UNCLASSIFIED
Defense Technical Information Center
Compilation Part Notice
ADPO 10740
TITLE: Measurement Requirements for Improved
Modeling of Arcjet Facility Flows
DISTRIBUTION: Approved for public release, distribution unlimited
This paper is part of the following report:
TITLE: Measurement Techniques for High Enthalpy
and Plasma Flows [Techniques de mesure pour les
ecoulements de plasma et les ecoulements a haute
enthalpie]
To order the complete compilation report, use: ADA390586
The component part is provided here to allow users access to individually authored sections
of proceedings, annals, symposia, ect. However, the component should be considered within
the context of the overall compilation report and not as a stand-alone technical report.
The following component part numbers comprise the compilation report:
ADP010736 thru ADP010751
UNCLASSIFIED
3A-1
Measurement Requirements for Improved Modeling of Arcjet
Facility Flows
Presented by
Douglas G. Fletcher
Reacting Flow Environments Branch
NASA Ames Research Center
Moffett Field, CA, 94035-1000
1. Introduction 3A-2
1.1 Historical Development of NASA Ames Arcjet Facilities 3A-2
1.2 Overview of Current NASA Arcjet Facilities 3A-4
1.3 Ames Aerodynamic Heating Facility Arcjet 3A-4
1.4 Arcjet Characterization Using Conventional Instrumentation 3A-5
2. CFD for Arcjet Flows 3A-8
2.1 Motivation for Arcjet Flow Modeling 3A-8
2.2 CFD Requirements for Arcjet Flow Simulations 3A-8
2.3 Strategies for Arcjet Flow Simulations 3A-9
3. Measurement Requirements for Arcjet Flow Modeling 3 A- 10
3.1 Enthalpy 3A-11
3.2 Arc Heater 3 A- 11
3.3 Arcjet Nozzle and Free-Stream Flow 3A-12
3.4 Blunt-Body Shock-Layer Flow 3A-12
3.5 Measurement Accuracy Requirements 3A-13
4. Experimental and Computational Investigation of Shock-Layer Flows 3 A- 14
4.1 Objectives of Investigation 3A-14
4.2 Experimental and Computational Approach 3 A- 14
4.3 High Pressure Case 3A-16
4.4 Low Pressure Case 3A-20
4.5 Spatially Resolved Measurements 3A-23
4.6 Lessons Learned 3A-23
5. Summary and Recommendations 3A-24
6. Acknowledgements 3A-25
7. References 3A-25
Paper presented at the RTO AVT Course on “ Measurement Techniques for High Enthalpy and Plasma Flows”,
held in Rhode-Saint-Genese, Belgium, 25-29 October 1999 , and published in RTO EN-8.
3A-2
1. Introduction
Current efforts to develop new reusable launch vehi-
cles and to pursue low-cost robotic planetary missions
have led to a renewed interest in understanding arc-
jet flows. Part of this renewed interest is concerned
with improving the understanding of arcjet test results
and the potential use of available computational-fluid-
dynamic (CFD) codes to aid in this effort. These CFD
codes have been extensively developed and tested for
application to nonequilibrium, hypersonic flow model-
ing. It is envisioned, perhaps naively, that the appli-
cation of these CFD codes to the simulation of arcjet
flows would serve two purposes: first, the codes would
help to characterize the nonequilibrium nature of the
arcjet flows; and second, arcjet experiments could po-
tentially be used to validate the flow models. These
two objectives are, to some extent, mutually exclusive.
However, the purpose of the present discussion is to
address what role CFD codes can play in the current
arcjet flow characterization effort, and whether or not
the simulation of arcjet facility tests can be used to
evaluate some of the modeling that is used to formu-
late these codes.
This presentation is organized into several sections.
In the introductory section, the development of large-
scale, constricted-arc test facilties within NASA is re-
viewed, and the current state of flow diagnostics using
conventional instrumentation is summarized. The mo-
tivation for using CFD to simulate arcjet flows is ad-
dressed in the next section, and the basic requirements
for CFD models that would be used for these simula-
tions are briefly discussed. This section is followed by
a more detailed description of experimental measure-
ments that are needed to initiate credible simulations
and to evaluate their fidelity in the different flow re-
gions of an arcjet facility. Observations from a recent
combined computational and experimental investiga-
tion of shock-layer flows in a large-scale arcjet facility
are then used to illustrate the current state of develop-
ment of diagnostic instrumentation, CFD simulations,
and general knowledge in the field of arcjet character-
ization. Finally, the main points are summarized and
recommendations for future efforts are given,
1.1 Development of NASA Ames Arcjet Facili-
ties
Development efforts that led to what we now classify
as arcjet test facilities began in the late 1950’s with the
goal of producing a continuously operable hypersonic
ground test facility. This need was driven by both
US Department of Defense and NASA mission plan-
ning requirements. From the NASA side, planetary
missions and the manned space program were push-
ing aerospace vehicles to higher aerothermodynamic
heating rates. Several excellent texts have been writ-
ten that include a much broader treatment of the his-
torical development of arcjet facilities and plasma arc
devices for propulsion. 1,2 However, for the purpose of
introducing the current topic, a brief recapitulation of
arcjet facility development activities at NASA Ames
Research Center is given below.
Fig. 1. NASA Ames concentric ring arcjet.
The rather ambitious target capabilities for develop-
ing the first Ames arcjet were: 1) 32 MJ/kg enthalpy;
2) 100 atm pressure; 3) 1 MW input power; and 4)
continuous and contaminant-free operation. The first
successful arcjet that even partially met some of these
goals was the Ames Concentric Ring Arcjet, 3 which
is depicted in Fig. 1. While the device could operate
at the intended high pressures, it had a very low effi-
ciency in terms of coupling the electrical energy to the
flow. As can be seen in Fig. 1, the arc region is quite
small, and most of the incoming air stream bypasses
the arc. This resulted in relatively low deposition of
energy into the test gas stream.
Fig. 2. Early Ames constricted- arc heater.
To improve the coupling of electrical energy into the
flow, the next round of heater configurations featured
more widely spaced electrodes separated by an orifice
plate that is intended to constrict the arc to a rela-
tively small region. It was hoped that forcing the flow
and arc through the same small region would improve
the electrical energy deposition and raise the stream
enthalpy. Figure 2 shows an example of this device,
which did show an improvement in energy deposition.
However, it proved to be nearly impossible to prevent
the arc from attaching at the edge of the orifice plate,
3A-3
and excessive arc- induced failures produced further de-
sign modifications.
Subsequent efforts resulted in the development of su-
personic arcjets, 4 which achieved high enthalpies and
low heat loss by extending the arc through the throat
region before attachment downstream in the low pres-
sure, expanded flow region. A schematic of one of the
earlier versions is shown in Fig. 3. Erosion of the down-
stream attachment point was minimal for this type
of arcjet because of the diffuse nature of the arc at
the low pressures of the supersonic flow region. The
constrictor diameter was only 6.4 mm, but the heater
performance was pretty much as predicted, and there
appeared to be a substantial gain in electrical energy
deposition. Shortly thereafter a second supersonic arc-
jet was developed with a 25.4 mm diameter constric-
tor and this device delivered enthalpies on the order of
900 MJ/kg on the flow centerline. 5 Unfortunately, the
stream was highly nonuniform and the excessive radial
gradients limited the application range of this heater.
Fig. 3. NASA Ames supersonic, constricted- arc facil-
ity.
In one of the more interesting developments that has
particular relevance to the current discussion, the
ARCFLO code was developed in 1967 to model pro-
posed arc heater configurations. 6 For the instrumental
technology available at the time, comparisons between
arcjet performance measurements and ARCFLO pre-
dictions were satisfactory. This led to the use of the
code in the development of new heater configurations.
Although impressively high enthalpy levels were gen-
erated in some of these early devices, there was no
great demand for routine operation at those condi-
tions. Instead, the emergence of the shuttle as the
primary launch and payload capability for NASA gen-
erated a significant demand for test capability in the
20 to 30 M J /kg range to develop and qualify shuttle-
related thermal protection materials. With the ex-
ception of meteor ablation studies and work involved
with the development of heat shields for planetary-
entry missions, 7 this test condition range has proven
satisfactory for a majority of the aerospace commu-
nity’s needs. Progress in providing robust test facil-
ities in the required performance range was enabled
by timely improvements in magnetically driven elec-
trode technology. 8 All of the successful heater designs
relied on magnetic fields to spin the arc attachment
point around the electrode to reduce the local heat-
ing. However, an optimal combination of geometry,
current load, and magnetic field strength leading to ex-
tended electrode lifetime could only be found through
trial and error, since theoretical models of the com-
bined fluid and plasma dynamics of the electrode were
inadequate at that time.
Using the new magnetically driven electrodes, the
Ames 20 MW Constricted Arc Jet was built in 1972.
A schematic rendering of the constrictor, downstream
electrode package and nozzle configuration is shown in
Fig. 4. This basic constricted-arc heater configuration
has been used continuously, with relatively little vari-
ation, in the Ames arcjet facilities since that time. An
excellent description of the electrode and constrictor
design and performance evaluation is given as part of
the report on the Ames 60 MW arcjet, 9 which is still
in use today.
Fig. 4. Current version of Ames constricted- arc heater.
In the late 60’s and early 70’s arcjets were in use at
aerospace companies and research centers around the
world. It appeared that arcjets would find extensive
use as aerot her mo dynamic test facilities where funda-
mental investigations of real gas phenomena could be
conducted. Although they did not provide perfect sim-
ulation of atmospheric flight environments, arcjets had
a significant advantage over impulse facilities in that
they could be operated at high enthalpy levels for long
periods of time. Unfortunately, it proved to be very
difficult to establish just what operating enthalpy level
was actually reached. In fact, the inability to charac-
terize the arcjet stream conditions ultimately limited
arcjets mainly to applications where complete knowl-
edge of stream conditions was not a requirement for
evaluating test results.
In a review of ground- test facility simulations of
poorly understood real-gas phenomena, Park 10 iden-
tified seven important problems: 1) determining aero-
dynamic parameters; 2) viscous/shock interactions; 3)
boundary layer transition; 4) understanding leeward
or base region flows; 5) nonequilibrium radiation; 6)
nonequilibrium ionization; and 7) surface catalysis.
Park then examined the capabilities of three types of
hypersonic ground-test facilities that could be used for
fundamental investigations of these problems: 1) im-
pulse facilities (including shock tunnels); 2) ballistic
ranges; and 3) arcjets. Arcjets were only deemed suit-
3A-4
able for studies of nonequilibrium radiation and sur-
face catalysis, and even then adequate specification of
the stream conditions was mentioned as a requirement
for improving the analysis of test results. 10
Before discussing the issue of stream conditions fur-
ther, it is useful to examine the current status of large-
scale arcjet facilities and their role in thermal protec-
tion material test and development. In addition, it is
instructive to examine the use of conventional stream
characterization instrumentation and how it is used in
the interpretation of test measurements.
1.2 Overview of Current NASA Arcjet Facili-
ties
Today, NASA’s large-scale arcjet facilities are used
mainly to simulate aero thermal heating environments,
although there is still some limited use in evaluating
supersonic air-breathing propulsion concepts. Our dis-
cussions will focus exclusively on facilities, modeling,
and measurements that relate to the principal appli-
cation: aerothermal heating simulation. Two NASA
Centers, Johnson and Ames, are currently operating
segmented- type constricted-arc heater facilities for this
application. This facility is the workhorse for the 20
to 30 MJ/kg enthalpy range of long-duration thermal
testing.
Fig. 5. Current version of JSC TP-1 constricted-arc
heater and nozzle.
Arcjet facilities at Johnson Space Center support TPS
testing requirements for manned missions. All of
the thermal protection materials for shuttle, includ-
ing tiles, coatings, and fillers, are qualified for use on
the basis of tests in these arcjets. Johnson has two 10
MW facilities JSC TP-1, which became operational in
1973, and JSC TP-2, which was upgraded to 10 MW in
1991. 11 Both facilities have segmented, constricted-arc
heaters. The TP-1 facility is usually arranged with a
conical nozzle configuration for st agnation- point test-
ing, while TP- 2 is typically configured with a rectan-
gular channel for flat-plate testing. A schematic of
the Johnson TP-1 heater and nozzle is presented in
Fig. 5, and it shows two noteworthy features. First,
because of the tungsten cathode, Oo is injected sep-
arately from N 2 further downstream in the heater to
prolong the useful life of the electrode. Although they
are injected separately, the two gases are thought to
be mixed by the time the downstream electrode pack-
age is reached. The second interesting feature is that
the throat diameter is larger than that of the constric-
tor, which causes some uncertainty regarding the sonic
location. Axial velocities in the arc column could actu-
ally be quite high, which may inhibit mixing of the Oo
and No streams. The facility is equipped with energy
balance instrumentation, which provides a measure of
the bulk enthalpy for each test.
The Arcjet Complex at Ames Research Center sup-
ports Ames’ role as lead NASA Center for thermal
protection material development. There are currently
three operating segmented, cons trie ted- arc facilities:
the Aerodynamic Heating Facility (AHF) and the
Panel Test Facility (PTF) are both rated at 20 MW ;
and the Interactive Heating Facility (IHF) is rated at
60 MW. There are also two operable Hue Is- type heater
facilities, 2x9 Turbulent Flow Facility (TFF) and the
Direct Connect Arcjet Facility (DCAF). Two arcjet fa-
cility buildings house the different arcjets, which share
common steam-ejector vacuum and water-cooling sys-
tems. With a shared vacuum system, only one facility
can operate at a time. However, facilities can operate
sequentially throughout the day with up to 8 runs dur-
ing a single operating shift. Note that the operating
frequency for an arcjet is greater than that of typical
large-scale impulse facilities.
A cross section of a typical Ames constricted- arc
heater configuration was shown above in Fig. 4. The
configuration is different from the JSC TP- 1 configura-
tion that was shown in Fig. 5. For the Ames heater, the
throat diameter is smaller than the constrictor diame-
ter, so the sonic point will always be located between
the converging and diverging sections of the nozzle.
