joining of carbon fibre reinforced plastics for automotive...

Joining of Carbon Fibre Reinforced Plastics

for Automotive Applications

Gordon KellyDepartment of Aeronautical and Vehicle Engineering

Royal Institute of TechnologySE–100 44 Stockholm, Sweden

TRITA-AVE.2004:25ISBN 91-7283-824-8

Joining of CFRP for Automotive Applications iii

Abstract

The introduction of carbon-fibre reinforced plastics in load bearing automotive struc-tures provides a great potential to reduce vehicle weight and fuel consumption. Toenable the manufacture and assembly of composite structural parts, reliable andcost-effective joining technologies must be developed. This thesis addresses severalaspects of joining and load introduction in carbon-fibre reinforced plastics based onnon-crimp fabric reinforcement.

The bearing strength of carbon fibre/epoxy laminates was investigated consid-ering the effects of bolt-hole clearance. The laminate failure modes and ultimatebearing strength were found to be significantly dependent upon the laminate stack-ing sequence, geometry and lateral clamping load. Significant reduction in bearingstrength at 4% hole deformation was found for both pin-loaded and clamped lami-nates. The ultimate strength of the joints was found to be independent of the initialbolt-hole clearance.

The behaviour of hybrid (bolted/bonded) joints was investigated both numeri-cally and experimentally. A three-dimensional non-linear finite element model wasdeveloped to predict the load transfer distribution in the joints. The effect of thejoint geometry and adhesive material properties on the load transfer was determinedthrough a parameter study. An experimental investigation was undertaken to deter-mine the strength, failure mechanisms and fatigue life of hybrid joints. The jointswere shown to have greater strength, stiffness and fatigue life in comparison to adhe-sive bonded joints. However, the benefits were only observed in joint designs whichallowed for load sharing between the adhesive and the bolt.

The effect of the environment on the durability of bonded and hybrid joints wasinvestigated. The strength and fatigue life of the joints was found to decrease sig-nificantly with increased ageing time. Hybrid joints demonstrated increased fatiguelife in comparison to adhesive bonded joints after ageing in a cyclic freeze/thawenvironment.

The strength and failure mechanisms of composite laminates subject to localisedtransverse loading were investigated considering the effect of the specimen size,stacking sequence and material system. Damage was found to initiate in the lami-nates at low load levels, typically 20-30% of the ultimate failure load. The dominantinitial failure mode was intralaminar shear failure, which occurred in sub-surfaceplies. Two different macromechanical failure modes were identified, fastener pull-through failure and global collapse of the laminate. The damage patterns and ulti-mate failure mode were found to depend upon the laminate stacking sequence andresin system. Finite element analysis was used to analyse the stress distributionwithin the laminates and predict first-ply failure.

Joining of CFRP for Automotive Applications v

Dissertation

This thesis contains an introduction to the research area and the following appendedpapers:

Paper A

Kelly G. and Hallstrom S., Bearing strength of carbon fibre/epoxy laminates: Effectsof bolt-hole clearance, Composites Part B, 35, p331-343, (2004).

Paper B

Kelly G., Load transfer in hybrid (bonded/bolted) composite single-lap joints, Toappear in Journal of Composite Structures.

Paper C

Kelly G., Quasi-static strength and fatigue resistance of hybrid (bonded/bolted)composite single-lap joints, Submitted for publication.

Paper D

Kelly G., Effect of the environment on the durability of bonded and hybrid (bonded/bolted) composite single-lap joints, Submitted for publication.

Paper E

Kelly G. and Hallstrom S., Strength and failure mechanisms of composite lami-nates subject to localised transverse loading, To appear in Journal of CompositeStructures.

Joining of CFRP for Automotive Applications vii

Division of work between the authors

Paper A

Kelly performed the experiments, numerical simulations and wrote the paper. Hall-strom supported the work and contributed to the paper with valuable commentsand suggestions.

Paper E

Kelly performed the experiments, numerical simulations and wrote the paper. Hall-strom supported the work, assisted in interpreting the results and made valuablesuggestions regarding the structure and content of the paper.

Joining of CFRP for Automotive Applications ix

Contents

Preface i

Abstract iii

Dissertation v

Division of work between the authors vii

1. Introduction 11.1. Materials for structural automotive parts . . . . . . . . . . . . . . . . 5

2. Joining techniques for composite laminates 62.1. Mechanical fastening . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.2. Adhesive bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132.3. Hybrid joining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

3. Localised transverse loading in composite laminates 20

4. Future work 21

References 23

Paper A A1–A24

Paper B B1–B14

Paper C C1–C19

Paper D D1–D18

Paper E E1–E23

Joining of CFRP for Automotive Applications 1

1. Introduction

Polymer composite materials are being used in a wide range of structural applica-tions in the aerospace, construction and automotive industries due to their lightweight and high specific stiffness and strength. A variety of materials are beingused ranging from lower performance glass fibre/polyester, used in small sail boatsand domestic products, to high performance carbon fibre/epoxy systems used inmilitary aircraft and spacecraft. One sector where the use of composite materials isstill evolving is the automotive industry. Composite materials offer great potentialin reducing vehicle weight, thus increasing fuel efficiency and reducing CO2 emis-sions. It is estimated that a reduction in vehicle weight of 100 kg would reduce fuelconsumption by approximately 0.3-0.4 litres per 100 km [1,2]. In addition to weightreduction, the number of individual parts can be significantly reduced making thehigh-volume composite car concept cost-effective. Current use of composite materi-als in high-volume automotive structures is limited to secondary exterior structuressuch as body panels, wheel housings or energy absorbing bumpers. Body panelsare commonly made from sheet moulding compound (SMC) which is based on ther-moset matrix reinforced by discontinuous glass fibre. Bumpers and front panels havebeen made from glass mat thermoplastic (GMT) which is commonly discontinuousglass fibre reinforced polypropylene. Both SMC and GMT composites are basedon a randomly oriented, discontinuous glass mat reinforcement with a maximumfibre volume content of around 30%. Both can be compared with the fast stamp-ing processes used to manufacture semi-finished sheet metal parts. Their structuralefficiency is typically low being only slightly higher than steel in the case of platebending. However, the advantages of good formability, high energy absorption andresistance to corrosion makes GMT and SMC appropriate for exterior automotivecomponents. A typical example of a composite automotive structure is the FordExplorer Sport Trac cargo area illustrated in Figure 1.

Figure 1: The Ford Explorer Sport Trac with SMC cargo area.

The Ford Explorer Sport Trac features a scratch-resistant, corrosion-proof SMCcargo area with integrated liner and side panels which have allowed for a reduction

2 G.Kelly

in weight of 20% in comparison to a sheet steel solution.However, there is a greater potential to utilise the benefits of polymer composite

materials in automotive structures. Introduction of fibre reinforced plastics in thebody structure could decrease the mass by 50-67% compared to current metallicbody structures. In comparison, the potential of alternative materials such as highstrength steel and aluminium are 25-35% and 40-55% respectively [3, 4]. Reducingthe weight of an automotive body structure can, in theory, lead to a secondaryweight reduction of other key systems. The chassis, gears and brakes can be designedlighter [5] and the potential exists to utilise new, alternative power-train systems.

