AE Faculty Play Major Role in $16 million Center for Simulating Plasma-Coupled Combustion

8/12/2013 Susan Mumm

The project has been awarded to the University of Illinois at Urbana-Champaign.

Written by Susan Mumm

Several Aerospace Engineering professors are participating in the new $16 million, 5-year Center for Exascale Simulation of Plasma-Coupled Combustion that has been awarded to the University of Illinois at Urbana-Champaign.

The new center will leverage extreme-scale computing to predict how plasmas – a gas that is transformed into a new state of matter when its atoms are ionized – could be used to control combustion. The research may pave the way for cleaner-burning combustors and more reliable and higher performance jet engines.

The National Nuclear Security Administration (NNSA) is funding the Center through the Predictive Science Academic Alliance Program II. One of three Multidisciplinary Simulation centers that NNSA is funding, the Urbana campus center is comprised of researchers from Illinois and the Ohio State University.

AE faculty participating are Profs. Jonathan Freund and Greg Elliott, Associate Prof. Dan Bodony, Research Prof. Tom Jackson, and Assistant Prof. Marco Panesi.
Plasma
Plasma

Freund, who also has an appointment in Mechanical Science and Engineering, co-leads the center with Computer Science Prof. William Gropp, the project’s Principal Investigator and director of the Parallel Computing Institute. Freund will orchestrate the predictive physics modeling and simulations, including the supporting experiments, within a framework of uncertainty quantification.

“Plasmas offer a little explored means of tuning combustion to meet engineering objectives of performance or efficiency,” he said. “Harnessing the power of forthcoming computer architectures, as is planned within this center, will enable truly predictive simulations that can advance this technology.”

Bodony is the Chief Software Architect and part of the Center’s Executive Committee with Gropp, Freund, Elliott, and Associate Prof. Luke Olson of Computer Science.

The purpose of the Center is to develop advanced software for future exascale computers that will be 1,000 times faster than petascale computers, like Blue Waters on the Urbana campus, Bodony said. Due to power consumption constraints, exascale computers will be made with a variety of computing elements. The expected arrangement will contain a mixture of multi-core central processing units (CPUs) and accelerators. In current-day computers, general purpose graphical processing units (GPGPUs) act like accelerators.

CPUs are good at managing the information flow within a computer but consume lots of power, while accelerators are good at performing arithmetic operations very quickly with little power consumption. Developing fast and efficient scientific software for future heterogeneous computers with processor counts beyond ten million is a leading edge topic in computational science, Bodony said.

The Center’s foundation will be a research code that Bodony led in developing. The code’s sufficient modularity and data structure abstraction provide flexibility the exascale architecture requires, and can be built on top of the pre-existing codebase.

Bodony’s specific role will be to oversee the Center’s primary deliverable – its scientific code and the tools used to develop it – by connecting the computer scientists with the physical application so that the Center can predict plasma coupled combustion using code harnessing future exascale computer’s performance.

Plasma with cross flow
Plasma with cross flow
Simulations of integrated multi-physics phenomena require the synthesis of model representation of many physical mechanics. These are often approximate, and therefore rely upon experimental effort. This is where Elliott’s expertise comes in. He will lead a mix of low-dimensional, physics-targeted experiments to develop and calibrate models, and full-scale, physics-integrated experiments of the ignition of a turbulent jet to evaluate the quality of the full simulation predictions.

Jackson will provide modeling expertise, which will be used both in interpreting the simulation results and in validating their correctness.

A number of model problems will be developed to provide additional verification and validation cases. These model problems will be semi-analytical in nature, in that simplified numerical methods can be used, and these solutions can then provide benchmark cases for code verification and validation, Jackson said.

These model problems also can be used to investigate underlying physical mechanisms, such as flame stability, flame propagation, plasma sheath modeling, electrode surface interactions, etc.

The presence of a plasma can couple with the combustion in one of three primary ways:

  • through a body force term in the momentum equation, called the Lorentz force
  • by producing radicals that can change the chemical pathways
  • or by localized heating by the plasma.

To better understand each mechanism in plasma-coupled combustion, these mechanisms can be switched on or off independent of each other in the model problems, Jackson said.

Panesi,who joined AE a year ago, is an expert in integrating kinetics models in to fluids simulations and quantifying their uncertainty, and will interact collaboratively with the center.

In a normal combustion event, many steps occur between the spark and the firing of an engine. Control of the intermediary steps is not possible with current technology. However, plasma has properties that enable intervention at intermediary steps. Plasma can create the same chemical species that occur during normal combustions and also can produce heat during the different phases, making the chemical process happen faster.  

By using plasmas as a control mechanism, researchers believe they can manage the chemical process, thereby reducing emissions of greenhouse gases into the environment. Plasmas could also help stabilize flames for hypersonic, high-speed jet engines, in which air passes through so fast that the flame can be extinguished.

Understanding how to manage plasma is a difficult problem, requiring three-dimensional, fluid computer simulations that can cover many space and time scales. To make reliable predictions, researchers need scalable petascale computational resources to model and analyze the physics components, which range from flow turbulence to electrodynamics.

“You have to be able to understand what’s happening at the atomic scale all the way up to the bulk flow in the plasma, which you can measure with a ruler,” said Gropp. “We can’t do this as one big computation, so we have to create new techniques that will help us stitch everything together.”

Established by Congress in 2000, NNSA, a semi-autonomous agency of the U.S. Department of Energy, is responsible for enhancing national security through the military application of nuclear science. NNSA maintains and enhances the safety, security, reliability and performance of the U.S. nuclear weapons stockpile without nuclear testing; works to reduce global danger from weapons of mass destruction; provides the U.S. Navy with safe and effective nuclear propulsion; and responds to nuclear and radiological emergencies in the U.S. and abroad.


Share this story

This story was published August 12, 2013.