WO2013030566A1

WO2013030566A1 - Method and apparatus for determining interlaminar shear mechanical properties of composite laminates

Info

Publication number: WO2013030566A1
Application number: PCT/GB2012/052113
Authority: WO
Inventors: Gang Zhou
Original assignee: Loughborough University
Priority date: 2011-09-02
Filing date: 2012-08-29
Publication date: 2013-03-07
Also published as: GB201115138D0; GB2494177A

Abstract

Method and Apparatus for Determining Interlaminar Shear Mechanical Properties of Composite Laminates A method and apparatus for testing interlaminar shear mechanical properties of composite laminates. The method involves preparing an elongate rectangular specimen and mounting said rectangular specimen longitudinally with three equally spaced contact points against a first surface and two contact points against the opposing, second, surface. A load is applied via the contacts to the specimen until the specimen delaminates indicated by an audible crack. Displacement of the specimen is measured at an inner region of the specimen at a section of pure shear. This measurement can be used to accurately determine interlaminar shear strength and modulus so the failure properties of the material are known for design consideration.

Description

Method and Apparatus for Determining Interlaminar Shear Mechanical Properties of Composite Laminates

TECHNICAL FIELD

The present invention relates to a method and apparatus for determining interlaminar shear mechanical properties of composite laminates. Such laminates structures are particularly used in the aerospace industry.

BACKGROUND TO THE INVENTION

Composite laminates are used extensively for load-bearing structures in various sectors of industry (e.g. aerospace, land transport, marine and wind energy) because of their light weight, high strength-to-weight and stiff-to-weight ratios, good corrosion resistance, superb fatigue strength limit and design versatility. Whenever composite laminates are used in industrial applications, structural design and stress analysis for the performance of a component (including finite element modelling) will require reliable and accurate interlaminar shear (ILS) properties of the laminates as input or prior information.

The majority of composite laminates (or structures) in use appear in sheets (plates or shells) with no fibre reinforcement in the thickness direction. When bent sufficiently, they suffer interior invisible damage such as delamination . Delamination not only shortens the designed life of the component and requires costly out- of-service repair but also posts a significant threat to its subsequent structural performance, if not detected in time. Current composite structural design practice in industry requires the residual performance of the composite component with a planar size of delamination of less than 30mm in diameter not to suffer any reduction (i.e. as good as new) . ILS mechanical properties are therefore a key set of properties used for evaluating the performance of the laminates. On the one hand, their reliable and accurate values must be available to and understood in a structural design and stress analysis such that possibility of the occurrence of delamination is minimised. On the other hand, the elimination of the underestimation of ILS properties using existing testing standards can lead to further weight-saving and cost- saving. Alternatively, developments of stitched composite materials could offer improved delamination resistance to some extent.

There are two main testing standards for the determination of ILS mechanical properties for fibre- reinforced composite laminates, namely double V-notch Iosipescu Shear (illustrated by Figure 1) and Short Beam Strength (SBS) (illustrated by Figure 2) methods. ILS mechanical properties to be measured consist of ILS strength and modulus. While Iosipescu can deliver both in a single test, SBS provides ILS strength only.

US Patent 3566681 describes the Iosipescu method but was originally intended for testing rocks and other building materials. It was later adapted for testing composite laminates in 1993.

Other known prior art relating to determining ILS properties includes US5280730 and US4926694. The reliability and accuracy of the testing methods and their ILS mechanical property data are evaluated against two technical criteria. The first is that a pure shear region must be induced within the gauge section of the loaded specimen. The second is that the induced pure shear region must initiate and lead to ultimate ILS failure (or delamination) in the gauge section of the specimen. Rigorously speaking, ILS property data are invalid if either the state of pure shear does not exist within the gauge section or the existing pure shear region does not initiate and lead to the occurrence of delamination .

In the fabrication of a typical Iosipescu specimen shown in Fig. 1, the laminate to be tested is first sandwiched (bonded) usually in the middle by two dummy stacks such that the laminate surfaces must be horizontal to the specimen surfaces with the laminate being in the middle of the specimen thickness after machining. All exterior surfaces of each specimen must be machined to exact dimensions and all the opposite surfaces must be machined to parallel. Two 90° V-notches are then cut symmetrically through the dummy stacks into 20% of the laminate thickness from each (interior) laminate surface. Alternatively, an Iosipescu specimen could directly be cut out of a very thick laminate panel (say, more than 19 mm) before being machined and having two V-notches cut.