Carbon fibre reinforced plastic (CFRP) offers high specific stiffness and strengthin comparison with materials such as steel and aluminium. CFRP parts are currentlyused in significant volumes in high end sport vehicles which are produced in volumesunder 500 units/year. A typical example of the use of lightweight materials in loadcarrying automotive structures is the CFRP monocoque and engine carrier of thePorsche Carrera GT. The vehicle design, floor panel and passenger cell are shownin Figure 2. The composite parts are made from layers of carbon preimpregnatedwith an epoxy matrix. The layers are laid up in a mould before being sealed ina bag and placed under vacuum. The sealed part is then placed in an autoclavewhere it is subjected to a specified temperature and pressure cycle required to curethe matrix material. The passenger cell also incorporates sandwich structures wherelayers of composite material are placed on either side of a plastic or aluminium foammaterial to provide an extremely stiff and lightweight part. The engine carrier atthe rear of the car is bolted to the passenger cell to form the main load carryingstructure. The external body panels are bolted to the chassis structure and are non-structural parts. The manufacturing process is labour intensive and not suitable forhigh volume production.

Figure 2: The CFRP monocoque structure in the Porsche Carrera GT.

In addition to the use of lightweight composite materials in high end sports vehi-


cles, technologies are being developed to allow these materials to be utilised in highervolume road vehicles. However, the current use of large structural composite partsin passenger cars has been limited to concept vehicles. European automotive man-ufacturers have presented demonstrator parts which have been developed to provetechnologies and concepts. The BMW Z22 concept car has a CFRP occupant com-partment and aluminium space frame structure. The body structure comprises ofapproximately 20 components including the CFRP side frame as shown in Figure 3.The side frame is manufactured using a liquid moulding technique. The weight ofthe occupant compartment is reduced by approximately 50% in comparison to asteel-based construction [6].

(a) BMW Z22 concept car. (b) CFRP side frame.

Figure 3: BMW Z22 concept car with carbon fibre reinforced plastic side panel.

The CFRP floor panel shown in Figure 4 was developed within the EuropeanUnion financed research project TECABS [7] (Technologies for Carbon Fibre Re-inforced Modular Automotive Body Structures). Automotive manufacturers withinthe project consortium include Volkswagen, Renault and Volvo. The demonstratorpart was designed and manufactured with the number of parts and weight reducedby 30% and 50% respectively utilising technologies which allow for 50 units/dayproduction. The part utilises technological advances in heavy tow carbon fibre,preform technology and resin transfer moulding (RTM).

An example of lightweight construction in series production is the CFRP roofstructure of the BMW M3 CSL. The roof structure is manufactured in three stageswith the first stage involving the preforming of carbon fibre fabric as shown inFigure 5(a). The second stage involves the resin injection process (RTM) where thepreformed fabric is placed into a closed mould and resin is injected into the materialunder pressure. Finally, the part is cured in a heated mould before being removedby a robot and coated with clear paint (see Fig 5(b)). The CFRP roof is 6 kg(50%) lighter than the conventional roof and due to its location, it lowers the car’scentre of gravity. In addition to the roof structure, the M3 CSL includes additionallightweight components made from composite materials. CFRP is used for the frontapron, doors and side trim, air intake, central console and diffusers. Glass fibrereinforced plastic (GFRP) is used for the rear bumper support, front seats and the

4 G.Kelly

Figure 4: CFRP floor panel demonstrator part developed within the TECABS project [7].

support structure for the through loading facility. The rear lid with the integratedspoiler is made from SMC. Altogether, the integration of the lightweight structuralcomponents has led to a reduction in vehicle weight of 200 kg in comparison to theM3 series production model [8].

(a) Preform manufacture for the CFRProof.

(b) Automated part removal from theRTM mould.

Figure 5: Series production of the CFRP roof in the BMW M3 CSL.

The introduction of load bearing composite parts in automotive structures placesnew demands on materials, manufacture and assembly processes. In order to realisehigh-volume production of composite automotive body structures, new material andmanufacturing concepts will have to be utilised. The following section presents thecurrent most promising material system for production of structural automotiveparts in fibre reinforced plastics.


1.1. Materials for structural automotive parts

The introduction of high performance reinforced plastic components in automotivestructures will be significantly dependent upon the raw material costs and the man-ufacturing process. While Formula 1 cars and luxury sports cars can motivate theuse of materials and technologies developed within the aerospace industry, new sys-tems will have to be developed if polymer composite materials are to compete withmetallic materials in the production of high-volume vehicles. A potential solutionto manufacture composite automotive structures involves the use of liquid mouldingtechniques together with non-crimp fabric (NCF) reinforcement. Non-crimp fabricreinforcement is manufactured from fibre tows which are placed side by side to form alayer. The layers are subsequently stacked on top of each other and stitched throughtheir total thickness. The resulting mat commonly consists of 2 to 4 layers and canbe easily handled and draped. The manufacturing of multi-axial reinforcement fromnon-crimp fabric material is illustrated in Figure 6.

Figure 6: Manufacturing of multi-axial non-crimp fabric reinforcement.

Non-crimp fabrics also provide additional advantages such as unlimited shelf-life,high deposition rates and relatively low cost. However, there is a general concernthat the mechanical properties of laminates made from non-crimp fabrics are lowerin comparison with laminates made from pre-preg material. A number of studieshave been conducted [9–15] and comparisons made to pre-preg laminates. Furtherresearch is required in order to quantify the difference in material properties basedon relevant base data.

A liquid moulding technique which can be used to mass produce large, complex,composite parts is resin transfer moulding (RTM). The process involves the infusionof a reinforcement mat with low viscosity resin which is injected under pressure.The dry (unimpregnated) reinforcement is commonly pre-shaped and oriented intoa skeleton of the actual part known as a preform, which is inserted into a matcheddie mould. The mould is then closed, allowing the resin to be injected. Heat isapplied to the mould to activate polymerisation mechanisms that cure the resin and

6 G.Kelly

solidify the part. Once the part develops sufficient handling or green strength, it ismoved or demoulded. Several benefits offered by the RTM process are :

• good laminate surface quality

• ability to manufacture large, complex shapes

• integration of ribs, cores and inserts

• use of a wide range of resin systems and reinforcements

• controllable fibre volume fraction

A major challenge in the design of structural automotive parts made from fi-bre reinforced plastics is the joining and load introduction aspect. This can arisethrough joining composite parts to metallic space-frames, or in attachment of sec-ondary structures such as seat mountings and suspension systems. Introducing loadsinto composite laminates has been found to be more complicated than in metallicmaterials due to material anisotropy, poor interlaminar strength and the absence oflocal stress relief such as plasticity. While a number of load attachments and jointscan be integrated into the components which are manufactured by RTM, the needto join or attach secondary structures is unavoidable. Thus, the development ofjoining techniques for composite structures is necessary.

2. Joining techniques for composite laminates

The joining of composite laminates has been traditionally achieved by mechanicalfastening or adhesive bonding. In selecting which method to adopt, the followingaspects should be considered:

• Structural/material constraints - material thickness, dissimilar materials, loadtype, load level, stress concentrations, operating temperature and environmen-tal exposure.

• Design/aesthetic constraints - weight, disassembly and need for smooth exter-nal joint surface.

• Manufacturing constraints - assembly time, cost, surface preparation and holedrilling.

