To investigate concrete drying shrinkage in high-altitude environments, moisture evaporation and shrinkage rates were examined under combined curing regimes of four temperatures (40 °C, 20 °C, 0 °C, −10 °C) and three relative humidities (RH40%, RH60%, RH80%). Curing temperature and humidity primarily regulate shrinkage deformation by altering the internal moisture evaporation rate. Both evaporation and shrinkage rates exhibited a rapid initial increase, followed by deceleration, and finally stabilization with increasing age. A strong positive correlation was observed between these two parameters. The high-temperature and low-humidity condition (40 °C, RH40%) induced the most severe shrinkage. Four machine learning algorithms (XGBoost, RF, ANN, and KNN) were used to construct prediction models. After hyperparameter optimization and cross-validation, the RF models exhibited superior generalization and robustness (test set R2 > 0.94). The model accurately captures the complex non-linear relationship between environmental parameters and shrinkage. SHAP analysis on the optimal models identified the moisture evaporation rate as the primary driving factor. The analysis quantified the non-linear contributions of temperature and age, alongside the inhibitory effect of humidity. The study verified the consistency between data-driven models and physical mechanisms. This study elucidates the shrinkage mechanism under extreme conditions. It provides a reliable reference for crack control and life prediction in high-altitude engineering.
Zhang et al. (Mon,) studied this question.