Ensemble Machine Learning Models to Predict the Compressive Strength and Ultrasonic Pulse Velocity of Sustainable Concrete

Currently, the concrete sector is experiencing a massive problem in adapting to the concept of sustainable development since the manufacturing of ordinary Portland cement (OPC) emits about 8% of CO₂, which is responsible for global warming. Thus, to reduce cement consumption, researchers are modifying conventional cement concrete by using various supplementary cementitious materials (SCMs). Among various SCMs available, fly ash (FlA) has been the most popularly used SCM. The use of FlA in the concrete industry not only promotes sustainable development by reducing cement consumption but also solves the problem associated with the disposal of FlA. In this study, various ensemble machine learning (EML) models such as random forest, AdaBoost, and gradient boosting have been developed to predict the compressive strength (CS) and ultrasonic pulse velocity (UPV) of FlA-based concrete. Database needed to develop the various models to predict the desired outputs was obtained by the experiments performed. In order to enhance the efficiency of the models, hyperparameter tuning was being done. All the developed models were able to predict the CS and UPV of FlA-based concrete. Comparison between the various models has been done on the basis of the model efficiency parameters and found that the gradient boosting was to be the best predictor whereas random forest to be the substandard.