Acoustic Emission Characterization of Matrix Damage Initiation in Woven CFRP Composites

In this paper, an experimental investigation is performed using acoustic emission (AE) technique to study the behaviour of crack initiation and propagation of woven carbon fibre reinforced plastic (CFRP) composites under static tensile loading. Three laminates with three plies each were manufactured and tested under tensile static loading. Damage was being monitored using AE technique with instantaneous AE hits energies being used as the damage parameter. Results showed that angle ply laminates experienced crack initiations at about 50% of ultimate tensile strength whereas cross ply laminates and non-conventional angle ply laminates having one cross ply laminate experienced initiation of cracks at about a third of ultimate tensile strength. This early cracks initiation is believed to be caused by the presence of 90 ̊ plies that cannot take much load transversely.


Introduction
The initiation and propagation of the failure modes in CFRP laminates under mechanical loads is usually difficult to predict and monitor.One of the principle reasons for that is the lack of transparency of CFRP composites that does not allow the inner damage evaluation by common inspection methods.For this reason, AE technique is used on composite materials to characterize damage failure modes under various loading conditions.AE technique is one of the principle NDT methods for distinguishing and identifying growing material defects [1].AE is defined according to ASTM standard [2] as "the class of phenomena where transient elastic waves are generated by the rapid release of energy from localized sources within a material, or the transient elastic waves so generated".In context of fibre reinforced composites, the AE waves are stress waves generated by the mechanical deformation of materials e.g., initiation/ growth of cracks, yielding, material dislocations, impacts, rubbing of contacting surfaces and fibre failure [3].Typically, mechanical stresses are produced by applying external forces causing the release of elastic energy "waves", e.g. by crack formation.The elastic wave propagating through the material is detected and converted into an electrical AE signal by the AE sensors.The AE system processes the AE signal, converts the received wave packets into feature data sets, determines the source location, calculates statistics, and displays them graphically and numerically.AE systems contain parametric channels to measure the environmental conditions as well as the external load as reference parameters for the detected AE hits [4].
Many researchers studied the influence of various parameters, such as fibre orientation, laminate thickness and stacking sequence on the acoustic waves response, often analysing signal count, amplitude and energy [5][6][7].AE technique was found to demonstrate excellent results on detecting and identifying initiations sites, cracking propagation and failure mechanisms of CFRP composite.Moreover, the use of AE is not only limited to study the initiation of cracks and its wave propagation in CFRP composites but also in natural fibre reinforced composite materials [8].In particular, AE is used to determine the strength of the interface between fibres and polymer matrices.Additionally, AE is also successfully used to monitor damage in other materials such as metals [9].Hufenbach et al. [10] carried out an experiment with differently structured multi-layered symmetric cross-ply Acoustic Emission Characterization of Matrix Damage Initiation in Woven CFRP Composites 510 and angle-ply laminates, for different textile fabrics.They found out that failure of a thermosetting and thermoplastic fibrous composite could be forecasted far before the real total failure by a strongly exponential increase of the AE activity.Their findings were in agreement with Bourchak et al. [7] results regarding crack initiation and failure threshold observed in CFRP composites with the use of AE technique.Lopresto et al. [11] executed an experiment in which extensional mode of propagation was analysed.The data collected were used to find the four elastic constants characterizing the inplane behaviour of the basic lamina.Whereas, Liu et al. [12] tested epoxy composite specimens used AE technique to study the effects of different lay-up patterns and mapping between the failure properties and the acoustic signal features.It was expected that the influence of complex lay-up patterns and sizes on the damage and failure properties of composites be represented by creating true mapping based on the AE technique.Mizutani et al. [13] used AE signal in combination with the modal analysis of lamb waves.The sequence of microscopic fracture mechanisms in locally loaded cross-ply carbon fibre composites were studied using microscopic and ultrasonic examination.Some researchers have used numerical model method to validate the results with the AE data.Johnson and Gudmunson [14] studied the transient wave propagation resulting from transverse matrix cracking in cross ply composite tensile test specimens that were investigated both theoretically and experimentally.The comparisons indicated a close resemblance between experimentally measured and numerically calculated signals.May et al. [15] carried out a series of tests which were conducted to act as validation cases for the numerical model to predict the initiation and propagation of damage in composite materials.The number of cracks identified by significant AE hits correlated well with the number of cracks identified by X-rays.Truong et al. [16] investigated the initiation of damage in laminates due to tensile loading using acoustic emission and X-ray radiography.The damage initiation and evolution in the composites reinforced with sheared fabrics depends on the shear angle of the reinforcement.Kam and Lai [17] carried out experimental and theoretical methods to study the first ply failure strength of laminated composite plates under different loading conditions.A finite element analysis, which was constructed based on the layer wise linear displacement theory and the Tsai Wu failure criterion, were used to predict the first ply failure strength of the plates.Havaldar et al. [18] examined the influence of architecture on the compressive behaviour of angle-ply composite laminates using dynamic load.The analytic predictions and experimental results both showed that the biaxial stress state that develops in angle-ply composite laminates has a significant effect on the axial compressive strength.Lopresto et al. [19] used AE to investigate the influence of the stacking sequence and direction on the acoustic response of the materials and suggested that a method based on attenuation, rather than arrival time, could be useful for source location in composite laminates.
Furthermore, angle ply composite laminates are an important class of laminates because they combine good properties in both axial and shear directions.Several investigators have evaluated the use of angle-ply laminates to determine compressive strength.Moreover, the properties of the composite with different angel ply can be found by the classical laminate analysis method, but in this paper we aim to study the crack initiation and propagation of AE waves which makes the effect of altering ply angle more distinctive.As part of this work, static tensile tests with AE technique are performed on woven CFRP composite laminates with different lay-up pattern to study the damage initiation and propagation.Some representative features such as the AE signals energy are extracted with the load to study the initiation of cracks and their propagation up to ultimate failure.The stress at which Acoustic Emission activity starts is referred here as the crack initiation stress.Ultimate tensile stress is the stress at total specimen failure.One of the objectives of this work is to shed light on any possible relationship between these two stresses.

