Innovative approach to SFRC analysis: Using gene expression programming (GEP) to predict stress-strain behavior of steel fiber-reinforced concrete under compression

The stress–strain diagram of SFRC is a crucial tool for evaluating its mechanical properties, as it clearly illustrates the enhanced strength, ductility, and energy absorption provided by the steel fibers. The present research presents a novel method for developing a predictive model for predicting the compression stress-strain curve of steel fiber reinforced concrete (SFRC), offering a reliable tool to capture its complex mechanical response. The Generalized Reduced Gradient Method (GRG) and Gene Expression Programming (GEP) were employed to develop a formula that incorporates the main influencing factors, including steel fiber volume percentage, yield stress, reinforcing index, and concrete compressive strength and strain. These characteristics were carefully investigated to develop a powerful formula capable of accurately predicting the stress-strain diagram of SFRC. The model is based on a large dataset of 182 experimental samples used for both development and validation. The models' accuracy was assessed through the root mean square error (RMSE), mean absolute error (MAE), and coefficient of determination ( ) by comparing actual and predicted values. The model's predictions correspond well with experimental data, demonstrating its robustness and reliability in predicting the stress-strain curves. This study sheds light on the intricate interaction of material characteristics in SFRC. It provides a powerful tool for engineers and researchers seeking to enhance the mechanical performance of fiber-reinforced concrete structures. • Developed a GEP model with the GRG method for predicting SFRC stress–strain behavior of SFRC. • Considered steel fiber volume, reinforcing index, and concrete strength and strain as key parameters. • Provides a practical tool for SFRC design and analysis.