Geometric Accuracy and Energy Absorption Characteristics of 3D Printed Continuous Ramie Fiber Reinforced Thin-Walled Composite Structures

Based on the existed research on continuous fiber 3D printing of composite materials, the specimens with three configurations in this study were first prepared using the printing parameters of a single layer thickness of 0.4 mm and a printing speed of 100 mm/min [35, 36]. The corresponded cross-sectional surfaces of the three structures are shown in Figure 2. In this study, the geometric accuracy of the printed specimens was used to evaluate the printability. It could be found in Figure 2 that the printed specimens with circular and hexagonal configurations presented relatively good shape accuracy. More specifically, low deviations between designed and printed paths for the continuous fiber were observed for the structures with circular and hexagonal shapes compared with the specimen with quadrangular configuration (as shown in Figure 2(a), (b)). These deviations were more significant at the corner area for the specimen with quadrangular configuration as shown in Figure 2(c). In this figure, the deposited continuous fiber strands presented a large offset toward the inner side of the specimens at the 90° corner position, resulting in an adverse effect on the printing quality of the printed structures. Many studies had illustrated that when printing polymer parts filled with continuous fiber, the shape precision was usually poor at corner regions, especially at sharp curvatures [37‐39]. In addition, it was well known that the printing quality of the 3D printed components was significantly affected by printing parameters [40‐42]. Therefore, this study tried to adjust the printing parameters to obtain the printed specimens with the better printing geometric accuracy. Among the three shapes of specimens in this study, the rectangular specimen was selected as the objective to obtain appropriate printing parameters because it had higher curvature compared with specimens of other shapes and accordingly showed the larger geometric deviations as discussed above.

Figure 3 shows rectangular specimens printed under different printing speeds and layer thicknesses, which were two key printing parameters influencing the printing quality. The tags F100 and L0.2 meant printed specimens prepared with a printing speed of 100 mm/min and a layer thickness of 0.3 mm, and other tags had the similar meaning. In this figure, it could be seen that at higher printing speeds (300 and 500 mm/min), the specimens presented poorer printing geometric accuracy (as shown in Figure 3(d)−(i)), with higher deviations between designed and printed paths. During the in-situ impregnation printing process, the matrix was extruded in a semi-molten state and the fibers stuck in the matrix were dragged and deposited with the matrix as the nozzle moved. During the printing process, the extruded semi-molten matrix did not hold the fibers well, so when the dragging force of the nozzle was high, the fibers were pulled out from the matrix. When printing sharp corners, the nozzle would decelerate and accelerate inevitably, leading to a non-uniform dragging force [43]. Especially, at a high printing speed, the velocity change of the nozzle and corresponding dragging force will be very high at the sharp corners. As a result, at high printing speeds, the fibers were easier to be pulled out from matrix when printing sharp corners. In addition, as the printing speed increased, the deposited material had less time to cool and solidify, which means that the matrix had less binding force on the fiber, resulting in the fiber being more easily pulled out from the matrix.

In addition, at high printing speeds (300 and 500 mm/min), printing geometric accuracy deteriorated with the increase of layer thicknesses. However, when the printing speed was low (100 mm/min), it seemed that shape accuracy of specimen printed at layer thickness of 0.25 mm (see Figure 4(b)) was superior to that at layer thickness of 0.2 mm (see Figure 4(a)). The layer thickness could affect the contact pressure between the material being extruded from the nozzle and the deposited material [24, 44, 45]. Lower layer thickness caused a tight space for depositing and a higher contact pressure between layers, leading to a better bonding of matrix and fibers as reported in Refs. [24, 28, 44]. Therefore, at low layer thickness, the fiber was not easily to be pulled out from the matrix. However, when the layer thickness was low and the printing speed was slow, the nozzle pressure would be high, pushing the matrix to both sides of the deposition path, and consequently resulting in reduced constraints of the matrix for the fiber, as well as matrix-rich region in the corner, as shown in Figure 4(a). Therefore, low layer thickness could also make the printing geometric accuracy at sharp corners worse.

