Comparative strength estimation model of recycled aggregate concrete modified with GGBS, Metakaolin, and fly ash

Recycled aggregates (RA) can provide a sustainable solution for replacing natural aggregates (NA) in the concrete mix. Nevertheless, because of a lack of databases, inspection experts and stakeholders are not confident in their ability to forecast their compressive strength (C_s). The majority of them concentrate only on using natural aggregates in the concrete mix. However, numerous scholars have offered other mix designs, recycled aggregate concrete (RAC) still cannot employ recycled aggregate (RA). One is the absence of a simple and effective RAC compressive strength forecast. This research predicts the C_s of concrete mixed with RA comprising fly ash, powdered granulated blast furnace slag, and metakaolin using integrated random forests (RF) models. A complete database for the creation of models was acquired from the literature. In this study, Dwarf Mongoose Optimization (DMO) and Flow direction algorithms (FDA), linked with the RF model purposed to ascertain the most favorable values for the hyperparameters of RF (RFDM and RFFD). By combining the strengths of RF models with optimization algorithms, the proposed approach offers a robust, accurate, and adaptable solution for predicting the compressive strength of recycled aggregate concrete, addressing the limitations of existing methods and facilitating the wider adoption of sustainable construction practices. The data clearly demonstrates that the RFFD and RFDM techniques have a remarkable capacity to reliably predict the Cs of RAC. During the training and testing stages, the R² values for the RFDM system were found to be, respectively, 0.9881 and 0.9921. Because the RFDM's OBJ was 1.6775 lower than the RFFD's OBJ of 2.5512, the findings of this index as a comprehensive one suggest that the RFDM may be the most effective framework.