Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69

Contents lists available at ScienceDirect

Journal of Quantitative Spectroscopy &

Radiative Transfer
journal homepage: www.elsevier.com/locate/jqsrt

Flow-radiation coupling for atmospheric entries

using a Hybrid Statistical Narrow Band model
Laurent Soucasse a,n, James B. Scoggins a,b, Philippe Rivière b, Thierry E. Magin a,
Anouar Soufiani b
a
von Karman Institute for Fluid Dynamics, B-1640 Sint-Genesius-Rode, Belgium
b
Laboratoire EM2C, CNRS, CentraleSupélec, Université Paris-Saclay, F-92295 Châtenay-Malabry Cedex, France

a r t i c l e i n f o abstract

Article history: In this study, a Hybrid Statistical Narrow Band (HSNB) model is implemented to make fast
Received 23 July 2015 and accurate predictions of radiative transfer effects on hypersonic entry flows. The HSNB
Received in revised form model combines a Statistical Narrow Band (SNB) model for optically thick molecular
12 April 2016
systems, a box model for optically thin molecular systems and continua, and a Line-By-
Accepted 12 April 2016
Available online 19 April 2016
Line (LBL) description of atomic radiation. Radiative transfer calculations are coupled to a
1D stagnation-line flow model under thermal and chemical nonequilibrium. Earth entry
Keywords: conditions corresponding to the FIRE 2 experiment, as well as Titan entry conditions
Atmospheric entry corresponding to the Huygens probe, are considered in this work. Thermal none-
Nonequilibrium plasma radiation
quilibrium is described by a two temperature model, although non-Boltzmann distribu-
Narrow-band model
tions of electronic levels provided by a Quasi-Steady State model are also considered for
Flow-radiation coupling
radiative transfer. For all the studied configurations, radiative transfer effects on the flow,
the plasma chemistry and the total heat flux at the wall are analyzed in detail. The HSNB
model is shown to reproduce LBL results with an accuracy better than 5% and a speed up
of the computational time around two orders of magnitude. Concerning molecular
radiation, the HSNB model provides a significant improvement in accuracy compared to
the Smeared-Rotational-Band model, especially for Titan entries dominated by optically
thick CN radiation.
& 2016 Elsevier Ltd. All rights reserved.

1. Introduction the density of the atmosphere is low. The numerical

simulation of hypersonic reactive plasma flows coupled
Spacecraft may undergo severe convective and radia- with radiative heat transfer is an active research topic for
tive heating during atmospheric entry at high velocities the design of thermal protection systems of future space
from the surrounding aerothermodynamic environment. missions. Past numerical investigations have shown sev-
Its accurate prediction during the design of such vehicles is eral major coupling effects including (i) radiative cooling of
therefore paramount for the success and safety of future the shock layer due to the strong emission of the plasma
[1–3], (ii) the production of precursor chemical com-
planetary missions. In particular, the radiative heat trans-
pounds ahead of the shock [4], and (iii) the promotion of
fer in the shock layer ahead of the vehicle is known to
ablation products released by the heat shield which may in
significantly alter the aerothermodynamic environment,
turn contribute to increased radiation blockage in the
especially early in the entry when the velocity is high and boundary layer [5]. At high altitudes corresponding to low
densities, the need to consider detailed nonequilibrium
n
Corresponding author. radiation appeared since the middle of the 1980s [6] and
E-mail address: [email protected] (L. Soucasse). thermodynamic and chemical nonequilibrium flowfield

http://dx.doi.org/10.1016/j.jqsrt.2016.04.008
0022-4073/& 2016 Elsevier Ltd. All rights reserved.
56 L. Soucasse et al. / Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69

solvers, coupled to radiative transfer, became common in are based on the Full-Spectrum Correlated-k approach
the 1990s [7,8]. (FSCK) previously developed for IR applications [32] and
The numerical simulation of radiative transfer is a use efficient tabulations and rescaling of the various
challenging problem because of the spatial, angular, and required distribution functions against temperatures,
spectral dependence of the radiation field. The reference molecular electronic state populations, and a typical Stark
approach for treating the spectral dependence is the Line- width of the atoms. The accuracy of these approaches was
By-Line (LBL) method which consists in finely discretizing demonstrated by successful comparisons with LBL results.
the radiative properties over the relevant spectral range. In the case of carbonaceous atmospheres [31], where only
These radiative properties depend on level populations three non-overlapping molecular band systems are con-
and on fundamental spectroscopic data gathered in spec- sidered, such an approach is very efficient and easy to
tral databases such as NEQAIR [9,10], SPRADIAN [11], implement. Moreover, it retains a description of radiative
MONSTER [12], SPECAIR [13]. In the present study, we use properties in terms of absorption coefficients and is there-
the HTGR database (High Temperature Gas Radiation), fore applicable to any radiation solver. For more arbitrary
which has been previously developed [14–18] for O2–N2 gas mixtures, including for instance ablation products, a
and CO2–N2 plasma applications. This database gathers large number of overlapping, non-weak molecular electro-
up-to-date atomic spectroscopic data from various sources nic systems, absorbing in the Voigt regime, and whose
(such as NIST [19] and TOPbase [20]) together with ab induced emission contribution might not be negligible,
initio calculations of diatomic molecular spectra and have to be accounted for. In this case, the multi-scale MS-
atomic line shapes. It includes bound–bound atomic and FSCK approach may become quite tedious to implement.
molecular transitions, bound-free transitions resulting Moreover, the spectral information is completely lost when
from various mechanisms, and free–free transitions. The using such full-spectrum approaches. This is not an intrinsic
covered spectral range is [1000–200,000 cm
1] and the limitation if one is only interested in heat transfer with gray
targeted maximum temperature is 30,000 K. The HTGR walls, but such models do not enable comparisons with
database has been used in several studies for LBL radiative experiments done in limited spectral ranges.
transfer calculations in hypersonic entries. In particular, Recently, Lamet et al. [33] have developed a Hybrid
Lamet et al. [21] performed uncoupled radiation simula- Statistical Narrow Band (HSNB) model, which combines a
tions of the FIRE II flight experiment using a two- Statistical Narrow Band (SNB) model for optically thick
temperature approach to model the thermal none- molecular systems with a box model for optically thin
quilibrium. More recently, Lopez et al. [22] carried out molecular systems and continua, and a LBL description of
coupled flow-radiation simulations of the relaxation atomic line radiation. Band parameters have been com-
behind a shock wave in Air with a consistent state-to-state puted using the HTGR database and tabulated against
modeling of the atomic electronic levels. translational–rotational and vibrational temperatures. The
A full LBL closely coupled flowfield-radiation model has HSNB model can easily include new radiating species and
been developed by Feldick et al. [23] for Earth hypersonic electronic systems, using the uncorrelation assumption
reentries. They used the tangent slab approximation and inside narrow bands, and arbitrary electronic populations
introduced optimized variable wavelength steps to may be specified. Also, it can be applied to predict the
decrease the computational costs. The full LBL simulations radiative flux in the case of non-gray walls. Nevertheless, it
were successfully compared to a hybrid line-by-line-gray might be less computationally efficient than k-distribution
model where molecular radiation in optically thin systems methods.
was assumed to be gray inside narrow bands. However, This study aims at showing the ability of the HSNB
although the LBL method is very accurate, the large model to predict accurately and efficiently coupled radia-
number of radiative transitions that have to be taken into tion effects on hypersonic entry flows. For this purpose, a
account makes it very computationally expensive and 1D stagnation-line flow model for blunt, hypersonic vehi-
impractical for coupled simulations in complex geome- cles [34] has been coupled with a newly developed radia-
tries. The Smeared-Rotational-Band (SRB) model is a tive transport code using the HSNB formulation. Earth entry
common way to simplify the calculation of molecular conditions corresponding to the FIRE 2 experiment, as well
radiation but its accuracy is restricted to small optical as Titan entry conditions corresponding to the Huygens
thicknesses. It has been used for instance in Ref. [24], probe, are considered. Thermal nonequilibrium is described
together with a LBL treatment of atomic radiation. by a two temperature model, although non-Boltzmann
More sophisticated approaches for radiative property distributions of electronic levels provided by a Quasi-
modeling include the k-distribution methods which are Steady State model are also considered for radiative trans-
based on the distribution functions of the absorption coef- fer. Section 2 details the 1D model for coupled flow and
ficient over the whole spectrum (see e.g. [25]) or over radiation along the stagnation-line of the vehicle. The HSNB
spectral narrow bands [26]. They have been widely used for model [33] is presented in Section 3 with new additional
modeling IR radiation in the field of atmospheric physics or features. The results are finally discussed in Section 4.
for combustion applications, but also for modeling visible,
UV and VUV radiation of astrophysical (Opacity Distribution
Function model of Ref. [27]) or thermal [28] plasmas. 2. Governing equations
Recently such models have been developed in the frame-
work of hypersonic nonequilibrium flows for air mixtures We consider a plasma flow constituted of atoms,
[29,30] and also in carbonaceous atmospheres [31]. They molecules and free electrons under chemical and thermal
 L. Soucasse et al. / Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69 57