It was found that after printing parameters optimization, the quadrangular tubes still presented the problem that the continuous fiber was not deposited according to the design path, as illustrated in Figure 4(b). Therefore, the printing geometric accuracy of thin-walled structure with quadrangular configuration still needed to be improved. In addition to varying printing parameters, Refs. [46, 47] found that shape accuracy could also be improved by adjusting the print path at the corners when printing pure matrix. These works indicated that when printing sharp corners, path smoothing could improve the printing accuracy at corners. Therefore, this study attempted to make path smoothing in rounding off the corners with different radii (R) by modifying the G-code. The rectangular specimens with rounded corners of different radii printed at a printing speed of 100 mm/min and a layer thickness of 0.25 mm are shown in Figure 5. The distance (δ) between the corner bisector of the designed right-angle corner and the actual deposited continuous fiber was used to quantify the printing geometric accuracy at the corner, as illustrated in Figure 5(a). The δ for printed specimens with rounded corners of different radii is presented in Figure 5(e). In this study, the radii of rounded corners were set to R = 1 mm, R = 2 mm and R = 3 mm. It could be found in this figure that compared with the specimens without amended rounded corners, the δ of the specimens with R = 1 mm and R = 2 mm were reduced by 39.9% and 10.7%, respectively. However, when R continued to raise to 3 mm, the value of δ was larger than that of the unrounded corner. As discussed above, the inevitable velocity changes of nozzle movement when printing sharp corners could lead to deviations between the actual fiber path and the designed fiber path. Rounding the sharp corners of the structure through path design and using rounded corners with the appropriate radius could mitigate this influence, thus improving deposition quality at the corners. Rounded corners with too large radius made the printed path too far from the desired right-angle path and did not achieve the desired shape. According to the above results, after printing parameters and printing path optimization, continuous ramie fiber filled thin-walled structures with circular, hexagonal, and quadrangular configurations were printed at a printing speed of 100 mm/min and a layer thickness of 0.25 mm. The thin-walled structures were successfully manufactured.

Figure 6 presents the deformation and collapse processes of thin-walled structures with three shapes under quasi-static compression. In this figure, it was found that thin-walled structures gradually collapsed under quasi-static compression loading. Specifically, the wall of the structure began to plastic deformation, buckle or break locally after the initial elastic stage, forming a folding wrinkle. As the compression continued, the folded part was compacted, and the next similar folding wrinkle began to form until the structure was completely compacted. Thin-walled structures with three shapes showed similar progressive deformation patterns but had some differences. For structure with circular configuration, the first folding wrinkle was formed by buckling near the upper plate, and there was no obvious second folding wrinkle before compaction. For structure with hexagonal shape, the first folding wrinkle was also near the upper plate formed by buckling, but as the compression continued, it was found that the second folding wrinkle was formed below the first folding. As for structure with quadrangular configuration, unlike the former two shapes structures, the first folding wrinkle appeared in the middle of the tube, and the distinct rupture of wall could be seen in Ref. [15].

Figure 7 shows the typical failure modes of the structures under quasi-static compression. In this figure, three typical failure modes could be seen: (a) Local buckling deformation and matrix cracking when local deformation was large, as shown in Figure 7(a); (b) fracture along the vertical axial direction of the tube wall and ramie fibers were pulled out from matrix and were fractured, as shown in Figure 7(b); (c) delamination along the horizontal direction of the structure wall (including matrix/matrix debonding and matrix/fiber debonding, as shown in Figure 7(c)). Large local buckling deformation was a common failure mode for thin-walled structures [48‐50]. Specifically, the printed structures absorbed energy through plastic deformation together with cracks and fractures during the quasi-static compression process. The thin-walled structures in this study were fabricated by in-situ impregnation fused deposition modeling technique, so relatively poor adhesion between layers might be one possible reason of fracture along the horizontal direction of the structure wall [44]. Such delamination could affect the load bearing capacity of the structure, as the composite layers affected by this phenomenon lose their capability to effectively distribute stresses. In addition, the layered cracking could prevent the excessive plastic bending required for the folding process (as in the structure with circular configuration) [51].

To investigate the mechanical performance of 3D printed continuous ramie fiber reinforced thin-walled structures with three shapes, representative force-displacement curves for three different structures are shown in Figure 8. Additionally, to better evaluate the variation on the mechanical performance of the tested samples, the force-displacement curves of thin-walled structures with the same cross-sectional shape are shown in Figure 9. It was found in Figure 8 that thin-walled structures with quadrangular configuration required smaller forces to initiate the crushing compared to the other structures. Moreover, it was found that all the force first rose rapidly to reach the first peak, which corresponded to the forming of the first wrinkle. Thereafter, the force-displacement curves declined to the first trough, indicating that the first wrinkle was compacted. Next, the curves continued to fluctuate, corresponding to the progressive folding process of the structures. The number of waves in the curves of thin-walled structures was in good agreement with the corresponded deformation and collapse processes as discussed in Figure 6. Specifically, the force-displacement curve for circular structure showed fewer waves than those of other structures, corresponding to the fewer formed compression folds of circular structure during compression as shown in Figure 6a. Due to the mix of different failure behaviors for energy dissipating, such as delamination, fiber breakage and matrix cracking, the force-displacement curves showed a few irregular sharp peaks.