Search

The development of design guidelines for structural recycled aggregate concrete (RAC) is hindered by the absence of a unified elastic modulus model focusing on the properties of diversely-sourced coarse recycled aggregates (CRA). This study develops a unified elastic modulus model of RAC and natural aggregate concrete (NAC) considering four inputs: CRA dosage, coarse aggregates/cement content, effective water/cement content, and coarse aggregates water absorption. Ten machine-learning algorithms are evaluated through various statistical approaches to explore the most suitable algorithm by establishing a database from 72 studies (i.e., 633 test results). Based on the results, CRA properties are significant for the elastic modulus prediction of RAC. All machine learning models perform better than existing models, and GBA, BRA, and XGBA models are the most suitable algorithms. The pioneering unified design equations proposed in this work can efficiently predict the elastic modulus of RAC and NAC, leading toward sustainable and cleaner concrete design.