Abstract
River sand supplies are decreasing due to overexploitation and illicit sand mining. One ton of Portland cement production (the main binder in concrete) emits about one ton of carbon dioxide into the atmosphere. Thus, this study replaced conventional cement and river sand (R sand) with recycled waste materials (fly ash and manufactured sand (M sand)). The concrete mix proportions were designed using M40 grade, and the Ordinary Portland cement (OPC) and R sand were replaced with 0-85 wt% of fly ash and 0-100 wt% of M sand. The concrete samples were tested for compressive strength after 3-90 days of curing. Furthermore, machine learning (ML) techniques were engaged to predict the compressive strength of the concrete samples using Extreme Gradient Boosting (XGBoost), Long Short-Term Memory (LSTM), Support Vector Machine (SVM), and Gaussian Process Regression (GPR). Besides, the concrete samples containing fly ash, M sand, and R sand were characterized for microstructures and elemental compositions using SEM-EDS. The results revealed improved concrete compressive strength by incorporating fly ash and M sand. After 28 days of curing, OPC and R sand were partially replaced with 25 and 50 wt% of fly ash and M sand attained the designed strength of M 40 grade concrete. XGBoost model yielded the most accurate performance metrics for forecasting the compressive strength in training and testing phases with R2 values equal to 0.9999 and 0.9964, respectively, compared to LSTM, SVM, and GPR. Thus, the XGBoost approach can be a viable technique for forecasting the strength of concrete incorporating fly ash and M sand. SEM-EDS analyses revealed compact formations with high calcium and silicon counts. Thus, the XGBoost approach can be a viable technique for forecasting the strength of concrete incorporating fly ash and M sand.