The loading and support arrangement indicated by arrows in Fig. 1 is referred to as "asymmetrical four-point bending". The notched region between the inner loader (on the right side of the top notch) and the inner support (on the left side of the bottom notch) is defined as the gauge section. Although, during a test, a state of pure shear is induced at the centre of the specimen or at the middle between the two V-notches (indicated by the location of a cross in Fig. 1), premature failure always occurs horizontally at one of the two notch roots. This is because the stress concentrations occur at the notch roots such that local maximum stresses are much higher than that at the centre between the two notch roots, causing the premature failure. This premature failure underestimates the ILS strength of the tested laminate and therefore becomes the most significant drawback of this method. Machining notches in the specimen could create micro-cracks at the notch roots and also it is very difficult to maintain the dimensional consistency of the notch root radius of 1.3 mm throughout machining. This not only causes premature failure but also creates a substantial variation of apparent ILS strength values. A two-element strain gauge rosette (placed at the position of the cross in Fig. 1) needs to be bonded at the pure shear spot on one side of the specimen with the strain gauge elements being parallel to the notch surfaces (or ± 45° directions) such that their readings allow ILS modulus to be determined.

It should be apparent that the aforementioned Iosipescu specimens are very expensive to fabricate and prepare in terms of required time and effort. However, the advantage of this standard is its well established and accepted status as the only standard in the field for generating ILS design data.

In a loaded SBS specimen shown in Fig. 2, a state of pure shear does not exist within the gauge section (region between the two supports) of the beam specimen. Although ILS failure (i.e. delamination) takes place, it always occurs outside the gauge section at one of its two ends or overhangs. This is because a 3-dimensional state of stress exists at the beam ends so that the delamination resistance is lower. As a result, this 'edge effect' limitation could underestimate the apparent ILS strength, as is the case with the Iosipescu specimen.

Due to this drawback, SBS is used only for a screening of composite laminates, but not for generating design data. However, the obvious advantage of this method over the Iosipescu method are its simplicity in use and low-cost.

Due to the significant drawbacks from both standards, current industrial practice in a composite structural design uses an increased knock-down or safety factor to deal with the lack of reliability and inaccuracy in ILS properties, thereby adding weight and cost. Weight savings in aircraft are particularly important in a current industrial environment characterised by high fuel costs and ever-more stringent emission standards. This requires composite structure design to be cost-effective, minimum-weight and damage-tolerant. This is achievable only when their ILS performance is thoroughly understood in terms of reliable and accurate ILS properties and when the underestimation of ILS properties is eliminated. Moreover, the effective evaluation of stitching in the development of stitched composite materials for improving their ILS resistance requires a more effective testing method than the two current ones.

DISCLOSURE OF THE INVENTION

The apparatus and method of the present invention seeks to overcome the drawbacks of the existing standards by providing more reliable and accurate ILS properties for conventional composite laminates and a more effective testing method for the evaluation of delamination resistance of new or novel composite materials such as stitched composite materials. Its use can offer a further weight saving in aircraft structures and thereby an improved fuel economy, a reduced emission and reduced aircraft operating cost. In one broad aspect the present invention provides a method for determining interlaminar shear mechanical properties of composite laminates, including:

preparing an elongate rectangular specimen with laminate layers running in a longitudinal direction;

mounting said rectangular specimen longitudinally with three equally spaced contacts on a first surface and two contacts on the opposing, second, surface,

applying a load via the contacts to one of the surfaces until the specimen delaminates indicated by an audible crack,

at the time of delamination, measuring the displacement of the specimen at an inner region of the specimen at a section of pure shear located between the central of the three contacts on the first surface and either one or both of the two contacts on the second surface .

Preferably the three contacts against the first surface form a support underneath the specimen and the two contacts against the second surface provide the load force above the specimen. In any event, it will be apparent that the invention utilises five contact points, rather than three or four as in the SBS and Iosipescu methods respectively. It is noteworthy that the invention does not exclude the implemention of additional contact points, e.g. triple identical spans (via seven contact points) or any other variation that utilises at least five contact points.

While it is preferable that the underneath three contacts form the specimen support surface and upper two contacts are adapted to provide the load force, the method or device of the invention could be configured the opposite way around, i.e. with a three contact load / two contact support .

Preferably the specimen is prepared with markers to indicate the placement points of the respective contacts. Preferably a central point of the specimen is aligned with the central contact.

