Experimental Evaluation of Low Velocity Impact Properties and Damage Progression on Bamboo/Glass Hybrid Composites Subjected to Different Impact Energy Levels

Six impact energy values, ranging from 2.5 J to 10 J, were applied to study the impact properties of neat epoxy and bamboo composites, while six impact energy values, ranging from 10 J to 35 J, were applied on bamboo/glass hybrid composites. Woven glass fibre was embedded at the outermost top and bottom layer of bamboo powder-filled epoxy composites, producing sandwich structured hybrid composites through lay-up and molding techniques. A drop weight impact test was performed to study the impact properties. A peak force analysis showed that neat epoxy has the stiffest projectile for targeting interaction, while inconsistent peak force data was collected for the non-hybrid composites. The non-hybrid composites could withstand up to 10 J, while the hybrid composites showed a total failure at 35 J. It can be concluded that increasing the filler loading lessened the severity of damages in non-hybrid composites, while introducing the woven glass fibre could slow down the penetration of the impactor, thus lowering the chances of a total failure of the composites.


Introduction
Generally, mechanical properties are the most important information that needs to be measured in materials [1]. However, in real life situations, impact is one of the very common phenomena experienced by all materials and structures. An instantaneous load applied on a surface during an impact event can cause unpredictable damages, which can sometimes lead to total structural failure. It is even worse if the low velocity impact event caused non-visible damages; the accumulated damages after several repeated events can lead to serious failures [2,3]. Experimentally, low velocity impact can be simulated using the drop test rig instrument. The Izod and Charpy impact testers are more descriptive of fracture toughness, which can be considered as the mechanical properties of materials [4].
It was cited from one source that low velocity impact is an impact event below 10 m/s, while intermediate, high and hypervelocity impacts correspond to the range of 10 m/s-50 m/s, 50 m/s-1000 m/s and 2 km/s-5 km/s respectively. Different ranges of impact velocities are very important to analyse, rather than saying that only high velocity needs more attention, as each structure has its own surroundings and working environment [5]. Different mechanisms of damage initiations

Characterisation of Impact Properties
The low velocity impact was simulated using an instrumented drop weight impact test machine, model IMATEK IM10, at the Faculty of Engineering, Universiti Putra Malaysia. The impact testing was conducted with five repeatability samples. The instrument was equipped with IMATEK Impact Analysis software, to record and process the impact results data. A hemispherical tip striker with a radius of 5 mm attached to a variable weight resulting in a total weight of 5.101 kg was dropped from several desired heights onto the clamped sample. Different heights were used to represent different magnitudes of the impact energy applied onto the samples, based on the following equation: where m is the total mass of the impactor, 5.101 kg; g is the gravitational acceleration, 9.81 m/s 2 ; and h is the height of the impactor. The raw data of the force, time, displacement, velocity and energy absorbed by the samples were recorded and calculated by the installed software. Table 2 shows the impact energy applied on the EP, EP-BF composites and EP/G-BF composites. Table 2. Different impact energies applied on the EP, EP-BF composites and EP/G-BF composites.

Composites EP and EP-BF Composites EP/G-BF Composites
Impact Energy (J) 2. Different ranges of impact energy were applied to the non-hybrid and hybrid composites so as to achieve the maximum impact energy withstood by each type of composite.

Characterisation of Impact Damages
The dye penetrant method was applied to observe the damages on the impacted samples. The length of the matrix cracking was measured to characterise the damage propagation as the response to different impact energies on the EP-BF composites, while the area of damage was measured for the EP/G-BF composites [21].

Force Displacement Analysis
Important information regarding the damage progression within the sample during an impact event can be obtained from the force-displacement graph. The movement of the impactor and the deformation of the impacted surface of the sample during contact with the impactor are marked as displacement values in the graph [22]. Figure 1 shows the force-displacement graphs of EP at the lowest impact energy level of 2.5 J.

