Answers are grounded in this paper and cite their sources.
The interlayer bond performance of self-consolidating concrete (SCC) strongly influences the structural integrity of multilayer concrete elements, particularly under delayed casting. Accurate assessment of interlayer bonding is challenging due to time-dependent rheology and stress-dependent failure mechanisms. This study combines experimental and machine learning (ML) approaches to investigate SCC incorporating supplementary cementitious materials (SCM), including Class F fly ash (FAF), ultra-fine Class F fly ash (UFAF), Class C fly ash (FAC), and granulated blast furnace slag (GBFS) at 15–60% replacement. Altogether, 2295 specimens were tested using slant shear strength (SSS), direct shear strength (DSS), and splitting tensile strength (STS) methods under five interlayer intervals and three curing ages. Results showed that longer interlayer intervals reduced bond strength, while extended curing enhanced bonding due to pozzolanic reactions. The SSS test showed limited sensitivity to interface degradation, whereas DSS was highly responsive to delays and SCM type, detecting weak interfaces. STS exhibited intermediate sensitivity, with UFAF providing the greatest tensile bond improvement. Five ML models Gradient Boosting Regression (GBR), Multilayer Perceptron, Extreme Gradient Boosting, Random Forest, and Decision Tree were developed to predict bond behavior, with GBR achieving the highest accuracy (R2 > 0.96). Comparative analyses indicated that dominant predictors varied across tests: curing age governed SSS, interlayer interval influenced DSS, and UFAF content affected STS. SHAP and partial dependence analyses confirmed these trends, emphasizing the stress-dependent nature of interlayer bonding. The proposed experimental-ML framework offers a reliable basis for performance-based SCC mix design and prediction of interlayer bonding in multilayer construction.
No takes yet. Share an insight, caveat, or question.
Reza Saleh Ahari (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: