urbulent flow arises in many important engineering applications involving aircraft, land vehicles, and various combustion and propulsion systems. It is important to understand and predict the nature of
these complex flows in order to improve the performance of such devices. Historically, most designs have been forced to rely heavily on experimental data for guidance because the computer power was not
available to numerically solve the governing equations without invoking limiting
physical approximations. This situation is beginning to change.
 |
| Velocity vectors showing flow pattern in a rib-roughened channel.
Pitch to height ratio of ribs was 7.2; Reynolds number based on hydraulic diameter, 20,000. Red denotes high velocity, blue, low. |
Professor Terry Simon of the University of Minnesota's Mechanical Engineering Department has been working on the experimental aspects of complex turbulent flows that occur in gas turbine engines for
several years. His research group currently includes University of Minnesota Graduate Students Richard Kaszeta, Rohit Oke, Hitoshi Sakamoto, Cynthia Satterness, and Kebiao Yuan. Collaborative researchers
from Iowa State University, including Professor Richard Pletcher and his Graduate Students Ning Meng, Ravikanth Avancha, and Joon Lee, are working on computer simulations of these flows. The research
effort is applying large-eddy simulation (LES) to turbulent flows in configurations characteristic of those occurring in critical turbine components of aeropropulsion systems. Such a method requires
large computer resources, but it is capable of achieving very realistic results because very little ad hoc modeling is employed to represent the effects of turbulence. The unsteady, three-dimensional
motion of the large eddies is resolved from first principles, and modeling is used only to account for the effects of eddies smaller than the computational grid itself. Such small eddies tend to be
nearly isotropic and universal. Modeling generally plays a fairly minor role in LES, and quite realistic results are expected. In the numerical simulations, it is possible to impose temperature
differences, heat flux levels, and even rotation effects difficult to establish in experimental programs. The simulations should reveal aspects of fundamental physics that are difficult or impossible
to measure. They should also provide information useful in developing or refining turbulence models for design-oriented computer codes. Comparisons will be made with the extensive experimental results
obtained by Professor Simon's research group.
The figures shown here are examples taken from building-block simulations already completed. These include turbulent flow through a rib-roughened channel and simple low-speed channel flow with high
heating and cooling that leads to significant variations in properties. The rib-roughened geometry is often used to enhance heat transfer rates in cooling passages of turbine blades. The simulations have
revealed significant density fluctuations that accompany high rates of heating or cooling even when the Mach number of the flow is very small. In general, heating was found to suppress turbulent velocity
fluctuations while cooling promoted velocity fluctuations.
Future simulations will be directed toward the more complex flow conditions observed in turbomachines. Planned studies include the evaluation of the effects of rotation on flows with heat transfer,
the simulation of film cooling flows, and work on transition and separation
under conditions representative of a low-pressure turbine. The simulations are intended to complement Professor Simon's
experimental studies.
 |
 |
| Contours of instantaneous streamwise velocity in channel flow with constant heat flux heating resulting in Tw/Tb = 1.49; Reynolds number, 10,810. Red indicates high values, blue, low.
|
Contours of instantaneous temperature in channel flow with constant heat flux heating resulting in Tw/Tb = 1.49; Reynolds number, 10,810. Red indicates high values, blue, low.
|
|
|