Force Displacement Analysis
Important information regarding the damage progression within the sample during an impact event can be obtained from the force-displacement graph. The movement of the impactor and the deformation of the impacted surface of the sample during contact with the impactor are marked as displacement values in the graph [22]. Figure 1 shows the force-displacement graphs of EP at the lowest impact energy level of 2.5 J. The overlapping of the three graphs in the figure shows the repeatability of the three samples for EP under the same magnitude of impact energy. Testing conducted on all samples, for each impact energy level, was also repeated three times to confirm the repeatability of the results. The closed curve from the force-displacement graph indicates the non-full penetration damage of the sample tested, which indirectly explains that the full penetration of the impactor into the sample will produce an open curve in the force-displacement graph [23].
The ascending and descending parts of the closed curve explain the loading and unloading conditions, respectively. The ascending part also provides information about the impact bending stiffness of the samples. The greater the peak force, the stiffer the projectile-to-target interaction, thus shortening the contact period of the impactor onto the surface [24]. The relationship between a greater peak force and a stiffer target is a representation of a situation in which greater force is needed to initiate damage in stiffer materials. Besides the target stiffness, the peak force also depends on the magnitude of the impact energy from the impactor.
The peak deflection or peak deformation occurring in the sample is the value of the peak displacement from the force-displacement graph. In most graphs, the point of the peak deformation value is almost the same as the point for the peak force. However, it is clearly understood that the peak deformation can be identified as the turning point at which the force curve returns to zero after the loading condition or ascending curve, while the peak force is the maximum value in the vertical direction of the graph. It is clear that both the peak force and the peak deformation values were obtained from two different points [25].
Another important value extracted from the force-displacement graph is the energy absorbed by the samples, which can be determined from the area under this graph. The energy absorbed is the kinetic energy transferred from the impactor to the samples during impact [24]. Both the energy absorbed throughout the impact event and the energy absorbed up to the peak deformation can be obtained from the force-displacement graph as shown in Figure 1. An example of a forcedisplacement graph for fully penetrated or perforated samples during impact can be seen in Figure  2. The overlapping of the three graphs in the figure shows the repeatability of the three samples for EP under the same magnitude of impact energy. Testing conducted on all samples, for each impact energy level, was also repeated three times to confirm the repeatability of the results. The closed curve from the force-displacement graph indicates the non-full penetration damage of the sample tested, which indirectly explains that the full penetration of the impactor into the sample will produce an open curve in the force-displacement graph [23].
The ascending and descending parts of the closed curve explain the loading and unloading conditions, respectively. The ascending part also provides information about the impact bending stiffness of the samples. The greater the peak force, the stiffer the projectile-to-target interaction, thus shortening the contact period of the impactor onto the surface [24]. The relationship between a greater peak force and a stiffer target is a representation of a situation in which greater force is needed to initiate damage in stiffer materials. Besides the target stiffness, the peak force also depends on the magnitude of the impact energy from the impactor.
The peak deflection or peak deformation occurring in the sample is the value of the peak displacement from the force-displacement graph. In most graphs, the point of the peak deformation value is almost the same as the point for the peak force. However, it is clearly understood that the peak deformation can be identified as the turning point at which the force curve returns to zero after the loading condition or ascending curve, while the peak force is the maximum value in the vertical direction of the graph. It is clear that both the peak force and the peak deformation values were obtained from two different points [25].
Another important value extracted from the force-displacement graph is the energy absorbed by the samples, which can be determined from the area under this graph. The energy absorbed is the kinetic energy transferred from the impactor to the samples during impact [24]. Both the energy absorbed throughout the impact event and the energy absorbed up to the peak deformation can be obtained from the force-displacement graph as shown in Figure 1. An example of a force-displacement graph for fully penetrated or perforated samples during impact can be seen in Figure 2. The open curves show that the displacement increased monotonically with a decreasing force. The penetration of the samples demonstrates a situation where the force applied exceeded the maximum allowable force for the samples [24]. Throughout the low velocity impact analysis, the full penetration events will not be discussed. Therefore, the force displacement graphs of all the samples were first analysed to exclude the data for the penetrated samples. This explained the different maximum impact energy levels for each type of sample discussed in this study. The perfect overlapping graphs in Figure 2 confirmed the repeatability of the tests and the consistency of the samples. The excluded data for the broken samples was confirmed after these three repeatability tests.
It can be seen that different samples can withstand different maximum impact energy levels. Obviously, the second range of impact energy, as listed in Table 2, was higher when compared to the first range, as the inclusion of woven glass fibre was assumed to slow down the penetration of the impactor, thus increasing the maximum allowable force impacted on the surface of the composites. For all composites, the data analysed from the force-displacement graphs is presented in Tables 3 and  4, which list the first and second ranges of impact energy applied on the samples, respectively. The relationship between the tabulated values will be further discussed in the following sections.  The open curves show that the displacement increased monotonically with a decreasing force. The penetration of the samples demonstrates a situation where the force applied exceeded the maximum allowable force for the samples [24]. Throughout the low velocity impact analysis, the full penetration events will not be discussed. Therefore, the force displacement graphs of all the samples were first analysed to exclude the data for the penetrated samples. This explained the different maximum impact energy levels for each type of sample discussed in this study. The perfect overlapping graphs in Figure 2 confirmed the repeatability of the tests and the consistency of the samples. The excluded data for the broken samples was confirmed after these three repeatability tests.
It can be seen that different samples can withstand different maximum impact energy levels. Obviously, the second range of impact energy, as listed in Table 2, was higher when compared to the first range, as the inclusion of woven glass fibre was assumed to slow down the penetration of the impactor, thus increasing the maximum allowable force impacted on the surface of the composites. For all composites, the data analysed from the force-displacement graphs is presented in Tables 3 and 4, which list the first and second ranges of impact energy applied on the samples, respectively. The relationship between the tabulated values will be further discussed in the following sections. Figure 3 shows the variation of the peak force for the EP and EP-BF composites at the first range of impact energy levels.