Specimens Preparation
The specimens were prepared using T300 plain 2D weave carbon fibre shown in Figure 1 and Epolam 2063 two-component epoxy resin.The Epolam 2063 epoxy resin has a low viscosity and a high reactivity.It also has long gel time and behaves well in damp environment.The formulated resin should yield good mechanical, electrical, chemical and water resistance properties.The Epolam 2063 provide excellent characteristics when used as poly epoxy resins.Successful applications include large filament wound booster and missile motor cases, pressure vessels, composite pipes, reverse osmosis tubular high pressure equipment, aerospace and automotive components and high temperature epoxy tooling, etc.The Epolam pre-mix ratio of resin to hardener is 48:52.The resulting matrix was mixed well using a magnetic stirrer shown in Figure 2. Three different batches of specimens were then prepared using resin infusion method as shown in Figure 3.The laminates were cured in a hot press using the manufacturer's recommended curing cycle.The latter consisted of three stages; preheating at 80˚C for 30 minutes, curing for six hours at 80˚C and then post curing for four hours at 80˚C.After the curing, average thicknesses were found to be approximately 0.6 mm for all three laminate.
Four specimens were cut from each laminate according to ASTM standard [20] using a diamond saw machine.The specimens were 20 mm wide and had a gauge length of 100 mm.The specimens are shown in Figure 4 (after testing) and have the following layups: 1) Laminate A specimens layup: (±45˚) 3 .