A qualitative comparison of mechanical fastening and adhesive bonding of com-posite laminates was presented by Astrom [16]. In addition to adhesive bondingand mechanical fastening, hybrid joints combining both techniques offer alternativedesign possibilities. Joints can be designed where the adhesive and bolt share theload with the load transfer distribution depending on the joint design and operatingenvironment. Limited published literature exists on the topic of hybrid joints incomposite laminates. This has resulted in a lack of knowledge regarding the loadtransfer in the joints and the strength and fatigue life. Paper B and Paper C inthis thesis are aimed at addressing this through characterising the load transfer,strength, failure and fatigue life of hybrid single-lap joints with CFRP adherends.

A summary of the important factors affecting the design of mechanically fas-tened, adhesive bonded and hybrid joints are presented in the following sections.


2.1. Mechanical fastening

Mechanically fastened joints, where two or more components are joined using boltsor rivets have been the subject of intense research during the last 30 years. The the-oretical and experimental research has mainly focused on high performance jointstailored to the aerospace industry. Significant contributions to the development ofmechanically fastened joints in composite laminates have been published in sympo-sium proceedings [17], comprehensive texts [18] and literature reviews [19–21].

Single-lap joint Double-lap joint

Single-strap joint

Figure 7: Mechanically fastened joint configurations.

Typical mechanically fastened joint configurations are shown in Figure 7. Themost commonly used joint configuration is the single-lap joint. While the structuralperformance of the joint may be poor as a result of the eccentric load path, thejoint is cost-effective and relatively simple to manufacture. This makes the single-lap joints attractive to the automotive industry and therefore the main part of thisthesis is devoted to this joint configuration.

There are many factors which influence the behaviour of fastened joints in com-posite materials. These factors have been highlighted by a number of authors [20–22]as :

1. Material parameters - fibre type and form, fibre orientation, resin type, stack-ing sequence and fibre volume fraction.

2. Design parameters - joint type, laminate thickness, geometry (width, holediameter, edge distance etc.), load direction and loading type.

3. Fastener parameters - fastener type, fastener size, fastener material, fastenerpattern, clamping force, washer size, fastener/hole fit.

Relevant material, design and fastener parameters which are of interest whenjoining composite laminates in automotive applications are investigated in Paper Aof this thesis.

Laminate bearing strength

The laminate bearing strength is a key factor governing the design of mechanicallyfastened joints. The maximum load which can be transferred by the fastened joint

8 G.Kelly

is governed by the ultimate bearing strength (σub ) of the laminate, which is defined

as

σub =

P u

dt(1)

where Pu is the ultimate failure load, d the hole diameter and t the thickness ofthe laminate. A common test used to determine the bearing strength compositelaminates is the pin bearing test, as shown in Figure 8. Relevant test standardswhich describe procedures to perform material bearing strength tests are ASTM [23]and ASTM D5961/D5961 M [24]. The pin bearing test essentially involves replacingthe fastener by a pin in a double lap joint configuration. Measurement of the loadand hole deformation is used to characterise the bearing strength of the material.The bearing strength can be determined by measuring in the stress at 4% holedeformation as outlined in [23] or by using the 2% offset strength procedure outlinedin [24].

Laminate

washer

pin

Figure 8: Pin bearing test set-up to measure material bearing strength.

There are several factors which influence the bearing strength of composite lami-nates. The joint geometry, laminate stacking sequence, clamping load and bolt-holeclearance have been shown to affect the laminate bearing strength. The effects ofeach of these factors are discussed in the following sections.

Geometrical effects

Geometrical parameters such as the bolt hole diameter d, laminate thickness t, widthw and edge distance e have been shown to significantly affect the strength of boltedjoints in composite laminates [25–30]. The commonly adopted terminology for jointgeometrical parameters is illustrated in Figure 9.

The bearing strength of a composite laminate has been shown to vary with theratio of edge distance to bolt diameter ratio (e/d) as shown in Figure 10. At lower


d

e

w

Bearing plane

Figure 9: Geometry related terminology for bolted joints in composite laminates.

values of e/d, the laminate fails in shear out mode where the high shear stressconcentration at 90◦ to the bearing plane results in the material behind the holebeing displaced. Shear-out failure can occur at larger end distances in laminatescontaining a high percentage of fibres in the 0◦ direction. At larger end distances,the material fails in bearing mode where compression loading leads to elongation ofthe hole. At higher values of e/d, the bearing strength remains essentially constant.

0 1 2 3 4 5 60

100

200

300

400

500

600

700

800

Ulti

mat

e B

earin

g S

tren

gth

(M

Pa)

σ

u b

e/d

???

[0/45/90/-45]s pin-loaded[0/45/90/-45]s clamped T=5Nm[0/45/90/-45]s2 pin-loaded[0/45/90/-45]s2 clamped T=5Nm

Shear-out Failure

Bearing Failure

Figure 10: Effect of the e/d ratio on the laminate bearing strength (from [30]).

The effect of the width to diameter ratio (w/d) on the bearing strength is shownin Figure 11. For small values of w/d, net-section failure occurs with the laminatefracturing in a direction perpendicular to the direction of applied loading. As thew/d ratio is increased, a transition to bearing failure mode occurs. At higher valuesof w/d, the laminate bearing strength remains unchanged.

The necessary values of e/d and w/d to achieve bearing failure have been shownto be strongly dependent upon the laminate stacking sequence. Collings [25] deter-mined the required ratios for a number of different stacking sequences for CFRPlaminates as listed in Table 1. Similar results for the [0◦/90◦] and [0◦/ ± 45◦] lami-

10 G.Kelly

0 1 2 3 4 5 6 7 80

100

200

300

400

500

600

700

800

w/d

???

[0/45/90/-45]s pin-loaded[0/45/90/-45]s clamped T=5Nm[0/45/90/-45]s2 pin-loaded[0/45/90/-45]s2 clamped T=5Nm

Ulti

mat

e B

earin

g S

tren

gth

(M

Pa)

σ

u b

Bearing Failure

Net-section Failure

Figure 11: Effect of the w/d ratio on the laminate bearing strength (from [30]).

nates have been shown by other authors [20, 31]. However, these values are depen-dent on the relative proportions of plies in each direction and may therefore differdepending on the tested laminate.

Table 1: Required w/d and e/d ratios for bolted joints in CFRP [25].

Stacking Sequence w/d e/d[0◦/90◦] >4 >5[±45◦] ≥8 >4

[0◦/ ± 45◦] ≥4 ≥4[0◦/60◦] ≥4 ≥4

Effect of stacking sequence

The laminate stacking sequence allows the designer to tailor laminates for appliedloads and this possibility has been used extensively in the design of composite joints.Stacking sequence has been shown to influence the strength of joints significantlywith regard to both strength and failure mode. Collings [25] in his early workreported that while stacking sequences had a lesser effect on the shear-out and net-section strengths, the bearing strength was sensitive to stacking sequence. Subse-quent research has shown that all failure modes are dependent on stacking sequence.Arnold [22] and Hamada [32] expressed the importance of 0◦ plies on the bearingstrength with Hamada proposing that the strength was dependent on the degreeof interlaminar fracture versus fibre buckling in the 0◦ plies. The pseudo-clampingeffect obtained through employing plies at 90◦ to the load direction at the outersurfaces, has been shown both experimentally [22, 33, 34] and numerically [33, 35]for both GFRP and CFRP laminates. Applying 90◦ plies at the outer surface actsin reducing interlaminar normal stresses and impedes delamination. In contrast to


the consensus in the published literature, Hamada [32] found that the highest jointstrengths were achieved through adopting 0◦ on the outer surfaces of the laminate.