Also, both the upstream and downstream electrodes
are copper, so oxygen does not need to be injected
separately for air tests. Since the overwhelming ma-
jority of arcjet tests at Ames Research Center are per-
formed using segmented- type, constricted- arc heaters,
Huels-type heaters will not be discussed further.
Even though they are both classified as segmented,
constricted- arc heaters, the different designs of the
JSC and Ames heaters illustrate the variety of elec-
trodes and nozzles that are in use today. There is
no standard design. Consequently, performance will
vary widely from facility to facility and characteriza-
tion of the performance of one facility is by no means
applicable to others unless the configuration is exaetty
duplicated.
1.3 Ames Aerodynamic Heating Facility Arcjet
The Aerodynamic Heating Facility (AHF) Arcjet at
NASA Ames Research Center is an example of cur-
rent large-scale, constricted-arc heater test facilities.
A schematic of the facility is shown in Fig. 6. Facil-
ity operation is initiated by evacuating the arcjet and
then striking an arc in a low-pressure argon stream. 12
3A-5
The test gas flow, usually air or nitrogen, is then in-
troduced through the segmented disks along the col-
umn, and the arc current is adjusted to achieve the
test conditions. Within the arc column, heating by
the electrical discharge causes substantial dissociation
and ionization of the test gas. The argon start-gas
stream is maintained during operation, and additional
argon is injected to protect the downstream electrode.
Each electrode package is made up of a series of al-
ternating copper rings and spacer disks. The rounded
rings are the actual electrodes, and they protrude into
the stream to move the arc attachment away from the
wall (see Fig. 4). Magnetic windings inside the elec-
trodes rotate the arc attachment point to reduce the
heat load on the electrodes. Each electrode can carry
up to 500 A of current. Typically, the anode is placed
at the upstream end of the arc column to benefit from
further cooling by the test gas.
Test Chamber
Fig. 6 Schematic of the NASA Ames AHF Arcjet.
Upon leaving the heater, the flow is accelerated to hy-
personic speed through a conical, converging-diverging
nozzle. During the expansion-driven acceleration, the
collision frequency decreases rapidly in the nozzle and
the thermochemical state of the flow departs from
equilibrium. At some point, the flow chemistry be-
comes frozen, and this may be followed by freezing
of the internal energy distribution of the molecular
species. Various nozzle sections can be used to pro-
vide expansion ratios ranging from 64 to 576. The
flow exits the nozzle and continues expanding into a
cabin where material tests are conducted. Material
samples are typically inserted into the stream 36 cm
downstream of the nozzle exit. Test durations of up to
20 minutes are possible, depending on the particular
conditions. During the tests, the stagnation pressure,
cabin pressure, and arc heater conditions are continu-
ously monitored.
In a typical test cycle, a preliminary analysis of the ex-
pected heat load in a flight application has been per-
formed and a candidate thermal protection material
has been selected for testing in an arcjet flow. The
test conditions are chosen to attempt to match the ex-
pected heat flux for a particular point on a predicted
trajectory, such as the peak heating point. Conven-
tional instruments, which will be discussed below, are
used to verify the test conditions. The test results
are then interpreted without the benefit of full knowl-
edge of the stream conditions. Currently, relating test
results from the arcjets to the intended flight applica-
tion is more of an art than a science, because the arcjet
stream conditions are not sufficiently characterized.
1.4 Arcjet Characterization Using Conven-
tional Instrumentation
The words “Arcjet Characterization” are typically un-
derstood to mean specifying the state of the arcjet test
stream, and they are referred to throughout this dis-
cussion in that context. Although it is important in
flow modeling, the need for arcjet characterization is
driven primarily by the needs of thermal protection
material developers, who need better specification of
the stream conditions to relate the results to flight en-
vironments. In addition, an improved understanding
of arcjet stream conditions in general may also make
arcjets more suitable for fundamental studies of real
gas phenomena.
The state of arcjet stream characterization in the early
90’s was summarized in an excellent and thoughtful
review article by Scott. 13 Both established and novel
instrumental techniques were critically reviewed in the
article. The article focused mainly on how various di-
agnostic techniques could be used to characterize the
most important stream variables: enthalpy and the de-
gree of nonequilibrium in the stream. Rather than re-
peat this review, some of the more widely used conven-
tional diagnostics are reviewed briefly below. The lim-
itations of these measurement techniques are discussed
to provide background for considering what measure-
ments are required to improve arcjet flow modeling.
Newer, less widely used spectroscopic techniques, such
as multiphoton spectroscopy will be mentioned later,
and are discussed more fully in the second article.
Traditional instruments that are used to obtain flow
property measurements include pitot probes and
calorimeters. Additional instruments, such as thermo-
couples and flow meters are used to measure coolant
flow rates and temperature rise to perform an energy
balance on the facility. Stream surveys are usually per-
formed with a traversing, sting- mounted probe, since
the facility can operate continuously and at a level
where the instrument can give an equilibrated response
to the quantity being measured.
Pitot measurements yield the stagnation pressure be-
hind a shock wave that is generated by the probe.
For much of the operating range of today’s large-scale
arcjet facilities, the pitot, or impact, pressure can be
related to the dynamic pressure of the flow, pv 2 / 2,
through the Rayleigh supersonic pitot relation 14 ,
M » L
(i)
3A-6
In the above expression, 7 is the ratio of specific heats
for the gas, M is the Mach number, p P is the pitot
pressure, p is the stream density, and v is the veloc-
ity. Although p and v are both important stream vari-
ables for arcjet flow characterization, a determination
of each variable cannot be made without an additional
measurement. For typical facility operating conditions
the flow velocity is a considerably larger quantity than
the stream density, so a strategy for determining both
variables should involve a velocity measurement.
Energy Balance - Most arcjet facilities are equipped
with instruments that can be used to perform an en-
ergy balance on the arcjet facility as a whole. Owing
to its simplicity, the energy balance approach remains
by far the most commonly used for characterizing the
arcjet stream. The basic principle of the measurement
is illustrated in Fig. 7, which shows the arc-jet opera-
tion measurements that must be acquired to perform
the energy balance. A simple first law relationship is
invoked for the system,
m h av g = VI - mc p (A T on - A T off ) , (2)
where m is the mass flow rate, /i av g is the bulk en-
thalpy, V is the arc voltage, I is the arc current, T
and c p are the coolant temperature and specific heat,
and the subscripts of AT refer to a measurement of
the temperature rise with the arc on and with the
arc off. This is required to account for the coolant
temperature rise that results from pumping a viscous
fluid through the cooling lines. An uncertainty anal-
ysis for typical measurement errors can be performed,
and this indicates that the average total gas enthalpy
can be determined fairly accurately. 15 However, there
are some important considerations. First, the larger
the facility, the more difficult it is to accurately mea-
sure the coolant temperature rise. Either a large num-
ber of measurements must be made in the smaller
coolant lines or the temperature distribution in a large
manifold must be resolved to determine the coolant
temperature rise. Second, the energy balance does
not account for further heat losses beyond the nozzle
that may reduce the bulk enthalpy value of the free
stream. Finally, although knowledge of the enthalpy
determined from an energy balance is important and
useful from a facility perspective, it is still an aver-
age, or bulk value. This average enthalpy value may
not be representative of that part of the test stream
actually impinging on the test article since gradients
in flow enthalpy that may develop in the arc column
persist owing to short residence times in the high pres-
sure region of the nozzle. Perhaps more importantly,
the energy balance approach provides no information
about the degree of nonequilibrium or how the energy
is apportioned in the free stream.
Fig. 7. Energy balance on a large-scale, constricted- arc
arcjet test facility.
Energy balance measurements can also be used to de-
termine heater efficiency values during facility opera-
tion. The heater efficiency, which is generally a func-
tion of arc pressure and current, is defined as
*Ih(p,I) = rh h wg /{VI). (3)
Once this is determined for the particular heater con-
figuration, it can be used to quickly estimate the bulk
enthalpy using the mass flow rate of the gas and the
arc voltage and current by simply rearranging the
equation. Because the efficiency is a function of the
arc current and the stagnation pressure, this mea-
surement must be carried out over the full range of
facility operation to develop an empirical correlation
that accounts for the dependence. 16 It is important
to understand that changes in electrode configuration,
or indeed, variation in electrodes themselves will di-
rectly influence the heater efficiency. Moreover, the
electrodes are typically the most frequently replaced
component of the facility, so efficiency values, and this
approach to estimating bulk enthalpy, should be used
with caution.
Electrode Package
Fig. 8. Sonic flow method for determining enthalpy in
an arcjet facility (after Winovich 17 ).
Sonic Flow - Another method used to determine the
total enthalpy is the sonic- flow method that was de-
veloped by Winovich. 17 The basic physical principle of
this method is that for any given equilibrium thermo-
dynamic state there is a unique value of the sonic mass
flow. Thus for a given enthalpy and pressure there is
3A-7
only one value of the choked mass flow. Conversely, for
a known pressure a measurement of the mass flow de-
termines the enthalpy. The graphical representation
of this approach is shown in Fig. 8. Assuming that
the flow is one-dimensional and in equilibrium, then
for both real and ideal gases a simple expression re-
lating mass flow and reservoir enthalpy can be derived
from the equations governing the flow from a reservoir
through a choked nozzle,
\/2l^ |> A v\ 1/2 '
(R T 0 ) [p 0 t h 0 J
In the above expression, A is the cross-sectional area,
the subscript 0 refers to stagnation conditions, and the
superscript * refers to conditions at the throat.
Simplified versions of this equation can be derived for
the case of thermally and calorically perfect gases, as
well as for calorically imperfect gases. For real gases
both 7 and R vary with temperature and pressure and
there are no simple closed form expressions that repre-
sent this variation. Consequently, the governing equa-
tions for the choked nozzle flow were solved iteratively
using an equation of state representing a dissociating
gas for a range of pressures and enthalpies. For the
range of pressures investigated ( 0.25 to 100 atm) all
solutions for the mass flow fell within 4 % of a mean
curve. A curve fit procedure was then used to develop
the empirical correlation,
777 _ C
(A p T ) h £ 97
( 5 )
where C is a constant factor whose value depends on
the system of units. The effects of boundary layer,
nonequilibrium (or frozen) chemistry, and variable
heat loss to the nozzle walls were examined in the orig-
inal work pertaining to this measurement approach. 17
While boundary layer and heat loss effects appear
to be small, the existence of nonequilibrium flow at
the throat leads to a systematically low estimate of
the flow enthalpy. As with the energy balance ap-
proach, the total enthalpy determined with the sonic-
flow method represents an average value, and there
is no information about the degree of nonequilibrium
where testing takes place beyond the nozzle exit.
presented an experimental investigation of Goulard’s
theoretical results for arcjet flows. During the same
time period, empirical correlations for stagnation point
heat transfer in any gas were published: 21,22
( 6 )
where k is a gas species dependent constant, R e jj
is the effective radius of the blunt-body article, and
Ah is the difference between the stream and cold
wall enthalpy. A clear advantage of this approach
is that it gives a spatially resolved measure of the
stream enthalpy at the test location. However, an im-
portant assumption in the use of the above correla-
tion is that the cat aly city of the surface of the heat
flux gauge is essentially full, i.e. all atoms impinging
on the surface recombine and deposit the excess en-
ergy from the exothermic reaction on the surface as
heat. It should be noted that full catalycity is rarely
achieved for calorimeters, and heat flux measure-
ments with gauges of different catalycity show wide
variation. 23,24 Oxidized, uncleaned surfaces, which are
typical on calorimeters that are in service, have sig-
nificantly lower catalycity. This means that calorime-
ters will generally under-measure the incident heat flux
when significant dissociated species are present at the
calorimeter surface. Since the inferred enthalpy is lin-
early dependent on the measured heat flux, this ap-
proach will lead to a lower estimate of the stream en-
thalpy level.
Fig. 9 . Stagnation point heat transfer measurements.
Stagnation Point Heat Flux - With certain as-
sumptions the total stream enthalpy can be inferred
from a simultaneous measurement of heat transfer and
impact pressure at the stagnation point of a blunt
body, such as a sphere or cylinder as depicted in Fig. 9 .
Boundary layer equations for stagnation point heat
transfer were developed by Fay and Riddell, 18 and
a subsequent modification of these results to include
nonequilibrium boundary layer chemistry and surface
catalytic effects was given by Goulard. 19 Later, Pope 20
All of the conventional approaches to arcjet stream
characterization that have been discussed in this sec-
tion share common attributes in that they infer en-
thalpy from other flow property measurements and
they provide no information on the degree of non equi-
librium. The ability of measurements made using
these approaches to guide and inform flow modeling
is therefore limited. Furthermore, it is not possible
to use these measurements to relate the arcjet stream
conditions to the intended flight application, because
3A-8
they do not address the nonequilibrium state of the
free stream.
New spectroscopic techniques that are currently under
development may improve this situation, and some of
these approaches will be discussed in the second lec-
ture. Judicious application of CFD codes may also
improve this situation by providing more insight into
the thermochemical state of the flow. However, any
CFD codes that are developed for this purpose must
be guided by experimental results, and this topic is
addressed in the following sections.
2. CFD for Arcjet Flows
Today, CFD is an important resource for aerospace
vehicle design, testing, and development. Investiga-
tions into new, or poorly understood, flow problems
are often undertaken with a combined experimental
and computational approach. Both the experiment
and the modeling benefit from the collaboration, since
the CFD simulations can evaluate a wide parameter
space quickly and efficiently, while the experimental re-
sults provide guidance for developing assumptions and
improving model fidelity. The general state of CFD for
a particular discipline in the wider field of aerospace
applications is periodically reviewed. Recent reviews
that are relevant to simulating arcjet facility flows can
be found in Refs. 25 and 26, which examine CFD for
high enthalpy test facilities and external flows, respec-
tively.
