Understanding the physics of complex, turbulent multiphase flows fascinates Olivier Desjardins. His research helps improve both traditional fuel-engine systems as well as newer biofuel combustion systems by focusing, among other things, on unraveling the physics behind the atomization process—in which fuel turns to tiny droplets before being burned. “I was always interested in this idea of being able to harness large-scale computer resources in order to reproduce what happens in nature—this idea of simulating real life,” Desjardins says.
Liquid atomization is still partially a mystery. “We don’t fully understand what’s happening” says Desjardins. “We can get droplets to form fairly efficiently, but we don’t know how to predictively model this process, how to reliably control the droplet size.” Without accurate models, the ability to test and improve combustions systems is greatly handicapped. Modeling atomization is also crucial to biofuel development, as these new and relatively untested fuels might behave differently than conventional fossil fuels. “We need to understand how those new fuels with new properties will affect the atomization performance of current engines,” Desjardins says. Traditionally, the industry simply builds a physical system and tests it—but this is expensive. Computational models can provide a thrifty alternative.
Desjardins helps to provide this alternative to commercial partners. Recently, his team helped General Electric embrace high performance computing to model their fuel injectors in order to ultimately reduce the development time for such systems and increase flexibility for biofuels. Desjardins is also creating models for Shell which determine the behavior of sand particles in oil pipelines. “Their question was, how strongly should we pump our mixture of oil, gas, and sand to avoid the formation of a static sand bed at the bottom of the pipeline?” he says. Using detailed models that calculate how these sand grains move and distribute through the turbulent slurry of oil and gas, Desjardins has some answers. “We try as much as possible to transfer the techniques we use to the industry,” he says.
Some of Desjardins’ work is too early-stage to be used in machines today, but it builds a foundation for advances in the future—such work includes electrohydrodynamic atomization, which harnesses the energy from an electric field to control how a fuel becomes atomized into droplets. With this capability, atomization could be improved and controlled, potentially increasing engine efficiency by leaps and bounds. It can also be used for space propulsion, in which droplets of fuel (controlled by an electric field) are ejected one at a time at high velocity to power highly efficient spacecraft. “We want to determine if we can use the electric field to help us control how big the droplets get, and where they go,” says Desjardins. “We’re adding electrophysics on top of the fluid equations, it’s a topic we’re pretty excited about.”
This gets at the core reason Desjardins got interested in this field to begin with—exploring the unknown. “These multiscale, multiphysics, and multiphase processes are so highly nonlinear and complicated and chaotic, that we have little idea what really happens. We can see the outcome—in the case of atomization, we see that we’re forming droplets—but we cannot provide a comprehensive theory describing how this takes place,” he says. “So when we create these simulations that no one has reproduced before, we can be pretty confident that we are the first people in the world to see how this works.”