nonequilibrium conditions. The thermal state of the the form

plasma is described according to a two-temperature model
∂U ∂Fc ∂Fd
in which the translation of heavy species and rotation of þ þ ¼ Sc þSd þ Sk þ Srad : ð5Þ
molecules are assumed to follow a Boltzmann distribution ∂t ∂r ∂r
at the temperature T, and the translation of electrons, U is the conservative variable vector. Fc and Fd are the
vibration of molecules, and electronic excitation of heavy convective and diffusive fluxes. Sc and Sd are convective and
species are assumed to follow a Boltzmann distribution at diffusive source terms resulting from the expansion along
the temperature Tve. The energy of a species s per unit the stagnation-line. Sk is the kinetic and energy transfer
mass is then defined by source term vector and Srad is the radiative source term
vector. Specifically, these vectors are written as
es ðT; T ve Þ ¼ etr ve ve f
s ðTÞ þ es ðT Þ þ es ; ð1Þ
 T
f U ¼ ρs ; ρur ; ρuθ ; ρE; ρeve ; ð6Þ
where etr ve
s , es , and es are the translation–rotational, vibra-
tion-electronic-electron, and formation energies respec-  T
tively, for species s. The rigid-rotor and harmonic- Fc ¼ ρs ur ; ρu2r þp; ρur uθ ; ρur H; ρur eve ; ð7Þ
oscillator models are used to describe rotational and
 T
vibrational modes when computing species thermo- ður þuθ Þ p
p1
Sc ¼
2ρs ; 2ρur ; 3ρuθ
2 ; 2ρH; 2ρeve ;
dynamic quantities, while species electronic energies are r ur þ uθ
computed from the electronic energy levels taken from ð8Þ
Ref. [35].
The pressure p of the plasma is modeled using the  T
Fd ¼ jrs ;
τrr ;
τrθ ; qr
τrr ur ; qve
r ; ð9Þ
perfect gas law
X 1
p¼ ρs rs T þ ρe re T ve ; ð2Þ Sd ¼
2j ; 2ðτθθ
τrr þ τrθ Þ; τθθ
3τrθ ;
sAH r rs
T
2ðqr
τrr ur
τrθ ur
τθθ uθ Þ; 2qve ; ð10Þ
where ρs is the species mass density, rs the perfect gas r

constant per unit mass of species s, H is the set of heavy h iT

species and the subscript e refers to free electrons. Sk ¼ ω_ ks ; 0; 0; 0; Ωve ; ð11Þ

2.1. Stagnation-line flow modeling h iT

Srad ¼ ω_ rad
s ; 0; 0; P
rad
; P rad;ve ; ð12Þ
We follow the methodology of Ref. [36] for deriving a 1D
where the overline symbol introduced in Eqs. (3) and (4) to
plasma flow model along the stagnation-line of a spherical
designate stagnation-line quantities has been omitted for
body (see Fig. 1). The problem is considered symmetric
the sake of clarity. The total energy and the vibrational-
according to the azimuth angle ϕ and the velocity compo-
electronic-electron energy per unit volume are defined by
nent uϕ is set to zero. Mass fractions and temperatures are P
assumed to be only a function of the radius r ρE ¼ s ρs es þ ρu2r =2 (only the radial velocity component
contributes to the kinetic energy due to the ansatz Eq. (4))
P
ys ¼ ys ðr Þ; T ¼ T ðr Þ; T ve ¼ T ve ðr Þ; ð3Þ and ρeve ¼ s ρs eves . Note that the formation energy is not

while velocity and pressure are split into radial and tan- included in eve according to the definition given in Eq. (1).
gential components following This choice is important for the source terms described
below. The total enthalpy is defined by H ¼ E þ p=ρ.
ur ¼ ur ðr Þ cos θ; uθ ¼ uθ ðr Þ sin θ; p
p1 ¼ p ðr Þ cos 2 θ; The radial species diffusion fluxes jrs are obtained by
ð4Þ solving a simplified Stefan–Maxwell equation for heavy
species [38]. The electron diffusion flux is computed from
where the pressure is assumed to follow a Newtonian
the ambipolar assumption that makes the electric current
approximation [37]. Introducing this decomposition into
equal to zero. The radial heat fluxes are defined by
the 3D Navier–Stokes equations in spherical coordinates
and then taking the limit when θ tends to zero leads to a set X tr ∂T ve ∂T
ve
qr ¼ jrs hs
λ
λ ; ð13Þ
of 1D equations for the stagnation-streamline quantities of s
∂r ∂r

X ve ∂T
ve
jrs hs
λ
ve
qve
r ¼ ; ð14Þ
s
∂r

where λtr and λve are the thermal conductivities of the

energy modes in equilibrium with the temperatures T and
Tve respectively, hs ¼ es þps =ρs , hs ¼ eve
ve
s for s ae and
he ¼ ee þ pe =ρe . The components of the viscous stress
ve ve

tensor τrr, τrθ and τθθ are given by

   
4 ∂ur ur þ uθ ∂uθ ur þ uθ 1
τrr ¼ μ
; τrθ ¼ μ
; τθθ ¼
τrr ;
3 ∂r r ∂r r 2
Fig. 1. Spherical body of radius R0 subjected to a hypersonic flow at u1 .
Azimuth and zenith angles are ϕ and θ, respectively. ð15Þ
58 L. Soucasse et al. / Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69

where μ is the viscosity. the stagnation-line and ϑ is the angle made between the
In Eq. (11) ω _ ks are the collisional species mass produc- optical path and the stagnation-line.
tion rates. Reaction rate constants are assumed to follow The radiative mass production rate in Eq. (12) corre-
an Arrhenius law. Forward reaction constants are taken sponds to the production or destruction of a species s
from the literature depending on the mixture considered. during a bound-free radiative process p (photoionization
Backward reaction constants are computed such that A þ hcσ ⇌A þ þ e
or photodissociation AB þ hcσ ⇌A þ B)
Rπ 
equilibrium relations are satisfied [39]. The source term Xνsp M s Z 1
2πκ pσ ðrÞ Iσ r; ϑ sin ϑ dϑ
4πηpσ ðr Þ
Ωve is written as ω_ rad
s ðr Þ ¼
0
dσ ;
p
NA 0 hcσ
  X
∂ur ur þ uθ ev0
ev et0
et ð18Þ
Ωve ¼
pe þ2 þ ρs s VT s þ ρe e ET e
∂r ∂r τ τ
X X
sAM s
where νsp is the stoichiometric coefficient for the bound
s ωs
ΔHp χ_ kp :
þ eve _k ð16Þ free process p and species s (positive for product, negative
s pAR for reactant), N A is the Avogadro number, h is the Planck
constant and c the speed of light.
The first term on the right-hand-side of Eq. (16) represents
The total radiative energy source term P rad is equal to
the internal work done by the electron pressure. The sec-
the opposite of the divergence of the radiative flux qrad,
ond term represents vibration-translation energy
which gives
exchange where ev0 s and evs are vibrational energies of
Z 1Z π
species s at the temperature T and Tve, respectively. The ∂qrad ðr Þ 
P rad ðr Þ ¼
; qrad ðr Þ ¼ 2π Iσ r; ϑ cos ϑ sin ϑ dϑ dσ :
associated relaxation time τs is computed from the
VT
∂r 0 0

experimental data of Ref. [40] taking into account the high ð19Þ
temperature correction from Ref. [41]. The third term For computing the source term P rad;ve , we first neglect the
corresponds to electron-heavy translation energy radiative energy exchanges with translational and rota-
t
exchange where et0 e and ee are translational energies of tional modes. Then, we take into account the fact that
electrons at the temperature T and Tve, respectively. The during a photoionization process, a part of the photon
expression of the relaxation time τET is given in Ref. [42]. energy goes into the formation energy [22]. This leads to
The fourth term represents the coupling between chem- X
istry and vibrational-electronic energy. Finally, the fifth P rad;ve ðrÞ ¼ P rad ðrÞ
Δhp ω_ rad
p;e ðrÞ; ð20Þ
pAJ
term accounts for the energy removed from the electron
bath due to the set R of electron impact ionization and where J is the set of photoinization processes, ω
_ rad
p;e is the
dissociation reactions, where χ_ kp is the molar rate of pro- electron mass production rate for the bound-free process
P
gress and ΔHp ¼ s νsp efs =M s is the chemical heat released p, and Δhp is the ionization energy per unit mass of
per unit mole (νsp is the stoichiometric coefficient for electron.
reaction p and Ms is the molar mass of species s) of reac-
tion p. The set of chemical reactions considered will be 2.3. Numerical implementation
specified for each case in Section 4.
Finally, ω _ rad
s is the radiative mass production rate of When radiation is ignored (Srad ¼ 0) we follow the
rad
species s, P is the total radiative energy source term and implementation of Munafò and Magin [34] for the
P rad;ve is the radiative energy source term for the energy numerical solution of Eq. (5) of which we recall the main
modes in equilibrium with Tve. features. Equations are discretized in space by means of
the finite volume method. The Roe scheme [45] is used to
2.2. Radiative source terms compute the convective flux at cell interfaces. Boundary
conditions, which will be specified in Section 4, are
For computing the radiative source terms of Eq. (12), implemented through ghost cells and are imposed in
we make use of the 1D tangent slab approximation terms of primitive variables. The time integration is fully
[43,44]. The radiative properties of the medium are implicit and is performed until steady state is reached. In
assumed to vary only along the stagnation-line direction order to advance the solution from the time-level n to the
and are assumed to be constant in planes that extend to time-level n þ 1, the fluxes and source terms are linearized
infinity, perpendicular to this direction. It is a reasonable around the solution at the time-level n by means of a
approximation in this work as we will consider shock layer Taylor-series expansion where the Jacobians are computed
of which the size is small compared to the radius of the numerically.
vehicle. When radiation coupling is considered the radiative
The radiative source terms are functions of the spectral source terms are added explicitly to the previous algo-
radiative intensity Iσ which is governed by the radiative rithm. They are not computed at each flow time step but
transfer equation are typically updated every 200 flow time steps. Starting
from an uncoupled solution, time integration is performed
∂I σ ðr; ϑÞ 
cos ϑ ¼ ησ ðr Þ
κ σ ðr ÞI σ r; ϑ ; ð17Þ until steady state is reached. In order to reduce the com-
∂r putational time, radiative calculations are carried out on a
for a non-scattering medium of optical index equal to 1, coarse mesh in which the convective cells are grouped by
where ησ and κ σ are the emission and absorption coeffi- five based on an extensive grid convergence analysis. The
cients at wavenumber σ, r designates the position along angular integration of the intensity field is achieved by
 L. Soucasse et al. / Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69 59