The present discussion is concerned mainly with the
impact of experimental measurements and instrumen-
tation on modeling, so detailed examinations of numer-
ical methods, particular models, and grid resolution,
which are familiar topics in the literature regarding
CFD, will not be covered. Rather, the intent is to
discuss shortcomings in current instrumentation and
available experimental data that make the task of pro-
ducing credible arcjet flow simulations exceedingly dif-
ficult, if not impossible. Although the conservation
equations and general numerical method are discussed
briefly below, they are invoked only to frame the dis-
cussion about what must be measured and how well.
The perspective is that of an experimental approach
to flow modeling that examines assumptions, model
inputs, and constraints in order to propose better ex-
perimental tests to resolve ambiguities and uncertain-
ties.
2.1 Motivation for Arcjet Flow Modeling
A major driving force behind arcjet flow modeling is
the desire to extract the most information from tests
of thermal protection systems in large-scale arcjet fa-
cilities. Testing costs are always a concern, and an
investment in computational resources to avoid test
article failures or to conduct a more efficient test cycle
represents a prudent strategy. Computational investi-
gations can often be undertaken at lower expense than
experimental efforts. Unless a complete computational
capability is being started from scratch, the costs of
employing state-of-the-art instrumentation for experi-
mental investigations is usually much higher, assuming
that manpower for both efforts is equivalent. If more
and better information could be obtained from arc-
jet testing, then substantial development cost savings
may be realized from a reduced dependence on flight-
experiments (e.g. FIRE 27,28 and Apollo 29 ) that are
often required to establish thermal protection system
effectiveness.
Minimizing thermal protection mass for current, low-
budget planetary missions is also an important motiva-
tion for generally improving the state of knowledge of
arcjet flow stream conditions. For these missions there
is neither time nor budget for flight testing a proto-
type before launching. If results from arcjet tests can
be extrapolated to flight conditions with quantifiable
uncertainties, then it may be possible to reduce the de-
sign safety margins that currently added to heat-shield
thickness. 30 It may ultimately be possible to estab-
lish flight performance of thermal protection materials
through arcjet testing if a sufficient understanding of
arcjet flows is developed. CFD modeling would play
an indispensable and enabling role in this effort.
Facility improvements and optimization for particular
test configurations could also benefit from the develop-
ment of CFD tools tailored to arcjet flow modeling. As
noted above, ARCFLO played an important role in the
early development of large-scale arcjet test facilities,
and there is a need for modern computational tools to
improve electrode designs, optimize heater configura-
tions, and design new nozzles for flat- plate test con-
figurations. These tools could also be used to design
test configurations that would provide the necessary
information at reduced cost and effort.
Additional motivation derives from the desire to im-
prove the general state of nonequilibrium flow mod-
eling and the understanding of real gas effects. As
mentioned above, the stable, relatively long-duration
arcjet operation at high enthalpies creates opportu-
nities for studying complex chemical and thermal in-
teractions that cannot be easily analyzed in impulse
facilities.
2.2 CFD Requirements for Arcjet Flow Simu-
lations
Before discussing measurement requirements for im-
proving computational simulations of arcjet facility
flows, it is useful to examine the CFD requirements
that have evolved from previous and ongoing efforts to
model arc-heated flows in large-scale facilities. Arcjet
flows are typically not in thermal and chemical equilib-
rium, except possibly in the constrictor and electrode
3A-9
package regions. Consequently, any attempt to model
the flow requires a CFD code that models nonequilib-
rium processes.
The conservation equations for hypersonic flows in
thermal and chemical nonequilibrium that are solved
by the LAURA CFD code 31 have been compiled
in a single reference publication by Gnoffo and his
co workers. 3 2 Eleven species that are typically encoun-
tered in simulations of hypersonic air flows were in-
cluded in the model: N 2 , O 2 , NO, N, O, Nj, O^", NO + ,
N + , 0 + , and e“ . Thus, eleven species continuity equa-
tions and three momentum equations must be solved
by the code. For this particular CFD code, three
separate energy equations are modeled to account for
nonequilibrium effects: vibrational energy, electronic
energy, and total energy. Thermodynamic data for
the eleven species and reaction rates for two different
models, Park 33 and Dunn and Kang 34 , were also given
in the report. In this CFD approach, which is repre-
sentative of those currently in use for nonequilibrium,
hypersonic flows, only the ground electronic states of
each species are modeled. When radiative energy flux
is important, it is typically treated separately or in a
loosely-coupled fashion. It should be emphasized that
there is no universally agreed upon model formulation,
particularly when it comes to nonequilibrium processes
and chemical reaction and energy transfer rates. In-
terested readers are referred to Refs. 35-38 for other
computational model formulations.
For arcjet flows in large scale facilities, argon must
also be considered since it is often added to the test
gas flow to protect electrode surfaces. If only the neu-
tral state is considered, this brings the total number
of species for air/argon flows to twelve. In addition,
thermodynamic and chemical reaction rate data must
also be included for argon. 39
2.3 Strategies for Arcjet Flow Simulations
Just as there is no universally accepted model for
nonequilibrium, hypersonic flows, there is no single
CFD code that can simulate the complete arcjet fa-
cility flow from the heater to the test article. Thus,
some reasonable modeling strategy must be developed
that matches available CFD models to flow regions in
an advantageous manner. To illustrate this point, sev-
eral modeling efforts that were concerned with either
arcjet characterization or interpretation of arcjet test
results are surveyed below. The presentation is or-
ganized by flow region, starting from the heater and
moving through the nozzle to the test article.
Arc Heater - The flows within the arc heater and
electrode packages are special cases, since the elec-
trodynamic processes occurring within these typically
subsonic flow regions are usually absent in hypersonic
flows (with the exception of MED device flows). How-
ever, a discussion of numerical studies of arc heaters
is included here for two reasons: first, the flow may be
in thermal and chemical equilibrium within the down-
stream electrode package; and second, if the flow and
discharge physics can be modeled correctly, then it
may be possible to compute inlet conditions for use
in nozzle calculations.
Within an arc- heater, the flow is typically subsonic
and is more properly described as a plasma owing to
the presence of the electrical discharge. To model this
portion of the flow accurately, a coupled solution of
the fluid dynamics, radiation, and electrodynamics is
required. The development of a CFD model for the
arc heater that included the necessary coupling was
undertaken at Ames Research Center several years
ago, 40,41 but the effort was eventually abandoned. In-
stead, the flow within constricted-arc heaters is still
modeled with either the ARCFLO code, which was
mentioned above, or a derivative. One of the deriva-
tive codes, SWIRL ARC, 42,43 has been modified to in-
clude the tangential component of gas injection that is
typically used to help stabilize the discharge in high-
pressure facilities. It should be noted that in any form,
ARCFLO does not attempt to fully simulate the physi-
cal processes within the heater. Rather, ARCFLO and
its derivatives employ a semi-empirical approach to
perform comparative studies and indicate trends that
might be useful for design studies.
Recently, there has been renewed interest in improv-
ing computational models of constrictors. A Navier-
Stokes formulation for a constrictor was developed and
implemented by Kim et al, 44 and an improved, fully-
coupled radiation model was applied to the study of
an arc heater by Sakai et al. 45 The main motivation
for this renewed activity is the need to increase the ef-
ficiency and performance capabilities of existing arcjet
facilities. Obviously, measurements will be required to
validate these newer flow models.
Nozzle and Free Stream - For studies relating to
arcjet characterization, arcjet test interpretation, or
general nonequilibrium flow modeling, the expanding
flow in an arcjet facility nozzle presents a challenge to
the computational community. The general strategy
for modeling arcjet nozzle flows relies on some means
for estimating the inlet conditions for the nozzle, par-
ticularly the stagnation enthalpy, and then using what-
ever experimental information is available from the
free stream to assess the fidelity of the simulation. De-
pending on the particular computational model, the
inlet conditions can be specified either in the subsonic
flow region upstream of the throat or in the super-
sonic portion of the nozzle. As was mentioned in the
description of a typical arcjet facility, the nozzle flows
are not in thermal or chemical equilibrium. Therefore,
the computational approach must model the thermo-
dynamic and chemical kinetic processes that govern
hypersonic, nonequilibrium flows.
3 A- 10
There have been several efforts aimed at simulating
flows in arcjet nozzles. At Ames Research Center
alone, three different numerical approaches have been
used recently to simulate nozzle flows in conical 46-48
and semi-elliptic 49 geometries. These particular stud-
ies were undertaken specifically to address arcjet char-
acterization issues. Additional investigations of con-
ical nozzle flows have been carried out in support of
arcjet surface catalysis experiments. 50 Details of the
different numerical approaches are given in each, of the
references. However, it is interesting to note the pro-
gression of the numerical models used in these stud-
ies. Babikian used a quasi- one- dimensional, multi-
temperature flow model, NOZNT, 51 to compare with
free stream temperature measurements in the Ames
AHF Arcjet Facility. 46 Gokgen performed simulations
of the nozzle flow with an axisymmetric, nonequilib-
rium Navier-Stokes solver in support of shock layer
experiments. 4 ' More recently, Loomis and his cowork-
ers used GASP, which is a general three-dimensional,
flow solver to simulate both conical and semi-elliptic
nozzle flows in support of thermal protection material
tests for the X-33. 49
Concurrent experimental and computational studies of
expanding, No/Ar plasma flows have also been car-
ried out by Schonemann and co workers. 52 The note-
worthy aspect of this particular investigation was the
use of experimental measurements at one axial loca-
tion to start the calculations and predict the rapidly
expanding flow properties at a second, downstream lo-
cation. This approach has the advantage of avoiding
some of the uncertainties that result from estimating
inflow conditions.
Flow Over a Test Article - As the current use of
large-scale arcjet facilities is aimed primarily at simu-
lating aerothermal heating, it is extremely important
to be able to model the flow over a test article ac-
curately. Test article flows can be classified into two
basic types: shock-layer flows over a test article in
a conical nozzle flow and bound ary- layer flow over a
flat plate for semi-elliptic, or rectangular, nozzle flow.
Since the flat plate is usually an extension of the nozzle
wall, the modeling requirements for simulating bound-
ary layer flows are identical to those for nozzle flows,
although the angle of attack is typically varied as part
of an experimental investigation. Shock-layer flows are
different, particularly for studies of stagnation point
heating. For this configuration, the flow undergoes
compression by a shock wave, whose strength depends
on the particular test conditions and geometry, before
impinging on the test- article surface. Thus, the free
stream conditions, which largely determine the char-
acteristics of the shock layer flow, must somehow be
known to carry out the simulation.
Typically, there are no stream measurements, other
than pitot pressure and heat flux, that could be used
to specify the stream conditions. For certain arcjet test
conditions, it is possible to estimate the stream condi-
tions using a combined equilibrium and frozen-flow an-
alytical approach, and then carry the analysis through
the shock layer based on measurements of the pitot
pressure, heat flux, and model surface temperature. 50
However, a more general approach involves simulating
the nozzle flow (again, an estimate of the initial en-
thalpy is required) with a numerical model and then
using those conditions as input to a shock layer calcula-
tion. An example of this latter approach can be found
in the work of Gokgen, 4 '’ 48 which will be discussed in
detail below. Inevitably, inaccuracies in modeling the
expanding nozzle flow affect the simulations of shock-
layer and boundary-layer flows in arcjet facilities.
The response of the test article to the shock layer flow
is also of considerable interest to the arcjet test and
material development communities, where much can
be gained by understanding the interaction between
the shock layer flow and the material. A review arti-
cle by Milos and Rasky 53 outlines the importance of
properly defining the boundary conditions that gov-
ern the interactions at the fluid/surface interface. The
authors also point out that since boundary processes
define the interaction of the fluid and solid computa-
tional models, their boundary conditions must agree.
This issue is especially important for understanding
the performance of charring and ablating thermal pro-
tection materials. Although it is very interesting, this
topic is outside the scope of the present discussion.
Measurement Requirements fur Arcjet Flow Modeling
Arcjet Flow Measurements j
Model Development Measurements
Starting Conditions
Simulation Validation
Three-body recombination rales
Third- body efficiencies
Spontaneous emission rates
Laser-excitation rates
Col lisional-radi alive model rates
Species thermodynamic data
Energy Transport Rates
Enthalpy
Pressure
Mass flows of
test gases
Inflow velocity
Contaminant
level
Turbulent or
laminar?
Gradients
Assessment of
equilibrium
Velocity
Species concentrations
Density
Pressure
Translational T
Rotational T
Vibrational T
Electronic state
populations
Post <no// le expansion
rate
Stream profiles
Fig. 10. Classification of measurements for arcjet flow
modeling.
3. Measurement Requirements for Arcjet Flow
Modeling
Measurements that can be used to improve compu-
tational models of arcjet flows can be separated into
general categories, which are illustrated in Fig. 10.
The first classification distinguishes between direct
measurements of properties of arcjet flows and more
generic measurements that can influence the develop-
ment of models for nonequilibrium, hypersonic flows.
Measurements of thermodynamic properties, species
concentrations, velocity, and enthalpy in arcjet flows
3 A- 11
would all fall into the first category of direct measure-
ments. More accurate determinations of important
reaction or energy transfer rates, which need not be
measured in arcjet flows, would fall into the second
category. While this category is probably of equal im-
portance in the improvement of arcjet flow modeling,
the majority of the discussion below is concerned with
direct measurements of primary arcjet flow quantities.
Within the first category of direct arcjet flow prop-
erty measurements, a further distinction can be made
between measurements that would be used to define
starting, or inflow-boundary, conditions and measure-
ments that could be used to assess the fidelity of the
simulation. Since the success of any flow modeling ef-
fort is inextricably linked to the accuracy with which
the starting conditions for the calculation are defined,
measurements of the input parameters are considered
to be of greater importance. Of the inflow parame-
ters for arcjet flows, the total enthalpy is the most,
important because it defines the total flow energy and
the initial composition and temperature. Despite its
importance, enthalpy has proven to be the most diffi-
cult parameter to characterize accurately, as discussed
above. Typically, the settling chamber pressure is
measured to reasonable precision for most arcjet tests,
so it is assumed herein that pressure is given. Other
primary measurements that define the starting con-
ditions are the mass flows of the test gases and the
configuration aiid geometry of the facility.
