Effect of Fibres on the Failure Mechanism of Composite Tubes under Low-Velocity Impact

Filament-wound composite tubular structures are frequently used in transmission systems, pressure vessels, and sports equipment. In this study, the failure mechanism of composite tubes reinforced with different fibres under low-velocity impact (LVI) and the radial residual compression performance of the impacted composite tubes were investigated. Four fibres, including carbon fiber-T800, carbon fiber-T700, basalt fibre, and glass fibre, were used to fabricate the composite tubes by the winding process. The internal matrix/fibre interface of the composite tubes before the LVI and their failure mechanism after the LVI were investigated by scanning electric microscopy and X-ray micro-computed tomography, respectively. The results showed that the composite tubes mainly fractured through the delamination and fibre breakage damage under the impact of 15 J energy. Delamination and localized fibre breakage occur in the glass fibre-reinforced composite (GFRP) and basalt fibre-reinforced composite (BFRP) tubes when subjected to LVI. While fibre breakage damage occurs globally in the carbon fibre-reinforced composite (CFRP) tubes. The GFRP tube showed the best impact resistance among all the tubes investigated. The basalt fibre-reinforced composite (BFRP) tube exhibited the lowest structural impact resistance. The impact resistance of the CFRP-T700 and CFRP-T800 tube differed slightly. The radial residual compression strength (R-RCS) of the BFRP tube is not sensitive to the impact, while that of the GFRP tube is shown to be highly sensitive to the impact.

square hollow section steel columns subjected to transverse impact to predict their behaviour and failure mechanism. The results showed that the strengthening method improved the impact resistance by reducing the lateral displacement of the strengthened column by approximately 58% compared with that of the bare steel column.
Owing to the particularity of the structure of the composite tubes, the factors affecting their impact resistance, including the material properties of the reinforced material, matrix material, structural properties of the layup mode, and tube size, are also very complex. These have not been sufficiently studied thus far. Mokhtar et al. [21] proposed that the maximum damage diameter of the basalt composite pipes increases significantly with the increase in the winding angle and impact energy, and the shape of the damaged area changes accordingly. The basalt composite pipes with a large winding angle absorb less energy. Liu et al. [7] reported that tubes with mixed ply angles exhibit better impact resistance in terms of energy absorption. Hassan et al. [22] proposed the study of LVI damage of woven fabric composites for different specimen thicknesses via finite element simulation and experimental verification. The simulation and experimental results were in agreement in terms of maximum contact force and time.
The above studies mainly focused on the impact response of composite tubes or metal/composite tubes to the environmental factors (temperature, humidity, and heat, etc.), impact energy, structural parameters, and other conditions. The effect of material properties (especially the reinforced fibre of the composites), on the impact performance and even the failure mechanism during the LVI have been investigated in few studies. Moreover, most of the residual mechanical strength after impact damage is based on the bending and torsional properties of the tubes; the radial residual compression strength (R-RCS) of the composite tubular structures, which is essential for pressure tubes, has rarely been reported. Herein, LVI damage responses and R-RCS of the four types of fibre (including carbon fibre-T800, carbon fibre-T700, basalt fibre and glass fibre) reinforced epoxy resin matrix composite tubes are investigated and discussed; computed tomography (CT) and scanning electric microscopy (SEM) are utilised to understand the inner damage and failure mechanism. This work aims to compare the impact behaviours of different fibre-reinforced composite tubes and understand how fibres affect the failure mechanisms of composite tubular structures. This may indirectly help in choosing and designing the materials to improve the anti-impact performance of composite tubular structures.

