1-s2.0-S1359835X98000554-main

7/30/2019 1-s2.0-S1359835X98000554-main

http://slidepdf.com/reader/full/1-s20-s1359835x98000554-main 1/7

A comparison of physical properties of

glass fibre epoxy composites produced by wetlay-up with autoclave consolidation and resintransfer moulding

D. Abraham, S. Matthews and R. McIlhaggerEngineering Composites Research Centre (ECRE), Department of Electrical and Mechanical Engineering, University of Ulster at Jordanstown, Shore Road,Newtownabbey BT37 0QB, Northern Ireland, UK

Comparisons are reported for composite samples of similar resin and fibre systems which were processed using thewet lay-up with autoclave consolidation and resin transfer moulding (RTM) by vacuum impregnation. Similardegrees of cure were obtained for laminates using the two methods of processing and the arising thermal andphysical properties (tensile, flexural, interlaminar shear strength (ILSS), void content and thickness variation)were measured. The fibre dominated properties (i.e. flexural and tensile strength) were found to be higher for theautoclaved samples due to the higher volume fraction arising from the superior compaction pressure, althoughwhen normalised on the basis of fibre volume fraction the results were similar. The matrix dominated ILSS valueswere higher for the RTM samples and this was attributed to improved wetting, reduced void content and a slightlylower degree of cure. Thermal analysis also indicated that the autoclaved (60% glass fibre by volume) compositeattained a slightly higher glass transition temperature than that achieved by RTM (50% fibre by volume) forsimilar cure times and cure temperatures. The significance of the results in an industrial context is discussed. ᭧1998 Elsevier Science Ltd. All rights reserved

(Keywords: composites manufacture; E. resin transfer moulding (RTM); E. autoclave)

INTRODUCTION

Aerospace composites manufacturing

The aerospace sector remains firmly rooted in traditional

autoclaving routes for processing the majority of

their composite parts. Autoclaving is a well understood

and mature technology which is capable of producing

material of consistently high quality, with high (greater than

55%) fibre volume fraction and low (less than 2%) void

content. Numerous studies of autoclaving have been

published, although most of these relate to pre-preg.

Stringer1 investigated autoclave consolidation for wet lay-

up carbon/epoxy laminates and reported a fibre volume

of 58% and less than 2% voidage by incorporating a

dwell period at the start of the cure cycle before applying

the consolidation pressure. Wenger2 has investigated

autoclave cure cycle optimisation for the facility used

in this investigation, varying the point of pressure

application, dwell temperature, dwell period and heat-upramp, and studying the quality of the final composite

component.

Aerospace resin transfer moulding

Despite the traditional use of the autoclave (almost

universally with pre-preg.), interest and confidence is

increasing for resin transfer moulding (RTM), usually

driven by materials costs. This has led to the production

of aircraft components such as radomes3, bullet fairings4,

and propeller blades5. Previous development studies for

structural components by aerospace operators includes

missile airframes6 and a highly loaded crank 7. With the

development of automated methods of preform produc-

tion8,9, parallel research on processing methods is aimed at

fast and reliable methods of impregnation with reduced

capital and running costs compared to either traditional wet

lay-up or vacuum bag and autoclave (pre-preg.) moulding.

RTM offers a useful alternative in several industrial sectors,

since it reduces the labour costs and environmental concerns

associated with wet lay-up and provides substantial savings

in materials costs when compared with the use of pre-pregs.

A further advantage of RTM is the wide range of processvariants, which range from relatively simple vacuum

impregnation processes requiring little by way of capital

Composites Part A 29A (1998) 795–8011359-835X/98/$-see front matter

᭧ 1998 Elsevier Science Ltd. All rights reserved.

PII: S1359-835X(98)00055-4

795

7/30/2019 1-s2.0-S1359835X98000554-main


equipment to hybrid RTM/SRIM operations and their

associated high investment.

Vacuum driven processes

Recent years have seen significant activity in vacuum

infusion processes for composites manufacture. Althoughapplications of this process have existed since the 1940s,

increasingly stringent legislation in several countries

concerning the emission of volatiles in the workplace has

led to investigations and numerous production examples of

vacuum infusion as a replacement for open mould

processes. This technique uses vacuum to transfer the

liquid resin into the mould cavity containing the fibre

reinforcement. The cavity itself may be formed by two

matched tools or a single hard tool and a vacuum film. In

common with all fabrication processes, it is important to

maintain voidage in the final laminate as low as possible,

and since RTM (generally) involves no further consolida-

tion, the impregnation process needs to be controlled withsome care. Increasing void content is well known to be

detrimental to mechanical performance, as illustrated for

example by Judd and Wright10, who reported that an

increase in the void content of 1% resulted in a 7% decrease

of the short beam shear properties (up to 4% void content).