Velocity, species concentrations, temperature(s), static
pressure, and density are examples of flow property
measurements that can be made at various locations
in the arcjet to assess the performance of a computa-
tional model. Flow quantities that are derived from
measurements of primary flow variables, such as the
dynamic pressure, specific heat ratio, Mach number,
and Reynolds number, are less important from the
perspective of evaluating computational models. How-
ever, these quantities are quite useful in specifying the
performance of the arcjet facility and for relating the
test conditions to the expected flight environment.
3.1 Enthalpy
It is readily apparent from even a casual reading of
the previous section on modeling requirements that all
simulations of the most important arcjet regions, the
nozzle and shock- layer, or boundary-layer, flows, re-
quire knowledge of the stagnation enthalpy. The state
of enthalpy determination using conventional instru-
mentation was examined in the introductory section,
and it was found to be inadequate for several reasons.
First, the conventional means for determining the flow
enthalpy can only give an estimate of the total value,
which does not specify the state of the essentially
frozen free-stream flow. Second, for the energy bal-
ance and sonic flow approaches, only the bulk enthalpy
value can be determined. While this is useful for mon-
itoring facility performance, the enthalpy value proba-
bly does not represent the free-stream core flow, where
stagnation- point tests are conducted, unless there are
no span wise enthalpy gradients. The assumption of
gradient-free flow appears to be questionable . 25 Third,
even when great care is taken with the treatment of
the calorimeter surface, enthalpy values derived from
heat flux measurements are likely to be systematically
low.
It should be noted that the measurements required
for determining flow enthalpy vary with flow region.
Moreover, the influence of the enthalpy determination
on the outcome of the flow simulations also depends
on where the enthalpy measurement is made. For arc
heater and nozzle flow simulations, a measurement of
the stream enthalpy within the downstream electrode
package, which also functions as a nominal settling
chamber, is appropriate. However, for shock layer sim-
ulations, inaccuracies are accumulated from simulating
both the nozzle flow and the shock layer flow. A more
appropriate enthalpy measurement location would be
the free stream, provided that the measurement could
quantify both the total enthalpy and the nonequilib-
rium state of the gas. With this information about
the free stream, the shock layer flow could be sim-
ulated independently of the nozzle flow. Obviously,
owing to the nonequilibrium nature of the flow, more
flow property measurements are required to determine
the thermodynamic and chemical state of the flow in
the free stream.
Finally, any enthalpy measurement must be spatially
resolved, and enthalpy gradients must be quantified to
remove potential ambiguity from the specification of
the starting conditions. This issue will be discussed
further below and the applicability of nonintrusive op-
tical diagnostics to enthalpy measurements will be ad-
dressed fully in the following lecture.
3.2 Arc Heater
The flow within the arc heater and electrode package is
usually subsonic and the enthalpy is mostly static, be-
ing comprised of thermal and chemical mode contribu-
tions. Because pressure is reliably known, a measure of
total density or temperature would permit a determi-
nation of the total enthalpy. Of the two variables, tem-
perature is more amenable to measurement through
optical means. Assuming that the flow is in thermal
equilibrium, then determination of a single tempera-
ture is sufficient for determining enthalpy. If there is
optical access to either the heater or electrode pack-
age region, then a spectrally resolved emission mea-
surement can be used to determine temperature. The
specific procedures for determining temperatures from
spectrally resolved emission are discussed in the fol-
lowing lecture.
3A-12
Useful information could also be derived from addi-
tional measurements of other flow variables in the elec-
trode package. These other flow properties include:
the axial flow velocity; the azimuthal velocity compo-
nent, which would quantify the amount of swirl at the
nozzle inlet; the total heat flux and radiative heat flux
to the wall; the amount of copper, which is introduced
into the stream by the process of electrode erosion; 54
and the electron number density downstream of the
arc. Although the axial extent of the electrode pack-
age region is usually not that large, the flow is cooling
as it moves toward the throat. Consequently, a de-
termination of the axial variation in any flow quantity
would provide some insight into the evolution of the
flow as it begins to accelerate.
Owing to limited accessibility, flow probes are not a vi-
able option. Their survival at typical large-scale arcjet
facility operating conditions is also an issue. Optical
access to the downstream electrode package can of-
ten be realized, 55 and measurements in this region are
particularly useful because this region provides the in-
flow to the nozzle. It may be possible to implement
optical measurements at two different axial locations
downstream of the arc termination to assess the rate of
evolution of the stream properties. Furthermore, since
large fluctuations in the magnitude of emission from
atomic transitions have been observed in the electrode
package, 55 it may be possible to develop a two point
correlation approach for velocity measurement.
3.3 Arcjet Nozzle and Free- St ream Flow
Nozzle - For the purposes of this discussion, the start-
ing point for nozzle flow is defined as the end of the
electrode package. Unlike the segmented arc heaters
and electrode packages of todays constricted-arc facili-
ties, the nozzle assemblies are typically fabricated in a
more monolithic manner. Because they are fabricated
with integral water cooling, there is little hope for in-
strumenting existing large-scale arcjet nozzles. This
essentially precludes in situ monitoring of the onset of
chemical and thermal freezing, which could then be
used as a starting point for frozen flow analysis.
Using smaller scale arcjet devices fabricated with seg-
mented nozzles it may be possible to address the on-
set of chemical, and possibly thermal, freezing for flow
conditions of interest in aerothermal testing applica-
tions. Note that the fluid dynamic expansion rate
plays an important role in determining the location
at which the flow freezes. Whatever studies are un-
dertaken in smaller facilities must address this issue.
Free Stream - Although the flow is usually chem-
ically and thermally frozen by the time it exits the
nozzle, the free stream region is often optically ac-
cessible, and measurements of many flow properties
are possible. Spatially resolved measurements of ve-
locity, translational temperature, density, pressure,
and species concentrations have ail been made using
laser-induced fluorescence (LIF) techniques. 56 ” 59 Re-
cently, measurements of enthalpy and its distribution
among thermal, chemical, and kinetic modes, were
demonstrated in N 2 /A 1’ 58 and air/Ar 59 flows using
two-photon LIF of atomic nitrogen. Although more
property measurements are required to determine en-
thalpy for nonequilibrium flow, the approach of using
LIF of the dissociated species to determine multiple
flow parameters appears capable of providing this in-
formation with the aid of certain assumptions. For-
tunately, the validity of the assumptions that are cur-
rently invoked can be evaluated experimentally. 59 Flow
property measurements using LIF techniques will be
discussed extensively in the second lecture.
Although further development of this approach is re-
quired to assess the assumptions and establish the
range of applicability, LIF based stream property mea-
surements may ultimately prove sufficient to establish
the enthalpy and degree of nonequi librium of the free
stream flow. This would provide a set of inflow con-
ditions that could be used to calculate the shock-layer
flow. A computational simulation of the flow over a
test article that was started from known free stream
conditions and compared to shock-layer property mea-
surements would allow a better assessment of the com-
putational modeling. Determinations of free stream
rotational and vibrational temperatures and assess-
ments of possible metastable atomic state populations
are needed to establish the validity of the LIF-based
approach. 59
In addition to establishing inflow conditions for shock
layer simulations, the two- photon LIF measurements
provide stream property information that can be used
to evaluate the fidelity of nozzle flow simulations.
Since total enthalpy is specified by the LIF measure-
ments, with a quantified uncertainty, that, value can be
used along with the constrictor pressure to start the
nozzle simulation. If the model used in the nozzle flow
simulation is accurate, it should reproduce the mea-
sured distribution of the total enthalpy into kinetic,
thermal, and chemical contributions in the nonequi-
librium free stream. Comparisons between nozzle sim-
ulations and free stream measurements are underway
for the chemically simpler N 2 /argon flow cases.
3.4 Blunt- Body Shock- Layer Flow
Even with the free stream conditions specified, much
is required in order to improve the general understand-
ing of shock- layer flows in the stagnation- point heating
configuration for aerothermal test applications. Finite
rate effects that vary in significance depending on the
test conditions and model geometry still control the
chemical and thermal state of the shock layer and im-
pact issues such as the difference in catalytic heating
between the arcjet test conditions and the flight envi-
ronment. Moreover, depending on the test geometry
3A-13
and conditions, merged shock layer and rarified flow
effects may be important, and may complicate the in-
terpretation of heat transfer data..
Instrumentation and techniques for making spatially
resolved flow property measurements are therefore re-
quired to establish the thermochemical state of the
gas in the shock layer as it moves toward the surface
of a test article. Spatial resolution is important be-
cause the flow is generally evolving from a nonequilib-
rium state toward an equilibrium state as it approaches
the surface. Understanding this evolution is important
from a modeling perspective as well as for aiding in the
interpretation of test results. Again, for nonequilib-
rium situations, multiple flow properties, including ve-
locity, species concentrations and temperatures must
be measured to specify the flow state. In contrast to
free steam conditions, shock layer temperatures (T r ,
T„, and T e ) can reach levels in excess of 5000 K and
pressures can be orders of magnitude higher. Many
internal energy levels of a number of species will be
populated, and the distributions of populations over
these energy levels may differ for different species (and
possibly electronic states). With an ablating material,
the situation is even more complex.
However, the goal of understanding the shock layer
thermochemistry is important because that is the envi-
ronment that must ultimately be related to flight con-
ditions. In addition, if instrumental approaches that
determine the thermochemical state of the shock-layer
can be developed, then it may be feasible to test all-
body vehicle configurations in the long-duration, arcjet
flow facilities.
3.5 Measurement Accuracy Requirements
The uncertainty in experimental measurements and in
computational predictions is an important consider-
ation in arcjet flow investigations. For the present
discussion, only uncertainties in experimental mea-
surements are considered. Experimental uncertainties
are estimates of errors in measurements that typically
arise from either systematic or random contributions,
or more typically, both. The systematic and random
contributions are manifestations of the more general
measurement attributes: accuracy and precision. Def-
initions of measurement accuracy and precision, which
are frequently confused, can be found in a variety of
reference publications, including a text on the sub-
ject by Bevington. 60 In the introductory discussion,
Bevington indicates that the accuracy relates to how
close a measured value is to the “true” value, while
precision provides information on how well something
can be measured, regardless of what that measurement
means. For arcjet flow property measurements that
would be used to improve flow modeling capabilities,
both accuracy and precision are important. However,
because flow properties are generally unknown, instru-
mentation must be developed to make measurements
with a minimal reliance on assumptions that cannot
be tested; otherwise it is impossible to assess accu-
racy. How closely a measured flow property represents
the real situation is important from the perspective
of predicting absolute flow property magnitudes, as in
the simulation of a single arcjet test.
In contrast, it is often easier to establish the precision
of a particular measurement from a number of different
observations at similar conditions. This can be done
without evaluating all of the assumptions that may go
into a particular measurement, and it may then be pos-
sible to use the measured quantity constructively with-
out knowing the absolute accuracy. Once the precision
is established, then the measurement could be made
for a number of different flow conditions where a single
control parameter, such as the arc current, is varied.
Simulations of these different flow conditions and com-
parisons with the experimental results would test the
ability of the model to predict trends and might ulti-
mately do more to establish confidence in the model-
ing than single-condition comparisons. Obviously, the
model would have to be optimized for an appropriately
chosen test condition. This approach may prove to be
more effective in advancing both modeling and instru-
mentation development, especially when one considers
the number of measurements required to document a
single test case completely. As the instrumental tech-
niques mature and as more is learned from parametric
comparisons, then it may be feasible to pursue a single,
well-documented test case.
From an experimental perspective, given the scarcity
of data and the fact that measurement results from one
facility cannot be directly transferred to another unless
the configuration is identical, any property measure-
ments that also have quantified uncertainty estimates
are useful in advancing the general state of knowledge.
Enthalpy measurements, as well as other inflow con-
ditions that are required to initiate simulations, are
examples of this type of flow property. Measurements
that are used to evaluate the validity of computational
models must be held to a higher standard, since their
determination may influence changes in the model for-
mulation. It is difficult to formulate a general state-
ment as to how high the standard should be, given the
complexity of present day CFD models. For some pa-
rameters an accuracy requirement can be postulated
easily. As an example, consider LIF-measured atomic
nitrogen concentrations, which currently have an esti-
mated uncertainty of ^ 30 % . 59 Although this uncer-
tainty appears to be large, the recombination rate for
the reaction,
A + A + M->A 2 + M ,
which largely determines the N atom concentration in
the chemically frozen free stream, is currently uncer-
tain by up to a factor of three. 62 Clearly even the
3A-14
relatively uncertain N atom concentration measure-
ment can be used to evaluate flow model performance.
To determine accuracy requirements for other flow
properties, an effort should be made to evaluate un-
certainties in quantities currently used in the models
and parametric studies with the computational model
should be performed to evaluate sensitivities.
4. Experimental and Computational Investiga-
tion of Shock-Layer Flows
Recent attempts to simulate arcjet flows and compare
the numerical predictions with experimental measure-
ments illustrate the current status of both simulations
and measurements. Knowledge gaps that affect the
comparisons between simulations and measurements
are readily apparent. The combined experimental and
computational investigation of blunt- body, shock- layer
flows in the Ames AHF Arcjet Facility was chosen for
this purpose because the investigation was motivated
by the need for improved characterization of arcjet
flows, and understanding the shock- layer flow is di-
rectly relevant to aerothermal testing of thermal pro-
tection materials. Although experimental investiga-
tions have also been undertaken in the electrode pack-
age and free stream regions of the same arcjet facility,
comparisons between measurements and simulations
for those studies are ongoing. Consequently, more can
be learned from examining the process and the results
of the documented shock-layer flow property compar-
isons, and from the results of those comparisons.
4.1 Objectives of Investigation The objectives of
this investigation were to: 1) determine whether a re-
gion of thermal and chemical equilibrium exists in the
shock layer formed over a flat- faced cylinder; 2) de-
termine the conditions required to establish the equi-
librium region; and 3) determine whether or not en-
thalpy measurements could be derived from spectrally
resolved emission emanating from the equilibrium re-
gion.
