M A S T E R ' S T H E S I S
Effects of fibre bundle size and stitch pattern on the static properties of unidirectional carbon-fibre non-crimp fabric composites
Marc Wouters
Luleå University of Technology
ghts reserved. No part of this publication may be reproduced and/or published by print, photoprint, microfilm or any other means without the ious written consent of SICOMP AB. In case this report was drafted on instructions, the rights and obligations are subject to the relevant agreement luded between the contracting parties. Submitting the report for inspection to parties who have a direct interest is permitted. © 2005 SICOMP AB.
Effects of fibre bundle size and stitch pattern on the static properties of unidirectional carbon-
fibre non-crimp fabric composites
Marc Wouters EEIGM Master Thesis
SICOMP AB, Box 104, SE-431 22 Mölndal, Sweden
Abstract
The use of Non-Crimp Fabrics (NCFs) as reinforcement in composites is relatively new, but the use of this type of materials is expected to increase a lot in the coming years. NCF-based composites can for example replace the very costly use of prepreg tapes, but before this will occur, the effect of stitching parameters, etc., on the performance of NCF-based composites must be fully understood.
In the present report, the variation in static (tension and compression) properties of unidirectional carbon-fibre NCF composites was studied as a function of stitch pattern and fibre bundle size. By studying the effects of stitch pattern and bundle size on the static properties, the aim was to be able to determine the most important stitch parameters to obtain good mechanical properties.
The results from the tests carried out show no big difference in properties between the NCF- based composites examined. Further investigations are therefore needed, with other types of fabrics (particularly different bundle size), before any real conclusions can be made regarding the effect of stitch pattern and bundle size on the mechanical properties of NCF-based composites.
Keywords: Non-crimp fabric composites, Carbon fibre, Unidirectional, Stitching,
Contents
1. INTRODUCTION ...5
2. NON-CRIMP FABRICS (NCF) ...6
2.1. FABRIC CONSTRUCTION...7
2.2. ADVANTAGES OF NCF ...10
2.3. COMPOSITE MANUFACTURING...15
2.3.1. Permeability...16
2.3.2. Drapeability (draping property) ...16
2.3.3. Compressibility ...17
2.3.4. Fabrication-induced defects ...18
2.3.5. Resin Transfer Moulding (RTM) ...18
2.3.6. Component manufacturing speed...20
2.4. MECHANICAL PROPERTIES OF NCF LAMINATES...21
2.4.1. Tension ...21
2.4.2. Compression...21
2.4.3. Impact ...22
2.5. COMPARISON OF WOVEN AND STITCHED 0/90 FABRICS...22
2.5.1. Mechanical properties ...24
2.5.2. Comparison of Drapeability Behaviour...24
2.5.3. Cost Comparisons ...25
2.5.4. Qualifications for Component Production ...25
2.6. BRIEF COMPARISON BETWEEN NCFS AND UDPT ...25
2.7. APPLICATIONS...26
3. EXPERIMENTAL...28
3.1. MATERIAL...28
3.1.1. Non-Crimp Fabrics ...28
3.1.2. Resin...29
3.2. MANUFACTURING OF PLATES...29
3.3. TENSILE TESTS...31
3.3.1. Test specimens: ...31
3.3.2. Test procedure ...32
3.3.2.1. Measurement of the stiffness and Poisson’s ratio ... 32
3.3.2.2. Measurement of ultimate strength ... 32
3.3.3. Calculations ...33
3.3.3.1. Tensile strength... 33
3.3.3.2. Tensile Stiffness ... 33
3.3.3.3. Poisson’s ratio ... 34
3.3.3.4. Ultimate tensile strain ... 34
3.4. COMPRESSION TESTS...34
3.4.1. Test specimens...35
3.4.2. Test procedure ...36
3.4.3. Calculations ...36
3.4.3.1. Compression strength ... 36
3.4.3.2. Compression stiffness ... 37
3.4.3.3. Ultimate compressive strain... 37
3.5. INFLUENCE OF THE STITCHING ON THE BUNDLE SHAPE...37
3.5.1. Material studied ...37
3.5.2. Equipments and study procedures...38
4. RESULTS AND DISCUSSION ...40
4.1. TENSILE TESTS RESULTS...40
4.1.1. The P0 material...42
4.1.2. Tensile strength ...42
4.1.3. Tensile stiffness...44
4.1.4. Poisson’s ratio...45
4.1.5. Ultimate tensile strain ...46
4.2. COMPRESSION TESTS RESULTS...47
4.2.1. Comparison of the P0 and P4 materials ...48
4.2.2. Compressive strength ...48
4.2.3. Compressive stiffness ...49
4.2.4. Ultimate compressive strain...50
4.3. INFLUENCE OF THE STITCHING ON THE BUNDLE SHAPE...51
5. CONCLUSIONS...56
ACKNOWLEDGEMENTS ...58
REFERENCES ...60
APPENDIX: ...62
1. Introduction
The traditional way to produce composite parts with high specific stiffness and strength has usually been to use prepreg tapes. The use of prepreg has, however, been mostly restricted to high-tech applications since the material and processing is very expensive. Today, when cost is of great concern even within the aircraft and aerospace industry, there is a large interest for development of cheaper materials and manufacturing techniques for high-tech applications.
