Predicting the mechanical properties of date palm wood fibre-recycled low density polyethylene composite using artificial neural network

Abstract

Background: Experimental determination of properties of engineering materials is quite expensive and time consuming. Computational methods of predicting the properties of materials, such as artificial neural network (ANN), are easier and bereft of complex mathematics that characterizes analytical methods. Also, Nigeria consumes a lot of bottled and sachet water. Most of the bottles and sachets are made with low density polyethylene (LDPE) and these sachets constitute a major source of pollution in Nigerian cities and towns. In addition, date palm is a major cash crop in Nigeria. Methods: In this study, an artificial neural network (ANN) approach is used to predict the mechanical properties of date palm wood fiber-recycled low density polyethylene composite. In the artificial neural network, multi layer perceptron architecture (MLP) with back-propagation is utilized. In ANN training module, the ground fibres weight percent (wt %) was used as input for various particles sizes (150, 212, 250 and 300 µm). The outputs consist of the ultimate tensile strength, elongation, tensile modulus, flexural strength, flexural modulus and hardness for the particle sizes: 150, 212, 250 and 300 µm. The artificial neural network system was trained using the prepared training set. After the training process, the test data were used to check the system accuracy. Results: The correlation coefficients of all predictions with experimental values were more than 0.99. Conclusion: These results show that artificial neural network is very successful in the prediction of the mechanical properties.