Numerical and experimental analysis of natural fiber reinforced biological composites
- (0) Download
https://www.eduzhai.net International Journal of M aterials Engineering 2012, 2(4): 43-49 DOI: 10.5923/j.ijme.20120204.03 Numerical and Experimental Analyses of Biocomposites Reinforced with Natural Fibres Leandro Joséda Silva1, Túlio Hallak Panzera1, AndréLuis Christoforo1,*, Luís Miguel Pereira Durão2, Francisco Antonio Rocco Lahr3 1Department of M echanical Engineering, Federal University of São João del Rei, Frei Orlando Square, 170 - São João del-Rei, M G, 36307-352, Brazil 2ISEP , School of Engineering Polytechnic of Porto, Centre for Research and Development in M echanical Engineering (CIDEM ), Rua Dr. António Bernardino de Almeida, 431 – 4200-072 Porto, Portugal. 3Department of Structural Engineering, University of São Paulo (EESC/U SP), São Carlos, 13566-590, Brazil Abstract In the last decades the biocomposites have been widely used in the construction, automobile and aerospace industries. Not only the interface transition zone (ITZ) but also the heterogeneity of natural fibres affects the mechanical behaviour of these composites. This work focuses on the numerical and experimental analyses of a polymeric co mposite fabricated with epo xy resin and unidirectional sisal and banana fibres. A three -d imensional model was set to analyze the composites using the elastic properties of the individual phases. In addition, a two-dimensional model was set taking into account the effective composite properties obtained by micro mechanical models. A tensile testing was performed to validate the numerical analyses and evaluating the interface condition of the constitu tive phases. Keywords Natural Co mposites, Sisal Fibres, Banana Fibres, Finite Element Analysis 1. Introduction The biocomposite structures constituted of polymeric matrices and natural fibres have been widely investigated in the last decades with high potential of applicat ion[1, 2]. Nowadays, these composites comprise one of the major areas of interest in composite materials research. According to, this interest became mo re effective due to the growing search for low cost materials fro m renewab le sources able to substitute traditional ones. Furthermore, several products made of rein forced plastics have been manufactured with bioco mposites with great success . One of the limitations of biocomposites is the difficu lty to predict the mechanical behaviour due to the interface conditions between the natural fibres and the polymeric matrices, which is e mphasized by the heterogeneity of natural fibre surface. Trad itionally, the macroscopic/ microscopic approaches have been used in the analys is of co mposite structu res. Instead of analytical methods, numerical simu lations have been used to predict the mechanical behaviour of co mposite structures. The Finite Element Method (FEM) has been used to analyze the global behaviour of composite structures and play an important role in detecting damage for laminated * Corresponding author: firstname.lastname@example.org (AndréLuis Christoforo) Published online at https://www.eduzhai.net Copyright © 2012 Scientific & Academic Publishing. All Rights Reserved composites. The FEM must be used with adequate geometric co mplacence in order to enhance the analyses. suggests avoiding triangular finite elements when the stress raiser factors are analysed and extreme mesh refinement is demanded. In this case, elements with four nodes should be also avoided due to shape approaches. The eight node elements can be used when a superior performance, related to geometric approaches, is desired. Several works found in the open literature apply the FEM to analyze the interactions between the constitutive phases based on the global behaviour of the co mposites [8-13]. Few works has been investigating the composites reinforced with natural fibres[14-16]. This work focuses on the numerical and experimental analysis of biocomposites reinforced with unidirectional sisal and banana fibres. Micro mechanical models and experimental tests were performed to obtain the mechanical properties of the individual co mposite phases and the biocomposites in order to evaluate the finite element model and the interface conditions. 