NASA Technical Reports Server (NTRS) 20010019054: Measured and Computed Hypersonic Aerodynamic/Aeroheating Characteristics for an Elliptically Blunted Flared Cylinder

22 Pages·1.1 MB·English

by  NASA Technical Reports Server (NTRS)

AIAA 2001 - 0562
Measured and Computed Hypersonic
Aerodynamic/Aeroheating
Characteristics for an Elliptically
Blunted Flared Cylinder
Francis A. Greene, Gregory M. Buck, and William A. Wood
NASA Langley Research Center
Hampton, Virginia 23681
39 th Aerospace Sciences Meeting & Exhibit
January 8-11, 2001
Reno, Nevada
For permission to copy or republish, contact the American Institute of Aeronautics and Astronautics
1801 Alexander Bell Drive, Suite 500, Reston, VA 20191-4344

AIAA 2001-0562
Measured and Computed Hypersonic Aerodynamic/Aeroheating
Characteristics for an Elliptically Blunted Flared Cylinder
Francis A. Greene*, Gregory M. Buck*, and William A. Wood*
NASA Langley Research Center, Hampton, VA 23681
Abstract
Computational and experimental h)personic aerodynamic /orces and moments and aeroheating levels/or Kistler Aerospace
Corporation's baseline orbiter vehicle at incidence are presented. Experimental data were measured in ground-based facilities
at the Langley Research Center and predictions were performed using the Langley Aerothermodynamic Upwind Relaxation
Algorithm code. The test parameters" were incidence (-4 to 24 degrees'), freestream Maeh nuntber (6 to 10)..9"eestream ratio of
specific heats (1.2 to 1.4), and freestream Reynolds n,tmber (0.5 to 8.0 million per Jbot). The effects of these parameters on
aerodynamic characteristics, as well as the effects of Reynolds number on measured heating levels are discussed Good
agreement between computational and experimental aerodynamic and aeroheating values were observed over the wide range of
test parameters examined. Reynolds number and ratio of specific heats were observed to significantly alter the trim L/D value.
At Mach 6. laminar flow was observed along the entire windward centerline up to the.[lare for all angles and Reynolds numbers
tested. Flow over the flare transitioned from laminar to transitional(cid:0)turbulent between 4 and 8 million per Jbot at 8 and 12
degrees angle of attack, and near 4 million per foot at 16 degrees angle of attack.
Nomenclature orbit (TSTO) space transportation system. By rolling
total axial force coefficient (includes base the rift vector, each stage is steered to a landing point
CA
that is near the launch site 3. The first-stage, referred to
pressure)
as the launch assist platform (LAP), returns shortly after
pitching-moment coefficient
CITI
launch. The second-stage or orbiter vehicle (OV)
(Gin),, longitudinal stability parameter, per radian
returns to the launch site after the payload has been
CN normal force coefficient
deployed and is the focal point of this paper.
Cp pressure coefficient
L/D lift-to-drag ratio Because the trim angle of the OV is determined
Q non-dimensional heat rate based on the balance of aerodynamic forces acting on
Re Reynolds number, per foot the vehicle unaided by reactive jets or movable control
surfaces, an accurate description of the aerodynamic
(t angle of attack, degree
environment is crucial if the vehicle is to land within a
radial angle, degree (0=leeside, 180=windside)
prescribed area surrounding the launch site. Similarly,
7 ratio of specific heats
for a successful flight, an accurate description of the
9 density, slug/fl 3
aeroheating environment is crucial for the proper
selection, sizing, and split-line locations of the thermal
Subscripts
protection system (TPS) material. Inaccuracy in the
_o freestream values
aeroheating data used in the design of the TPS or in the
2 post-shock values
aerodynamic data used in the guidance and control
algorithms could result in vehicle failure or limit
Introduction
vehicle capabilities in its operational mode. To avoid
To significantly lower Earth to orbit (ETO) costs, inaccuracy, factors that influence flight aerodynamic or
aircraft like operations are being incorporated into the aerothermodynamic values should be identified and
designs of fully reusable space transportation system accounted for.
concepts. Several entrepreneurial start-up companies The influence of reacting gas chemistry within the
are stepping in to fill the low-cost launch void by shock layer has a significant impact on aerodynamic
building reusable launch vehicles (RLV) to service the and aeroheating values and is quantified best using
low and medium Earth orbit launch markets _. Their computational methods. Before a computational fluid
intention is to capitalize on the growth in the dynamics (CFD) tool is employed to numerically
commercial space sector-" in traditional and new predict data used in the design of flight vehicle, the
industries. The RLV under development by Kistler credibility of the CFD tool should be established.
