ANN-Based Mechanical Property Prediction of Bio-Fibre for Wind Turbine Blade Materials with FEM Validation

The increasing demand for renewable energy highlights the need for sustainable materials in wind turbine blade design. Conventional fiberglass blades, while effective, present environmental and disposal challenges, motivating the exploration of bio-composites as greener alternatives. This study aims to develop and validate an integrated framework that combines experimental validation, Finite Element Method (FEM) pre-screening, Artificial Neural Networks (ANN), and Rule of Mixtures (RoM) validation to evaluate the feasibility of bio-fibre wind turbine blades Mechanical properties of flax, hemp, sisal, jute, pineapple fiber, and resin are obtained from previously published experimental studies available in the literature, with resin content fixed at 90% and permutations generated for ANN training. Experimental tensile testing on a 90% resin–10% pineapple fiber composite yields 131 MPa, closely matching the permutation prediction of 118.6 MPa, confirming dataset reliability. FEM simulations are then employed to pre-screen potential maximum performance values within the dataset range, ensuring the physical feasibility of ANN input properties. Using these validated inputs, the ANN predicts feasible bio-composite compositions, which are further compared against RoM estimations. The results show that ANN predictions remain within a 7% deviation from RoM values, demonstrating consistency with micromechanical theory. This integrated framework highlights that FEM-based input screening enhances ANN prediction reliability, and pineapple-based bio-composites can serve as sustainable and technically viable alternatives for wind turbine blade applications.