computing the intensity at each point for 20 regularly 3.1. Statistical narrow band model for optically thick mole-
spaced cos ϑ values between
1 and 1. This relatively cular systems
coarse angular discretization was found sufficiently accu-
rate for the planar geometry associated with the tangent For an optically thick molecular system, the emission
slab approximation. The spectral integration over the coefficient and transmissivity are strongly correlated.
wavenumber σ in Eqs. (18) and (19) is achieved using the However, the ratio ησ =κ σ can be considered uncorrelated
Hybrid Statistical Narrow Band model, which will be to the transmissivity τσ , allowing us to write
detailed in Section 3. Finally, radiative source terms are Δσ Δσ Δσ
Δσ η ðs Þ ∂τσ ðs ; sÞ
k 0 k 0 η ðs Þ k 0 ∂τkσ ðs0 ; sÞ
linearly interpolated from the cell centers of the radiation ηkσ ðs0 Þτkσ ðs0 ; sÞ ¼ σk 0 C σk 0 :
κ σ ðs Þ ∂s0 κ σ ðs Þ ∂s0
mesh to the cell centers of the flow mesh.
ð23Þ
This assumption is valid at thermal equilibrium where the
3. The Hybrid Statistical Narrow Band model for ratio ησ =κ σ is equal to the Planck function which is nearly
radiative transfer constant within a narrow band. For thermal none-
quilibrium conditions, Lamet et al. [33] checked that this
The spectral radiative intensity at an arbitrary point s of assumption remains satisfactory for atmospheric entry
an optical path starting at point s¼0 is given by the flow applications. Discretizing the optical path into
homogeneous cells of size Δsi ¼ si þ 1
si , the contribution
solution of the radiative transfer equation, such that
of optically thick molecular systems to the mean intensity
Z s is thus written as
I σ ðsÞ ¼ I σ ð0Þτσ ð0; sÞ þ ησ ðs0 Þτσ ðs0 ; sÞ ds0 ; ð21Þ
0 Δσ
 Rs I thick
σ ðsj Þ
00
where τσ ðs ; sÞ ¼ exp
s0 κ σ ðs00 Þ ds is the spectral trans-
0
j
1   Δσ
XX Δσ Δσ ηkσ
0
missivity between points s and s. We search for an ¼ τkσ ðsi þ 1 ; sj Þ
τkσ ðsi ; sj Þ ∏ τkσ ðsi ; sj Þ
0 Δσ
;
Δσ kAT i ¼ 0
κ kσ 0
k AS
expression of the averaged intensity I σ ðsÞ over a spectral i 0
k ak

narrow band Δσ . First, the radiative mechanisms are ð24Þ

grouped into different contributions: e.g. a molecular where T is the set of optically thick molecular systems and
electronic system, a set of atomic lines, or a continuum S is the set of all the systems. A mean equivalent point si is
process. These contributions are assumed to be statistically introduced to simplify the spatial integration between si
uncorrelated, which allows us to write and si þ 1 , such that
Δσ
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Δσ Δσ
I σ ðsÞ ¼ I σ ð0Þ ∏τkσ ð0; sÞ τkσ ðsi ; sj Þ
Δσ
¼ τkσ ðsi þ 1 ; sj Þ
Δσ
τkσ ðsi ; sj Þ
Δσ
: ð25Þ
k

XZ s Δσ Δσ 0 This pragmatic choice does not cause any significant loss of

ηk ðs0 Þτkσ ðs0 ; sÞ ∏ τkσ ðs0 ; sÞ
0
þ ds ; ð22Þ accuracy.
0 0
k k ak
From statistical assumptions concerning the intensity
where the index k refers to a radiative contribution. and the position of the lines within a narrow band Δσ , the
Numerical tests have shown that the uncorrelation SNB model [46] provides an expression for the mean
assumption between atomic lines and molecular lines was transmissivity of a homogeneous column of the form
Z !
valid with an accuracy of about 1%. Note that we also Δσ 1 W
assume in Eq. (22) the mean intensity at the starting point τσ ðlÞ ¼ expð
κ σ lÞ dσ ¼ exp
; ð26Þ
Δσ Δσ δ
of the path (s¼0) to be uncorrelated with the total
transmissivity as we will only consider gray radiation where δ is the mean spacing between the line positions
leaving the boundaries of the domain. within Δσ and W is the mean black equivalent width of
Δσ
For the evaluation of the term ηkσ ðs0 Þτkσ ðs0 ; sÞ in Eq. these lines. The mean black equivalent width can be set as
(22), we use different procedures which are presented in a function of three band parameters: the mean absorption
the following subsections: (i) a SNB model for optically coefficient κ σ Δσ of the absorbing species and two over-
thick molecular systems; (ii) a box model for optically thin Δσ Δσ
lapping parameters βD and βL related to Doppler and
molecular systems and continua; (iii) a LBL treatment for Lorentz broadening. For addressing non-homogeneous
atomic lines. When all contribution types are included, the optical paths, both the Curtis–Godson and the Lindquist–
resulting method is named the Hybrid Statistical Narrow Simmons approximations have been considered. The
Band (HSNB) model. Lindquist–Simmons approximation is known to be more
The criterion retained to decide whether a molecular accurate but also more computationally expensive com-
system is thick or thin is based on the maximum value of pared to Curtis–Godson. Numerical tests have shown that
the optical depth κ σ l for a plasma at thermodynamic the precision of the Lindquist–Simmons approximation
equilibrium with T ¼8000 K, p ¼2 atm and l ¼10 cm. If the was required during coupled calculations because it pre-
maximum value of κ σ l is greater than 0.1, the molecular vents numerical instability in the free-stream region.
system is considered as thick. Details are given in Appendix A.
60 L. Soucasse et al. / Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69

The required band parameters for radiative transfer where ns is the number density of species s, Qs is the
Δσ
calculations (ησ =κ σ , κ σ , β L , βD ) have been cal- Δσ Δσ Δσ electronic partition function of species s, ξ is the volu-
culated in Ref. [33] for thick molecular systems of air and metric translational partition function of free electrons,
in Ref. [47] for thick molecular systems in Martian atmo- Eion is the ionization energy of species A and kB is the
spheres. They have been tabulated according to two tem- Boltzmann constant. Similar relations can be derived for
peratures T and Tve for 199 spectral bands of constant size photodissociation processes.
Δσ ¼ 1000 cm
1 in the range [1000–200,000 cm
1]. As For both continua and thin molecular systems, the size
explained by in Ref. [33], the parameters can be converted of the spectral bands and the spectral range are the same
for treating arbitrary electronic level populations as each as for thick molecular systems.
electronic molecular system is treated independently.
3.3. Line-by-line treatment of atomic lines
3.2. Box model for optically thin molecular systems and
continua For atmospheric entry applications, many atomic
radiative transitions are optically thick. Attempts have
If a molecular system k is optically thin for all been made to derive a SNB model for atoms but the results
were not sufficiently accurate due to the weak spectral
wavenumbers σ (κ kσ l⪡1, l being a typical length of the
density of atomic lines [33]. Therefore, the contribution of
problem), the mean transmissivity over a spectral narrow
atomic lines to the mean intensity is treated in a LBL
band can be simply expressed by
  manner,
Δσ Rs Δσ 00
τσ ðs ; sÞ
k 0 ¼ exp
s0 κ σ ðs Þ ds . In addition, the cor-
k ″

∣
Δσ
Δσ X
j
1 at
ησ 
I at
σ ðsj Þ ¼ τat
σ ðsi þ 1 ; sj Þ
τ σ ðsi ; sj Þ
at
relation between the emission coefficient and the trans- i¼0
κσ
at
i
missivity is weak such that one can write Δσ
 ∏ τkσ ðsi ; sj Þ
0

