1.3 Limitations and Constraints of Numerical Methods
11
1.3 Limitations and Constraints of Numerical Methods
Until recently, the validation of numerical methods was mainly done by comparing calculated results with overall aerodynamic forces measurements and properties
measured at the wall, essentially pressure. For many studies this type of comparison
was sufficient. Such are the “traditional” methods, empirical or based on a simplified
modelling approach, predicting only wall properties such as pressure, skin friction,
heat flux and the overall performance of the vehicle. These methods could also give
an idea of the overall flow-field, for example, predicting the size of a separated
region and the location of a point of separation, but this information was considered
more or less qualitative. The scope of flow prediction has changed with the advent
of theoretical models based on solving the Navier-Stokes equations (see above) or
lower-order models formulated from simplification of these equations. It is clear
that this approach is the only one capable of calculating complex flows containing shock waves and expansion fans, detached regions and vortex structures, where
the dissipative regions being turbulent in almost all practical situations. Not only
the wall properties are calculated, but other quantities in the flow-field including the
average velocity and the complete turbulent field. However, in its current state the
Navier-Stokes approach is still far from being free from criticism, many difficulties
remains in the numerical methods and in the physical modelling, especially for the
prediction of the full turbulence field. There is therefore a strong need to validate
codes prior to their routine use for design purposes.
Although prediction of wall properties remains an essential goal for most computational methods, since lift, drag and in some cases wall temperature are the quantities of greatest practical interest, it soon became clear that the comparison with wall
properties was insufficient to validate most of the advanced predictive methods. In
general, the Navier-Stokes codes give a faithful and impressive image of the aerodynamic field structure. However, a closer look at the results shows that the situation is
far from entirely satisfactory. Thus, it can be seen that a fairly good prediction of the
wall pressure can coexist with a poor quantitative description of the velocity field.
Frequently, the extent of a separated region is underestimated, sometimes considerably. In addition, the turbulent quantities are poorly predicted, especially if the flow
is in great part separated. These discrepancies may make the validity of the method
suspect, since they reveal a certain deficiency either in their numerical scheme, or
in the turbulence model, or both. Conversely, a satisfactory prediction of the field
can be accompanied by significant disagreements in the calculation of the surface
properties affecting the transfer quantities, such as skin friction and heat transfer.
In some applications, knowledge of the external far field itself is of primary
interest, as in the case of the infrared signature where the detailed description of
the hot propellant jet, with a precise location of the Mach disks, is essential. Air
pollution studies require good prediction of the flow field to allow for a proper
assessment of chemical reactions and species concentration as well as pollutant
dispersion. The same is true for the prediction of vehicle noise, where the prediction
of the aerodynamic noise level and the analysis of its origin being an important issue
11
1.3 Limitations and Constraints of Numerical Methods
Until recently, the validation of numerical methods was mainly done by comparing calculated results with overall aerodynamic forces measurements and properties
measured at the wall, essentially pressure. For many studies this type of comparison
was sufficient. Such are the “traditional” methods, empirical or based on a simplified
modelling approach, predicting only wall properties such as pressure, skin friction,
heat flux and the overall performance of the vehicle. These methods could also give
an idea of the overall flow-field, for example, predicting the size of a separated
region and the location of a point of separation, but this information was considered
more or less qualitative. The scope of flow prediction has changed with the advent
of theoretical models based on solving the Navier-Stokes equations (see above) or
lower-order models formulated from simplification of these equations. It is clear
that this approach is the only one capable of calculating complex flows containing shock waves and expansion fans, detached regions and vortex structures, where
the dissipative regions being turbulent in almost all practical situations. Not only
the wall properties are calculated, but other quantities in the flow-field including the
average velocity and the complete turbulent field. However, in its current state the
Navier-Stokes approach is still far from being free from criticism, many difficulties
remains in the numerical methods and in the physical modelling, especially for the
prediction of the full turbulence field. There is therefore a strong need to validate
codes prior to their routine use for design purposes.
Although prediction of wall properties remains an essential goal for most computational methods, since lift, drag and in some cases wall temperature are the quantities of greatest practical interest, it soon became clear that the comparison with wall
properties was insufficient to validate most of the advanced predictive methods. In
general, the Navier-Stokes codes give a faithful and impressive image of the aerodynamic field structure. However, a closer look at the results shows that the situation is
far from entirely satisfactory. Thus, it can be seen that a fairly good prediction of the
wall pressure can coexist with a poor quantitative description of the velocity field.
Frequently, the extent of a separated region is underestimated, sometimes considerably. In addition, the turbulent quantities are poorly predicted, especially if the flow
is in great part separated. These discrepancies may make the validity of the method
suspect, since they reveal a certain deficiency either in their numerical scheme, or
in the turbulence model, or both. Conversely, a satisfactory prediction of the field
can be accompanied by significant disagreements in the calculation of the surface
properties affecting the transfer quantities, such as skin friction and heat transfer.
In some applications, knowledge of the external far field itself is of primary
interest, as in the case of the infrared signature where the detailed description of
the hot propellant jet, with a precise location of the Mach disks, is essential. Air
pollution studies require good prediction of the flow field to allow for a proper
assessment of chemical reactions and species concentration as well as pollutant
dispersion. The same is true for the prediction of vehicle noise, where the prediction
of the aerodynamic noise level and the analysis of its origin being an important issue