In practice each contact will be a surface transverse to the longitudinal direction of the specimen.

Preferably displacement is measured at one or both of the sections of pure shear by an LVDT device. The section (s) of pure shear is/are located at an inner region of the specimen between the middle support contact and each of the two load contacts. It is determined as a point of maximum interlaminar shear stress where the bending stress is effectively zero due to the opposing directions of adjacent contacts (hence "pure" shear) . Preferably an elongate side of the specimen is covered in a light coloured (e.g. white) paint layer over the "sandwich" of laminate layers such that the location of delamination is more readily visible for further study if needed . The method according to the invention can test any laminates . However, the test examples discussed herein are with reference to 4mm thick laminates. All data is calculated on that basis. It should be noted that both loader and support distances on a jig set up according to the invention can be adjusted to suit individual needs.

In a second broad aspect the invention provides an apparatus for determining interlaminar shear mechanical properties of composite laminates including:

a support surface comprised of three support contacts ;

a load surface comprised of two load contacts; loading means for applying a load to the two load contacts; and

a displacement measurement device for measuring displacement of a specimen at a pure shear location between a central support contact and either or both of the load contacts spaced therefrom.

In practice, the pure shear locations tend to be at a point equidistant between the central support contact and either of the load contacts. This is the case for symmetrical laminates, however, it may not apply to non^¬ symmetrical laminates; i.e. in that case the pure shear locations are not at a point equidistant between the central support and load contacts. The general methodology of the invention has been (and will be herein) described as "double span bending" (DSB) or the "double span bending test" (DSBT) . In practice, DSBT consists of the DSB method (or theory) , a DSB mechanical apparatus and PC-based software that is programmed according to the DSB method. Software is thus used to calculate ILS modulus, as without it the calculation could take many long hours even to a trained engineer with a basic knowledge of classical lamination theory and DSB method.

The current DSB method overcomes all the drawbacks of the existing Iosipescu Shear and SBS methods. Specifically, a state of pure ILS can be induced within the two inner regions of the gauge section of the beam specimen. As a consequence, ILS strengths obtained according to the invention are greater in magnitude than those obtained using either the SBS or Iosipescu methods because the data is unaffected by premature failure.

BRIEF DESCRIPTION OF DRAWINGS Figure 1 illustrates the Iosipescu method according to the prior art for determining ILS;

Figure 2 illustrates the short beam strength (SBS) method according to the prior art;

Figure 3 illustrates a specimen prepared and mounted according to the method of the present invention;

Figure 4 illustrates the specimen from Figure 3, further including indications of the stress pattern induced by the testing method;

Figure 5 illustrates, schematically, an example of a test apparatus according to the invention;

Figure 6 illustrates both standard and modified double span bending set-ups according to the invention;

Figure 7 shows a photograph of an experimental set-up with a suitable jig; Figure 8 illustrates the gauge section of a specimen;

Figure 9 illustrates ILS failure locations in specimens; Figure 10 shows a photograph of the occurrence of delamination in a specimen; and

Figure 11 illustrates a bar chart summarising ILS strengths from both carbon/epoxy in Double Span Bending according to the invention.

MODE(S) FOR CARRYING OUT THE INVENTION

Referring to Figures 3 and 4, the method according to the invention takes a beam specimen B that is effectively twice the length of a comparable beam from the SBS method of Figure 2. The beam B is mounted upon three equally spaced supports (arrows S) and a transverse load is applied at positions indicated by arrows L. Such a loading and support set-up creates two identical or double spans, which in turn are further split into four equal longitudinal regions (either side of each load point L between any two adjacent supports S, resulting in an inner and outer region for each side) , as illustrated in Fig. 3.

The section of a beam specimen B that is covered either by the two spans or by the four regions is defined as the gauge section. Although all four regions under load L have constant levels of ILS stress, the two inner regions experience greater levels of ILS stresses than the two outer regions, as indicated by the schematic ILS stress "graph" in Fig. 4. In particular, the two inner regions contain the sections of pure shear. Under the higher interior ILS stresses, the occurrence of ILS failure (or delamination) in any one of the two inner regions is therefore guaranteed.

