14
1 The Experimental Approach in Aerodynamic Design
1.6 Industrial Aerodynamics Testing: Combining Tests
and Numerical Simulation
The turn of the century has seen tremendous progress in numerical simulation methods as a result of the even more staggering improvement in computing resources.
RANS methods are used extensively in industry regardless of the complexity of the
geometries. The scope of these methods is well defined and gives the designer the
opportunity to carry out a considerable amount of the design work by numerical simulation mainly in the automobile, aircraft, propulsion and turbomachinery industry.
However, this approach has its limits: the zones of strong interaction, of separation
on a curved surface or of massively detached flow are notoriously difficult to predict with precision. In addition, since the automotive and aerospace industries are
highly competitive and will innovate through concepts for hypersonic flight, active
flow control or radically new geometrical configurations, this may involve physical
phenomena for which the tools have been inadequately or not validated at all. It
will then be necessary to revert to wind tunnel tests at the design stage in order to
validate the numerical tools to handle the new phenomena or to combine numerical
and experimental approaches (see Chap. 13).
If we take the example of the flight of an air-breathing hypersonic vehicle, the
propulsion system is so integrated that it is difficult to determine its air-propulsive
balance by separating the propulsion on one side and the drag on the other. Here, the
designer is faced with a situation where not only it is difficult to carry out ground
tests, that are fully representative of the real flight, but also the validation of the
calculation means is flawed as the airframe has to be decoupled with the propulsion
system. It is therefore essential to develop an approach that closely links numerical
simulation and testing at all stages of design. The engineer will have to implement
this approach on a more fundamental level as well to account for issues related to
boundary layer transition at very high altitude flight and high Mach numbers (quiet
wind tunnel tests to validate the transition mechanism are presented in Sect. 5.4),
rather than just the global aspects such as aero-propulsive assessment. This can be
evaluated either by means of calculations coupling the external aerodynamics, the
internal aerodynamics, the combustion chamber and the nozzle, or by tests in free
jet installations, often partially representative of the flight conditions.
For a more conventional vehicle, the wind tunnel remains a very effective and
reliable means of characterisation. It is indeed possible to acquire within a reasonable time a very large amount of data that would be difficult to obtain by numerical
calculation. However, once again, the experimental approach suffers from approximations due to support effects, wall effects, Reynolds number corrections and model
deformations that are not representative of the flight, to mention a few. We would like
to emphasise on the fact that neither experimental approach nor numerical simulation
are means to an end, they are in fact complimentary to each other and if employed
rationally it will get us as close as possible to reality (see Chap. 13).
1 The Experimental Approach in Aerodynamic Design
1.6 Industrial Aerodynamics Testing: Combining Tests
and Numerical Simulation
The turn of the century has seen tremendous progress in numerical simulation methods as a result of the even more staggering improvement in computing resources.
RANS methods are used extensively in industry regardless of the complexity of the
geometries. The scope of these methods is well defined and gives the designer the
opportunity to carry out a considerable amount of the design work by numerical simulation mainly in the automobile, aircraft, propulsion and turbomachinery industry.
However, this approach has its limits: the zones of strong interaction, of separation
on a curved surface or of massively detached flow are notoriously difficult to predict with precision. In addition, since the automotive and aerospace industries are
highly competitive and will innovate through concepts for hypersonic flight, active
flow control or radically new geometrical configurations, this may involve physical
phenomena for which the tools have been inadequately or not validated at all. It
will then be necessary to revert to wind tunnel tests at the design stage in order to
validate the numerical tools to handle the new phenomena or to combine numerical
and experimental approaches (see Chap. 13).
If we take the example of the flight of an air-breathing hypersonic vehicle, the
propulsion system is so integrated that it is difficult to determine its air-propulsive
balance by separating the propulsion on one side and the drag on the other. Here, the
designer is faced with a situation where not only it is difficult to carry out ground
tests, that are fully representative of the real flight, but also the validation of the
calculation means is flawed as the airframe has to be decoupled with the propulsion
system. It is therefore essential to develop an approach that closely links numerical
simulation and testing at all stages of design. The engineer will have to implement
this approach on a more fundamental level as well to account for issues related to
boundary layer transition at very high altitude flight and high Mach numbers (quiet
wind tunnel tests to validate the transition mechanism are presented in Sect. 5.4),
rather than just the global aspects such as aero-propulsive assessment. This can be
evaluated either by means of calculations coupling the external aerodynamics, the
internal aerodynamics, the combustion chamber and the nozzle, or by tests in free
jet installations, often partially representative of the flight conditions.
For a more conventional vehicle, the wind tunnel remains a very effective and
reliable means of characterisation. It is indeed possible to acquire within a reasonable time a very large amount of data that would be difficult to obtain by numerical
calculation. However, once again, the experimental approach suffers from approximations due to support effects, wall effects, Reynolds number corrections and model
deformations that are not representative of the flight, to mention a few. We would like
to emphasise on the fact that neither experimental approach nor numerical simulation
are means to an end, they are in fact complimentary to each other and if employed
rationally it will get us as close as possible to reality (see Chap. 13).