Fabrication of Composite Tubes by Wet-Filament Winding
The wet-filament winding method is generally used to fabricate symmetric structures with rotating bodies and positive curvatures. The process flow diagram is shown in Figure 1.  The material parameters of the FRP materials are listed in Table 2. The material parameters including Young's modulus, shear modulus, and Poisson's ratio have been measured according to GB/T 1458-2008. The composite tubes with a length of 1200 mm were prepared by filament winding processing, as shown in Figure 2. A cutting machine was applied to cut the tubes into short tubes of 100 mm length, 70 mm inner diameter, and about 3.0 mm thickness. The curing stage of the resin system played an important role in the moulding process. The subsection heating scheme shown in Figure 3 was set as the optimal heating solution. In this study, carbon fibre-T800, carbon fibre-T700, basalt fibre and glass fibre are the four types of fibres that were each wound with epoxy resin to fabricate a total of four composite tubes with a winding angle of [±25 4 /90 2 ]. The mechanical properties of the four types of fibres are listed in Table 1. The material parameters of the FRP materials are listed in Table 2. The material parameters including Young's modulus, shear modulus, and Poisson's ratio have been measured according to GB/T 1458-2008. The composite tubes with a length of 1200 mm were prepared by filament winding processing, as shown in Figure 2. A cutting machine was applied to cut the tubes into short tubes of 100 mm length, 70 mm inner diameter, and about 3.0 mm thickness.  The material parameters of the FRP materials are listed in Table 2. The material parameters including Young's modulus, shear modulus, and Poisson's ratio have been measured according to GB/T 1458-2008. The composite tubes with a length of 1200 mm were prepared by filament winding processing, as shown in Figure 2. A cutting machine was applied to cut the tubes into short tubes of 100 mm length, 70 mm inner diameter, and about 3.0 mm thickness. The curing stage of the resin system played an important role in the moulding process. The subsection heating scheme shown in Figure 3 was set as the optimal heating solution.  The curing stage of the resin system played an important role in the moulding process. The subsection heating scheme shown in Figure 3 was set as the optimal heating solution.

Low-Velocity Impact Tests
To simulate the LVI process on the composite tubes, the MTS ZCJ1302-AD drop hammer tester (MTS Industrial Systems (China) Co., LTD, Shenzhen, China) was used to affect the specimens ( Figure 4). The impactor had a hemispherical tip with a mass of 2 kg and a diameter of 12.5 mm. A PCB Quartz ICP force sensor (PCB Piezotronics, Inc.Buffalo, New York, NY, USA) was used to measure the force-time responses (in terms of the test time, impact force, impactor velocity, displacement, and energy) of the samples. The impactor was dropped from a certain height to achieve the impact energy level of 15 J by striking once only. During the experiment, large deformation was observed in the short tubes subjected to impact, and layer overflow also occurred easily. This contributed to the lack of restraints at both ends. However, this does not agree with the actual working conditions. To address this problem, metal kits ( Figure 5a) were installed on both ends of the tubes as shown in Figure 5b. During the LVI tests, composite tubes were placed on the V-groove to restrain the slippage of the tubes and improve the accuracy, as shown in Figure 5c.

Low-Velocity Impact Tests
To simulate the LVI process on the composite tubes, the MTS ZCJ1302-AD drop hammer tester (MTS Industrial Systems (China) Co., LTD, Shenzhen, China) was used to affect the specimens ( Figure 4). The impactor had a hemispherical tip with a mass of 2 kg and a diameter of 12.5 mm. A PCB Quartz ICP force sensor (PCB Piezotronics, Inc.Buffalo, New York, NY, USA) was used to measure the force-time responses (in terms of the test time, impact force, impactor velocity, displacement, and energy) of the samples. The impactor was dropped from a certain height to achieve the impact energy level of 15 J by striking once only.

