Acta Astronautica 217 (2024) 349–362

Contents lists available at ScienceDirect

Acta Astronautica
journal homepage: www.elsevier.com/locate/actaastro

Research paper

Design space investigations of scramjet engines using reduced-order

modeling
Ali Can Ispir a , Bora O. Cakir b,c ,∗,1 , Bayindir H. Saracoglu b ,1
a von Karman Institute for Fluid Dynamics, Aerospace and Aeronautics Department, Belgium
b
von Karman Institute for Fluid Dynamics, Turbomachinery and Propulsion Department, Belgium
c
Lund University, Department of Energy Sciences, Sweden

ARTICLE INFO ABSTRACT

MSC: Conceptual design studies performed with reduced-order approaches allow feasibility and sizing considerations
00-01 of air-breathing engines to be configured at an affordable computational cost. The present study is devoted to
99-00 exploring the design space of a dual-mode ramjet engine operating in scramjet mode by means of reduced-
Keywords: order analysis to assess the effects of propulsive system design configurations on component level and overall
Hypersonic aircraft performance characteristics. The approach proposed in this work combines axisymmetric flow configuration
Scramjet used for the design of supersonic/hypersonic intakes and solutions of one-dimensional flow governing equations
Conceptual design
coupled with finite-rate chemistry and thermophysical properties tables in the numerical domains of the
Thermal choking
combustor and nozzle components. The scramjet design space is generated by varying parameters which are
High-speed propulsion
flight Mach number and altitude, intake truncation angle, intake exit Mach number and equivalence ratio.
Performance outputs of total pressure recovery factor, compression ratio, captured air mass flow rate, intake
startability index, thrust, specific impulse, fuel consumption and overall efficiency are computed for each design
scenario. The generated database is visualized via performance maps and analyzed in terms of propulsive
characteristics. A feature importance study is also conducted to quantify the effects of design parameters on
the propulsive performance.

1. Introduction and operating conditions to fundamental performance parameters [3,

4]. There are several efforts to establish representative thermodynamic
High-speed airbreathing engines such as ramjets and scramjets, can schemes comprised of simplified governing flow equations coupled
achieve high propulsive efficiency at elevated speeds due to the absence with multiple reduced-order models. In this regard, regenerative cool-
of limitations induced by turbomachinery components and utilizing ing calculations are coupled with 1D reactive flow equations for investi-
the high ram pressure of the incoming air instead of carrying an on- gating temperature and pressure rises in a scramjet engine operating at
board oxidizer storage. The primary difference of scramjet engines over 27 km altitude with Mach 7.7 flight speed and, the effect of heat trans-
ramjets is that the flow maintains its supersonic character throughout fer taking place between the supersonic flow and combustor cooling
the entire engine whereas, in ramjet engines, the incoming air flow channels on the overall performance [5]. Torrez et al. used a reduced-
is slowed down to subsonic conditions before the fuel injection. Even order model coupling one-dimensional governing equations with finite-
though a scramjet engine is composed of a reduced number of dynamic rate chemistry approach and similar to analyze pre-combustion shock
mechanical elements, there are multiple phenomena which contribute and dissociation effects in an ethylene fueled scramjet engine [6]. This
to the flow complexity inside the propulsive duct and they must be investigation involved examining the shock train structure through
investigated in detail to ensure the system reliability and mission the isolator and combustor in the ram, early scram, and late scram
integrity [1]. modes observed during the acceleration phase of the flight. Addition-
In this regard, system level design approaches, reduced-order mod- ally, Birzer and Doolan utilized a one-dimensional ordinary differential
els and preliminary performance analysis are considered cost-effective equation (ODE) set to model a hydrogen-fueled scramjet combustor and
solutions as compared to high resolution numerical and/or experimen- validated the proposed methodology with experimental data obtained
tal analysis [2] to characterize the influence varying design constraints from Hyshot-II test campaigns [7].