ησ τσ Δσ C ησ Δσ τσ Δσ . These two simplifications also hold ; ð31Þ

k0 A S
k0 a at
for continua because of their weak spectral dynamics.
Therefore, the contribution to the mean intensity of the where the first average is computed exactly over the
set of optically thin molecular systems and continua, B, is narrow band.
then written as High resolution atomic radiative properties are com-
Δσ XX
j
1 puted according to
Δσ
I box
σ ðsj Þ ¼ ηkσ Δσ
∏ τkσ ðsi ; sj Þ Δsi : ð27Þ X Aul se
ησ ¼ nu f ul ðσ
σ ul Þhcσ ;
0
kAB i ¼ 0 i k AS
ð32Þ
ul
4π
Δσ Δσ
Band parameters (ηkσ , κ kσ ) for optically thin mole-
Xh i
cular systems have been tabulated according to two tem- κσ ¼ nl Blu f ul ðσ
σ ul Þ
nu Bul f ul ððσ
σ ul ÞÞ hσ ;
a ie
ð33Þ
peratures T and Tve. As for thick molecular systems, they ul
can be converted for treating arbitrary electronic level where Aul, Bul and Blu are the Einstein coefficients related
populations. to spontaneous emission, induced emission and absorp-
Band parameters for continua have been tabulated tion of the transition u-l, nu and nl are the number den-
according to one temperature (Tve for free–free processes,
sities of the upper and lower levels and σul is the wave-
photoionization and photodetachment, and T for O2 pho- se a ie
number of the transition, ful , ful and ful are the line profiles
todissociation). For bound-free processes, three para-
associated to spontaneous emission, absorption and
meters have been tabulated in order to treat chemical non-
induced emission, respectively. The line shapes are related
equilibrium between the species involved in the process:
Δσ to one another to retrieve equilibrium at Tve [46]
the spontaneous emission coefficient ηeq
σ , the true
σ ul 3
f ul ðσ
σ ul Þ ¼ f ul ðσ
σ ul Þ
ie se
Δσ ; ð34Þ
absorption coefficient κ abs
σ and the induced emission σ
Δσ  
coefficient κ ie;eq
σ . For example, for a photoionization σ ul 3 hcðσ
σ ul Þ
f ul ðσ
σ ul Þ ¼ f ul ðσ
σ ul Þ
a se
exp : ð35Þ
process ðA þhcσ ⇌A þ þe
Þ, the spontaneous and induced σ kB T ve
emission coefficients have been tabulated according to the
A Voigt profile is considered for the spontaneous emission
partial pressure of species A, assuming chemical equili- se
line shape ful . Einstein coefficients are taken from the NIST
brium between A and A þ concentrations. Under chemical database and collisional broadening data are taken from
nonequilibrium conditions, the actual radiative properties Refs. [15,48].
can be retrieved according to The LBL treatment of atomic lines is not too penalizing
Δσ Δσ because of the small number of atomic lines (of the order
ηkσ ¼ ηk;eq
σ χ neq ; ð28Þ
of few thousand) compared to the number of molecular
Δσ Δσ Δσ lines (of the order of several million). Furthermore, the
κ kσ ¼ κ k;abs
σ
κ k;ie;eq
σ χ neq ; ð29Þ spectral grid dedicated to atomic radiation can be much
  smaller as compared to the LBL spectral grid including all
nA þ ne
QA Eion
χ neq ¼ exp ; ð30Þ radiative contributions. For this work, an adaptive spectral
nA 2Q A þ ξ kB T ve grid which combines the 11 point stencil per line proposed
 L. Soucasse et al. / Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69 61

in Ref. [49] and a refinement procedure between two lines Table 1

in order to accurately capture the far wing regions has Conditions for the trajectory points t ¼1634 s and t ¼1642.66 s of the FIRE
2 experiment and for the trajectory point t ¼191 s of the Huygens probe
been implemented. More details regarding this procedure
entry. Radius of the vehicle R0, wall temperature Tw and free stream
are given in Appendix B. conditions (temperature T 1 , velocity u1 , total mass density ρ1 and mass
fractions y1 ).

4. Results Case FIRE 2 FIRE 2 Huygens

(1634 s) (1642.66 s) (191 s)

Three hypersonic entry conditions have been studied: R0 (m) 0.935 0.805 1.25
two conditions of Earth entry corresponding to the tra- Tw (K) 615.0 480.0 1000.0
jectory points t¼1634 s and t ¼1642.66 s of the FIRE T 1 (K) 195.0 273.0 183.0
u1 (km/s) 11.36 10.56 4.788
2 experiment and one condition of Titan entry corre-
ρ1 (kg/m3) 3:72  10
5 7.17 10
4 3:18  10
4
sponding to the trajectory point t¼191 s of the Huygens
y1 (%) 77.0–23.0 77.0–23.0 98.84–1.16
probe entry. The FIRE 2 experiment has been the subject of (N2–O2) (N2–O2) (N2–CH4)
many numerical studies because of the availability of flight
data [50,51]. In addition, the study of the first trajectory
point t ¼1634 s allows to investigate strong thermal
nonequilibrium effects. The Titan test case has been cho-
sen to show strong molecular radiation effects coming calculations have been performed with the LBL, HSNB-
from the CN-Violet system [39,52]. Weak and HSNB models from the same flowfield corre-
Boundary conditions for the numerical simulations are sponding to the coupled result obtained with the HSNB
given in Table 1. The wall of the vehicle is assumed to be model. Table 3 gives the incident radiative flux at the wall,
non-ablative, non-catalytic and isothermal at T ¼ T ve ¼ T w together with the total computational time for one radia-
and a no slip condition for the velocity is prescribed. In the tion calculation, for the different combinations of models
free stream, temperature, velocity and mass densities are and test cases. Compared to the reference LBL solutions,
imposed. For radiation, we assume that the boundaries of the HSNB model provides an accurate prediction of the
the computational domain are black walls at Tw and T 1 . incident radiative flux, with an error between 3% and 5%
For Earth entries, a mixture of 11 species (e
, N, N þ , O, and a speed up factor around 80 for the computational
þ þ
O þ , NO, N2, N2 , O2, O2 , NO þ ) is considered. Chemical time. Most of the computational gain comes from the
reactions and rates are taken from Ref. [53]. The radiative calculation of LBL molecular spectra which is very expen-
systems taken into account are listed in Table 2. For pho- sive due to the large number of molecular lines.
todetachment processes, the computation of N
and O
From Table 3, the HSNB-Weak model provides reason-
concentrations are based on a chemical equilibrium ably accurate results for Earth entry with a difference of
assumption with N and O. For Titan entries, a mixture of 13 3.5% and 4.5% for the two trajectory points. However, for
species (N, C, H, N2, C2, H2, CN, NH, CH, CH2, CH3, CH4, the Titan entry case, the incident radiative flux is over-
HCN) at thermal equilibrium is envisaged. For the trajec- predicted by 26%. These are expected results based on a
tory point considered, thermal nonequilibrium effects are previous assessment of Smeared-Rotational-Band models
weak and ionization is insignificant. Chemical reactions in Ref. [24]. Concerning the computational times, the
and rates are taken from Ref. [54]. Finally, all radiative HSNB-Weak model is 5 times faster than the HSNB model
systems taken into account are listed in Table 2. for Earth entry cases. The Lindquist–Simmons approx-
imation used for computing mean transmissivities over
4.1. Accuracy and efficiency of the HSNB model non-homogeneous paths for thick molecular systems (see
Section 3.1 and Appendix A) is responsible for the lower
In order to assess the accuracy and the efficiency of the computational efficiency of the HSNB model.
HSNB model, a comparison with the rigorous Line-By-Line The spectral and cumulated incident radiative fluxes at
(LBL) method is carried out. In LBL calculations, radiative the wall are displayed in Fig. 2 for the early trajectory point
properties of the plasma are computed from the spectro- (1634 s) of the FIRE 2 experiment. It can be seen that the
scopic HTGR database [18] on a high resolution spectral complex structure of the LBL spectral flux is correctly
grid of 4:4  106 points in order to capture correctly all the captured by both HSNB and HSNB-Weak models, with a
atomic and molecular lines. good agreement on the total cumulated flux (see Table 3
It is also interesting to compare the HSNB model with for numerical values). The incoming radiation mostly
the Smeared-Rotational-Band model, which is often used arises from molecular and atomic transitions in the
as a simple model to treat molecular radiation but may Vacuum Ultraviolet. The accuracy of the HSNB model
lead to a strong overestimation of radiative fluxes. For this should also be assessed regarding the total radiative
purpose, we implemented a model similar to the energy source term along the stagnation line. Fig. 2 also
Smeared-Rotational-Band that will be called hereafter shows this distribution together with the difference with
HSNB-Weak. It consists in computing the mean transmis- LBL calculations normalized by the absolute maximum
Δσ
sivity of thick molecular systems according to τkσ ðs0 ; sÞ value of the total radiative source term. The differences do
Rs Δσ 00
¼ exp
s0 κ σ ðs Þ ds
k ″ instead of Eq. (26). not exceed 5% for both HSNB and HSNB-Weak models. The
For the three entry conditions described previously highest discrepancies are located near the shock position,
(FIRE 2 (1634 s, 1642.66 s) and Huygens (191 s)), where the radiation emission is at a maximum.
62 L. Soucasse et al. / Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69

