Laboratory-Based and Data-Driven Learning of Concrete Enhanced with Modified Jute Fiber Reinforcement for Sustainable Flexural Prediction

In this work, only 36 concrete beams were prepared, to study the performance of concrete reinforced with modified jute fiber, which focus on its mechanical properties preferably flexural strength and the predictive capabilities of machine learning (ML) models. Sodium hydroxide (NaOH) was used for the treatment of Jute fibers, then were uniformly cut to 20 mm lengths and added to M30-Concrete grade in 0%, 1%, 1.5%, and 2%. Material tests, for sieve analysis, specific gravity, and water absorption, were conducted on aggregates and jute fibers, with the concluding data showing high moisture absorption. Slump tests was done for the fresh concrete, demonstrated reduced workability as fiber content increases. Mechanical tests showed that, 1% jute fiber content revealed optimal improvements in flexural strength Tests. The results were quantitatively analyzed, the hypothesis test was performed using ANOVA revealed that modified jute fibers do not significantly decrease flexural strengths, the p-values for all mechanical properties were greater than 0.05 level of significance, leading to the conclusion that the null hypotheses could not all be rejected. Machine learning models, encompassing multiple linear regression and Random Forest regression, were implemented to predict concrete properties based on fiber content and curing ages, with R-squared values of 0.879 for flexural strength. The results suggest that chemically modified jute fibers enhance flexural properties, and machine learning can effectively model these improvements.