2. Materials and Methods Two bioco mposites were investigated in this work, one fabricated with sisal fibres and another with banana fibres. The sisal fibres (Agave Perrine) were supplied by Sisalsul Co mpany (São Pau lo-Brazil). The banana fibres were manually extracted fro m pseudo steam of banana plant (Musa balbisiana). The thermoset matrix phase, epo xy resin 44 Leandro Joséda Silva et al.: Numerical and Experimental Analyses of Biocomposites Reinforced with Natural Fibres (Diglicid il etér of bisfenol A), was supplied by Alpha Resiqualy Co mpany (São Paulo -Brazil). Tensile tests were ca rried out to characterize the fibres and the matrix phase according to ASTM  and ASTM  respectively. The elastic properties were used to perform the micro mechanical analyses of the composites reinforced with 30% o f fibre volu me fract ion. Micro mechanical models such as “Rule of mixture” and “Halpin-Tsai” were used to estimate the effective properties of the bioco mposites. The Rule of mixture considers a perfect interface condition (Voigt model) and the Halp in-Tsai considers a semi-empirical model. Table 1 shows not only the modulus of elasticity (E) of the fibres and the matrix, but also the results obtained via micro mechanical analyses. It is noted that the banana fibres (31.56 GPa ) are stiffer than sisal fibres (16.40 GPa). Standard BS EN ISO. A machine speed of 2 mm.min-1 was set during the tensile test (Figure 3). Figure 2. Backscattering electron images of sisal composite: (a) and banana composite (b) Table 1. Const it ut ive phases and composites (micromechanical analysis) p ro pert ies Ma teri al Meth o d E (GPa) Sisal fibre 16.40 (±2.51) Banana fibre Exp eriment al 31.56 (±2.80) Epoxy resin 0.83 (±0.05) Composites Sisal fibre Rule of Mixture Banana fibre Halpin-Tsai Sisal fibre Banana fibre 5.50 10.05 2.40 3.77 A composite lamina reinforced with sisal/banana fibres were fabricated manually by the use of a metallic frame to align the fibres, avoiding the presence of residual stress after cure, as shown in Figure 1. Figure 3. T ensile test for sisal fibre composite. The experimental stress-strain curve was co mpared to the Fin ite Element Analyses (FEA) results. The FEA was conducted based on the elastic properties obtained via micro mechanica l models and tensile test. Table 2 shows the modulus of elasticity of the co mposites obtained by tensile testing. The banana fibre composites (8.44 GPa) exhib it a superior modulus of elasticity to the sisal fibre co mposites (6.22 GPa). Table 2. Modulus of elasticity of the composites Com posi te E (GPa) Epoxy/sisal Epoxy/banana 6.22 (±0.38) 8.44 (±0.83) 3. Finite Element Analyses (FEA) Figure 1. Assembly of natural fibres using a metallic structure A scanning electron microscopic (SEM ), Hitachi T-3000 model, was used to observe the cross section of the biocomposites. Figure 2 shows the back scattering electron images of the sisal (a) and banana (b) fibre co mposites, res p ectiv ely . The testing machine Shimad zu-Autograph was used to determine the modulus of elasticity and strength of the composites based on the recommendations of Brit ish In order to verify the efficiency of the FEA, experimental tests were conducted for the composites fabricated with 30% of sisal and banana fibres. The software Ansys® 12.1 was used to simu late the biocomposites under tensile loadings using two-dimensional (2D) and three-dimensional (3D) mo d els . Some approaches were carried out in the FEA model, mainly when the 3D analysis was performed. The fibres were set as circular cross-section, paralle l d istributed and uniform along its length; the matrix phase was set as homogeneous, continuous and isotropic; and a perfect adhesion between the constitutive phases was set. The numerical model was fixed in one of the ends and the load was uniformly distributed in the cross section (pressure) of the other end, increasing its intensity from 0 up to 100 MPa, with increments of 10. The loading levels were chosen to assure a linear-elastic analysis during the simu lation. International Journal of M aterials Engineering 2012, 2(4): 43-49 45 The pressure increments were applied in the nu merical specimen, and the displacements were collected. Based on these results the specific strain ( ) was calcu lated by Eq. 1, where ∆L is the length variation (mm) and L is the original length (mm) of the specimen. L (1) L The numerical stress/strain curves were plotted and compared with the experimental tensile test in order to evaluate the interface condition. The FEA was carried out using a 2D and 3D model. 