An alternative way to manufacture composites with continuous fibres has emerged during the last couple of years. The orientation of the reinforcement is set already in the textile manufacturing stage by stitching several layers of unidirectional plies of fibres. The fabrics produced in this way are usually called Non-Crimp Fabrics (NCF), since their fibres are more or less without crimps. For this reason, this type of fabric present better properties than woven fabrics, since the fibres are straight and effective from the beginning of loading. Furthermore, as it is possible to assemble several layers of unidirectional fibres at the fabric manufacturing stage, the lay-up of NCF reinforcements is a lot quicker than the lay-up of prepreg tapes.
But what about the mechanical properties of NCF composites? To be of interest for high-tech
applications, these materials must present high quality with low variation in properties, and
more knowledge is therefore needed about NCF composites before they can be used with
confidence within e.g. aerospace. The objective of this study was to investigate how different
fibre bundle size and different stitch parameters affect the static properties of unidirectional
carbon-fibre NCF composites.
2. Non-Crimp Fabrics (NCF)
One of the reasons to add fibres to plastics is to increase the mechanical properties. As it is a big variation in properties among different types of fibres, the properties of fibre-reinforced plastics varies a lot. Individual fibres or fibre bundles can only be used in very few manufacturing processes. For most applications the fibres need to be arranged in some way to facilitate their handling. In polymeric composite terms, a fabric is defined as a manufactured assembly of long fibres in the form of a flat sheet with one or more layers of fibres. These layers are held together either by mechanical interlocking of the fibres themselves or by a secondary material that binds the fibres together and hold them in place. In the case with secondary materials, some manufacturers use an adhesive as secondary material, but it is far more common to find that the fibres are stitched together, since this maintains the drapeability of the fabric [1]. The stitched fabrics can be manufactured utilising almost any type of reinforcing fibre. The most common types of stitched fibres are based on glass, carbon or aramid fibres. The different ways for putting fibres together into sheets, and the variety of fibre orientations possible, leads to a huge variety of fabrics, each type having its own characteristics.
In recent years the interest for using multiaxial fabrics in design of composite components has increased. Stitch fabrics, produced by “advanced warp knitting technology”
are often called "Directionally Orientated Structure Fabrics (DOS Fabrics) or "Non-Crimp- Fabric (NCF)" because there are no crimp on the inserting yarns [8]. The NCF fabrics consist of one or more layers of long fibres held in place by a secondary non-structural thread. The structural fibres can be glass, carbon or aramid fibres, or a combination of these fibres. The stitching process allows a variety of fibre orientations, beyond the simple 0/90 woven fabrics, to be combined into different types of fabrics. Considering unidirectional fabrics, these fabrics offer oriented strengths, something that often is preferred in high performance applications.
However, multiple orientations are sometimes beneficial to obtain a quasi-isotropic reinforcement.
The mechanical properties of NCF composites are controlled by the type, amount, and
orientation of the fibre being used. There are today many different types of fibres
commercially available to meet the design requirements needed. The ability to tailor the fibre
architecture allows for optimised performance, which translates into weight and cost savings.
2.1. Fabric construction
Stitched fabrics, also known as non-woven, non-crimped, or knitted fabrics have good and repeatable mechanical properties that are predictable because of the fibre architecture.
Stitched fabrics are produced by putting together layers of aligned fibres, see Figure 1.
Typically, the fibre orientations are; 0º (warp), 90º (weft or fill), and ±45º (bias) [2]. Other orientations can be manufactured on order, but as they necessitate major machine changes they are not especially common [1].
Figure 1. Stitched triaxial and quadraxial fabrics.
The stitching is carried out on special machines that are based on the knitting process, see Figure 2. Each machine varies in the precision by which the fibres are laid down, particularly with reference to keeping the fibres parallel [1].
Figure 2. LIBA copcentra MAX 3 CNC + chopper [3].
The machines have a frame that draws in the fibres for each axis/layer until the required
layers have been put together, and then stitches the layers together as shown in Figure 3.
Figure 3. Reinforcement example.
The 0º fibres in a stitched fabric (fibres running in the direction of the roll, or warp direction) are fed from spools located on stationery "creel" racks behind or above the machine. Cross plies are fed by carriages that are moving back and forth across the width of the roll (or weft direction). The stitching thread is usually polyester due to its combination of good properties (for binding the fabric together) and low cost. The width of the fabric is set by the width of the conveyor table (distance between hooks) [3].
The structure of the fabric have a great importance since the fabric must be stable enough to be handled, cut and transported to the mould, but drapeable enough to conform to the mould shape and contours. Apart from the properties of the fibre being used, the characteristics of NCF fabrics are determined by two main factors [1]:
(i) the combination of Tex and thread count of the roving, and
(ii) the method of stitching (particularly the stitch gauge, stitch length and stitch style)
The stitch gauge (the frequency of stitches across the fabric width) is usually 5 or 6 stitches
per inch (gauge 5 and 6) or 10 or 12 per inch (gauge 10 and 12). The lower number of stitches
(lower gauge number) gives lower stability and allows the fabric to be more drapeable. The
stitch length (distance between two stitches in the roll direction or 0 °axis) can vary by several
millimetres. A smaller stitch length will result in a fabric with better drape qualities, but the
slower production rate result in higher cost. The stitch style (stitch pattern) also affects the
handling properties of the fabric. A linear (or chain) stitching gives good drapeability, while a
zig-zag (or tricot) stitch makes the fabric more stable. A description of stitch gauge, stitch
length and stitch style is presented in Figure 4. Some stitching styles can be highly elastic,
which leads to extremely high drape of the fabric. Tricot stitch is necessary in fabrics that
contains 0° layer. Often a combination of tricot and linear stitch (modified tricot) is used to
get the optimum combination of drape and stability.