Early investigations of arcjet facility flows included ef-
forts to characterize the shock layer flow using emis-
sion spectroscopy. 62,63 If the flow is in thermal and
chemical equilibrium, then a measurement of the tem-
perature from spectrally resolved emission and a con-
current pressure measurement would uniquely specify
the thermo chemical state of the flow and its enthalpy.
A relatively recent analysis of shock- layer emission ap-
peared to indicate the presence of an equilibrated re-
gion within the shock layer at a lower pressure than
had been expected. 46 Therefore, a major goal of the
present investigation was to verify the existence of the
equilibrated region, and begin the task of defining the
conditions that produce the equilibrium flow. By un-
dertaking this investigation, issues associated with the
development of an “enthalpy meter” based on mea-
surements of shock layer emission could also be as-
sessed.
At the outset, several areas of uncertainty were iden-
tified that had to be addressed in the investigation.
These areas included: 1) what criteria are used to
identify a region of thermal and chemical equilibrium;
2) how to interpret emission measurements with cer-
tainty; and 3) the effects of spatial intensity gradients
on measurements that are integrated along the line-of-
sight. It quickly became apparent that computational-
flow modeling could address some of these issues and
help guide the experimental investigation. Conversely,
it was realized that the experimental measurements
might also provide some assessment of the computa-
tional model validity, but this was not the primary
objective.
4.2 Experimental and Computational Approach
Experiment - The shock layer emission experiments
were carried out in the Ames AHF Arcjet Facility,
which was previously described in the introductory
section. To generate the highest shock layer pressure
values the facility was configured with the 30.5 cm di-
ameter nozzle, which produces the least free stream
expansion. A 15.2 cm diameter, flat- faced cylinder
made of copper was placed in the stream to generate
the shock layer. Two different test conditions were
surveyed, and these are referred to as the high pres-
sure and low pressure cases. Test conditions for the
two cases are summarized in Table 1.
Table 1. Arcjet test conditions for shock layer inves-
tigation.
Test Conditions
Case
Pressure
Current.
Voltage
atm
A
V
Low
1.70
1141
2657
High
6.80
2075
5630
To obtain as much information from a single facility
run as possible, line-of-sight emission spectra were ac-
quired from multiple axial locations along the central
stagnation streamline using a spectrograph and CCD
camera. The model was placed at two different axial
locations in the stream, 34.5 and 36.9 cm downstream
of the nozzle exit (forward and back positions, respec-
tively), to allow full coverage of the shock layer emis-
sion with the finite viewing area of the CCD and spec-
trograph system. Thus, two separate facility runs at
the same nominal operating conditions were required
to obtain the full shock layer emission profile for each
test case.
Emission spectra were acquired at several grating po-
sitions covering the UV to near- 1 R wavelength range
and the measured signals were converted to absolute
intensities through calibration with standard spectral
lamps. Each grating position was chosen to measure
3A-15
certain spectral features that could be used to ascer-
tain temperature or species information using spectral
analysis techniques. 64,65 Particular attention was given
to developing methods for determining rotational and
vibrational temperatures using spectral feature ratios
that minimized the influence of uncertainties in the
measurements. This was desirable because agreement
between the measured temperature values was thought
to be a good indicator of the presence of a thermally
equilibrated flow region. Portions of the Nj (1,2),
(0,1), and (0,0) vibrational bands were found to yield
vibrational and rotational temperatures with mini-
mum uncertainty based on spectral simulations. Note
that the temperature values are derived from emis-
sion that is integrated along the line-of-sight, so the
inferred flow properties actually represent intensity-
weighted, spatially averaged values. Further details of
the experimental configuration and the spectral anal-
ysis can be found in Refs. 65 and 66.
Computational Modeling Approach - The ex-
perimental measurements consisted of spectrally re-
solved, absolute intensities from multiple axial loca-
tions within the shock layer during a single facility
run. Consequently, the computational simulation had
to be able to address issues relating to emission, which
required the use of a radiative transport code. In ad-
dition, two different CFD models were required to
predict the shock layer flow. The first model was
used to simulate the nozzle flow to determine the free
stream conditions ahead of the shock layer, which was
then simulated with a second, separate computational
model. Flow properties predicted by the shock layer
model were then used to calculate the radiative trans-
port.
The two flow models that were used for the simulation
were developed by Gokgen. 67,68 Both models use an
a xisym metric formulation, which is appropriate for the
conical nozzle flow and the shock layer flow over a flat-
faced cylinder. Twelve chemical species: N 2 , O 2 , NO,
N, 0, Nj , O*, NO + , N + , 0 + , e” , and Ar; are modeled
for these flows, and three temperatures: translational,
rotational, and vibrational; are used to represent the
thermal state of the gas mixture. The reactions and
rate coefficients that are used in the model are derived
from the multi- temperature model of Park and Lee. 51
Turbulence is not included in either flow model; the
flow is assumed to be laminar throughout the facility.
Further details of the computational models can be
found in Refs. 67 and 68, and more information about
the nozzle and shock layer computations can be found
in Refs. 47 and 48.
To calculate the emissive flux for the shock layer flow,
the NEQAIR 69 radiative transport model was used.
Inputs to the model, which include species concentra-
tions and temperatures, were obtained from the flow
solution by interpolating between calculated quantities
at the known measurement locations. For all of the cal-
culations, the emitting level populations were assumed
to follow Boltzmann distributions, albeit with poten-
tially differing values of rotational and vibrational tem-
peratures. The electronic temperature was assumed to
equal the vibrational temperature.
Starting Conditions for the Calculations The
most important inputs to the CFD model are the start-
ing conditions, and, as discussed above in the intro-
ductory section, the flow enthalpy and its spatial dis-
tribution are not known in the free stream or at the
entrance to the nozzle. Therefore, some means of esti-
mating the starting conditions based on facility mea-
surements was required for the nozzle and shock layer
simulations. For this investigation, measurements of
stagnation point heat flux and shock layer pitot pres-
sure were used to estimate the total stream enthalpy
based on an empirical correlation of the form presented
in Eq. (6). At the time of this work, the argon mass
flow rate had not been recently measured, so it was
assumed to be equal to 5 % of the total mass flow for
the high pressure case. Using the estimates of flow
enthalpy and argon mass flow along with the facility
measurement of the arc heater pressure, the nozzle
flow could be calculated. Based on the low level («
3 ppm) measured in the stream of the 60 MW arcjet
facility, 9 copper was not included in the simulations of
the AHF arcjet nozzle and shock layer.
Upon exiting the nozzle, the flow regions that are
within the local Mach angle near the periphery ex-
pand into the test box at a rate that is different from
that in the nozzle, unless the nozzle exit static pressure
matches the ambient pressure. Owing to this further
expansion, the calculated flow property values at the
exit of the nozzle could not be used directly as the in-
flow conditions for the shock layer simulations. This
effect is well documented for perfect gas flows, 70 but
the rates of expansion for the nonequilibrium arcjet
flows at various operating conditions are not known.
Therefore, the calculations were continued in the axial
direction at the nozzle expansion rate until the cal-
culated dynamic pressure matched the value derived
from the shock layer pitot pressure measurement us-
ing Eq. (1). Although this procedure does not exactly
reproduce the fluid dynamics of the full free stream,
it produced acceptable inflow conditions for the shock
layer simulations without requiring a complete simu-
lation of the flow in the test box.
It should be noted that the enthalpy value that was
ultimately used in the simulations was greater than
the value derived from the stagnation point heat flux
and pressure measurements for both the high and low
pressure test cases. After performing initial compar-
isons between the calculated and measured emission
spectra, the enthalpy was increased for both test cases
to bring the calculated spectra into closer agreement
3A-16
with the measurements. The totai increase for the high
pressure case was limited arbitrarily to 10 %, although
it could have been increased further, as will be seen be-
low. Adjusting the total flow enthalpy estimate to a
higher value is justified if the calorimeter surface has a
low catalycity and the recombining atom flux is signifi-
cant, because correlations such as Eq. (6) apply to fully
catalytic surfaces. While the catalycity of the partic-
ular calorimeter used for these measurements is not
precisely known, the surface of the constantan foil was
known to have a ceramic oxide coating, so the gauge
was considerably less than fully catalytic. Given the
uncertainty in the total flow enthalpy, the poor agree-
ment found in the initial comparisons between calcu-
lated and measured spectra could not be attributed to
the flow model. This clearly illustrates the importance
of quantifying the stream enthalpy, since using the ex-
perimental measurements to guide the adjustment of
the input conditions compromises any assessment of
the flow model.
4.3 High Pressure Case
Calculated Shock- Layer Properties - Using the
inflow conditions and two-model computational ap-
proach the flow properties were calculated for the
shock layer at the high pressure case conditions. Axial
profiles of the flow properties along the central stream-
line where the measurements were made are shown in
Fig. 11 and 12. The axial distributions of pressure and
temperature are shown in Fig. 11 as a function of non-
dimensional ized distance from the test article surface.
Note that the nozzle solution predicts that the flow
is vibration ally frozen ahead of the shock, but the ro-
tational energy is predicted to be in equilibrium with
translation, The rotational and translational temper-
atures rise significantly near the shock and decrease
as the blunt- body surface is approached. A lesser in-
crease is exhibited by the vibrational temperature, and
all three temperatures are seen to converge to a value
that is very close to the equilibrium temperature for
these conditions at about 0.1 R upstream of the sur-
face. Thus, the calculations suggest that there is a
region of thermal equilibrium within the shock layer
at the high pressure conditions.
Axial profiles of species mass fractions are shown in
Fig. 12, also as a function of the normalized distance
upstream of the test article surface. Outside of the
surface boundary layer, at the point where the tem-
perature values converge, the species mass fractions
are quite close to their equilibrium values, which are
indicated on the right-most vertical axis. According to
the simulation, the flow is also very nearly in chemical
equilibrium at these test conditions.
Distance, x/R
Fig. 11. Temperature and pressure profiles within the
shock layer for the high pressure case.
Fig. 12. Species mass fraction profiles within the shock
layer for the high pressure case.
In the description of the experimental effort above, the
use of Nj spectral features to determine vibrational
and rotational temperature values within the shock
layer was noted. For an isolated rovibrational transi-
tion, the measured intensity for thermal equilibrium
conditions can be expressed as
h = ^e- E «' kT , (7)
where I\ is the spectral intensity, K\ represents the
line shape function and the transition strength, n is
the species density, Q is the partition function, L is
the line-of-sight path length, and E u is the total upper
state energy. This expression shows that the inten-
sity has a linear dependence on the emitting species
density and an exponential dependence on tempera-
ture. From the computed axial property profiles of
Figs. 11 and 12, it is apparent that all temperatures
and the Nj concentration are higher near the shock
front. Consequently, the measured N* emission could
3A-17
easily be dominated by contributions from emission at
the shock front, where nonequilibrium effects are more
likely to be present. Temperatures are derived from
intensities with the implicit assumption that the ro-
tational and/or vibrational level populations are each
thermally equilibrated. If the measured intensity is
dominated by emission from nonequilibrium regions,
then this assumption is untenable.
The prediction of significantly higher temperatures
and N* concentrations near the shock front led to
a further analysis of the flow property distributions
along the emission measurement sight lines. Recall
that one of the objectives was to use temperature mea-
surement comparisons to determine whether or not the
flow reached thermochemical equilibrium. This re-
quires that the temperature values derived from the
spectral analysis actually represent the central, core-
flow region, rather than the shock front. To assess this
potential problem in interpreting the spectral data,
computed, flow property profiles along the spanwise
flow direction were extracted from the shock layer so-
lution at selected axial measurement locations.
Distance along the line of sight, m
Fig. 13. Temperature variation along the line-of-sight
for the high pressure case.
Profiles of rotational and vibrational temperatures
along the line-of-sight direction, which is normal to
the flow axis, are shown in Fig. 13 for two of the mea-
surement positions. The axial location of the sightlines
is given on the figure in terms of the distance from the
surface of the test article. Again, this will make it dif-
ficult to draw conclusions about the state of the flow
from temperatures derived from the emission spectra.
Distance along the line of sight, m
Fig. 14. number density along the line-of-sight for
the high pressure case.
Computed Nj number densities are shown in Fig. 14 as
a function of distance along the line-of-sight at three
different axial measurement locations, two of which
correspond to the locations of the temperature profiles
in the previous figure. As with temperatures, there is
a significant increase in N* number density near the
shock front. For the two flow properties that govern
emission from Nj, the shapes of the spanwise pro-
files are predicted to be far from the idealized top hat-
distribution that is assumed to exist in the spectral
analysis.
Comparisons With Measurements - Despite the
indication that strong spatial gradients would com-
promise the derivation of flow properties from the
measured intensities, values of line-of-sight integrated
rotational temperature, vibrational temperature, and
number density values were extracted from the spec-
tral data. To make a meaningful comparison, the cal-
culated flow field emission was averaged in the same
manner as the measured intensity. The approach used
to derive these values from the flow property and emis-
sion calculations is described at length in Ref. 47.
Comparisons of the measured and calculated line-
of-sight (LOS) averaged temperatures are shown in
Fig. 15. Measured and calculated vibrational and ro-
tational temperatures are shown as a function of the
normalized distance from the surface of the test arti-
cle. Despite the predicted strength of the shock- front
region emission, the computed, LOS- averaged temper-
atures still appear to nearly converge near the test ar-
ticle surface. Compared to the measured temperature
values, the computed LOS averaged values appear to
approach convergence faster and to a greater degree.
Although the overall trends appear to be consistent
between the measured and calculated values, the two
sets of LOS-averaged temperatures do not agree. In
view of this disparity and because the measured tem-
perature values do not converge, the existence of an
3A-18
equilibrated flow region could not be ascertained from
the comparison.
Fig. 15. Predicted and measured line-of-sight averaged
temperatures for the high pressure case.
