High-Fidelity Prediction of Temperature and Time-Dependent Heat of Hydration of Blended Cementitious

Описание: High-Fidelity Prediction of Temperature and Time-Dependent Heat of Hydration of Blended Cementitious Pastes Using Machine Learning: A Mass Concrete Application
Presented By: Luna Al Hasani, Georgia Institute of Technology
The temperature and time-dependent heat of hydration of cementitious pastes is a fundamental concept in understanding the driving mechanisms of cement hydration. Higher curing temperatures increase the rate of heat release, which despite leading to an increase in mechanical properties at early age, may also result in durability issues when concrete structures are massive. Models which currently depict the heat of hydration rely on empirical fits or kinetic based approaches, which consider the different physiochemical characteristics of the cementitious pastes and curing temperatures in their formulation. To advance the durability and sustainability of concrete, mineral additives have been widely promoted as partial replacements of ordinary Portland cement, which can exert varying effects on hydration kinetics. This adds a level of complexity to the modeling efforts. In recent years, machine learning has emerged as a promising potential to optimize the prediction of properties of different material systems by learning through data and statistical methods. Here, Gaussian process regression (GPR) has been used to predict the time and temperature dependent heat of hydration for cementitious pastes, by taking the curing temperature and physiochemical properties of the paste as input features. Results show that GPR can predict the output criteria with high-fidelity when compared with experimental heat of hydration time histories. The predicted outputs show good performance when upscaled for use in mass concrete thermal modeling. This creates an opportunity to apply data-driven approaches to perform property predictions of cement pastes and facilitates the selection of mixture designs to satisfy certain performance criteria.