Aaron Morris' NSF CAREER grant will focus on simulating the flow of irregularly-shaped particles of many different sizes. Advancing this fundamental science could apply to pharmaceuticals, food science, manufacturing, agriculture, and many other fields.

Morris, an assistant professor of mechanical engineering, focuses on developing mathematical models to predict the behavior of granular flows. “Understanding how particles flow is very challenging,” said Morris. “In particle flows, there are many instabilities and flow structures, such as clusters, that are not observed in traditional fluids like liquids or gases.  Such structures and large-scale behavior depend on interactions that occur at the scale of an individual particle.  Understanding the physics at the particle scale, such as collisions between irregularly shaped grains, and how they are interconnected to the bulk flow dynamics, is a key challenge for modeling granular flows.”

Most current theory and models of particle flows assume a perfectly spherical shape for the particles, which makes calculations easier but doesn’t accurately reflect how they interact. “When irregularly-shaped particles collide with each other, they exchange energy in many different ways between rotational and translational mode,” he said. “Those exchanges of energy in a granular system make for a lot of interesting physics. My work is on the fundamental side, to develop models for that behavior, and connect those models to existing theory.”

For his NSF project, Morris will focus on one shape in particular: a spherocylinder, which is a cylinder with hemispherical caps on both ends. “It’s essentially a sphere that’s been extended along one axis,” said Morris. “That allows us to start with a foundation of existing theory about sphere-shaped particles, and as the cylinder gets larger, we work our way toward more irregular shapes and the more complex physics involved.”

His work will utilize Purdue’s powerful research computing clusters, to dig into the mathematics of both physics and probability. This is out of necessity, because simulating individual particles is extremely computationally intensive. “Current discrete element methods for modeling non-spherical particles are limited to fewer than a million particles,” said Morris. “By comparison, a single cup of sand has more than 100 million particles. So we start by building high-fidelity discrete element simulations, and then use machine learning to build probability distribution functions, which can then predict the dynamics of billions of particles.”

Morris is also looking forward to building some small-scale demonstration experiments, showing some of the practical applications of this theory as an outreach activity. “Many people may not realize how pervasive particle technology is used in the products we all use on a daily basis," said Morris. "From coffee beans and cooking spices to baby powder and laundry detergent, they all utilize particle technology. For example, many pharmaceutical companies use 'fluidized beds,' mixing solids to create a highly uniform product." As part of his NSF project, Morris will work with students to build small-scale devices that are both fun and demonstrate the rich physics that arise in particle flows.

Writer: Jared Pike, jaredpike@purdue.edu, 765-496-0374

Source: Aaron Morris, abmorris@purdue.edu, 765-494-0020