By assuming that the vibrational temperature deter-
mined from the spectral analysis of the measured in-
tensities represented the electronic temperature, val-
ues of LOS-integrated N2+ number density could be
determined. As was done for temperature, an ap-
proach for deriving a comparable quantity from the
calculated flow properties was also developed. 71 Mea-
sured and calculated values of the LOS-integrated N*
number density are compared in Fig. 16. Some of the
disagreement between the two sets of number densi-
ties can be attributed to differences in spatial gradi-
ents along the line-of-sight. Until the spatial gradient
effects are investigated experimentally, it is not pos-
sible to determine whether the difference seen in the
comparison nearer the test- article surface is caused by
spatial averaging from the optical system or by inac-
curate modeling of the dissociative recombination
processes.
Fig. 16. Predicted and measured LOS-integrated
number density for the high pressure case.
Based on the comparisons between the measured and
calculated flow properties above, it was not possible to
determine whether an equilibrated flow region exists
within the shock layer for these conditions. Such a de-
termination clearly requires an approach that resolves
the spatial intensity gradients to extract information
from the relevant flow region in the core of the shock
layer. Once that is done, then the impact of other
assumptions can be examined, and the processes that
lead to equilibration can be investigated. Although
the question of an equilibrated region was not conclu-
sively resolved, it was still possible to address whether
or not emission- based measurements could be used to
determine the thermo chemical state of the flow.
For these investigations, the experimental instrumen-
tation did not actually measure flow properties. In-
stead, flow property information was derived from an
analysis of measured emissive intensities. To address
the issue of using emission to evaluate the flow en-
thalpy and to understand how the experimental ap-
proach and the computational modeling might be im-
proved, comparisons were made between calculated
and measured spectral data at selected measurement
locations.
Fig. 17. Measured and computed emission spectra for
the 240 nm grating position, 20.7 mm upstream of test
article.
Comparisons between the measured and calculated
emission spectra are presented below at several grating
positions for a single axial location, 20.7 mm upstream
of the test article. For the 240 nm grating position, the
comparison is shown in Fig. 17. At this spectral loca-
tion, the emission is mainly from NO 7 and S with
probable contributions from the j3 and e systems. Ow-
ing to the overlap of the emission from the different
electronic states of NO, extracting temperature infor-
mation from this spectral region is not feasible. Except
for emission below 2100 A, the calculated intensity is
less than the measured value.
Fig. 18. Measured and computed emission spectra for
the 310 nm grating position, 20.7 mm upstream of test
article.
A similar comparison of measured and calculated emis-
sion spectra at the 310 nm grating position is shown
in Fig. 18. The off-scale spectral features are Cu atom
3A-19
transitions. Copper is present in the stream because
of electrode erosion, and it is not included in the com-
putational model. Emission from molecular species in
this spectral region is mainly from NO at shorter wave-
lengths, and N 2 (2+) and N 2 (1-) systems at the longer
wavelengths. The strongest emission peaks aside from
those due to Cu emission are from N 2 (2+). Agree-
ment between the calculations and the measurements
is reasonable good in shape, but the overall signal level
from the computational spectra appears to be low. Re-
call that the calculated signal levels are exponentially
dependent on the electronic temperature. If the calcu-
lated electronic temperature, which is nearly 6000 K
at this measurement location, was increased by 200 K,
the N 2 (2+) emission would nearly double.
Fig. 19. Measured and computed emission spectra for
the 345 nm grating position, 20.7 mm upstream of test
article.
For the 345 nm grating position comparison, which is
shown in Fig. 19, the emission is mainly from molec-
ular species: N 2 (2+), N 2 (1-), and CN violet. The
CN in the flow comes from dissociation of CO 2 that
is present naturally in air and the subsequent recom-
bination of C and N. Although it is truly a minor
species, the transition strength is large and it is a sig-
nificant emitter, as seen in the region near 3550 A
where several of the measured peaks are not repro-
duced by the calculated spectra. Cyanogen was not in-
cluded as a species in the computational model. Some
of the under-prediction of the intensity magnitude can
therefore be attributed to the exclusion of CN from
the calculation. For the N* (1-) emission, a 200 K
increase in the electronic temperature would produce
a roughly 20 % increase in the calculated intensity.
Comparisons were also done for the 415 and 450 nm
grating positions, which contained mostly molecular
emission, and the agreement between the measured
and calculated intensity magnitude is better, although
the calculated levels are still low. The improved agree-
ment for these grating positions is likely due to their
use to guide the adjustment of the estimated stream
enthalpy.
The shock layer flow also contained significant O and
N atom populations, and atomic transition intensi-
ties were recorded at two near-IR grating positions.
The measured and computed emission from O atomic
transitions at 777 nm and 845 nm are compared in
Figs. 20a and 20b. As with the grating positions at the
shorter wavelengths, the calculated intensity is gener-
ally lower than the measured intensity. At a calculated
electronic temperature of 6000 K, an increase of 200
K would nearly double the atomic emission.
a.)
Experiment
1.
Computation
1
1
u—
b.)
7700 7720 7740 7760 7780 7800
Wavelength, A
20
15
10
5
0
8400 8420 8440 8460 8480
Experiment
Computation
Wavelength, A
Fig. 20. Measured and computed emission from the
a) 777 nm O transitions and from the b) 845 nm O
transition, 20.7 mm upstream of test article.
Similar comparisons were done for N atom transitions
and these are shown in Fig. 21a and 21b for the 744
and 868 nm N transitions, respectively. The calculated
emission for the N atom transitions is also low and
because the emitting states are at energy levels that
are similar to those of the O atom transitions above, an
increase in the electronic temperature of 200 K would
also result in a near doubling of the intensity for these
transitions.
a.)
Wavelength, A
b.)
Fig. 21. Measured and computed emission from the
a) 745 nm N transition and from the b) 868 nm N
transition, 20.7 mm upstream of test article.
3A-20
Based on these spectral comparisons, it appears that
the calculated emission was generally low for all grat-
ing positions at this measurement location. Since the
emission is exponentially dependent on temperature,
it is likely that the calculated flow temperature was
low. Again, the most likely culprit for this discrep-
ancy is the enthalpy, which was probably not raised
to the proper level. Clearly, the exponential depen-
dence of the emission on temperature makes emission
a very sensitive indicator of flow temperature. From
an instrumentation development perspective, this im-
plies that emission- based measurements have both the
signal magnitude and sensitivity that are necessary
to measure temperature, and ultimately flow enthalpy
(with the assupmtion that velocity is negligible within
the shock layer), reasonably well. However, the spa-
tial gradients must be resolved for this approach to
succeed.
A different perspective on the LOS-integrated
number density distribution that was presented in
Fig. 16 can be obtained by comparing measured and
calculated (1-) emission for a single grating posi-
tion at each of the measurement locations on the cen-
tral stagnation streamline. This comparison is shown
for the 426 nm grating position in Fig. 22a for the
back position of the test article, and in Fig. 22b for
the forward position. There is an easily distinguish-
able difference between the evolution of the measured
signal and the calculated signal. By performing a di-
rect comparison of measured and calculated emission
spectra, uncertainties introduced in the analysis that
was performed to derive flow properties from measured
intensities are avoided. However, possible differences
between measured and calculated flow property gradi-
ents are still present and will influence the comparison.
Of particular concern is the possibility of additional
averaging of the measured intensities that may have
been caused by the optical collection system, 65 This
possibility has not been accounted for in these com-
parisons, so only qualitative statements can be made
regarding the differences between the calculated and
measured spectra.
a.)
b4=35.3 mm
b3-33.2 mm
b 1-27.3 mm
Distance from
model surface
b.)
Fig. 22. Nj emission at 420 nm grating position for all
of the axial measurement locations for the high pres-
sure case: a.) back position; b.) forward position.
In Figs. 22a and 22b, it is apparent that the measured
rate of evolution and decline of the Nj emission as the
flow goes from the shock to the test article is less rapid
than predicted by the calculations. Although the lim-
itations in the spatial resolution of the optical system
preclude further statements about the axial distribu-
tion of the Nj emission, the comparison does illustrate
the possibility of using emission spectra to evaluate the
population dynamics of important shock layer species.
4.4 Low Pressure Case
Starting Conditions - The lower pressure case test
conditions were chosen to maximize the change in
shock layer pressure, which was reduced by a factor of
« 3. At the lower pressure, the collision frequency in
the shock layer is reduced significantly. Consequently,
the flow is less likely to be in thermal or chemical equi-
librium. If the degree of departure from equilibrium
could be determined at these test conditions, then
progress could be made in defining the test conditions
that lead to equilibration within the shock layer.
As was found in the comparisons between the simu-
lated and measured emission spectra for the high pres-
sure case, the comparisons for the low pressure case
indicated that the stream enthalpy value derived from
the stagnation point heat flux and pressure measure-
ments was probably low. Therefore, additional calcu-
lations were performed at total enthalpy values that
were 14 % and 32 % higher than the estimated values.
3A-21
In addition to the uncertainty in the stream enthalpy
that was present for both the high and low pressure
test cases, the uncertain argon mass flow rate became
an issue for the simulations of the low pressure con-
ditions. The total mass flow rate for the low pressure
case was reduced by a factor of « 4, based on the re-
duction in pressure, while the start and shield argon
mass flow remained constant. For the high pressure
case, the argon mass flow was assumed to be 5 % of the
total mass flow. This implied that the relative argon
mass flow could be 20 % of the total mass flow for the
low pressure case. At this level, the uncertainty in the
argon mass flow becomes more important because of
its increased participation in the reaction kinetics. For
example, in three-body recombination reactions, Ar is
less efficient than N 2 as the third body. 61 To address
this additional uncertainty and attempt to bound its
influence, flow simulations were performed for three
different argon mass fractions: 5 %, 10 %, and 30 %.
Thus, owing to the uncertain starting conditions a to-
tal of five different simulations of the shock layer flow
for the low pressure case were computed. The start-
ing conditions and computed free stream properties for
each of these simulations are summarized in Table 2.
At nearly constant enthalpy, increased argon mass flow
is seen to increase the temperatures and free stream
velocity slightly, while increasing the dissociation frac-
tion for nitrogen substantially. Comparing cases that
have the same argon mass fraction, increasing the total
enthalpy produces results that are similar to increasing
argon mass fraction at constant enthalpy. This uncer-
tainty in the starting conditions clearly create difficul-
ties for comparisons of simulations and experimental
measurements.
Table 2. Starting conditions and free stream proper-
ties for the low pressure test case simulations
Parameter Case 1 Case 2 Case 3 Case 4 Case 5
po, atm
1.7
1.7
1.7
1.7
1.7
/io,MJ/kg
15.2
15.15
17.15
17.26
20.07
w Ar
.05
.30
.05
.30
.10
«oo, km/s
4.12
4.19
4.29
4.39
4.53
Poo, Pa
62.
57.5
58.5
57.1
57.6
Too, K
727
775
741
823
786
T v 00 , K
2960
3360
3100
3550
3370
Wn 2
.67
.45
.63
.42
.53
w N
.06
.09
.10
.12
.16
WO
.22
.16
.22
.16
.21
Calculated Shock-Layer Properties - Using the
Case 3 conditions, axial profiles of temperatures and
pressure for the shock layer flow were computed, and
these are shown in Fig. 23 as a function of the nor-
malized distance from the test article surface. While
the rotational and translational temperatures are still
higher near the shock, the increase over the levels
nearer the test article is not as great as was seen for
the high pressure case (see Fig. 11). At the lower
shock layer pressure, the vibrational and translational-
rotational temperatures do not appear to converge
outside of the boundary layer of the test article. This
is in contrast to the results of the simulation for the
high pressure case, where the temperatures clearly
converged as the model surface was approached. Ac-
cording to the simulation, the shock layer is not in
thermal equilibrium, except within the boundary layer
at these simulated conditions.
Fig. 23. Axial profiles of temperature and pressure in
the shock layer.
Fig. 24. Axial profiles of species populations in shock
layer.
For the same starting conditions, the axial profiles of
the neutral species and Nj are plotted in Fig. 24, again
as a function of the nondimension al distance from the
surface. As was seen in the high pressure case, there
appears to be a significant peak in the Nj concen-
tration near the shock front. However, for the low
pressure conditions of this simulation, the mass frac-
tion near the shock front is only « 5 times higher than
the mass fraction nearer the test article surface (as
opposed to « 50 times for the high pressure case, see
3A-22
Fig. 12). The mass fractions of N and N 2 do not reach
a limiting value as the surface is approached, so the
simulation indicates that the shock layer is also not in
chemical equilibrium.
a.)
Distance along the line of sight, m
b.)
Distance along the line of sight, m
Fig. 25. Computed flow property profiles for Case 3
conditions at selected axial locations for a) tempera-
tures and b) N2+ number density.
Although the starting enthalpy and argon mass flow
values were less certain for the low pressure case, the
gradients in the Nj and temperature profiles appeared
to be less severe than found for the high pressure case.
To assess the spatial gradients along the optical sight
lines at the measurement locations, temperature and
species profiles were extracted from the shock layer so-
lutions, and these are shown for selected axial locations
in Figs. 25a and 25b, respectively. For both temper-
atures and N t number density, the computed profiles
along the lines-of-sight are much closer to the ideal-
ized top-hat distributions that are required to derive
temperatures that are representative of the core flow
region from the spectral analysis of the measured in-
tensities. Unfortunately, it appears that the majority
of the shock layer flow is likely to be in nonequilib-
rium, which may violate the other major assumption
of the spectral analysis. The degree of departure from
equilibrium and its impact on the distributions of pop-
ulations over the various energy levels is difficult to
quantify.