∗ Corresponding author.
E-mail address: [email protected] (B.O. Cakir).
1
Waterloosesteenweg 72, Rhode-Saint-Genese, B-1640, Belgium.

https://doi.org/10.1016/j.actaastro.2024.01.036
Received 8 December 2023; Received in revised form 25 January 2024; Accepted 28 January 2024
Available online 29 January 2024
0094-5765/© 2024 The Authors. Published by Elsevier Ltd on behalf of IAA. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).
A.C. Ispir et al. Acta Astronautica 217 (2024) 349–362

the heat addition decelerates the supersonic flow as well as the com-
Nomenclature
pression when it passes through the propulsive duct, the Mach number
A Area [m2 ] can possibly drop under the sonic condition. This phenomenon, referred
𝐶𝑓 Skin friction coefficient to as thermal choking [10], arises issues of stability and performance
𝐶𝑝 Heat capacity at constant pressure [J kg−1 degradation especially in scramjets which are commonly not equipped
K−1 ] with a convergent channel downstream of the burner to re-accelerate
𝐶𝑣 Heat capacity at constant volume [J kg−1 the subsonic flow back to supersonic conditions. Accordingly, there
K−1 ] are several attempts to offer a robust methodology for ramjet engine
CPR Compression pressure ratio design to detect thermal choking position along propulsive duct for
𝐷 Hydraulic diameter [m] given operational and flight conditions [8,11,12]. As the flow in ramjet
ER Equivalence Ratio engines modulates between different flow regimes, the location of
F𝑢 Uninstalled thrust thermal choking is preferred to be near the inlet of diverged noz-
I𝑠𝑝 Specific impulse [s] zle [13]. However, design variables in a scramjet engine such that
JANAF Joint Army-Navy-Air Force the air flow through the propulsive duct the kept far away from the
𝑚̇ Mass flow rate [kg s−1 ] thermal choking limit (particularly in the combustion chamber). In
M Mach number or Reaction number the design space analysis of scramjet propulsion systems, captured air
mass flow rate and fuel injection which describe engine specifications
𝑀𝑊 Molecular weight [kg mol−1 ]
are mainly optimized to meet flight trajectory requirements. This is
P Pressure [Pa]
done by balancing propulsive performance in terms of specific impulse
𝑃𝑟 Prandtl number
and intake performance outputs such as compression ratio, startability
𝑄 Reaction progress rate
index and total pressure recovery factor. Nevertheless, an important
𝑅 Gas constant [J K −1 kg−1 ]
factor limiting temperature rise, hence the thrust production in scramjet
𝑅𝑢 Universal gas constant [kg mol−1 K −1 ]
burner is related to thermal choking phenomenon [14]. Thus, the
SHAP Shapley Additive Explanations
investigating the impact of the design space of a scramjet engine on
SUNDIALS SUite of Nonlinear and DIf-
thermal choking is essential.
ferential/ALgebraic equation
Moreover, among the scramjet engine parameters, operating flight
Solvers
conditions not only change thermal choking location and possible
T Temperature [K]
occurrence, but also the intake flow topology and performance which
TR Temperature ratio
must establish suitable pre-combustion flow conditions to accommo-
TPR Total pressure recovery factor
date auto-ignition in the burner and overall system performance [15–
U Velocity [m/s]
17]. Comprehending the effects of scramjet engine design parameters
𝑋 Molar concentration [mol m−3 ]
on each system component and the overall performance by mapping
Y Species Mass Fraction
the propulsive outputs to detect thrust production capabilities is re-
Greek symbols stricted mainly due to thermal choking for given design parameters.
This requires a detailed design space investigation performed by com-
𝛽 Truncation angle
putationally affordable techniques (e.g. ones established by appropriate
𝜖 Flow direction parameter
reduced-order approaches). Thus, there is a lack of comparative assess-
𝛾 Specific heat ratio
ments of the flight properties and propulsive system design parameters
𝜌 Density [kg/m3 ] for scramjet engine which the current study aims to address.
𝜔̇ Molar production rate [mol m−3 s−1 ]
Therefore, a thorough investigation is proposed in order to ex-
Subscripts plore the design space of a dual mode ramjet engine in scramjet
operation. Axisymmetric flow templates are utilized to configure the
0 Stagnation
scramjet intake for given flight conditions. The corresponding intake
𝐶, 𝐼 Combustor inlet
design parameters and one-dimensional flow equations coupled with
𝑖 Species index
detailed-chemistry closure are solved for the combustor and nozzle
𝐼, 𝐸 Inlet exit
parts. A design database is created including five independent design
𝑤 Wall variables: intake truncation angle and intake exit Mach number, fuel–
air equivalence ratio for the combustor, and flight conditions which
are altitude and Mach number. The propulsive performance parameters
Moreover, Cakir et al. [8] proposed a holistic design and analysis which are uninstalled thrust, specific impulse, fuel consumption, left-
tool which combines one-dimensional solutions of the flow devel- over fuel mass fraction, and overall efficiency in each design scenarios
opment within the combustor and nozzle components in order to are computed. Moreover, intake design space for scramjet engines
investigate performance impact of ramjet intakes. The introduced ap- is evaluated by assessing total pressure recovery, captured air mass
proach is utilized to investigate the impacts of intake design variables flow rate, startability index and compression ratio. Accordingly, perfor-
such as exit Mach number and truncation angle on the ramjet propul- mance maps of uninstalled thrust with respect to the specific impulsive
sive performance. In another work studying intake design parameters, and propulsive efficiency are extracted to illustrate the design envelope
flow turning and shock angles of a supersonic intake were analyzed constrained by thermal chocking and non-ignition in terms of intake
via multi-dimensional numerical approaches to achieve shock-on-lip configuration parameters and flight conditions. Finally, discussions
condition which promises maximum intake efficiency and mass capture regarding the influence of supersonic combustion dynamics on the
area with minimum duct length [9]. On the other hand, one of the realizable propulsive domain are provided and supported by a feature
major challenges in the scramjet engine design process is related to importance study performed with the Shapley Additive Explanations
maintaining the Mach number along the flow path above unity. Since (SHAP) framework [18].