Several detailed reviews of void formation in RTM have

been presented11,12. Hayward and Harris13 investigated the

effects of vacuum assistance and reported a reduction

in voidage during positive pressure RTM. Abraham

and McIlhagger14 considered several injection strategies

and gating arrangements, and found these to affect

the degree of visual porosity and moulding translucency.

Qualitative comparison showed that vacuum impregnation

and peripheral gating produced the best quality

mouldings.

Clearly, the properties of vacuum impregnated mouldings

are also strongly dependent upon the fibre volume fraction

which is achieved. Unlike conventional RTM, which iscarried out in a fixed height cavity, the use of a flexible

membrane to close the cavity means that the degree of fabric

compaction (thus the fibre volume fraction) depends upon

the pressure difference across the membrane15.

Autoclaving versus RTM from a manufacturing perspective

Table 1 summarises the main features that characterise

autoclaving and RTM. It is evident that autoclaving is the

favoured method when the need for high product quality

dominates over production rate (as evidenced by its

continued dominance of aerospace manufacture), while

RTM appears to offer lower cost manufacture and entry tohigher volume markets, typified by the growing interest

from within the automotive sector.

In addition to the savings on materials costs compared

with pre-pregs and the improvements in the working

environment compared with wet lay-up, RTM offers

considerable advantages over either alternative, due to a

reduction in the labour content and higher production rates.

Operator exposure to liquid resins is also much reduced. An

approximate time analysis for the two (laboratory) pro-

cesses is shown in Table 2, which suggests that RTM offers

a significant time saving of 1.5 h (25%) over the equivalent

Wet lay-up and resin transfer moulding: D. Abraham et al.

796

Table 1 Main characteristics of RTM and autoclaving

Property RTM Autoclave

Production output Moderate (10 2–10 5 /annum) Low (10 1–10 2 /annum)Void content Low ( Ͻ 2%) Low ( Ͻ 2%)Labour intensity Low–moderate HighAchievable V f Moderate (30–65%) High (50–70%)Typical applications Low to medium volume, automotive parts, non-structural components

(e.g. automotive spoilers), structural items (e.g. propeller blades, missile boxes)Aerospace industry, Formula 1automotive, sporting goods

Table 2 Comparison of Wet lay-up with autoclave consolidation and RTM: breakdown by operations for laboratory scale processes

Wet lay-up and autoclave RTM

Production step Time (min) Production step Time (min)

Mould preparation H 22.5a Mould preparation H 22.5 a

Fabric cutting H (including preparation) 25 Fabric cutting H (including preparation) 25Resin mixing H I (including preparation) 25 Resin mixing H I (including preparation) 25Hand lay-up H I 120b Resin degassing 60Vacuum bagging H 30 Tool assembly þ testing H 30Autoclave cycle 192 Resin injection H 12Autoclave debagging H 10 Resin curing 99

Tool cooling 30Tool ejection and preparation for next run H 30

Total processing time 424.5 Total processing time 333.5Resin post-curing at 145ЊC 452 Resin post-curing at 145ЊC 368.4Total production time 876.5 Total production time 701.9Manual processing time (% of total production time) 232.5 (26.5%) Manual processing time (% of total production time) 144.5 (20.6%)Direct Contact with resin (% of total production time) 145 (16.5%) Direct contact with resin (% of total production time) 25 (3.6%)

aAutoclave and RTM mould preparation takes 90 min, but the preparation will endure a minimum of four production runs.bAutoclave hand lay-up takes two operators working for 60 min.H, these steps involve manual processing.I, these steps involve direct operator contact with the resin.

