicant assumptions made

during the calculations
Exceptions and limitations: (to specications) Calculations: the detailed calculations showing how the reserve factors
are obtained. Apart from the arithmetic this section should justify the assumptions by giving the reasons for their use and it should show by reference where the data used in the calculations originated. Compiled by: Checked by:
The presentation of reports varies from company to company, and the easiest course before attempting to write ones rst is to look to see how previous reports were set out. It is a personal view that typing is unnecessary and, because of the prevalence of mathematical formulae, not always successful. This does not apply to the Type Record which summarises all the reports for the whole aircraft. Alternatively, typing everything down to and including the Exceptions and Limitations but leaving the actual calculations handwritten makes a neat job.

When analysing the stresses in a structure we start by carrying out a critical, non-mathematical review where we ask ourselves questions such as:
Do we have a structure at all or is it a mechanism? Is it a redundant structure (i.e. one that has more members than are
necessary)?
Is the design suitable for reacting the type of loads that are being
applied to it (e.g. folded channels or other open sections that are poor at resisting torsion)? Next, identify all of the loads being applied to the structure (and they will differ from one load case to another). Identify all of the points at which the loads and reactions act and assess their ability to resist loads in
194
various directions (e.g. a pin joint can react direct loads but not a moment). Now calculate, numerically, the loads, reactions and moments and ensure that they balance. If they dont balance either youve missed something or the whole structure is moving. Using the appropriate formulae (from simple load/area to complex thin-wall shear ow sums) calculate the stress or load on each part of the structure. Derive a reserve factor by dividing the allowable stress for the material by the stress being applied. The RF should be greater than one (RF > 1) but not by much otherwise the structure is unnecessarily heavy. Finally collate all of the calculations into a formal stress report and have the report independently checked for accuracy.
12.6 References
There are many books on stress analysis and the theory of structures and eventually everyone nds their favourite. In the following list of books, each contains its own bibliography so that readers are set off on a path of study which can be as detailed as they wish. The comments are my own but I think most stressmen would agree with them. Bruhn, E.F., The Analysis and Design of Flight Vehicle Structures. Tri-State Offset. Expensive in the UK but a master reference and guide. Full of easy to follow worked examples but entirely professional. Roark, R.J. and Young, W.C., Formulas for Stress and Strain. New York: McGraw-Hill. A standard reference book stuffed full of practical and usable data. Early editions are closer to the aircraft industry than the 5th edition which gives wide coverage of all structural work. Peery, D.J., Aircraft Structures. New York: McGraw-Hill. A classic textbook with just the right blend of practical illustration and mathematical explanation. All the above are written using imperial units. For people used to SI units: Megson, T.H.G., Aircraft Structures for Engineering Students. London: Edward Arnold.
Chapter 13
Presentation of Modications and Repairs
13.1 Denitions
As the title says, this chapter concerns modications and repairs and by that we mean changes of greater or lesser degree to an existing aircraft. When an aircraft is rst designed and built its features and characteristics are all recorded on drawings and reports, the structure in particular being documented in great detail. All these records are an integral part of the machinery of approval for the aircraft, approval which is given by the Airworthiness Authority (see Section 11.1) and which is required by law or by government order, before the aircraft can enter service. If the existing aircraft is now changed in any way, however small, it is no longer exactly the aircraft which was approved and the records can prove it. Therefore, when the change is made, if the aircraft is to continue to y legally, the change