The Department of Defense is supporting the
development of robust adaptive refinement methods
for unsteady geometrically complex moving body
problems by means of the High Performance Computing
Modernization Program (HPCMP) Initiative
known as CHSSI. The object of the work is to exploit
the computational advantages inherent in structured
data to solve this important class of problems on
parallel scalable computer platforms.
The physical domain of complex problems is
decomposed into near-body and off-body regions.
The near-body domain is discretized with "Chimera"
overset grids that need extend only a short distance
into the field. The off-body domain is discretized with
overset structured Cartesian grids (uniform) of varying
levels of refinement. The near-body grids resolve
viscous boundary layers and other flow features
expected to develop near body surfaces. Off-body
grids automatically adapt to the proximity of near-body
components and evolving flow features. The
adaptation scheme automatically maintains solution
accuracy at the resolution capacity of the near-body
system of grids. The approach is computationally
efficient and has high potential for scalability. Grid
components are automatically organized into groups
of equal size, which facilitates parallel scale-up on
the number of groups requested. The method has
been implemented in the computer program known
as OVERFLOW-D.
For example, OVERFLOW-D was used in FY99
to obtain a time-accurate simulation of the V-22 tilt-rotor
aircraft in high-speed cruise conditions. Tempo-ral
resolution of the simulation provided 2,000
time-steps per revolution of the rotor blades. Nearly
30 million grid points are used to spatially resolve the
problem domain. An important result of the simula-tion
is the capture of the rotor-tip vortices as part of
the solution. As indicated in the figure, the vortices
are evident in the field a full body length downstream
of rotors. The simulation was carried out using 65
processors on an IBM-SP. Post-process analysis of the
large unsteady data set was carried out on an SGI Origin 2000.
The V-22 result is significant in several respects.
Accurate simulation of the rotor wake system is
important in predicting rotor-airframe interactional
dynamics, as well as aircraft acoustics. Combined
with existing flight data and scheduled tunnel data,
the result provides the basis for demonstrating design-to-
flight analysis capability for general tilt-rotor
aircraft. The result is a high-fidelity baseline data-set
that can be used to evaluate future tilt-rotor concepts
such as variable-diameter tilt rotors, quadrotors, and
future defense transport tilt rotor configurations. The
capabilities illustrated by this simulation support
important defense and civil priorities relating to
rotary-wing aircraft.
Point of Contact: R. Meakin
(650) 604-3969
bmeakin@mail.arc.nasa.gov
Back To Top
Previous Paper
Return to Revolutionary Technology
Next Paper 
