|
|
|
Research in Computational Fluid Dynamics
|
|
Affiliated Faculty:
David Caughey,
Lance Collins,
David Erickson,
Steve Pope,
Jane Wang
|
|
Computational fluid dynamics (CFD) is the branch of fluid mechanics devoted
to the development and application of computer-based tools to solve the
partial differential equations describing fluid flow. The field of CFD
includes those aspects of numerical analysis and computer science relevant
to the numerical solution of partial differential equations and mesh
generation, the development of physically-based models for those phenomena
that cannot be computed directly, and the application of these tools to
important problems in fluid mechanics and fluids engineering.
|
|
False-color image of the chemical species OH in a
turbulent mixing layer between cold hydrogen and hot air. From
large-eddy simulations performed by Lu, Ren & Pope (2004).
(Courtesy Steve Pope)
|
|
|
Lattice-Boltzmann simulations of a sheared flow of gas and spherical particles.
(Courtesy Rolf Verberg and Don Koch) |
|
CFD has evolved over past decades from a specialized research area,
studied primarily by those developing new algorithms for specific fluid
flow problems, to a broadly-used practical tool used by engineers in
virtually every industry. The ability of CFD to solve fluid flow problems
in arbitrarily complex geometries has led to rapid growth in it use in
industry. At the same time, the ability of CFD to provide essentially
exact answers to (at least) a limited class of flow problems, with a wealth
of detailed data on the local behavior of the fluid, makes it an attractive
tool to study the nature of fluid flows, themselves.
|
|
CFD is still an actively evolving field. Work at Cornell is directed
at developing coupled finite-volume/particle-based probability-density
function (PDF) methods for turbulent combustion, large-eddy simulation
(LES) techniques for inert and reacting turbulent flows, direct
numerical simulations (DNS) of multi-phase turbulent flows, with
application to particulate clustering and cloud formation, and the
development of efficient algorithms for steady and unsteady transonic
flow fields. In addition to these efforts directed at developing basic
algorithms, CFD is also used as a tool in many, if not all, of the
fluid mechanics research projects at Cornell and elsewhere.
|
|
Airfoil surface pressure contours for flow past the RAE 2822 airfoil at
3.0 degrees incidence and Mach 0.75. Figure demonstrates convergence of
an accelerated multigrid algorithm showing negligible differences between
the fully converged solution (solid lines) and the solution obtained using
only 5 cycles of the multigrid algorithm.
(Courtesy Dave Caughey) |
|
|
|
For some time, faculty in M&AE have had a special relationship with
Fluent, Inc., one of the leading developers of CFD software systems
for industry. Several of our graduates have played pivotal roles in
the development and implementation of key aspects of the algorithms,
and Cornell faculty have worked with Fluent engineers to incorporate
the results of their research into recent releases of Fluent. Cornell
faculty also see the value of introducing students to modern engineering
simulation tools, such as Fluent and ANSYS (a finite-element code for
structural analysis), and these are used in a variety of undergraduate
and graduate courses in M&AE at Cornell.
|
|