needs approval from the Airworthiness Authority and documents of its own to supplement the original records of the aircraft type. Changes to a design need some denition (and here we are dealing only with structure). Repairs clearly involve a change unless new identical spare structural components can be tted using the original xings and xing holes. Most aircraft have a repairs manual which lists and illustrates standard repair schemes which are already approved and can be incorporated without further discussion or paperwork. Damage which is outside the range of the repairs manual will need a new repair scheme designed by an Approved Design Organisation (see Section 11.2) and accepted by EASA. In many repair situations the Approved Design Organisation will be the original aircraft manufacturer. (Note: In this section we are describing the European system. Constant reference to two different procedures would be confusing but the systems discussed are fairly general. The American system uses consulting engineers (Designated Engineering Representatives) to monitor design output, whereas the European system accepts design output only from an approved organisation.) Modications are changes and under EASA procedures can be either major or minor. In essence, for structures, a modication will be major if it has an appreciable effect upon primary structure or lifed items (i.e. those parts with a life dened in numbers of ying hours or cycles). The
196
classication of major or minor will be carried out by the approved design organisation using procedures and criteria agreed by the regulator. A classication either way will have no effect upon the quantity or quality of work necessary to show compliance with the airworthiness requirements. However, minor modications or repairs are approved by the approved design organisation whereas major ones have to be approved by EASA. In Section 5.8 we dened Class 3 components; if the modication concerns a Class 3 part it can safely be classed as minor. Modications as they affect structure or structures designers are usually initiated by some operational requirement. They can range from enlarging a fuselage door, through designing new brackets and attachments for the latest navigation aid, to substituting a new type of rivet for one which is unobtainable. The rst of these is obviously major but we cannot denitely say that either of the others is minor without rst comparing them with the detailed criteria in the organisations approved handbook. When an operator requires a modication on one of his aircraft, or on all the aircraft in a eet, he approaches an Approved Design Organisation, which may be inside his own company. If he needs to go outside his company he must be sure the ofce he approaches is competent and approved to work in the particular eld in question; for instance, if the modication is structural, it is no good talking to experts in fuel systems. If the modication involves two disciplines (for instance, a new navigation aid may require major structural modications to house an aerial in addition to electrical system changes), then one design ofce is picked to supervise the work and subcontract those parts which it is not qualied to handle. The approval of organisations is mentioned in Chapter 11 with an indication of their necessary qualications. In the context of this chapter, an organisation is recognised by EASA as being competent to report to them that a proposed change to an aircraft will in no way produce a conict with the Airworthiness Requirements. How the Approved Design Organisation justies its condence in its claim to itself and to EASA is to some extent its own affair, but its methods and records will be open to scrutiny by EASA as part of their assessment of competence. To summarise: (1) Prior to a modication (or a repair) the aircraft (or other equipment such as galley) has been examined and approved by a recognised airworthiness regulator. (2) After modication the aircraft has changed, so the original approval can no longer be valid. Therefore the approval must be reestablished (EASA do this, for major modications, by way of a Supplemental Type Certicate or STC, see below). (3) The conventional way of describing what modication is to be made and of recording for the future what was done, is by engineering drawings.
197
13.2 The essential paperwork associated with modications