Aerospace Corporation is a fully reusable two-stage-to- While this is a multi-step process, generally, a first step
"Aerospace Technologist, Aerothermodynamics Branch. Aerodynamics. Aerothermodynamics and Acoustics Competency, Senior Member AIAA
rAerospace Technologist. Aerothermodynanucs Branch, Aerodynamics. Aerothermodynamics and Acoustics Competency
Copyright © 2001 by the American Institute of Aeronautics and Astronautics, Inc. No copyright is asserted in the United States under Title 17,
U.S. Code. The U.S. Government has a royalty-free license to exercise all rights under the copyright claimed herein for Governmental Purposes.
All other rights are reserved by the copyright owner.
1
American Institute of Aeronautics and Astronautics Paper 2001-0562
is to ensure the computational tool can reproduce covered the trailing wake (21x37x101). For each
perfect gas wind tunnel data and flow physics when solution, the outer grid domain was adapted to conform
applied to the subject vehicle at incidence. If the CFD to the bow shock. Also, points in the direction normal
tool is not successful at this step, then its credibility for to the surface were redistributed such that the cell
predicting data at flight conditions is in question Reynolds number at the wall was of order one and the
Computational and experimental hypersonic grid stretching in the boundary layer was less than 1.2.
aerodynamic forces and moments and aeroheating Using this criterion, LAURA has accurately predicted
levels are presented for Kistler Aerospace surface heat transfer and aerodynamics at wind tunnel
Corporation's baseline OV configuration at incidence. and flight conditions 7'8,9.10
Continuum laminar CFD predictions for CF4
(tetrafluoromethane) and perfect gas air were obtained Computational Method
using the Langley Aerothermodynamic Upwind LAURA is a finite-volume shock-capturing point-
Relaxation Algorithm (LAURA) 4 code. Aerodynamic
implicit structured grid solver that relaxes either the
coefficients and aeroheating levels from LAURA are Euler, thin-layer Navier-Stokes, or full Navier-Stokes
compared with corresponding measured data obtained equations in pseudo time to asteady state. The inviscid
at Langley Research Center's (LaRC) first-order flux is upwind biased using Roe's flux-
Aerothermodynamic Facilities Complex (AFC) 5'6. difference splitting 11with Harten's entropy fix12and is
Specifically, the experimental data were measured in extended to second-order with Yee's Symmetric Total
the AFC 20-1nch Mach 6 CF4, 20-Inch Mach 6 Air, and Variation Diminishing (STVD) approach 13. Viscous
31-Inch Mach 10 Air Tunnels and covered an angle of terms are incorporated via second order central
attack (AOA) range of-4 to 24 degrees, and a Reynolds difference approximations. LAURA operates
number range of 0.5 to 8 million per foot. Comparing efficiently in a macro-tasked or parallel computing
data from the AFC tunnels allows the influence of
environment and is capable of predicting laminar or
vehicle attitude, Mach number, Reynolds number, and turbulent t4 aerothermodynamics for perfect gas 15,
ratio of specific heats (7) on aerodynamic and tetrafluoromethane (CF4), equilibrium 16, or non-
aeroheating values to be assessed. The influence of equilibrium 17flows. The LAURA predictions presented
these parameters on aerodynamic values is discussed. correspond to the wind tunnel conditions listed in Table
Furthermore, flow visualizations from experiment, 1 and were made assuming laminar viscous (thin-layer
compared with corresponding images from CFD Navier-Stokes) flow for CF4, and perfect gas air. The
predictions, are presented. Finally, the influence of wall temperature was constant for all cases; it was set to
Reynolds number on aeroheating levels is presented. 540 Rankine.
Aerodynamic and aeroheating reference values and the
center of gravity location were withheld at Kistler's Experimental Model
request.
The aerodynamic and aeroheating data were
measured on one-percent scale models. Aeroheating
Computational Model and Grid data were obtained using the two-color, relative
A numerical representation of the baseline OV and intensity, phosphor thermography technique is. The
the symmetry plane grid is shown in Figure 1. The aeroheating data were then reduced and analyzed using
numerical predictions were performed on a the IHEAT TMcode. The global data produced by this
computational geometry that did include the engine technique permit the resolution of complex flow
nozzle and wake, but did not include raceways like phenomena such as transition fronts, vortex structures,
those shown in Figure 2. The OV is reminiscent of the and shock interactions. To obtain accurate heat transfer
Polaris re-entry stage: a flare stabilized elliptically measurements with the phosphor thermography
blunted cylinder. The inner-mold line of the full-scale technique, models are cast from a silica ceramic
baseline OV measures 611 inches in length from nose material that has a low diffusivity, is well defined, and
to flare end; the cylinder radius is 82.3 inches; the has uniform thermal properties. These models are
maximum flare radius is 132.0 inches, and the flare coated with a mixture of phosphors suspended in a
angle is 21 degrees. The nose radius is 165.3 inches silica-based colloidal binder. The phosphor models
and a 25.5-inch radius curve is used to transition from tested, one of which is shown in Figure 2, included the
the nose to the cylinder. A 27.9-inch radius curve fills raceways, but did not include a representation of the
the cylinder-flare junction. engine nozzle.
