Resistance Welding of Thermoplastic Composites

(1)

Resistance welding

of

thermoplastic composites

Process and performance

(2)

(3)

Resistance welding

of

thermoplastic composites

Process and performance

Proefschrift

ter verkrijging van de graad van doctor aan de Technische Universiteit Delft,

op gezag van de Rector Magnificus prof. ir. K.C.A.M. Luyben, voorzitter van het College voor Promoties,

in het openbaar te verdedigen op vrijdag 21 november 2014 om 12:30 uur

door

Huajie SHI

Aeronautical and Astronautical Manufacturing Engineer geboren te Yangquan, Shanxi, China

(4)

Prof. dr. ir. R. Benedictus

Copromotor: Dr. ir. H.E.N. Bersee and Dr. I.F. Villegas

Samenstelling promotiecommissie:

Rector Magnificus, Voorzitter

Prof. dr. ir. R. Benedictus, Technische Universiteit Delft, promotor Dr. ir. H.E.N. Bersee, Technische Universiteit Delft, copromotor Dr. I.F. Villegas, Technische Universiteit Delft, copromotor Prof. dr. S.L. Corre, Polytech Nantes, France

Prof. dr. A. Maffezzoli, University of Salento, Italy Prof. dr. S.J. Picken, Technische Universiteit Delft Dr. B. Defoort, Airbus Defence and Space, France Prof. dr. R. Curran, Technische Universiteit Delft, reservelid

ISBN: 978-94-6186-381-2 Copyright c 2014 by H. Shi

All rights reserved. No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means, electronic or mechani-cal, including photocopying, recording or by any information storage and retrieval system, without the prior written permission of the author.

Cover design: H. Shi Photography: F. Oostrum

(5)

(6)

(7)

Summary

One way to improve the emission efficiency of air transportation is to reduce the weight of aircraft, and for this purpose composite materials are been increasingly used in aircraft to replace the metals. Cheaper and faster manufacturing methods for composite structures are therefore required to meet the fast growth of the de-mand of airplanes. Compared to thermoset composites, thermoplastic composites are drawing more and more attention by aircraft industries not only due to their excellent material properties but also due to their potentials to reduce cycle time and structure cost by using low-cost manufacturing technologies such as welding. Resistance welding has been regarded as one of the most promising welding tech-niques owing to the low energy consumption, simplicity of welding operation and capability for scaling up. However, deeper knowledge of resistance welding has to be gained to mature this technology for wider applications in aircraft.

Previous researches on resistance welding of thermoplastic composites are mainly focused on understanding the welding mechanisms and characterizing the welding qualities. The heat transfer, consolidation and crystallization mechanisms of static resistance welding have been analysed, and various testing methods have been used to characterize the mechanical performance of the resistance welded joints. How-ever, there are still some gaps in the current knowledge:

• Firstly, compared to the main welding parameters, such as power input and heating time, less is known on how the specific properties of adherends or heating element can influence the welding quality.

• Secondly, the definitions of processing windows mainly rely on trial-and-error tests, which are costly and time consuming.

• Finally, continuous resistance welding has been invented to scale-up the cur-rent used welding technique, while little is known about the heating mecha-nism of this welding process or how to define the welding parameters involved. The research work was aimed to gain deeper knowledge of resistance welding, in both micro-level and macro-level. In micro-level study, the specific properties that influence the welding quality were discussed, such as the surface properties of adherends, the weld line properties and the welding induced voids. In macro-level study, the emphasis was put on improving the current welding process, for both static resistance welding and continuous resistance welding.

Resistance welding of woven fabric reinforced thermoplastic composites was in-vestigated, with a special attention paid to the surface properties of the adherends.

(8)

Both the type of the majority fibres (warp yarns or weft yarns) and the apparent orientation of the majority fibres on the adherend surface were found to influence the failure mode and lap shear strength. Fibre sizing was found to be crucial for a good fibre-matrix adhesion, and therefore it was crucial for a good weld.

Due to the negative effect of process induced voids on the mechanical perfor-mance of resistance welded joints, the void formations during resistance welding were studied. Other than fibre de-compaction, the residual volatiles inside the ad-herends was found to be a main cause of the voids in the joints for a welding process performed under a moderate welding pressure. Non-uniform void distribution was observed inside the joints, with void concentrations near the middle of weld over-lap. The voids could be reduced by using pre-dried adherends or using a higher welding pressure.

The weldline, formed during welding by the heating element and the surround-ing neat resin, plays an important role in load transfer, therefore, its effect on the weld performance was investigated. A thinner weldline, usually obtained by using a thinner heating element, was found to be preferable to a thicker one. The relatively weaker welding quality near the edges of the joints was found to be a limitation of weld performance, but this could be improved by tailoring heat generation at the weld overlap or creating resin fillets near the edges.

Apart from the studies in a micro-level, a macro-level study was also performed with a focus on the welding process. The possibility of using displacement mea-surement data for process monitoring and processing window definition was inves-tigated, and it showed ability to detect voids generation and resin squeeze flow during welding and to construct processing windows.

The process of continuous resistance welding was analysed, and a model was developed to simulate the heat generation and heat transfer during welding. The model predicted welding temperatures showed a close agreement with the experi-mental results. Compared to static resistance welding, non-uniform heat genera-tions and dissimilar temperature distribugenera-tions were found in continuous resistance welding. The effects of welding parameters, such as welding voltage, welding speed and width of copper connectors, on the welding temperature were analysed.

(9)

Samenvatting

Eén manier om de uitstootefficiëntie van luchttransport te verbeteren is het verlagen van het gewicht van vliegtuigen. Met dit doel worden composietmaterialen in toenemende mate in vliegtuigen gebruikt om metalen te vervangen. Goedkopere en snellere fabricagemethoden voor composietconstructies zijn dus nodig om de snelle groei van de vraag naar vliegtuigen te beantwoorden. In vergelijking met thermoharders trekken thermoplastische composieten steeds meer aandacht van luchtvaartindustrieën, niet alleen vanwege hun uitstekende materiaaleigenschappen, maar ook door hun potentie om productietijd en constructiekosten te verlagen door het gebruik van goedkopere fabricagetechnologieën, zoals lassen.

Weerstandslassen wordt beschouwd als ´e´en van de meest veelbelovende lastech-nieken, wat te danken is aan het lage energieverbruik, de eenvoud van de lasoperatie en de mogelijkheid tot opschaling. Er moet echter diepere kennis van weerstand-slassen worden vergaard om deze technologie rijp te maken voor bredere toepassing in vliegtuigen.

Eerdere onderzoeken naar weerstandslassen van thermoplastische composieten zijn voornamelijk gericht op het begrijpen van het lasmechanisme en het karakteris-eren van de laskwaliteiten. De warmteoverdracht, consolidatie en kristallisatiemech-anismen van statisch weerstandslassen zijn geanalyseerd en verschillende testmeth-oden zijn gebruikt om de mechanische prestatie van weerstandsgelaste verbindingen te karakteriseren. Er zijn echter nog een aantal gaten in de huidige kennis:

• Ten eerste, vergeleken met de belangrijkste lasparameters, zoals vermogensin-voer en opwarmtijd, is er minder bekend over hoe de specifieke eigenschap-pen van het verbonden materiaal of het warmte-element de laskwaliteit kan be¨ınvloeden.

• Ten tweede, de definities van proceslimieten vertrouwen vooral op ’trial-and-error’ proeven, welke duur en tijdrovend zijn.

• Ten slotte, continu weerstandslassen is uitgevonden om de huidige gebruikte lastechniek op te schalen, alhoewel er weinig bekend is over het opwarm-mechanisme van dit lasproces of hoe de betrokken lasparameters gedefinieerd dienen te worden.

