Compressive Strength Prediction of Green Concrete with Recycled Glass-Fiber-Reinforced Polymers Using a Machine Learning Approach
Pouyan Fakharian, Reza Bazrgary, A Ghorbani, Davoud Tavakoli, Younes Nouri
Abstract
Fiber-reinforced polymer (FRP) materials are increasingly used in the construction and transportation industries, generating growing volumes of waste. This study applied a machine learning model to predict the compressive strength of eco-friendly concrete incorporating recycled glass-fiber-reinforced polymer (GFRP) waste. Based on 119 laboratory mixes, the model achieved a good prediction accuracy (R2 = 0.8284 on the test set). The analysis indicated that compressive strength tends to decrease at higher GFRP dosages, with relatively favorable performance observed at low contents. The two most influential factors were the water-to-cement ratio and the total GFRP content. The physical form of the recycled material was also important: powders and fibers generally showed positive effects, while coarse aggregate replacement was less effective. This machine learning-based approach offers preliminary quantitative guidance on mix design with GFRP waste and highlights opportunities for reusing industrial by-products in more sustainable concretes.