The predictions were computed on a three-block Pictured in Figure 3 are the baseline model,
"medium" grid system. The first block covered the raceways, a cylinder extension, and a flare extension
main body (89 streamwise x 37 circumferential x 65 used for force and moment tests. The extensions were
points between the body and outer domain); the second used to construct variants of the baseline configuration;
covered the nozzle wake (29x37x85), and the third
-2-
American Institute of Aeronautics and Astronautics Paper 2001-0562
thesevariantwseretestedb,utarenotthesubjecotfthis regarding tunnel operating ranges and specifications
document.Ontheflightvehicle,theracewayasre can be found in Refs. 5and 6.
located
+30 degrees off the leeward symmetry plane.
The force and moment data presented were measured
on the baseline OV model without raceways or an
engine nozzle. In the air and CF4 facilities, a stainless
steel nose was mated to an aluminum body. This setup
reduced conduction heating effects on the strain-gage
balance.
-f ]iling Wake
,Wake
Figure 2. One-percent scale thermographic phosphor
model for the K-1 OV.
Main Body
Figure 1. Computational model - baseline OV &
symmetry plane grid.
Figure 3. One-percent scale force and moment model
Experimental Facilities for the K-10V.
The experimental data presented in this document
were measured in three hypersonic facilities that are Data Uncertainties
part of the LaRC AFC. Nominal flow conditions for For aerodynamic data on the baseline OV, repeat
each facility are listed in Table 1. All three facilities, runs were made, references for similar configurations
the 20-Inch Mach 6 CF4, 20-Inch Mach 6 Air, and 31- were consulted, and the measured (i.e. experimental)
Inch Mach 10Air Tunnels are similar in their operation. data were compared with predictions from CFD. A
Each is a blow-down-to-vacuum wind tunnel that uses a statistical uncertainty analysis has not been performed
filtered test gas. The gas is transferred from a high- to date; however, the error in aerodynamic values
pressure storage facility to a heater. To prevent expressed as a percentage (0.5 percent) of the full-scale
liquefaction, the air and CF4 gas are heated. The gas is balance load is presented in Table 2. The experimental
then regulated to the desired pressure and passed aerodynamic values were corrected for sting
through filters to remove particles (> 5-microns) before deflections. During aerodynamic testing, the state of
entering the settling chamber. From the settling the boundary layer on the test model in the facilities is
chamber, the filtered gas expands through a nozzle and not explicitly known. However, the LAURA
enters the test section at the design Mach number. predictions at wind tunnel conditions presented in this
Once through the test section, the gas is slowed to a report were made assuming laminar flow. The
subsonic velocity and collected in vacuum spheres. The appropriateness of the laminar assumption for air is
CF4 tunnel is a closed-circuit tunnel. The gas is supported by subsequently measured heating trends
recovered from the vacuum spheres, impurities with Reynolds number. Due to the low operating
removed, and the gas stored for subsequent use. Details Reynolds numbers in the CF4 facility, the flow is
-3-
American Institute ofAeronautics and Astronautics Paper 2001-0562
believedtobelaminar.Toaddconfidencien the
redundancy, comparisons at other Reynolds numbers
computationsaollutionsa,checkforgridconvergence
for air data at Mach 6 and 10 are not included. Also,
wasconductedIn.comparinagerodynamcoicefficients
contributions to the aerodynamics from the engine
fromthemeshusedwithvaluesfroma finermesh
nozzle were not included in the computed data, since
(surfacdeimensioninscreasebdyafactoorftwo),less
the nozzle was not part of the experimental
thana onepercentchangewasobservedin all
configuration.
aerodynamfoicrcesandmoments.
The numerical predictions compare well with the
Asdiscusseidn Ref.18,theaccuracoyf the
experimental data. For Mach 6air, differences between
phosphotrechniquies dependeonnt thetemperature
measured and computed values are generally within six
riseonthesurfaceofthetestmodel.Forheating
percent for CA, L/D, and Cm, and within four percent for
measuremeonntwsindwarsdurfaceth,eaccuracoyfthe
Cry. Experimental Mach 10 values, shown in Figure 6,
phosphosrystemisestimatetodbeapproximately
+ 8 for CAand CNagree with the corresponding CFD values
percent, and the overall experimental uncertainty of the within four percent, while differences of 7-10 percent
heating data due to all factors is estimated to be +15 exist for Cm and L/D. Shown in Figure 7 is data at the
percent. Because the leeside temperature increase is Mach 6 CF4 condition. The percentage agreement is
only a few degrees during a test, the experimental similar to that observed for the Mach 10 data. In
uncertainty for leeside aeroheating rates increases to general, predicted and measured coefficient values of
+25 percent. Computed aeroheating levels at three CN and CA are within four percent, and Cm and L/D are
longitudinal distributions for the medium and fine grids within 10 percent. Good agreement between computed
are shown in Figure 4. Medium and fine grid and measured aerodynamic coefficients is observed.
