Comparative analysis of the influence of partial replacement of cement with supplementing cementitious materials in sustainable concrete by using machine learning approach

Abstract

Abstract Cement manufacturing is a major contributor to climate change because of the greenhouse gas carbon dioxide released into the atmosphere throughout the process. In this paper, cement content of concrete has been partially replaced by using two supplementing cementitious materials (SCMs) materials like Silica Fume and Fly Ash. Characterizations of both materials has been conducted for their end use utilization in concrete applications. Extensive experimentation has been conducted to ensure the effect of partial replacement on the performance characteristics of concrete through compressive strength, flexural strength, and split tensile strength of concrete. It was observed that both the waste material has the ability to replace cement content without changing the performance of concrete. Finding indicating that replacement with proper mix design can improve the strength of green concrete. Silica fume have better response as compared to fly ash replacement on the strength characteristics of green concrete. Accuracy of experimental data has been validated by using machine learning approach. Experimental results are used to train the machine learning models. Metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), R2 Score, and Cross Validations are used to evaluate the performance of models. According to the findings, the extreme Gradient Boosting Regression model performs better than any of the other models when it comes to predicting and validating the compressive strength, flexural strength, and Split tensile strength of green concrete mixtures. It achieves an R2 value of 0.9811 for the prediction of the split tensile strength, 0.9818 for the flexural strength, and 0.9127 for the compressive strength. The findings of this research shed light on the usefulness of regression models for predicting the properties of green concrete and for validating such predictions with experimental results in terms of accuracy. The replacement of 10–15% for both SCMs resulted good agreements for strength characteristics.