Low-Velocity Impact Tests
To simulate the LVI process on the composite tubes, the MTS ZCJ1302-AD drop hammer tester (MTS Industrial Systems (China) Co., LTD, Shenzhen, China) was used to affect the specimens ( Figure 4). The impactor had a hemispherical tip with a mass of 2 kg and a diameter of 12.5 mm. A PCB Quartz ICP force sensor (PCB Piezotronics, Inc.Buffalo, New York, NY, USA) was used to measure the force-time responses (in terms of the test time, impact force, impactor velocity, displacement, and energy) of the samples. The impactor was dropped from a certain height to achieve the impact energy level of 15 J by striking once only. During the experiment, large deformation was observed in the short tubes subjected to impact, and layer overflow also occurred easily. This contributed to the lack of restraints at both ends. However, this does not agree with the actual working conditions. To address this problem, metal kits ( Figure 5a) were installed on both ends of the tubes as shown in Figure 5b. During the LVI tests, composite tubes were placed on the V-groove to restrain the slippage of the tubes and improve the accuracy, as shown in Figure 5c. During the experiment, large deformation was observed in the short tubes subjected to impact, and layer overflow also occurred easily. This contributed to the lack of restraints at both ends. However, this does not agree with the actual working conditions. To address this problem, metal kits ( Figure 5a) were installed on both ends of the tubes as shown in Figure 5b. During the LVI tests, composite tubes were placed on the V-groove to restrain the slippage of the tubes and improve the accuracy, as shown in Figure 5c.

Damage Characterization
An X-ray micro-CT system (GE, phoenix v|tome|x m, GE Sensing & Inspection Technologies, Cincinnati, OH, USA) was employed to characterise the internal damages of the composite tube after the LVI and R-RCS tests. The tomography system equipped with a 300 kV micro focus X-ray tube and a highly dynamic DXR digital detector array provided detailed three-dimensional information about the specimens.

Radial Residual Compression Strength Test
R-RCS tests were conducted to study the radial compressive strength loss of the composite tubes tested on the LD23 universal testing machine (Shenzhen Rambo Sansi Material Testing Co. LTD, Shenzhen, China). A constant displacement rate of 10 mm/min was employed and the tests were carried out according to GBT 5352-2005 standard. As shown in Figure 6, the composite tubes underwent radial compression between two parallel load plates with the dimensions of 200 mm × 150 mm × 30 mm. The impact position was located along the largest transverse size. The experiment was stopped when the deformation rate of the diameter reached 60% or the structure was clearly damaged. The deformation rate P was calculated using the following equation: where ∆Y is the change in the variation value of the diameter and d is the mean inner diameter of the composite tube.

Damage Characterization
An X-ray micro-CT system (GE, phoenix v|tome|x m, GE Sensing & Inspection Technologies, Cincinnati, OH, USA) was employed to characterise the internal damages of the composite tube after the LVI and R-RCS tests. The tomography system equipped with a 300 kV micro focus X-ray tube and a highly dynamic DXR digital detector array provided detailed three-dimensional information about the specimens.

Radial Residual Compression Strength Test
R-RCS tests were conducted to study the radial compressive strength loss of the composite tubes tested on the LD23 universal testing machine (Shenzhen Rambo Sansi Material Testing Co. LTD, Shenzhen, China). A constant displacement rate of 10 mm/min was employed and the tests were carried out according to GBT 5352-2005 standard. As shown in Figure 6, the composite tubes underwent radial compression between two parallel load plates with the dimensions of 200 mm × 150 mm × 30 mm. The impact position was located along the largest transverse size. The experiment was stopped when the deformation rate of the diameter reached 60% or the structure was clearly damaged. The deformation rate P was calculated using the following equation: where ∆Y is the change in the variation value of the diameter and d is the mean inner diameter of the composite tube.

Damage Characterization
An X-ray micro-CT system (GE, phoenix v|tome|x m, GE Sensing & Inspection Technologies, Cincinnati, OH, USA) was employed to characterise the internal damages of the composite tube after the LVI and R-RCS tests. The tomography system equipped with a 300 kV micro focus X-ray tube and a highly dynamic DXR digital detector array provided detailed three-dimensional information about the specimens.