The role of ±45 plies in diffusing the load path in bolted joints has been reportedby a number of authors [22, 27, 32]. This has led to the current belief that efficientjoint design is only achieved when quasi-isotropic stacking sequences are used. Hart-Smith suggested the stacking sequence design guideline illustrated in Figure 12.The diagram has resulted from the author’s comprehensive experimental researchon joints in carbon fibre-epoxy laminates. Hart-Smith [27] also reported the effect ofreducing stress concentrations through use of glass fibre strips around bolt holes inCFRP. Through locally substituting 0◦ CFRP fibres by softer GFRP strips (aboutfour hole diameters wide), experimental evidence has shown the improved strengthof bolted joints in these hybrid laminates.

Figure 12: Selection of ply stacking sequence for fibrous composites [28].

It has also been reported [36] that plies of similar orientation should be dis-persed throughout the laminate and not ‘blocked’ in order to achieve higher bearingstrength. Thus the most efficient laminates, considering all possible failure modes,have a dispersed quasi-isotropic stacking sequence. Cross-ply laminates have beenstudied in tension by Marshall et al. [35] and in bending by Chen and Lee [33]with both papers concluding that [90◦n/0◦n]s laminates exhibit higher ultimate jointstrength compared to [0◦n/90◦n]s laminates.

Lateral clamping load

The effect of lateral (through-thickness) load around the bolt hole has been shownto be beneficial with regard to joint bearing strength [20, 25–28, 30, 31, 35, 37, 38].Collings [25] assumed the stress induced by clamping to be constant over the washerarea and given by

σzz =T

0.2 d π4

(D2 − d2)(2)

where T is the applied torque, D is the outer diameter of the washer, d is the holediameter and ‘0.2’ is a torque coefficient determined by Stewart [39]. Strength im-

12 G.Kelly

provements for 3 mm thick CFRP laminates of up to 170% were found for a constantlateral pressure of 22 MPa [25]. It was noted that little further improvement wasachieved through increasing the lateral pressure. In contrast, Smith et al. [31] foundthat for 1 mm CFRP laminates covering a range of stacking sequences, pressuresgreater than 35 MPa were required to achieve maximum bearing strength. Akay [40]supported the latter findings in a study on various woven and unidirectional CFRPlaminates. Slightly different behaviour was discovered for aramid fibre reinforcedplastic (AFRP) laminates with Akay and Kong Ah Mun [38] reporting that bear-ing strengths could be increased by approximately 100% by mere finger-tightening(1-2 MPa) in comparison to the pin-loaded condition. Transition from bearing totensile failure was also noted upon applying finger-tight lateral pressure. Similartransitions have also been noted for CFRP [40] but at significantly higher lateralpressures (>25 MPa) depending on the stacking sequence.

Bolt-hole clearance

The clearance between the bolt and the hole significantly influences the behaviourof mechanically fastened joints. Several authors have investigated the effect of bolt-hole clearance on the behaviour of mechanically fastened joints in composite lami-nates [41–48]. In general, increased clearance was found to reduce the contact areabetween the bolt and the hole thus increasing the stress levels. However, it is diffi-cult to compare the results from different authors as the problem is non-linear beingdependent upon the applied load, friction and the bolt-hole contact area.

The majority of the published work in this area has considered the pin-loadedjoint representing a double-lap joint configuration. However, the effect of bolt-holeclearance in a single-lap joint is potentially more important as the clearance allowsthe bolt to rotate in the hole. Bolt rotation has a detrimental effect on joint perfor-mance as a result of the reduction in the bolt-hole contact area. Ireman [47] investi-gated the effect of different design parameters on the strength of a CFRP/aluminiumsingle bolt, single-lap joint using a three dimensional finite element model. Whileclearance was found to decrease the joint strength, the laminate thickness, friction,clamping force and bolt configuration were found to have a stronger influence on thejoint strength. The effect of clearance in single-bolt, single-lap joints has been in-vestigated by McCarthy et al. [48]. CFRP joints were manufactured with clearancesranging between neat-fit and 240 µm. The 2% offset strength, ultimate strength andultimate strain were determined according to ASTM D5961/D5961M [24]. Increasedbolt-hole clearance was found to decrease the joint stiffness and increase the ulti-mate failure strain. The effect of clearance on the 2% offset strength and ultimatestrength was found to be small for the clearance levels considered. An alternativedefinition of strength was derived based on a decrease in the joint stiffness.

Bolt-hole clearance can significantly affect the load distribution in multi-rowjoints. A neat fit between the bolt and the hole will ensure that the load is transferredimmediately. In contrast, the load transfer is delayed in clearance fit joints untilcontact has been established. Stanley et al. [49] investigated the load distributionin multi-row single-lap joints concluding that relatively small amounts of clearancecan have substantial effects on the bolt load distribution.

With mass production of modular composite automotive structures, the manu-


facturing and assembly tolerances will become important parameters which will haveto be considered in the design process. While larger tolerances may be advantageousfor automated manufacture and assembly operations, knowledge of the subsequenteffect on the mechanical properties of fastened structures is necessary. The effectsof bolt-hole clearance on the bearing strength of carbon fibre epoxy laminates wasinvestigated in Paper A.

2.2. Adhesive bonding

Adhesive bonding has been used for many years as a method of providing highstrength structural joints. Kutscha and Hofer [50] and Kutscha [51] reviewed muchof the early research in this field prior to 1969 including many of the classical an-alytical methods. Lehman and Hawly [52] also provided one of the earlier papersanalysing and comparing different joint configurations. Oplinger [53] and Vinson [54]have since provided general reviews of the technology of adhesive bonded joints incomposite structures. The science of adhesive bonding encompasses many fields suchas surface physics, chemistry and mechanics. Research in these fields has led to theproposal of different theories which describe the mechanisms governing adhesion.The theories are presented in the comprehensive text by Kinloch [55].

Single-lap joint Double-lap joint

Scarf jointStepped-lap joint

Figure 13: Adhesive bonded joint configurations.

A wide variety of joint configurations are possible when bonding composite struc-tures. Figure 13 illustrates the most common joint configurations. The single-lapand double-lap configurations are the most commonly found joints in practice andare applicable for joining relatively thin adherends. The more advanced stepped-lap and scarf configurations are used to transfer high loads in joints with thickeradherends. The manufacture of stepped-lap and scarf joints is more complicated,requiring low machining tolerances.

The most extensively studied joint configuration in terms of experimental andtheoretical research is the single-lap joint. The joint is illustrated in more detail inFigure 14. The joint is commonly used due to its simple design and easy assemblyin a similar manner to the mechanically fastened single-lap joint. Figure 14 alsohighlights the stress distribution in the adhesive as a result of in-plane loading.The highest stresses occur at the ends of the overlap as shown. As the load is notcollinear in the joint, a moment is generated at the overlap and the joint rotateswhen deformed. This rotation induces additional peel stresses on the adhesive which

14 G.Kelly

can reduce the strength of the joint significantly. Methods to limit or overcome peelstresses and increase joint strength are discussed in a later section discussing endeffects. Additional factors affecting the strength of bonded joints are discussed inthe following sections.

x=0

Load path through the joint

Undeformed joint

Deformed joint

−15 -10 −5 0 5 10−10

−5

0

5

10

15

20

25

30

35

40

45

x (mm)

Str

ess

(MP

a)

Shear stressPeel stress

15

Stress distribution in the adhesive

Figure 14: Structural behaviour of bonded single-lap joints.