Figure 4. Stitching parameters.
The fabric is penetrated by a stitch bar that contains up to 700 needles with polyester stitch yarn (in some cases glass stitch yarn). In simple terms, the stitch density can be modified to suit the application the fabric shall be used for. Fabrics that needs substantial handling before laminate manufacture needs greater integrity and will therefore have a high stitch density, while fabrics that needs to conform to complex three dimensional shapes will have low stitch density.
Once the fabric is stitched, it is cut to specified lengths and taken up on rolls that are ready for packaging and shipment. The end-use manufacturing process determines the size and length of the rolls. Continuous in-line processing, like pultrusion, needs long material runs and the fabrics are therefore shipped on large "transbatchers" that are weighing several thousands of pounds instead of the standard 200 pound roll [6].
NCFs have two main advantages. Firstly, as no crimp is induced in the process, see Figure 5, (in fact a small amount of crimp is always present) the in-plane mechanical properties are improved compared to woven fabrics. Secondly, the stitching yarns also provide some out-of-plane reinforcement, which translates into better through thickness properties and impact performance [5]. However, polyester fibres do not bond very well to some resin systems and the stitching can therefore act as failure initiator. Sometimes “high performance” fibres such as glass or aramid can be used as stitching thread to improve the interfacial strength between the stitching thread and the matrix.
Figure 5. Through thickness stitching.
2 Stitch length
Stitch Style
Stitch Gauge
While the most common is to have the same amount of fibre in each axis, “unbalanced”
fabrics with different weights per layer can also be produced. Unbalanced quadraxial fabrics, containing twice the weight in the 0º-and 90º axes, are becoming very popular among boat builders. One of the benefits with quadraxial fabrics is labour savings through replacement of individual 0º and 90º layers with one single layer of quadraxial fabric. Another common
“unbalanced” fabric is the warp triaxial (+45°, 0°, -45°) fabric in which the 0° layer contains 50% of the fibres and the remaining 50% is spread equally between the ±45 layers.
It is also possible to mix different fibre types, either between different layers or inside individual layers, to form hybrid fabrics. For example, when low weight and good impact tolerance is a priority, a combination of aramid and E-glass can provide tough and abrasion resistant laminates.
Another important criterion for the fibre fabric is that it shall allow for optimum resin flow when composites are manufactured. Properly designed, the fabric shall allow for quick wet out and wet through without wetting agents, and shall stay in place once the resin is applied. For example, extremely large weight fabrics (enabling large quantities of fibre to be incorporated rapidly) can be difficult to impregnate without some automated process.
However, NCFs generally enhance resin flow far more than chopped strand mats and woven roving, and the stitching helps the resin to migrate through the layers.
2.2. Advantages of NCF
The two key improvements with stitched multiaxial fabrics compared to woven fabrics are:
(i) Better mechanical properties, primarily from the fact that the fibres are always straight and non-crimped. In the case of woven fabrics, the fibres embedded in resin will straighten during loading/impact, which causes resin cracks. When load is transmitted along the many kinks in woven roving a concentration of stress occurs due to the shearing between fibre and resin. If repeated loading and unloading takes place this will cause a more rapid breakdown of the laminate (fatigue). Furthermore, in the case of NCF, more orientations of fibres are available from the increased number of layers of fabric.
(ii) Improved component build-up speed, based on the fact that fabrics can be made thicker
and with multiple fibre orientations so that less layers are needed.
The properties of non-crimp fabric-based composites are determined by:
i) Fibre properties. The mechanical properties of most fibre-reinforced resins are considerably higher than those of un-reinforced resins. The mechanical properties of fibre/resin composites are therefore dominated by the contribution of the fibre to the composite. In Table 1 some average properties for fibres currently used in composites are presented.
Table 1. Typical fibre properties.
In this study, the fibre used was carbon fibre. The carbon fibre is produced by controlled oxidation, carbonisation and graphitisation of carbon-rich organic precursors, which are already in fibre form. The most common precursor is polyacrylonitrile (PAN), because it gives fibres with good strength and modulus values up to 550-600 GPa [1]. They also offer excellent compression strength for structural applications. Pitch fibres are made from petroleum or coal tar pitch. Pitch fibres present extremely high modulus values (up to 970 GPa) and their favourable coefficient of thermal expansion make them interesting for space/satellite applications.
Carbon fibres are usually grouped according to the modulus band in which their
properties fall. These bands are commonly referred to as: high strength (HS), intermediate
modulus (IM), high modulus (HM) and ultra high modulus (UHM), see Table 2 [1]. These
different grades of fibres are obtained according to variations in the graphitisation process
(high strength fibres at ~ 2600 °C and high modulus fibres at ~ 3000°C, with other types in
between).
Table 2. Commercial PAN-based carbon fibre properties.
Carbon fibres are more expensive than glass fibres, but offer an excellent combination of
properties; low weight (they have the highest specific stiffness of any commercially available
fibres), high strength (in both tension and compression), high resistance to corrosion, high
resistance to creep, and high resistance to fatigue. Carbon fibres are more brittle (less
elongation to break) than glass or aramid fibres. Their tensile strength is equal to glass while
their modulus is about three to four times higher than glass. Their impact strength, however, is
lower than for both glass and aramid, with particularly brittle characteristics for HM and
UHM fibres, see Appendix A. Moreover, carbon fibres can cause galvanic corrosion when
used next to metals. A barrier material such as glass and resin is therefore often used to
Carbon fibres are supplied in a number of different forms, from continuous filament tows to woven and knitted fabrics. Usually tows of continuous carbon filament contain 1000 to 75.000 individual filaments, while the filament diameter of most types is about 5-7 µm [1].