aeroheating values compare within three percent, with This good agreement, which exists at incidence with
the exception of the leeward flare region. The overall significant three-dimensional effects and over a wide
error for a grid converged LAURA solution is range in flow conditions, demonstrates the ability of the
approximately +10 percent. CFD tool to reproduce data at varied wind tunnel
conditions and augments the credibility of LAURA to
predict aerodynamic data at flight conditions.
Results and Discussion
Aerodynamic and aeroheating data were measured
Angle of attack effects
at Langley Research Center's 20-Inch Mach 6 CF4, 20-
With increased incidence, factors such as flow
Inch Mach 6 Air, and 31-Inch Mach 10 Air Tunnels on
asymmetry between the windward and leeward
one-percent scale models of Kistler Aerospace
Corporation's baseline orbiter vehicle. The data was surfaces, changes in separation and reattachment points,
and the movement of the bow shock relative to the
obtained over the wide range of incidence, Mach
number, Reynolds number, and ratio of specific heats vehicle surface, affect aerodynamics by altering surface
listed in Table 1. In this section, these measured data pressure distributions. In addition, as the flow is
processed through the flare shock, pressure levels on
are compared corresponding computational data from
the flare are elevated in a non-linear manner.
the LAURA code and the effects of attitude, Reynolds
The behavior of aerodynamic coefficients with
number, ratio of specific heats, and Mach number on
angle of attack is displayed in Figures 5-7. These
aerodynamic characteristics are described. Following
figures show the qualitative behavior of the coefficients
this, computed and measured aeroheating levels, as well
with angle of attack is independent of Mach number,
as aeroheating trends with Reynolds number are
Reynolds number, and ratio of specific heats. At all
presented.
flow conditions, a change in the slope of the
aerodynamic values occurs near 10 degrees. Near this
Predicted vs. measured aerodynamics
angle, pressure levels on the flare have increased to a
Displayed in Figures 5-7 are plots of experimental
point where their influence on the aerodynamic
and computational aerodynamic data. Measured values coefficients is noticeable. This influence continues to
are represented by open symbols and predicted values
increase non-linearly with angle of attack and is
by a solid line through a filled symbol. The axial force
primarily responsible for the steeper slopes in the
coefficient, normal force coefficient, pitching moment
aerodynamic data between 10 and 20 degrees. In the
coefficient, and lift-to-drag ratio for air at Mach 6 (Re =
Mach 10 air and the Mach 6 CF4 data, a second change
4.0 million per foot) are presented in Figure 5. The
in slope is present near 20 degrees. This second change
same quantities for air at Mach I0 (Re = 2.0 million per
is caused by the bow-shock flare-shock interaction and
foot), and for CF4 at Mach 6(Re = 0.4 million per foot) is not observed in the Mach 6 air data because the
are presented in Figures 6 and 7, respectively. Because
responsible interaction occurs at an angle that isbeyond
agreement comparable to that in the above figures was
the displayed angle of attack range. An expansion fan
observed across the Reynolds number range, to avoid
generated by the shock-shock interaction reflects off the
-4-
American Institute of Aeronautics and Astronautics Paper 2001-0562
flaresurfacea,ndwithincreasinagngleof attack, million per foot, the experimental values asymptotically
resultisnadecreasinewindsidepressuorentheaftend approach the inviscid values. The difference between
oftheflare.Previousstudieosnsimilacronfigurationsthe experimental axial coefficient data at the highest
havenotedthesamephenomenl'o. nFigure8isaplot Reynolds number and the inviscid CFD data is two
of the computedpressurecoefficientalongthe percent or less. This implies that the influence of
windwardcenterlineforMach6 CF4at20and24
viscosity on CA is small beyond 4 million per foot.