Table 2 The spectral and cumulated incident radiative fluxes at

Radiative systems considered for Earth and Titan entries. the wall and the total radiative source term along the
stagnation line are shown for the Titan test case in Fig. 3.
Earth entries
Atomic lines N, N þ , O, O þ While the HSNB model reproduces with a good accuracy
Thick molecular N2 (Birge–Hopfield 1 and 2, Worley–Jenkins, the LBL calculation, both the spectral flux and the total
systems Worley, Caroll–Yoshino), O2 (Schumann– radiative source term are strongly over-predicted by the
Runge), NO (β, β0 , γ, γ 0 , δ, ε)
HSNB-Weak model. This failure comes from an incorrect
Thin molecular N2 (first and second positive), NO (11,000 Å,
systems
þ
infrared), N2 (first and second negative, treatment of the CN-Violet molecular system in the spec-
Meinel) tral range [25,000–29,000 cm
1]. For this case, the value
Bound-free processes Photoionization (N, O, N2, O2, NO), photo- of the HSNB model is clearly realized.
dissociation (O2), photodetachment An additional assessment of the accuracy of the HSNB
(N
, O
)
Free-free processes N, O, N þ , O þ , N2, O2
model is proposed in Appendix C where we have repro-
duced the LBL results of Ref. [31] concerning the trajectory
Titan entries
Thick molecular N2 (Birge–Hopfield 1 and 2, Worley–
point t ¼189 s of the Huygens probe entry.
systems Jenkins, Worley, Caroll–Yoshino), CN violet,
C2 Swan 4.2. Analysis of radiation effects on flow and heat transfer
Thin molecular N2 (first and second positive), CN (red,
systems LeBlanc), C2 (Philips, Mulliken, Deslandres–
D'Azambuja, Ballik and Ramsay, Fox– In this section, the results of coupled simulations
Herzberg) obtained with the HSNB model are compared to the results
of uncoupled simulations to show how radiative transfer
affects the aerothermodynamic fields and the heat fluxes
at the wall of the vehicle.

4.2.1. FIRE 2 (1634 s)

The FIRE 2 flight conditions of the early trajectory point
Table 3
Comparison between LBL, HSNB-Weak and HSNB models for FIRE 2 t¼1634 s correspond to a high velocity entry into a low
(1634 s), FIRE 2 (1642.66 s) and Huygens (191 s) cases. Incident radiative density atmosphere (see Table 1). The plasma flow around the
fluxes and computational times for one radiation calculation. vehicle is then in strong thermal nonequilibrium that may not
be correctly described by multi-temperature models [42].
FIRE 2 (1634 s) LBL HSNB-Weak HSNB
146.78 151.94 150.85 In order to investigate non-Boltzmann effects, we
qrad;i
w (W/cm2)
t CPU (s) 20,480 41 242 implemented for this particular case the Quasi-Steady State
(QSS) model proposed by Johnston [55]. For each radiation
FIRE 2 (1642.66 s) LBL HSNB-Weak HSNB
553.78 578.81 581.97 calculation, non-Boltzmann populations of electronic levels
qrad;i
w (W/cm2)
t CPU (s) 19,140 51 250 of N and O as well as the first electronic levels (X, A, B, C) of
þ
N2 and N2 are determined from simple correlations
Huygens (191 s) LBL HSNB-Weak HSNB
82.68 104.26 86.24 depending on the macroscopic state of the flow (electron
qrad;i
w (W/cm2)
t CPU (s) 13,158 5 105 temperature, total number densities). An additional
assumption is made concerning N2 VUV systems (Birge–

Fig. 2. FIRE 2 (1634 s). Comparison between LBL, HSNB-Weak and HSNB models. Left: spectral and cumulated incident fluxes at the wall. Right: total
radiative source term along the stagnation line and differences with LBL calculations normalized by the maximum absolute value.
 L. Soucasse et al. / Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69 63

Fig. 3. Huygens (191 s). Comparison between LBL, HSNB-Weak and HSNB models. Left: spectral and cumulated incident fluxes at the wall. Right: total
radiative source term along the stagnation line and differences with LBL calculations normalized by the maximum absolute value.

Table 4 plasma density increases and the heat fluxes decrease because
FIRE 2. Standoff distance δ, conductive flux at the wall qw, radiative flux at the the temperature levels are lower. In the coupled QSS case, the
rad
wall qw and incoming radiative intensity at the wall Iw over spectral intervals
    same trends are obtained but to a lesser extent. Table 4 also
Δσ 1 ¼ 16; 667
33; 333 cm
1 and Δσ 2 ¼ 2500
50; 000 cm
1 . shows the incoming radiative intensity at the wall over two
specific spectral ranges corresponding to the experimental
FIRE 2 (1634 s) Uncoupled Coupled Coupled QSS Flight
data flight data, given with an uncertainty of 20%. All uncoupled,
coupled, and coupled QSS results are far from the flight data,
δ (cm) 5.36 5.05 5.21 – although the QSS case is much closer than the two others.
qw (W/cm2) 94.9 76.3 76.5 – Fig. 4 displays temperatures along the stagnation line for
rad
qw (W/cm2) 203.6 150.0 74.7 – the uncoupled, coupled, and coupled QSS cases. The uncou-
Iw(Δσ 1 ) (W/cm2/sr) 2.05 2.12 1.28 0.1 pled temperature profile can be split into four regions from
Iw(Δσ 2 ) (W/cm2/sr) 8.57 7.16 4.71 1.3 right to left: the free-stream, the shock, the equilibrium
FIRE 2 (1642.66 s) Uncoupled Coupled Flight data
plateau and the boundary layer close to the wall. Because of
δ (cm) 4.06 4.01 – the low density, the shock region is wide and in strong
qw (W/cm2) 635.6 617.0 – thermal nonequilibrium. In the boundary layer, Tve is slightly
rad greater than T, because of atomic recombination which cre-
qw (W/cm2) 791.3 581.7 –
Iw(Δσ 1 ) (W/cm2/sr) 11.65 9.28 10.5 ates vibrational energy. When radiation is considered (cou-
Iw(Δσ 2 ) (W/cm2/sr) 71.01 53.63 63 pled case), the temperature distributions are significantly
affected. The shock layer spreads out and the equilibrium
zone is shortened. Radiative cooling lowers the peak and
Hopfield 1 and 2, Worley–Jenkins, Worley, Caroll–Yoshino): plateau temperatures. In particular, the maximum of Tve
the population of the upper energy levels of these transi- decreases from 14,670 K (uncoupled) to 13,470 K (coupled).
tions, which are above the dissociation limit, are computed Another interesting feature is that the free-stream region is
according to a chemical equilibrium assumption with atomic no longer at thermal equilibrium because the radiation
Nitrogen. The non-Boltzmann populations are taken into absorption from the shock increases electronic and vibra-
account in the HSNB model using Eqs. (32) and (33) for tional energy.
atomic radiation and using expressions given in Ref. [33] for Fig. 4 also shows the species molar fractions along the
band parameters of molecular systems. It should be under- stagnation line. For the uncoupled case, the main chemical
lined that this QSS model used for radiation is not consistent mechanisms are the dissociation of molecular nitrogen
with the flow modeling, though it provides a first approx- and oxygen and the ionization of atomic nitrogen and
imation of thermal non-equilibrium effects. The full con- oxygen through the shock. A significant amount of Nitro-
sistent state-to-state coupling between flow and radiation gen monoxide is also produced in the shock region. The
has been recently achieved for instance in Ref. [22] for ionization level is quite important in the plateau as the
atomic electronic states. electron molar fraction reaches 0.15. In the boundary layer,
Table 4 compares the uncoupled, coupled, and coupled the ionization level drops down and atomic nitrogen starts
QSS results concerning the shock standoff distance, the con- recombining. For the coupled case, the fall of the two
ductive flux, and the radiative flux at the wall. The coupling temperatures slows down the ionization reactions and the
effect is to decrease all of these quantities due to the radiative electron molar fraction reaches a maximum of 0.08. The
cooling associated with the strong radiative emission in the free-stream region ahead of the shock becomes chemically
shock layer. The shock layer thickness decreases because the reacting under the effect of radiation: atomic oxygen is
64 L. Soucasse et al. / Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69