The fibre price is usually considerably higher for small fibre diameters. The fibre diameter can impact the physical, mechanical and aesthetic qualities of carbon fibre reinforced laminates. Using fibres with smaller diameters mean higher fibre content and less resin. This improves strength and can reduce weight. There can also be a cost saving realised through less resin. In terms of fabrics, smaller diameters translate into better drape, improved mechanical properties and superior surface finish (less print through) [11].
ii) Resin properties [1, 2]. The resins in thermoset composites are very important for the properties and process characteristics. The primary functions of the resin are to transfer stress between the reinforcing fibres, act as a glue to hold the fibres together, and protect the fibres from mechanical and environmental damage. Thermoset resins are usually liquids or low melting point solids in their initial form. When used to produce finished goods, these thermosetting resins are “cured” by the use of a catalyst, heat or a combination of heat and catalyst. Unlike thermoplastic resins, cured thermosets will not melt and flow but will soften when heated, and once formed they cannot be reshaped. The most common thermosetting resins used in the composites industry are unsaturated polyesters, epoxies, vinyl esters and phenolics. In this study, the resin used was epoxy.
Epoxy resins have a well-established record in a wide range of composite parts, structures and concrete repair. The structure of the resin can be engineered to yield a number of different products with varying levels of performance. A major benefit of epoxy resins over unsaturated polyester resins is their lower shrinkage. Epoxy resins can also be formulated with different materials or blended with other epoxy resins to achieve specific performance features. Cure rates can be controlled to match process requirements through the proper selection of hardeners and/or catalyst systems. Generally, epoxies are cured by addition of an anhydride or an amine hardener as a 2-part system. Different hardeners, as well as quantity of a hardener, produce different cure profiles and give different properties to the finished composite.
Epoxies are primarily used for fabricating high performance composites with good mechanical properties, resistance to corrosive liquids and environments, excellent electrical properties, good performance at elevated temperatures, and good adhesion. However, epoxy resins do not have particularly good UV resistance.
Epoxy resins are used with a number of fibre reinforcements, including glass, carbon and
aramid. Epoxies are compatible with most composite manufacturing processes.
iii) Fibre/resin interface. The interfacial properties between fibres and resins are controlled by the degree of bonding that exists between the two. This is heavily influenced by the treatment given to the fibre surface (sizing).
In our case the carbon fibre used had a surface treatment (sizing) applied to improve matrix bonding and to protect it during handling. Finishes, or sizes, for carbon fibres used in structural composites are generally epoxy based, with varying levels being used depending on the end use of the fibre. For weaving, the size level is about 1-2% by weight, whereas for tape prepregging or filament winding (or similar single-fibre processes), the size level is about 0.5- 1%. The chemistry and level of the size are important not only for protection and matrix compatibility but also because they affect the degree of spread of the fibres. Fibres can also be supplied unsized but these fibres will be prone to break during handling. Most carbon fibre suppliers offer 3-4 levels of sizings for each grade of fibre [1].
iv) Fibre volume fraction. The ratio of the fibre to resin derives largely from the manufacturing process used, but is also influenced by the type of resin system used, and the form in which the fibres are incorporated. In general, since the mechanical properties of the fibres are much higher than those of resins, the higher the fibre volume fraction the higher the mechanical properties of the composite. In practice there are, however, limits since the fibres need to be fully wetted by the resin to be effective, and there will be an optimum packing of fibres. In addition, the manufacturing process used leads to varying amounts of imperfections and air inclusions. Typically, with a common hand lay-up process, widely used in the boat- building industry, the upper limit for fibre volume fraction is approximately 30-40%. For higher quality, more sophisticated and precise processes are used, and the fibre volume fraction can then approach 70%.
v) Geometry and orientation. The geometry of the fibres in a composite is important since the fibres have their highest mechanical properties along their lengths. This leads to highly anisotropic (direction specific) properties of fibre-reinforced composites. This means that it is very important when considering the use of composites to understand both the magnitude and the direction of the applied loads. When correctly accounted for, these anisotropic properties can be very advantageous since it is only necessary to put material where loads will be applied, and thus redundant material or parasitic material that is put in orientations where there is little or no load is avoided. Fabrics allow for this precise placement of the reinforcement, whereas this cannot be obtained with milled fibres or chopped strand mats.
Due to the continuous nature of the fibres in most fabrics, the strength to weight ratio is much
higher than for cut or chopped fibre versions. Stitched fabrics allow for customised fibre
orientations within the fabric structure. This can be of great advantage when designing for
components involved in the composite. This type of construction allows for load sharing between fibres so that higher mechanical properties are typically obtained.
Some characteristic properties of carbon fibre NCFs are presented in Table 3.
Table 3. Estimated laminate properties of Vectorply’s NCF products.