Het onderzoek was gericht op het verkrijgen van diepere kennis van weerstand-slassen, op zowel micro- als macro-niveau. Bij het micro-niveau onderzoek werden de specifieke eigenschappen die de laskwaliteit be¨ınvloeden besproken, zoals de oppervlakte-eigenschappen van de verbonden materialen, de laslijn-eigenschappen

(10)

en de door de las ge¨ınduceerde leegtes. Bij het macro-niveau onderzoek werd de nadruk gelegd op het verbeteren van het huidige lasproces, voor zowel statisch weer-standslassen als continu weerweer-standslassen. Weerweer-standslassen van gewoven textiel versterkte thermoplastische composieten is onderzocht, met speciale aandacht voor de oppervlakte-eigenschappen van de verbonden materialen. Zowel het type van de meerderheidsvezels (scheringgaren of inslaggaren) en de schijnbare ori¨entatie van de meerderheidsvezels op de oppervlakte van het te verbinden materiaal bleken de modus van falen en de schuifsterkte te be¨ınvloeden. Oppervlaktebehandeling van de vezels bleek cruciaal te zijn voor een goede vezel-matrix hechting en was daarom cruciaal voor een goede las.

Door het negatieve effect van door het proces ge¨ınduceerde leegtes op de mecha-nische prestaties van weerstandsgelaste verbindingen, werd de vorming van leegtes tijdens weerstandslassen bestudeerd. Naast de-compaction van de vezels bleken de overgebleven vluchtige stoffen in de te verbinden materialen een hoofdoorzaak van de leegtes in de verbindingen voor een lasproces dat werd uitgevoerd onder een gemiddelde lasdruk. Niet-uniforme verdeling van de leegtes werd waargenomen in de verbindingen, met concentraties van leegtes bij het midden van de lasoverlap. De leegtes konden worden verminderd door het gebruik van voorgedroogde materialen of het gebruik van een hogere lasdruk.

De laslijn, die tijdens het lassen wordt gevormd door de het verwarmingsele-ment en de omliggende pure hars, speelt een belangrijke rol in overdracht van de belasting. Daarom werd zijn effect op de lasprestatie onderzocht. Een dunnere laslijn, meestal te verkrijgen door een dunner verwarmingselement te gebruiken, bleek te verkiezen boven een dikkere. De relatief mindere laskwaliteit bij de randen van de verbinden bleek een beperking van de lasprestatie, maar dit kon worden verbeterd door het op maat maken van warmteopwekking bij de lasoverlap of het cre¨eren van een harsstrook bij de randen.

Naast de onderzoeken op een micro-niveau, werd er ook een macro-niveau on-derzoek uitgevoerd met een nadruk op het lasproces. De mogelijkheid van het ge-bruiken van verplaatsings meetdata voor procesmonitoring en proceslimiet definitie werd onderzocht en toonde de mogelijkheid aan om het ontstaan van leegtes en kni-jpvloei van hars tijdens lassen te detecteren en om proceslimieten te construeren.

Het proces van continu weerstandslassen werd geanalyseerd en er werd een model ontwikkeld om de warmteopwekking en warmteoverdracht tijdens lassen te simuleren. De door het model voorspelde lastemperaturen lieten een nauwe overeenkomst zien met de uitkomsten van de experimenten. Vergeleken met statisch weerstandslassen, werden bij continu lassen niet-uniforme warmteopwekking en on-gelijke temperatuur verdelingen gevonden. De effecten van lasparameters, te weten: lasspanning, lassnelheid en de breedte van de koperen verbinden, op de lastemper-atuur werden geanalyseerd.

(11)

List of Tables

2.1 Characteristics of the two types of GF/PEI prepregs used in this study. . . 15 2.2 Mechanical tests performed in this study (LSS: lap shear strength

test; DCB: double cantilever beam test; ENF: end notched flexure test; 3 PB: three point bending test; SBS: short beam shear test). . 18 3.1 Material properties at room temperature for heat transfer model. . 39 3.2 Temperature dependent material properties of GF/PEI and PEI. . . 40 3.3 Material properties at room temperature for stress analysis. . . 41 3.4 The predicted temperature dependent modulus of GF/PEI in the

direction through thickness. . . 42 3.5 Constants used for the analysis of moisture induced void nucleation

and growth. . . 50 3.6 GF/PEI laminates with different drying conditions. . . 53 4.1 Specifications of stainless steel meshes used for welding. . . 63 4.2 Material properties at room temperature for the material used in

heat transfer simulation. . . 66 4.3 Temperature dependent material properties of GF/PPS and PPS. . 67 6.1 Room temperature material properties used in the heat transfer model.112 6.2 Temperature dependent properties of GF/PPS and PPS. . . 112

(16)

(17)

List of Figures

1.1 Thermoplastic composites for aircraft applications. . . 2 1.2 Schematic of welding/fusion bonding of thermoplastic at interface [12]. 3 1.3 Classification of welding techniques for thermoplastics [9]. . . 4 1.4 Principle of resistance welding [14]. . . 4 1.5 Application of resistance welding in aircraft manufacturing, i.e. the

fixed leading edge of Airbus A380-800. . . 5 1.6 Schematic of the flow of this thesis. . . 7 2.1 Schematic drawing of [[(0◦/90◦)]n]sand [[(90◦/0◦)]n]slaminates stacked

with different surface sides: Type I-warp, Type I-Weft, Type II-warp and Type II-Weft. . . 15 2.2 The experimental welding setup. . . 16 2.3 Positions of the thermocouples used for the temperature

measure-ments during the resistance welding of GF/PEI. . . 17 2.4 Schematic drawing of single lap shear GF/PEI samples with different

fibre orientations at the welding surfaces (warp yarn shown as black and weft yarn shown as grey in the figure). . . 19 2.5 Schematic diagrams of the DCB and ENF samples for plain GF/PEI

specimen and mesh embedded GF/PEI specimen (a), and the four different fibre orientations at the crack propagation interfaces for plain GF/PEI specimens (b) (warp yarn shown as black and weft yarn shown as grey in the figure). . . 20 2.6 The welding temperatures (maximum temperatures measured

dur-ing the welddur-ing) and lap shear strength values for the SS0303 GF/PEI joints (Type I-warp) welded under a constant power input of 80 kW/m2 _{and with different heating times (using 80% of the}

maxi-mum weld strength as the benchmark). . . 21 2.7 Fracture surfaces and SEM micrographs of the fracture surfaces of

the SS0303 GF/PEI joints (Type I-warp) welded under a constant power input of 80 kW/m2 _{and with a heating time of 30 s (under}

welded joints). . . 22 2.8 Fracture surfaces and SEM micrographs of the fracture surfaces of

the SS0303 GF/PEI joints (Type I-warp) welded under a constant power input of 80 kW/m2 and with a heating time within the pro-cessing interval: (a) a joint welded for 55 s, (b) a joint welded for 40 s, (c) a joint welded for 90 s. . . 23

(18)

2.9 Crack resistance curves (R curves) of plain SS0303 GF/PEI speci-mens and mesh embedded SS0303 GF/PEI specispeci-mens. . . 23 2.10 SEM micrographs for the fracture surfaces of Mode I specimens: (a)

SS0303 GF/PEI laminate, (b) mesh embedded SS0303 GF/PEI. . . 24 2.11 Fracture surfaces and SEM micrographs of the fracture surfaces of

the SS0303 GF/PEI joints (Type I-warp) welded under a constant power input of 80 kW/m2 _{and with a heating time of 100 s (over}

heated joints). . . 24 2.12 Cross-section micrograph of a SS0303 GF/PEI joint (Type I-warp)

welded at 80 kW/m2 _{and 100 s that shows how the crack is deviated}

towards the mesh in the centre of the overlap. . . 25 2.13 Fracture surfaces of SS0303 GF/PEI joints (Type I-warp) made of

adherends with residual moisture and welded at 80 kW/m2 _{for 55 s.}

The SEM micrograph (right) evidences the presence of porosity in the crack path. . . 25 2.14 Cross-section micrographs of SS0303 GF/PEI joints (Type I-warp)

welded at 80 kW/m2 _{for 55 s by using laminates of different}

mois-ture conditions: (a) laminate with residual moismois-ture, (b) fully dried laminate. . . 26 2.15 Different fibre sizings for the glass fabrics of GF/PEI prepreg [23]:

