Integrated formwork removal decision framework for concrete slabs using FBG sensors and machine learning

Accurate prediction of early-age concrete strength is critical for ensuring construction safety and optimizing formwork removal schedules. This study presents an integrated decision-making framework for curing quality management, employing Fiber Bragg Grating (FBG) sensors to monitor internal temperature and drying shrinkage for 28 days. To validate hydration kinetics, a thermo-mechanical coupled analysis was conducted, and a machine learning framework using Gradient Boosting was explored to predict compressive strength. The results indicated that a literature-based dataset limited to 55 points led to an R 2 approaching 1, revealing inherent overfitting due to reliance on ambient rather than internal core temperatures. These observed deviations highlight the necessity of in-situ monitoring. The study concludes that enhancing predictive robustness requires larger datasets synchronized with internal hydration heat records to mitigate overfitting in field applications.