13.2.1 The specication
Every job starts with a specication and an order. In the case we are discussing the specication must come from the operator of the aircraft. It will probably be verbal and it may only say we have a problem here which must be solved within EASA regulations, but hopefully it would be more complete. A well-written specication virtually designs the job. If the specication is verbal the design organisation would be prudent to write it down and feed it back to the operator for his approval (by signature). The order (so far as the technician is concerned) need only say Proceed with a design to my specication, but again this should be written and signed.
13.2.2 Recording modications

The next step, if the modication is major, is for the approved design organisation to apply to EASA for modication approval and this sets the approval process in motion. The information required at the time of application will be along the lines of:
Aircraft type Applicants modication number Name, address and Design Organisation Approval number Description of the modication Timescale for the modication programme Statement of the airworthiness requirements with which compliance is to be shown Certication plan showing how compliance is to be shown with the affected parts of the airworthiness requirements.
Knowing what the modication will entail and the airworthiness requirements which are to be met allows the designer to study those requirements and produce a draft compliance check list (CCL). The CCL will list all of the airworthiness requirements affected by the modication, note the method by which compliance is to be shown (e.g. analysis, test or design) and the document which demonstrates compliance for that particular aspect. For the structural aspects the documents listed will probably be stress or test reports. With the CCL drafted, the designer will have a pretty good idea of the compliance reports that will have to be produced for the modication. The designer will also be in a position to give a detailed brief to the design team, including a distillation of the airworthiness requirements such as the necessary strength, reliability, etc.
198
Now is a good time to create a preliminary drawing structure showing the necessary installation, assembly and parts drawings required to fully dene the desired modication. It is good practice to split the drawings into two types:
drawings dening work that can be done in the absence of the aircraft;
and drawings that dene work which requires the aircraft to be available for it to be carried out.
The rst category will, broadly, dene manufacturing work, while the second details what can be termed installation activities. The advantages of making this distinction can be numerous. First, manufacturing and installation tasks are often carried out under different regulatory approvals and might even be undertaken by different organisations or divisions of the same organisation. The manufactured part will usually need to be inspected and then booked into stores to await the arrival of the aircraft. Once the aircraft arrives the part will be issued from stores, either to the installation department or to the customers maintenance provider for installation, and so there is a clear separation as to the procedures used for each function and which inspectors are taking responsibility for the different stages of the work. Perhaps more importantly the separation allows the greatest possible percentage of the programme to be completed and signed off before the aircraft arrives in the hangar, thus minimising aircraft down-time. Aircraft on the ground are not earning money and so time on the ground should be kept to the absolute minimum. Structural drawings need to communicate to the workshops all of the details of the nished part. The required dimensions, material specications, fastener and other hardware part numbers, heat treatment and other processes, corrosion protection measures and part marking must all be clearly and unambiguously shown. It has been said that the draughtsman is in the communications industry and drawings are the method by which he communicates with the workshops. Before a drawing can be considered complete it must be independently checked by appropriately qualied and authorised specialists for each of the disciplines covered by the drawing. Structural drawings will require to be signed off by a stressman but they might also affect other aspects such as performance, aerodynamics, systems, etc., and so will need to be checked and signed off by a specialist from each of these areas. Ultimately, someone must put their signature on the drawing to say that all of the necessary investigations have been carried out by the appropriate specialists and that, overall, the drawing complies with all of the relevant airworthiness requirements. This function is provided by a Compliance Verication Engineer or CVE. Once the signature of the CVE has been obtained the drawing can be issued and copies sent to the workshops for manufacture.
199
13.2.3 Reports
As we noted above, the drawing will be signed by the stressman to indicate his approval of the strength of the change. Before he signs he must have justied his condence to himself and he must also be in a position to justify his condence to EASA. To do this he must either make a satisfactory written analysis of the strength and write a stress report (i.e. he has got to stress it himself) or he must have available a satisfactory report of a physical test of the strength. In some instances he will have a strength test covering one particular loading case and the other cases will be covered by stressing. A test report will need to be preceded by a test specication set out with the following information: (a) (b) (c) (d) a reference number an issue or revision letter a date a title Test Specication
Then the following headings and information: (e) Subject (f) Reference documents and relevant specications (g) Object of test (usually to show compliance with specications (f)) (h) Method of test (i) Equipment needed (j) Loads to be applied (k) Methods of recording load A test report will need the following: (a) (b) (c) (d) a reference number an issue or revision letter a date a title Test Report
Then the following headings and information: (e) (f) (g) (h) (i) (j) (k) (l) (m) Subject Object of test Date of test Witnesses Test specication reference number Account of test Results achieved Conclusions (comparison with test specication) Signature of tester and a witness
200
The stress report and its contents were discussed in Chapter 12. Whatever other reports might be necessary to cover various aspects of the airworthiness requirements (safety analyses, electrical reports, ight manual supplement, post installation test schedules, etc.) it is invariably wise to create an over-arching design report for the modication. This report should give the reviewer (maybe EASA or the customer) a broadbrush overview, describing all of the main features of the programme. It will tie the other, disparate, documents together in a clear and logical manner as well as providing a home for a whole range of minor details relating to the modication. Lastly, it should allow a reader to quickly ascertain that all of the necessary certication work has been done and where to nd it. A design report, if written properly, acts as a guide book to the modication.
The Modication Leaet