Factors affecting the strength of bonded joints

There are many factors which affect the strength of adhesive bonded joints. Inthe following sections, the most important factors affecting the strength of bondedsingle-lap joints in automotive applications are discussed.

Overlap length

The overlap length of an adhesive joint defines the length by which the adherendsoverlap one another. The overlap length is an important design parameter whichinfluences joint rotation, the adhesive stress distribution and ultimately the strengthof the joint. Increasing the overlap length will serve to reduce the rotation of thejoint. The effect of the overlap length on the stress state in an adhesive joint has beenstudied by Hart-Smith [56–58] and Grant [59,60] within research programs aimed atdeveloping design philosophies for aircraft construction. Hart-Smith’s work withinthe PABST (Primary Adhesive Bonded Structures) program and ESDU (Engineer-ing Science Data Unit) programs based on Grant’s work at British Aerospace have


served to provide guidelines with regard to overlap lengths in adhesive joints togetherwith analysis techniques for several common joint configurations. The effect of over-lap length on the adhesive shear stress in a single-lap joint is shown in Figure 15.

-15 -10 -5 0 5 10 150

5

10

15

20

25

30

35

40

45

50

Overlap length, L (mm)

She

ar S

tres

s(M

Pa)

Figure 15: Shear stress distribution in a bonded single-lap joint.

The magnitude of the shear stress in the adhesive is shown to increase withdecreasing overlap length. The stress in the central region of the overlap is lowwith the peak stresses occurring at the ends of the overlap. In the case of shortoverlap joints, the adhesive may yield along the entire bond line when the maximumtheoretical load has been applied. Increasing the overlap length allows for an “elastictrough” to develop at the centre section of the joint. The length over which the bondyields is constant and dependent upon the load being transferred. Further increasein the overlap length serves only to increase the size of the elastic region and reducethe stress at the centre of the joint. The design philosophy adopted by Hart-Smithin the PABST program was that the adhesive should never be the weak link in thestructure. The elastic trough region is deemed essential if the joint is to overcomethe effects of creep and the plastic zones are designed to be long enough to supportthe ultimate load under the worst operating conditions. A combination of moistureand high temperature provide the worst operating conditions for adhesives due tothe softening which takes place. It was recommended that the minimum operatingstress (in the elastic trough region) be no greater than 10% of the adhesive shearstrength. While an overlap length to adherend thickness ratio of 8:1 is suitable forlaboratory lap-shear testing, a ratio of 80:1 was chosen in the PABST program as adesign criterion for aircraft structures.

16 G.Kelly

End effects

Critical regions exist at the ends of the overlap region in bonded joints. The maxi-mum stress occurs in these regions as illustrated in Figure 15. Several authors [61–64]have modified the adherend or adhesive geometry at the end of the overlap as ameans of increasing the joint strength.

Adams and Harris [64] investigated the modification of the overlap ends in alu-minium/epoxy single-lap joints. The performance of joints with square edged ad-herends (basic design in Fig. 16), rounded adherends and adhesive fillets were com-pared. The joints with the square edges demonstrated the lowest strength with astrength increase of 28% being observed in the joints with the adhesive fillet. Thehighest strength was found in the joints with the rounded adherends and adhesivefillet where a strength increase of 60% was observed in comparison to the squareedged joints. The strength predictions from finite element analysis were found to bein good agreement with the experimental results.

An additional problem which can be alleviated through tapering the adherendsat the ends of the overlap is that of peeling stress. Hart-Smith [61] showed thatprofiling the adherend could reduce the peel stresses, change the failure mode, andhence increase the joint strength. The author noted that the taper must be con-tinued to a fine edge if the benefits are to be gained. Adherend tapering may bedifficult to accomplish in practice for thin adherends. Hildebrand [63] and Adams etal. [62] investigated various joint designs for single and double-lap joints respec-tively. A selection of alternative designs for the ends of the overlap region are shownin Figure 16.

Basic design Adhesive fillet Outside taper

Inside taper Inside taper and adhesive fillet

Figure 16: Alternative modifications to the end of adhesive lap joints (Redrawn from [62]).

The possibility to modify to the ends of the overlap depends on the adherendthickness, the manufacturing process and cost. In general, adherend tapering maybe difficult to achieve but ensuring a well profiled fillet is essential in the design ofa good adhesive joint.

Effect of the surface ply fibre orientation

One factor which can easily be overlooked when designing joints is the orientationof the fibre direction in the surface ply. The surface ply fibre orientation has animportant influence on the resulting joint strength. Ideally, the fibre orientation ofthe surface ply should be aligned in the direction of applied loading. However, whenmanufacturing large complex shapes with a minimal number of preforms, the surface


fibre directions may be dictated by part manufacture considerations. As joints maybe required at different locations in the same structure, the surface fibre directionsmay differ resulting in non optimal surface fibre directions for some bonded surfaces.

The load-displacement response of bonded joints with surface fibre directionsin the loading direction and perpendicular to the load direction is illustrated inFigure 17.

0 0.5 1 1.5 2 2.50

2

4

6

8

10

12

14

Grip displacement (mm)

Fra

ctur

e lo

ad (

kN)

Surface ply parallel to the load direction

Surface ply perpendicular to the load direction

Figure 17: Effect of the surface ply orientation on joint strength. Single-lap joint withCFRP adherends [0/45/90/-45]s and Epibond 1590 A/B adhesive.

When the surface fibre direction coincides with the loading direction, the strainon the surface of the laminate is limited due to presence of the fibres. The strengthof the ply is also highest in the fibre direction. Hence, the reduced strain andhigh strength act to transfer the load from the adhesive into the laminate in anefficient manner. In contrast, the surface strain in the loading direction of jointswith the surface ply aligned perpendicular to the direction of the applied load ishigher due to the absence of any reinforcement with the strain being governed bythe matrix mechanical properties. The strength is also significantly lower whichresults in fracture and delamination of the surface ply at relatively low load levels.As shown in Figure 17 for the case of a single-lap joint bonded with Epibond 1590A/B epoxy adhesive, the surface fibre orientation can result in a decrease in strengthof approximately 50%. A possible alleviation of the surface ply orientation problemin complex structures made using a limited number of preforms is to use a wovenfabric on the surface of the laminate.

Surface preparation techniques

Surface preparation of the adherends is of critical importance when adhesive bond-ing. Polymers are low surface energy materials and adhesives do not bond easilyto polymer surfaces. The adhesion is also affected by the additives contained in

18 G.Kelly

many industrial materials including plasticisers and anti-oxidants. Release agentsused during manufacture also serve to contaminate the bonding surfaces. A numberof techniques to improve the bonding of polymers and polymer matrix compositeshave been developed. These methods essentially remove or modify the low molecu-lar weight layer which exists at the surface of many polymeric materials. It is alsobeneficial to modify the surface topology to allow capillary action by the adhesiveon the surface. Several of the most commonly used surface preparation methodsfor polymer-matrix composites are solvent wiping and abrasion, corona dischargetreatment, cold plasma treatment and use of peel ply during composite laminatemanufacture. The surface preparation technique adopted for the laminate in thisthesis was solvent wiping and abrasion.

Fatigue behaviour and environmental effects

The performance of adhesive bonded joints has been investigated extensively duringthe last 30 years which has led to an increased understanding of the joint behaviourand factors affecting joint strength. However, significant challenges remain to de-velop theoretical models to allow prediction of the behaviour of aged joints and jointssubject to fatigue loading. The effects of temperature and moisture have been shownto decrease the strength and fatigue resistance of adhesive bonded joints [65–79].The environment can alter the properties of the adhesive, the composite laminateor the adhesive/adherend interface. The combination of elevated temperature andmoisture is particularly detrimental with moisture diffusing into the material at afaster rate due to the increased mobility of the polymer chains.