Peak Force Variation with Impact Energy
In Figure 3, it can be seen that the EP has the highest peak force when compared to EP-BF10 and EP-BF30 for all impact energy levels. This leads to the first conclusion that epoxy has the stiffest projectile-to-target interaction [24]. Comparing EP-BF10 and EP-BF30 gives a different trend at different impact energy levels. At an impact energy of 2.5 J and 4.4 J, the EP-BF10 shows a higher peak force compared to EP-BF30. However, at an impact energy of 3.75 J, the EP-BF10 has a lower peak force compared to EP-BF30 with a significant value. These inconsistent values were caused by the random orientation of the bamboo powder in the epoxy and the agglomeration of powder that might be occurring, which results from the poor distribution of the bamboo powder during fabrication [18].
Based on the first conclusion for epoxy, a comparison of the different loadings of bamboo composites made at impact energies of 2.5 J and 4.4 J is more consistent when compared to the comparison made at an impact energy of 3.75 J; these comparisons show that lower bamboo filler loading gives a stiffer projectile-to-target interaction [24]. However, at an impact energy of 5 J, the EP-BF30 could still withstand the force without breaking, while the EP-BF10 experienced total damage. This situation shows that EP-BF30 has good strength but lower stiffness, while the EP-BF10 shows the opposite trend. In the second range of impact energy levels, the hybrid composites respond consistently to the force applied, as illustrated in Figure 4.  10.00 ---  Figure 3 shows the variation of the peak force for the EP and EP-BF composites at the first range of impact energy levels.