The top level document for a modication will be called something like the modication leaet, modication certicate or type record. This document has several purposes and these are, broadly:
To show the modication number, title, issue and aircraft type To list the reference number and issue of the applicable airworthiness
requirements To act as an index for all of the documentation and drawings which dene the modication and which are needed to show compliance with the appropriate airworthiness requirements To record the modication classication (major or minor) Most importantly, to certify that the modied aircraft complies with the airworthiness requirements stated above. This certication will normally be signed by the Head of Design.
If the modication was classied as minor then the signing of the modication leaet signies nal approval of the modication (assuming that approval of minor modications is within the design organisations Scope of Approval issued by EASA). Approval of major modications is the prerogative of EASA and this will be signied by way of a Supplemental Type Certicate or STC following receipt of the completed modication leaet from the design organisation.
Updating of Modications
Different design organisations will control modication updates in different ways. However a common factor is that if the essential character of the modication is changed, if any key assumptions are invalidated, or if the t, form or function of any of the important elements is affected then the whole modication will need to be reviewed and re-approved.
201

An aircraft modication or repair is a change to the approved design standard and so requires its own separate investigation and approval. Design of modications or repairs in Europe and some other countries is carried out by approved design organisations. Minor modications can be approved by the design organisation but major modications require an STC from the local regulator. In America consulting engineers (called Designated Engineering Representatives) provide the approval function. The classication of a modication or repair as major or minor has no effect upon the quantity or depth of certication work required; it affects only who is to approve the change. For major modications application for approval should be made to EASA noting key details and timescales. Before commencing design of a modication an agreed, clear specication should be in place. The modication paperwork will include a range of required reports and a drawings package and it is invariably a good idea to include a design report to provide a modication overview. In order to minimise aircraft down-time it is helpful to split drawings into those that dene work to be done off the aircraft and those that dene work to be done on the aircraft. All drawings have to be independently checked and signed-off by the appropriate specialists and then signed by a CVE. The top-level document to a modication is the Modication Leaet. It acts as a modication index and certies compliance with the airworthiness requirements. When updating a modication, if any of the key features is affected then the whole modication needs to be reviewed and re-approved.
13.4 Conclusion
Reading this book can at best set somebody on the road towards being a stressman or a structures designer. In fact, almost all the subjects he would need to study in depth are mentioned, but the qualication that no book can give is experience. Some readers who have reached this point in the book will have reached senior positions in their own eld and a broad look at structural engineering will be all they want of it. Some other readers may go on to join a design organisation and acquire experience and expertise. A third group will comprise the airline engineers, on the ground or in the air, who at times want the services of a capable design organisation to make an operational improvement, design a repair or even to bring an old aircraft up to new standards. This section was written for each group to read in its own way. Group one may have read it for interest or as notes on the type of paperwork which should be emerging from the design ofce next door. Group three will have read it in much the same way, but with the sharper interest of knowing that if paperwork of this
202
