Advanced Methods Of Impact Damage Analysis In Composites Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Detecting impact damage is an important factor for maintaining the structural integrity of aerospace composite structures. Since impacts can cause severe reductions in stiffness and strength of the composite structures, there is a need to look at the material's stiffness and strength after an impact event. This paper discusses the classification of the extent and type of damage based on the use of Non-destructive Evaluation (NDE) techniques. The composite materials chosen for this research are Carbon Fiber Reinforced Plastics (CFRP) with a MTM57 resin system. The materials were manufactured using standard hand lay-ups to produce plate specimens measuring 250 mm - 150 mm with 11, 12 or 13 layers. To perform the damaging impact tests, an instrumented drop test machine was used. The impact energy was set from 0.37 J to 41.72J. The impacted specimens were later examined using X-ray and SEM techniques. A damage model was developed from this work which can provide sufficient information on the type and extent of damages. The model can be used to provide the fundamental understanding and the prediction of damage and failure progression in CFRP as a function of the number of layers and impact energies.


A. Carbon Fibre. A. Prepreg. B. Impact behaviour.

1. Introduction

Damage in composite structures is usually caused by the development of different failure modes induced in the zone surrounding the impact point. The behaviour of composite structures under the impact shock is complex due to simultaneous concurrent damage mechanisms [1-3]. This failure mode is a function of the type of impact loading, lay-up sequence and boundary conditions of the structure during the event of an impact. For an effective design of a composite structure, all failure modes must be taken into account to predict the type of damage with specific regards to the impact loading. It is not possible to identify all failure modes taking place during and after the impact through visual observation [4]. Types of damage are arranged in a complicated pattern such that the whole series of impact involved and the damage size cannot be predicted analytically. Damage in composite structures will first initiate in the form of matrix cracks and breakage, and then induce delaminations at ply interfaces at the next stage [1, 5], due to the effect of impact parameters such as large contact force and high velocity impact. Delamination is the type of failure mode that may cause reductions in the mechanical properties of the material. However, as the size of delamination increases linearly with the impact energy, damages become more severe and arrive at fibre failure. The important factor leading to this damage mode (fibre failure) is the effect of micro structural randomness on the delamination behaviour at the impacted surface. The damage may further expand within the laminates, reduce the strength and stiffness of the structure, and significantly degrade the load carrying capabilities [6-9]. The internal damage is often barely visible and termed as barely visible impact damage (BVID), however to some extent it may also be barely detectable, and this event is known as barely detectable impact damage (BDID) [1].

There are various methods used for detecting the presence of impact damage. It is important to determine the existence and location of damage and its extent. Usually by carrying out visual observation directly on the impacted surface, the presence of matrix cracks and the size and shape of delaminations can be predicted. Previous studies revealed that this observation technique has lead to a great deal of understanding of the impact damage and its development [10-11]. However, this is clearly not a true measure of the failure modes prediction and it was considered a possible proxy for the true damage before the non-destructive evaluation (NDE) techniques were performed. The NDE techniques consist of x-ray radiography, magnetic particle crack detection, dye penetrant testing, ultrasonic flaw detection and electro-magnetic testing. Since there are many NDE techniques performed for inspecting damages, two most commonly used were the ultrasonic and x-ray radiography method. There are also varieties of imaging techniques that may be used for materials characterisation depending on the type of information needed. For example, x-ray analysis will generally give information about the overall area of damage, however it will not provide information on different damage types. Imaging techniques can be used with more accuracy [12]. For applications related to high-resolution surface, two types of imaging techniques can be used; Atomic Force Microscopy (AFM) and Scanning Electron Microscopy (SEM).

Therefore, this current work discusses experimental results of the low velocity impact in structures made from woven CFRP prepreg plates of different layer thicknesses, under varying impact energy levels. This study also explains the method for producing the test specimens and the set-ups of equipments for performing the dynamic impact test using an instrumented drop test rig. Further results were also made on the non-destructive evaluation (NDE) using the x-ray radiography as well as the micro imaging techniques using the SEM fractography to inspect the topographies of impacted specimens at high magnification. The study also discusses correlation between the damage modes and the impact energy. A progressive damage model was also been developed which can be used to assess the extent and types of damage.

2. Impact Tests in Composite Structures

In this section, material used to perform the impact test is discussed. Further explanations are also provided about the drop test rig used in this work, the test procedures implemented and damage measurements.