(a) Chromium methacrylate and (b) aminosilane. . . 27 2.16 The flexural strength, short beam shear strength (a) and the failure

modes (b) of the GF/PEI laminates with different fibre sizings. . . . 27 2.17 R-curves of DCB tests for GF/PEI specimens with different fibre

sizings . . . 28 2.18 (a) Lap shear strength of SS0303 and TC7781 GF/PEI joints (Type

I-warp) welded with 80 kW/m2 _{and 55 s, and (b) SEM micrographs}

for the fracture surfaces of both types of joints. . . 28 2.19 Lap shear strengths of SS0303 GF/PEI joints with different fibre

orientations welded with 80 kW/m2 _{and 55 s.} _{. . . .} ₂₉

2.20 Fracture surfaces and SEM micrographs for the SS0303 GF/PEI joints with different fibre orientations on the joining surfaces and welded with 80 kW/m2 _{and 55 s: (a) Type I-warp, (b) Type}

II-warp, (c) Type I-weft, and (d) Type II-weft. . . 30 2.21 Comparison of interlaminar fracture toughness for the SS0303 GF/PEI

laminates with different fibre orientations at the crack propagating interfaces. . . 30 3.1 Thermal expansion measurements using Thermomechanical Analysis

for ambient conditioned PEI specimens (a) and fully dried GF/PEI laminates specimens (b). . . 37 3.2 Geometry and boundary conditions for heat transfer analysis of

re-sistance welding. . . 39 3.3 Geometry and boundary conditions for internal stress analysis. . . . 40 3.4 Cross-section micrographs of GF/PEI laminate before fibre de-compaction

(19)

LIST OF FIGURES ix

3.5 Mechanical equilibrium of the compressed fibre in a composite. . . . 43 3.6 Schematic diagram of void growth in a shell of viscous fluid. . . 45 3.7 (a) The temperature distributions, and (b) stress distributions for

resistance welding of GF/PEI, at a heating time of 55 s. . . 47 3.8 Compressibility of eight layers assembly of 7781 glass fabric. . . 48 3.9 Distributions of critical stress for the GF/PEI joints welded with

different pressures of 0.1 MPa, 0.2 MPa, 0.4 MPa and 0.8 MPa (the negative sign in the plots indicates a compressive stress). . . 48 3.10 Cross-section micrographs of GF/PEI joints welded with different

pressure of 0.1 MPa, 0.2 MPa, 0.4 MPa and 0.8 MPa. . . 49 3.11 The average void contents (a) and the void distributions along the

weld overlaps (b) for GF/PEI joints welded with different pressures of 0.1 MPa, 0.2 MPa, 0.4 MPa and 0.8 MPa. . . 49 3.12 Effects of temperature (a) and welding pressure (b) on void

nucle-ation and growth. . . 51 3.13 Distributions of void nucleation rate (a) and void growth (b) inside

the joints for the maximum welding temperature, t = 55s. . . 52 3.14 Cross-section micrographs of GF/PEI joints using laminates in

dif-ferent drying conditions: Type-A, Type-B, Type-C and Type-D. . . 52 3.15 The average void contents (a) and the void distributions along the

weld overlaps (b) for GF/PEI joints welded using laminates in dif-ferent drying conditions: Type-A, Type-B, Type-C and Type-D. . . 53 3.16 Cross-section micrographs of resistance welded joints using

oven-dried Type-A GF/PEI laminates (a) and Type-B GF/PEI laminates (b). . . 53 3.17 Cross-section micrographs of Type-A GF/PEI joints welded with

different pressures of 0.8 MPa, 1 MPa, 1.2 MPa and 1.5 MPa. . . . 54 3.18 The average void contents (a) and the void distributions along the

weld overlaps (b) for Type-A GF/PEI joints welded with different pressures of 0.8 MPa, 1 MPa, 1.2 MPa and 1.5 MPa. . . 54 4.1 Schematic of two-dimensional heat transfer simulation for resistance

welding of GF/PPS with a metal mesh heating element. . . 65 4.2 DSC heat flow curve of PPS (heated up to 350◦C at a heating rate

of 10◦C/min, dwell time 10 min at 350◦C, and then cooled down to room temperature naturally). . . 66 4.3 Comparison of lap shear strength (a) and distributions of welding

temperatures across the weld overlap (b) for the GF/PPS joints welded with different meshes of M200, M24, M12 and M8 as the heating element, under a constant power input of 80 kW/m2 _and

(20)

4.4 (a) The fracture surfaces of GF/PPS joints welded with different meshes of M200, M24, M12 and M8, under a constant power input of 80 kW/m2 _{and a heating time of 60 s; (b) illustrations of crack}

propagating paths and the SEM pictures for the joints welded with a M200 mesh and a M24 mesh as the heating element (joints welded with a M12 or M8 mesh showed a similar crack propagating path as that of joints welded with a M24 mesh). . . 69 4.5 Load-displacement curve of the GF/PPS joints welded with a M200

mesh as the heating element; the crack propagation of the joints under a tensile load close to the ultimate load is shown on the right side of the picture. . . 70 4.6 (a) The relationship between the lap shear strength and the thickness

of weld line for the GF/PPS joints welded using different meshes of M200, M24, M12 and M8 as the heating element; (b) the evolutions of joint rotation angle during single lap shear tests for the GF/PPS joints welded using different meshes of M200, M24, M12 and M8 as the heating element. . . 70 4.7 (a) Comparison of lap shear strength for the GF/PPS joints welded

with the same type of mesh, M200, but with different layers of neat PPS films, i.e. 0, 1, 4 and 6 layers; (b) comparison of lap shear strength for the GF/PPS joints welded with different meshes that have the same wire diameter but different opening gaps, i.e. M24, M26 and M45 meshes. . . 72 4.8 The distributions of model simulated welding temperatures for the

GF/PPS joints welded with a M200 mesh as the heating element with different heating times of 60 s, 80 s and 120 s under a constant power input of 80 kW/m2. . . 73 4.9 (a) The lap shear strength of GF/PPS joints welded with a M200

mesh heating element under a constant power input of 80 kW/m2

and for different heating times of 60 s, 80 s and 120 s; (b) the fracture surfaces of the joints. . . 73 4.10 The cross section micrographs of GF/PPS joints welded with a M200

mesh heating element under a constant power input of 80 kW/m2

and for different heating times of 60 s, 80 s and 120 s. . . 74 4.11 (a) Schematic drawing of resistance welded joints with a folded M200

mesh and with a wider M200 mesh as the heating elements, and (b) a comparison of temperature distributions at the weld interface of the joints welded using different meshes as the heating element: an un-folded M200 mesh, a folded M200 mesh and a wider M200 mesh. The joints were welded under a constant power input of 80 kW/m2

and a heating time of 60 s. . . 75 4.12 (a) The single lap shear strength and (b) fracture surfaces of the

GF/PPS joints welded using different M200 meshes as the heating element: an un-folded mesh, a folded mesh and a wider mesh. . . . 76

(21)

LIST OF FIGURES xi

4.13 The cross-section micrographs of the GF/PPS joints welded using different M200 meshes as the heating element: an un-folded mesh, a folded mesh and a wider mesh. . . 77 4.14 The strain distributions along the overlaps of the GF/PPS joints

welded using different M200 meshes as the heating element: an un-folded mesh, a un-folded mesh and a wider mesh, under a tensile load of 2000 N. . . 77 4.15 (a) Schematic drawing of resistance welded GF/PPS joints with resin

fillets added to the edges of the joints, either with ragged edges or with flat edges; (b) comparison of lap shear strength values for the GF/PPS joints with different fillet situations: without fillet, with fillets bonded to flat adherend edges, and with fillets bonded to ragged adherend edges. . . 78 4.16 The fracture surfaces of resistance welded GF/PPS joints: (a) with

fillets bonded to flat edges and (b) with fillets bonded to ragged edges. 79 4.17 Comparison of strain distributions for the resistance welded GF/PPS

joints with different fillet situations: without fillet, with fillets bonded to flat adherend edges, and with fillets bonded to ragged adherend edges, under a tensile load of 2000 N. . . 79 5.1 Illustration of intimate contact process in two adjacent surfaces. . . 90 5.2 Schematic drawing of resistance welding and displacement

measure-ment. . . 93 5.3 The experimental setup of displacement measurement during

resis-tance welding: 1) weld stack, 2) thermal insulation block, 3) a ref-erence plane on the top of thermal insulation block for displacement measurement, and 4) laser displacement sensor. . . 93 5.4 The temperature and displacement curves for resistance welding of