350
A.C. Ispir et al. Acta Astronautica 217 (2024) 349–362

Fig. 1. Schematics for the internal compression flow along a Busemann diffuser.

2. Methodology turn relate to the local temperature.

𝑟𝑑𝑉 𝜃 𝑑𝑉 ⎛ ⎞
2.1. Intake design module 𝑑𝜌 2 𝑉𝑟 𝑑𝜃 + 𝑉𝜃 𝑑𝜃 ⎜ 𝛾𝑝 𝛩 ⎟
𝑑𝜃
=−
𝛾 𝑉2 −𝑉2−𝑉2 ⎜𝛾 − 1 − ( 𝛩 )⎟ (5)
𝑚𝑎𝑥 𝑟 𝜃 ⎜ 𝑝 𝑇 𝑒𝑇 − 1 ⎟
Modern hypersonic cruiser designs are based on generic waverider ⎝ ⎠
geometries which necessitate a propulsion system that is integrated to
Afterwards, combining the mass and momentum conservation equa-
the fuselage of the vehicle. This necessity puts a strict constraint on
tions which are both represented in terms of local temperature and
the flow compatibility between the external geometry of the vehicles
velocity components, one obtains the T-M equations for calorically
and the intake flow field. For this purpose, axisymmetric flow fields
imperfect gas as Eq. (6).
are demonstrated to meet the requirements induced by structural com-
𝛾 2 𝑑𝑉𝜃 𝑑𝑉 𝑑𝑉
ponents as well as the combustion process downstream [19]. In this (𝑉 − 𝑉𝑟2 − 𝑉𝜃2 )(2𝑉𝑟 + 𝑉𝜃 𝑐𝑜𝑡(𝜃) + ) − 𝑉𝜃 (𝑉𝑟 𝑟 + 𝑉𝜃 𝜃 )
regard, the intake contouring is determined based on the axisymmetric 2 𝑚𝑎𝑥 𝑑𝜃 𝑑𝜃 𝑑𝜃
𝛾𝑝 𝛩
flow templates of a Busemann diffuser [20] which is of high popularity ×( − ( 𝛩 )) = 0 (6)
𝛾𝑝 − 1
among various propulsion system design studies [21,22] (see Fig. 1). 𝑇 𝑒𝑇 − 1
The Busemann diffuser is a streamtube representation of three
dimensional axisymmetric conical flow field [23] whose behavior is Finally, in order to formulate the equation set that is going to be
described by means of Taylor–Maccoll equations [24]. Nevertheless, solved numerically in combination with the conically symmetric flow
the original T–M equations are described for calorically perfect ideal conditions where the flow properties are only varying with the angular
𝑑𝑉