Materials and Test Specimens

The composite material chosen for the research reported here is a woven Carbon Fibre Reinforced Polymer (CFRP) prepreg manufactured by ACG (Advanced Composite Group). The type of material used was MTM resin systems (42%RW) with CF2900 fabric (280 g/m2, 12K and 2 - 2 twill fabric). The term MTM57 means that the prepreg component is based on a 120°C curing epoxy matrix resin whilst 42% RW indicates that the prepreg has 42% of resin contents in it. The code CF2900 specifies the Generic Specification of the material whilst 280 g/m2 and 12K indicate the fabric weight and number of filaments of the prepreg used. This material was supplied in a 1.25 m - 16 m roll and the total area of the material was 20 m2. It was stored in a refrigerator at -18°C except during the manufacturing processes. The composite plate was fabricated by a hand lay-up method and the curing processes used a standard vacuum-bagging procedure with the application of elevated temperature and pressure in an autoclave (cured for 30 minutes at 120°C at 5.8 bar). An approximate equation for the thickness of the layers is as follows,


where wf = fabric weight (280 g/m2), = fiber density (1.8 g/m3) and vf = fibre volume fraction (49%). The value of cpt obtained was 0.317 mm per ply and therefore the thickness calculated for 11, 12 and 13 layers are 3.5 mm, 3.8 mm and 4.1 mm respectively; the reader can consult [13] for details. Nominal size of the test specimens was 250 mm long and 150 mm wide. PZT sensors of type Sonox P5 were placed at three different positions of the test specimens in order to record the responses from the impact event.

A total of 32 plates were used to perform the damaging tests at 22 different energy levels for the 12-layers specimen. The impact energies for the 12-layers specimens were set to range from 0.37 J to 41.72 J. The remaining 10 plates were used to check the repeatability of the test by carrying out further tests at four selected energies (three tests each at 41.72 J and 20.86 J and two tests each at 31.29 J and 10.43 J). The reason for performing the repeatability test was to check the accuracy of the measurement conducted and it was only carried out in the 12-layers specimens due to material and cost limitation. For 11-layers and 13-layers, 4 plates were used to perform the impact test at four different energy levels (41.72 J, 31.29 J, 20.86 J and 10.43 J). A simple nomenclature was defined to identify the plate. The 12-layers plate, labelled A, B, C and F were named accordingly (for example, the plate B were labelled 12LB1 to 12LB8) with each of the label has 8 plates whilst for 11 and 13-layers plate, they were labelled as D and E (for example, the plate D were labelled 11LD1 to 11LD4).

Drop Test Apparatus and Test Methods

A standard drop-weight impact-testing rig was used to implement low-velocity impact damage in CFRP laminates. The overall features of the testing apparatus are shown in Figure 1. This testing apparatus consisted of a large platform equipped with an instrumented impactor that is movable in both x and y direction that allows varying location of impacts to be carried out. This instrumented impactor was allowed to fall from a determined height guided by a rail to strike onto the test specimen clamped horizontally on both sides. The impactor was a hemispherical steel cylinder of Young's modulus equal to 21 MPa. This impactor was controlled electro-magnetically during the test whilst a simple mechanical catching system was designed to prevent multiple impacts on the test specimen. The required impact energy was attained by selecting an appropriate combination of impactor mass and drop height. For example, an impactor of total mass m = 2.25 kg was raised to a height, h = 1.89 m to perform an impact of 41.72 J. Therefore, a series of impact tests (ranging from 0.37 J to 41.72 J) were performed in order to introduce damage to the specimen. Similar tests were performed on different sample thicknesses to achieve different contact force histories and responses.

Figure 1.

Most of the impact tests were based on the conservation of energy principle where the potential energy (PE) before the impact event is equal to the kinetic energy (KE) after the impact event [14-15] as shown in Equation 2,


where m = mass of the impactor, h = drop height, g = acceleration of gravity and v = velocity at impact. The velocity at the impact location (as shown in Equation (3)) can be calculated by solving Equation (2) under the assumption that drag force caused by air resistance can be neglected and the impact velocity is independent of the mass.


The test specimen was clamped horizontally at both edges using the clamping mechanism and the damaging test was carried out by hitting its centre using the hemispherical head at an impact energy selected before starting the tests. Each test specimen was hit once at a time and after each hit, the impactor was seized by the catcher system to avoid further damage due to rebound. This process continued until a series of varying features of damage development were acquired. The impact history was recorded by the LMS Testlab Impact Modal Analysis which provides force-time plots and the impact responses were traced from the signals captured from PZT sensors which present voltage-time plots. After performing the impact test, all the test specimens were visually inspected to access the damage extent.

3. Non-destructive Evaluation

This section briefly describes NDE techniques used for damage detection in composite structures. The focus is on advantages and limitations of using these NDE techniques, especially the x-ray radiography and SEM fractography.