7/30/2019 1-s2.0-S1359835X98000554-main


autoclave processing time. Further gains are evident when

post-curing is taken into account. The major portion of the

time savings using RTM arises because the majority of the

initial processing takes place at a higher temperature than

for wet lay-up with autoclave consolidation; thus the

lengthy heat-up times are eliminated. RTM is also

significantly less labour intensive and eliminates almost1.5 h of manual processing time. The operator exposure to

wet resin is limited to the relatively short time required for

mixing. In contrast, the wet lay-up alternative involves

exposure for the entire laminating process (2 h). It must be

recognised that the data presented here are for research

laboratory processes, which are not necessarily

representative of conventional industrial practice.

Nonetheless it is likely that the general, rather than

the specific, findings read over to industrial

operations. Clearly, on an industrial scale, where batch

mixing would serve several mouldings, the level of

automation would be higher, and certain procedures

(e.g. preforming, resin degassing) could be conductedoff-line, the advantages of the closed mould process

would be magnified.

A quantitative comparison of RTM and autoclaving

In addition to the manufacturing requirements, it is useful

to compare the quality and physical properties arising from

RTM laminates with those of competing processes. Few

direct comparisons of autoclaving and RTM have been

reported. Hayward and Harris16 compared properties arising

from vacuum assisted RTM with several alternative

processes, although different resin systems were used ineach case. The main comparison here has been with an

autoclave route which, although used conventionally with

pre-pregs, was used in the present study with a dry fabric/

wet lay-up approach, in order that the same materials could

be used in both processes. Tensile, flexural and interlaminar

shear strength (ILSS) material tests were used to compare

the physical properties arising from the two processing

routes. In addition void contents, volume fraction of fibre

and the thickness variation within the laminate were

examined. The degree of cure, as determined using dynamic

mechanical analysis (DMA) and differential scanning

calorimetry (DSC)17,18, was also assessed although, in

order to minimise the influence on properties, cure timeswere adjusted in an attempt to achieve similar glass

transition temperatures T g.

In order to examine the potential of RTM composites as

replacements for laminates which are traditionally

autoclaved, the performance of two representative

materials has been compared. These were produced using

identical materials using laboratory equipment and

processes as defined below. Vacuum impregnation was

used for the RTM plaques in this work to achieve a glass

fibre volume fraction in the region of 50%. Parallel studies

using wet lay-up with subsequent autoclave consolidation at

12 bar pressure used a similar arrangement with a flexiblemembrane with the mould cavity evacuated to

approximately 1 bar vacuum.

EXPERIMENTAL DETAILS

Materials

The reinforcement material used was E-glass, 136 tex

yarn. This was woven by CS-Interglas Ltd. into a plain

weave structure of 7.1 ends/cm in the warp direction and7 picks/cm in the weft direction to give an areal density of

196 g/m2. Twelve fabric layers were cut from the roll stock

and stacked so that weft and warp were aligned in the same

direction for every layer. The fabric preform dimensions

were 355 ϫ 355 mm and 440 ϫ 168 mm for the RTM and

wet lay-up plaques respectively. The matrix was an epoxy

resin supplied by Ciba Polymers. This was a two-part

system of LY-564 resin and HY-2954 hardener which is

intended for both RTM and hand lay-up applications. The

mix ratio used (as recommended by the manufacturers) was

100:35 parts by weight resin to hardener.

Processing

The RTM technique consisted of first mixing the resin

and then degassing it for 1 h at 30ЊC to remove air

suspended in the liquid. It was then injected into the tool at

75ЊC at a supply pressure of 750 mbar (absolute). After

injection, the temperature was increased to 100ЊC at 0.64ЊC/

min (this relatively slow heating rate was a consequence of

the small oven and high thermal mass of the mould). The

resin supply pressure was increased to atmospheric and the

composite cured at this temperature for 1 h. The plaque was

then demoulded and subsequently post-cured at 145ЊC for

368 min. These temperature and pressure conditions areshown in Figure 1a. The tool, as seen in Figure 2, consisted

of an aluminium base containing thermocouples for process

monitoring and oven temperature control. Resin inlet and

outlet ports were also positioned on the base plate. In order

to achieve a good surface finish on both sides of the

moulding a flat metallic caul plate was used. The area of the

caul plate was slightly larger than the fabric to form a

peripheral gap which would permit the resin flow front to

surround the preform and form a peripheral gate. The tool

was sealed using a rubber gasket and nylon bag arrangement

to which a vacuum was applied. The purpose of the vacuum

was twofold; firstly vacuum only was used to impregnate the

dry fabric, and secondly to compact the fibre to obtain ahigher volume fraction.

