Collaborative research: a discrete modeling framework for fracture and damage prediction in structural steel

The National Science Foundation Division of Civil, Mechanical, and Manufacturing Innovation awarded the University of Arkansas $155,000 on January 1, 2027, under the Engineering program (CFDA 47.041) to develop a discrete modeling framework for predicting fracture and damage in structural steel.

The research integrates a discrete particle representation of steel microstructure within finite element simulations to capture crack initiation, propagation, and arrest with mechanistic fidelity. The work comprises five coordinated tasks: developing a mesoscale representation of steel microstructure from microscopy observations; performing high-throughput microscale experiments to quantify mechanical properties of individual phases and interfaces; and formulating and implementing a discrete lattice model embedded within conventional finite element simulations. The framework will produce open-source software, educational materials, and publicly available guidance for researchers, practicing engineers, and standards organizations. Graduate students will receive interdisciplinary training in infrastructure and materials engineering.

More accurate and broadly applicable fracture predictions will enable engineers to design structures that perform more reliably during earthquakes and extreme weather while reducing reliance on costly physical testing and accelerating development of stronger, more fracture-resistant steels.

Performance occurs at Fayetteville, Arkansas, with a period of performance from January 1, 2027, through December 31, 2029.

Abstract

Failures caused by fracture or cracking in steel are difficult to predict, making IT harder to design buildings, bridges, and other critical infrastructure that are safe, economical, and resilient. Current computer models often simplify how steel behaves under stress; consequently they often produce unreliable predictions of fracture and failure when materials or loading conditions change. This project develops a new way of modeling steel that captures how its internal material structure influences the formation and growth of cracks. By making predictions more accurate and broadly applicable, the research will help engineers design structures that perform more reliably during earthquakes, extreme weather, and other severe events. The resulting advances will support safer and more resilient infrastructure while reducing the need for costly physical testing and accelerating the development of stronger, more fracture-resistant steels. Open-source software, educational materials, and publicly available guidance will make these tools widely accessible to researchers, practicing engineers, and standards organizations, while graduate students will receive interdisciplinary training that prepares them to address future challenges in infrastructure and materials engineering. The research develops a multiscale computational framework that integrates a discrete particle representation of steel microstructure within conventional finite element simulations to capture crack initiation, propagation, and arrest with mechanistic fidelity. The work includes five coordinated tasks: developing a mesoscale representation of steel microstructure from microscopy observations; performing high-throughput microscale experiments to quantify the mechanical properties of individual phases and interfaces; formulating and implementing a discrete lattice model that is embedded within continuum structural analyses; validating the framework through a comprehensive experimental program spanning multiple fracture scenarios; and releasing open-source software, calibration tools, and instructional materials. The framework establishes direct connections between microstructural morphology, constituent material properties, and structural-scale fracture behavior while maintaining computational efficiency for engineering applications. By replacing empirical calibration with experimentally measured microstructural information, the research establishes a new class of predictive models that is applicable across many different loading conditions and fracture modes. The resulting methodology will provide a foundation for computationally designing and optimizing future steel microstructures for improved fracture resistance while advancing the fundamental understanding of fracture mechanics across multiple length scales. 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.- subawards are not planned for this award.