Radial Residual Compression Strength Test
R-RCS tests were conducted to study the radial compressive strength loss of the composite tubes tested on the LD23 universal testing machine (Shenzhen Rambo Sansi Material Testing Co. LTD, Shenzhen, China). A constant displacement rate of 10 mm/min was employed and the tests were carried out according to GBT 5352-2005 standard. As shown in Figure 6, the composite tubes underwent radial compression between two parallel load plates with the dimensions of 200 mm × 150 mm × 30 mm. The impact position was located along the largest transverse size. The experiment was stopped when the deformation rate of the diameter reached 60% or the structure was clearly damaged. The deformation rate P was calculated using the following equation: where ∆Y is the change in the variation value of the diameter and d is the mean inner diameter of the composite tube.

Scanning Electron Microscopy (SEM) Characterization
The SEM analysis of the tubes was carried out on a Hitachi SU8010 microscope (Hitachi, Tokyo, Japan) to observe their internal matrix/fibre interface and the dispersion of the fibres in the resins after breaking off in liquid nitrogen.

Low-Velocity Impact Response
The diagrams for contact force, displacement, impactor velocity, and absorbed energy were studied to assess the impact behaviour of the four types of fibre-reinforced composite tubes subjected to LVI. Displacement is obtained by recording the vertical displacement of the impactor nose. Figure 7 exhibits four closed contact force-displacement (F-D) curves, indicating that the impacted tube is not perforated; an open curve is observed in the presence of perforation. Small oscillations in the rising phase can be observed on the force-time curve in Figure 7 due to localized damage induced by the impactor, which indicates that the composite tubes are constantly damaged. The contact force near the value of 1250 N drops sharply at CFRP-T700 and CFRP-T800 tubes. The contact force at the GFRP tube decreases violently under about 2250 N, which may be caused by serious damage. Compared with the other three tubes, the contact force at BFRP tube is relatively stable, of which no big fluctuations appear during the whole process. During the loading phase of impact, all curves exhibit the same trend. The area under the curve is the deformation energy that is initially progressively transferred from the impactor to the tubes; subsequently, the deformation energy is returned from the tubes to the rebounding impactor. The area inside the loop refers to the energy absorbed during the impact. After the tubes absorbed certain impact energy, the damage mechanism was activated, and then a fluctuant rising of the force appears. As displayed in Figure 7, during the impact and the rebound process, the contact force values of the FRP tubes at the same displacement show the following trend: GFRP > BFRP > CFRP-T700 > CFRP-T800.

Scanning Electron Microscopy (SEM) Characterization
The SEM analysis of the tubes was carried out on a Hitachi SU8010 microscope (Hitachi, Tokyo, Japan) to observe their internal matrix/fibre interface and the dispersion of the fibres in the resins after breaking off in liquid nitrogen.

Low-Velocity Impact Response
The diagrams for contact force, displacement, impactor velocity, and absorbed energy were studied to assess the impact behaviour of the four types of fibre-reinforced composite tubes subjected to LVI. Displacement is obtained by recording the vertical displacement of the impactor nose. Figure  7 exhibits four closed contact force-displacement (F-D) curves, indicating that the impacted tube is not perforated; an open curve is observed in the presence of perforation. Small oscillations in the rising phase can be observed on the force-time curve in Figure 7 due to localized damage induced by the impactor, which indicates that the composite tubes are constantly damaged. The contact force near the value of 1250 N drops sharply at CFRP-T700 and CFRP-T800 tubes. The contact force at the GFRP tube decreases violently under about 2250 N, which may be caused by serious damage. Compared with the other three tubes, the contact force at BFRP tube is relatively stable, of which no big fluctuations appear during the whole process. During the loading phase of impact, all curves exhibit the same trend. The area under the curve is the deformation energy that is initially progressively transferred from the impactor to the tubes; subsequently, the deformation energy is returned from the tubes to the rebounding impactor. The area inside the loop refers to the energy absorbed during the impact. After the tubes absorbed certain impact energy, the damage mechanism was activated, and then a fluctuant rising of the force appears. As displayed in Figure 7, during the impact and the rebound process, the contact force values of the FRP tubes at the same displacement show the following trend: GFRP > BFRP > CFRP-T700 > CFRP-T800.

