Optimizing compressive strength in sustainable concrete: a machine learning approach with iron waste integration

The current research delves into enhancing the sustainability of construction materials by incorporating iron waste into concrete mixtures. The primary aim revolves around predicting the compressive strength of such innovative concrete formulations, a critical factor in maintaining the structural integrity of constructions. By employing various machine learning algorithms—Support Vector Machine (SVM), Random Forest (RF), Gradient Boosting (GB), Extreme Gradient Boosting (XGB), and Multilayer Perceptron (MLP)—the study determines the most efficacious models for predicting compressive strength. Notably, Random Forest emerges as the most proficient, as evidenced by its exceptional R² (= 0.972) and CPI score (= 0.250). A meticulous sensitivity analysis further elucidates the principal factors influencing compressive strength, notably the incorporation ratios of Iron Waste and Fine Aggregate, alongside the concrete’s age. This investigation meticulously navigates from data preprocessing to the final model selection and sensitivity analysis, ensuring the robustness of the predictive models. Moreover, the study extends its utility beyond academic realms by developing an accessible graphical user interface (GUI), hosted on GitHub, to facilitate the application of these findings. The inclusion of iron waste not only propels the construction industry towards more sustainable practices but also valorizes waste materials. Consequently, this research contributes substantially to the domain of sustainable construction by providing a reliable methodology for the integration of iron waste in concrete, thereby fostering the development of eco-friendlier construction practices. The additional creation of a GUI significantly amplifies the impact of this research, making its insights accessible to a broader audience, thus benefiting the society and construction industry at large.