3.1. Two-Di mensional Simulation In the two-dimensional (2D) model the composite behaviour was estimated using the elastic properties provided via micro mechanical models: “Rule of mixture” and “Halp in-Tsai”. The material was considered isotropic and homogeneous. The Poisson ratio was calculated by the rule of mixture based on the Poisson ratio of the sisal fibres (zero) and the epoxy resin (0.4). 3.1.1. Sisal Fibre Co mposites The maximu m displacements were used to obtain the specific strain of the composites. Table 3 exh ibits the numerical strain values obtained for each stress level. The absolute error corresponds to the percent difference between the experimental and numerica l results. Figure 4 shows the stress/strain curves obtained via two-dimensional (2D) analyses for the numerical models (Ru le of mixture and Halpin-Tsai) and experimental test. The curve identified as „Experimental modulus‟ represents the numerica l simu lation data based on the elastic properties via mechanical testing. The large percent error ind icates that the model was not able to predict the mechanical behaviour of the composites, even when the experimental elastic property was used (Table 2). Table 3. Specific strain for the sisal fibre composites – two-dimensional model Stress (MPa) 0 10 20 30 40 50 60 70 80 90 100 Exp. 0 0.001238 0.002414 0.003714 0.005105 0.006513 0.007935 0.009348 0.010766 0.012218 0.013692 Rule of Mixture 0 Er % 0.001815 46.55 0.003629 50.33 0.005444 46.58 0.007258 42.18 0.009073 39.30 0.010887 37.20 0.012702 35.89 0.014517 34.83 0.016331 33.66 0.018146 32.53 Specific strain (mm/mm) Hal pin-Ts ai 0 Er % 0.004160 235.98 0.008320 244,65 0.012481 236.05 0.016641 225.96 0.020801 219.37 0.024961 214.55 0.029121 211.54 0.033281 209.12 0.037442 206.44 0.041602 203.85 Exp. Modulus Er % 0.001597 28.99 0.003195 32.33 0.004792 29.02 0.006389 25.15 0.007986 22.62 0.009584 20.77 0.011181 19.61 0.012778 18.69 0.014376 17.66 0.015973 16.66 Figure 4. Stress/strain plot model of the sisal fibre composites for 2D model 46 Leandro Joséda Silva et al.: Numerical and Experimental Analyses of Biocomposites Reinforced with Natural Fibres The FEA based on the elastic properties from the “Rule of Mixture” exhib ited better results compared to Halpin -Tsai model. This behaviour can be attributed to the good adhesion quality between the constitutive phases. The simu lation using the experimental modulus of elasticity achieved better results compared to the micro mechanical analysis model. However, a large percent error was observed, which can be attributed to the Poisson ratio predicted by the rule of mixtu re. The Po isson ratio of the sisal fibres and epoxy resin was considered zero and 0.4, respectively. 3.1.2. Banana Fibre Co mposites The specific strains for the banana fibre co mposites are shown in Table 4. The specific strains determined by the rule of mixture and the experimental modulus exhibited a good agreement with the experimental results, except for the Halp in-Tsai model. The error is higher when the applied force is lower. This behaviour can be attributed to the sample adjustment in the beginning of the test. Figure 5 shows the stress/strain curves obtained by the e xperimental tensile test and 2D nu merica l simu lations. The experimental and nu merical results revealed a good agreement, except for Ha lpin-Tsai model. Halpin -Tsai model is a semi-emp irical equation wh ich considers a non-perfect interface condition. In this way, it is possible to conclude that