degreeasngleofattack.The figure shows the decrease (This conclusion was also reflected in unpublished data
in pressure coefficient over the aft end of the flare as at Mach 20 in Helium, which had a larger Re variation
the angle of attack is increased. compared with the air data at Mach 6. The Reynolds
The nose, cylinder, and flare components to the numbers (based on body length) for Helium were
aerodynamic coefficients from the LAURA solutions at equivalent to flight values). At low angles of attack (<
Mach 6 in air are plotted in Figure 9. At low angles, the 10 degrees), experimental values of Cm have a small
contribution from the sum of the nose and cylinder to dependence with Reynolds number and the difference
the aerodynamic coefficients is comparable to or is between the high Reynolds number experimental data
greater than the value from the flare. With increased and the inviscid data is also small, but observable. In
angle of attack, the flare contribution to the this region, the high Reynolds number data exhibit a
aerodynamic coefficients grows more rapidly than the larger instability (larger Cm values). While the
contribution associated with the nose and cylinder. dependence on pitching moment with Reynolds number
With the exception of a very slight non-linearity in CN is small, a change is present in trim angle of attack
from the cylinder (when compared with non-linear between the high Reynolds number experimental data
behavior on flare), nose and cylinder aerodynamic and the inviscid data. For Mach 6 air, approximately a
contributions are essentially linear with angle of attack. two-degree change was observed between the high
The CFD data in Figure 9 indicate the non-linear nature Reynolds number experimental data and computational
in the aerodynamic coefficients with angle of attack is inviscid data. This small change in trim angle is
derived from the flare. This non-linear behavior is significant because it occurs at a location were the L/D
consistent with similar blunt-nosed flare-stabilized value varies rapidly with angle of attack. This change
configurations 2° and is due to the variation of the in trim angle translates into an increase in trimmed L/D
pressure distribution for the flare. with Reynolds number at Mach 6 and is indicated on
Figure 10(d). Finally, longitudinal stability, measured
by (Cm),_, while difficult to infer from Figure 10(c), is
Reynolds number effects
The effects due to the variation in Reynolds slightly enhanced with increasing Reynolds number.
number on longitudinal aerodynamics are conveyed
Mach number effects
through plots of CA, CN, Cm, and L/D as displayed in
Just as differences in shock layer properties
Figure 10. Measured Mach 6 air data at Reynolds
numbers equal to 0.5, 1.0, 2.0, and 4.0 million per foot, induced by high temperature chemistry can affect the
and Mach 6 inviscid air CFD predictions from the Data- aerodynamic character of a vehicle traveling at hyper-
Parallel Lower-Upper Relaxation (DPLUR) 21 flow velocities, differences in shock layer properties induced
solver are shown. The inviscid data represent the bound by changes in Mach number under perfect gas
in the data with increasing Reynolds number. It should conditions can also affect aerodynamic character.
be noted, the DPLUR solution was non-inclusive of the Shock standoff and shock angle are a function of the
conditions across the shock and variations in standoff or
wake; consequently, the experimental axial force
coefficient data in this section does not include base angle can alter surface pressure.
pressure contributions. (The base pressure contributes In Figure 1l, the effects on aerodynamics that
approximately 2 to 4 percent to the total axial force at result from changes in Mach number are displayed
this condition). through plots of CA, CN, Cm, and L/D data measured at
The data presented in Figures 10(b) and 10(d) show Mach 6 and l0 at a Reynolds number of approximately
the effect of Reynolds number on CN, and L/D is, in 2.0 million per foot. Initially, values of CN and CAare
general, negligible across the angle of attack range generally lower for Mach 10, with the gap closing as 20
considered. However, the data presented in Figures degrees is approached. Near of 20 degrees, Mach l0
values exceed those at Mach 6; however, at the largest
10(a) and 10(c) indicate Reynolds number induced
changes exist in Cm and exist as expected in CA. At a angle considered, the bow-shock flare-shock interaction
reverses the order, causing values of CNand CAat Mach
given angle of attack, the difference between the
maximum and minimum experimental axial force 6 to exceed their respective values at Mach 10. Lift-to-
coefficient value at Mach 6 in air is approximately 10 drag trends follow those noted for Cr_and CA, except at
percent. The largest portion of this percentage occurs the highest angle of attack. The characteristics of the
between 0.5 and 1.0 million per foot. Beyond 1.0 pitching moment curve are also affected. Data in
5
American Institute of Aeronautics and Astronautics Paper 2001-0562
Figure1l(c)reveal
C m increases with Mach number for negligible differences due to 3'. Past 8 degrees, the
angles of attack less than 15 degrees. At angles greater influence of 7 becomes significant and differences are
than 15, Mach 10 values are more negative. At the produced which increase rapidly with angle of attack.
highest angles (_ > 20), the change in the value of Beyond 20 degrees, the change in Cm character
Mach 10 pitching moment coefficient is a result of the exhibited in CFo is a result of the shock-shock
bow-shock flare-shock interaction influencing flare interaction. Due to the higher density ratio in CFa, the
pressures. The interaction onset angle is lower for shock detachment distance is less than that for air;
Mach 10 due to its smaller bow shock standoff distance. therefore, the influence of the expansion fan on the flare
The effect of Mach number on longitudinal stability, occurs at lower angle of attack. Differences in shock
(Cm),_, is inferred from Figure ll(c). For angles structure are discussed inthe next section.