produced by photodissociation and electrons are produced 20%. Temperatures and composition along the stagnation
mainly by photoionization of molecular oxygen. line are shown in Fig. 6. For the uncoupled case, the flow is
These coupling effects on temperature and compo- in thermal equilibrium everywhere except in the shock
sition are also noticeable in Fig. 4 for the coupled QSS region. A chemical equilibrium zone with flat molar frac-
case, however they are much weaker. In order to tion profiles is clearly distinguishable between the
understand this behavior, the radiative source term boundary layer and the shock. The electron molar fraction
along the stagnation line is plotted in Fig. 5 and split is around 0.07 in this zone. When radiation is considered,
according to the atomic, molecular, and continua emis- as for the previous trajectory point, the vibration-
sion contributions. From uncoupled to coupled calcula- electronic-electron temperature Tve increases and a slight
þ
tions, the peak of the radiative source term is decreased fraction of O, e
and O2 are produced in the free stream
by a factor two due to radiative cooling. From coupled to by photodissociation and photoionization of O2. The tem-
coupled QSS calculations, the peak of the radiative perature is slightly decreased in the shock layer, leading to
source term is further reduced and atomic contribution a lower ionization level.
almost vanishes at the shock location. The reason is that
the Tve Boltzmann distribution leads to an over- 4.2.3. Huygens (191 s)
estimation of the population of the highest electronic This Titan entry case has been selected to show the
þ
energy levels of atomic N and O, as well as N2 and N2 . In ability of the HSNB model to handle strong optically thick
particular, for the N2 VUV systems, the dissociation molecular radiation. For early trajectory points, a state-to-
equilibrium assumption makes the population of the state electronic specific model of the CN molecule is
upper electronic energy level associated with these required [39,55]. As thermal nonequilibrium modeling has
systems close to zero and thus cancels their contribution been already discussed for the FIRE 2 t¼1634 s case, we
to radiation. The remaining molecular emission peak in will focus here on an quasi-equilibrium case at the tra-
Fig. 5 comes mostly from NO radiation, of which elec- jectory point t ¼191 s.
tronic energy levels are assumed to be populated When coupled radiation effects are taken into account,
according to the temperature Tve. An incorrect treatment the standoff distance decreases from 10.77 (uncoupled) to
of the thermal state of NO might be responsible of the 10.33 cm (coupled), the conductive heat flux decreases from
remaining discrepancies between the coupled QSS 24.2 to 22.0 W/cm2 and the radiative flux decreases from
results and the flight data. 95.6 to 80.57 W/cm2. The radiative flux obtained in the
uncoupled case is in agreement with Ref. [52]. They found a
4.2.2. FIRE 2 (1642.66 s) radiative flux of 75 W/cm2 but they applied to their results a
The FIRE 2 flight conditions of the trajectory point correction coefficient of 0.75 to model the 3D effects.
t ¼1642.66 s correspond to a quasi-thermal equilibrium The temperature and CN molar fraction along the stag-
situation. Due to the higher density, the kinetic energy nation line are plotted in Fig. 7. In the shock layer, we can
transfer between the internal energy modes is much faster notice a lower temperature due to radiative cooling and a
than for the trajectory point t ¼1634 s. Thus, the QSS higher CN molar fraction because of lower dissociation rates.
model has not been considered in this case. The role of CN radiation, especially the CN-Violet
As it can be seen in Table 4, the standoff distance and system, has been highlighted when discussing the accu-
the conductive flux slightly decrease and the radiative flux racy of the HSNB model in Section 4.1. Indeed, it can be
diminishes by around 25% when radiation is coupled to the seen in Fig. 8 that the dominant contributors to
flow. In addition, coupled results are in fair agreement radiation are the CN-Violet, followed by the CN-Red
with flight data which are given with an uncertainty of molecular systems. N2 radiation also contributes

Fig. 4. FIRE 2 (1634 s). Temperatures and composition along the stagnation line.
 L. Soucasse et al. / Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69 65

Fig. 5. FIRE 2 (1634 s). Total radiative source term along the stagnation line.

Fig. 6. FIRE 2 (1642.66 s). Temperatures and composition along the stagnation line.

Fig. 7. Huygens (191 s). Temperature and CN molar fraction along the stagnation line.

significantly at the emission peak. However, C2 radiation on hypersonic entry flows. The HSNB model reproduces
is weak. When radiation is coupled to the flow, the emis- the LBL results with an accuracy better than 5% and a
sion is reduced at the peak because of the lower speed up of the computational time around two orders of
temperature. magnitude. Concerning molecular radiation, the HSNB
model provides a significant improvement compared to
the Smeared-Rotational-Band model in the case of Titan
5. Conclusion entry dominated by optically thick CN radiation. Taking
into account the coupling with radiation, both convective
We have shown in this paper the ability of the HSNB and radiative heat fluxes at the wall decrease. The standoff
model to accurately and efficiently predict radiation effects distance is reduced due to radiative cooling of the shock
66 L. Soucasse et al. / Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69

Fig. 8. Huygens (191 s). Total radiative source term along the stagnation line.

layer. For Earth entry cases, we observed slower ionization Acknowledgments

þ
levels as well as the appearance of O, e
and O2 in the free
stream produced by the photodissociation and the pho- This research was sponsored by the European Research
toionization of O2. Council Starting Grant no. 259354 and by the European
This study was focused on a 1D flow-radiation model in Commission Project ABLAMOD (FP7-SPACE No. 312987).
order to demonstrate the feasibility of coupled simulations The authors gratefully acknowledge Dr. Alessandro
with the HSNB model. However, 2D or 3D simulations may Munafò for his previous developments on the stagnation-
be required to determine the spatial distribution of the line flow solver.
heat flux on the wall. For these geometries, coupled cal-
culations are not practical with a LBL description of
radiative properties. Though through the use of paralleli- Appendix A. Curtis–Godson and Lindquist–Simmons
zation, the HSNB model may be used to compute radiative approximations for SNB model
source terms accurately in a reasonable amount of time.
The Monte Carlo method would be probably the best For a homogeneous optical path, analytical expressions
algorithm to solve the radiative transfer equation in can be derived for the mean black equivalent width W =δ
this case. introduced in Eq. (26) for mean transmissivity calculations.
The analysis of the first trajectory point of the FIRE These expressions depend on the broadening mechanism
2 experiment has shown that the two temperature and a prescribed distribution law for line intensities within
approach was not satisfactory to model the thermal the narrow band and are functions of two parameters: a
nonequilibrium state of the plasma. The results obtained mean absorption coefficient k per unit partial pressure of
using non-Boltzmann electronic populations provided by the absorbing species and an overlapping parameter β .
the QSS model of Johnston [55] were closer to the flight For a non-homogeneous optical path, the Curtis–God-
data even if some discrepancies still remain. For such entry son approximation consists in using the expressions
conditions (high velocity, low density), a self-consistent derived for homogeneous media with averaged para-
 
electronic specific collisional-radiative model should be meters k and β that we define according to
developed. The HSNB model could be used since it is Z s
00
compatible with arbitrary populations of electronic states. u¼ pa ð s00 Þ ds ; ðA:1Þ
s0
Finally, it is worth mentioning the usefulness of the
presented HSNB model for studying coupled ablation- Z s
 1 00
radiation phenomena in the boundary layer of future k ¼ pa ðs00 Þkðs00 Þ ds ; ðA:2Þ
u s0
entry vehicles. Light-weight carbon-phenolic ablators have
Z s
already been extensively used for several recent missions  1
β ¼  βðs00 Þpa ðs00 Þkðs00 Þ ds00 ; ðA:3Þ
and are likely to dominate future space missions as we uk s0
look forward to evermore challenging entry environments.
where u is the mean pressure path length and pa the
One notable effect that ablators may have on the radiation
partial pressure of the absorbing species. Expressions of
field is the ability of blown ablation products to absorb
the mean black equivalent width W L ðs0 ; sÞ=δ and
radiant energy from the shock layer and carry this energy
W D ðs0 ; sÞ=δ for both Lorentz and Doppler broadening are
down stream of the stagnation region. As several of the
given in Table 5. In order to obtain the mean black
relevant carbonaceous species have already been included
equivalent width in the Voigt broadening regime, we use
in the HTGR database and their HSNB band parameters
the expression proposed of Ref. [56]
have already been presented, the HSNB model could be
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
used to significantly reduce the cost of ablation-radiation W V ðs0 ; sÞ 
1=2
¼ uk 1
Ω ; ðA:4Þ
studies in the future. δ
 L. Soucasse et al. / Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69 67

Table 5
Expressions of mean black equivalent width for Curtis–Godson approximation and its derivative for Lindquist–Simmons approximation. The first column
indicates the broadening type and the distribution function considered for line intensities. For Doppler broadening, the tailed-inverse exponential dis-
tribution is used for air systems (N2, O2, NO) and the exponential distribution is used for carbonaceous systems (CN, C2).

Approximation W ðs0 ;sÞ ∂W ðs″ ;sÞ

Curtis–Godson δ Lindquist–Simmons
1δ ∂s″

rffiffiffiffiffiffiffiffiffiffiffiffiffiffi   pffiffiffiffiffiffiffiffiffiffi
Lorentz, tailed-inverse exponential  pa s″ kðs″ Þ 2xr þ ð1
r2 Þpffiffiffiffiffiffiffiffiffi
1 þ 2x ffi
2βL 1 þ uk 
1 ð1
r2 þ 2xÞ 1 þ 2x
βL

″
x¼ uk
; r ¼ βL ðs Þ
2βL βL
Doppler, exponential 
βD R þ 1 xexpð
ξ2 Þ pa ðs″ Þkðs″ Þ R þ1 expð
ξ2 Þ
pffiffiffi
1 1 þ xexpð
ξ2 Þ dξ pffiffi dξ
π π
1 ð1 þ xexpð
r2 ξ2 ÞÞ
 
″
x¼ uk
 x¼ uk
; r ¼ βD ðs Þ
βD βD βD
Doppler, tailed-inverse exponential βp

Dffiffi
R þ1   pa ðs″ Þkðs″ Þ R þ1 expð
ξ2 Þ
π
1 ln 1 þ xexp
ξ2 dξ pffiffi dξ
π
1 ð1 þ xexpð
r2 ξ2 ÞÞ
 
″
x¼ uk
 x¼ uk
; r ¼ βD ðs Þ
βD βD βD

2 (1634 s) test case. The difference between the two

approximations is very tiny and for most of the points is
much lower than the difference between HSNB Lindquist–
Simmons and LBL calculations. However, we can see a
much larger discrepancy in the free stream: the radiative
source term predicted by the HSNB Curtis–Godson model
becomes negative while it remains positive for the HSNB
Lindquist–Simmons and LBL models. A negative energy
source term in this cold region is a major computational
issue for coupling because negative temperatures can be
predicted after few iterations. This is why the Lindquist–
Simmons approximation has been chosen for the coupled
simulations even though it requires around three times
more computational time.

