Theme 1: High Fidelity Simulation of Helicopter Interactions
Helicopter aerodynamics represents one of the most scientifically challenging and industrially rewarding problems for CFD,
with flow speeds ranging from subsonic to supersonic, vortices and wakes which must be resolved and maintained over long times,
and dynamic flow separation effects. Work funded under the rotorcraft aeromechanics DARP and by Westland Helicopters Ltd (WHL)
is pushing forward the development of a validated CFD code for use within WHL for the resolution of all relevant flow regimes.
The consortium will provide this effort with the stimulus of targeting the understanding of multi component interactions,
including the main rotor/tail rotor interaction responsible for noise and handling problems and the fuselage influence on rotor
flows. These simulations will have a significant impact on current efforts to feed full CFD simulations into simpler models for
design in the short term and will be exploited to raise the profile of CFD within WHL and other industrial companies for the
longer term. This work programme will be led by the University of Glasgow.
Theme 2: Simulation of a Free Flying Flexible Aircraft
Computational Fluid Dynamics can provide aerodynamic data for the simulation of the aero-elastic and flight mechanics
response of aircraft for conditions beyond the capability of current simplified empirical and linear models. Work within
the PUMA DARP has developed time-marching flutter analysis methods where the Euler-based time-marching simulation of a
Hawk aircraft has been achieved, with an immediate extension to the simulation of flight mechanics. Building on this, world
leading CFD based fast flutter methods, based around Hopf Bifurcation and system identification techniques, have been
developed that allow instabilities to be detected around two orders of magnitude faster than using time marching techniques.
The HPCx consortium resources will facilitate the following work. Firstly, the simulation of a free-flying deforming Hawk at
high incidence will be possible. The simulation will serve as a test bed for an initial study of aerodynamically induced
nonlinear aeroelasticity and flight mechanics, a field that is virtually untouched using CFD. Secondly, calculations on a
NASA F16-XL aircraft will be made on refined grids. This work is being done by Glasgow within a unique NASA run study to
compare CFD predictions with flight-test data. Both strands of work will be significantly enhanced by access to HPCx where
the extra computing power will allow grids of sufficient refinement to be used. This programme of work will be led by the
University of Glasgow.
Theme 3: Simulation of Vertical Landing Aircraft
The vertical landing phase of aircraft such as the Joint Strike Fighter (JSF) creates a highly unsteady and complex
aerodynamic problem. In the transition from wing-borne to jet-borne flight there are considerable geometry changes as intake
doors open, flaps deploy and nozzles rotate. In addition, ground vortex and fountain flows create the potential for hot gas
ingestion. Work towards a simulation of this problem has been supported within the PUMA DARP. The consortium resources should
allow the demonstration of the capability for a geometrically complete descending aircraft, including geometry changes.
Experience gained from the LESUK and
UKTC consortia for LES on structured and unstructured meshes will be
utilised. Numerical challenges include the handling of complex moving geometries, mesh adaption for unsteady flows and widely
varying flow speeds
requiring special preconditioning techniques. The wide variation of time scales present in the domain creates the requirement
for extremely long run times. The partners will focus on developing unsteady RANS and DES techniques for moving geometries in
the transition phase, whilst Loughborough will concentrate on the ground effect regime. A final joint calculation will
demonstrate the simulation of an aircraft from transition down to ‘wheels on’ with seamless switching between unsteady RANS,
Detached Eddy Simulation and Large Eddy Simulation. Such a calculation would be at the forefront of current international
research and is only feasible with the capability offered by HPCx.