between 0 and 10 degrees, the stability of the baseline The variation in aerodynamic coefficients between
OV is reduced with increasing Mach number. This air and CF4 arise from _' induced pressure differences.
finding is consistent with data contained in Ref. 22 for These pressure differences are small on the nose and
similar configurations at 0 incidence. At trim, the cylinder and are significant on the flare. Therefore, at
vehicle is more stable at Mach 10 than at Mach 6, low angles, where the nose and cylinder contribute most
indicating enhanced longitudinal stability with to the aerodynamics, aerodynamic variations due to ),
increasing Mach number. This trend continues to 20 are small. At higher angles, when aerodynamic
degrees angle of attack. The largest Mach number contributions from the flare become significant,
induced differences in lift-to-drag are found in the differences in aerodynamic characteristics are
region surrounding the trim angle. While the delta in prominent.
L/D appears significant, a Mach number induced Consistent with past sources, the data show
change in trim angle of attack causes the trim L/D value changes in _, can significantly alter the aerodynamic
to increase slightly with Math number. character. An increase in flare effectiveness is observed
in CF4 until the bow-shock flare-shock interaction
effects influences pressures on the flare; furthermore, trim
Under certain conditions, the CF4 Tunnel can angle and trim L/D are affected. The trim angle in
provide a simulation of the effect ofy and density ratio flight will be determined by the balance of aerodynamic
and thereby provide good estimates of flight forces, unaided by moveable control surfaces;
aerodynamics 23. However, in general, it is not assumed consequently, knowing the influence of 3', or any
that testing with CF4 gas will predict flight variable, on the trim angle is critical with respect to L/D
aerodynamics. But rather, by comparing Mach 6 air and aerothermal issues. Based on the data in Figure 12,
with Mach 6 CF4 data, the two Facilities provide insight not accounting for high temperature chemistry would
on how the aerodynamic characteristics of a returning result in an over prediction in trim angle of attack.
space transportation vehicle is influenced by )' and Inaccuracy in trim angle would affect the guidance and
density ratio. Using a heavy gas, CF4, as the test control algorithms and the split line layout for the TPS.
medium, post normal-shock values of 7 (-1.1) and post Also shown is a considerable over prediction in L/D at
normal-shock values of density ratio (=12) on the order trim, which would affect cross-range estimates.
of those encountered during hyper-velocity flight are
created in a ground-based facility. Consequently, shock Flow visualization
detachment distance and pressure compression/ The schlieren images shown in Figure 13 convey
expansion behavior similar to that encountered during the OV bow and flare shock locations in Mach 6 air at a
flight are produced. The lower values of 7 associated 2 million per foot Reynolds number condition at 0, 10,
with reacting gas air affect aerodynamics by producing and 20 degrees angle of attack. The dashed line
higher compressions and lower expansions in pressure overlaid on each schlieren image indicates the
relative to those produced by perfect gas air. Because numerically predicted bow shock location from the
of this, it is especially critical to quantify this behavior LAURA code for Mach 6 air at a freestream condition
when the vehicle has a compression or expansion corresponding to experiment. The dot-dashed line
surface _-4. overlaid on each image indicates the numerically
The effect of y on longitudinal aerodynamic values predicted bow shock location for Mach 6 CF4.
is presented in Figure 12. In the figure, CA, CN, Cm.and Comparing the two lines illustrates the dissimilar shock
L/D data measured at Mach 6for a Reynolds number of detachment distance that result from changes in density
approximately 0.5 x 106 per foot in air and CFa are ratio. The normal-shock density ratio for CF4 is more
plotted. The data in Figure 12(a) indicate )' has a small than twice the value for air. For the same angle of
(3-4 percent) effect on CA. For angles up to attack, compared with air, the shock detachment
approximately 8 degrees, Cr,, Cm, and L/D show distance is smaller for CFa and more closely simulates a
-6-
American Institute ofAeronautics and Astronautics Paper 2001-0562
shockstructureencountered during Earth entry. The effect of Reynolds number at Mach 6 in air on
Evident in Figure 13 is the outline of the OV, the non-dimensional heating along the OV windward
location of the bow and flare shock, and the bow-shock centerline at 8, 12, and 16 degrees is presented in
flare-shock intersection at the highest angle of attack. Figure 15. At all angles considered, the aeroheating
The agreement of prediction with experiment for the data along the entire length of the cylinder is
flare shock location is excellent. From this agreement it independent of Reynolds number, indicating laminar
can be concluded that the numerical method is flow. At 8 degrees (see Figure 15(a)), the flow over
accurately capturing the shock and is reproducing the flare remains laminar up to 4 million per foot, and
density ratios across the shock in agreement with appears transitional/turbulent at 8 million per foot. The
experiment. The agreement between measured and measured aeroheating data on the flare at 12 degrees
computed bow shock location for Mach 6 CF4 is (see Figure 15 (b)) varies with Reynolds number, but
comparable to that displayed for Mach 6 air and images the trend is not indicative of transitional/turbulent flow
are not presented to avoid redundancy. Based predicted since the heating levels do not increase with Reynolds
CF4 shock locations, in flight near trim, it appears the number. However, measured heating levels were near
shock-shock interaction will not affect the those for laminar CFD predictions. Approximately half
aerothermodynamic or aerodynamic characteristics of way down the flare, the 16-degree aeroheating data at 4
the OV. million per foot (see Figure 15(c)) rapidly increase
Surface heatin_ compared with values at lower Reynolds numbers.