GF/PEI laminates, with 0.1 wt% residual moisture, under a power input of 80 kW/m2 _{and with a heating time of 100 s; the}

displace-ment curve can be divided into five stages: i) intimate contact, ii) thermal expansion, iii) squeeze flow, iv) cooling contraction, and v) stabilization. . . 95 5.5 Temperature evolution at different locations away from the welding

interface as predicted by a heat transfer model of the resistance welding process [15]. . . 95 5.6 (a) The temperature and displacement curves for resistance welding

of GF/PEI laminates, with 0.1 wt% residual moisture, under a power input of 80 kW/m2 _{for different heating times of 30 s, 55 s, and 100 s,}

and (b) the evolution of displacement in relationship with temperature. 96 5.7 Cross-section micrographs of GF/PEI welds (0.1 wt% residual

mois-ture, 80 kW/m2 _{power input, 0.8 MPa welding pressure) under}

(22)

5.8 The displacement curves and pictures for resistance welding of GF/PEI laminates, with 0.1 wt% residual moisture, at a constant power in-put of 80 kW/m2 _{and under different welding pressures of 0.2 MPa,}

0.8 MPa and 1.5 MPa. . . 98 5.9 The cross-section micrographs and lap shear strength (b) of

resis-tance welded GF/PEI joints, with 0.1 wt% residual moisture, at a constant power input of 80 kW/m2 for 55 s, and under different pressures of 0.2 MPa, 0.8 MPa and 1.5 MPa. . . 98 5.10 Fracture surfaces of resistance welded GF/PEI joints with different

welding pressures: (a) 0.8 MPa, and (b) 1.5 MPa. . . 99 5.11 The displacement curves of resistance welding of GF/PEI laminates

with 0 wt%, 0.1 wt%, and 0.3 wt% residual moisture at a power input of 80 kW/m2 _{for 55 s. . . 100}

5.12 (a) The cross-section micrographs and lap shear strength (b) of joints welded using GF/PEI laminates with 0 wt%, 0.1 wt%, and 0.3 wt% residual moisture contents at a power input of 80 kW/m2 _{for 55 s. . 100}

5.13 (a) Optical microscopy (OM) and scanning electric microscopy (SEM) images for the fracture surfaces of joints welded with laminates with 0 wt% residual moisture (joints welded with laminates with 0.1 wt% residual moisture showed similar fracture surfaces); (b) OM and SEM images for the fracture surfaces of joints welded with laminates with 0.3 wt% residual moisture. . . 101 5.14 (a) The processing window defined using welding displacement curve

and (b) the processing window defined from mechanical tests for re-sistance welding of GF/PEI laminates with 0.1 wt% residual mois-ture at a constant power input of 80 kW/m2_{. . . 101}

5.15 Correlation between lap shear strength and welding displacement curve for resistance welding of GF/PEI laminates with 0.1 wt% resid-ual moisture at a power input of 50 kW/m2 _{(a), and at a power input}

of 120 kW/m2 (b). . . 103 5.16 Schematic drawing of the bonding area between heating element and

adherends. . . 103 5.17 Correlation between lap shear strength and welding displacement

curve for resistance welding of fully dried GF/PEI laminates at a power input of 80 kW/m2. . . 104 6.1 A schematic diagram of continuous resistance welding of single-lap

joints. . . 110 6.2 Continuous resistance welding setup, consisting of 1) pneumatic

sys-tem for welding compaction, 2) block connectors, 3) laminates, 4) motion system, 5) pneumatic system for clamping, 6) insulators, and 7) power supply. . . 113 6.3 Geometry of continuous welding setup for the welding of single lap

shear joint. . . 114 6.4 A flowchart of the process modelling for CRW. . . 114 6.5 The relationship of mesh resistance with clamping pressure. . . 116

(23)

LIST OF FIGURES xiii

6.6 (a) Comparison of the total resistance with the mesh resistance and (b) relationship between the fluctuations of total resistance and the movement of the copper wheels. . . 116 6.7 The relationship between the mesh resistance and the length of the

mesh, for calculating the equivalent resistivity of the metal mesh. . 117 6.8 (a) Predicted voltage distribution in the mesh heating element under

a constant current input of 1.92 A, and (b) a comparison of the model prediction with the measurement results, at different X positions of d = 0 mm, 63.5 mm and 127 mm. . . 118 6.9 (a) Distribution of resistive heating for the stainless steel mesh

be-tween the two block connectors using the average value of resistive heating for each sub-area, as simulated under a constant welding voltage of 1V, and (b) distribution of resistive heating for the stain-less steel mesh between the two block connectors, as simulated di-rectly from the electrical model under a constant welding voltage of 1V. . . 119 6.10 Average resistive heating rate in sub-area 5 for different positions of

the wheel connectors (on block connectors 1, 5, 10 and 20) under a 1 V constant voltage. . . 120 6.11 Schematic diagram of contact interface between the stainless steel

mesh and the block connectors. . . 121 6.12 The distribution of welding temperature at the weld overlap for the

welding performed at a constant welding voltage of 4.3 V and using a welding speed of 1.27 mm/s. . . 122 6.13 Positions of the thermocouples used for temperature measurement. 123 6.14 Comparisons of experimentally measured welding temperatures with

predictions of process model for GF/PPS joints welded at a constant welding voltage of 4.3 V and welding speeds of (a) 1.27 mm/s and (b) 0.85 mm/s. . . 124 6.15 The influence of welding voltage on welding temperature (a), at a

constant welding speed of 0.85 mm/s; and influence of welding speed on welding temperature (b), at a constant welding voltage of 4.3 V. 125 6.16 Selections of weld ing speed and welding voltage for different

pre-defined welding temperatures of 250 ◦C , 320 ◦C , 400 ◦C and 550

◦_{C for TC5. . . 125}

6.17 (a) The influence of contact resistance at the block connector-stainless steel mesh interfaces on the temperature distribution across the welding area 10 (x = 120.65 mm), and (b) the influence of heat transfer coefficient at the block connector-stainless steel mesh inter-faces on the temperature distribution across the welding area 10 (x = 120.65 mm), without contact resistance. . . 126

(24)

6.18 (a) The effect of the width of block connectors on the resistance of stainless steel mesh heating element and welding temperature using a welding voltage of 4.3 V and a welding speed of 0.85 mm/s; (b) selection of welding speed and required power inputs for the welding process using different widths of block connector to obtain a welding temperature of 320◦C at TC5, at a constant welding voltage of 4.3 V.127

(25)

Chapter 1 Introduction

Air transport, the fastest mode of transport except space travel, makes long-distance travel more efficient and convenient, however, it contributes about 2% to the total of man-made carbon dioxide (CO2) emissions [1]. As air transport

demand is expected to double in the next 15 years [2], eco-efficient aircraft must be designed, manufactured and operated to meet the goal of cutting CO2 emissions in

half by 2050 when compared to 2005 levels [1].