Peak Force Variation with Impact Energy
. Figure 3. Variation of the peak force for the EP and EP-BF composites at the first range of impact energy levels.
In Figure 3, it can be seen that the EP has the highest peak force when compared to EP-BF10 and EP-BF30 for all impact energy levels. This leads to the first conclusion that epoxy has the stiffest projectile-to-target interaction [24]. Comparing EP-BF10 and EP-BF30 gives a different trend at EP-BF30 could still withstand the force without breaking, while the EP-BF10 experienced total damage. This situation shows that EP-BF30 has good strength but lower stiffness, while the EP-BF10 shows the opposite trend. In the second range of impact energy levels, the hybrid composites respond consistently to the force applied, as illustrated in Figure 4. The peak force increased as the impact energy increased for both loadings of hybrid composites. The inclusion of woven-type glass fibre on the outermost surface of the composites helps with a better force distribution when compared to the random orientation of fibre, resulting in a more consistent data trend [22]. The highest impact energy marked from the first range of impact energy levels in Figure 3, which was 10 J, was the lowest value in the second range of impact energy levels in Figure  4. At this impact energy, only the EP was comparable to the EP/G-BFC, while the EP-BFC samples failed and experienced total damage. From the values shown in both Figures 3 and 4, the hybrid composites tend to have a higher peak force to initiate damages as expected. Moreover, woven-type glass fibres are good in impact resistant, and it is expected that a woven-type fibre of any material will have better resistance towards impact when compared to other types of fibre such as random and unidirectional fibres [26].

Energy Absorbed Variation with Impact Energy
The impact energy supplied during an impact event is converted into two fractions, which are the loss elastic energy and the energy absorbed by the sample. The absorbed energy is presented by the damage mechanisms on the sample or structure, where more severe damage can be an indication of more energy being absorbed [24,26]. The severity of the damage is subject to the mechanical properties of the reinforcement and the matrix, the shape of the impactor tip, the fibre orientation, the sample's geometry and the impact energy levels [21].
Each sample absorbed a different amount of energy at each impact energy level, and thus a direct relationship of impact energy with the amount of absorbed energy was not advisable for explaining the severity of the damage on the samples. It is understood that a higher impact energy will cause a higher amount of impact energy to be absorbed, as depicted in Tables 1 and 2 [25]. Therefore, the percentage of energy absorbed by each sample at the respective impact energy level presents a better relationship of the impact energy with the absorbed energy and explains the severity of damage on the samples.  The peak force increased as the impact energy increased for both loadings of hybrid composites. The inclusion of woven-type glass fibre on the outermost surface of the composites helps with a better force distribution when compared to the random orientation of fibre, resulting in a more consistent data trend [22]. The highest impact energy marked from the first range of impact energy levels in Figure 3, which was 10 J, was the lowest value in the second range of impact energy levels in Figure 4. At this impact energy, only the EP was comparable to the EP/G-BFC, while the EP-BFC samples failed and experienced total damage. From the values shown in both Figures 3 and 4, the hybrid composites tend to have a higher peak force to initiate damages as expected. Moreover, woven-type glass fibres are good in impact resistant, and it is expected that a woven-type fibre of any material will have better resistance towards impact when compared to other types of fibre such as random and unidirectional fibres [26].