Comparisons With Measurements - As with the
high pressure case, rotational and vibrational tem-
perature values were derived from an analysis of
spectral features. 66 Using an intensity- weighted aver-
aging approach, 47 temperature values that could be
compared with the experimental values were extracted
from the computed flow properties at the axial mea-
surement locations for some of the different simula-
tion cases. Comparisons between the computed and
measured LOS-averaged temperatures are shown in
Figs. 26 and 27, for the simulation conditions of Case
3 and Case 4, respectively. For these two cases, the
enthalpy levels were in the middle of the range of sim-
ulations and were nearly in agreement. However, for
Case 3 the argon mass fraction was 0.05, while for Case
4, the argon mass fraction was 0.3. In Fig. 26, for the
Case 3 condition, there are significant differences in
both trends and magnitudes between the computed
and measured temperatures. First, for the experimen-
tal values, the rotational and vibrational temperatures
do not appear to overlap, except perhaps accidentally
at one or two measurement locations. Owing to low
signal levels near the shock front, the measured val-
ues are highly uncertain. Consequently, the discussion
of trends will be restricted to the positions nearer to
the surface than x/R — -0.4. For those locations, the
trends in the computed and measured vibrational tem-
peratures appear to be reasonably similar, although
the measured values are generally greater in magni-
tude. In contrast, the measured rotational tempera-
tures do not appear to decrease significantly from the
values near the shock front, while the computed rota-
tional temperatures clearly show evidence of relaxation
going toward the test article.
Fig. 26. Computed and measured LOS-averaged tem-
peratures for the conditions of Case 3 (ho = 17.15
MJ/kg, w Ar = 0.05).
3A-23
Fig. 27. Computed and measured LOS- aver aged tem-
peratures for the conditions of Case 3 (ho = 17.26
MJ/kg, 10 Ar = 0.30).
At the high argon mass flow conditions of Case 4, the
comparison between measured and computed L OS-
averaged vibrational temperatures shows improved
agreement on the magnitude, while maintaining rea-
sonable agreement on the shape of the distribution
at locations closer than x/R = -0.4. The agreement
between magnitudes of the computed and measured
LOS- averaged rotational temperatures also appears to
be improved with the increased argon mass flow, but
the difference between the distributions is unaffected.
It should be noted that the influence of streamwise
and spanwise spatial averaging by the optical collec-
tion system on the measured values has not been fully
accounted for in these comparisons. Considering the
uncertainty in the starting conditions for the simula-
tions and the uncertainty in the unresolved spatial in-
tensity gradients for the experiment, the general lack
of agreement is not surprising.
4.5 Spatially Resolved Measurements
In a recent set of experiments conducted at the low
pressure test conditions, Park acquired emission spec-
tra from multiple locations along the spanwise direc-
tion as a single axial position within the shock layer. 71
Several separate emission measurements were recorded
simultaneously by the spectrograph and CCD system,
and a series of adjusted collection mirror positions were
used to cover the radial extent of the shock layer dur-
ing a single facility run. An Abel-inversion was then
used to obtain spatially resolved emission spectra from
the LOS-integrated intensities. Finally, temperatures
were derived from the Abel-inverted spectra using an
analytical method that involved ratios of N* spectral
features and ratios of 0 atom transitions. A fuller
description of the experiment and the analytical ap-
proach is given in Ref. 71.
Radial distributions of the rotational, vibrational, and
electronic temperatures that were derived from the
spectrally resolved emission are shown in Fig. 28. The
rotational and vibrational temperatures were deter-
mined using the same analytical approach that was
used to derive temperatures from the line-of-sight in-
tensities, above. As explained in Ref. 71, two sets of
atomic oxygen transitions were used to calculate elec-
tronic temperature. Thus, the two electronic temper-
ature distributions are labeled by the shorter wave-
length transition used in each intensity ratio. Ow-
ing to an unresolved background contribution at the
8446.5 A transition, electronic temperatures derived
from the intensity of that transition are systematically
low. Electronic temperature values obtained using the
7773.4 A transition are believed to be valid.
7500
7000
. 6500
Q)
3
£ 6000
Q)
Cl
E
,CD 5500
5000
4500
0 2 4 6 8 10
Distance from the center, cm
Fig. 28. Radial temperature profiles in the shock layer
at 12.7 mm upstream from the test article. Labels for
T e indicate the shorter wavelength 0 transition of the
pair.
The most striking aspect of the radial temperature dis-
tribution shown in Fig. 28 is that the rotational, vibra-
tional, and electronic (for the 7773.4 A pair) tempera-
tures appear to overlap within their respective uncer-
tainties. This may indicate a region of thermal equilib-
rium within the shock layer for the low pressure condi-
tions, contrary to the prediction of the computational
simulation. The shapes of the rotational and vibra-
tional temperature distributions are reasonably similar
to the shapes of the predicted radial temperature dis-
tributions at 9 mm that were shown in Fig. 25a. (Note
that the 9 mm axial location of the predictions was
closest to the 12. 7 mm axial position of the measure-
ments.) The temperatures and the emission spectra
from these experiments are still being analyzed
4.6 Lessons Learned
First, and foremost, the futility of attempting to per-
form detailed computational simulations of arcjet flows
without adequate specification of the starting condi-
tions, mainly the enthalpy, has to be recognized. Un-
Tr
Tv
L L L LMff,
- k Te_7773.4
Te_8446.5
I I I —
3A-24
til this issue is resolved, the knowledge gained from
performing combined experimental and computational
investigations will be marginal. The limitations of us-
ing stagnation point heat flux and impact pressure
measurements to estimate enthalpy were clearly illus-
trated. Although the unspecified argon mass flow rate
did affect the low pressure test simulations, this inflow
parameter is a more tractable problem. Most large-
scale arcjet facilities routinely measure the mass flows
of the test gases, and the Ames Research Center Arcjet
Facilities have recently been equipped with improved
mass flow sensors and control capability.
The second important lesson to derive from this ex-
ercise concerns the use of diagnostic techniques and
approaches that do not resolve spatial gradients. At-
tempts to compare simulation predictions with mea-
surements of spatially integrated quantities can lead to
misleading conclusions. For the low pressure case, the
LOS-averaged temperature distributions suggest that
the flow is not in thermal equilibrium, while the radial
distribution of spatially resolved temperatures suggest
the opposite. An investment of additional effort into
acquiring emission spectra in the radial direction to
obtain Abel- inverted intensities has a far better (and
more certain) return than deriving the comparable in-
tegrated flow properties from a number of computa-
tional simulations. It was fortunate that the com-
parisons between the predicted and measured L OS-
averaged temperatures agreed so poorly for the high
and low pressure cases; otherwise, the temptation to
“correct” the measurements using the computational
results may have proven overwhelming.
Finally, the possibility that a portion of the shock layer
is in thermal equilibrium at the low pressure test condi-
tions contradicts the computational predictions, which
showed extensive thermal nonequilibrium for all of the
different low pressure simulations. Further knowledge
of the chemical state of the shock layer, which is cur-
rently under investigation, and verification of equilib-
rium would provide much- needed insight into the na-
ture of the shock layer flow.
5. Summary and Recommendations
The question of the state of CFD simulations of arc-
jet flows is still dominated by the lack of knowledge
about the flow enthalpy. Any other consideration is
secondary. Conventional methods for estimating the
flow enthalpy, including energy balance, sonic flow,
and stagnation point heat transfer, all provide insuffi-
cient specification of the flow enthalpy for simulation
purposes. If adequate resources and dedicated effort
are brought to bear on this problem, then eventually
it will be resolved and arcjet facility simulations will
become much more meaningful. New LIF-based ap-
proaches to enthalpy measurement may improve this
situation. 58 ’ 59
Although this premise cannot be rigorously tested un-
til enthalpy can be more accurately determined, it
appears that current nonequilibrium, hypersonic-flow
computational models are able to provide reasonable
simulations of arcjet flows. This observation is based
on a qualitative assessment of the comparisons be-
tween the shock layer measurements and predictions,
combined with the fact that no obvious shortcomings
in the flow models could be identified. Despite hav-
ing to estimate the enthalpy and use two CFD models
and a radiative transport code to predict intensity, the
comparisons with measured values were generally fa-
vorable; particularly for the high pressure case.
While the flow enthalpy must be accurately specified to
enable detailed comparisons between simulations and
measurements for a single arcjet facility test condition,
measurements and simulations of relative trends in ar-
cjet characteristics are not similarly constrained. Mea-
surements and simulations of the response of sensitive
flow properties to changes in arcjet control parameters
during a single facility test can add substantial infor-
mation to the knowledge base at the present time. The
approach to this involves using diagnostic instrumen-
tation to monitor stream parameters, when conditions
have stabilized, as a single control variable, such as
the arc current, is changed. (Examples of this type
of experiment will be discussed extensively in the sec-
ond lecture.) Making the best possible estimate of the
total enthalpy for a single condition, a computational
simulation is essentially calibrated at that condition.
Subsequent conditions are then simulated using fur-
ther estimates of the total enthalpy, without chang-
ing the other parameters of the model. Comparisons
are then made between the measured and predicted
trends. This approach avoids the larger uncertainties
that pertain to measurements of absolute quantities.
Some of the more obvious parametric studies to per-
form include: 1.) varying pressure to assess impact on
chemistry; 2.) varying the arc current, which varies
the initial ionization level; and 3.) varying the test
gas composition to investigate relative third-body ef-
ficiencies in N 2 recombination.
To establish the validity of diagnostic approaches,
comparisons between measurements made using multi-
ple independent instruments would be extremely use-
ful. This observation is particularly relevant to de-
veloping new approaches for determining the flow en-
thalpy, which is the most important parameter to mea-
sure accurately, and which ultimately determines how
useful and relevant arcjet testing will become.
At present, it is too early to propose code validation
experiments for large-scale arcjet facility flows. How-
ever, it is appropriate to begin thinking about how to
develop diagnostics and strategies that may eventually
enable code validation experiments in these facilities.
The most important advantage that arcjets have over
3A-25
impulse facilities is the test duration. Steady state
flow conditions and material response can be achieved
and documented. The long test time allows for mul-
tiple property measurements that can be temporally
averaged. Because there is no diaphragm, these mea-
surements can be repeated in multiple runs to assess
facility repeatability with a rapid turn-around time be-
tween tests.
Test guidelines for validation experiments have been
suggested by Mehta in his descriptions of methods for
producing credible computational simulations. 72 Per-
forming validation experiments in arcjet facilities will
require considerable maturation of currently available
diagnostic techniques to ensure adequate specification
of inflow conditions for either nozzle or shock-layer flow
simulations. In addition to improving the accuracy of
inflow condition measurements, sufficient spatial cov-
erage must be attained to allow approximate integra-
tion of the flow properties for comparison with other
measurements of mass flow, energy balance, etc. With-
out this type of internal accuracy check of the experi-
mental results, confidence in the measurements would
not be sufficient to motivate extensive computational
simulations or efforts to improve physical models.
6. Acknowledgements
Many colleagues have contributed to the lecture ma-
terial contained herein, but special thanks are due to
Tahir Gokgen and Chung Park of Thermosciences In-
stitute (Eloret), Mark Newfield of Ames Research Cen-
ter, and James Donohue of UTRC for their excellent
work on the combined computational and experimen-
tal investigation of arcjet flows. John Balboni of the
Thermophysics Facilities Branch at Ames Research
Center provided many useful references on arcjet fa-
cilities in general. The entire staff of the Aerodynamic
Heating Facility Arcjet provided vital test support and
Frank Hui’s efforts as test engineer deserve special
recognition. Many of the thoughts expressed in this
document resulted from conversations with colleagues
about arcjet flows and testing. While any erroneous
statements are attributable solely to the author, the
contributions of Chul Park, Joan Pallix, Raj Venkat ap-
athy, Ellis Whiting, and Jochen Marschall of Thermo-
sciences Institute (Eloret) to what has been written are
greatly appreciated. Similar conversations with Paul
Kolodziej, Jeff Bull, Dave Stewart, Joe Hartman, Joe
Olejniczak, Dave Olynick, Stephanie Langhoff, Sur-
rendra Sharma, and George Raiche of NASA Ames
Research Center have also proven to be very helpful in
writing this document.
7. References
1. P. R. Dennis, C. R. Smith, D. W. Gates, and J.
B. Bond, Editors, Plasma Jet Technology , NASA
Report SP-5033, National Aeronautics and Space
Administration, Washington, DC, October, 1965.
2. D. A. Gerdeman and N. L. Hecht, Arc Plasma Tech-
nology in Materials Science , Springer- Verlag, New
York, 1972.
3. H. A. Stine, “The Hyperthermal Supersonic Aero-
dynamic Tunnel”, presented at International Sym-
posium on High Temperature Technology, Asilo-
mar, CA, 8-11 September, 1963.
4. C. E. Shepard, V. R. Watson, and H. A. Stine,
“Evaluation of a Constricted- Arc Supersonic Jet”,
NASA Technical Note TN D-2066, 1964.
5. C. E. Shepard, “Advanced High- Power Arc Heaters
for Simulating Entries into the Atmospheres of the
Outer Planets”, AIAA Paper No. 71-263, AIAA & h
Aerodynamic Testing Conference”, (1971).
6. V. R. Watson and E. B. Pegot, “Numerical Cal-
culations for the Characteristics of a Gas Flowing
Axially Through a Constricted Arc”, NASA TN D-
4024, 1967.
7. C. Park, J. H. Lundell, M. J. Green, W. Winovich,
and M. A. Covington, “Ablation of Carbonaceous
Materials in a Hydrogen-Helium Arcjet Flow”,
AIAA J., 22, pp. 1491-1498, October, (1984).
8. J. R. Jedlicka, “The Shape of a Magnetically Ro-
tated Electric Arc Column in an Annular Gap”,
NASA Technical Note TN D-2155, 1964.
9. W. Winovich and W. C. A. Carlson, “The 60-
MW Shuttle Interaction Heating Facility”, pre-
sented at the 25 t/l International Instrument Sym-
posium, Anaheim, ISBN 87664-434-5, May, (1979).
10. C. Park, “Laboratory Simulation of Aerothermo-
dynamic Phenomena: A Review”, AIAA Paper
No. 92-4025, AIAA 11 th Aerospace Ground Testing
Conference, Nashville, TN, (1992).
11. R. K. Smith, D. A. Wagner, and J. W. Cunning-
ham, “ A Survey of Current and Future Plasma
Arc- Heated Test Facilities for Aerospace and Com-
mercial Applications”, AIAA Paper No. 98-0146,
36th Aerospace Sciences Meeting, Reno, NV, 12-15
January, 1998.