A reduction in fuel burn can be achieved in the new generation of aircraft us-ing several different strategies, such as improvus-ing aerodynamic efficiency, reducus-ing aircraft weight and reducing engine fuel consumption. In these cases the aircraft efficiency can be improved by:

• improving aerodynamic efficiency by developing new configuration concepts, such as blended wing body, joined wing and lifting body, and using laminar flow technology [3]

• reducing aircraft weight by using lighter materials, lower weight increment joining methods and structure optimization

• reducing engine (fossil) fuel consumption by developing more efficient engines and using sustainable alternative fuels [1]

1.1 Thermoplastic composites for aircraft

struc-tures

Composite materials, in particular carbon fibre reinforced polymer (CFRP), have been increasingly used for applications in aircraft structures due to their great potential for weight reduction, typically 20% lighter than aluminium structures [4]. Compared to the widely used thermoset composites (TSCs), thermoplastic composites (TPCs) have the same weight reduction capabilities but showing more advantages due to their inherent properties. Without cross-linking in the polymer-ized thermoplastic, thermoplastic polymer can be re-melted and re-shaped easily, making it possible to manufacture and join thermoplastic composite structures us-ing low-cost manufacturus-ing technologies, such as press formus-ing and weldus-ing, or

(26)

fusion bonding. By eliminating the use of an autoclave, the manufacturing cost of thermoplastic composites can be greatly reduced. Moreover, thermoplastic compos-ites offer more advantages, such as lower storage requirements, good environment resistance, excellent fire, smoke and toxicity (FST) performances and recyclability. An increasing number of structures made of thermoplastic composites are being used in the current generation of aircraft, for example: the press-formed ribs of the Dornier 328, the resistance welded fixed wing leading edges of the Airbus A340 and A380, the pressure bulkhead floors of the Gulfstream G550 and G650, the cockpit floor of the A400M, the induction-welded rudder and elevators of the new Gulf-stream G650 [5], and the press-formed brackets, clips and cleats of the Boeing 787 and Airbus A350 [6]. Some examples of aircraft structures made of thermoplastic composite are shown in Figure 1.1. In order to extend the application of thermo-plastic composites from aircraft secondary structures to primary structures, the relevant key manufacturing technologies, such as press-forming, co-consolidation, welding and automated fibre placement, are currently being researched, and many projects have been set up, such as the Thermoplastic Affordable Primary Aircraft Structures (TAPAS) project [7], to reach this goal.

Figure 1.1: Thermoplastic composites for aircraft applications.

1.2 Joining methods for TPCs

Due to the geometric complexities of the aircraft structures and the limitations placed on the size of parts that can presently be manufactured, joining is a necessary process when manufacturing thermoplastic composite structures and the joining

(27)

1.3. Resistance welding of TPCs 3

methods influence the mechanical performance, manufacturing costs and weights of such structures. Mechanical fastening, i.e. riveting and bolting, and adhesive bonding are the two most commonly used joining methods for aircraft structures. Mechanical fastening increases the total weight of composite structures and also the final costs of a structure [8, 9], moreover, stress concentrations are always found in the vicinity of the holes in the material. Adhesive bonding, although demonstrating the advantages of lower stress concentrations and lighter structures compared to those with mechanical fastenings, is difficult to achieve when high performance thermoplastic composites, such as polyetheretherketone (PEEK) and polyphenylene sulphide (PPS), are used, due to their relatively low surface energy [10].

Thanks to the inherent properties of thermoplastic polymers, in particular remeltability and self-healing, thermoplastic composites can be welded using lo-cally heating, molecule diffusion and consolidation [11–13], as schematilo-cally shown in Figure 1.2. Welding of thermoplastic composites shows the potential to reduce overall manufacturing costs of aircraft structures and is expected to be a better solution for the joining of thermoplastic composites [9]. As shown in Figure 1.3, many different heating mechanisms can be used as a heat source for welding, among which resistance welding, ultrasonic welding and induction welding are three of the most promising welding techniques [8].

Figure 1.2: Schematic of welding/fusion bonding of thermoplastic at interface [12].

1.3 Resistance welding of TPCs

As one of the most promising and efficiency welding technologies, resistance weld-ing shows many advantages, such as low-cost equipment required, lower energy consumption, and ease of welding operation. The principle of resistance welding of thermoplastic composites is schematically shown in Figure 1.4. In resistance welding, an electrically resistive heating element, normally a carbon prepreg or a metal mesh, is sandwiched between two adherends to be welded. When an elec-trical current is applied, the heating element will heat up due to Joule heating, and as a result the temperature at the interface of the two adherends will increase, allowing molecule diffusion at the interface. Under a moderate welding pressure, a

(28)

Figure 1.3: Classification of welding techniques for thermoplastics [9].

solid weldline can be obtained at the overlap area after the following cooling stage. Resistance welding has been successfully used for joining secondary structures of aircraft, such as the outboard fixed leading edges of the A340-500/600 and A380-800 which are made of continuous glass fibre reinforced polyphenylene sulphide (GF/PPS), see Figure 1.5.

Figure 1.4: Principle of resistance welding [14].

Current research on resistance welding of thermoplastic composites is mainly focused on understanding welding mechanisms and characterizing weld quality. The heating mechanism of welding has been analysed using heat transfer models and the temperature evolution during the welding process can be predicted [15– 21]. Based on heat transfer simulation, analysis of other welding mechanisms, such as consolidation [13, 20, 22, 23] and crystallization [13, 15–17, 20, 24–26], has also been performed. Different characterization methods have been used to evaluate the quality of resistance welded joints to understand the effect of welding parameters on the weld quality. Optical microscopy, C-scan, B-scan and thermal imaging have been used for qualitative analysis [14, 27–31]. Due a lack of standard mechanical testing methods for the welded joints, various mechanical testing methods have been used, such as single lap shear test [17, 29–36], short beam shear test [27], double cantilever beam test [32, 33], end-notched flexural test [37], three- or

(29)

four-1.4. Research objective 5

Figure 1.5: Application of resistance welding in aircraft manufacturing, i.e. the fixed leading edge of Airbus A380-800.

point bending test [38].

The current resistance welding technique, normally regarded as static resistance welding, is suitable to weld small-to-medium size structures, however it is not straightforward to scale it up for large size structures [2, 14], i.e. the stringer-skin assembly of an aircraft, because there will be substantial requirements for power input and clamping pressure. In order to solve this problem, the concept of sequential resistance welding is proposed [18, 39, 40], however, there is too much hand work required during the preparation for the welding process, i.e. positioning of the multiple heating elements. More recently, a continuous resistance welding technique has been invited by the Canada National Research Council (NRC) [41]; using only one heating element during welding, it shows great potential to reduce the intensive labour involved in welding preparation and to achieve automation of the welding process.

1.4 Research objective

To gain a deeper knowledge on resistance welding of thermoplastic composites and to extend its application in aircraft structures, we need to answer the questions: What and how the specific properties of adherends and heating element influence the resistance welding of thermoplastic composites at the micro-level/material-level, and can the welding process be improved at the macro-level/process-level? Even though a lot of effort has gone into understanding the welding mechanisms of thermoplastic composites and characterizing the weld quality, there are still some gaps in our knowledge:

• Firstly, most of the previous research into the quality of resistance welding has focused on the effects of the welding parameters, i.e. power input and heating time, while less emphasis has been given to the micro/material level, i.e. the effects of adherend properties, mesh properties and the local geometry of the joints.

(30)

• Secondly,the definitions of processing windows usually rely on trial-and-error tests, which are costly and time consuming.

• Finally, continuous resistance welding was invented to scale-up the current welding technique, however, little is known about the heating mechanism of this welding process or how to define the welding parameters involved. The objective of research work presented here was to:

• gain an in-depth understanding of what and how the specific properties of ad-herends and heating element influence the resistance welding of thermoplastic composites

• save the time and cost spent in determination of welding parameters, for both static welding process and continuous welding process

1.5 Thesis outline

The thesis is divided into 7 Chapters and 2 main parts: a micro-level study and a macro-level study, as shown in Figure 1.6. In part I, Chapters 2, 3 and 4, the specific properties that influence the welding quality will be discussed. In part II, Chapters 5 and 6, the emphasis will be on improving the current welding processes. Conclusions are given in Chapter 7.

• In Chapter 2, an experimental study on the effect of fibre-matrix adhesion and fibre orientation on the strength and failure modes of resistance welded thermoplastic composites joints will be presented

• In Chapter 3, the process induced voids in resistance welding will be investi-gated, and suggestions for reducing the process induced voids will be given • In Chapter 4, strategies to improve the strength of single lap resistance welded

joints will be discussed, and a focus is set on the effect of weld line properties • In Chapter 5, a new method for process monitoring and processing window

definition based on displacement detection will be discussed

• In Chapter 6, modelling of heat generation and heat transfer for continuous resistance welding process will be given

• Conclusions will be drawn in Chapter 7, and recommendations will be given for future work.