Energy Absorbed Variation with Impact Energy
The impact energy supplied during an impact event is converted into two fractions, which are the loss elastic energy and the energy absorbed by the sample. The absorbed energy is presented by the damage mechanisms on the sample or structure, where more severe damage can be an indication of more energy being absorbed [24,26]. The severity of the damage is subject to the mechanical properties of the reinforcement and the matrix, the shape of the impactor tip, the fibre orientation, the sample's geometry and the impact energy levels [21].
Each sample absorbed a different amount of energy at each impact energy level, and thus a direct relationship of impact energy with the amount of absorbed energy was not advisable for explaining the severity of the damage on the samples. It is understood that a higher impact energy will cause a higher amount of impact energy to be absorbed, as depicted in Tables 1 and 2 [25]. Therefore, the percentage of energy absorbed by each sample at the respective impact energy level presents a better relationship of the impact energy with the absorbed energy and explains the severity of damage on the samples. Figures 5 and 6 present the percentage of energy absorbed by the EP, EP-BF and EP/G-BF composites, respectively.
It was seen that an inconsistent trend was represented by the EP-BF composites compared to the EP in Figure 5. This is due to the random orientation of the short bamboo fibres, since the force being applied cannot be distributed evenly. At the same time, the agglomeration of bamboo powder that might be located within the structure tends to either slow down the damage propagation or worsen the damage by integrating a larger surface damage [8,13]. The inconsistent response of randomly oriented bamboo composites towards impact causes an inconsistent energy to be absorbed by the samples. This concept is illustrated in Figure 7.  It was seen that an inconsistent trend was represented by the EP-BF composites compared to the EP in Figure 5. This is due to the random orientation of the short bamboo fibres, since the force being applied cannot be distributed evenly. At the same time, the agglomeration of bamboo powder that might be located within the structure tends to either slow down the damage propagation or worsen the damage by integrating a larger surface damage [8,13]. The inconsistent response of randomly oriented bamboo composites towards impact causes an inconsistent energy to be absorbed by the samples. This concept is illustrated in Figure 7.   It was seen that an inconsistent trend was represented by the EP-BF composites compared to the EP in Figure 5. This is due to the random orientation of the short bamboo fibres, since the force being applied cannot be distributed evenly. At the same time, the agglomeration of bamboo powder that might be located within the structure tends to either slow down the damage propagation or worsen the damage by integrating a larger surface damage [8,13]. The inconsistent response of randomly oriented bamboo composites towards impact causes an inconsistent energy to be absorbed by the samples. This concept is illustrated in Figure 7.    It was seen that an inconsistent trend was represented by the EP-BF composites compared to the EP in Figure 5. This is due to the random orientation of the short bamboo fibres, since the force being applied cannot be distributed evenly. At the same time, the agglomeration of bamboo powder that might be located within the structure tends to either slow down the damage propagation or worsen the damage by integrating a larger surface damage [8,13]. The inconsistent response of randomly oriented bamboo composites towards impact causes an inconsistent energy to be absorbed by the samples. This concept is illustrated in Figure 7.  Compared to non-hybrid composites, a clearer relationship between the percentage of energy absorbed and the different impact energy values of hybrid composites can be seen in Figure 6. As the impact energy increased, the percentage of energy absorbed increased. Besides this, at all impact energy levels, the percentage of energy absorbed for the EP/G-BF30 is lower than for the EP/G-BF10. This is in good agreement with the damage found on the samples, where EP/G-BF10 experienced more severe damage when compared to EP/G-BF30.

Damage Analysis on The Impacted Samples
The damage analysis will be separated based on the types of samples and their response towards different impact energy levels. Figure 8 shows the damages on the EP samples. This is in good agreement with the damage found on the samples, where EP/G-BF10 experienced more severe damage when compared to EP/G-BF30.

Damage Analysis on The Impacted Samples
The damage analysis will be separated based on the types of samples and their response towards different impact energy levels. Figure 8 shows the damages on the EP samples. From the lowest energy values of 2.50 J to 5.00 J, the samples were free from cracks, and no dented surface was visible. However, at an impact energy level of 10.00 J, the EP sample was totally broken. This observation leads to the conclusion that neat polymer plates will experience total failure at a certain impact energy level without experiencing minor cracks [21]. This situation therefore lowers the dependency and safety of a product in real life applications. It might withstand a higher load, but an unexpected total failure might happen at any limit without any preliminary sign of damage.
The damage on the non-hybrid EP-BF composites after the low velocity impact was observed with the aid of the dye penetrant and is shown in Figures 9 and 10.  From the lowest energy values of 2.50 J to 5.00 J, the samples were free from cracks, and no dented surface was visible. However, at an impact energy level of 10.00 J, the EP sample was totally broken. This observation leads to the conclusion that neat polymer plates will experience total failure at a certain impact energy level without experiencing minor cracks [21]. This situation therefore lowers the dependency and safety of a product in real life applications. It might withstand a higher load, but an unexpected total failure might happen at any limit without any preliminary sign of damage.
The damage on the non-hybrid EP-BF composites after the low velocity impact was observed with the aid of the dye penetrant and is shown in Figures 9 and 10.
As depicted in Figure 9, matrix cracking was detected on the EP-BF10 samples at 2.50 J, 3.75 J and 4.40 J impact energy levels. The matrix cracking propagates from the top surface to the bottom, and from the centre (where the impactor was dropped) to the sides of the rectangular plates. The EP-BF10 samples can withstand the first three impact energy levels; the propagation of matrix cracking stopped before reaching the sides of the samples. However, 10.00 J of impact energy enabled the matrix cracking to propagate until the end sides of the samples, thus breaking the rectangular plates into pieces. Compared to the visual look of the broken EP samples in Figure 9, the broken sample of EP-BF10 indicates a less severe damage as it broke into four large pieces that can be laid out, instead of numerous smaller pieces. Micro-sized fibres with a random orientation limit the analysis of the damage, as no trend can be suggested concerning the relationship of the impact energy levels with the severity of the damage caused [27].
Polymers 2020, 12, x FOR PEER REVIEW 9 of 14 Compared to non-hybrid composites, a clearer relationship between the percentage of energy absorbed and the different impact energy values of hybrid composites can be seen in Figure 6. As the impact energy increased, the percentage of energy absorbed increased. Besides this, at all impact energy levels, the percentage of energy absorbed for the EP/G-BF30 is lower than for the EP/G-BF10. This is in good agreement with the damage found on the samples, where EP/G-BF10 experienced more severe damage when compared to EP/G-BF30.