Low-Velocity Impact Damage Characterisation
To demonstrate the above failure behaviour and to better understand the failure mechanisms after impact, especially when the damage is underneath the material surface and cannot be seen by the naked eye, CT images of specimens can be utilised. Three circumferential-sectional views (A-A', B-B' and C-C') and three axial-sectional views (a-a', b-b' and c-c') are chosen from the centre to the edge of the impact location in the tubes, as displayed in Figure 9.

Low-Velocity Impact Damage Characterisation
To demonstrate the above failure behaviour and to better understand the failure mechanisms after impact, especially when the damage is underneath the material surface and cannot be seen by the naked eye, CT images of specimens can be utilised. Three circumferential-sectional views (A-A', B-B' and C-C') and three axial-sectional views (a-a', b-b' and c-c') are chosen from the centre to the edge of the impact location in the tubes, as displayed in Figure 9.

Low-Velocity Impact Damage Characterisation
To demonstrate the above failure behaviour and to better understand the failure mechanisms after impact, especially when the damage is underneath the material surface and cannot be seen by the naked eye, CT images of specimens can be utilised. Three circumferential-sectional views (A-A', B-B' and C-C') and three axial-sectional views (a-a', b-b' and c-c') are chosen from the centre to the edge of the impact location in the tubes, as displayed in Figure 9.  Figure 9 displays the CT images of composite tubes after impact, wherein, images (a) to (d) is for CFRP-T800, CFRP-T700, BFRP and GFRP tubes, respectively. From CT images, more types of damage mechanisms can be observed in greater details, including fibre breakage, fibre rupture and delamination [23]. Figure 9a illustrates the failure mechanism of CFRP-T800 tube after impact. The degree of the damage weakens along with the distance from the impact centre to the edge. Figure 9 (b) displays the failure mechanism of CFRP-T700 tube after impact. Similarly to the CFRP-T800 tube, obvious delamination damage is observed and the global fibre breakage occurs around the impact centre. This is related to the lower elongation at break and the greater brittleness of the carbon fibres. During the LVI process, the contact force of both CFRP-T800 and CFRP-T700 tubes drops sharply when it reaches 1300 N (Figure 7), which may be related to the global fibre breakage damage, as seen in Figure 9a,b. It is clear that the impact behaviour of the two composite tubes is similar at 15 J energy. Figure 9c displays the failure mechanism of the BFRP tube under LVI. Differing from the CFRP tubes, the failure mechanism of the BFRP tube is mainly in the form of delamination, while the fibre  Figure 9 displays the CT images of composite tubes after impact, wherein, images (a) to (d) is for CFRP-T800, CFRP-T700, BFRP and GFRP tubes, respectively. From CT images, more types of damage mechanisms can be observed in greater details, including fibre breakage, fibre rupture and delamination [23]. Figure 9a illustrates the failure mechanism of CFRP-T800 tube after impact. The degree of the damage weakens along with the distance from the impact centre to the edge. Figure 9 (b) displays the failure mechanism of CFRP-T700 tube after impact. Similarly to the CFRP-T800 tube, obvious delamination damage is observed and the global fibre breakage occurs around the impact centre. This is related to the lower elongation at break and the greater brittleness of the carbon fibres. During the LVI process, the contact force of both CFRP-T800 and CFRP-T700 tubes drops sharply when it reaches 1300 N (Figure 7), which may be related to the global fibre breakage damage, as seen in Figure 9a,b. It is clear that the impact behaviour of the two composite tubes is similar at 15 J energy. Figure 9c displays the failure mechanism of the BFRP tube under LVI. Differing from the CFRP tubes, the failure mechanism of the BFRP tube is mainly in the form of delamination, while the fibre breakage is observed locally. Special attention should be paid to the fact that the fibre breakage damage of the BFRP tube occurs in the impact centre, which is different to the CFRP tubes, indicating that the fibre has an influence on the FRP tubes when subjected to LVI. Figure 9d displays the failure mechanism of the GFRP tube after impact. Similarly to the BFRP tube, the delamination damage is also the main failure mechanism and localized fibre breakage is observed. From the above, it can be seen that, when subjected to the impact energy of 15 J, fibre breakage and delamination are the main forms of damage in the FRP tubes, while fibre rupture is rarely seen, which requires a large amount of energy [24]. The interfacial adhesion between fibres and resins differ a lot, as well as the mechanical properties of composite materials reinforced by different fibres. This results in the damage degree of composite tubes when subjected to LVI.
