An integrated Artificial neural network technique to optimize the various parameters of Pineapple/SiO2/epoxy-based nanocomposites under NaOH treatment

Abstract

• Optimized the mechanical properties of pineapple/SiO₂/epoxy nanocomposites using artificial neural networks (ANN) and Taguchi methods, focusing on tensile, flexural, and impact strength. • Introduced ANN for precise parameter optimization in natural fiber composites, achieving a prediction accuracy with R 2 >0.97, a first in PALF/SiO₂/epoxy studies. • Optimal configuration (20 % fiber content, 10 mm fiber length, 5 wt % SiO₂, 6 h NaOH treatment) provided significant enhancements: tensile strength (45.01 MPa), flexural strength (67.23 MPa), and impact resistance (34.51 kJ/m²). • Utilized eco-friendly pineapple leaf fibers and chemical treatments to produce high-strength, lightweight composites with minimal environmental impact. • Developed composites are ideal for automotive interiors, aerospace components, and structural applications requiring lightweight and high-strength materials. Natural fibre and nano-filler-based composites are gaining popularity because of their high strength, recyclable nature, high strength-to-weight ratio, and other benefits. The primary goal of this study is to maximize the many factors that influence the mechanical characteristics of hybrid nanocomposites. The nanocomposites were created using the standard hand lay-up method with the following conditions: (i) fibre content, (ii) length of pineapple fibre in mm, (ii) weight proportion of nano SiO 2 ; and (iii) alkali treatment periods. Following manufacture, the composite's mechanical characteristics were tested according to ASTM standards. Quantitative optimisation using the Taguchi technique Depending on the S/N ratio of Taguchi refinement, the nanocomposite with the maximum mechanical performance contains 20 % pineapple fibre with a length of 10 mm, 5 % nano SiO 2 , and 6 h of NaOH treatment. The ANN predictive algorithm and the Taguchi L 27 array show that the experiment and projected data for tensile, flexural, and impact are within 3 % and 4 %. According to the ANOVA study, the NaOH approach provides around 45 %, followed by the pineapple fibre content of 25 %, fibre length of 20 %, and filler content of 5 %. The similarity between artificial neural networks and test findings expands the scope of ANN for forecasting strength properties. This approach will assist designers in predicting system failures concerning the operating duration and may be utilised in numerous industrial industries to analyse toughness attributes.