Measured surface heating levels, inferred using Phosphor images displayed as a function of Reynolds
phosphor thermography, were obtained at Mach 6 and number are shown in Figure 16. A localized increase in
10 in air at the conditions listed in Table 1. These heating is visible on the aft end of the flare at the 4
measured data, along with predicted laminar values million per foot condition. The onset of smooth body
from the LAURA code, are presented in non- transition is believed to be responsible for the increase
dimensional form in Figure 14. Primary emphasis is and schlieren images (see Figure 17) confirm the bow-
given to the windward surface. Data along the shock flare-shock interaction is not responsible. The
windward centerline at three angles of attack that are in size of the separated flow region in front of the flare did
the vicinity of the expected trim angle of attack are not change significantly with Reynolds number and the
shown. Specifically, measured data at 8, 12, and 16 boundary layer state on the flare transitioned at a lower
degrees are considered. While there is a 1 to 2 degree Reynolds number as the angle of attack increased.
difference in angle of attack between measured and Based on measured heating data at 8 and 16 degrees, it
computed values, in general, the two techniques exhibit is speculated that the flow at 12 degrees transitions
similar trends and have good qualitative agreement. from laminar to transitional/turbulent beyond 4 million
On the cylinder, the computed heating values at Mach 6 per foot.
are higher than measured values (see Figure 14 (a));
however, good quantitative agreement is well within the Concluding Remarks
stated uncertainty of the measured data. On the flare, Aerodynamic and aeroheating data measured at
the computed data at 10 degrees are bound by the Langley Research Center's 20-Inch Mach 6 CF4, 20-
experimental data at 8 and 12 degrees and good Inch Mach 6 Air, and 31-Inch Mach 10 Air Tunnels on
agreement is within the uncertainty of the measured a one-percent scale model of Kistler Aerospace
data. At 15 degrees, computed data over the first half Corporation's baseline orbiter vehicle have been
of the flare is in qualitative agreement with measured presented and compared with corresponding
data at 16 degrees; like the data at 10 degrees, good computational data predicted using the LAURA code.
agreement is within the uncertainty of the measured For aerodynamic coefficients and aeroheating levels,
data. Over the last half of the flare, heating trends with good agreement was observed between measured and
Reynolds number at 16 degrees that are presented in computed values. This good agreement, which exists
Figure 15(c) indicate transitional/turbulent. Computed over awide range in incidence (-4 to 24 degrees), Mach
and measured data at Mach 10 agree well with each number (6 to 10), Reynolds number (0.5 to 4.0 million
other along the cylinder and the flare. The agreement per foot), and ratio of specific heats (1.2 to 1.4),
of the experimental data with the laminar predictions is demonstrates the ability of the CFD tool to reproduce
indicative of the measured data being laminar on the data at varied wind tunnel conditions and augments the
windward surface. The agreement of measured with credibility of LAURA to predict aerodynamic data and
computed data is well within the uncertainty limits. aeroheating data at flight conditions.
Like the Mach 6 data at 10 degrees, the Mach 10 data at A non-linear behavior with incidence was observed
10 degrees are also bound by the experimental data at 8 in the measured aerodynamic data and was examined
and 12 degrees. with CFD. It was determined from computational data
7
American Institute ofAeronautics and Astronautics Paper 2001-0562
thattheflow
over the flare was responsible for the non- Acknowledgements
linear behavior; furthermore, the computational data The authors wish to acknowledge Christopher J.
showed that the contributions to aerodynamic Riley formerly of NASA Langley for providing the
coefficients derived from the flow over the nose and
Mach 6 DPLUR inviscid predictions and Stephen J.
cylinder were linear versus incidence. At small angles Alter of NASA Langley for providing the surface and
of attack, the aerodynamic characteristics of the OV volume grids.
were primarily driven by pressure over the nose and References
cylinder. Near of 10 degrees angle of attack, the
iProctor, P., "Kistler Seeks to Create 'UPS of Space,' "
influence from the flare on aerodynamic characteristics
Aviation Week and Space Technology, June 30, 1997,
was noticeable; beyond this angle, aerodynamic
I_P"53-55
characteristics were primarily driven by pressure over
- Moorman, T., "The Explosion of Commercial Space
the flare.
and the Implications for National Security," AIAA
The influence of Mach number, Reynolds number,
Paper 98-0002, January 1998.
and ratio of specific heats on aerodynamic
3 Mueller, G. E., "A Low Cost Aerospace Vehicle."
characteristics was also examined. Trim angle and trim
Acta Astronautica, Vol. 39, No. 1-4, 1996, pp. 239-244.