The mechanical properties of the composite structures play an important role in the impact behaviour of composite tubular structures. The matrix/fibre interfacial properties of the composite structures play an important role in the structural impact resistance. To examine the internal matrix/fibre interface of the un-impacted reinforced tubes, their SEM images were obtained ( Figure 10). It can be clearly observed from the SEM images that most of the basalt fibres were pulled out from the resins and a small amount of resin adhered to the pulled-out fibres, indicating that a weak interfacial adhesion existed between the fibres and resin in the case of the BFRP tubes. It could be used to explain the failure mechanism of the BFRP tubes when subjected to impact. Furthermore, the relatively low mechanical properties of BF and BFRP caused the localized fibre breakage to occur. In the SEM image of the CFRP-T800 tube, a small amount of fibres was pulled out with partial resin adhesion, but the size of the pulled-out fibres was small. Similarly to the case of the CFRP-T800 tube, only a few fibres were pulled out in the case of the CFRP-T700 tube, and a neat and packed cross-section could be observed. Relating to the mechanical properties, it can be understood that the failure mechanism of the two is similar when subjected to impact. In particular, in the SEM image of GFRP tube, partial glass fibres were pulled out from the resin because of their large elongation; whereas a large amount of resin adhered to the pulled-out fibres, indicating that the adhesion between the glass fibre and resin matrix was strong and the glass fibres were completely incorporated into the resin. Therefore, the GFRP tube shows the best impact resistance, which related to the infiltration of fibres in resins, and the toughness of the fibres. Figure 11 represents the energy-time (E-T) diagrams of the four types of fibre-reinforced composite tubes during the impact process. The internal energy in the tubes first increase and subsequently decrease to a stable value. The maximum internal energy of the four tubes reinforced by CFRP-T800, CFRP-T700, BFRP and GFRP are 14.56 J, 14.49 J, 14.46 J and 14.59 J, respectively; as soon as the internal energy reaches a maximum value, the impactor starts to rebound. The initial impact energy set in the experiment is 15 J. The maximum internal energies of the four tubes obtained by the impact sensor are about 14.5 J, and the error is within the acceptable range, implying that the impact sensor is accurate. The final energies in the tubes are 10.62 J, 10.69 J, 9.7 J and 11.05 J, respectively, which are the absorbed energies of the tubes. To compare the impact damage behaviour, the percentage of energy absorption (r) is calculated as where E a is the absorbed energy and E i is the maximum internal energy. The calculated values of r of the four tubes are 72.7%, 73.7%, 67.1% and 75.7%, respectively. Figure 12 clearly shows that the GFRP tube has the maximum peak force as well as the highest value of r, which indicates that the strong interfacial adhesion between the fibres and resin, and the elongation of the glass fibres improves the structural impact resistance of the tubes. The value of r of the CFRP-T800 and CFRP-T700 tubes are close, which is related to the similarity of the previous failure mechanism. Fibre breakage occurs globally around the impact centre and thus absorbs more energy. The GFRP tube absorbs plenty of energy through delamination and deformation, of which the residual deformation ratio of the GFRP tube reaches the most (Figure 8). In contrast, the value of r of the BFRP tube is much lower than the other three tubes, indicating that the BFRP tube showed the lowest structural impact resistance. It can be seen from Figure 9c that the BFRP tube mainly realizes energy absorption through delamination damage. Although localized fibre breakage occurs, the energy absorbed by delamination is much smaller [25]. Therefore, the BFRP tube absorbed the least energy. Relevant information about the impact behaviour of all composites is summarized in Table 3. Materials 2020, 13, x FOR PEER REVIEW 11 of 16 Figure 10. Sectional SEM images of the un-impacted FRP composite tubes reinforced by the four types of fibres. Figure 11 represents the energy-time (E-T) diagrams of the four types of fibre-reinforced composite tubes during the impact process. The internal energy in the tubes first increase and subsequently decrease to a stable value. The maximum internal energy of the four tubes reinforced by CFRP-T800, CFRP-T700, BFRP and GFRP are 14.56 J, 14.49 J, 14.46 J and 14.59 J, respectively; as soon as the internal energy reaches a maximum value, the impactor starts to rebound. The initial impact energy set in the experiment is 15 J. The maximum internal energies of the four tubes obtained by the impact sensor are about 14.5 J, and the error is within the acceptable range, implying that the impact sensor is accurate. The final energies in the tubes are 10.62 J, 10.69 J, 9.7 J and 11.05 J, respectively, which are the absorbed energies of the tubes. To compare the impact damage behaviour, the percentage of energy absorption (r) is calculated as delamination is much smaller [25]. Therefore, the BFRP tube absorbed the least energy. Relevant information about the impact behaviour of all composites is summarized in Table 3.    delamination is much smaller [25]. Therefore, the BFRP tube absorbed the least energy. Relevant information about the impact behaviour of all composites is summarized in Table 3.

