Episode 15: This Time, it Actually is Rocket Science: Computational Tools for Modeling Combustion

Navigate Left
Navigate Right
  • flame.1

Navigate Left
Navigate Right

A.J. Fillo is in his final year of his PhD in Mechanical Engineering in the School of Mechanical, Industrial, and Manufacturing Engineering, within the College of Engineering. Working with Dr. Kyle Niemeyer. A.J. is studying combustion, or how things burn; specifically, A.J. is working to better understand how the microscopic motion of molecules impacts the type of combustion that we use in jet engines.

            To understand combustion, first it’s helpful to understand energy.  If you drop a ball at the top of a hill, it will roll to the bottom, if you place a tea bag into a hot glass of water, the flavors will move through the water until you have tea. Both of these processes take something from its high energy state, to a more stable lower energy state. In our tea cup, molecular diffusion is what moves that energy around. Diffusion is the process of molecules becoming evenly dispersed by moving from high to low concentration and happens at very small scales, and affects everything around us including the combustion that we use in jet engines.

Diffusion is only part of the story though.  In fluid mechanics, the study of how gasses and liquids move around, diffusion controls the smallest aspects of motion but what processes control motion on a larger scale? To answer that A.J. used the example of an airplane wing. In physics class, many of us have seen a drawing or a demonstration of an airplane wing with smooth streaks of air flow over it, we call those smooth air streaks streamlines.  These smooth streamlines represent something called laminar flow, which is very smooth and predictable, but fluid flows are rarely predictable, usually they are swirly, changing, and chaotic.  These chaotic flows are called turbulence and exist all around us, they cause planes to bounce around when we fly through rough air, they drive the little vortex tornado the forms when our sink drains, and they can even impact the motion, structure, and chemistry of a jet fuel flame.

Both turbulence and diffusion work to move energy around in combustion, but we don’t yet have a firm understanding of how these two different processes interact to control the combustion we use to propel us through the air.

Turns out, flames are hard to study because as you can imagine, anything you would use to measure a flame, does not want to be in a flame; measurement tools like thermocouples and pressure transducers can melt, or even combust themselves.  But there is another tool at our disposal.  We can use super computers to simulate how combustion is happening in jet engines and even use it to study how turbulence and diffusion interact, or how molecules are moving around during combustion.

A.J.’s research focuses on developing computational tools to look at these effects. The sum total of reactions happening during jet fuel combustion are large and complex, meaning that the equations are not easy to solve, and trying to do so can take thousands of computer cores for several days. By developing a more efficient computer algorithm to look at these reactions we can make these simulations faster, more efficient, and less expensive.

In reality, Jet fuel is a mixture of hundreds of different chemicals, so to simplify things, A.J. uses fuels like hydrogen (H2), n-heptane (H3C(CH2)5CH3), and toluene (C6H5CH3) as representative fuels. Although a single, simpler compound, even as simple as just hydrogen, has hundreds of chemical reactions and dozens of different radical molecules that form during its combustion. To get around the limitation of computer memory and speed up how quickly his simulations run, A.J. created an algorithm to optimize how the computer handles the math to make sure things run as smoothly as possible.  You can think of it a bit like going to the DMV, usually the line takes forever because people are rarely ready with their paper work in hand when they get to the front of the line, instead people must get out of line, get more paper work, and start over.  Using this analogy, A.J.’s algorithm works to make sure everyone in line arrive with their paper work completed, ready to hand off, and let the next person through. This reduces dramatically reduced the amount of computer memory needed to solve these combustion simulations and speeds up the math.

3D simulation results for a turbulent, premixed, hydrogen air flame looking at peak flame temperature colored by chemical composition. Flow is from back to front

A.J. became interested in mechanical engineering because of his love of magic. A.J. started his academic journey at the University of Missouri Columbia as a journalism major but transferred to OSU for the engineering program. A.J. has always loved performing, which is why science outreach has been such a large part of his graduate school experience. Partnered with the Corvallis Public Library, A.J. hosts LIB LAB, a hands-on multimedia educational YouTube series focused on STEAM (science, technology, engineering, arts, and mathematics) education, which he previously talked about on our GRADx event.

To learn more about A.J. Fillo’s research, or more about Inspiration Dissemination, visit blogs.oregonstate.edu/inspiration or tune in Sundays at 7 PM at 88.7 KBVR-FM.

 (Icon Image by A.J. Fillo’s masters work, and an photo-art series A.J. did on combustion, Turbulent, premixed jet fuel air Bunsen burner with a fuel rich jet fuel air flame. Fuel is commercially available ‘Jet-A.’ Photo shot at 1/8000 second shutter speed and aperture of f/2.8)