Damage Analysis on The Impacted Samples
The damage analysis will be separated based on the types of samples and their response towards different impact energy levels. Figure 8 shows the damages on the EP samples. From the lowest energy values of 2.50 J to 5.00 J, the samples were free from cracks, and no dented surface was visible. However, at an impact energy level of 10.00 J, the EP sample was totally broken. This observation leads to the conclusion that neat polymer plates will experience total failure at a certain impact energy level without experiencing minor cracks [21]. This situation therefore lowers the dependency and safety of a product in real life applications. It might withstand a higher load, but an unexpected total failure might happen at any limit without any preliminary sign of damage.
The damage on the non-hybrid EP-BF composites after the low velocity impact was observed with the aid of the dye penetrant and is shown in Figures 9 and 10.  Energy levels (J) 2.50 3.75 4.40 5.00 Distance (mm) 53 61 70 break Similar damage behaviour was observed on the EP-BF30 samples compared to the previous EP-BF10. Dye penetrant was used, and the observation of the damage is shown in Figure 10. Matrix cracking was detected as propagating from the top to the bottom and from the centre to the sides of the impacted samples. However, small differences can be seen between the damage on the EP-BF30 and that on the EP-BF10, i.e., the number of lines of matrix cracking is lower than the number found on the EP-BF10. For the EP-BF30, only two obvious lines of matrix cracking were observed from the centre of impact, while in the EP-BF10 the damage propagated into four observable lines from the centre of impact. Figure 10 shows that the EP-BF30 withstood the 5.00 J impact energy Although no trend can be suggested regarding the propagation of matrix cracking, the analysis was presented in terms of the distance of the matrix cracking propagation from the centre to the side of the rectangular samples. It was found that the distance increased as the impact energy increased from 2.50 J to 4.40 J. These distances were measured on the bottom surface of the EP-BF10 samples, and the longest distance from the centre is recorded in Table 5. Similar damage behaviour was observed on the EP-BF30 samples compared to the previous EP-BF10. Dye penetrant was used, and the observation of the damage is shown in Figure 10.
Matrix cracking was detected as propagating from the top to the bottom and from the centre to the sides of the impacted samples. However, small differences can be seen between the damage on the EP-BF30 and that on the EP-BF10, i.e., the number of lines of matrix cracking is lower than the number found on the EP-BF10. For the EP-BF30, only two obvious lines of matrix cracking were observed from the centre of impact, while in the EP-BF10 the damage propagated into four observable lines from the centre of impact. Figure 10 shows that the EP-BF30 withstood the 5.00 J impact energy but failed at 10.00 J, which is one level higher than for the EP-BF10. The damage analysis for the EP-BF30 is presented in Table 6. The shorter distance of the damage propagation for the EP-BF30 suggested that increasing the bamboo fibre loading in the epoxy matrix improved the impact resistance of the composites. The impact resistance of the hybrid EP/G-BFC was expected to be higher compared to the non-hybrid EP-BF composites. The inclusion of woven glass fibres at the top and bottom outermost layers of the composites was believed to help in slowing down the impact absorption into the plates, thus reducing the damage [13]. Figures 11 and 12 show the damage propagation on the hybrid EP/G-BF10 and on the EP/G-BF30, respectively. Dye penetrant was not used as the damage is clearly visible and the area can directly be calculated from the surface of the impacted samples. Energy levels (J) 2.50 3.75 4.40 5.00 10.00 Distance (mm) 27 40 59 63 break The shorter distance of the damage propagation for the EP-BF30 suggested that increasing the bamboo fibre loading in the