The hand lay-up/autoclave technique consisted of mixing

the resin and then using it in 12 approximately equal

portions to impregnate each layer of the cut fabric. The resin

was poured onto a sheet of nylon film and rolled out to the

size of the fabric, which was then placed carefully onto the

thin film of resin. The fabric was then lightly rolled in order

to assist wet-out and then transferred to the tool base plate.

All 12 layers were treated in the same way and laid up on top

of each other. Release film was placed over the top fabric

layer followed by a metal caul plate and finally two layers of

breather cloth. The assembly was then vacuum bagged andtransferred to the autoclave, where it was subjected to the

temperature and pressure conditions shown in Figure 1b.

797


7/30/2019 1-s2.0-S1359835X98000554-main


The temperature was ramped at 4ЊC/min to 50ЊC, and then

held at this temperature for 1 h before being further ramped at

3ЊC/min to 100ЊC, where it was held for 1.5 h before beingfinally ramped down at 4ЊC/min to 40ЊC. As well as there

being a vacuum applied to the mould cavity, the autoclave

pressure was raised to 12 bar (gauge). After the thermal cycle,

the tool was removed from the autoclave and the flat

plaque demoulded before subsequent post-cure at 145ЊC for

452 min.

Preparation and analysis

The plaques were sectioned using a diamond wheel.

Specimens were then polished to the required dimensions

using 320 grit wet and dry paper. A grid size of 10 mm wasdrawn on the plaques to map thickness variation and the

thickness was measured at the line intersections using a

micrometer fitted with barrels of 6.8 mm in diameter. The

Composite Research Advisory Group (CRAG) standards for

tensile, flexural and ILSS methods 302, 200 and 100respectively were used19. The DMA method required a

60 mm ϫ 10 mm sample which was a flexure bar cut from

the composite plaques, whereas the DSC technique only

required a 10 mg sample weight.

Void content was calculated by comparing the composite

density calculated by resin burn off (to ASTM D2584-94)

and the value calculated by the liquid displacement method.

The latter used a Sartorius Density Determination Kit

(Model Number YDK 01-OD) coupled to a Sartorius MC1

series balance (Model Number RC 210 D). Prior to the tests,

the samples were conditioned in a humidity chamber at

23ЊC and 50% relative humidity for 40 h. Five samples were

taken from along a length of the mouldings, in order to give

a representative void content.


798

Figure 1 (a) i, RTM temperature cycle; ii, RTM pressure cycle. (b) i, Autoclave temperature cycle; ii, Autoclave pressure cycle

7/30/2019 1-s2.0-S1359835X98000554-main


RESULTS

The properties of the RTM and wet-layup/autoclave plaques

are compared in Table 3. The results (Table 3b) show that

autoclave moulding achieved an extra 13.5% fibre volume

fraction due to the superior compaction pressure of 13 bar.

This produced increases in flexural and tensile strength and

modulus. However, it is worth noting that the RTM plaque

was compacted and impregnated at a fibre volume fraction

of 50% using only the pressure difference caused by

evacuating the mould cavity. It is also interesting that the

ILSS arising from the RTM technique was 70% higher than

that from the autoclave method. Since ILSS is generally

accepted to be a good indicator of matrix quality and void

content, this suggests that RTM more than adequately

matches the autoclave process in terms of matrix quality.

However the slightly lower degree of cure of the RTM

sample (Table 3c) is also likely to be a contributing factor

here. In order to determine the influence of the processing

method on the other mechanical properties, two approaches

were considered. The first was to normalise the properties

on the basis of the volume fraction of fibre. The nominal

fibre volume fraction was calculated from the tex, fibre

density and yarn spacing in the fabric (which provided the

reinforcement superficial density) and the volume of the

laminate, using the mean thickness values from Table 3b.

The normalised flexural strength and modulus for the

autoclaved laminates are 9% and 6% higher than those

achieved using RTM while in tension, the situation isreversed with the RTM laminates exhibiting increases of 4

and 8% in normalised strength and modulus respectively

over the autoclaved laminates. The second method involved

the estimation of tensile properties using a rule of mixtures.

