Particulate media, or granular materials, are random aggregates of a large number of solid particles. They come in all sizes and encompass a broad range of naturally occurring materials, such as gravel, sand, and soil, and a broad range of industrial and agricultural products, such as grains, fertilizers and powders, and pigments. These materials pose special challenges to handle on a large scale because they can exhibit an array of solid-like and fluid-like behaviors. A key question in designing such materials is how to pick the constituent particles to achieve desired material properties of the aggregate. This project will develop a new approach that will allow practitioners to engineer particle types that will yield predetermined, target outcomes for the behavior of the aggregate. The project will involve a combination of computations and experiments to solve the problem of particle design and to develop a set of general design rules. The project will provide opportunities for students at all academic levels to participate in research. The researchers will participate in a Bridge Program that encourages undergraduates, especially those from underrepresented groups, to continue to graduate work. In addition, the team will participate in several outreach activities to increase the public's appreciation of science and engineering, including the Physics with a Bang! event, which demonstrates science principles to broad audiences.
To tackle the inverse problem of identifying appropriate particle shapes for given performance targets, the project implements a computer-aided design and optimization method that makes use of ideas from evolutionary computation. Going beyond devising specific solutions to specific design targets, the project introduces new capabilities by developing general design rules. Such rules apply across a whole range of related targets and, once established, enable rational design by providing a mapping from overall behavior of the particulate medium to particle-level attributes, such as shape. For the design of random composite or particulate systems with complex particle shapes this is a new approach with transformative potential. In this project,arbitrary convex and non-convex shapes will be represented by sets of bonded spheres of varying radius and overlap, whose configuration can be mutated and evolved to optimize performance. Directions to be investigated include distributions of particle sizes and shapes, particle breakage, particles combining hard and soft materials, designing the stress response in the nonlinear regime of incipient failure, and optimizing not only the particles but also the processing pathway. In a second stage of the project this methodology is extended to designing flow behavior, focusing on concentrated particle suspensions. For calibration and validation the computational effort is closely coupled to systematic experiments that use high-resolution 3d-printing to fabricate particles and use x-rays and ultrasound for non-invasive imaging.
