The effectiveness of surrogate models in predicting the long-term behavior of varying compressive strength ranges of recycled concrete aggregate for a variety of shapes and sizes of specimens

One of the most fundamental developments in improving the mechanical properties of concrete is the introduction of recycled coarse aggregate, which offers an environmentally preferable substitute for traditional waste management techniques. Using recycled coarse aggregate and a small number of mix proportions for the concrete components, a few studies looked at the mechanical properties of concrete. To assess the impact of recycled coarse aggregate on the long-term compressive strength of concrete at various mix proportions and different compressive strength ranges, this study analyzed four models, including linear regression (LR), nonlinear regression (NLR), pure quadratic (PQ), and full quadratic (FQ). Three datasets training, testing, and validating, each containing 314 data points culled from various studies, were used to apply the models. The recycled coarse aggregate (RA) density ranged from 0 to 1240 kg/m3, and the curing time (t) varied from 1 to 90 days. While the predicted compressive strength of the models ranged between 5 and 75 MPa, the compressive strength of the data gathered from the experimental work of several studies ranged from 8 to 78 MPa. The models’ accuracy was assessed using several metrics, including the coefficient of determination (R2), the root-mean-square error (RMSE), the scatter index (SI), the objective (OBJ), and the mean absolute error (MAE).

---

Opracowanie rekordu ze środków MEiN, umowa nr SONP/SP/546092/2022 w ramach programu "Społeczna odpowiedzialność nauki" - moduł: Popularyzacja nauki i promocja sportu (2022-2023)