This was done by considering the effective volume fraction

of fibre in either the warp or the weft direction (approxi-

mately 50% in each case) and neglecting the reinforcing

effect of the transverse fibres. The effective fibre volume

fraction in each of the two yarn directions was determined

by the following expression:

V f ¼nsestexs

10t c r

s

where V f is the fibre volume fraction expressed as a percen-

tage, ns is the number of plies, e s is the number of yarns per

metre, texs is the linear density in g/km, t c is the final com-

posite thickness (mm) and rs is the fibre density in kg/m3 .

The matrix properties (tensile strength and modulus)

were taken from the supplier data sheet and the fibre tensile

modulus was also taken from supplier information. The fibre

tensile strength was calculated by measuring the tensile

strength of one yarn. The constituent properties and the

warp direction fibre volume fractions used in the ROM

calculations are shown in Table 4.

For both processing techniques the laminate properties

were over-estimated. The estimated tensile strengths were

20% and 28% higher than the measured values for the RTM

and autoclaved plaques respectively while the equivalent

tensile moduli approached 50% error in both cases. This is

attributable largely to the simple nature of the ROM

approach and highlights the difficulty of predicting proper-

ties in woven structures. The simplifying assumptions used

in the calculation include; a non-crimp fibre architecture;

zero void content; continuous and unidirectional fibres (in

each yarn direction); that the load was applied along the

799


Table 3 Laminate properties arising from wet lay-up with autoclave consolidation and RTM

Property RTM Autoclave RTM% differ-ence

(a) Mechanical PropertiesILSS (MPa) 17.73 (%CV ¼ 4.9), normalised ¼ 35.16 10.45 (%CV¼ 2.4), normalised ¼ 16.35 þ69.7%Flexural strength (MPa) 415.7 (%CV ¼ 3.8), normalised ¼ 824.4 571.2 (%CV¼ 5.3), normalised ¼ 893.8 ¹27.2%Flexural modulus (GPa) 19.63 (%CV ¼ 3.2), normalised ¼ 38.93 26.43 (%CV¼ 5.6), normalised ¼ 41.36 ¹25.7%Tensile strength (MPa) 274.0 (%CV ¼ 8.2), normalised ¼ 543.4 332.9 (%CV¼ 6.6), normalised ¼ 521.0 ¹17.7%Tensile modulus (GPa) 13.91 (%CV ¼ 7.9), normalised ¼ 27.58 16.36 (%CV¼ 10.5), normalised ¼ 25.60 ¹15.0%Calculated theoretical tensile strength (MPa) 348.3 421.6 ¹17.4%Calculated theoretical tensile modulus (GPa) 19.87 24.61 ¹19.3%

(b) Physical propertiesCalculated V f from mean thickness (%) 50.4 63.9 ¹21.1%Mean thickness (mm) 1.775 (%CV ¼ 4.3) 1.394 (%CV ¼ 2.1) þ27.3%Voidage (%) 1.507 (%CV ¼ 13.4) 1.566 (%CV ¼ 14.1) ¹3.8%

(c) Thermal propertiesT g (DMA) 145.0 (%CV ¼ 0.4) 155.6 (%CV ¼ 0.6) ¹6.8%T g (DSC) 142.9 (%CV ¼ 2.1) 149.9 (%CV ¼ 2.2) ¹4.7%

Figure 2 RTM tool construction

7/30/2019 1-s2.0-S1359835X98000554-main


entire length of the fibre; that the fibres and matrix wereperfectly bonded; that all fibres broke simultaneously after

application of the tensile load; and that there were no

variations in constituent properties or fibre volume fraction.

In practice, crimp and misalignment of the yarns during

handling means that some departures from the ideal

architecture are inevitable. Consequently, the properties

will be reduced compared to the estimated values. The use

of a plain weave, which maximises crimp, magnifies these

effects. Also, prior to separation, damage mechanisms such

as disbonding and delamination mean that the tensile load

will no longer be shared by the two phases in the simple way

assumed by the ROM. In addition, manufacturing induced

defects such as fibre breakage, matrix cracking and voidageall contribute to reductions in measured strength and

modulus.

Micrographs taken from representative areas of each

laminate type are shown in Figure 3. The effects of the

higher autoclave pressure are immediately apparent from

the high degree of compaction which is evident in both the

90Њ fibre bundles which are visible and in the lenticular 0Њ

tows running out of the section. Also, there are fewer visible

resin rich areas in the autoclaved sample, which is again

attributed to the higher pressure which caused nesting of the

plies and compaction of the tows. These two effects are

consistent with the higher fibre volume fraction of theautoclaved sample. It is also interesting to note that neither

micrograph displays any significant voidage, which con-

firms the quantitative results produced by the density

measurements at approximately 1.5% in both cases.