order is not being produced, then the improvements, repairs, etc., which they are asking for will not be given airworthiness approval. For people in the middle group, this chapter, like the rest of the book, is a rough definition of the terms and expressions used in their industry and perhaps an aide mmoire for the rst time they are out on their own, offering up the drafts of the papers needed to get approval for a change to someone elses design.
13.5 References
Certication Specications (various parts) EASA website, www.easa.eu.int Federal Aviation Regulations (various parts) FAA website, www.faa.gov
Appendix 1
Strength of Rivets in Single Shear
Fig. A.1
Lap-jointed plates held with one rivet.
(1) The data in Table A.1 indicate joint strength per rivet in a joint which has at least two rivets. (2) The data apply to solid, mushroom or universal head rivets in 2117 alloy, i.e. British Standard SP 85, American Standard MS 20470 AD and to Cherrylock Blind Rivets type CR 2249. (3) The gures are accurate enough for design purposes but not for check stressing calculations to be offered for approval.
Table A.1
Joint strength (lbf) of rivet and sheet metal combinations Rivet diameter (in.)
3 32 1 8 5 32
3 16
0.022 (24 SWG) 0.028 (22 SWG) 0.036 (20 SWG) 0.048 (18 SWG) 0.064 (16 SWG)
Thickness t of the thinner sheet (in.)
180 210 220 295 380 390 390 540 595 595 625 835 860
INDEX
Index Terms
Links
A
acceleration coefficient factor acrylics adhesive bonding aerodynamic centre aerodynamic moment aerodynamics aeroelasticity aerofoil ageing, of light alloy aileron aileron reversal air density airflow airframe airworthiness airworthiness authorities Alclad allowable stress Alocrom aluminium alloy angle of attack annealing anodising application for modification 44 53 53 101 119 53 53 29 50 48 91 31 50 48 32 3 179 174 128 52 130 89 39 105 130 197 48 38 33 50 164
This page has been reformatted by Knovel to provide easier navigation.
Index Terms
approval of modification approval of organisations approved design organisation aramid aspect ratio assumptions asymmetry of beams
Links
196 196 196 167 48 193 84 197
B
balance, aerodynamic of loads of moments balancing load of tailplane batch numbers beams beam depth bearing stress bend radius bending bending moment biplanes blind rivets bolted joints bolts bonded stores bonded structures bonding, adhesive box structures bracing wires brazing Brinell hardness British Standards 42 183 41 53 176 72 75 71 133 61 42 4 146 153 144 177 99 119 58 4 119 71 109 78 6 73
Index Terms
brittleness buckling compression inter-rivet load buttock line
Links
71 73 86 73 86 34 35 77
C
CAA cadmium plating canard carbon fibre cast aluminium cementite centre of pressure centrifugal force centroid channel sections chord Class 1, 2, 3, (structures) cleats clips co-curing composite materials compression buckling concessions condition (heat treatment) control loads control surfaces corrosion, electrochemical et seq. This page has been reformatted by Knovel to provide easier navigation. 174 128 31 159 90 105 48 54 184 84 34 87 154 154 166 159 59 86 177 105 49 31 124 49 85
Index Terms
electrolytic fretting galvanic shielding stress cost (of material) couple crash loads CRES cut-outs
Links
125 127 125 127 125 103 75 47 93 155 127
D
data sheets datum, horizontal deflection of beams degreasing de Havilland DER design organisation, approved dihedral dimensions of aircraft dissimilar metals distribution of loads drag drain holes drawings DTD ductility Duralumin 77 35 73 131 14 195 196 34 34 126 186 39 204 198 108 90 96 107 197 100 119 189 37
Index Terms
Links
E
EASA edge distances E-glass elastic limit electrolytic corrosion elevator emergency alighting loads empennage end load (due to shear) epoxy primer epoxy resin etch primer Euler 174 150 167 70 125 31 47 30 76 132 169 131 86 33 49 195
F
factor, acceleration proof reserve ultimate fail safe structure failure modes FAA fasteners fatigue faying fibreglass fin finite element fitting factors flanged holes flaps 45 52 191 52 14 62 174 144 91 127 100 31 188 53 135 32 This page has been reformatted by Knovel to provide easier navigation. 189
Index Terms
flexural centre floor beams flutter fly press folding force et seq. forging forming frame free body concept fretting corrosion fuselage
Links
83 50 31 116 117 38 50
90 114 10 44 127 31
G
g galvanic corrosion glass reinforced plastic glued joints graphite fibre grip length guillotine gusts 45 125 100 164 167 144 115 46 161
H
handed parts hardening hardness heat treatment honeycomb Hookes Law horizontal datum 112 105 71 105 15 67 37 48