Radial Compression after Impact Response and Behaviour
Radial residual compression strength (R-RCS) tests were conducted to investigate the loss of compressive strength of the specimen. Figure 13 shows the compression load-displacement curve for the four types of impacted and un-impacted fibre-reinforced tubes, including CFRP-T800, CFRP-T700, BFRP and GFRP, during R-RCS tests. The curves show that the LVI behaviour has different effects on the residual compression performance of the four fibre-reinforced tubes. To compare the impact damage behaviour of these four types of fibre-reinforced composite tubes, the percentages of R-RCS ratios (R) [26,27] can be calculated as where F max is R-RCS peak load. The values of R of the CFRP-T800, CFRP-T700, GFRP and BFRP tubes of the fibre-reinforced tubes are 88.1%, 90.7%, 97.3%, and 73.8%, respectively. The R-RCS of the impacted BFRP tube showed little variation compared to the un-impacted tube. This is related to the fact that the BFRP tube absorbed less energy. As the failure mechanism and absorption energy of CFRP-T800 and CFRP-T700 tube differ little under LVI, their retention rates of compression performance after impact are relatively close. Due to the occurrence of global fibre breakage, the compression performance after impact decreases significantly. It should be noted that during the compression process of the impacted CFRP-T800 tube, a severe load drop appears at the position of I', indicating that the LVI has a great impact on the radial compression performance of the CFRP-T800 tube. On the contrary, the GFRP tube absorbed more energy leading to a lower R-RCS. Among the four types of tubes, the R-RCS of the BFRP tube is not sensitive to impact and that of the GFRP tube is highly sensitive to impact. To compare the damage evolution during the R-RCS test of the tubes with different fibres, images of the composite tubes under radial compression are shown in Figure 14. There appears to be elastic deformation at the initial stage of the compression process, as shown in Figure 14. The compression process of four types of FRP tubes differs little, taking the CFRP-T700 tube as an example. With deepening of the deformation, the compression load reaches the maximum value and thereafter begins to fall drastically. Evident delamination appears at the impact position A in image (b), indicating that the strength of tubes decreased and thus caused damage. Fibre fracture and warp up at position B in image (c) can be attributed to the bending stress concentration in the central face under the compression stage; position I' corresponds to this in Figure 13. In Figure 14d, the upper and lower cambered surfaces are clearly bent and damaged; position I" in Figure 13 corresponds to the damage on the opposite side of the impact centre. After the compression displacement exceeds (d), the degree of damage of the tube increases and the force drops continuously until the structures on both sides are completely destroyed; after the upper and lower arch surfaces are fully stressed, the force begins to increase once again. According to the order of damage, damage occurs first in impact position during compression. Therefore, LVI has a considerable influence on the R-RCS of composite tubes.
CFRP-T800 and CFRP-T700 tube differ little under LVI, their retention rates of compression performance after impact are relatively close. Due to the occurrence of global fibre breakage, the compression performance after impact decreases significantly. It should be noted that during the compression process of the impacted CFRP-T800 tube, a severe load drop appears at the position of I', indicating that the LVI has a great impact on the radial compression performance of the CFRP-T800 tube. On the contrary, the GFRP tube absorbed more energy leading to a lower R-RCS. Among the four types of tubes, the R-RCS of the BFRP tube is not sensitive to impact and that of the GFRP tube is highly sensitive to impact. To compare the damage evolution during the R-RCS test of the tubes with different fibres, images of the composite tubes under radial compression are shown in Figure 14. There appears to be elastic deformation at the initial stage of the compression process, as shown in Figure 14. The compression process of four types of FRP tubes differs little, taking the CFRP-T700 tube as an example. With deepening of the deformation, the compression load reaches the maximum value and thereafter begins to fall drastically. Evident delamination appears at the impact position A in image (b), indicating that the strength of tubes decreased and thus caused damage. Fibre fracture and warp up at position B in image (c) can be attributed to the bending stress concentration in the central face under the compression stage; position I' corresponds to this in Figure 13. In Figure 14d, the upper and lower cambered surfaces are clearly bent and damaged; position I" in Figure 13 corresponds to the damage on the opposite side of the impact centre. After the compression displacement exceeds (d), the degree of damage of the tube increases and the force drops continuously until the structures on both sides are completely destroyed; after the upper and lower arch surfaces are fully stressed, the force begins to increase once again. According to the order of damage, damage occurs first in impact position during compression. Therefore, LVI has a considerable influence on the R-RCS of composite tubes.