epoxy matrix improved the impact resistance of the composites. The impact resistance of the hybrid EP/G-BFC was expected to be higher compared to the non-hybrid EP-BF composites. The inclusion of woven glass fibres at the top and bottom outermost layers of the composites was believed to help in slowing down the impact absorption into the plates, thus reducing the damage [13]. Figures 11 and 12 show the damage propagation on the hybrid EP/G-BF10 and on the EP/G-BF30, respectively. Dye penetrant was not used as the damage is clearly visible and the area can directly be calculated from the surface of the impacted samples.    Energy levels (J) 2.50 3.75 4.40 5.00 10.00 Distance (mm) 27 40 59 63 break The shorter distance of the damage propagation for the EP-BF30 suggested that increasing the bamboo fibre loading in the epoxy matrix improved the impact resistance of the composites. The impact resistance of the hybrid EP/G-BFC was expected to be higher compared to the non-hybrid EP-BF composites. The inclusion of woven glass fibres at the top and bottom outermost layers of the composites was believed to help in slowing down the impact absorption into the plates, thus reducing the damage [13]. Figures 11 and 12 show the damage propagation on the hybrid EP/G-BF10 and on the EP/G-BF30, respectively. Dye penetrant was not used as the damage is clearly visible and the area can directly be calculated from the surface of the impacted samples.   Generally, the damaged areas on both surfaces, top and bottom, were seen to increase as the impact energy level increased. A significant difference in the area was clearly observed on the top surfaces, while on the bottom perforation was detected [24,26]. Compared to the non-hybrid EP-BFC, the EP/G-BFC samples did not break into pieces during the full penetration event. The inclusion of woven glass fibres improved the properties of the composites in terms of impact damage resistance. This improvement is very beneficial for real life applications as the severity of failure is lower when compared to the total failure in the non-hybrid composites. The woven glass fibres can hold the structure in one piece during the highest impact incident. The damage area of the hybrid EP/G-BFC is presented in Table 7. For both types of samples, EP/G-BF10 and EP/G-BF30, the damage area increased significantly from 10 J to 20 J, and a smaller difference was calculated as the energy level increased further. From 20 J to 35 J, the damage was seen to propagate more towards the bottom surface compared to the propagation from the centre to the sides of the samples.

Conclusions
The greater the peak force, the stiffer the projectile-to-target interaction, thus shortening the contact period of the impactor on the surface of the composites. The non-hybrid EP-BF10 composites have a stiffer projectile-to-target interaction compared to the EP-BF30 composites. However, neat epoxy samples showed the stiffest projectile-to-target interaction among the three samples, EP, EP-BF10, EP-BF30, at all impact energy levels. The non-hybrid EP-BF10 composites exhibited good stiffness but lower strength, while the EP-BF30 composites showed the opposite relation. Damage initiation and propagation in the EP-BF30 composites was slower and less severe when compared to the EP-BF10 composites. The non-visible damage in the bamboo composites, which occurred after the low velocity impact, can be analysed through the force undulations in the force-time graphs and can be observed using the dye penetrant method.
A significant improvement was observed with the inclusion of woven glass fibres in the composites. The non-hybrid composites broke into pieces during the highest impact energy that is applied, while the hybrid composites experienced only perforation and the structure did not totally break. The distance of the matrix cracking was shorter for the EP-BF30 when compared to the EP-BF10 at the same impact energy level, suggesting that increasing the bamboo fibre loading can improve the impact resistance, although in short fibre-reinforced composites.