The results suggest generally that the only inferiority in

the vacuum impregnation technique, compared to the wet

lay-up and autoclave alternative for glass fibre/epoxy, arises

from the lower fibre fraction which was achieved. Clearly,

the fibre fraction could be raised (for vacuum impregnation)

by introducing a secondary consolidation process, although

it remains to be seen if such laminates can be filled

satisfactorily using the low pressure gradient which is

available from vacuum impregnation if the entire compac-

tion process is carried out beforehand. If this is not the case,

then a shift to conventional, pressure-driven RTM process is

indicated. Mould stiffness is critical in such processes to

control the fibre volume fraction and minimise the part

thickness variation. However, tooling costs are required to

be minimised to reduce the part cost and increase the

flexibility in production volumes. In this study, a flexible

upper tool was used. Atmospheric pressure was used to

compact and impregnate the fabric reinforcement.

Comparison of the thickness distribution compared with

that for an autoclave consolidated plaque (Table 3b) shows

that the coefficient of variation (CV) for the RTM plaquewas twice that from the autoclave. It seems reasonable to

expect corresponding variations in the fibre volume fraction

and mechanical properties within the plaque (particularly in

intensive properties such as the strength values). However,

no significant difference in the variability of properties is

evident from Table 3.

DMA and DSC were used to determine the T g of both

composites. DMA is often reported18 to give T g results that are

10ЊC higher than those given by DSC. Five tests on eachmoulding were completed for both methods and the mean

values are listed in Table 3c. The results from both techniques

suggest that the T g of the RTM composite is slightly lower than

that from the autoclave with the DMA results, suggesting a

greater difference between the two processes. This difference

arises not from any inherent differences in the processing

methods, but rather in the different cure schedules which were

used. The results indicate that the RTM post-cure cycle was

inadequate. This was attributed to the reliance on data from

simulated thermal cycles using a dielectric analyser (DEA) to

determine the post-cure requirements. The latter were

conducted using similar fibre volume fractions, whereas thedifferent consolidation pressures during the two processes

produced different volume fractions. Clearly, this factor will


800

Figure 3 (a) Autoclave micrograph at 35 ϫ magnification. (b) RTMmicrograph at 35 ϫ magnification

Table 4 Constituent properties

Constituent Tensile strength (MPa) Tensile modulus (GPa) Warp V f (RTM) Warp V f (autoclave)

Fibre 1139 71 0.253 0.322Matrix 80 2.5 0.747 0.678

7/30/2019 1-s2.0-S1359835X98000554-main


influence the thermal properties and therefore the progression

of the cure cycle. The RTM plaques had a higher resin content

with lower thermal conductivity, a higher heat capacity, and

resulted in a lower degree of cure. The intrinsic differences

between DSC and DMA measurements must also be

considered. The former measures a thermochemical change

in the resin matrix, and the difference in T g measurements isdue only to the different states of cure present in the two

sample types. However, since the DMA measures a thermo-

mechanical effect, this is more likely to be influenced by the

fibre content, and this may magnify the apparent differences

between the two processes. It is also worth noting that the CV

for the DMA results is lower than that for the DSC. The

improved repeatability may be related to the effects of sample

size and the somewhat variable fibre content in the relatively

small DSC samples.

CONCLUSIONS

RTM, based upon vacuum impregnation, offers materials and

labour cost savings compared with conventional autoclaving,

operational and health and safety benefits compared to wet

laminating and reductions in tooling costs compared to normal

(positive pressure driven) RTM. This study has shown that

laminate quality from vacuum impregnation is comparable

with autoclave consolidation when the same materials are

used. However, the fibre fractions available from vacuum

impregnation are somewhat lower due to the lower con-solidation pressure and this influences the mechanical and

thermal properties of the laminate. A secondary consolidation

process would overcome this problem.

ACKNOWLEDGEMENTS

The authors would like to thank CS-Interglas for supplying

the glass fibre and also David Hurry of Ciba Polymers for

supplying the epoxy resin. Also, thanks are due to RoyCarton and Maurice Jamieson for their assistance in

processing and testing the samples respectively. Financial

support from the European Regional Development Fund

(ERDF) and the Technology Development Programme

(TDP) is also gratefully acknowledged.

