Energy absorption and damage prediction in natural fibre composites under low velocity impact using machine learning and FEA.

Abstract

This study investigates the energy absorption and damage prediction of banana fiber composite laminates under low-velocity impact using a combination of experimental testing, finite element analysis (FEA), and machine learning (ML). Banana fiber composites are a promising eco-friendly alternative to synthetic materials in structural applications due to their sustainability, high strength, and energy absorption properties. The laminates, fabricated using the hand layup technique, were subjected to low-velocity impact tests to measure their energy absorption, force-displacement behavior, and damage progression. FEA simulations were conducted to model the impact response, and ML models, including logistic regression and Naive Bayes, were developed to predict the impact behavior. The results show that banana fiber composites exhibit significant energy absorption, with an experimental value of 14.36 kJ at a drop height of 1.8 m. Both FEA and ML models closely predicted this energy absorption, with minor deviations, validating the robustness of the methodologies. The study highlights the integration of ML as a powerful tool for predicting composite material behavior, achieving an accuracy of 1.0 in predicting energy absorption and damage initiation. The findings provide valuable insights into the potential of banana fiber composites for use in lightweight, high-strength materials for the automotive and aerospace industries.