2.3. Composite manufacturing
The parameters associated with composite manufacturing are, of course, dependant on the
type of infiltration process used. For instance, the compaction forces experienced by the fibre
stack inside the cavity of an RTM (Resin Transfer Moulding) mould might be quite different
from the compaction forces in a vacuum bag during a RFI (Resin Film Infusion) or RIFT
(Resin Infusion Flexible Tooling) process. It is expected that the different nature of these
compaction forces will influence the level of nesting and crimp of the NCF yarns as well as
the final cross-sectional shape of the tows. The resin flow during impregnation is also very
different in the case of RTM (mainly in-plane) when compared to RFI and RIFT (dominant
through-thickness flow). The variables that control the resin flow, such as the resin injection
pressure, the vacuum pressure, the resin viscosity, and the local distribution of fibres
influences the flow pattern, which in turn affects the formation of voids and other defects such
as fibre wash.
When comparing the meso-structure of NCF-based laminates manufactured by RFI with NCF laminates manufactured by prepreg there is a considerable difference. In the case with prepreg the tows appear to spread more evenly due to the rolling. The consequence of this is that the meso-structure of NCF prepreg laminates are relatively similar to laminates made by prepreg-tape, i.e. low crimp and uniform fibre distribution. The tows tend to retain their shape during infusion and there is therefore more crimp, tow nesting and resin rich areas in RTM and RFI manufactured laminates. The differences in meso-structure translate into differences in failure modes and impact performance.
The processability of a particular fabric is usually defined in terms of three characteristics. Firstly, the ability of the fabric to be impregnated by a low-viscosity resin during infusion (permeability). Secondly, the capacity of the fabric to adopt to the shape of the mould or tool without excessive and unacceptable wrinkling (drapeability) and, thirdly, the ability of the fabric to be compacted to the designed fibre volume fraction (compressibility) [5].
2.3.1. Permeability
The literature on this subject is divided into either experimental papers that characterise the fabric permeability by means of more or less standard impregnation tests or analytical papers that try to model the fabric permeability taking into account the architecture of the fabric unit cell. In both cases, permeability is normally characterised only in-plane, probably because this is the expected flow configuration in RTM, but very little attention has been paid to the through-the-thickness permeability (not only in NCF but also in other fabrics), which most often is the dominant flow direction in RFI and RIFT.
Although a few papers report measurements of permeability of some NCFs, there are no comprehensive experimental studies on the effect of NCF architecture on the through-the- thickness permeability. The effects of the inter-tow gap, which offers a preferential path for resin flow, and of the stitch tension, which might be responsible for very low permeability areas within the reinforcement, are expected to greatly influence the impregnation behaviour of NCFs. As the competition between the micro-flow across the fibre bundles and the meso- flow through the inter-tow channels controls air entrapment and void formation, control of the fabric architecture is essential to ensure final composite quality [5].
2.3.2. Drapeability (draping property)
The style and density of the stitching pattern, and the quantity of non-structural stitch fibres
drapeability is for instance required than for the flat roof of a bus. Modelling tools are available for predicting of the fibre orientation after that a woven fabric has been deformed over the surface of a tool to with a curved shape. Most of these tools are based on the fact that the dominant mode of deformation is shear-rotation of the fibre bundles around their crossover points. Simple algorithm-based tools can both predict the final fibre orientation and local variations in permeability. Although these models work reasonably well with most of the simpler woven structures, they do not work satisfactory when applied to NCF structures.
There is experimental evidence of complicated mechanisms involved in the deformation process of dry NCFs, i.e. tow spreading and buckling. Some attempts have been made to simulate the complex in-plane deformation of NCFs, by combining traditional pure kinematics with energy minimisation algorithms. Although the results from these simulations are promising, the basic assumptions of these models are quite restrictive and some refinement is still required (the effect of stitching is only taken into account indirectly).
Nevertheless, the current alternative to these models is FE analysis of very complex geometry’s with highly non-linear mechanical properties, which generally is very computationally time-consuming [5].
2.3.3. Compressibility
Preform compaction during tool closure is a key issue in liquid composite moulding. In fact, the final preform microstructure depends to a certain extent on its compaction to the designed thickness. The mechanisms that govern the compaction behaviour of fibre arrays are very complex and include fibre bending, tow slipping and friction, tow nesting and tow flattening, among others. As a result of the interaction between these different mechanisms, the overall response of dry fibrous preforms when subject to compaction pressure is highly non-linear, with an apparent compressive modulus that increases with volume fraction.
A common approach to simulate non-linear compaction is by means of regression of compaction experiments. The empirical models thus generated relate normally fibre volume fraction (or porosity) with applied compaction pressure, usually with a power-law form that includes several parameters with no clear physical meaning, which have to be determined experimentally. However, none of these models account for the effect of the knitting thread in NCF structures.
A detrimental effect in preform compaction of warp-knitted and stitched structures is the
buckling of the knitting/stitching thread, which usually translates into a loss of effectiveness
of the through-the-thickness reinforcement. This effect is more pronounced in loosely knitted
structures than in tightly stitched fabrics [5].
2.3.4. Fabrication-induced defects
Whereas conventional laminated composites based on prepreg tapes possess a relatively homogeneous structure and consequently a small amount of fabrication-induced defects (mainly voids and resin microcracks) this may not always be the case for NCF-based composites. The stitching process, unless carefully controlled, can squeeze the fibres together (particularly in the 0° direction) creating resin-rich areas in the laminate, which usually constitute areas prone to cracking due to the combined effect of residual stresses and stress concentrations. Furthermore, the impregnation of heterogeneous fibre regions is likely to generate complex flow fronts that may develop voids inside or outside the tows.
