Prediction of the mechanical strength of modified kenaf fiber reinforced polymer incorporating nanographene using ANN
S. Jothi Arunachalam, Nagaraj Ashok
Abstract
Artificial Intelligence (AI) methods, such as artificial neural networks (ANN) and machine learning, have found a common use in solving numerous challenges in engineering. The present research work consisted of the making of a hybrid polymer nanocomposite through reinforcing a polymer composite with natural fibers of kenaf (Hibiscus cannabinus) fiber (KF) and Nanographene. A chemical treatment to help in changing the surface property of the fibres facilitate its adhesion and interaction with the polymeric matrix was conducted using a potassium permanganate (KMnO 4 ) solution in acetone (C 3 H 6 O). Originally, the ANN model was applied and trained to predict and optimize the tensile strength (TS) of the resultant KF/nanographene hybrid nanocomposite (KFN). The utilized model was the architecture of a single-layer perceptron with the configuration 3-5-1, and the hidden layer with five neurons. The Central Composite Design (CCD) was used to design the experiments in such a way that it gave a systematic approach towards knowing the impact of the specific variables on the tensile strength. Scanning electron microscopy (SEM) studies supported the great influence of the modification of KMnO 4 on the fiber-matrix interface and hence on the mechanical properties of the composite. The outcomes of the mechanical testing carried out using analysis of variance (ANOVA) proved that the major factors had a significant impact on the tensile strength, whilst the model fitted excellently with a value of coefficient of determination R 2 = 0.9805. By using the ANN-CCD model, the optimum tensile strength value of 46 MPa was predicted; this value was simply the nearest match to the experimental validation test method’s experimentally obtained value of 45.4 MPa, which translates to a near 98.45% accuracy in predicting the model’s result. This paper brings out the usefulness of using ANN with CCD to quickly get reliable estimations of mechanical properties and therefore save experimental design time and cost of production and resources of developing composite materials. The use of natural resources such as kenaf fiber combined with nanographene promotes a sustainable approach by enhancing material performance while supporting sustainable development goals in materials engineering.