Prediction of Flexural Strength of Mini-Scale Reinforced Concrete Beam Using Feed Forward Neural Network

This research investigates the partial substitution of cement with rice husk ash (RHA) and coconut fiber (CF) to enhance flexural strength in concrete. Ordinary Portland cement (OPC) and standard cement were combined with RHA and CF in 5%, 10%, and 15% ratios, followed by flexural strength testing and statistical analysis using One-Way ANOVA and One-Sample t-tests. Results showed that the 5% mixture had a mean flexural strength of 4.04 MPa, below the standard of 5.5 MPa, while the 10% and 15% mixtures exceeded the standard, achieving 7.50 MPa and 7.60 MPa, respectively, with significant differences at a p-value < 0.05. Neural network models were used to predict flexural strength, with the best configuration yielding an MSE of 0.3496, MAE of 0.489, and R of 0.9777. Thus, the 15% RHA and CF mix proves to be an effective, sustainable, and economical substitute for conventional cement, offering high flexural strength beneficial to the construction industry.