3.1 X-ray Radiography

The type of radiography used here was Hewlett-Packard Faxitron® of model 43855A. To produce image of high quality, certain parameters need to be satisfied. The energy level was set to 22kV at 10 mA of constant tube current whilst the film to focus distance was set to 720 mm. Exposure time for producing satisfactory images were set to 1 minutes and 45 seconds. Radiography depends upon absorption of different degrees of radiation by the impacted specimen. The technique is based on absorption of penetrant (zinc iodide, ZnI2) by the impacted or damaged surface [16-17]. The main drawback in this technique is that, the penetrant only absorbs well in the areas where damage is clearly visible. For damages that are not severe or only occur at the upper surface of a specimen, it is hard to produce an x-ray image since the penetrant is not absorbed. As a result, the x-ray radiation would not be able to pass through it and there would be no images recorded. It was also discovered that for internal defects where it is impossible to fill with the penetrant, the damages may also remain undetected in the x-ray film [18].

3.2 SEM Fractography

For this research work, JEOL JSM 6400 was used to determine the type of failure mechanisms observed from the composite laminates after the low-velocity impacts. The samples were then coated with gold in a Polaron EmScope SC500 sputter coater to provide a gold layer of approximately 20nm thickness and prevent charging effects during the imaging process. Each specimen was focused at different magnification to distinguish the specimen type and failure mechanism. The magnifications used for this current work were 35, 100, 500, 1000, 2500 and 5000. At 50-100 magnification, the fractographic analysis illustrates the overall view of the damage area, while at 500-750 magnification it allows one to look at the failure mechanism. Finally, at 1000-2500 magnification it gives one a chance to look at the failure mechanism in more detail.

The next procedure of examining the impacted surface was by looking at the cross sectional defects. This can be done by cutting along the damage area into half and examined the cross sectional damages. Once the samples were cut into half, they need to undergo another crucial process that is the grinding and polishing. The samples were need to be grinded and polished well to produce a specimen that has no deformation or smearing of the sample, good edge and freedom from any type of scratches which may confuse microimages interpretation. All preparation methods are equally important; starting from the initial cut to the final polish must be made carefully. For this cross sectional inspection, only selected specimens were used for investigation of which 4 specimens each from 11, 12 and 13 layers. (11LA1-11LA4, 12LA1-12LA4, 13LA1-13LA4). Each sample was viewed at 20, 100, 500 and 2500 magnification to look at the cross sectional failure modes more precisely and to measure the depth of cracks induced from the test.

4. Results and Discussions

The analyses of the acquired results are presented in this section. Experimental results are divided into three main categories; results obtained from the contact force histories, results from the x-ray radiography and results retrieved from the SEM fractography both from the plain and cross-sectional views.

Results from the Contact Force History

Contact force history can be divided into two types; load-displacement curve and force-time curve. Numerous studies have been conducted on the behaviour of contact force history during the impact on composite structure [19-20]. The force-time history can yield significant information pertaining to damage initiation and growth [22-25]. A schematic view for a typical force-time curve obtained from this research is shown in Figure 2. It was found that as the impactor hits the specimen, the contact force recorded from the transducer increases in a sinusoidal like manner with time. Non-destructive evaluations (x-ray and micrograph images) have revealed that matrix cracking, matrix breakage and small amount of delamination growth have occurred once the impactor hits the specimen. The failure modes observed from the impacted specimens varied with the peak force recorded. As the contact force increases and reaches the failure load F1, it can be seen that there is a sharp drop on the contact force indicating a sudden decrease in the transverse stiffness [21, 26] of the impacted specimen. This stiffness loss might be the result of matrix crack and breakage which may later lead to delamination growth. After the sudden drop, the contact force will continue to increase further and reach the maximum peak force. Once the impactor begins to rebound, the contact force decreases in the same sinusoidal pattern until the whole contact force is lost. Previous studies revealed that the failure load, F1, varies depending on the impactor mass, impact velocity, varying plate size and thickness and the boundary conditions [26-27]. The contact duration varies for different type of drop height, impactor size and thickness. For large impact mass such as the 2.25 kg, the force-time curves were composed of several small oscillations due to the plate vibrating against the impactor during the impact event and the contact duration was longer. It is vice-versa for the force-time curves obtained for a smaller mass such as the 0.75 kg where the contact duration was much shorter and the oscillations did not have much time to develop. This type of response is typical for a drop test application which uses a smaller mass and shorter drop height to perform the test. Figure 3(a-b) shows the difference between these two types of impact mass with regards to the contact duration.

Figure 2.

Figure 3.

