The durability of cementitious materials in aggressive acidic environments remains a critical challenge for infrastructure systems, primarily due to the complex and nonlinear interactions between mixture design parameters and degradation mechanisms. Despite extensive experimental investigations, the predictive understanding of these interactions and their combined influence on mechanical performance is still limited. In this study, an integrated experimental–computational framework is developed to systematically evaluate and predict the mechanical behavior of cementitious composites under acidic exposure conditions (pH = 3 and 6). A comprehensive experimental program was conducted to measure compressive, tensile, and flexural strengths at different curing ages, alongside durability-related indicators.To address the limitations of conventional empirical approaches, several machine learning models including. Random Forest, Gradient Boosting, XGBoost, and Elastic Net were implemented to capture the nonlinear relationships between input variables and performance responses. Among these, the Random Forest model exhibited the highest predictive accuracy (R2 up to 0.95), demonstrating its robustness in modeling complex material behavior. Experimental results showed that compressive strength increased from approximately 20 MPa to 60 MPa with curing time. Exposure to acidic conditions resulted in a reduction of mechanical properties by approximately 20–30%. The results reveal that the interaction between mixture composition and exposure conditions governs performance more significantly than individual parameters alone, while excessive incorporation of supplementary constituents may reduce long-term efficiency due to diminished matrix reactivity. The developed machine learning model achieved a high prediction accuracy with a coefficient of determination (R2) of up to 0.95. The prediction errors were limited to approximately 2.0 MPa for compressive strength and 0.10 MPa for tensile and flexural strength, indicating high reliability. Feature importance analysis further identifies curing time and environmental conditions as the dominant controlling factors. The proposed framework provides new insights into the data-driven design of durable cementitious systems and offers a reliable basis for optimizing material performance in aggressive service environments, particularly for infrastructure applications such as pavements and industrial structures. Sensitivity analysis indicated that curing time was the most influential parameter (45%), followed by environmental condition (35%) and material composition (20%).
Nasr et al. (Thu,) studied this question.