Flexural strength prediction of steel fibre reinforced concrete beams: mechanical and ANN approach

Concrete, a fundamental construction material, excels in compression but exhibits brittleness due to its weakness in tension. To address this limitation, researchers have explored the use of various fibres, particularly steel fibres, to enhance its tensile properties, impart ductility, and improve flexural strength. Although several models have been proposed to predict the flexural strength of fibre-reinforced concrete, achieving consistent accuracy across diverse datasets remains a challenge. This research proposes two predictive approaches: a mechanics-based model grounded in physical principles and an artificial neural network (ANN) trained on a compiled database of 72 experimentally tested steel fibre reinforced concrete (SFRC) beams. The mechanics-based model demonstrates reliable performance with a root mean square error (RMSE) of 0.97 and a normalised prediction error of 1.27%. The ANN model also shows strong predictive capability, achieving a coefficient of determination (R2) of 0.99 and RMSE of 0.98. These results confirm the reliability of both approaches. The novelty of this work lies in the development of a simplified mechanics-based model tuned for SFRC beams without longitudinal reinforcement, and its systematic comparison with a data-driven ANN model using multi-source experimental data.