(31)

BIBLIOGRAPHY 7

Figure 1.6: Schematic of the flow of this thesis.

Bibliography

[1] Airbus. Environmental innovations from Airbus. Technical report, 2013. [2] Airbus’ Global Market Forecast for 2013-2032. Technical report, 2013.

[3] NASA: Energy, Efficiency & Emissions. Technical report, http://www.aeronautics.nasa.gov/, 2009.

[4] Autar K. Kaw. Mechanics of composite materials (second edition). CRC Press, Taylor & Francis Group, Boca Raton, FL. USA, 2006.

[5] Arnt Offringa. New thermoplastic composite design concepts and their auto-mated manufacture. JEC Composites magazine, 58(June-July):45–49, 2010. [6] A. Offringa, J.W. van Ingen, and A. Buitenhuis. Butt-joined, Thermoplastic

Stiffened-skin Concept Development. SAMPE Journal, 48(2):7–15, 2012. [7] http://www.tapasproject.nl/.

[8] D. Stavrov and H. E. N. Bersee. Resistance welding of thermoplastic composites-an overview. Composites Part A: Applied Science and Manufac-turing, 36(1):39–54, 2005.

[9] Ali Yousefpour, Mehdi Hojjati, and Jean-Pierre Immarigeon. Fusion Bond-ing/Welding of Thermoplastic Composites. Journal of Thermoplastic Com-posite Materials, 17(4):303–341, 2004.

[10] S J Shaw, J Comyn, and L Mascia. Surface treatment and bonding of ther-moplastic composites. In Agard Conference Proceedings 590 - Bolted/Bonded

(32)

Joints in Polymeric Composites. 83rd Meeting of the AGARD Structures and Materials Panel, Florence, Italy, 2-3 September, 1996 1996.

[11] Woo Il Lee and George S. Springer. A model of the manufacturing pro-cess of thermoplastic matrix composites. Journal of Composite Materials, 21(11):1017–1055, 1987.

[12] Alfred C. Loos and Min-Chung Li. Non-isothermal autohesion model for amor-phous thermoplastic composites. Journal of Thermoplastic Composite Mate-rials, 7(4):280–310, 1994.

[13] Christophe Ageorges, Lin Ye, Yiu-Wing Mai, and Meng Hou. Characteristics of resistance welding of lap shear coupons.: Part II. Consolidation. Composites Part A: Applied Science and Manufacturing, 29(8):911–919, 1998.

[14] M. Hou, L. Ye, and Y. W. Mai. An Experimental Study of Resistance Welding of Carbon Fibre Fabric Reinforced Polyetherimide (CF Fabric/PEI) Compos-ite Material. Applied ComposCompos-ite Materials, 6(1):35–49, 1999.

[15] Tom B. Jakobsen, Roderic C. Don, and John W. Gillespie. Two-dimensional thermal analysis of resistance welded thermoplastic composites. Polymer En-gineering & Science, 29(23):1722–1729, 1989.

[16] A M. Maffezzoli, J M. Kenny, and L. Nicolais. Welding of PEEK/carbon fiber composite laminates. SAMPE Journal, 25(1):35–40, 1989.

[17] X.R. Xiao, S.V. Hoa, and K.N. Street. Processing and Modelling of Resistance Welding of APC-2 Composite. Journal of Composite Materials, 26(7):1031– 1049, 1992.

[18] Scott T. Holmes and John W. Gillespie. Thermal analysis for resistance weld-ing of large-scale thermoplastic composite joints. Journal of Reinforced Plastics and Composites, 12(6):723–736, 1993.

[19] Christophe Ageorges, Lin Ye, Yiu-Wing Mai, and Meng Hou. Characteristics of resistance welding of lap shear coupons. Part I. Heat transfer. Composites Part A: Applied Science and Manufacturing, 29(8):899–909, 1998.

[20] Ziya Seyhan Colak, Fazil Onder Sonmez, and Vahan Kalenderoglu. Process modeling and optimization of resistance welding for thermoplastic composites. Journal of Composite Materials, 36(6):721–744, 2002.

[21] ´Edith Talbot, Pascal Hubert, Martine Dub´e, and Ali Yousefpour. Optimiza-tion of thermoplastic composites resistance welding parameters based on tran-sient heat transfer finite element modeling. Journal of Thermoplastic Com-posite Materials, 26(5):699–717, 2013.

[22] L. J. Bastien and J. W. Gillespie. A non-isothermal healing model for strength and toughness of fusion bonded joints of amorphous thermoplastics. Polymer Engineering & Science, 31(24):1720–1730, 1991.

(33)

BIBLIOGRAPHY 9

[23] Christine A. Butler, Roy L. Mccullough, Ranga Pitchumani, and John W. Gillespie. An analysis of mechanisms governing fusion bonding of thermoplas-tic composites. Journal of Thermoplasthermoplas-tic Composite Materials, 11(4):338–363, 1998.

[24] R. Carbone and A. Langella. Numerical Modeling of Resistance Welding Pro-cess in Joining of Thermoplastic Composite Materials Using COMSOL Mul-tiphysics. In COMSOL Conference, 2009.

[25] Susan C. Mantell and George S. Springer. Manufacturing process models for thermoplastic composites. Journal of Composite Materials, 26(16):2348–2377, 1992.

[26] Christophe Ageorges, Lin Ye, Yiu-Wing Mai, and Meng Hou. Characteristics of resistance welding of lap-shear coupons. Part III. Crystallinity. Composites Part A: Applied Science and Manufacturing, 29(8):921–932, 1998.

[27] M. Dub´e, P. Hubert, A. Yousefpour, and J. Denault. Resistance welding of thermoplastic composites skin/stringer joints. Composites Part A: Applied Science and Manufacturing, 38(12):2541–2552, 2007.

[28] Christophe Ageorges, Lin Ye, and Meng Hou. Experimental investigation of the resistance welding for thermoplastic-matrix composites. Part I: heating element and heat transfer. Composites Science and Technology, 60(7):1027– 1039, 2000.

[29] I. Howie, J.W. Gillespie, and A.J. Smiley. Resistance Welding of Graphite-Polyarylsulfone/Polysulfone Dual-Polymer Composites. Journal of Thermo-plastic Composite Materials, 6(3):205–225, 1993.

[30] Roderic C. Don, John W. Gillespie, and Cynthia L. T. Lambing. Experi-mental characterization of processing-performance relationships of resistance welded graphite/polyetheretherketone composite joints. Polymer Engineering & Science, 32(9):620–631, 1992.

[31] Eric C. Eveno and John W. Gillespie. Resistance welding of graphite polyetheretherketone composites: An experimental investigation. Journal of Thermoplastic Composite Materials, 1(4):322–338, 1988.

[32] Meng Hou, Mingbo Yang, Andrew Beehag, Yiu-Wing Mai, and Lin Ye. Re-sistance welding of carbon fibre reinforced thermoplastic composite using al-ternative heating element. Composite Structures, 47(1-4):667–672, 1999. [33] Christophe Ageorges, Lin Ye, and Meng Hou. Experimental investigation

of the resistance welding of thermoplastic-matrix composites. Part II: opti-mum processing window and mechanical performance. Composites Science and Technology, 60(8):1191–1202, 2000.

(34)

[34] D Stavrov, H.E.N Bersee, and A Beukers. The influence of the heating ele-ment on resistance welding of thermoplastic composite materials. In ICCM-14 Conference, 2003.

[35] A Yousefpour, M Simard, M.A Octeau, M Lamare´ee, and M. Hojjati. Effects of mesh size on resistance welding of thermoplastic composites using metal mesh heating elements. In SAMPE-Europe, 2004.

[36] Martine Dub´e, Pascal Hubert, Jan NAH Gallet, Darko Stavrov, Harald EN Bersee, and Ali Yousefpour. Metal mesh heating element size effect in resis-tance welding of thermoplastic composites. Journal of Composite Materials, 46(8):911–919, 2012.

[37] Willem Hamer. An experimental study on the influence of the metal mesh heating element on the mechanical performance in resistance welded joints. PhD thesis, 2011.