It will be apparent from Figure 4 that the sections of pure shear arise at the point where the bending stress resulting from the downward force load L acting on the beam B between the supports S passes through zero on the representative graph. In practice, the two transverse loads L are applied by cylindrical line loaders (each of, for example, 6.4 mm in diameter for 4mm thick beams) mounted on a jig as illustrated by Figure 5. Line loaders are bolted onto a single load cell, which measures force. One or two LVDT (linear variable differential transformer) devices (shown in Figures 4 and 6) that are erected under the beam surface in between two cylindrical supports S (each of, for example, 6.4 mm in diameter) measure beam deflections. At the end of each test a load-deflection curve is recorded and a recorded peak or critical force measured at the occurrence of delamination is used to calculate ILS strength. A linearised dominant region that is established through two points selected on the load-deflection curve will be used for the calculation of ILS modulus through supporting software.

A specimen beam B is prepared according to the following methodology : (i) A large laminate plate is fabricated and individual beam specimens are cut out . The method of the invention is independent of the type of composite material system, be it UD tape- based, two or three dimensional fabric-based or stitched fabric. (ii) The length dimension of these specimens varies, depending on the thickness of the laminate. While the width of specimens is not critical, it must remain constant and at least 20% greater than the laminate thickness. The length of overhang (regions of the beam outside the two outer supports S indicated by letter 0 in Fig. 3) needs to be just greater than the laminate thickness . (iii) Each individual beam specimen can be cut with its longitudinal direction being coincided with either 0° or 90° fibres. If specimens in their lay-ups do not contain either 0° or 90° fibres, the orientations of angle plies must be referred to the longitudinal direction of the specimens. (iv) Each specimen preferably has one longitudinal side "whited-out " (i.e. white paint or correction fluid) and the opposite side marked out with vertical lines for all loading and support contact locations. The whited area makes visual inspection of a crack in the sample much easier. However, white-out areas may not be necessary for glass fibres, aramid fibres or any other reinforcement . (v) The longitudinal side edges of the specimens, after cutting, must be either parallel or perpendicular to the 0° fibres in the specimens. (vi) Both thickness and width dimensions are measured at three longitudinal locations, namely, under each loader and at the central support. Then average values of the three measurements are taken respectively for thickness and width of each specimen in the calculations of ILS strength and modulus .

(vii) The ratio of a single support span to the specimen thickness should be kept between 4 and 6, preferably 5.

(viii) All cylindrical line loaders and supports must be parallel to one another and the line loaders must be situated at the middle of each of the two spans. The diameters of cylindrical line loaders and supports do not need to be identical with the former generally being greater than the latter. Several sets of loader and support are thus necessary to cater for different thicknesses of practical composite materials. In order to mitigate local crushing, thin shim pads for both loaders and supports can be used for testing relatively soft/ductile composite material systems or composite material systems with relatively low through-the-thickness compressive strengths. In any event, the thickness of shims must be controlled such that their use will not invalidate the results obtained by the invention.

Mechanical testing of the sample is then undertaken according to the following steps: (ix) A laminate beam specimen prepared in accordance with steps (i) to (viii) above is placed on the supports of the apparatus with its front side aligned with the marked position line. This is aided by an 90° angle tool such that while its inner side is pushed against one end of the supports for alignment, the edge of its flat element of the tool on the top of the supports is flush with or parallel to the marked line on the supports. Then the specimen is simply pushed against the edge of the angle tool. (x) The properly aligned specimen is then held down at two locations very close to the two outer supports . (xi) One LVDT device is positioned under the specimen surface at the middle of each span (as indicated by parameter d in Fig. 4) and at the middle of its width to measure beam deflection. The spring-loaded cylindrical shaft of the LVDT must be ensured to be perpendicular to the specimen surface. If two LVDT devices are used, the average of two beam deflections should be taken. A height adjustment mechanism for the LVDT may be used to ensure that the ball bearing of the LVDT is in contact with the specimen. (xii) To start a test, load is applied at the two mid- spans at a constant speed of no more than 3mm/min, preferably 1 mm/min or less, and is measured by a load cell. (xiii) Testing is allowed to continue until an audible cracking sound associated with the occurrence of delamination is heard. Then the testing is stopped. In the case of the lack of a clear audible cracking sound from the specimen, a noticeable load drop of more than 0.5 kN is considered sufficient to stop the test. A complete load- deflection history is recorded for each specimen. (xiv) The location of delamination will be clearly indicated on the whited-out side of the tested specimen. For specimens with glass fibres, it should be self-evident due to their translucent nature .

Data from the test can then be processed, according to the following steps:

(xv) From a recorded load-deflection history for the tested specimen, the peak load or load corresponding to the occurrence of delamination must be identified.