AE Damage Detection System
Static tensile tests were then performed as per ASTM standard [20] using a MTS 809 testing machine at a crosshead speed of 2 mm/min.The data was analysed to examine the initiation of cracks stress levels and how this stress threshold is affected by altering the plies angles.Damage was being monitored using Vallen AMSY-5 AE system seen in Figure 5.The AE equipment is composed of the transducers, the preliminary amplifiers and software for signal gathering and processing, recording and display as shown in Figure 5.  Considering the stress sensitivity for composite laminates, two AE transducers (sensors) are symmetrically placed on the specimen ends 85 mm apart and slightly away from the machine grips as shown in Figure 6.
The difference in the arrival time of the AE signals received by these two transducers is used to find the source locations of damage events.The preliminary amplifier employs the enlarging circuits with a wide frequency.The signal gathering and processing systems describe the energy time curve, the counting time curve and the amplitude time curve.To improve the test precision, the sensitivity and consistency of the AE system are calibrated before experiments.A 0.5 mm pencil lead break tests were carried out prior to each test to calibrate the hits location accuracy of the AE system as per ASTM standard [21].The pencil lead was broken against the CFRP laminates and the produced elastic waves are recorded by two transducers and its results analysed to work out the speed of sound on the laminate.-9 show the AE energy vs. tensile stress for all laminates A, B and C specimens respectively.It can be seen clearly how AE energy fluctuates whenever there is damage happening on the specimen up to final failure.However, for the objective of this work, only the stress level at the point of crack initiation is of importance.All figures show a distinct behaviour where no AE activity is recorded earlier in the loading curve until the matrix cracks in the specimens start to occur.After that, there are different peaks of AE energy that represent the severity of damage where fibre breaks usually cause the highest energy levels [7,22,23].It is also noticed that AE energy activity slows down indicating saturation in the laminate damage towards ultimate failure.Most of the AE activity and consequently early damage is contrib.uted to the presence of off-axis (i.e., off loading axis) plies.
Data extracted from Figures 7-9 are summarized in Tables 1-3 for laminates A, B and C respectively.These tables show the tensile tests results for all tested specimens in terms of average stress levels at which matrix cracks initiated and also the ultimate tensile stress.Even though fibre reinforced composite materials are known to have significant statistical scatter in mechanical test data due to the complex interaction between fibres and resin as well as the uncontrollable manufacturing variations [24], the results of this work were consistent from speci-men to specimen as shown in Tables 1-3.
Figure 10 shows a summary of results.As can be seen, the matrix crack initiation stress is almost half (55%) the ultimate stress in the (±45˚) 3 angle ply specimens whereas it is about a third (32%) of ultimate tensile stress in the (0˚/90˚) 3 cross ply specimens and also about a third (36%) of ultimate tensile stress in the unconventional angle ply laminate C. Unlike Laminate A specimens, the similarities in matrix crack initiation stress level in laminates B and C are contributed to the presence of 90˚ fibres.It is also noticed how the strength of angle ply specimens is significantly reduced by using only (±45˚) plies.

Conclusions
Changing ply angles in CFRP composites is known to significantly affect the ultimate tensile strength.However, the presence of 90˚ plies can cause the matrix to crack early and at a similar stress level regardless of the presence of other non 90˚ plies.However, to detect this important threshold is necessary to use a sensitive damage detection method such as AE technique.Additionally, this damage threshold is found to be as little as a third of the ultimate tensile strength in woven CFRP composite laminates having some of its fibres laying on the 90˚ angle.Consequently, results of this work indicate that in woven CFRP composite laminates there is a nonlinear relationship between fibre angle, ultimate tensile strength and matrix crack initiation stress level.Moreover, this work focused specifically on the general alteration of ply angles and how this affect the crack initiation stress with some sequences chosen for experimental investigation.However, due to the nature of composite materials, vast amount of possible stacking sequences could be designed.A more precise investigation is recommended to make the relationship between crack initiation stress and ultimate tensile stress more generic regardless of the stacking sequence.This can only be done with a comprehensive tests program that includes a large number of specimens.

Figure 6 .
Figure 6.Composite specimen positioned for AE test.

Figures 7
Figures 7-9show the AE energy vs. tensile stress for all laminates A, B and C specimens respectively.It can be seen clearly how AE energy fluctuates whenever there is damage happening on the specimen up to final failure.However, for the objective of this work, only the stress level at the point of crack initiation is of importance.All figures show a distinct behaviour where no AE activity is recorded earlier in the loading curve until the matrix cracks in the specimens start to occur.After that, there are different peaks of AE energy that represent the severity of damage where fibre breaks usually cause the highest energy levels[7,22,23].It is also noticed that AE energy activity slows down indicating saturation in the laminate damage towards ultimate failure.Most of the AE activity and consequently early damage is contrib.uted to the presence of off-axis (i.e., off loading axis) plies.Data extracted from Figures7-9are summarized in Tables 1-3 for laminates A, B and C respectively.These tables show the tensile tests results for all tested specimens in terms of average stress levels at which matrix cracks initiated and also the ultimate tensile stress.Even though fibre reinforced composite materials are known to have significant statistical scatter in mechanical test data due to the complex interaction between fibres and resin as well as the uncontrollable manufacturing variations[24], the results of this work were consistent from speci-

Figure 10 .
Figure 10.Summary of ratios of crack initiation stress to ultimate tensile stress.