REFERENCES

1. Stringer, L. G., Optimization of the wet lay-up/vacuum bag processfor the fabrication of carbon fibre epoxy composites with high fibrefraction and low void content. Composites, 1989, 20(5), 441–452.

2. Wenger, W., Investigations into 3-D reinforcements for compositematerials. D. Phil. thesis, University of Ulster, 1993.

3. Unger, E. W. and Wiedmann, J., Application of RTM techniques in

the manufacturing of a radome design for high rain resistance. In 8th European Electromagnetic Structures Conference, Nottingham,UK, 1995.

4. Becker, W., RTM simultaneous design and tooling reduces cost andhandling time. In 23rd International SAMPE Conference, KiameshaLake, New York, 1991, pp. 601–608.

5. McCarthy, R. F. J., Haines, G. H. and Newley, R. A., Polymercomposite applications to aerospace equipment. Composites Man-

ufacturing, 1994, 5(2), 83–93.6. Marchbank, I., Automated RTM for an airframe component, In

Materials and Processing—Move into the 90s, Elsevier ScienceB.V., Amsterdam, 1989, pp. 75–86.

7. Gottesman, T., Green, A. K., Itah, Y., Rimoch, Y. and Solomon, L.,Development of resin transfer moulding for manufacture of highlyloaded components. In Advanced Materials: Cost Effectiveness,

Quality Control, Health and Environment . SAMPE/ElsevierScience B.V., 1991, pp. 337–346.

8. Chestney, J. A. and Sarhadi, M., A prototype manufacturing cell forautomation assembly of fibre reinforced composite performs. In 4th

International Conference on Automated Composites, The AlbertHall, Nottingham, 1995, pp. 439–446.

9. Turner, M. R., Rudd, C. D., Long, A. C., Middleton, V. and McGee-hin, P., Automated fibre lay-down techniques for preform manufac-ture. In 4th International Conference on Automated Composites,The Albert Hall, Nottingham, 1995, pp. 431–438.

10. Judd, N. W. C. and Wright, W. W., Voids and their effects on themechanical properties of composites—An appraisal. SAMPE Jour-

nal, 1978, 14, 10–14.11. Abraham, D. and McIlhagger, R., A review of liquid injection tech-

niques for the manufacture of aerospace composite structures. Poly-mers and Polymer Composites, 1996, 4(6), 437–443.

12. Staffan Lundstrom, T., Void formation and transport in manufactur-ing of polymer composites. Doctoral Thesis 1996:184D, Lulea Uni-

versity of Technology, Sweden.13. Hayward, J. S. and Harris, B., Processing factors affecting the qual-

ity of resin transfer moulded composites. Plastics and Rubber Pro-cessing Applications, 1989, 11, 191–198.

14. Abraham, D. and McIlhagger, R., Vacuum assisted resin transfermoulding for high performance carbon fibre composites. In 4th

International Conference on Automated Composites, The AlbertHall, Nottingham, 1995, pp. 299–306.

15. Williams, C., Summerscales, J. and Grove, S., Resin infusion underflexible tooling (RIFT): a review. Composites, Part A, 1996, 27A,517–524.

16. Hayward, J. S. and Harris, B., Effect of process variables on thequality of RTM mouldings. SAMPE Journal, 1990, 26(3), 39–46.

17. Akay, M., Cracknell, J. G. and Farnham, H. A., Detection of under-cure in carbon fibre/epoxy composites by dynamic cure analysis.Polymers and Polymer Composites, 1994, 2(5), 317–322.

18. O’ Neal, H. R., Welch, S., Rogers, J., Guilford, S., Curran, G. and

Menard, K. P., Comparison of T g values for a graphite epoxy com-posite by differential scanning Calorimetry (DSC), thermo-mechan-ical analysis (TMA) and dynamic mechanical analysis (DMA),

Journal of Advanced Materials, 1995, 49–54.19. Curtis, P.T. (ed.), Test Methods for the Measurement of Engineer-

ing Properties of Fibre Reinforced Plastics, Royal Aerospace Estab-lishment, Composites Research Advisory Group (CRAG),Technical Report 88012, Issue 3 (Final), 16th February 1988.

801


1-s2.0-S1359835X98000554-main

Documents