(xvi) The peak load in addition to the specimen dimensions is used for calculating the ILS strength of the laminate specimen.

(xvii) To determine ILS modulus supporting software is required to perform the calculation. The following information or data, in addition to the specimen dimensions, will be required as input: Physical characteristics: a lay-up of the laminate, number of plies in the laminate, (cured) ply thickness;

Basic mechanical properties: E_llr E₂2, G₁₂ and v₁₂;

Testing set-up characteristics: loading arm length and location of the LVDT; and Measured data: Ignore an initial 10% and final

25% of loads from each load-deflection curve. Select two points on the curve through two loads and corresponding beam deflections that correspond to a linearised region. Connect two points to establish a linearised region that will be used for the calculation of ILS modulus

The two longitudinal sections (or locations) of pure shear are illustrated by two vertical dashed lines drawing through the thickness of the specimen in Fig. 4. As a result, the occurrence of ILS failure (or delamination) in any one of the two regions (see Fig. 3) is guaranteed. For the first time, the method of the invention provides reliable and accurate ILS strength and modulus data in a single test. In addition, as it is based on the bending of simple laminate beams (and not the elaborate preparation of an Iosipescu sample) , the double span bending method of the invention maintains all the low-cost and simple-to-use characteristics of the established SBS method.

By comparison to SBS, the method of the invention has three significant advantages:

(a) it possesses two pure shear regions (see Fig. 4); (b) the location of delamination (ILS failure) coincides with one of these two pure shear regions;

(c) it does not suffer premature failure as its ILS stresses around the two pure shear regions are less affected by close-to-zero bending stresses; and

(d) it produces an ILS modulus.

When compared to Iosipescu, the method of the invention has the following six advantages:

(a) the location of delamination (ILS failure) coincides with one of the two pure shear regions;

(b) it does not suffer premature failure, unlike the notches in Iosipescu specimens that often initiate premature failure;

(c) the preparation of a specimen is significantly simpler ;

(d) a variation of specimen thickness is not limited by the jig;

(e) it requires neither notch cutting nor strain gauge bonding; and

(f) it is significantly low-cost and easy to use.

Mathematical Analysis : ILS Strength and Modulus Major mathematical results in terms of equations for the double span bending method of the invention are outlined below, e.g. ILS strength in the laminate beams is calculated by:

Taking into account the fact that material properties and lamina configuration of the laminate are inseparable, Eq. (1.0) can thus become:

in which b is the beam width, D the flexural rigidity, E_xiA_i and y' the axial rigidity and the mid-plane distance to the near surface of a sub-laminate, respectively.

The ILS modulus G at a specified location x is calculated by :

Composite Materials and Specimen Manufacture In an experimental application of the invention, two types of composite systems have been used, namely, unidirectional (UD) prepreg-based LTM45/34-700 carbon/epoxy and woven fabric-based Cycom 5250-2/G803-40 carbon/BMI . For the former, of primary interest, composite panels of 32 plies were fabricated in three different lay-ups, i.e. UD in [ (0°) ] i_6s, cross-ply in [(0°/90°)]_8s and quasi-isotropic in (-45^ο/0°/+45^ο/90°) _4s . All panels were cured in an autoclave using the manufacturer's recommended curing cycle of 18 hours at 60°C under a pressure of 0.62 MPa (90 psi) and using a ramp rate of 2°C/min. A nominal thickness of these panels is 4.0 mm with a cured nominal ply thickness of 0.128 mm. The UD mechanical properties of this composite system were determined as En of 127 GPa, ¾2 of 9.1 GPa, Gz of 5.6 GPa, and V₁₂ of 0.31.

The latter composite material is 5-harness satin woven fabric with a nominal ply thickness of 0.275 mm. Laminate panels of both 8 and 16 plies were fabricated in a quasi-isotropic lay-up of [ ( 0⁰ / 90⁰ ) _F ( ±45⁰ ) _F] _2s and [ ( 0⁰ / 90⁰ ) F (±45⁰ ) _F] 4s to make up a nominal thickness of 2 mm and 4 mm, respectively. In the lay-up process, sub- laminates were first prepared with each consisting of 4 plies and then sub-laminate stacks were individually debulked so that a caul-plate was not used in the autoclave cure. The sub-laminate stacks were finally assembled. These panels were cured in an autoclave with a curing cycle of 3 hours at 190°C under a pressure of 0.25 MPa (36 psi) and a ramp rate of 1.5°C/min. They were post-cured for 8 hours at 225°C. Its mechanical properties are E of 62.0 GPa, ¾2 of 62.0 GPa, G of 2.4 GPa, and V₁₂ of 0.05. Experimental Procedures