The effect of stitching on the void formation during RFI manufacturing has been studied using ultrasonic C-scanning and optical microscopy. The void content was considerably higher in stitched regions due to the heterogeneity and anisotropy of these regions. The flow pattern appeared to be very complex in the stitched regions. The first flow was through-the- thickness along the stitch-direction followed by in-plane impregnation of the fibre layers at a rate that depended on the fibre orientation. Voids ranging from 100 µm to 500 µm were concentrated in the stitched regions, and although the voids could have been formed due to poor degassing, they were attributed to low consolidation pressure in the stitched regions. The void content varied with type of resin, but was always higher in stitched regions than in un- stitched regions [5].
2.3.5. Resin Transfer Moulding (RTM)
Resin Transfer Moulding (RTM) is a “Closed Mould Process” in which fabrics are laid up as dry material stacks between two matching mould surfaces – one male and one female. These fabrics are sometimes pre-pressed to fit the shape of the mould, and held together by a binder.
These fabric preforms are then more easily laid into the mould. The matching mould set is
then closed and clamped and a low-viscosity thermoset resin is injected under moderate
pressures (3 to 7 atm typically) into the mould cavity through one port or a series of inlets in
the mould. A schematic of the RTM process is presented in Figure 6. The reinforcements
commonly used may include a variety of fibre types, in various forms such as continuous
fibres, mat or woven constructions, as well as hybrids that consists of more than one fibre
type. Vacuum can also be applied to the mould cavity to reduce void formation. This is known
as Vacuum Assisted Resin Injection (VARI). Once all the fabric is wet out, the resin inlets are
closed, and the laminate is allowed to cure. Both injection and cure can take place at either
ambient or elevated temperature, but the injected composite part is typically cured under elevated temperature.
Figure 6. RTM principle.
The main advantages with RTM are [1,2]:
i) Laminates with high fibre volumes can be obtained with very low void contents.
ii) Good health, safety, and environmental control due to a closed process.
iii) Possible labour reductions. This process can produce parts faster, as much as 5-20 times faster, than open mould techniques.
iv) Both sides of the component have a moulded surface. The mould surface can produce a high quality finish (like those on an automobile).
v) The RTM process produces tighter dimensional tolerances (± 0,1 mm).
vi) Complex mould shapes can be achieved. Cabling and other fittings can be incorporated into the mould designs.
And the main disadvantages are [1, 2]:
i) Matched tooling is expensive and heavy in order to withstand pressures. High production volumes are required to divide the high tooling costs on more parts.
ii) The size of the parts is limited by the mould size.
iii) Unimpregnated volumes can occur resulting in very expensive scrap parts. Thus,
reinforcement materials are limited due to the flow and resin saturation of the
fibres.
2.3.6. Component manufacturing speed
The current long fibre reinforced composite manufacturing techniques are often too expensive to compete with conventional materials in series production. The basis of a composite component production process is the raw materials (fibre and resin). During the manufacturing process, the textile fabric is cut, preformed, folded or sewn to make it fit into the mould. Then the fabric is placed into the mould and the resin injected. After cooling (thermoplastic matrix) or curing (thermoset resin), the produced part has to be machined to obtain the final component shape. Non-destructive testing (e.g. ultrasonic) is often required as a final task. Many manual procedures are therefore necessary to achieve the final composite component.
By using NCF, the cost effective solution begins with engineering the laminate requirements at the fabric manufacture stage. Previously, the composites industry has tried various plastics technologies such as tape layering or filament winding. These plastics techniques have, however, limitations, and the composites industry has therefore turned to textile techniques for further progress. By creating pre-made pre-designed textile fabrics with the required fibre orientation the many steps in the assembly of parts can be reduced. Below is an example of cost saving presented by DUROPLASTIC (South Africa):
Take an example of a standard laminate of 1x300 CSM (Chopped Strand Mat) + 1x450 WR (Woven Roving) + 1x450 CSM and one with Quadriaxial 1050 g/m².
Standard:
Fabric: 750 g CSM at 9.00 SEK/kg = 6.75 SEK 450 g WR at 9.50 SEK/kg = 4.30 SEK Resin: 2.4 kg at 5.00 SEK/kg = 12.00 SEK
Labour: 3 Layers at 15min/layer at 25 SEK/hr incl O/Heads = 18.75 SEK Total: 41.80 SEK
Quadriaxial:
Fabric: 1050 g at 23.50 SEK/kg = 24.67 SEK Resin: 1 kg at 5.00 SEK/kg = 5.00 SEK
Labour: 1 layer at 15 min/layer at 25 SEK/hr incl O/heads = 6.25 SEK Total: 35.92 SEK
This shows a saving of nearly 6.00 SEK/m
2, due to the fact that fewer layers require less lay-
up time. The second most important saving is in weight. The standard weight was 3.6 kg
while the quadriaxial weight was approximately 2 kg, i.e. nearly half the weight, and the tensile strength of the quadriaxial was almost 30 % higher than the standard laminate [4].
However, the benefits of this new class of materials can only be achieved if the total process chain of material selection, i.e. fabric customisation, part design, manufacturing, and maintenance is optimised. Furthermore, much of the design work related to tape winding or lay-up is shifted to the stage of fabric construction. The fabric production process can be slow and the cost of the machinery high. This, together with the fact that good quality stitched fabrics (low Tex fibres are required to get good surface coverage for the low weight fabrics) are expensive, explains why the cost of NCFs can be relatively high compared to woven fabrics.