One practical problem of conducting environmental tests is the long exposuretimes which are required. In order to overcome time constraints, laboratory ageingconditions are commonly modified to allow for shorter ageing times. In order topredict the long term performance of adhesive bonded joints, accelerated ageing isoften performed by subjecting the joints to more severe conditions than would beexpected in normal service. The ageing of adhesive bonded joints may be acceleratedby environmental conditions such as mechanical stress, high temperature, moistureand aggressive media. While providing a useful guide to the long-term behaviourof joints, care must be taken to ensure that accelerated ageing does not introduceadditional factors such as different failure modes than those encountered duringservice.

The strength and fatigue resistance of CFRP single-lap joints with a structuralpolyurethane adhesive were investigated in Paper D. Accelerated ageing of the jointswas performed in different environments considering the effects of constant and cyclictemperature and humidity levels. The effect of environmental ageing was found todecrease the strength and shorten the fatigue life of the joints significantly. Both thepolyurethane adhesive and carbon fibre/epoxy laminates were affected by ageing incyclic freeze/thaw and hot/wet environments.

Currently there are no generalised models developed due to the vast range ofadherend and adhesive materials available. The joint configurations, environmen-tal conditions and material history are also factors which make prediction of jointperformance a significant challenge.


2.3. Hybrid joining

An alternative method to mechanical fastening and adhesive bonding is hybrid join-ing which combines both techniques. Previous investigations of hybrid joints incomposite laminates have focused on aerospace applications and not considered thetechnique to offer any significant benefits except for limiting damage propagation inadhesive bonded joints [80–82]. However, the combination of mechanical fasteningand adhesive bonding has potential advantages in production of automotive struc-tures with the parts being fixed in position during the bonding process, allowingthe joined part to continue to the next stage in production without having to waitfor the adhesive to harden. Another important factor is the behaviour of joints dur-ing crash loading. Bonded joints with relatively short overlap lengths can debondcausing separation of the parts. Hybrid joining provides an alternative method toabsorb energy during crash loading and can limit part separation.

A hybrid single-lap joint is illustrated schematically in Figure 18. The distribu-tion of load between the adhesive and the bolt is related to the relative displacementbetween the adherends. The adhesive provides the stiffest load path and thus trans-fers the majority of the applied load. However, as the joint deforms and the relativedisplacement between the adherends increases, load is transfered to the bolt. Theaim of Paper B was to investigate the parameters which govern load transfer in hy-brid joints and identify joint configurations where the technique is viable. A finiteelement simulation model was developed and used to predict the load distributionin hybrid joints. A parameter study highlighted the effect of the joint geometry andadhesive material behaviour on the load transferred by the adhesive and the bolt.

Bolt

Adhesive

Adherend

Figure 18: Hybrid (bolted/bonded) single-lap joint.

In addition to predicting the load transfer in hybrid joints, it is important tocharacterise the strength and fatigue life of hybrid joints. The strength, failuremodes and fatigue life of hybrid joints were investigated in Paper C. It was shownthat the joints could be designed to have greater strength and fatigue life in com-parison to adhesive bonded joints. However, it was found that benefits were onlyobtained in joints which allowed for load sharing between the bolt and adhesive.The relative strengths of the bolted and bonded joints was also shown to influencethe ultimate failure mode and fatigue life.

In Paper D, hybrid joints demonstrated extended fatigue life in comparison toadhesive bonded joints when the joints were subject to environmental ageing.

20 G.Kelly

3. Localised transverse loading in composite lam-

inates

Attachments in CFRP structures often require that loads are introduced into thelaminate in the transverse or thickness direction. This is arguably the most difficultload case for composite laminates, given the poor interlaminar and through-thicknessstrength of the material. Localised transverse loads are commonly avoided in com-posite laminates and this is reflected in the volume of published literature.

Important contributions regarding this topic have been made by Waters andWilliams [83] and by Banbury et al. [84,85]. Waters and Williams [86] conducted anearly experimental investigation of bolt push-through failure in composite laminates.The authors investigated the response and failure modes of a wide range of compositelaminates. Clamped circular specimens of diameter 25.4 mm and 50.8 mm wereloaded in the transverse direction through a countersunk fastener. The characteristicfailure modes were found to be intralaminar fibre failure in the vicinity of the bolthole and intralaminar and interlaminar matrix failure in the laminate interior. It wasalso concluded that the strength of transversely loaded composite laminates couldbe improved through the use of tough resins, high-strain fibres and low modulusadhesive interleaving.

Banbury and Kelly [84] conducted a two part study investigating the pull-throughstrength of carbon fibre reinforced epoxy laminates made from woven fabric andunidirectional tape pre-preg material. The authors provided a detailed investiga-tion of the observed failure modes and discussed the mechanisms responsible forpull-through failure. The internal damage pattern in the laminates was likened tothat observed in composite laminates subject to low velocity impact. Failure wasattributed to a critical level of matrix strain which was independent of the laminateand fastener geometry or material type for a given resin system. The authors alsoconducted a two-dimensional axi-symmetric finite element analysis to simulate thefastener pull-through tests [85]. A progressive damage model based on the maximumstrain criterion was used to predict failure in the laminates although no consider-ation was given to interlaminar failure modes. The results from the model wereconsistent with the experimental observations and the predicted failure loads wereshown to be in good agreement with the experimentally determined values.

The introduction of localised transverse loads in composite laminates based onnon-crimp fabric reinforcement was investigated in Paper E. The effect of the lami-nate stacking sequence, resin system and specimen size on the strength and failuremechanisms were investigated experimentally. In addition, a three dimensional fi-nite element model was developed to analyse the stress distribution in the laminateand predict the first-ply failure load.

There is currently no widely accepted test method for evaluating the perfor-mance of transverse load introductions. Two different methods are suggested in theComposite Materials Handbook (MIL-HDBK-17) [87], however, it is acknowledgedthat these methods may not be representative of actual load introductions. The firstmethod is an adaptation of MIL-STD-1312 Test 8 and this method is suggested forscreening purposes only. The second proposed method is similar to that employedin Paper E and this method is more suitable for design purposes.


4. Future work

Localised transverse loading remains a challenge in designing automotive structuresusing laminated composite materials. The work performed in Paper E highlightedthat damage initiates in composite laminates at low load levels. The propagationof damage during fatigue loading and residual strength are topics which need to beexplored. The attachment of secondary structures may involve loading the laminatein the thickness direction. Future research in this area should focus on transferringthe load into the plane of the laminate with a tailored geometry or fibre reinforcementarchitecture. Textile reinforcement with fibres in three directions offer a potentialsolution to this problem.

Future automotive structures will be modular and incorporate a variety of dif-ferent materials, each tailored to meet the required structural demands. The de-velopment of technologies for joining multi-material structures combining metallic,plastic and composite parts will be necessary. Adhesive bonding and hybrid join-ing are attractive methods to join such structures in a cost-effective manner. Asa continuation of the work presented in Paper B and Paper C in this thesis, thenext step would be to apply the outlined principles on joints with different adherendmaterials.


References

[1] Woisetschlaeger E. Keine monokultur. Automobil Entwicklung, September 2001.

