Application of optimization-based regression analysis for evaluation of frost durability of recycled aggregate concrete

M Esmaeili-Falak and RS Benemaran, STRUCTURAL CONCRETE, 25, 716-737 (2024).

DOI: 10.1002/suco.202300566

Concrete constructed using recycled aggregates in place of natural aggregates is an efficient approach to increase the construction sector's sustainability. To improve recycled aggregate concrete (RAC$$ \mathrmRAC $$) technologies in permafrost, it is essential to certify the stability in frost-induced conditions. The main goal of this study was to use support vector regression (SVR$$ \mathrmSVR $$) methods to forecast the frost durability (DF$$ \mathrmDF $$) of RAC$$ \mathrmRAC $$ on the basis of durability agent value in cold climates. Herein, three optimization methods called Ant lion optimization (ALO$$ \mathrmALO $$), Grey wolf optimization (GWO$$ \mathrmGWO $$), and Henry Gas Solubility Optimization (HGSO$$ \mathrmHGSO $$) were employed for indicating optimal values of SVR$$ \mathrmSVR $$ key parameters. The results depicted that all developed models have capability in predicting the DF$$ \mathrmDF $$ of RAC$$ \mathrmRAC $$ in cold regions. The values of OBJ$$ \mathrmOBJ $$ as a comprehensive index depicted that the GWO-SVR$$ \mathrmGWO-\mathrmSVR $$ model has the higher value at 0.0571 as the weakest model, then ALO-SVR$$ \mathrmALO-\mathrmSVR $$ at 0.0312 recognized as the second-highest model, and finally the HGSO-SVR$$ \mathrmHGSO-\mathrmSVR $$ system at 0.0224 mentioned as outperformed model. ALO-SVR$$ \mathrmALO-\mathrmSVR $$ and GWO-SVR$$ \mathrmGWO-\mathrmSVR $$ approaches were likewise capable of accurately forecasting the DF$$ \mathrmDF $$ of RAC$$ \mathrmRAC $$ in cold regions, but the created HGSO-SVR$$ \mathrmHGSO-\mathrmSVR $$ method outperformed them all when taking into account the explanations and justifications.

Return to Publications page