Fig. 9. FIRE 2 (1634 s). Comparison between Curtis–Godson and Lind-

Appendix B. Adaptive spectral grid for atomic radiation
quist–Simmons approximations.
This appendix describes the method for generating an
adaptive spectral grid for the LBL treatment of atomic
2 !2 3
2 2 !2 3
2 radiation in the HSNB model (Section 3.3). We start from
1 W L ðs0 ; sÞ 5 1 W D ðs0 ; sÞ an 11 point stencil [49] defined by σul, σ ul 7 Δ, σul being
Ω ¼ 41
þ 41
5
1:
uk

δ uk

δ the line center and Δ the 5 point half-stencil
 
ðA:5Þ 1 1 25
Δ ¼ γ V ; γ V ; Δσ W ; Δσ FW ; γ V : ðB:1Þ
8 2 2
An alternative approach for treating non-homogeneous
optical path is the Lindquist–Simmons approximation [57]. The estimated distances from the line center to the line
It consists in finding expressions for the space derivative of wing Δσ W , and to the far wing Δσ FW are computed by
  
the mean black equivalent widths in Lorentz ∂W L ðs00 ; sÞ=∂s00 Δσ W ; Δσ FW ¼ π2 1 þ ζ γ L þ αγ D , where γL and γD are the
and Doppler ∂W D ðs″ ; sÞ=∂s00 broadening regimes. The half widths at half maximum of the line related to Lorentz
expressions used is in this work are given in Table 5. They and Doppler broadening respectively. The values of the ζ,
involve both local parameters kðs″ Þ, βðs″ Þ and averaged
  α constants are taken to be f1; 1:8g for Δσ W , while they are
parameters k , β . The non-uniform mean black equivalent qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
widths in each broadening regime are then obtained by chosen as f2:6; 5:8g for Δσ FW . γ V ¼ γ 2D þ γ 2L is the esti-
spatial integration, according to mated Voigt half width at half maximum. It was shown
Z s that such stencil provides a reasonably accurate resolution
W L=D ðs0 ; sÞ 1 ∂W L=D ðs″ ; sÞ 00
¼ ds ; ðA:6Þ of line intensities for low pressure atmospheric entry
δ s0 δ ∂s″
conditions due to the low degree of line broadening in that
and Eqs. (A.1) and (A.2) are used again to get the mean regime [49].
black equivalent width in the Voigt broadening regime. This fixed point method is likely to work well in spectral
A comparison between the HSNB Curtis–Godson and regions with a high number of electronic transitions and
the HSNB Lindquist–Simmons is given in Fig. 9 for the FIRE with a large degree of line overlap because the majority of
68 L. Soucasse et al. / Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69

the points are distributed around the line centers. For areas
in which there are large distances between neighboring line
centers, the method is likely to provide poor estimates of
spectral quantities due to the large error in interpolating the
spectral values in the far line wing regions.
For this reason, we implemented an adaptive refinement
procedure in order to accurately compute the far line wing
regions for atomic spectra which have a relatively weak
spectral density. To begin, the complete line list for atoms
considered is first ordered by ascending line center values.
Then, each region between two consecutive lines is con-
sidered. For each consecutive line pair, an adaptive mesh is
created based on the two corresponding line shapes. We will
denote the left line shape properties with the superscript L,
and the right properties with an R. First, the approximate
center point between each line is defined simply as
Fig. 10. Huygens (189 s). Comparison between LBL, HSNB and LBL results
σ LR ¼ ðσ Lul þ σ Rul Þ=2. Next, the following set of points are from Bansal and Modest [31]. Contribution of the CN red and violet
added to the mesh based on the 11 point stencil, but systems to the radiative source term along the stagnation line.
ensuring that points added for each line do not overlap one
another: the Huygens probe entry from uncoupled flowfield taken
from Johnston [55]. The spectroscopic constants were
σ L
ul þ Δ;
L
8 Δ o σ LR
σ
L L
ul ðB:2Þ
taken from Laux [59] and nonequilibrium populations of
σ Lul
ΔR ; 8 Δ o σ Rul
σ LR
R
ðB:3Þ the CN electronic states were considered.
In order to reproduce their results we have computed
For lines which are sufficiently close, the above proce- the radiative source term along the stagnation line with
dure will prevent unnecessary points from being added to
both the HSNB model and the LBL approach. For the CN
the spectral grid. For lines which are very far apart in
electronic states, both Boltzmann populations and none-
comparison to their line widths, the above set of points are
quilibrium populations based on the QSS model of Bose
augmented by adding points recursively to the center
region by successively bisecting the two intervals closest et al. [60] have been taken into account. The comparison is
to the last points added by the stencil above until the shown in Fig. 10 for the CN red and violet systems. First of
spacing between the outermost two stencil points for each all, we can see that for a given population assumption, our
line is at least half the size of the spacing between the LBL and HSNB results give similar results. When the QSS
outermost stencil point and the next point. In other words, model is used, we obtain a good agreement with the LBL
two bisection fronts are propagated towards the line results of Bansal and Modest [31] for both LBL and HSNB.
centers until the spacing between points matches that of Note that nonequilibrium effects are not negligible in the
the two outermost stencil points of each line. considered simulation and lead to about 16% difference at
The accuracy of this adaptive procedure has been suc- the peak value of the radiative source term.
cessfully compared with the fine LBL spectral grid of 4:4 
106 spectral points that we use for full LBL calculations [58].
The differences we obtained were negligible compared to the
accuracy of the full HSNB model which is around a few References
percent (see Section 4.1). The spectral size of the adaptive
mesh for atomic radiation was around 4  104 points for the [1] Gökçen T, Park C. The coupling of radiative transfer to quasi-1D flows
applications considered in this paper, which allows the with thermochemical nonequilibrium. In: AIAA 29th aerospace sci-
computational time to be two order of magnitude less. Thus ences meeting. 1991. p. 570.
[2] Hartung L, Micheltree R, Gnoffo P. Coupled radiation effects in
this adaptive spectral mesh procedure constitutes a decisive thermochemical nonequilibrium shock-capturing flowfield calcula-
development for the practical use of the HSNB model. tions. J Thermophys Heat Transf 1994;8(2):244–50.
[3] Wright MJ, Bose D, Olejniczak J. Impact of flowfield-radiation cou-
pling on aeroheating for Titan aerocapture. J Thermophys Heat
Transf 2005;19(1):17–27.
Appendix C. Accuracy of the HSNB model against lit- [4] Stanley SA, Carlson LA. Effects of shock wave precursors ahead of
erature results hypersonic entry vehicles. J Spacecr Rocket 1992;29(2):190–7.
[5] Johnston CO, Gnoffo PA, Sutton K. Influence of ablation on radiative
heating for Earth entry. J Spacecr Rocket 2009;46(3):481–91.
The accuracy of the HSNB model has been assessed in [6] Park C. Radiation enhancement by nonequilibrium in earths atmo-
Section 4.1 by comparison with LBL results, where both sphere. J Spacecr Rocket 1985;22:27–36.
[7] Hartung LC. Development of a nonequilibrium radiative heating
HSNB parameters and LBL radiative properties were
prediction method for coupled flowfield solutions. J Thermophys
obtained from the HTGR spectroscopic database. The pur- Heat Transf 1992;6:618–25.
pose of this appendix is to compare the HSNB model with [8] Sharma S. Modeling of nonequilibrium radiation phenomena: an
reference results obtained by other researchers. assessment. J Thermophys Heat Transf 1996;10:85–396.
[9] Whiting E, Park C, Liu Y, Arnold J, Paterson J. NEQAIR96, none-
We have considered the LBL radiation simulation of quilibrium and equilibrium radiative transport and spectra program:
Bansal and Modest [31] of the trajectory point t¼189 s of users manual. NASA Reference Publication; 1996.
 L. Soucasse et al. / Journal of Quantitative Spectroscopy & Radiative Transfer 180 (2016) 55–69 69

