Automated analysis and design for recycled aggregate reinforced concrete beams.

The goal of this study is to propose a robust method for the analysis and design of recycled aggregate reinforced concrete (RAC) beams. For this purpose, a machine‐learning framework, named, GA‐XGBoost was designed. A comprehensive dataset of 276 tests of RAC beams was compiled from available literature for training, validating, and testing the models. The results confirmed that the proposed models can offer a precise tool for failure mode classification and capacity prediction of RAC beams. For the design framework, a Generative Adversarial Network for tabular data was utilized to produce synthetic data. These synthetic data can provide options for the selection of design variables. In this framework, the design variables are selected by specifying the design constraints, the objective function to be optimized, and validating the capacity requirements using the developed GA‐XGBoost models. Lastly, a web application was developed to make the proposed GA‐XGBoost models appropriate for practical design. [ABSTRACT FROM AUTHOR]