Index Terms
horizontal stabiliser
Links
30
I
inertia inertia factors intergranular corrosion inter-rivet buckling 45 53 125 86
J
joggles jointing compound joints bolted Junkers, Professor Hugo 133 127 153 153 92 138
K
kinetic heating 51 184
L
landing loads lanolin lathe lift distribution light alloy limit loads limit of proportionality load cases deflection distribution in joints 46 78 183 152 51 132 114 29 44 90 52 69 38 48 53 56
Index Terms
load (Cont.) landing limit path permissible transfer ultimate working load factors loading cases longeron lugs
Links
51 52 189 191 42 52 52 46 46 6 156
53
56
186
M
machined parts mainplane maintenance manoeuvre cases envelope maraging steel margin of safety MIL specifications milling mode of failure modifications et seq. application for approval of modification leaflet Modulus of Elasticity Modulus of Rigidity 197 196 200 67 69 113 29 180 44 46 46 96 191 108 115 62 195 189
Index Terms
moment aerodynamic monocoque
Links
42 53 10
N
Newton, Sir Isaac normalising notches 60 105 71 139 143
O
off sets organisations, approval of oxide films 35 196 128
P
paint part numbering passivating pearlite permanent set phenolic resins pitching of rivets plastic deformation plastic materials plate plating, cadmium Poissons ratio polycarbonate polyester resin post buckling pound force 131 112 129 105 70 171 31 150 70 99 113 128 71 101 169 73 42 178
Index Terms
precipitation treatment pre-preg pressure, centre of pressurisation loads primary structure Principle of Superposition profile washer proof factor load stress
Links
91 160 48 50 87 78 151 52 52 52 71
Q
quality et seq. quarantine store 176 173
R
reactions et seq. redundant members Redux bonding reinforced plastic release note reserve factor residual stress resultant, of forces rib rivet length rivet pitch riveted joints rivets in tension 184 119 99 176 191 125 40 10 146 151 150 144 150 159 38
Index Terms
Rockwell hardness rolling routing rovings rubber press rudder rust
Links
71 31 112 159 117 31 124 49
S
sandwich panel sandwich structure section lines semi-monocoque sensory structures S-glass shear centre due to bending due to torsion lag modulus strain sheet material shielding corrosion short beam bending skin cracking skin joints slat slot smart structures soldering solution treatment 100 14 34 10 102 167 42 83 72 76 74 69 67 113 127 190 156 139 33 32 101 119 91 60 67 59 23 164
Index Terms
span spar spoiler stabiliser stability, of structure stainless steel stalling speed statically indeterminate structures station lines steel steel alloys stiff path strain strain gauge strength-to-weight ratio stress bearing permissible proof report ultimate stress corrosion stress raisers stress to strain relationship stressed skin stressing stringer cut-outs stringers struts superposition, principle of
Links
34 6 33 30 86 93 54 184 37 92 93 157 63 102 101 63 71 191 52 190 52 97 192 65 10 182 137 8 86 78 155 10 155 67 59 186 125 161 65 127
Index Terms
Links
T
tabs tailplane action of balancing loads tempering of steel tension et seq. testing thrust Timoshenko, Professor titanium tolerances torsion box of tubes traceability transfer, of loads triangle of forces turning lathe manoeuvre moment type record 114 45 39 200 54 199 38 73 97 176 59 82 76 178 39 40 61 76 32 30 42 53 105 60
U
ultimate factor load stress undercarriage loads unidirectional tapes 52 52 52 56 160
Index Terms
Links
V
V-n diagrams vector Vickers hardness vinyl ester resin 46 40 71 170
W
Wagner, Professor warping of boxes Warren girder waterline web, shear weight welding wicking wing area canard section wire-braced structure wire locking work hardening for forming working loads Wright brothers wrought alloy 73 82 46 34 73 38 119 167 29 32 31 48 4 146 70 118 52 3 90 5 96 48 182 77
X
x, y, z axes 188
Index Terms
Links
Y
Y-alloy yaw yield point Youngs Modulus 107 31 69 67
Z
zero balance zinc alloys zinc chromate 188 107 131

Understanding Aircraft Structures

Uploaded by

Copyright:

Available Formats

Understanding Aircraft Structures

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Understanding Aircraft Structures

Uploaded by

Copyright:

Available Formats

What is the purpose and scope of the book?

What is the purpose and scope of the book?

What are some of the major topics covered in the book?

What are some of the major topics covered in the book?

Understanding Aircraft Structures