the banana fibres provide a good adhesion with the matrix phase, showing a correlation between the numerical (rule of mixture and experimental modulus) and the experimental results. Table 4. Specific st rains for the banana fibre composit es – t wo-dimensional model Stress (MPa) 0 10 20 30 40 50 60 70 80 90 100 Exp. 0 0.001486 0.002504 0.003546 0.004583 0.005566 0.006784 0.007902 0.009006 0.010231 0.011441 Rule of Mixture 0 Error % 0.000993 33.13 0.001987 20.65 0.002980 15.95 0.003974 13.30 0.004967 10.76 0.005960 12.14 0.006954 12.00 0.007947 11.76 0.008941 12.61 0.009934 13.17 Specific strain (mm/mm) Hal pin-Ts ai 0 Error % 0.002648 78.27 0.005297 111.56 0.007945 124.07 0.010594 131.15 0.013242 137.92 0.015891 134.25 0.018539 134.60 0.021188 135.25 0.023836 132.99 0.026485 131.49 Experimental Modulus 0 Error % 0.001049 29.36 0.002099 16.16 0.003149 11.21 0.004198 8.40 0.005248 5.72 0.006297 7.17 0.007347 6.25 0.008396 6.78 0.009446 7.67 0.010495 8.27 Figure 5. Stress/strain plot for the two-dimensional analyses of banana fibre composites International Journal of M aterials Engineering 2012, 2(4): 43-49 47 3.2. Three-Di mensional Simulati on Based on the scanning electron microscopic images (Figure 2) of the composite cross section, it was possible to design a three-dimensional model (3D). In this case, the composite behaviour was investigated using the individual properties of the constitutive phases (matrix and fibres) reported in Table 1. The discretization of the 3D model was performed with the fin ite element So lid 185. Th is finite element is designed by eight nodes situated at the corner of the cube. Each node has six degree of freedo m, three translations and three rotations around X, Y and Z axis. According to the SEM image of the sisal co mposite cross section detailed in Figure 2a, a 3D model was mounted with 55 parallel fibres oriented in the same direct ion of the tensile load (Figure 6). In order to minimize the computational work, only a part of the structure was simu lated, as seen in Fig. 6. Figure 2b shows a good align ment and a uniform distance between the banana fibres. The banana fibre co mposites are more ho mogeneous than the sisal fibre co mposites. The 3D model for the banana fibre co mposites was designed as shown in Figure 7. The longitudinal moduli of e lasticity of the phases (matrix and fibres) were used for the FEA. The determination of Poisson ratio of natural fibres is very comp lex. For this reason different levels of Poisson ratio (fro m 0.00 up to 0.35) were investigated in order to identify whether there is significant effect on the mechanical behaviour of the composites. According to the Poisson ratio of the epo xy resin was taken as 0.4. Table 5. Specific st rains for the sisal fibre composit es – Three-dimensional model Stress (MPa) 0 10 20 30 40 50 60 70 80 90 100 Poisson ratio of the fibre Exp. 0.0 0.25 0 0.00124 0.00241 0.00371 0.00511 0.00651 0.00794 0.00935 0.01077 0.01222 0.01369 0 0.00217 0.00434 0.00651 0.00869 0.01086 0.01303 0.01520 0.01737 0.01954 0.02171 Er (%) 75.36 79.89 75.40 70.13 66.69 64.18 62.60 61.34 59.94 58.59 0 0.00219 0.00437 0.00656 0.00875 0.01093 0.01312 0.01531 0.01750 0.01968 0.2187 Er (%) 76.34 81.19 76.67 71.36 67.90 65.36 63.78 62.51 61.10 59.74 Poisson ratio of the fibre 0.3 0.35 0 Er (%) 0 Er (%) 0.00219 76.71 0.00219 76.71 0.00438 81.26 0.00438 81.27 0.00656 76.74 0.00656 76.75 0.00875 71.44 0.00875 71.44 0.01094 67.97 0.00109 67.98 0.01313 65.44 0.01313 65.44 0.01532 63.85 0.01532 63.86 0.01750 62.58 0.01750 62.59 0.01969 61.17 0.01969 61.18 0.02188 59.81 0.02188 59.81 Exp. Modulus 0 Er (%) 0.00160 28.90 0.00319 32.22 0.00479 28.92 0.00638 25.05 0.00798 22.52 0.00956 20.67 0.01117 19.52 0.01277 18.59 0.01436 17.56 0.01596 16.57 Figure 8 shows the stress/strain curves related to numerical and experimental results. The 3D simu lation data revealed that the Poisson ratio values of the sisal fibres do not affect significantly the linear-elastic behaviour of the co mp o s ites . Figure 6. Three-dimensional (3D) model of the sisal fibre composite Figure 7. Three-dimensional (3D) model