CAREER: Physics-Informed, AI-Driven Prediction and Optimization of Concrete for Resilient Infrastructure

This Faculty Early Career Development Program (CAREER) award will advance the predictive design of concrete, the most widely used construction material, by developing new data-driven approaches to improve performance, cost efficiency, and resource utilization. Concrete production currently relies on empirical, trial-and-error methods that often lead to suboptimal mixtures and increased costs. These limitations are becoming more significant as the industry incorporates diverse supplementary materials derived from industrial byproducts and natural resources, introducing greater variability and complexity. This project will integrate data science with fundamental materials science to enable faster, more reliable, and more adaptable concrete design. By supporting the use of locally available materials and reducing reliance on standardized formulations, the work will enhance efficiency and flexibility in infrastructure development. The project also integrates research and education through a novel design challenge that engages undergraduate and high school students in solving real-world engineering problems while developing skills in data science and materials design. These efforts will expand participation in science and engineering and contribute to a future-ready workforce aligned with national priorities. The project will develop physics-informed, data-driven frameworks for the predictive and inverse design of blended cement concrete systems. The research will: (1) build a large-scale, open-access data infrastructure through automated literature mining and guided experiments; (2) develop mechanistic descriptors of binder reactivity and porosity by integrating atomistic simulation, diffraction-based characterization, thermodynamic modeling, and machine learning; and (3) incorporate these descriptors into predictive models to enable accurate property prediction and multi-objective optimization of concrete mixtures. These models will support rapid identification of optimal mixtures that meet performance targets while minimizing cost and environmental impact and maximizing resource efficiency across diverse material systems. The resulting framework will improve the reliability, interpretability, and transferability of data-driven models and establish a foundation for next-generation materials design. This work will advance fundamental understanding of cementitious materials while enabling scalable, automated approaches to infrastructure material design. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria. NSF Award ID: 2543840 | Program: 01002627DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Kai Gong | Institution: William Marsh Rice University, HOUSTON, TX | Award Amount: $669,852 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2543840 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2543840.html