[2] Friedrich H.E. Leichtbau und werkstoffinnovationen im fahrzeugbau. AutomobiltechnischeZeitschrift, March 2001.

[3] Gjostein N.A. Technology needs beyond PNGV. In Basic Needs for Vehicles of the Future,New Orleans, Louisiana, 1995.

[4] Eusebi E. Composite intensive vehicles - past, present and future. PNGV Symposium onstructural material challenges, 1995.

[5] Parr T., Wallentowitz H., Wohlecker R., Wynands D., and von Zengen K-H. Leichtbaupoten-zial eines aluminium-intensiven fahrzeugs. Automobiltechnische Zeitschrift, March 2003.

[6] BMW on the road to carbon car. Reinforced Plastics, April 2001.

[7] Technologies for Carbon-Fibre Reinforced Modular Automotive Body Structures (TECABS).Competative and Sustainable Growth Project: G3RD-CT-2000-00102.

[8] BMW takes carbon fiber technology to the next level. Automotive Engineer, November 2003.

[9] Bibo G.A., Hogg P.J., and Kemp M. Mechanical characterisation of glass- and carbon-fibrereinforced composites made with non-crimp fabrics. Composites Science and Technology,57:1221–1241, 1997.

[10] Wang Y., Li J., and Do P.B. Properties of composite laminates manufactured with E-glassmulti-axial non-crimp fabrics. Journal of Composite Materials, 29(17):2317–2333, 1995.

[11] Hogg P.J, Ahmandnia A., and Guild F.J. The mechanical properties of non-crimp fabricbased composites. Composites, 24(5):423–432, 1993.

[12] Dexter H.B. and Hasko G.H. Mechanical properties and damage tolerance of multiaxial warp-knit composites. Composites Science and Technology, 56:367–380, 1996.

[13] Godbehere A.P., Mills A.R., and Irving P. Non-crimped fabrics versus pre-preg CFRP com-posites - a comparison of mechanical performance. In 6th International Conference on FibreReinforced Plastics (FRC’94), University of Newcastle-upon-Tyne, 1994.

[14] Drapier S. and Wisnom M.R. Finite element investigation of the compressive strength ofnon-crimp fabric based composites. Composites Science and Technology, 59:1287–1297, 1999.

[15] Drapier S. and Wisnom M.R. Finite element investigation of the interlaminar shear behaviourof non-crimp fabric based composites. Composites Science and Technology, 59:2351–2362,1999.

[16] Astrom B.T. Manufacturing of Polymer Composites. Chapman & Hall, London, UK, 1997.

[17] Kedward K.T. Joining of composite materials, ASTM Special Publication 749. AmericanSociety for Testing and Materials, 1980.

[18] Matthews F.L. Joining fibre-reinforced plastic composites. Elsevier Applied Science, 1987.

[19] Poon C. Literature review on the design of composite mechanically fastened joints. TechnicalReport AENOE5, National Aeronautical Establishment, National Research Council, Canada,1986.

[20] Godwin E. W. and Matthews F. L. Review of the strength of joints in fibre-reinforced plastics:Part 1 Mechanically fastened joints. Composites, 11(3):155–160, 1980.

[21] Camanho P.P. and F.L. Matthews. Stress analysis and strength prediction of mechanicallyfastened joints in FRP: a review. Composites Part A, 28(6):529–547, 1997.

24 G.Kelly

[22] Arnold W.S., Marshall I.H., and Wood J. Optimum design considerations for mechanicallyfastened composite joints. Composite Structures, 16:85–101, 1990.

[23] ASTM D 953-87 Standard Test Method for Bearing Strength of Plastics. American Societyfor Testing and Materials, 1987.

[24] ASTM D 5961/D 5961 M Standard Test Method for Bearing Response of Polymer Ma-trix Laminates. American Society for Testing and Materials, 1996.

[25] Collings T.A. The strength of bolted joints in multi-directional CFRP laminates. Composites,8(1):43–55, 1977.

[26] Kretsis G. and Matthews F.L. Strength of bolted joints in glass fibre/epoxy laminates. Com-posites, 16(2):92–102, 1985.

[27] Hart-Smith L.J. Mechanically fastened joints for advanced composites - phenomenologicalconsiderations and simple analysis. Douglas Paper 6748, Douglas Aircraft Company, 1978.

[28] Hart-Smith L.J. Design and analysis of bolted and riveted joints fibrous composite structures.Douglas Paper 7739, Douglas Aircraft Company, 1986.

[29] Chutima C. and Blackie A.P. Effect of pitch distance, row spacing, end distance and boltdiameter on multi-fastened composite joints. Composites Part A, 27(2):105–110, 1996.

[30] Kelly G. and Hallstrom S. Bearing strength of carbon fibre epoxy laminates: Effects of bolthole clearance. Composites Part B, 35(4):331–343, 2004.

[31] Smith P.A. and Pascoe K.J. The effect of stacking sequence on the bearing strengths ofquasi-isotropic composite joints. Composite Structures, 6(1):1–20, 1986.

[32] Hamada H., Haruna K., and Maekawa Z. Effects of stacking sequences on mechanicallyfastened joint strength in quasi-isotropic carbon-epoxy laminates. Journal of CompositesTechnology and Research, 17(3):249–259, 1995.

[33] Chen W.H., S.S. Le, and Yeh J.T. Three dimensional contact stress analysis of a compositelaminate with a bolted joint. Composite Structures, 30:287–297, 1995.

[34] Quinn W.J. and Matthews F.L. The effect of stacking sequence on the pin bearing strengthin GFRP. Journal of Composite Materials, 11:139–145, 1987.

[35] Marshall I.M., Arnold W.S., Wood J., and Mousley R.F. Observations on bolted connectionsin composite structures. Composite Structures, 13(1):133–151, 1989.

[36] Arnold W.S. PhD thesis, Department of Mechanical Engineering, Paisley College of Technol-ogy, Paisley, Scotland, 1989.

[37] Stockdale J.H. and Matthews F.L. The effect of clamping pressure on bolt bearing loads inglass fibre reinforced plastics. Composites, 9:34–38, 1976.

[38] Akay M. and Kong Ah Mun S. Bearing strength of autoclave and oven cured kevlar/epoxylaminates under static and dynamic loading. Composites, 26(6):451–456, 1995.

[39] Stewart W.C. Metals engineering design. ASTM Handbook Second Edition, 1965.

[40] Akay M. Bearing strength of as-cured and hygrothermally conditioned carbon fibre/epoxycomposites under static and dynamic loading. Composites, 23(2):101–108, 1992.

[41] Naik R.A. and Crews J.H. Stress analysis method for a clearance-fit bolt under bearingconditions. AIAA Journal, 24:1348–1353, 1985.

[42] Naik R.A. and Crews J.H. Stress analysis method for a clearance-fit bolt under bearingconditions. In AIAA/ASME/ASCE/AHS 26th Structures, Structural Dynamics and MaterialConference, number AIAA Paper 85-0746, Orlando, Florida, USA, 1986.

[43] Eriksson L.I. Contact stresses in bolted joints of composite laminates. Composite Structures,6(1):57–75, 1986.


[44] Hyer M.W., Klang E.C., and Cooper D.E. The effects of pin elasticity, clearance and frictionon the stresses in a pin-loaded orthotropic plate. Journal of Composite Materials, 21(3):190–206, 1987.