12. A. Balter-Peterson, F. Nichols, B. Mifsud, and W.
Love, “Arc Jet Testing in NASA Ames Research
Center Thermophysics Facilities”, AIAA Paper No.
92-5041, th AIAA International Aerospace Planes
Conference , (American Institute of Aeronautics and
Astronautics, New York, 1992).
13. C. Scott, “Survey of Measurements of Flow Proper-
ties in Arcjets”, Journal of Thermophysics and Heat
Transfer ,7, pp. 9-24, (1993).
14. H. W. Leipmann and A. Roshko, Elements of Gas-
dynamics , John Wiley & Sons, New York, 149,
(1957).
15. D. A. Gerdeman and N. L. Hecht, Arc Plasma Tech-
nology in Materials Science , Springer- Verlag, New
York, pp. 94-97, (1972).
16. J. Balboni, Thermophysics Facilities Branch, NASA
3A-26
Ames Research Center, private communication,
(1997).
17. W. Winovichj “On the Equilibrium Sonic-Flow
Method for Evaluating Electric-Arc Air-Heater Per-
formance”, NASA TN D-2132, NASA, Washington,
DC, 1964.
18. J. A. Fay and F. R. Riddell, “Theory of Stagnation
Point Heat Transfer in Dissociated Air”, J. Aero-
nautical Sciences , 25 , pp. 73-85, (1958).
19. R. Goulard, “Catalytic Recombination Rates in Hy-
personic Stagnation Heat Transfer”, Jet Propulsion )
28 , pp. 733-745, (1958).
20. R. B. Pope, “ Stagnation-Point Convective Heat
Transfer in Frozen Boundary Layers”, AIAA Jour-
nal , 6, pp. 619-626, (1968).
21. E. V. Zoby, “Empirical Stagnation-Point Heat-
Transfer Relation in Several Gas Mixtures at
High Enthalpy Levels”, NASA TN D-4799, NASA,
Washington, DC, (1968).
22. R. B. Pope, “Measurements of Enthalpy in Low-
Density Arc-Heated Flows”, AIAA J., 6, pp. 103-
110, (1968).
23. E. L. Winkler and R. E. Sheldahl, “Influence of
Calorimeter Surface Treatment on He at- Transfer
Measurements in Arc-Heated Test Streams”, AIAA
J., 4 , pp. 717-716, (1966).
24. L. A. Anderson, “Effect of Surface Catalytic Ac-
tivity on Stagnation Heat-Transfer Rates”, AIAA
Journal , 11 , pp. 649-656, (1973).
25. C. Park, ” Evaluation of Real- Gas Phenomena in
High- Enthalpy Aero thermal Test Facilities: A Re-
view”, J. Thermophusics and Heat Transfer , 11, pp.
330-338, (1997).
26. G. Candler, "Chemistry of External Flows”, in
Aerothermochemistry for Hypersonic Technology ,
VKI Lecture Series 1995-04, von Karman Institute
for Fluid Dynamics, Rhode Saint Genese, (1995).
27. D. L. Cauchon, “Project Fire Flight I Radiative
Heating Experiment”, NASA TM X-1222, (1966).
28. D. L. Cauchon, “Radiative Heating Results from
the Fire II Flight Experiment at a Reentry Velocity
of 11.4 Km/s”, NASA TM X-1402, (1967).
29. D. B. Lee and W. D. Goodrich, “The Aerother-
mo dynamic Environment of the Apollo Command
Module During Suborbital Entry”, NASA TN D-
6792, (1972).
30. D. Olynick, Y.-K. Chen, and M. E. Tauber, “Fore-
body TPS Sizing with Radiation and Ablation for
the Stardust Sample Return Capsule”, AIAA Paper
No. 97-2474, June, (1997).
31. P. A. Gnoffo, “Application of the Program LAURA
to Three-Dimensional AOTV Flowfields”, AIAA
Paper NO. 86-0565, January, (1986).
32. P. A. Gnoffo, R. N. Gupta, and J. L. Shinn, “Con-
servation Equations and Physical Models for Hyper-
sonic Air Flows in Thermal and Chemical Nonequi-
librium”, NASA TP-2867, February, (1989).
33. C. Park, “Assessment of a Two- Temperature Ki-
netic Model for Ionizing Air”, AIAA Paper No. 87-
1574, June, (1987).
34. M. G. Dunn and S.-W. Kang, “Theoretical and
Experimental Studies of Reentry Plasmas”, NASA
CR-2232, (1973).
35. J.-H. Lee, “Basic Governing Equations for the
Flight Regimes of Aeroassisted Orbital Transfer Ve-
hicles”, in Thermal Design of Aeroassisted Orbital
Transfer Vehicles , H. F. Nelson, ed., Volume 96 of
Progress in Astronautics and Aeronautics, AIAA,
New York, pp. 3-53, (1985).
36. G. V. Candler and R. W. MacCormack, “The Com-
putation of Hyperonsic Ionized Flows in Chemical
and Thermal Nonequilibrium” , AIAA Paper No.
88-0511, Jan., (1988).
37. K. G. Brown, “Chemical and Thermal Nonequilib-
rium Heat Transfer Analysis for Hypervelocity, Low
Reynolds Number Flows”, AIAA Paper No. 85-
1033, June, (1985).
38. C. Park, Nonequilibrium Hypersonic Aerothermo-
dynamics , John Wiley & Sons, New York, (1990).
39. R. J. Gessman, C. 0. Laux, and C. H. Krueger, “Ex-
perimental Study of Kinetic Mechanisms of Recom-
bining Atmospheric Pressure Air Plasmas”, AIAA
Paper No. 97-2364, AIAA 28 th Plasmadynamics
and Lasers Conference, Atlanta, GA, June, (1997).
40. P. Durgapal, “Electrode Phenomena in High Cur-
rent, High Pressure Arc Heaters”, J. Thermophysics
and Heat Transfer , 7 , pp. 412-417, (1993).
41. P. Durgapal, “Strongly Coupled Radiative Transfer
and Joule Heating in an Arc Heater Cathode”, J.
Thermophysics and Heat Transfer , 8, pp. 730-736,
(1994).
42. S. F. Shaeffer, “SWIRLARC: A Model for Swirling,
Turbulent, Radiative Arc Heater Flowfields”, AIAA
Paper No. 78-68, (1978).
43. W. N. MacDermott and E. J. Felderman, “Arc
Heater Scaling Parameters Predicted with the
SWIRLARC Code”, AIAA Paper No. 93-2797,
AIAA 28 th Thermophysics Conference, July, (1993).
44. K. H. Kim, 0. H. Rho, and C. Park, “Assessment
of ARCFLO Code and Computations of Arc Heater
Using Navier-Stokes Code”, AIAA Paper No. 99-
0736, AIAA 87 th Aerospace Sciences Meeting, Jan.,
(1999).
45. T. Sakai, K. Sawada, and M. Mitsuda, “Application
of Planck-Rosseland-Gray Model for High Enthalpy
Arc Heaters”, AIAA Paper No. 98-2838, (1998).
46. D. S. Babikian, N. K. J. M. Gopaul, C. Park, “Mea-
surement and Analysis of Nitric Oxide Radiation
in and Arcjet Flow”, J. Thermophysics and Heat
Transfer , 8, pp. 737-743, (1994).
47. T. Gokgen, C. S. Park, M. E. Newfield, and D.
G. Fletcher, “Computational Simulation of Emission
Spectra from Shock Layer Flows in an Arc- Jet Facil-
ity”, J. Thermophysics and Heat Transfer 12, 180-
3A-27
189, (1998); also AIAA Paper No. 97-0135.
48. T. Gokgen, C. S. Park, and M. E. Newfield, “Com-
putational Analysis of Shock Layer Emission Mea-
surements in an Arc-jet Facility”, AIAA Paper No.
98-0891, AIAA 36 th Aerospace Sciences Meeting,
Jan., (1998).
49. M. P. Loomis, S. Polsky, E. Venkat apathy, D. K.
Prabhu, and F. C. L. Hui, “Arcjet Semi-Elliptic
Nozzle Simulations and Validation in Support of X-
33 TPS Testing”, AIAA Paper No. 98-0864, AIAA
36 t/l Aerospace Sciences Meeting, Jan., (1998).
50. D. A. Stewart, Y.-K. Chen, D. J. Bamford, and A.
B. Romanovsky, “Predicting Material Surface Cat-
alytic Efficiency Using Arc-Jet Tests”, AIAA Paper
No. 95-2013, AIAA 30* ^ Thermophysics Confer-
ence, June, (1995).
51. C. Park and S. H. Lee, “Validation of Multi-
temperature Nozzle flow Code”, J. Thermophysics
and Heat Transfer , 9, pp. 9-16, Jan. -March, (1995).
52. A. T. Schonemann, M. Auweter-Kurtz, H. A.
Habiger, P. C. Sleziona, and T. Stocle, “Analysis
of the Argon Additive Influence on a Nitrogen Arc-
jet Flow”, J . Thermophysics and Heat Transfer , 8,
pp. 466-472, (1994).
53. F. S. Milos and D. J. Rasky, “Review of Numeri-
cal Procedures for Computational Surface Thermo-
chemistry”, J. Thermophysics and Heat Transfer ,
8 , pp. 24-34, (1994).
54. W. N. MacDermott, D. D. Horn, and C. J. Fisher,
“Flow Contamination and Flow Quality in Arc
Heaters Used for Hypersonic Testing”, AIAA Pa-
per No. 92-4028, (1992).
55. J. M. Donohue, D. G. Fletcher, and C. S. Park,
“Emission Spectral Measurements in the Plenum
of an Arc-jet Facility”, AIAA Paper No. 98-2946,
1 th AIAA/ASME Joint Thermophysics and Heat-
Transfer Conference, June, (1998).
56. D. J. Bamford, A. O’Keefe, D. S. Babikian, D.
A. Stewart, and A. W, Strawa, “Characteriza-
tion of Arc- Jet Flows Using Laser-Induced Fluo-
rescence”, J. Thermophysics and Heat Transfer 9,
26-33, (1995).
57. D. J. Bamford and A. Romanovsky, “Velocity and
Chemical Composition Measurements in an Arc Jet
Flow”, AIAA Paper No. 95-2039, AIAA 30 t/l Ther-
mophysics Conference, June, (1995).
58. D. G. Fletcher, ’’Arcjet Flow Properties Determined
from Laser-Induced Fluorescence of Atomic Nitro-
gen”, AIAA Paper No. 98-0205, 36 t/l Aerospace
Sciences Meeting, (1998).
59. D. G. Fletcher and D. J. Bamford, “Arcjet Flow
Characterization Using Laser- Induced Fluorescence
of Atomic Species”, AIAA Paper No. 98-2458,
7 th AIAA/ASME Joint Thermophysics and Heat
Transfer Conference, (1998).
60. P. R. Bevington, Data Reduction and Error Analysis
for the Physical Sciences . McGraw-Hill, New York,
pp. 3-7, (1969).
61. D. L. Baulch, D. D. Drysdale, D. G. Horne, and A.
C. Lloyd, Evaluated Kinetic Data for High Temper-
ture Reactions. VoL 2 , Butterworth Group, Lon-
don, pp. 25-53, (1973).
.62. W. Winovich, “Total Radiation Measurements at
the Stagnation Point of Blunt Bodies at Stagnation
Temperatures to 15000 K”, AIAA Paper No. 68-
405, AIAA 3rd Aerodynamic Testing Conference,
April, (1968).
63. A. F. Okuno and C. Park, “Stagnation Point Heat
Transfer Rate in Nitrogen Plasma Flows: Theory
and Experiment”, Journal of Heat Transfer , pp.
372-384, August, (1970).
64. C. O. Laux, “Optical Diagnostics and Radiative
Emission of Air Plasmas” , Stanford High Temper-
ature Gasdynamics Laboratory Report No. HTGL
T-288,' Stanford University, August, (1993).
65. C. S. Park,. M. E. Newfield, D. G. Fletcher, T.
Gokgen, and S harm a, S. P., “Spectroscopic Emis-
sion Measurements within the Blunt Body Shock
Layer in an Arc-Jet Flow”, AIAA Paper No, 97-
0990, Jan., 1997; also J. Thermophysics and Heat
Transfer , 12 , pp. 190-197, (1998).
66. C. S. Park, M. E. Newfield, D. G. Fletcher, and T.
Gokgen “Spectroscopic Measurements of the Flows
in an Arc- Jet Facility”, AIAA Paper No. 98-0893,
36 t/l Aerospace Sciences Meeting; also J. Thermo-
physics and Heat Transfer , 13, pp. 60-67, (1999).
67. T. Gokgen, “Computation of Nonequilibrium Vis-
cous Flows in Arc- Jet Wind Tunnel Nozzles”, AIAA
Paper No. 94-0254, 32 nrf Aerospace Sciences Meet-
ing, (1994).
68. T. Gokgen, “Effects of Freest ream Nonequilibrium
on Convective Heat Transfer to a Blunt Body” , J.
Thermophysics and Heat Transfer , 10, April- June,
pp. 234-241, (1999).
69. E. E. Whiting, J. O. Arnold, and G. C. Lyle, “A
Computer Program for a Line-by-Line Calculation
of Spectra from Diatomic Molecules and Atoms As-
suming a Voigt Line Profile”, NASA TN D-5088,
March, (1969).
70. E. Venkat apathy, J. W. Naught on, and D. G.
Fletcher, “Experimental and Computational Study
of Sonic and Supersonic Plumes”, AIAA Paper No.
95-3496, AIAA Atmospheric Flight Mechanics Con-
ference, August, (1995).
71. C. S. Park, D. G. Fletcher, and J. M. Donohue,
“Spatially Resolved Shock Layer Emission Measure-
ments and Analysis in an Arc-Jet Facility”, AIAA
Paper No. 99-1046, 37 t/l Aerospace Sciences Meet-
ing, Jan., (1999).
72. U. B. Mehta, “Guide to Credible Computer Simula-
tion of Fluid Flows”, J. Propulsion and Power , 12,
pp. 940-948, (1996).
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Books to Borrow
Open Library
TV News
Understanding 9/11