Table 1 summarises the results obtained from impact test for the whole 12-layers specimen. It also demonstrates the repeatability test carried on selected impact energy for the 12-layers specimen. Table 1 confirms that the repeatability was excellent, at least in terms of the peak contact force recorded. The reason for performing the repeatability test was to ensure that the contact force is behaving in a similar manner under the same impact conditions.

Table 1.

Table 2 shows the results for the 11-layer plates. Four different energy levels were tested at this point and the results revealed that the measured peak force showed an excellent correlation with the impact energy. Due to material limitations, repeatability tests were not conducted for 11 and 13 layers.

Table 2.

The results for the 13-layer plates are shown in Table 3. Note that the impact energies used here were the same as for 11 layers. From Table 3, it can be concluded that as the amount of impact energy increased, the amount of force detected also increased and showed a positive correlation between the impact energy and force detected.

Table 3.

Results from the X-ray Radiography

There were two different methods used to estimate damage. The first method utilised visual observation of the damage size using a vernier calliper directly on the test specimen whilst the second method used x-ray radiography. The damage was found to be circular in shape on visual inspection, however the x-ray film showed the damage to be rectangular. Table 4 summarises the results obtained for different types of estimation made on all the specimens. From these tables, it can be seen that damages were not captured in the x-ray film at the impact energy of 10.43 J. This is because damages below this energy level are categorised as surface defect which consist of only matrix cracks and matrix breakage. As a result, the x-ray radiation would not be able to pass through and there could be no images recorded. Only when the energy level reaches 20.86 J, the damages are visually captured for all the specimens. Therefore, x-ray radiography process was not carried out for specimens that were impacted at lowest impact energy (0.37 J-2.22 J) due to the fact that there were no damages spotted at the impact energy below than 10.43 J for all the specimen. Figure 4(a-b) illustrates the results obtained for all the layers and the differences recorded in term of damage area measured for all the layers. Results reveal that, at the highest impact energy (41.72 J), the damage area measured for 11-layers specimen was large, followed by 12-layers and 13-layers. This is because specimens with a smaller amount of layers are prone to more damage compared to specimens with larger amount of layers. The general trend that was observed indicates that, as the plate thickness increases, the damage detected on the specimens decrease due to their ability to absorb the impact energy. Therefore it can be concluded that the critical threshold for damages captured in x-ray film is above 20.86 J for all the specimens. To determine precisely on the failure modes, SEM is therefore needed and the results were discussed in the next section.

Table 4.

Figure 4.

Results from the SEM Fractography

Damages that were produced in this study can be categorised into three different energy levels. The matrix crack or matrix breakage was identified at impact energy below 20.86 J, interlaminar delamination at the impact energy between 20.86 J and 31.29 J whilst fibre crack and fibre breakage was identified at impact energy above 31.29 J.

Figure 5(a-i) shows the results of SEM micrographs (both the surface and cross sectional defects) and radiography images captured at the highest impact energy (41.29 J). These high-resolution micrographs confirm the hypothesis that fibre crack and fibre breakage was present at this point and there is no significant difference in the damage features captured for all the layers. These types of damages could reduce the stiffness and toughness of the material and leads to catastrophic failure of the structure at extreme scenario. Cross sectional micrograph images reveal that, as the impactor hits the test specimen, there was a crack developed inside the laminate and was arranged in a complicated pattern that would be very difficult to measure. For this research, the crack length was measured along the crack line on the specimen cross sectional inspection. This crack initiation and growth depends on the amount of contact force introduced to the test specimen. It was observed that, the length of crack differ with the plate thicknesses. As the plate thickness increases, the crack length measured on the test specimens decrease. Although radiograph images was used to estimate the damage area, preliminary prediction of failure modes can also be made by looking at the different colour contrast produced on the radiograph images.

Figure 5.

The same trend was also found at the impact energy of 31.29 J as revealed in Figure 6(a-i). This is because all those failure modes observed above this impact energy is categorised as critical failure modes which changes the mechanical response of the material [28]. It is also found that the amount of severity for these fibre failures increased with the raise of the impact energy and peak force. At highest impact energy, there are large numbers of fibres being damaged compared to lower impact energy. This can be clearly seen by comparing the amount of damaged fibres captured in Figure 5(a, d, g) and Figure 6(a, d, g). It was revealed that, as the amount of contact force decreases, there was a clear and consistent decrease observed in the crack length as shown in Figure 6(a-i).

Figure 6.

Figure 7(a-i) illustrates the SEM images captured at 20.86 J. At this impact energy, it can be seen that 11-layers specimen faces a severe matrix breakage and there were also some spot of fibre crack. For the 12-layers specimen, the type of damages that were observed includes matrix crack, matrix breakage and fibre crack. There were quite large numbers of matrix cracks observed at 13-layers specimen due to the ability to absorb higher level of impact energy.