L/D increased slightly with increased Mach number.
Cheatwood, F. M. and Gnoffo, P. A., "User's Manual
At Mach 6 in air, a Reynolds number effect was
for the Largely Aerothermodynamic Upwind
observed; trim angle and trim L/D increased with
Relaxation Algorithm (LAURA)," NASA-TM4674,
Reynolds number. The ratio of specific heats had the
April 1996.
greatest influence on aerodynamic values at trim.
5Micol, J.R., "Langley Aerothermodynamic Facilities
Upon comparing trim angle of attack and L/D at trim in
Complex: Enhancements and Testing Capabilities,"
CF4 with perfect gas air values, a significant reduction
AIAA Paper 98-0147, January 1998.
in both was observed in CF4. Pressure levels over
6 Micol, J. R., "Hypersonic Aerodynamic
compression surfaces, like the OV flare, vary greatly
/Aerothermodynamic Testing Capabilities at Langley
with ratio of specific heats. The observed trends were a
Research Center: Aerothermodynamic Facilities
result of these pressure variances over the flare. If
Complex," AIAA Paper 95-2107, June 1995.
aerothermo-chemistry effects were not accounted for, a
7 Weilmuenster, K. J., Gnoffo, P. A., Greene, F.
significant over prediction in trim angle and L/D at trim
A.,"Navier-Stokes Simulations of Orbiter Aerodynamic
would result. Because ideal and reacting gas flows
Characteristics Including Pitch, Trim, and Bodyflap,"
have dissimilar pressure expansions/compressions and
Journal of Spacecraft and Rockets, Vol. 31, No. 3,
shock structures, all factors that can significantly
May-June 1994, pp.355-366.
influence aerodynamics, the influence of )' must be s Gnoffo, P. A., Weilmuenster, K. J., and Alter, S. J.,
accounted for when designing, screening, or evaluating
"Multiblock Analysis for Shuttle Orbiter Re-Entry
any entry vehicle.
Heating From Mach 24 to Mach 12journal of
Measured aeroheating levels were inferred from the
Spacecraft and Rockets, Vol. 31, No. 3, May-June
thermographic phosphor technique in air at Mach 6.
1994, pp.367-377.
These aeroheating data were in the vicinity of the trim 9Gnoffo, P. A., Braun, R. D., Weilmuenster, K. J., and
angle of attack and varied from 0.5 to 8.0 million per Mitcheltree, R. A., "Prediction and Validation of Mars
foot in Reynolds number. Along the entire windward
Pathfinder Hypersonic Aerodynamic Data Base," AIAA
centerline of the OV cylinder, the flow was laminar for
Paper 98-2459, June 1998.
all Reynolds numbers and angles considered. At 8 l0 Hamilton, H. H., Greene, F. A., and Weilmuenster,
degrees, measured aeroheating data on the flare were
K. J.,"Comparison of Heating Calculations with
independent of Reynolds number up to 4 million per
Experimental Data on a Modified Shuttle
foot, suggesting laminar flow. At 8 million per foot, the
Orbiter,"Journal of Spacecraft and Rockets, Vol.29,
data were no longer independent of Reynolds number
No.2, March-April 1992, pp. 208-215.
and the flow appeared transitional/turbulent. At 12
11 Roe, P. L., "Approximate Riemann Solvers,
degrees, the measured heating data on the flare varied Parameters Vectors, and Difference Schemes," Journal
with Reynolds number and had a trend with Reynolds
of Computational Physics, Vol. 43, No. 2, 1981, pp.
number that was not indicative of transitional/turbulent
357-372.
flow. However, measured heating levels were in the
12 Harten, A., "High Resolution Schemes for
vicinity of those for laminar CFD predictions. At 16
Hyperbolic Conservation Laws, " Journal of
degrees, the flow over the aft end flare is believed to be
Computational Phvsics, Vol. 49, No. 3, 1983, pp. 357-
transitional/turbulent at 4 million per foot. When 393.
compared with data at lower Reynolds numbers, a rise 13 Yee, H. C., "On Symmetric and Upwind TVD
in measured aeroheating values was observed at 4
Schemes," NASA TM-86842, September 1985.
million per foot condition.
-8-
American Institute of Aeronautics and Astronautics Paper 2001-0562