[38] M. Dub´e, P. Hubert, A. Yousefpour, and J. Denault. Fatigue failure charac-terisation of resistance-welded thermoplastic composites skin/stringer joints. International Journal of Fatigue, 31(4):719–725, 2009.

[39] C. L. T. Lambing, R. C. Don, S. M. Andersen, S. T. Holmes, B. S. Leach, and J. W. Gillespie. Design and manufacture of an automated resistance welder for thermoplastic composites. In 49th Annual Technical Conference (ANTEC ’90), pages 2527–2531, 1991.

[40] Steven H. McKnight, Scott T. Holmes, John W. Gillespie, Cynthia L. T. Lamb-ing, and James M. Marinelli. Scaling issues in resistance-welded thermoplastic composite joints. Advances in Polymer Technology, 16(4):279–295, 1997. [41] Ali Yousefpour and Marc-Andre Octeau. Resistance welding of thermoplastics,

(35)

Part I

A micro-level study

(36)

(37)

Chapter 2 Effect of fibre-matrix adhesion

and fibre orientation

As the surfaces of adherends are directly welded to the heating element during the resistance welding process and they together form an important interface for load transfer in the welded joints, the effect of specific properties of adherends, especially the surface properties such as the type of surface fibres, the fibre orientation and the quality of fibre-matrix adhesion, on the mechanical performances of resistance welded joints is investigated in this chapter.

2.1 Introduction

In the latest decades, the mechanical properties of high grade thermoplastics have reached a level as good as, if not better than, thermosetting plastics, resulting in thermoplastic composites being increasingly used in the aircraft industry [1]. Due to the intrinsic properties of thermoplastic resins, thermoplastic composites can be welded in a relatively short time. And among the available welding techniques, resistance welding has been regarded as one of the most promising welding tech-niques because of its advantages such as low manufacturing cost and simplicity of processing. Resistance welded joints are characterized by a heterogeneous weld line composed of a heating element and some extra neat resin.

So far, research on resistance welding of thermoplastic composites has been mainly focused on understanding the influence of the welding parameters, com-parative analysis of different types of heating elements and comcom-parative analysis of different adherends. The temperature and degree of consolidation have been acknowledged as the main factors that influence the quality of the welds, and they have been found dependent on the welding parameters such as power input, welding pressure and heating time [2–4]. Processing models have been developed to predict the welding temperature and consolidation degree during welding [3, 5–8]. The heating element is not only responsible for heat generation during welding but it also becomes a part of the weld line of the final joints, and therefore it has an effect on the quality of the welded joints. Stavrov [9] compared a metal mesh heating el-ement with a carbon fibre heating elel-ement and concluded that the former provided

(38)

better mechanical performance and weld repeatability. Further research on the ef-fect of the characteristics of stainless steel mesh heating elements on the strength of the joints was carried out by Dub´e [10]. The diameter of the metal wires and the open area were shown as the most important parameters influencing the weld quality. Likewise, the effect of different adherend materials, namely carbon fibre reinforced polyetheretherketone (CF/PEEK), carbon fibre reinforced polyetherke-toneketone (CF/PEKK), glass fibre reinforced polyetherimide (GF/PEI), carbon fibre reinforced polyetherimide (CF/PEI), and glass fibre reinforced polyphenyle-nesulfide (GF/PPS), on the strength of the resistance joints has been investigated [10–13]. Among them unidirectional CF/PEKK welded with a metal mesh heating element resulted in the highest lap shear strength of 52 MPa [10]. However, no study has been reported so far on the influence of basic characteristics of the ad-herends such as the fibre orientation on the welding surfaces and the fibre-matrix adhesion on the mechanical properties of the resistance welded joints, even though intralaminar failure has been acknowledged by different researchers as one of the typical failure modes in resistance welded joints [10–12]. Since fibre-matrix ad-hesion and fibre orientation are known to influence the intralaminar properties of the composites [14–17], those parameters are expected to have an effect on the performance of resistance welded joints.

The aim of the present work is to gain a deeper insight into the effect of fibre-matrix adhesion and fibre orientation on the mechanical performance of resistance welded joints. An analysis of the influence of those factors on the strength of resistance welded GF/PEI joints was performed with a thorough evaluation of the failure modes. The mechanical properties of the welded joints were evaluated via single lap shear testing and the failure modes of the joins were analyzed through visual inspection, optical microscopy and scanning electron microscopy (SEM).

2.2 Experimental

2.2.1 Laminates

The material used in this study was 8HS glass fibre reinforced PEI composite sup-plied by TenCate Advanced Composites, The Netherlands. Two types of GF/PEI prepreg, SS0303 and TC7781, were used, which only differ in the glass fibre sizing, as listed in Table 2.1. Due to the weave pattern of the satin woven fabric, the ratio of warp and weft yarns per unit area is different for the two sides of a single ply. In this paper, “warp side” was used to indicate the side where a higher ratio of warp yarns can be seen, while “weft side” was used to indicate the side where a higher ratio of weft yarns can be seen. Furthermore, the main apparent orientation of the fibres (the main direction of the longest visible fibre bundles) is also differ-ent on either side of a single ply. Therefore, a laminate with a stacking sequence of [[(0◦/90◦)]n]s , where 0◦ is the direction of warp yarns and 90◦ is the direction

of weft yarns, can be built with two different outer surfaces (see Figure 2.1): (1) “Type I” -the main apparent orientation of the fibres in the 0◦ direction of the laminate and (2) “Type II” -the main apparent orientation of the fibres in the 90◦

(39)

2.2. Experimental 15

direction of the laminate. Similarly, the laminate: [[(90◦/0◦)]n]s can also be built

either Type I or Type II. Since both Type I [[(0◦/90◦)]n]sand Type II [[(90◦/0◦)]n]s

laminates have warp outer surfaces while both Type II [[(0◦/90◦)]n]s and Type I

[[(90◦/0◦)]n]s laminates have weft outer surfaces, in this paper, Type I-warp, Type

II-warp, Type I-weft and Type II-weft were used to refer to these four types of laminates, respectively.

Figure 2.1: Schematic drawing of [[(0◦/90◦)]n]s and [[(90◦/0◦)]n]s laminates stacked with

different surface sides: Type I-warp, Type I-Weft, Type II-warp and Type II-Weft.

Product code SS0303 (050 003 8463 34000) TC7781 (P127 001 8463 34000) Resin type Ultem 1000 PEI Ultem 1000 PEI

Glass fabric style US Style 7781 US Style 7781

Yarn type EC6 68tex EC6 68tex

Mass of fabric 300 g/m2 _{300 g/m}2

Fabric manufacturer TenCate Hexcel

Fibre treatment Volan A (chromium methacrylate) Hexcel TF970 (aminosilane)

Table 2.1: Characteristics of the two types of GF/PEI prepregs used in this study.

A hot platen press was used to fabricate the laminates. Before consolidation, the prepreg plies were dried in an oven at 260 ◦C for 3 hours in order to fully remove any residual moisture and N-Methyl-2-pyrrolidone (NMP, a solvent used in the impregnation process). Then, the plies were laminated and consolidated in the hot platen press at 320 ◦C under 2.0 MPa pressure for 20 min. After the consolidation process, the adherends were immediately cut out from the laminates using a water cooled diamond saw and then welded or tested. If welding was not performed immediately after the cutting operation, the composite adherends were stored in a desiccator.

(40)

2.2.2 Heating element

A stainless steel (AISI 304L) mesh with a plain woven pattern was used as the heating element. In this study, a M200 metal mesh (0.04 mm wire diameter and 0.09 mm open gap width) was used since it is known to provide resistance welded joints with excellent properties [10]. The heating elements were cut out from the mesh sheet into strips of 13 mm in width and 250 mm in length. In order to fill the open areas of the mesh and provide a resin rich area between the heating element and the laminates, the mesh was sandwiched between two layers of 60 µm thick PEI resin films prior to the welding process (no pre-consolidation of the mesh and the extra resin was carried out).