of the banana fibres composite 3.2.1. Sisal Fibre Co mposites The 3D simu lation results of the sisal fibre co mposites are exhibited in Table 5. As well as observed for the 2D simu lation, high percent erro rs were obtained. Figure 8. Stress/strain curves of the sisal fibres composites for 3Dmodel The 2D and 3D models were not able to simu late acceptably the mechanical behaviour of the sisal fib re composites. This fact can be attributed to the modelling hypothesis, such as the circular cross section and the uniform 48 Leandro Joséda Silva et al.: Numerical and Experimental Analyses of Biocomposites Reinforced with Natural Fibres distribution of sisal fib res in the matrix phase. As it verified in the SEM image (Figure 2), the composite reinforced with sisal fibres e xh ibits an irregular c ross section with heterogeneous fibre distribution, wh ich hinders the design of 3D modelling. The 3D simu lation carried out using the experimental modulus of elasticity showed better results, however it was not able to simulate the experimental behaviour of the composite. The 2D and 3D models for the sisal fibre co mposites e xhibited simila r results, as can be noted by the percent error values showed in Table 2 and 5, respectively. The 2D simu lation, wh ich was based on the properties of the “Rule of Mixture model”, well described the composite behaviour here analysed. As shown in Table 6, the numerica l results exh ibited good estimative showing percent errors lower than 10%. Fig. 9 reveals that the Poisson ratio variation does not affect the FEA results. The FEA performed using the experimental modulus of elasticity validates the 3D nume rical model for the banana fibre co mposites. 3.2.2. Banana Fibre Co mposites Table 6 shows the experimental and nu merical specific strain data obtained by the 3D co mposite model. The results reveal a good correlation between the specific strains, except for the lower pressure (10MPa), wh ich can be explained by the clamping effect in the beginning of the test, as seen in Figure 9. Fi gure 9. St ress/st rain curves of banana fibre composit es for 3D model 4. Conclusions Table 6. Specific strain for the banana fibre composites – Three-dimensional model Stress (MPa) 0 10 20 30 40 50 60 70 80 90 100 Poisson ratio of the fibre Exp. 0 0.00149 0.00250 0.00355 0.00458 0.00557 0.00678 0.00790 0.00901 0.01023 0.01144 0.0 0 0.00120 0.00241 0.00361 0.00482 0.00602 0.00722 0.00843 0.00963 0.01083 0.01204 Er (%) 18.97 3.84 1.85 5.07 8.14 6.47 6.64 6.93 5.90 5.22 0.25 0 0.00121 0.00243 0.00364 0.00486 0.00607 0.00729 0.00850 0.00972 0.01093 0.01215 Er (%) 18.25 2.98 2.76 6.00 9.11 7.42 7.59 7.88 6.85 6.16 Poisson ratio of the fibre 0.3 0.00122 18.21 0.00243 2.93 0.00365 2.81 0.00486 6.06 0.00608 9.16 0.00729 7.48 0.00851 7.64 0.00972 7.94 0.01094 6.90 0.01215 6.21 0.00122 18.21 0.35 0 Er (%) 0.00122 18.22 0.00243 2.95 0.00365 2.79 0.00486 6.04 0.00608 9.14 0.00729 7.46 0.00850 7.62 0.00972 7.92 0.01093 6.88 0.01215 6.20 Exp. Modulus 0 Er (%) 0.00105 29.39 0.00210 16.21 0.00315 11.25 0.00420 8.45 0.00524 5.77 0.00629 7.22 0.00734 7.08 0.00839 6.82 0.00944 7.72 0.01049 8.31 The uniformity and the arrangement of the natural fibres along the matrix phase are primary to achieve significant results from the fin ite element analysis. The 2D simulat ion using the elastic p roperties predicted by the Rule of mixture was able to describe the biocomposites behaviour, revealing a strong interface condition. The FEA jo ined with the micro mechanical analysis was able to characterize the interface condition of the co mposite p h as es . The 3D model was ab le to simu late the experimental results of the banana fibre co mposites. The uniformity and homogeneity of banana fibre co mposites enhanced the model prediction efficiency. The variation of Po isson ratio (fro m 0.0 up to 0.4) of the fibres does not affect the FEA results. Otherwise, the 3D simulat ion was not able to predict the mechanical behaviour of sisal fibre co mposites which was attributed to the non-uniformity and heterogeneity of the fib res . The FEM can be applied to estimate the mechanical behaviour of bioco mposites reinforced with unid irect ional natural fib res, main ly when there is regularity and symmetry of fibre arrangements. REFERENCES  K. Van De Velde, P. Kiekens. “Thermoplastic pultrusion of natural fibres reinforced composites.” Composite Structures, Vol. 54, p. 355–360, 2001.  