Improving Predictive Precision in Compressive Strength of Steel Fiber-Reinforced Concrete: Utilizing Optimized Deep Learning Methods

This study highlights the critical importance of steel fiber-reinforced concrete (SFRC) inconstruction for its enhanced strength and crack resistance. It addresses the challenges inaccurately predicting SFRC's compressive strength due to complex interactions with differentfiber types, which conventional regression models often fail to capture. To improve predictionaccuracy, deep learning (DL) techniques such as One-Dimensional Convolutional NeuralNetworks (1D-CNN) are employed. The research aims to enhance the precision of predictingSFRC’s 28-day compressive strength using Observer-Teacher-Learner-Based Optimization(OTBLO) to optimize deep learning models. This approach demonstrates the potential of DL tostreamline traditional concrete testing processes. The effectiveness of these models is verifiedwith indicators showing strong performance in predicting SFRC strength, with high Pearsoncorrelation coefficients and low error metrics for 1D-CNN.

Improving Predictive Precision in Compressive Strength of Steel Fiber-Reinforced Concrete: Utilizing Optimized Deep Learning Methods Keywords:

Steel fiber-reinforced concrete , 1-dimensional convolutional neural networks , compressive strength , Observer-Teacher-Learner-Based optimization