2.2.3 Resistance welding

The in-house developed resistance welding setup was used in this study, see Fig-ure 2.2. It consists of a power supply unit, two pneumatic systems, two copper connectors, two thermal insulators and a data acquisition (DAQ) system. A com-puter controlled power supply unit, Delta Elektronika, with a maximum DC output of 45 A and 70 V, was used to provide the welding energy. Two pneumatic cylin-ders were used to provide the welding pressure, and two other pneumatic cylincylin-ders were used to apply the clamping pressure. High-density fibre (HDF) wood blocks covered with a layer of 127 µm thick polyimide film were used as thermal insula-tors. The resistance welding process was controlled by a labview program, and the welding parameters, such as welding voltage and welding current, were recorded by a DAQ system during the welding process.

Figure 2.2: The experimental welding setup.

(41)

used for the welding. Adherends with dimensions of 100 mm × 192 mm were single lap welded with an overlap length of 13 mm, where the 0◦ direction of the original laminates was kept parallel to the shortest side of the adherends, as shown in Figure 2.3. Different welding times of 30 s, 40 s, 55 s, 90 s, 100 s and 120 s, corresponding to welding energies of 2.4 MJ/m2_{, 3.2 MJ/m}2_{, 4.4 MJ/m}2_{, 7.2}

MJ/m2_{, 8.0 MJ/m}2 _{and 9.6 MJ/m}2_{, were used for the analysis of basic failure}

modes of the resistance welded joints. Later, the heating time was set to 55 s for the investigation on the effect of fibre-matrix adhesion and fibre orientation.

Figure 2.3: Positions of the thermocouples used for the temperature measurements during the resistance welding of GF/PEI.

Extra welds were performed in order to determine the welding temperatures at each heating time. In these welds two K-type thermocouples (φ 0.1 mm), insulated with polyimide tape, were placed in the middle of the overlap, as schematically shown in Figure 2.3. The welding temperature for each heating time was calculated by using the average of the maximum reading of the two thermocouples in at least three different experiments.

2.2.4 Mechanical testing and characterization methods

Single lap shear tests were used to evaluate the strength of the resistance welded joints. The preferred propagation paths for cracks at the weldline were assessed via both Mode I and Mode II interlaminar fracture tests. The effect of fibre sizing on the laminate properties was quantified through flexural tests (or three point bending tests) and short-beam shear tests. Detailed information on each type of test as well as the samples used is given below and summarized in Table 2.2.

(42)

Failure mode Effect of fibre-matrix Effect of fibre analysis adhesion orientation LSS Type of samples Welded Welded Welded

Prepreg type SS0303 SS0303 and TC7781 SS0303 Stacking sequence [(0◦_/90◦_)]

4s [(0◦/90◦)]4s [(0◦/90◦)]4sand [(90◦/0◦)]4s

Main apparent fibre orientation Type I Type I Type I and Type II DCB Type of samples Press-consolidated Press-consolidated Press-consolidated & ENF Prepreg type SS0303 SS0303 and TC7781 SS0303

Stacking sequence [[(0 ◦_/90◦_)] 4s]sand [[(0◦_/90◦_)] 4s]s [[(0◦_/90◦_)] 4s]sand [[(0◦_/90◦_)] 4s/M 200/P EI]s [[(90◦/0◦)]4s]s

Main apparent fibre orientation

Type I Type I Type I and Type II (Crack opening surfaces)

3PB Type of samples - Press-consolidated -Prepreg type - SS0303 and TC7781 -Stacking sequence - [(0◦_/90◦_)]

4s

-Main apparent fibre orientation - Type I -SBS Type of samples - Press-consolidated -Prepreg type - SS0303 and TC7781 -Stacking sequence - [[(0◦_/90◦_)]

4s]s

-Main apparent fibre orientation - Type I

-Table 2.2: Mechanical tests performed in this study (LSS: lap shear strength test; DCB: double cantilever beam test; ENF: end notched flexure test; 3 PB: three point bending test; SBS: short beam shear test).

Single lap shear test

Single lap shear tests were performed according to the ASTM D1002 standard. Six test specimens were cut from each weld using a water cooled diamond saw into final dimensions of 187.3 mm long and 25.4 mm wide. The tests were performed at room temperature (23 ± 3◦C ) and relative humidity of 50 ± 5%. A Zwick/Roell 250KN testing machine was used with a constant crosshead speed of 1.3 mm/min. At least six samples were tested for each individual welding setting. The apparent lap shear strength (LSS) of the joints was calculated as the maximum load divided by the total overlap area. The facture surfaces of the welded specimens were examined visually and by SEM.

GF/PEI laminates made of different types of prepreg and with various stacking sequences were used in this study. For the analysis of the basic failure modes, SS0303 Type I laminates with a [(0◦/90◦)]4s stacking sequence (Type I-Warp) were

used, as shown in Figure 2.4. For the investigation of the effect of fibre-matrix ad-hesion, SS0303 and TC7781 Type I laminates with a [(0◦/90◦)]4s stacking sequence

(Type I-Warp) were used. Finally, for the analysis of the effect of fibre orientation, SS0303 Type I and Type II laminates with both [(0◦/90◦)]4s and [(90◦/0◦)]4s

stack-ing sequences (Type I-warp, Type I-weft, Type II-warp and Type II-weft joints) were used.

Fracture toughness tests

Mode I and mode II interlaminar fracture toughness energies, GIC and GIIC, were

measured through double cantilever beam (DCB) and end notched flexure (ENF) tests following the ASTM D 5528-01 and Airbus AITM 1.0006 standards, respec-tively. The modified beam theory was used for calculating GIC values as

recom-mended by the ASTM D 5528-01 standard. A pre-crack, 40 mm long for DCB or 50 mm long for ENF, was created by inserting a layer of 25 µm thick Kapton film

(43)

Figure 2.4: Schematic drawing of single lap shear GF/PEI samples with different fibre orien-tations at the welding surfaces (warp yarn shown as black and weft yarn shown as grey in the figure).

in the middle plane of the specimens. The longitudinal edges of the DBC and ENF specimens were coated with a thin layer of water-based white fluid to make the visual detection of crack tip more obvious. A Zwick/Roell 20KN testing machine was used for the tests with cross-head speeds of 3 mm/min and 1 mm/min for DCB and ENF tests, respectively. Six samples were tested for each type of specimen.

Two types of tests were carried out (see Figure 2.5), namely fracture toughness measurements for cracks propagating through the laminate and fracture toughness measurements for cracks propagating within the weld line. For the analysis of cracks propagating through the laminate, both SS0303 and TC7781 GF/PEI laminates with stacking sequences of [[(0◦/90◦)]4s]s and [[(90◦/0◦)]4s]s were built, with either

Type I or Type II surfaces at the crack opening interfaces, namely Type I-warp, Type II-warp, Type I-weft and Type II-weft. For the analysis of cracks propagat-ing within the weld line, press consolidated instead of welded samples were used in order to exclude the undesirable effects caused by a non-uniform weld quality in crack propagation [18]. SS0303 laminates with a [[(0◦/90◦)]4s/M 200/P EI]s

stack-ing sequence were built for this purpose, where PEI denotes one layer of 60 µm thick PEI resin film. In order to keep the cracks propagating along the mesh-matrix interface and to keep the specimen symmetric, two metal meshes were consolidated inside the GF/PEI laminates with two PEI films in between (middle line). The pre-crack was created in between the two PEI layers.

Resistance Welding of Thermoplastic Composites

Resistance welding

of

thermoplastic composites

Process and performance

Resistance welding

of

thermoplastic composites

Process and performance

Summary

Samenvatting

Contents

I

A micro-level study

11

II

A macro-level study

85

List of Tables

List of Figures

Chapter 1

Introduction

1.1

Thermoplastic composites for aircraft

struc-tures

1.2

Joining methods for TPCs

1.3

Resistance welding of TPCs

1.4

Research objective

1.5

Thesis outline

Bibliography

Part I

A micro-level study

Chapter 2

Effect of fibre-matrix adhesion

and fibre orientation

2.1

Introduction

2.2

Experimental

2.2.1

Laminates

2.2.2

Heating element

2.2.3

Resistance welding

2.2.4

Mechanical testing and characterization methods