University of Minnesota
University Relations
http://www.umn.edu/urelate
612-624-6868

Minnesota Supercomputing Institute


Log out of MyMSI

Research Abstracts Online
January 2008 - March 2009

University of Minnesota Twin Cities
Institute of Technology
Department of Mechanical Engineering

PI: Sean C. Garrick, Associate Fellow

Modeling and Simulation of Turbulent, Reacting, Multi-phase, Multi-scale Flows

Computational fluid dynamics (CFD) approaches are attractive in that they are well established, with many algorithms and schemes that have been ported to number of performance computing platforms and are used to solve, model, and investigate a wide variety of phenomena. These researchers are developing models and methodologies to improve the prediction and fidelity of turbulent, reacting, multi-phase flows. At the nano-scale, for example, knowledge of interaction between the fluid and particle fields is scant at best. The underlying multi-scale/physics nature of the physical processes are often absent. Current research underway includes the development of hybrid discrete-continuous approaches to predict nano-particle size (volume) and shape (surface-area). In addition to the mathematical/physical models, a great deal of effort is directed towards developing the appropriate algorithms to efficiently couple the two approaches resulting in the resolution of all appropriate length and time scales with a high degree of fidelity to real world processes. Scalable algorithms and the deployment on tera- and peta-scale computing platforms make this combined approach feasible. The results are computational tools that provide a level of detail otherwise unavailable. This includes information about the time/temperature-history, internal structure, chemical composition, crystalline phase, of the particles and the effects of fluid turbulence on these properties.

These computational tools are directed towards addressing two pressing challenges: the need to reduce anthropogenic emissions of greenhouse gases and toxic compounds due to combustion of fossil fuels and the need to meet an expanding global demand for energy. For example, it is estimated that through the condensation of mercury on powder activated carbon injected into coal-fired power plant exhaust, and the removal of the particles via electro-static precipitation, reductions of at least 85% in anthropogenic mercury emissions can be achieved. Unfortunately the mercury-to-particle condensation process — especially in turbulent flows — is not very well understood as current models have not been validated and full-scale tests are prohibitively expensive. Simulation has the potential to speed-up research and development significantly. Research thus far suggests that while smaller particles are better dispersed and therefore better mixed with the mercury vapor, larger particles actually remove more mercury from the combustion exhaust gas. The use of nano-structured particles appears to be promising. The idea is to engineer particles with large pores that would increase the rate of mercury absorption and allow more mercury to be removed from the exhaust gases. Such applications lie in the "sweet-spot” of mechanical engineering and the role that CFD and supercomputing play in turning nano-scale science and engineering into tomorrow’s technology is significant.

Group Members

Michael Buhlmann, Graduate Student
Shankhadeep Das, Graduate Student
Takumi Hawa, Research Associate
Juha Kurkela, Research Associate
James Linden, Undergraduate Student
Jun Liu, Graduate Student
Jason Loeffler, Undergraduate Student
Pat Mathiesen, Undergraduate Student
Saurav Mitra, Graduate Student
Nathan Murfield, Graduate Student
Jouni Pyykonen, Research Associate