Theme 4: Aeroelasticity Studies for Aero-Engine Core-Compressors
Unsteady turbulent high-speed compressible flows often give rise to complex aeroelastic phenomena by influencing the dynamic
behaviour of structures on which they act. The problem is particularly severe for aero-engines where virtually all bladerows
are susceptible to large-amplitude vibration under unsteady flow effects. Due to their wide operating envelope, the most
complex and the least understood aeroelasticity phenomena occur in multi-stage core compressors which suffer from a mixture
of aeroelastic instabilities such as acoustic resonances cavity resonances, flutter, high and low engine-order forced response,
buffeting, vortex shedding and rotating stall. Such phenomena are believed to be caused by at least one bladerow undergoing
severe stall but the overall compressor still managing to function because of the overall pressure ratio. There are no
established design methods that can predict the amplitude of vibration and most engine development work is carried out by
testing actual hardware with a view to redesigning those bladerows that have high vibration amplitude.
Over the last ten years, the Aeroelasticity Group at Imperial College, which forms part of the Rolls-Royce sponsored
Vibration University Technology Centre (UTC), has established itself as a world-leading research unit for large-scale
modelling of turbomachinery aeroelasticity. The consortium will allow a detailed study of a specific core compressor that
has a well-documented case history for a variety of aeroelastic instabilities, including detailed measured data. This programme
will be led by Imperial College, London.
Theme 5: Simulation of Internal Air System
A wide variety of flow phenomena occur in internal air systems. Rotating disc cavity flows, which are often dominated by
rotational effects, are particularly important. While CFD is now established for a range of rotating disc problems, there
are still two particular areas for which the flow physics is not yet well understood and for which application of CFD is limited.
These are main gas path and air system interaction and buoyancy affected flow in high-pressure compressor disc cavities. Both
classes of flow are three-dimensional and unsteady. For the cavities the unsteadiness is due to instability and occurs even when
the boundary conditions are fully steady and axisymmetric. For annulus flow interaction, it is well known that unsteadiness
arises due to passing of the rotating blades, and it has recently been shown that further unsteadiness can arise at frequencies
unrelated to the blade passing and that these can have an important influence on the mean flow. This unexpected result has
stimulated considerable interest in industry, but further studies are required to understand this flow feature and its implications.
The HPCx resources will allow the simulation of a compressor disc cavity with axial flow. It is expected that use of HPCx
will allow solutions with 1 to 2 orders of magnitude more mesh points than is currently feasible, allowing the rotational
Rayleigh number to be extended into the engine representative range. Successful predictions would give new insight into this
complex flow, and improved predictive methods for this type of flow are eventually expected to lead to less engine tests being
needed in development programmes. The proposed calculations will be at the forefront of international research and the planned
work will be led by the University of Surrey.
Theme 6: Simulation of Store Separation
The prediction of the trajectory of a store after its release from an aircraft is of great interest to the designer. The store
can often flip and hit the wing or the tail of the aircraft endangering the life of the pilot. The trajectory of the store
is dependent on its mass and on the aerodynamic forces the store is subjected to after its release. The accurate prediction of
the forces requires the solution of the unsteady Navier-Stokes equations in the presence of moving geometry. Since large boundary
movement has to be handled, mesh adaptivity is required. In this case, special care has to be taken to ensure geometric
conservation. In addition, higher order interpolation of the flow variable in the adapted regions may need to be utilised to
guarantee second order accuracy of the overall scheme. The simulation of a store release problem on a fully loaded fighter is only
possible with the capability offered by HPCx. For improved efficiency, any procedure needs to be enhanced to maintain the original
load balancing after adaptivity has taken place. The mesh adaption and dynamic load balancing would put such a demonstration at
the forefront of what is being done globally and will be led by the University of Wales, Swansea.
Theme 7: Support from Daresbury Laboratory
Daresbury Laboratory has a long record of providing support for Engineering Supercomputing Consortia. The computations will
all generate large data sets and manipulating and extracting meaning from these will entail significant difficulty. Daresbury
will investigate the parallel implementation of feature extraction algorithms for vortices and shock waves which are in the
literature and which have started to appear in commercial visualisation codes. The most suitable algorithms will be implemented
in some of the consortium codes to allow on-the-fly data reduction and for the computational and visualisation performance to be
evaluated.
|