[45] DiNicola A.J. and Fantle S.L. Bearing strength of clearance fit fastener holes in toughenedgraphite/epoxy laminates. In Composite Materials:Testing and Design Vol.11, pages 220–237.ASTM STP 1206, 1993.

[46] Lanza Di Scalea F., Capello F., and Cloud G.L. On the elastic behaviour of cross-ply com-posite pin-joint with clearance fits. Journal of Thermoplastic Composite Materials, 12:13–33,1999.

[47] Ireman T. Three-dimensional stress analysis of bolted single-lap joints. Composite Structures,43(3):195–216, 1998.

[48] McCarthy M.A., Lawlor V.P., Stanley W.F., and McCarthy C.T. Bolt-hole clearance effectsand strength criteria in single-bolt, single-lap, composite bolted joints. Composites Scienceand Technology, 62(10-11):1415–1431, 2002.

[49] Stanley W.F., McCarthy M.A., and Lawlor V.P. Measurement of load distribution in multi-bolt composite joints, in presence of varying clearance. Plastics, Rubber and Composites,31(9):412–417, 2002.

[50] Kutscha D. and Hofer K.E. Feasibility of joining advanced flight vehicle structures. TechnicalReport AFML-TR-68-391, US Air Force, US, 1969.

[51] Kutscha D. Mechanics of adhesive bonded lap type joints: survey and review. TechnicalReport AFML-TDR-64-298, 1964.

[52] Lehman G.M. and Hawley A.V. Investigation of joints in advanced fiberous composites foraircraft structures. Technical Report AFFDL-TR-69-43, 1969.

[53] Oplinger D.W. Proceedings of the fourth army materials technology conference - advances injoining technology. 1975.

[54] Vinson J.R. Adhesive bonding of polymer composites. Polymer Science and Engineering,29(19):1325–1331, 1989.

[55] Kinloch A.J. Adhesion and Adhesives: Science and Technology. Cambridge University Press,1987.

[56] Hart-Smith L.J. Adhesive bonded single lap joints. Technical Report CR-112236, NASA,Langley Research Center, 1973.

[57] Hart-Smith L.J. Adhesive bonded double lap joints. Technical Report CR-112235, NASA,Langley Research Center, 1973.

[58] Hart-Smith L.J. Adhesive bonded joints for composites - phenomenological considerations. InTechnology Conferences Associates - Conference on Advanced Composites Technology, 1978.

[59] Grant P. Stress and strength analysis of bonded joints. Technical Report 50(R)P, BritishAircraft Corparation, 1976.

[60] Grant P. Analysis of adhesive stresses in bonded joints. In Symposium: Jointing in fibrereinforced plastics. IPC Science and Technology Press Ltd., 1978.

[61] Hart-Smith L.J. Design and analysis of adhesive bonded joints. Technical Report Report No.6059A, McDonnell Douglas Company, 1972.

[62] Adams R.D., Atkins R.W., Harris J.A., and Kinloch A.J. Journal of Adhesion, 29, 1986.

[63] Hildebrand M. International Journal of Adhesion and Adhesives, 14, 1986.

[64] Adams R.D. and Harris J.A. International Journal of Adhesion and Adhesives, 7, 1987.

[65] Ashcroft I.A., Hughes D.J., Shaw S.J, Abdel Wahab M.M., and Crocombe A.D. Effect of thetemperature on the quasi-static strength and fatigue resistance of bonded composite double-lap joints. Journal of Adhesion, 75, 2001.

26 G.Kelly

[66] Adams R.D., Coppendale J., Mallick V., and Al-Hamdan H. The effect of temperature on thestrength of adhesive joints. International Journal of Adhesion and Adhesives, 12(3):185–190,1992.

[67] Parker B.M. The durability of adhesive bonded carbon fibre composite. In CompositesBonding ASTM STP 1227. ASTM, 1994.

[68] Brewis D.M., Comyn J., and Tegg J.L. The durability of some epoxide adhesive bonded jointson exposure to moist warm air. International Journal of Adhesion and Adhesives, 35(1):35–39,1980.

[69] Gledhill R.A. and Kinloch A.J. Environmental failure of structural adhesive joints. Journalof Adhesion, 6:315–330, 1974.

[70] Bowditch M.R. The durability of adhesive joints in the presence of water. InternationalJournal of Adhesion and Adhesives, 16(2):73–79, 1996.

[71] Fernando M., Harjoprayitno W.W., and Kinloch A.J. A fracture mechanics study on theinfluence of moisture on the fatigue behaviour of adhesively bonded aluminium-alloy joints.International Journal of Adhesion and Adhesives, 16:113–119, 1996.

[72] Ashcroft I.A., Digby R.P., and Shaw S.J. The effect of the environment on the performanceof bonded joints. In Joining and repair of plastics and composites. IMechE, 1999.

[73] Ashcroft I.A., Abdel Wahab M.M., Crocombe A.D., Hughes D.J., and Shaw S.J. The effect ofthe environment on the fatigue of bonded composite joints: Part 1 testing and fractography.Composites Part A, 32, 2001.

[74] Abdel Wahab M.M, Ashcroft I.A., Crocombe A.D, and Shaw S.J. Diffusion of moisture inadhesively bonded joints. Journal of Adhesion, 77:43–80, 2001.

[75] Gilmore R.B. and Shaw S.J. The effect of temperature and humidity on the fatigue behaviourof composite bonded joints. In Composites Bonding ASTM STP 1227. ASTM, 1994.

[76] Mackie R.I and Su N. The effect of ageing and environment on the static and fatigue strenghof adhesive jonts. Journal of Adhesion, 42:191–207, 1993.

[77] Briskham P. and Smith G. Cyclic stress durability testing of lap shear joints exposed tohot-wet conditions. International Journal of Adhesion and Adhesives, 20:33–38, 2000.

[78] Ferreira J.A.M, Ries P.N., Costa J.D.M., and Richardson M.O.W. Fatigue behaviour ofcomposite adhesive lap joints. Composites Science and Technology, 62:1373–1379, 2002.

[79] Knox E.M. and Cowling M.J. Durability aspects of adhesively bonded thick adherend lapshear joints. International Journal of Adhesion and Adhesives, 20:323–331, 2000.

[80] Hart-Smith L.J. Bonded-bolted composite joints. Journal of Aircraft, 22(11):993–1000, 1985.

[81] Hart-Smith L.J. Design methodology for bonded-bolted composite joints. Technical ReportAFWAL-TR-81-3154, Douglas Aircraft Company, 1982.

[82] Chan W.S. and Vedhagiri S. Analysis of composite bolted/bonded joints used in repairing.Journal of Composite Materials, 35(12):1045–1061, 2001.

[83] Waters W.A. and Williams J.G. Failure mechanisms of laminates transversely loaded by boltpush through. Technical report, NASA Technical Memorandum 87603, 1985.

[84] Banbury A. and Kelly D.W. A study of fastener pull-through failure of composite laminates.Part 1: Experimental. Composite Structures, 45(4):241–254, 1999.

[85] Banbury A., Kelly D.W., and Jain L.K. A study of fastener pull-through failure of compositelaminates. Part 2: Failure prediction. Composite Structures, 45(4):255–270, 1999.

[86] Waters W.A. and Williams J.G. Failure mechanisms of laminates transversely loaded by boltpush through. NASA Technical Memorandum 87603, 1985.

[87] Composites Material Handbook 17. US Army Research, 1999.

joining of carbon fibre reinforced plastics for automotive...

Documents