Predictive performance assessment of recycled coarse aggregate concrete using artificial intelligence: A review

Abstract

Recycled coarse aggregate concrete enables the creation of environmentally friendly and cost-effective mixes. It helps address the disposal problem of demolition concrete waste, meeting demand while improving product functionality and reusability. The abundance of obsolete buildings in cemeteries contributes to Construction and Demolition waste. Recycled Concrete Aggregate (RCA) from demolished structures can be utilized as aggregates, albeit with concerns about its impact on compressive strength due to absorption issues. This review aimed to study and develop the different Artificial Intelligence (AI) model for the prediction of the compressive strength of concrete with varying RCA content and natural coarse aggregate content as input parameters while compressive strength as output parameter. The range of the input parameters is 0 % to 100 % while the range output parameter is 28 MPa to 70.3 MPa. Experimental data from literature articles used to train and validate the model development. Engineers and researchers can utilize these models to predict compressive strength by changing the input parameters. XGBoost Regression Model performed well with R2 0.93594 followed by Random Forest Model with R2 0.92766, and Gradient Boosting Model with R2 0.90616 respectively. Ridge Regression, Lasso Regression, and Linear Regression Models were not performed well in predicting the compressive strength of RCA concrete with R2 0.57657, 0.57558, 0.57675 respectively. ANN also performed significant in prediction of RCAC compressive strength with R2 0.8039. Future research could focus on optimizing the mechanical properties of concrete containing RCA using AI models. Furthermore, the study extends its analysis to explore the application of AI in predicting the strength of various types of concrete, highlighting the versatility and potential of AI-driven approaches in enhancing concrete mix design.