Conclusions
This study investigates the effect of fibres on the LVI behaviour, failure mechanism, and RCS of tubular composite structures reinforced by carbon fibre-T800, carbon fibre-T700, basalt fibre, and glass fibre. The following conclusions can be drawn from the results obtained. i.
Delamination and fibre breakage damage are the main failure mechanisms of FRP tubes when subjected to LVI. The damage degree depends on the interfacial adhesion between the fibres and resins, and the mechanical properties of the tubes. Global fibre breakage damage occurs in the

Conclusions
This study investigates the effect of fibres on the LVI behaviour, failure mechanism, and RCS of tubular composite structures reinforced by carbon fibre-T800, carbon fibre-T700, basalt fibre, and glass fibre. The following conclusions can be drawn from the results obtained. i.
Delamination and fibre breakage damage are the main failure mechanisms of FRP tubes when subjected to LVI. The damage degree depends on the interfacial adhesion between the fibres and resins, and the mechanical properties of the tubes. Global fibre breakage damage occurs in the CFRP-T800 and CFRP-T700 tubes, while localized fibre breakage occurs in the BFRP and GFRP tubes. ii.
The residual deformation ratios of the composite tubes show the following trend: GFRP tube > CFRP tubes > BFRP tube-the same as the percentages of energy absorption. The residual deformation ratio and the energy absorption percentage of CFRP-T700 and CFRP-T800 tubes differ little, which related to their mechanical properties. The GFRP tube absorbs the most energy through delamination and large deformation, which shows the best impact resistance. The BFRP tube absorbs the lowest energy, mainly by delamination, which shows the poorest impact resistance. iii.
LVI has a considerable influence on the R-RCS of composite tubes. The retention rates of R-RSC of the four tubes are 88.1%, 90.7%, 97.3%, and 73.8%. The BFRP tube is not sensitive to impact, and the GFRP tube is highly sensitive to impact.