Figure 7.

Figure 8(a-i) demonstrates the SEM images captured at 10.43 J. The damage hypothesis described here is well in line with the observations made using cross sectional SEM fractography. It was found that, any damages that were observed below 20.86 J can be classified as not severe since they are in the form of matrix crack and matrix breakage. Efforts were made to obtain a fractograph showing matrix cracks in support of the above hypothesis, but without much success. It was also observed that, there were no internal laminate defects recorded in the cross sectional microimages. Most of these damages will lead to delamination if the amount of impact energy increases. It can be observed that, there were also no images captured in the x-ray radiography below this threshold level. This supports the hypothesis that below the impact energy of 10.43 J, there will be no images captured in the x-ray radiography.

Figure 8.

Results based from this SEM micrograph images allowed to identify three critical impact energy thresholds. The first threshold is for determining matrix crack and matrix breakage which was identified below the impact energy of 20.86 J whilst for fibre crack the threshold was classified between 31.29 J and 20.86 J. Finally the critical threshold for most of the damages was distinguished above 31.29 J.

Table 5 concludes the crack length measured along the crack line. It was observed that, as the amount of impact energy increased, the length of crack measured also increased. It was also found that as the number of layers is increased, the length of crack decreases. This support the hypothesis that, the more layers the sample has the more energy it can absorb. Therefore, the crack length measured for 13 layers has a lower value compared to 11 and 12 layers.

Table 5.

More information about the damage caused to this material can be retrieved by simply inspecting the cross sectional defects of the damaged area. Three types of different damages were observed through the cross sectional defects. It consists of interfacial micro-cracking, parallel cracks and single catastrophic crack. A damage pattern was generated from the cross sectional inspection as shown in Figure 9. This pattern was constructed from the observation made on the cross sectional fractography of the damaged area as illustrated in Figure 10. For impact energies which are less than 20.86 J, most of the damages were categorised as interfacial micro-cracking resulting in a multiple 'branching' pattern on the upper layer of the laminate (Figure 9a). As the impact energy increases and varied between 20.80 J and 31.29 J, damage progresses from top down within the laminate in a parallel manner surrounded by micro-cracking. This results in a parallel 'branching' pattern as revealed in Figure 9(b). It was observed that, as the impact energy reached 31.29 J, damages become more severe which starts a pattern of a single crack that leads to a single 'branching' pattern as shown in Figure 9(c).

Figure 9.

Figure 10.

5. Damage Model

Once damage occurs, the material may undergo some lost in mechanical properties on the impacted area [29]. The lost in the mechanical properties are strongly dependent on the failure modes resulting from the impact damage. Therefore, the damage model proposed here as shown in Figure 11 is developed on the basis of failure modes observed from the cross sectional defects. For interfacial micro-cracking in a layer, most of this failure mode was observed at the impact energy less than 20.86 J. However, as the impact energy increases, the failure mode becomes severe and arrived at the parallel cracks. This was observed at the impact energies between 20.86 J and 31.29 J. When fibre failures are predicted, the damaged area becomes larger and this can be obviously seen for specimens that were damaged at the impact energy above 31.29 J. It was found that, as the impact energy increases, these parallel cracks join and produce a single catastrophic crack. This trend was observed for all the layers.

Figure 11.

6. Conclusion

The fractographic research presented here clearly shows the failure modes such as matrix damage and fibre failure in CFRP on its impact damage strength. The degradation in material strength and stiffness due to the existence of damage is in line with the damage behaviour of most of the composite materials. The damage propagation described here clarifies not only how the damage initiates and grows, but also describes the phenomenon of damage progression and its final failure.

The main objective of this study was to perform a series of low energy impacts in carbon fiber reinforced polymer composites, and then to carry out a sequence of SEM investigations to evaluate the failure modes. From the work conducted, it can be concluded that (in agreement with previous fundamental research), as both impact energy and peak force detected increase, the number of distinct failure modes also increases. The experimental results allowed the identification of three critical impact energy thresholds. The threshold for matrix cracking and matrix breakage was identified as below the impact energy of 20.86 J whilst for fiber cracking it was identified between 31.29 J and 20.86 J. It is worth stating that the threshold energy for major damage was identified when the specimens were subjected to impact energy of more than 33.89 J.

A damage model was developed from this research work. It can provide information on the types and extent of damage. As a result, the model can be use to provide fundamental understanding of damage and failure mode progression in carbon fibre reinforced composites under varying layer thickness and impact energies.