2.4. Mechanical properties of NCF laminates
2.4.1. Tension
The response of a NCF laminate to tensile loading is very dependent on the tensile stiffness and strength of the reinforcing fibres, since the stiffness and strength of the fibres are much higher than the stiffness and strength of the resin. In the case of NCF, because the fibres are set straight within the matrix, when tension loading is applied the fibres can go into tension without working against the matrix, which results in higher tensile strength than for woven fabrics and chopped strand mats. The combination of straight fibres and a high fibre-to-resin ratio provides NCFs with very good tensile properties [4].
2.4.2. Compression
In studies on the compressive behaviour of NCFs, the approach has been to study the relationship between microstructural parameters (geometrical and material) and the macroscopic compressive strength. First of all, the stitching across the blanket thickness seems not to affect the compressive strength. The results of most studies show that as in classical laminates the defect maximum angle and the resin characteristics are the most important features, which define the compressive strength [12]. The adhesive and stiffness properties of the resin system are crucial, as it is the possibility of the resin to maintain the fibres straight and prevent them from buckling that determines the compressive properties.
Furthermore, the use of NCF permit to avoid the main disadvantage of woven fabrics which
have a degree of crimp in the fibre yarns, which translates into poor in-plane properties
(particularly in compression).
2.4.3. Impact
The impact and compression after impact properties of NCFs are better than for woven fabrics due to the use of straight fibres. The performance of NCFs is further improved by the quasi- isotropic nature of quadriaxial fabrics, having only a 45-degree property translation between each layer of fibres [4].
2.5. Comparison of woven and stitched 0/90 fabrics
Woven fabrics are manufactured with fibres laid at right angles to each other. The two sets of interlaced continuous fibres are crimped and not straight. They are fabricated on looms in a variety of weights, weaves, and widths. In a plain weave, each fill yarn or roving alternately crosses over and under each warp yarn allowing the fabric to be more drapeable and conform to curved surfaces.
However, 0/90 fabrics can also be made by stitching, see Figure 7. This process effectively combines two layers of unidirectional fibres into one fabric. Running at much higher speeds than the full multiaxial machine, a simpler variant of this machine is used to produce these fabrics, which provides a cost-effective alternative to 0°/90° woven fabrics.
Figure 7. Aspect of a [0/90] NCF (glass fibres).
Considering the fibre orientation, 0/90 NCFs is comparable to 0/90 woven fabrics, see Figure 8. However, with 0/90 NCFs there is the advantage that the yarns are directly oriented and lie absolutely straight. Thus, 0/90 NCFs can offer mechanical performance improvements of up to 20% in some properties compared to 0/90 woven fabrics [1]. The reason to this is explained by:
1. Parallel non-crimp fibres bear the load immediately upon being loaded.
2. Stress points found at the intersection of warp and weft fibres in woven fabrics are eliminated.
3. A higher density of fibres can be packed into a laminate with NCFs compared to woven fabrics. In this respect the fabric behaves more like layers of unidirectional fibres.
(a) (b)
Figure 8. (a) Woven fabric structure, and (b) NCF structure.
Other benefits of NCFs compared to woven fabrics are that heavy fabrics can be easily produced with more than 1kg/m² of fibres, and the better packing can reduce the amount of resin required [6].
It is in the heavy tow size that NCF competes best with woven fabrics. As woven fabrics gain in weight, so does the amount of crimp, and the strength therefore reduce on a comparative weight basis. Heavy weight non-crimp fabrics do not lose in strength but grow in fibre density when compared to woven fabrics.
At present woven fabrics and multi-axial multiply fabrics are the main techniques in
competition for composite components. The two fabrics can be compared with regards to
mechanical properties, drapeability behaviour, production costs and individual industry
applications.
2.5.1. Mechanical properties
Non-impregnated Fabrics
Some investigations have been made on dry, non-impregnated textile fabrics, particularly in relation to qualifying fibre and filament damage, which would adversely affect the mechanical properties of the composite component [3]. Tensile testing carried out on non-impregnated structures demonstrates that the multi-axial multiply fabric offers much better properties than woven fabrics.
The tensile strength was around 20 per cent higher and the Young’s modulus improved by about 15 per cent. The results show that almost no filament damage occurs in the multi-axial multiply fabric due to the penetration of needles during the knitting process depending on the fabric construction.
Laminates
Investigations of laminated textiles also resulted in improved performance in favour of multi- axial multiply fabric structures. In this case the tensile strength demonstrated better values (over 10 %) and stiffness (over 20 %) due to the straight reinforcement fibres. Even the interlaminar properties of these structures can be increased by around 10 %, by using a high performance yarn for the loop system, since they are based on true three-dimensional reinforcement.
2.5.2. Comparison of Drapeability Behaviour
Since the drapeability is an important parameter in production, pre-examination of the deformation behaviour of reinforcement textiles to avoid wrinkling is important. An undefined covering of the component can lead to inferior mechanical laminate properties. It is therefore necessary to qualify the deformation behaviour of different textile structures in order to estimate their suitability for specified component shapes. There are no known methods of drapeability testing which take into account the special properties of high performance fibres.
For bi-axial reinforcement structures a test method based on the shear principle was used.
The fabric was secured in a square frame and deformed in a conventional drawing machine
until wrinkling occurred.
In contrast to woven fabrics, the drapeability of multi-axial multiply fabrics can be adjusted via the loop system independently of the ply construction. Deformation can be assisted by larger loop sizes and slacker density of the knitted values. This offers higher drapeability values than those for woven fibres. Multi-axial reinforcement textiles were also subjected to another testing method. This was designed with four clamps for each fibre sheet of the multi-axial fabric. These clamps were then moved in a contra-directional manner to achieve a shearing deformation.
