Novel Mechanical Strength Prediction Models of Fibre Reinforced Concrete Using Statistical Analysis

Abstract

Fibre Reinforced Concrete (FRC) has emerged as a promising construction material due to its enhanced mechanical properties and improved performance under various loading conditions. This study focuses on the development of Non-Linear Regression (NLR) models for predicting the 28-day mechanical characteristics like Compressive (CS), Splitting tensile (STS) and Flexural strengths (FS) of FRC. Through an extensive review of existing literature and empirical data, various factors affecting the mechanical properties of FRC have been identified that include the cement content, fine and coarse aggregate content, super plasticizer content, fibre content and water cement ratio. Leveraging this comprehensive understanding, NLR equations have been formulated to capture the complex relationships between these variables and the 28-day resulting mechanical strengths. The accuracy and reliability of the models have been rigorously assessed through statistical analysis and performance evaluation metrics like R2, Root Mean Square Error (RMSE) and probability (p) value. The NLR model for CS demonstrated an R2 of 0.933, RMSE of 5.6 MPa and a p-value of 4.39e-34, similarly the NLR model for STS demonstrated an R2 of 0.932, RMSE of 1.07 MPa and a p-value of 5.22e-11 and finally FS model of FRC demonstrated an R2 of 0.94, RMSE of 1.23 MPa and a p-value of 7.54e-15 respectively.