Performance Prediction of Eco-Friendly Concrete with Artificial Neural Networks (ANNs)

Concrete is renowned for its durability and versatility in construction, making it essential for global infrastructure development. Its extensive use contributes significantly to carbon emissions and environmental harm. In response, eco-friendly concrete has developed as a viable option, including elements such as Alccofine and Graphene oxide to improve performance while lowering environmental effect. In this study Alccofine, which accounts for 10% of the mix, replaces a portion of the Ordinary Portland cement with a supplemental substance obtained from industrial slag, minimizing the concrete's carbon footprint. Graphene oxide, at 0.045%, improves mechanical strength potentially increasing the concrete's lifespan and lowering maintenance requirements when compared to typical mixes. Artificial Neural Networks (ANNs) serve as a reliable way for properly estimating the compressive strength of environmentally friendly concrete. By training ANNs on 80% of the datasets containing composition variables, curing conditions, and other important parameters, the models capture complicated, complex relationships and was tested on the remaining 20% to forecast compressive strength with minimal error. The Decision Tree Regressor scored a training precision of 0.4679 and a testing precision of 0.2955, while the Random Forest Regressor scored a training precision of 0.4592 and a testing precision of 0.3010. Based on these findings, The Random Forest Regressor's higher accuracy in prediction establishes it as the more effective model for this purpose. According to the results, the ANN can effectively learn and recognise patterns to forecasting the compressive strength of environmentally friendly concrete. This demonstrates the potential of machine learning techniques to optimize environmentally friendly concrete mixtures and propel advancements in concrete technology.