[10] Cruden BA, Brandis AM. Updates to the NEQAIR radiation solver. In: [35] Gurvich LV, Veyts IV, Alcock CB. Thermodynamic properties of
Proceedings of the 6th high temperature gas radiation workshop, individual substances. 4th ed., New York: Hemisphere Pub. Corp;
ESA, St. Andrews, UK. 2014. 1989.
[11] Fujita K, Abe T. SPRADIAN, structural package for radiation analysis: [36] Klomfass A, Müller S. Calculation of stagnation streamline quantities
theory and application. ISAS report no. 669, 1997. in hypersonic blunt body flows. Shock Waves 1997;7:13–23.
[12] Kuznetsova L, Surzhikov S. Absorption cross sections of diatomic [37] Anderson JD. Hypersonic and high temperature gas dynamics. New
molecules for problems of radiative heat transfer in low- York: McGraw-Hill Book Company; 1989.
temperature plasma. High Temp 1999;37(3):374–85. [38] Magin TE, Degrez G. Transport properties of partially ionized and
[13] Laux CO. Radiation and nonequilibrium collisional-radiative models. unmagnetized plasmas. Phys Rev E 2004;70:046412.
VKI special course on physico-chemical models for high enthalpy [39] Magin TE, Caillault L, Bourdon A, Laux CO. Nonequilibrium radiative
and plasma flows modeling, 2002. heat flux modelling for the Huygens entry probe. J Geophys Res
[14] Chauveau S, Perrin MY, Rivière Ph, Soufiani A. Contributions of 2006;111:E07S12.
diatomic molecular electronic systems to heated air radiation. J [40] Millikan R, White D. Systematics of vibrational relaxation. J Chem
Quant Spectrosc Radiat Transf 2002;72:503–30. Phys 1963;39(12):3209–13.
[15] Chauveau S, Deron CH, Perrin MY, Rivière Ph, Soufiani A. Radiative [41] Park C. Review of chemical-kinetic problems of future NASA mis-
transfer in LTE air plasmas for temperatures up to 15000 K. J Quant sions, I: Earth entries. J Thermophys Heat Transf 1993;7(3):385–98.
Spectrosc Radiat Transf 2003;77:113–30. [42] Panesi M, Magin TE, Bourdon A, Bultel A, Chazot O. Fire II experi-
[16] Babou Y, Rivière Ph, Perrin MY, Soufiani A. High-temperature and ment analysis by means of a collisional-radiative model. J Thermo-
nonequilibrium partition function and thermodynamic data of dia- phys Heat Transf 2009;23(2):236–48.
tomic molecules. Int J Thermophys 2009;30(2):416–38. [43] Olynick DR, Henline WD, Chambers LH, Candler GV. Comparison of
[17] Babou Y, Rivière Ph, Perrin MY, Soufiani A. Spectroscopic data for the coupled radiative flow solutions with project Fire II flight data. J
prediction of radiative transfer in CO2
N2 plasmas. J Quant Spec- Thermophys Heat Transf 1995;9(4):586–94.
trosc Radiat Transf 2009;110:89–108. [44] Johnston CO, Hollis BR, Sutton K. Nonequilibrium stagnation-line
[18] Soufiani A, Rivière Ph, Perrin MY. High temperature gas radiation radiative heating for Fire II. J Spacecr Rocket 2008;45(6):1185–95.
(HTGR) database and models. Von Karman Institute lecture series [45] Roe PL. Approximate Riemann solvers, parameter vectors, and dif-
(STO-EN-AVT-218), 2013. ference scheme. J Comput Phys 1981;43:357–72.
[19] Kramida A, Ralchenko Y, Reader J. NIST ASD Team. NIST Atomic [46] Goody HR, Yung Y. Atmospheric radiation. New York: Oxford Uni-
Spectra Database (version 5.2). 2014 〈http://physics.nist.gov/asd〉. versity Press; 1989.
[20] Cunto W, Mendoza C, Ochsenbein F, Zeippen CJ. TOPbase at the CDS. [47] Depraz S, Rivière Ph, Perrin MY, Soufiani A. Band models for radia-
Astron Astrophys 275 〈http://cdsweb.u-strasbg.fr/topbase/topbase. tive transfer in non-LTE diatomic molecules of CO2
N2 plasmas. In:
html〉. Proceedings of the 14th international heat transfer conference, no.
[21] Lamet JM, Babou Y, Rivière Ph, Perrin MY, Soufiani A. Radiative IHTC14-22301, Washington, USA. 2010.
transfer in gases under thermal and chemical nonequilibrium con- [48] Rivière Ph. Systematic semi-classical calculations of Stark broad-
ditions: application to Earth atmospheric re-entry. J Quant Spectrosc ening parameters of NI, OI, NII, OII multiplets for modelling the
Radiat Transf 2008;109:235–44. radiative transfer in atmospheric air mixture plasmas. J Quant
[22] Lopez B, Perrin MY, Rivière Ph, Soufiani A. Coupled nonequilibrium Spectrosc Radiat Transf 2002;73:91–110.
flowfield-radiative transfer calculation behind a shock wave. J [49] Lino da Silva M. An adaptive line-by-line statistical model for fast
Thermophys Heat Transf 2013;27(3):404–13. and accurate spectral simulations in low-pressure plasmas. J Quant
[23] Feldick AM, Modest MF, Levin DA. Closely coupled flowfield radia- Spectrosc Radiat Transf 2007;108:106–25.
tion interactions during hypersonic reentry. J Thermophys Heat [50] Cauchon DL. Radiative heating results from the FIRE II flight
Transf 2011;25(4):481–92. experiment at a reentry velocity of 11.4 km/s. NASA TM X-1402,
[24] Johnston CO, Hollis BR, Sutton K. Spectrum modeling for air shock- 1967.
layer radiation at lunar-return conditions. J Spacecr Rocket 2008;45 [51] Cornette ES. Forebody temperature and calorimeter heating rates
(5):865–78. measured during project FIRE II reentry at 11.35 km/s. NASA TM
[25] Denison MK, Webb BW. The spectral-line-based weighted-sum-of- X-1305, 1967.
gray-gases model in nonisothermal nonhomogeneous media. J Heat [52] Caillault L, Walpot L, Magin TE, Bourdon A, Laux CO. Radiative
Transf 1995;117:359–65. heating predictions for Huygens entry. J Geophys Res 2006;111:
[26] Goody R, West R, Chen L, Crisp D. The correlated-k method for E09S90.
radiation calculations in nonhomogeneous atmospheres. J Quant [53] Park C, Jaffe RL, Partridge H. Chemical-kinetic parameters of
Spectrosc Radiat Transf 1989;42:539–50. hyperbolic Earth entry. J Thermophys Heat Transf 2001;15(1):76–90.
[27] Mihalas D, Weibel-Mihalas B. Foundations of radiation hydro- [54] Gökçen T. N2
CH4
Ar chemical kinetic model for simulation of
dynamics. Mineloa: Dover Publications, INC; 1999. Titan atmospheric entry. J Thermophys Heat Transf 2007;21(1):
[28] Rivière P, Soufiani A, Perrin MY, Riad H, Gleizes A. Air mixture 9–18.
radiative property modelling in the temperature range 10,000- [55] Johnston CO. Nonequilibrium shock-layer radiative heating for earth
40,000 K. J Quant Spectrosc Radiat Transf 1996;56(1):29–45. and Titan entry [Ph.D. thesis]. Blacksburg: Virginia Polytechnic
[29] Bansal A, Modest MF. Multiscale part-spectrum k-distribution Institute and State University; 2006.
database for atomic radiation in hypersonic nonequilibrium flows. J [56] Ludwig CB, Malkmus W, Reardon JE, Thomson JAL. Handbook of
Heat Transf 2011;133(12):122701. infrared radiation from combustion gases. Technical report NASA
[30] Bansal A, Modest MF, Levin DA. Multi-scale k-distribution model for SP-3080, 1973.
gas mixtures in hypersonic nonequilibrium flows. J Quant Spectrosc [57] Young SJ. Nonisothermal band model theory. J Quant Spectrosc
Radiat Transf 2011;112:1213–21. Radiat Transf 1977;18:1–28.
[31] Bansal A, Modest MF. Modeling of radiative heat transfer in carbo- [58] Scoggins JB, Soucasse L, Rivière Ph, Soufiani A, Magin TE. An adaptive
naceous atmospheres using k-distribution models. J Thermophys hybrid statistical narrow band model for coupled radiative transfer
Heat Transf 2013;27(2):217–25. in atmospheric entry flows. In: Proceedings of the 8th European
[32] Modest MF. Radiative heat transfer. 3rd ed., Boston: Academic Press; symposium on aerothermodynamics for space vehicles, Lison, Por-
2013. tugal. 2015.
[33] Lamet JM, Rivière Ph, Perrin MY, Soufiani A. Narrow-band model for [59] Laux CO. Optical diagnostics and radiative emission of air plasmas
nonequilibrium air plasma radiation. J Quant Spectrosc Radiat Transf [Ph.D. thesis]. Stanford: Standford University; 1993.
2010;111:87–104. [60] Bose D, Wright MJ, Bogdanoff DW, Raiche GA, Allen Jr. GA. Modeling
[34] Munafò A, Magin TE. Modeling of stagnation-line nonequilibrium and experimental assessment of CN radiation behind a strong shock
flows by means of quantum based collisional models. Phys Fluids wave. J Thermophys Heat Transf 2006;20(2):220–30.
2014;26:097102.