Machine learning-based integrated maturity model for strength estimation of blended cement concrete

While conventional maturity models utilize the temperature history of concrete to predict strength development in construction environments, no globally applicable model currently exists. Consequently, model parameters must be experimentally determined for each new mix design, presenting a significant limitation in practical applications. To address this issue, the present study proposes a machine learning-based approach to predict strength development parameters, enabling generalized application across a wide range of mix designs. A series of practical mortar and concrete mixtures were prepared to derive strength development parameters using three different maturity models. A comprehensive parametric analysis was conducted to evaluate the effects of key variables, including supplementary cementitious material contents and curing temperatures. The derived parameters were validated against existing literature to ensure their reliability. A Random Forest regression model was then developed to predict strength development parameters based on mix proportion, age, and temperature history. The model demonstrated the highest prediction accuracy for compressive strength using the time–temperature-dependent apparent activation energy model.