It’s no secret that the use of fossil fuels comes with a high cost to society. The emission of pollutants and noise-generation that result from fuel combustion in aircraft, automotive engines, and power plants are taking a toll. Take just one example: the U.S. military spends more than a billion dollars in health expenses each year to treat the hearing damage that results from the enormous noise-exposure levels on aircraft carriers. The impact isn’t limited to military personnel, however. People living near airports experience higher stress levels and are at greater risk for coronary heart disease and stroke. Assistant Professor Matthias Ihme (pronounced EE-may) is tackling these issues by focusing on improving combustion technology.
Observing and learning about the issues present during combustion isn’t simple. As Ihme points out, “because of the sheer size, complexity and cost, it is difficult to investigate gas turbines in a lab.” His solution is computational modeling: high-fidelity simulation-methods that study combustion as if they were physical processes in realistic environments that would otherwise be very difficult to reproduce in small-scale experimental systems.
The focus of Ihme’s research is understanding the coupling of turbulence, combustion chemistry, pollutant formation and noise emission. As Ihme explains, these detailed simulations enable the resolution of a large range of spatial and temporal scales, which is critical for understanding the complex interplay between different processes. Ihme hopes to reduce not only noise emission from aircraft, but to decrease pollutants and improve fuel economy as well. In this view, even seemingly small improvements add up to large victories. Every percentage of improved fuel economy, extrapolated over millions of miles and gallons, is a major savings, especially with the rising price of oil and ever increasing air traffic.
Because of the complex interaction between turbulence and reaction chemistry in combination with the enormous number of chemical species (the simplest hydrocarbon fuel alone requires the consideration of more than 40 chemical compounds), simulations of turbulent combustion become computationally demanding. Such detailed simulations are only possible with high-performance computing (HPC). Ihme uses ORCI’s shared HPC resource, Flux, to complement his own dedicated 128-core Linux cluster and resources at Oak Ridge National Laboratory and XSEDE (formerly known as TeraGrid). “Providing resources on demand is very useful– they’re easily accessible and scalable,” said Ihme. Flux also proved helpful in satisfying short-term demand during the weeks when Ihme hosted a visiting researcher from Switzerland and China.
With the growing speed and availability of HPC, Ihme sees great potential in high-fidelity simulation techniques. “Not so long ago, combustion and turbulence were both poorly understood. High-fidelity simulation techniques can potentially provide highly accurate simulation results. I am very excited about it — there is enormous opportunity to impact society.”