Both standard and modified (with the loader distance of 16 mm) double span bending set-ups according to the invention have been used, as shown in Fig. 6. All cylindrical supports and loaders are 6.4 mm in diameter. An overall support span of 40 mm was fixed for all 4-mm thick specimens. For each 4-mm thickness specimen, this provided a nominal support span-to-depth ratio of 5. A support span-to-depth ratio for the thinner specimens was about 8 as the cured laminate was about 2.5 mm thick. On one side of the specimen, all the anticipated contact areas were marked up with the vertical lines being drawn through the thickness. On the opposite side, the gauge section was painted with a thin layer of white correction liquid so that the occurrence of delamination could readily be visible (see Fig. 10) .

Once a specimen was placed over the supports, a spirit level and a set square were used to ensure that the specimen was perpendicular to the supports and loaders, which were aligned with a mid-span marker. An experimental set-up with a suitable jig is shown in Fig. 7. All the tests were carried out on a universal testing machine (e.g. a MAND universal testing machine) at the crosshead speed of 3mm/min for ILS specimens. Load and crosshead displacement and deflection were recorded through a data acquisition system (e.g. an Orion delta 3530D) at a sampling rate of 1 Hz. For comparison, three-point bending tests were also carried out for the specimens in UD, cross-ply and quasoi-isotropic lay-ups. Exerimental Results and Discussion - Failure characteristics of composite beams utilising DSB method according to the invention Some of the early experimental results obtained from both standard and modified set-ups of Figure 5 are summarised in Tables 1-3 for carbon/epoxy laminates and in Table 4 for carbon/BMI laminates. ILS strength values were calculated by using Eq. (1.0) . Towards the end of each test, once a cracking sound was heard, the test was halted immediately to allow a post^¬ mortem examination of the failed specimens to be conducted and photographic evidence to be taken. For carbon/BMI quasi-isotropic laminates with two different thicknesses, the average ILS strengths are 69.2±11.8 MPa for 16-ply laminates and 74.9 MPa for 8-ply laminates. The difference of about 8% between the two thicknesses is small .

Exploring the modified DSB set-up was intended to find the 'best' set-up such that the probability of ILS failure in the two inner regions would be increased. In modified DSB, ILS stress in the two inner regions was increased by 15% while the corresponding bending stress under each of the two loaders was decreased by about 20%. Consequently, the local stress concentrations under the loaders become less. Meanwhile, the bending stress at the central support is marginally increased (by just 2%) . The average ILS strengths of carbon/epoxy laminates obtained are 66.5±8.3 MPa for cross-ply and 63.1±8.9 MPa for quasi-isotropic. The ILS strength difference between the two lay-ups is almost negligible. For the reasons above, the post-mortem examination of the longitudinal and through-the-thickness locations of ILS failure has therefore become the most important aspect of the investigation into the ILS behaviour of composite beams when utilising the method of the invention. For the convenience of discussion, the 3-dimensional space of the gauge section of each specimen has been split into 16 regions, 4 in the longitudinal direction and 4 in the through the thickness direction, as illustrated in Fig. 8.

In general, all the longitudinal locations of ILS failure were found to be within the two inner regions (IL and IR) of the beams where a state of pure shear existed and the highest ILS stresses were analytically predicted to occur, as can be seen in Tables 1-4. All specimens from different composite materials, in different lay-ups and under both standard and modified DSB failed in one of inner regions. This significant finding has confirmed that the 'edge effect' associated with the location of ILS failure in three-point bending, the key weakness of SBS and Iosipescu methods, can be eliminated by using the method of the invention. In the through-the-thickness direction, the overwhelming majority of the ILS failure locations are in the two interior quarters, as can be seen in Tables 1-4 and Fig. 9 statistically. This includes about one third of ILS failure location at the mid-plane of the laminate specimens. Those failures that occurred outside the two interior quarters are either in the upper quarter associated with the loaders in standard DSB or in both quarters associated with the loaders and central support in modified DSB. In standard DSB, the local stress concentrations under the loaders could be the cause, as none of the specimens failed on the top of the central support. In modified DSB, the local stress concentrations affected the ILS failure locations not only associated with the loaders but also associated with the central support due to the slightly increased local normal stress. Although the testing was stopped each time (within the first three seconds) after an audible cracking sound, post-testing inspection of the failed specimens seemed to show that about 37.1% (13/35) of the tested specimens had multiple through-the-thickness delamination locations.

