In the context of the global environmental challenges facing humankind, there is an increasing need to develop new environmentally friendly products and processes to minimise or avoid the use of polluting materials. Natural fibres have become attractive to researchers, due to their low cost, relatively good mechanical properties, high speciﬁc strength, non-abrasive, eco-friendly and bio-degradability characteristics. They are now being exploited in replacement of organic and synthetic fibres, such as glass, aramid and carbon in the composite material design. For the past two decades, several plants such as flax, sisal, cotton and coconut have been used by various industry groups such as Mercedes for the development of composite products.
In this thesis, a structured six-stage approach has been developed to study the potential of using fibres from different endemic Pandanus leaves for the manufacture of a polyester composite. For each stage, experimental tests and measurement were carried out using well-calibrated devices. As a first stage, the mechanical and chemical characteristics of the Mascarene endemic Screwpine Pandanus (Pandanaceae) fibres have been studied according to several physiological properties of this species (tree maturity, leaf maturity, sunlight exposure, state of natural degradation of the leaf, position of a fibre along the leaf). This first study includes the selection of the species with the highest fibre tensile strength among the 21 existing in Mauritius, the optimisation of alkaline-treatment parameters (soaking time and NaOH concentration) of the fibres from the selected Pandanus species. The purpose of the alkaline-treatment is to increase the strength of the fibre and to improve its adhesion to the matrix. This study revealed that the best species having the highest tensile strength (TS) is the P. iceryi (251 ± 149 MPa) followed by P. utilis (153 ± 81 MPa). An optimum improvement of 23.2 and 36.2 % in the tensile strength compared to untreated fibres were obtained after 0.5 wt. % NaOH for 8 h and 0.5 wt. % NaOH for 14 h respectively for these two species. At the end of this part, a new method for predicting the fibre tensile strength has been developed through the correlation between FT-IR (specific IR peaks ratio relative to the cellulose crystallinity) results and the fibre tensile strength. In the rest of the study, the P. utilis was chosen because of its relatively high strength and its higher availability/abundance than P. iceryi.
In the second stage, P. utilis treated fibres having the optimum tensile strength were converted into non-woven preformed fibre-mats. A method similar to the pulp alignment process, used in papermaking was used to orient the fibres randomly on the web. Hydroxyethylcellulose (HEC) was used such as a fibre-bonding agent.
Then, the third stage consists to determine the interfacial shear stress (IFSS) improvement after alkaline-treatment, the effect of using HEC fibre-bonding agent and the fibre critical length using a single fibre pull-out test (fibre – polyester). The results have shown that HEC has no adverse influence on fibre strength and the alkaline-treatment followed by a coating of 2 wt. % of HEC increases the resistance of the shear-stress by 36 % compared to the untreated one (7.3 ± 1.9 and 4.6 ± 1.2 MPa respectively). Morphological properties such as inter-fibre void spacing, fibre orientation and fibre dispersion were compared to that of a commercial E-glass mat and the results showed a high similarity of these properties.
During the fourth stage, non-woven fibre-mat composites at different volume fractions (0, 10.7, 21.6 and 32.4 % of fibres) were manufactured using the vacuum infusion process. After performing tensile and flexural tests (ASTM D638 and ASTM D790 respectively), the composite having the highest tensile and flexural strengths (32.4 % of fibre/matrix volume fraction for both tests) was chosen for the rest of the study. The results of the composite sample tests showed that the tensile and flexural strength increased at relatively low values of 10 and 16 % respectively compared to the unreinforced matrix. Both the tensile and flexural moduli significantly increase by 51 % and 70 % respectively.
The optimal composite was chosen to manufacture a 550 mm Air-X 403 blade model using the vacuum infusion process. A flap wise static test to determine the load at normal operation, load at worst-case and load at failure was performed as per BN-EN 61 400-2: 2006 standard. The failure was achieved by applying a force of 105 N at a distance of 2/3 of the blade length from the blade root. This maximum force is more than enough to cover the normal operation (45.8 N) and the worst-case scenario (64.8 N). The blade flexural stiffness obtained using Macaulay's equation was 30.1 kN m2 knowing that this value is 53 kN m2 for a unidirectional E-glass / polyester blade composite according to the literature. Three replicated blades were tested to check the repeatability of the test and the blade and the result shows that the variation in the force and deflection at the failure of the blades were 4.5% and 9.5% respectively.
The main part of the fifth stage consists of the development of a homogenised mathematical model of the optimum composite material using the finite element method (FEM). The model was used to predict the maximum theoretical value of the composite strength. Three homogenisation methods using ANSYS Mechanical APDL (ANSYS Parametric Design Language) were performed including the tensile curve fitting method, ANSYS Material Designer tool and direct analysis of a Representative Elementary Volume (REV). The latter takes into account the interfacial shear stress between fibre and matrix, and specific parameters (dynamic and static frictional coefficients) were obtained through a simulation of the pull-out test using ANSYS APDL. Fibre architecture of the REV was obtained by using an image reconstitution technique. Result of the pull-out simulation showed that dynamic friction and a ratio (static to dynamic frictional coefficient) of 0.1 and 9.5 respectively allowed to obtain the targeted IFSSs relative to the fibre bond-slip behaviour (relative error of 0.95 % and 6.7 % for the maximum shear stress and sliding stress respectively compared to the experimental result). The comparison of the three-homogenisation methods showed that the fitting method is the most suitable when optimising an application, the Material Designer tools when optimising the fibre architecture and volume fraction, and the direct REV when optimising the micromechanical parameters such as IFSS. The result of the simulation on a 3-D composite sample showed that the composite should have a maximum tensile strength of 72 MPa, i.e. 30.5 % more than that of the physical model. The reliability of the mathematic model was confirmed by the result of a flexural test. Generally, the correlation coefficient between the theoretical and the experimental curve obtained in this stage was R2 = 0.98.
In the sixth stage, the homogenised material was used to model a 550 mm 3-D solid-blade in view of determining the maximum theoretical flap wise force, optimising the internal structure design of a skinned blade and proposing an optimal ratio blade skin-thickness to blade length. The simulation of the flap wise tested as per BN-EN 61 400 -2: 2006 standard revealed that the maximum force at failure should be twice that of the physical model (200 N and 105 N respectively. It was found that to resist a worst-case scenario; a non-solid blade model must have at least a skin-thickness of 1.75 mm. However, 1 mm of blade-skin-thickness will be sufficient when using a 3mm thick reinforcement flange. A material saving of 35.5 % compared to a solid-blade will be obtained in the latter case.
The numerical model allows the prediction of the mechanical behaviour of the bio-based (fibre) composite for the specified application, thus reducing the need to perform several expensive physical tests.
Based on the findings of the different experimental and theoretical models, it is concluded that P. utilis fibres can effectively be used as a reinforcing material in the polyester matrix for the manufacture of a small-scale WTB, promoting the use of green energy. It also offers an opportunity for the creation of a micro-enterprise. A critical analysis of the future work, perspective and possible improvement has been presented in the last chapter.
Keywords: alkaline-treatment, ANSYS APDL, Hydroxyethylcellulose, interfacial shear stress, natural fibre polyester composite, non-woven mat, Pandanus utilis, pull-out test, Representative Elementary Volume, tensile strength, flexural strength, wind-turbine blade.