|
|
|||
|
Executive summaries of Action Groups of Group of Responsables in Aerodynamics ·
AD(AG34) ‘Aerodynamics
of supersonic air intakes’ Action
Group 34 started in 1999 with participants from ONERA, Dstl (Chairman), EADS (D),
BAe Systems, MBDA (F), NLR and QinetiQ. A
wide range of flow problems can be encountered in the design of supersonic
air intakes and their integration into flight vehicles. CFD methods hold
immense potential for reducing risk in the intake design process as well as
helping to find and evaluate to new approaches to their design. This
collaboration improved the understanding of two basic intake compression
surface types and put in place a calibrated bleed modelling capability within
the industrial CFD codes of each participating organisation. Practitioners
were brought together from both fundamental and applied areas of aerodynamics
research. The
Group was asked to consider a form of collaboration that centred on the use
of prediction methods alone, namely CFD methods, to characterise new or
particularly challenging intake engineering problems. The Group chose to
pursue two such cases. The first was an intake compression surface, or
‘bump’, flow that would be approached using CFD alone. The second
was a 2D wedge intake featuring an external compression surface with a
distributed bleed system. The latter case would be approached using elements
of the familiar prediction method calibration and comparison exercise
followed by application of selected methods to a generic intake design.
Navier-Stokes based methods were employed by all participants throughout the
collaboration. Little
detail was known about the boundary layer development on a conical bump-type
compression surface (Figure 1),
particularly at off-design conditions. The bump geometry chosen was an exact
stream surface of the inviscid, Mach 1.8 flow past a 23° cone at zero incidence and
sideslip.
Figure 1 Surface
pressure distribution with near-surface streamlines, Mach 1.6, zero sideslip A
matrix of calculations was completed by the Group. Each participant computed
the effect of a different parameter, yielding a significant multiplication of
the effort. Effects included Mach number (0.6 to 3.5), sideslip, turbulence
model and Reynolds number. Code-to-code comparison cases were also computed
at key points in the matrix. This aspect of the collaboration worked very
well. Shared
experience of the extraction of boundary layer displacement thickness at an
arbitrary location in a The
consistency of results obtained for the two code-comparison cases was
encouraging. All the codes gave acceptable and consistent flow physics though
with detailed differences. An appropriate mesh density was identified and
this will provide a starting point for similar applications in the future.
The boundary layer diversion characteristics of the bump (Figure 2) were characterised over the full matrix. Some of these
were unexpected. Figure 2 Effect
of Mach number on bump surface boundary layer displacement thickness The
work is considered to have significantly improved the understanding of this
class of flow. Problems that are likely to be encountered when designing an
intake in conjunction with such a bump diverter feature were inferred from
the study. Experimental verification of the findings is required and a highly
focussed experiment could be defined on the basis of the current work.
Further work is also required to examine the integration of such a bump with
intake entries of various types. It was felt by the Group that integration
with a forward-swept intake entry would yield the most favourable aerodynamic
interaction effect. The
overarching aim of second case was to improve the understanding and ability
to model flows that limit the performance of a wedge type external
compression supersonic intake (Figure
3). This case is directly relevant to both combat aircraft and air
breathing weapons and the underlying modelling capability has more general
applicability to the control of shock-boundary layer interaction. The main
aim was to improve the understanding of flow development as an intake normal
shock wave passed over a porous bleed surface as the mass flow into the
intake was reduced. In more three-dimensional wedge intake designs such as
prismatic intakes, very complex shock patterns may be encountered at such
off-design conditions. Calibrated CFD methods would be very valuable to the
designer in this situation.
Figure 3 Examples
of wedge intake configurations A
precursor to this exercise, and one of the main achievements of the Group,
was the implementation and extensive calibration of porous wall bleed models
within the CFD capabilities of the participating organisations. Shock /
boundary layer interaction data from an ONERA channel flow experiment (Figure 4) were used for this purpose.
Figure 4 Test
Case: Passive control of shock / boundary layer interaction Effects on modelling of code, grid density, turbulence
model and bleed model were then studied on an idealised Study Case. This
consisted (Figure 6, inset) of a 7° wedge-type compression surface with distributed bleed and
a lip positioned to give a
nominal design shock-on-lip Mach number of 1.8. Computations of the effect of
bleed quantity across a range of intake operating mass flow ratios were
executed (Figure 5), using what was
judged to be the best modelling approach found from the earlier calibration
exercise.
Figure 5 Effect
of mesh characteristics on predicted Mach number distributions Solutions were obtained in all of the desired flow regimes
including supercritical operation. In general, a high degree of sensitivity
to turbulence model was encountered (Figure
6) at bleed mass flows of 1% or less of the reference capture area. The
flow in this regime was typically characterised by the presence of a large
flow separation on the ramp.
Figure 6 Comparison
of predicted intake performance Relative to the original objectives, the collaboration has, first of all, shown that solutions can be readily obtained in the flow regimes of interest. It has also shown that consistency can be obtained for a range of approaches, especially where the flow is well-ordered. Prospects for the effective use of CFD in the design of intakes with porous bleed systems appear to be good, though comparison against a good quality experimental test case will be required to consolidate these findings. Such an exercise would also be required to determine which of the modelling approaches would best predict the onset of large-scale flow separation and which, if any, would be applicable to regimes where there was extensive flow separation. In
final summary, new predictive capabilities have been implemented and all
participants have benefitted from the shared experience of applying CFD
methods to challenging and highly relevant intake problems. Successful use of
CFD by a team drawn from different organisations to explore a new design and,
also, to expand knowledge of the design space around familiar design has been
demonstrated. Both the benefits and the limitations of this way of working
have been observed in the course of this Action Group. Overall, the outcome
of the collaboration has been very positive.
|