M . A. Dweib, B. Hu, A. O‟donnell, H. W. Shenton, R. P. Wool. “All natural composite sandwich beams for structural applications”. Composite Structures, Vol. 63, p. 147–157, International Journal of M aterials Engineering 2012, 2(4): 43-49 49 2004.  A. S. Virk, J. Summerscales, W. Hall, S. M . Grove, M . E.  D. Nabi Saheb, J. P. Jog. “Natural Fibre Polymer Composites: M iles. “Design, manufacture, mechanical testing and numerical modeling of an asymmetric composite crossbow A Review”. Advances in Polymer Technology, Vol. 18, No. 4, p. 351–363, 1999. limb”. Composites: Part B, Vol. 40, p. 249–257, 2009.  G. Wimmer, C. Schuecker, H. E. Pettermann. “Numerical  R. V. Silva. “Compósito de resina poliuretana derivada de óleo de mamona e fibras vegetais”. 2003, 139p. Tese (Doutorado) - Escola de Engenharia de São Carlos, simulation of delamination in laminated composite components – A combination of a strength criterion and fracture mechanics”. Composites: Part B, Vol. 40, p. 158–165, Universidade de São Paulo, São Carlos. 2009.  K. M . M ohan Rao, K. M ohana Rao. “Extraction and tensile properties of natural ﬁbres: Vakka, date and bamboo”. Composite Structures, Vol. 77, p. 288–295, 2007.  K. Alnefaie. “Finite element modeling of composite plates with internal delamination”. Composite Structures, Vol. 90, p. 21–27, 2009.  J. E. Saliba. “Use of Finite Element in M icromechanics of Natural Composites”. Computers & Structures, Vol. 61, n. 3, 415-420, 1996.  L. M . Junior, P. Brøndsted. “M icromechanical modeling of damage and fracture of unidirectional fibre reinforced composites: A review”. Computational M aterials Science. Vol. 44, p. 1351–1359, 2009a.  S. J. Kim, C. S. Lee, H. J. Yeo, J. H. Kim, J. Y. Cho. “Direct numerical simulation of composite structures”. Journal of composite materials, Vol. 36, n. 24, p. 2765 – 2785, 2002.  G. Z. Kang, Q. Gao. “Tensile properties of randomly oriented short δ-Al2O3 fibre reinforced aluminium alloy composites: II. Finite element analysis for stress transfer, elastic modulus and stress-strain curve”. Composites Part A, Vol. 33, p. 657 – 667, 2002.  L. M . Junior, P. Brøndsted. “M icromechanisms of damage in unidirectional fibre reinforced composites: 3D computational analysis”. Composites Science and Technology, Vol. 69, p. 1036 – 1044, 2009b.  A. A. Shaikh, S. A. Channiwala. “Experimental and Analytical Investigation of Jute Polyester Composite for Long Continuous Fiber Reinforcement”. Journal of Reinforced Plastics and Composites, Vol. 25, n. 8, p. 863 – 873, 2006.  S. Lee, S. Wang, G. M . Pharr, H. Xu. “Evaluation of interphase properties in a cellulose fiber-reinforced polypropylene composite by nanoindentation and finite element analysis”. Composites: Part A, Vol. 38, p. 1517–1524, 2007.  S. K. Al-Oraimi, A. C. Seibi. “M echanical characterisation and impact behaviour of concrete reinforced with natural fibres”. Composite Structures, Vol. 32, p. 165-171, 1995.  American Society For Testing And M aterials. ASTM . “Standard test method for tensile properties of single textile fibers”. ASTM D3822, Annual book of ASTM standards, 2007.  American Society For Testing And M aterials. ASTM . “Standard test method for tensile properties of plastics”. ASTM D638, Annual book of ASTM standards, 2003.  British Standard. BS. “Glass fibre reinforced plastics: tensile test”. BS 2747, 1998.  V. Tita. “Análise dinâmica teórica e experimental de vigas fabricadas a partir de materiais compósitos poliméricos reforçados”. 1999, 155 p. Thesis. Escola de Engenharia de São Carlos da Universidade de São Paulo. São Carlos - SP.
... pages left unread,continue reading
Free reading is over, click to pay to read the rest ... pages
0 dollars，0 people have bought.
Reading is over. You can download the document and read it offline
0people have downloaded it