2.5.3. Cost Comparisons
The economic view shows that biaxial layer fabrics (0/90) cannot be produced less expensive than woven fabrics. The high investment costs of this technology do not allow cheaper fabrics. In addition, higher costs for staff must be calculated due to the long preparation and set-up times. But the results for multiaxial layer fabrics demonstrate that they can be produced less expensive than woven fabrics if the full potential of this technology is used. For example, the break-even point in a textile manufacturing cost calculation for a special 4-layer fabric of 1200 g/m² is reached when production volume reaches 1400 metres of fabric length (in dependence of special production parameters).
When component processing production costs are taken into consideration, the use of multi-axial multiply products also shows cost advantage. Since multi-axial fibre sheets are pre-designed, the waste factor is much lower.
2.5.4. Qualifications for Component Production
The use of multiaxial multiply fabrics can lead to more rationalised composite component production compared with woven fabrics. This is being proved in current applications such as city buses, ferry boats and windmill rotor blades. Promising future applications in the automotive field lie in floor panels, crash component parts and body shells. Within the building industry there are prospects for earthquake resistant concrete reinforcement.
2.6. Brief comparison between NCFs and UDPT
Compared with unidirectional prepreg tape (UDPT), the recent development of non-crimp
fabrics (NCF) has brought together the advantages of textile technology (high deposition
Experimental results and economic considerations show that non-crimp fabrics constitute a more than adequate alternative to traditional prepreg materials for structural applications.
NCFs offer processing advantages derived from the reduction of labour costs associated with lay-up, draping and refrigerated storage [5].
Whereas most in-plane mechanical properties of NCF laminates show slightly lower values than UDPT, this is not the case for the engineering properties that determine the design allowables. The compression after impact and open-hole compression strengths are in most cases comparable or even better for NCF-based structures [5].
2.7. Applications
Current projects see the use of NCF in very different applications ranging from sporting goods to aerospace applications. Below are some applications presented [8, 9,10].
• Sports articles: Snowboards, sledges, sail and surfing boards, and chutes.
Figure 9. NCFs in sport equipments.
• Automotive: Car body and bumpers, trucks, busses.
Figure 10. NCFs in the automotive industry.
• Wind power: Blades for wind power generators.
Figure 11. NCFs in the wind power industry.
• Ship manufacture: Fishing boats, powerboats, yachts and naval ships.
• Aviation and aerospace: Perspectives of applications as structural composites, development of high quality carbon non-crimp fabrics suitable for the primary structure of a large wing-box by DEVOLD AMT AS (Norway) and BAE Systems Airbus (UK) [7].
• Defence projects: FA-18 and Collins Class submarines in USA, and the Swedish Navy’s new YS 2000, Visby Class “stealth” corvette, built at the Karlskrona shipyard of Kockums AB. With a 73 m LOA (length overall) and a beam of 10,4 m, it is believed to be the largest sandwich composite structure built to date. Six vessels of this type have been ordered for the Swedish Navy and they are supposed to enter service in 2004 [5, 7].
Figure 12. The Visby YS 2000 Corvette.
3. Experimental
3.1. Material
3.1.1. Non-Crimp Fabrics
Six different types of Non Crimp Fabrics (NCFs), manufactured by DEVOLD AMT AS (Norway), have been tested. Pictures of the NCFs are presented in Apppendix B. The characteristics of the NCFs are presented in Table 4. Each fabric contains only one unidirectional layer of fibres. All fabrics had the same gauge platina (10 threads per inch), but different stitch patterns (stitch style and stitch length) and different bundle size (large or small). Furthermore, there were some small differences in areal weight.
• Gauge platina: Gauge where the fibres enter the machine, here 10 yarns per inch in the weft direction.
• Stitch style: 10/12// Ù Pure zig-zag 10/10/12/12// Ù Mixed zig-zag / straight
• Stitch length: 2,5 or 3,5 mm distance between the stitching points in the warp direction.
• Gauge stitching: Gauge in the stitching part, here 10 or 5 stitch points per inch in the weft direction (set by the number of needles per inch in the stitching part)
• Areal weight: Expected to be 315 g/m
2(but there were some small differences between the expected and the measured weights).
The fibre that was used to manufacture the fabrics was the T700 SC 50C 12K from Toray.
This is a standard quality fibre with intermediate modulus. The fibre characteristics are presented in Appendix C.
• T 700 SC: The fibre type is the T 700, and is used as untwisted yarns.
• 50 C: The sizing used is based on a Bisphenol A epoxy and an unsaturated polyester (Sizing code 5 in TORAYCA Classification). The size is supposed to match epoxy resin.
The sizing amount is about 1,0 %.
• 12K: 12000 filaments per roving (bundle).
Fabric P0 P1 P2 P3 P4 P5 Designation (T) L350- C10-A (/ 0 grader L 320)
Fibre Type (T) ? T700 50C 12K
Tex (T) [g/1000m] ? 800
Areal Weight (T) [g/m2] ? 315
Areal Weight (R) [g/m2] 353,18 315,78 325,03 324,43 358,57 361,23 Stitch style (T-R) 10/12// 10/12// 10/10/12/12// 10/10/12/12// 10/12// 10/12//
Stitch length (T-R) 2,5 3,5 3,5 2,5 2,5 2,5
Gauge platina (T-R) ? 10
Gauge stitching (T-R) 10 10 10 10 10 5
T: Theoretical, and R: Real/measured.
