PSI - Issue 31

M. Gljušćić et al. / Procedia Structural Integrity 31 (2021) 116 – 121

117

2

M. Gljuš ć i ć et al. / Structural Integrity Procedia 00 (2019) 000–000

of additive manufacturing and significantly weaker mechanical properties in comparison with traditional fabrication methods still limit their application. Hence, a modified FDM method allows the structure to be selectively reinforced with short or continuous fiber reinforcement, resulting in composite structures with improved mechanical properties as shown in Abadi et al. (2018); Agarwal et al. (2018); Blok et al. (2018); Hart et al. (2018); Melenka et al. (2016); Miguel et al. (2019); Oztan et al. (2018); Somireddy et al. (2020) with indications of notable deviation from the CLT theory and the rule of mixture, as well as a significant influence of manufacturing defects on the overall mechanical propertiesMei, (2019); Werken et al. (2019; Yu et al. (2019). Although the results in the mechanical behavior of these materials are scarce today, authors up to date have documented mechanical properties under tensile testing without interruption Agarwal et al. (2018); Blok et al. (2018); Melenka et al. (2016). Those experimental data lack the stress strain relationship response in the unloading stages, thus resulting in a lack of possibility of plastic strain determination. Finally, the study on the mechanical behavior of composite materials is still an ongoing research as shown in Abdo et al. (2019); Chowdhury et al. (2019); Eskandari et al. (2019); Takacs et al. (2020) hence, it is necessary to examine the additively manufactured laminates from both composite and additive manufacturing perspectives in order to develop new products, or reverse engineer complex geometries Babić et al. (2020). 2. Specimen preparation and experimental procedure Therefore, eight sets of continuous carbon fiber reinforced thermoplastic composite (CFRTP) test specimens were designed and additively manufactured according to ASTM D3039 for different fiber volume fractions and layer stacking sequences. The ASTM D3039 instructions result in specimens which are less prone to stress concentrations caused by voids in fiber direction change, as documented in (Hart et al., 2018), resulting in rectangular test specimens with average dimensions of 250 mm in length, 29 mm in width. They contain 15 layers with 130 μm thickness, out of which 8 layers are of polyamide, and 7 layers are of continuous carbon fibers embedded in nylon matrix, enriched with short carbon fibers for stability. The laminate stacking sequence (LSS) was selected to ensure the maximal volume fraction of continuum carbon fibers ( V ff ), while covering a wide range of commonly used fiber direction angles, resulting in cross ply and angle-ply laminates with fiber orientations of 0/90, -30/30, -45/45, -60/60, while the polyamide layers were deployed on constant -45/+45 raster angle. Moreover, as commercial 3D printers often print an unnecessary contour around the printed part, specimens had to be prepared by machining the unnecessary edges. These edges were out of polyamide material only, thus the milling had to be done using sharp cutter under with lower rpm, thus the resulting average specimen width was 24 mm. According to ASTM D3039, glass-epoxy angle-ply gripping tabs with ( l x w x t) dimensions of 55mm x 24mm x 2mm were produced and bonded on the specimens ends with an epoxy-based glue. These tabs were indeed necessary, as testing without them resulted in damage induced by grip pressure necessary for test execution. Moreover, since the behavior was planned to be monitored using digital image correlation (DIC) system, the layup surface was treated with black and white color to achieve a stochastically paint pattern necessary for GOM-Aramis DIC system software. Finally, the prepared specimens were statically tested over three consecutive phases of progressive tensile loading and unloading using an axial testing machine. Three equal specimens were tested for each LSS for three sequential steps of loading and unloading during constant strain rate, except for of LSS -60/60 which was tested with two steps due to its extremely low stiffness. The entire process was monitored and recorded using the GOM-Aramis DIC system with 10 Hz frequency, capturing average displacements, as well as specific local strains at the material layup connections. Moreover, fiber angle rotation during the yielding phase in most angle-ply laminates was also recorded. The average strains were calculated in two axes, y being the loading axis, and x being the axis in the specimen width direction, perpendicular to loading axis, while making sure of eliminating the effects of recorded rigid body motion. The strains in thickness direction z were not recorded, thus neither calculated. 3. Specimen preparation and experimental procedure The tensile behavior has been recorded for every LSS at every 0,1 second, but for simplicity, diagrams show the data for every second. Moreover, the stress-strain diagrams, d) on Figures 1-4, show the average effective stress vs. average effective strain calculated as the average of the three identical specimens for every LSS