1.4 Some Constraints for Wind Tunnel Test
13
and turbulence distributions). The Reynolds number influences the upstream development of the boundary layer according to a history effect, not by a local effect.
There are many observations in favour of this quasi-independence with regard to the
Reynolds number: formation of vortices on a delta wing, vortex breakdown, base
flows, shock-induced separation, cavity flow to list a few. The important thing is to
achieve an established turbulent regime for the upstream flow and to provide a precise definition of the boundary layer. Further examination of the physics shows that
the thickness of the boundary layer is the appropriate length scale if the extension of
the separated domain is small. On the other hand, if the separation is extended, the
characteristic scale becomes the distance from the point of separation, the influence
of the initial boundary layer being quickly forgotten.
1.5 Deformation of Models
Another important problem for representativeness of tests or calculations is geometric
conformity. Small scale models, most often representative of the overall shape, cannot
reproduce the levels of detail such as surface defects (screws, rivets, structural joints,
etc.) which can, themselves, vary between two versions of the same production
vehicle. The general forms are themselves difficult to define, for instance the wing
tips of an Airbus A380 deflect by several metres in flight, significantly altering the
flow around the wings and this alters trimming characteristics of the aircraft. The
wind tunnel models are also not infinitely rigid and are normally subjected to high
aerodynamic loading due to the large dynamic pressure required to get closer to
the flight Reynolds numbers. Therefore they are bound to deflect and deform. Thus
the extrapolation of wind tunnel measurements to the real operational condition
is a complex problem even beyond the complexity related in achieving dynamic
similarity. During wind tunnel testing, model deformation measurement techniques
have been developed and are used to transpose the results to reference geometry
(see Sect. 9.6). In flight, the precise identification of the shape of the aerodynamic
surface in different flight conditions remains a difficult problem which is generally
only approached by aeroelastic modelling. For CFD calculations, the geometry is
known and controlled but rarely corresponds to the actual geometry in flight, which
itself varies according to the flight conditions.
It is theoretically possible to design flexible models that deform in the same
way as the aerodynamic surface of the vehicle in operation. In practice this is not
feasible on complex models, but can be done on simplified models to study critical
phenomena of aeroelastic coupling such as flutter and force response. These models
are expensive and require very specific testing procedures. These difficulties are
more rarely encountered in the testing of ground vehicles where most often full scale
prototypes are tested, except while testing very large vehicles or structures where
once again model scaling is required, for instance trains and other very large size
vehicles, including ships or buildings.
13
and turbulence distributions). The Reynolds number influences the upstream development of the boundary layer according to a history effect, not by a local effect.
There are many observations in favour of this quasi-independence with regard to the
Reynolds number: formation of vortices on a delta wing, vortex breakdown, base
flows, shock-induced separation, cavity flow to list a few. The important thing is to
achieve an established turbulent regime for the upstream flow and to provide a precise definition of the boundary layer. Further examination of the physics shows that
the thickness of the boundary layer is the appropriate length scale if the extension of
the separated domain is small. On the other hand, if the separation is extended, the
characteristic scale becomes the distance from the point of separation, the influence
of the initial boundary layer being quickly forgotten.
1.5 Deformation of Models
Another important problem for representativeness of tests or calculations is geometric
conformity. Small scale models, most often representative of the overall shape, cannot
reproduce the levels of detail such as surface defects (screws, rivets, structural joints,
etc.) which can, themselves, vary between two versions of the same production
vehicle. The general forms are themselves difficult to define, for instance the wing
tips of an Airbus A380 deflect by several metres in flight, significantly altering the
flow around the wings and this alters trimming characteristics of the aircraft. The
wind tunnel models are also not infinitely rigid and are normally subjected to high
aerodynamic loading due to the large dynamic pressure required to get closer to
the flight Reynolds numbers. Therefore they are bound to deflect and deform. Thus
the extrapolation of wind tunnel measurements to the real operational condition
is a complex problem even beyond the complexity related in achieving dynamic
similarity. During wind tunnel testing, model deformation measurement techniques
have been developed and are used to transpose the results to reference geometry
(see Sect. 9.6). In flight, the precise identification of the shape of the aerodynamic
surface in different flight conditions remains a difficult problem which is generally
only approached by aeroelastic modelling. For CFD calculations, the geometry is
known and controlled but rarely corresponds to the actual geometry in flight, which
itself varies according to the flight conditions.
It is theoretically possible to design flexible models that deform in the same
way as the aerodynamic surface of the vehicle in operation. In practice this is not
feasible on complex models, but can be done on simplified models to study critical
phenomena of aeroelastic coupling such as flutter and force response. These models
are expensive and require very specific testing procedures. These difficulties are
more rarely encountered in the testing of ground vehicles where most often full scale
prototypes are tested, except while testing very large vehicles or structures where
once again model scaling is required, for instance trains and other very large size
vehicles, including ships or buildings.