As a whole, changing the testing configuration from standard DSB to modified DSB has not, statistically, brought any noticeable benefit for carbon/epoxy laminates as shown in Table 5. For carbon/BMI quasi-isotropic laminates, the average ILS strengths are 69.2±11.8 MPa in standard DSB and 63.9±7.7 MPa in modified DSB. Again, the difference of about 8% between the two set-ups is less than the respective data scatters of 17% and 12%.

Results and Discussion - Effect of lay-ups on ILS strength

The common consensus for the nature of ILS properties of composite laminates is that they are resin-dominated mechanical properties. This implies that the lay-up of composite laminates should not affect the magnitude of the ILS strengths. Some results seemed to agree with this, whereas all others have shown over the years that the lay-up always affects the magnitude of ILS strengths in the SBS method. This is not entirely surprising, as the ILS failures in SBS via three-point bending are always affected by bending stresses in addition to local normal stress under the loader (s) and possibly interlaminar tensile stress at the free ends due to the lack of a state of pure shear in those methods.

From the bar chart in Fig. 11 (and also individual Tables 1-4), the average ILS strengths of carbon/epoxy laminates obtained from standard double span bending are 98.0±6.6 MPa for UD, 88.8±8.4 MPa for cross-ply laminates, 84.4±6.2 MPa for quasi-isotropic laminates. While the scatters of the individual group data are reasonably good (within 9% in the worse case) , the steady reductions of the ILS strengths from UD, cross ply, to quasi-isotropic are more or less in the same range as the scatters of the individual group data.

When 50% of fibres were directed to the in-plane transverse direction, the loss in ILS strength of cross ply laminates is only 9%. A further reinforcement reduction of 25% in the longitudinal direction led to just 5% loss in ILS strength. Even from UD to quasi- isotropic, the reduction in ILS strength is only about 14%. As the earlier discussion on ILS failure locations indicates that about 12% of ILS failures, which occurred outside the two interior quarters, were not driven by ILS stresses, this percentage seems in qualitative accordance with the percentage loss of the ILS strengths from UD to quasi-isotropic lay-up. After all, a state of pure shear existed only over the two cross sections, in which the maximum ILS failure was most likely to occur within the two interior through-the-thickness quarters. ILS Modulus

Calculation of ILS modulus must be made through software which requires a considerable amount of basic data in addition to testing data as input, and also shown through Eqs (2.2) - (2.6) . Early trials indicate that the values are close to those produced by losipescu method.

Claims

CLAIMS:

1. A method for determining interlaminar shear mechanical properties of composite laminates, including :

preparing an elongate rectangular specimen with laminate layers running in a longitudinal direction; mounting said rectangular specimen longitudinally with three equally spaced contacts against a first surface and two contacts against the opposing, second, surface,

applying a load via the contacts to the specimen until the specimen delaminates indicated by an audible crack,

at the time of delamination, measuring the displacement of the specimen at an inner region of the specimen at a section of pure shear located between the central of the three contacts on the first surface and either one or both of the two contacts on the second surface.

2. The method of claim 1 wherein the three contacts against the first surface form a support underneath the specimen and the two contacts against the second surface are able to provide a downward load force from above the specimen.

3. The method of claim 1 or 2 wherein the specimen is prepared with a marker to indicate placement of the respective contacts.

4. The method of claim 3 wherein a central marker point of the specimen is aligned with the central contact .

5. The method of any of the preceding claims wherein displacement is measured at one or both of the sections of pure shear by an LVDT device.

6. The method of claim 2 wherein the points of pure shear at which to take the displacement measurement ( s ) are determined between the middle support and each load contact by taking into account the thickness of the specimen.

7. The method of any of the preceding claims wherein an elongate surface of the specimen is covered in a paint layer over the "sandwich" of laminate layers such that the location of delamination is more readily visible.

8. An apparatus for determining interlaminar shear mechanical properties of composite laminates including :

a support surface comprised of three support contacts ;

9. The apparatus of claim 8, further including a specimen holding means.

10. The apparatus of claim 8. wherein the displacement measurement device is an LVDT.