position (𝑉𝜃 = 𝑑𝜃𝑟 ), a velocity of 𝑣 = 𝑉 ∕𝑉𝑚𝑎𝑥 is defined that is
gases whose validity is prone to violations in consideration of the √
high Mach numbers scramjets are designated to fly at. Hence, it is represented as 𝑣 = 𝑣2𝑟 + 𝑣2𝜃 . Hence, deriving the ordinary differential
necessary to derive the T-M equations to achieve validity of the de- equation for the angular velocity,
scribed flow fields of the intakes throughout the entire design space
⎛ ⎞ ( )
range of scramjet engines. Accordingly, Busemann flow is comprised of 𝑣𝑟 𝑣2𝜃 ⎜ 𝛾 −1
𝛾𝑝
− ( 𝛩 )⎟ −
𝛩 𝛾
1 − 𝑣2𝑟 − 𝑣2𝜃 (2𝑣𝑟 + 𝑣𝜃 𝑐𝑜𝑡(𝜃))
conical axisymmetric isentropic conditions which requires the entropy ⎜ 𝑝 𝑇 𝑒 𝑇 −1 ⎟
2
𝑑𝑣𝜃 ⎝ ⎠
and total enthalpy to be preserved (𝛥𝑆 = 0 and 𝛥ℎ0 = 0). = (7)
𝑑𝜃 ⎛ ⎞
2
− 𝑣2𝜃 ) − 𝑣2𝜃 ⎜ 𝛾 −1
𝛾𝑝
𝑉𝑚𝑎𝑥 ( 𝛩 )⎟
𝛾 𝛩
𝑉2 (1 − 𝑣2𝑟 −
ℎ0 = =ℎ+ (1) 2 ⎜ 𝑝 𝑇 𝑒 𝑇 −1 ⎟
2 2 ⎝ ⎠
Here, 𝑉𝑚𝑎𝑥 refers to the maximum velocity a flow can be expanded
Nevertheless, the dependency on local temperature at each step
adiabatically to absolute zero temperature (𝑇 = 0 K) whilst the local
of the angular integration procedure necessitates an iterative process
static enthalpy of a flow at temperature 𝑇 is given by Eq. (2).
to determine temperature and temperature dependent variables which
⎡ ⎤ dictates the chemical properties of flow. Thus, utilizing the isentropic
𝑎2 ⎢ 𝛾𝑝 𝛩 ⎥ flow equations for calorically imperfect gas, the velocity can also be
ℎ=
𝛾 ⎢𝛾 − 1 − ( 𝛩 )⎥ (2)
⎢ 𝑝 𝑇 𝑒𝑇 − 1 ⎥ defined as a function of temperature in which total temperature (𝑇𝑡 ),
⎣ ⎦ gas constant (𝑅) and the ratio of specific heats for calorically perfect
Then, expressing
√ local velocity in terms of radial and angular compo- gas (𝛾𝑝 ) are constant for air.
nents (𝑉 = 𝑉𝜃2 + 𝑉𝑟2 ), Eq. (2) can be recast. [ ( ) ( )]
2𝑇 𝛾𝑝 𝑇 𝛩 1 1
𝑣2 = 𝑡 1− + 𝛩
− 𝛩
(8)
⎡ ⎤ 𝑅 𝛾𝑝 − 1 𝑇𝑡 𝑇𝑡
𝑒 𝑇𝑡 − 1 𝑒 𝑇 − 1
1 2 𝑎2 ⎢ 𝛾𝑝 𝛩 ⎥
(𝑉𝑚𝑎𝑥 − (𝑉𝜃2 + 𝑉𝑟2 )) = ( 𝛩 )⎥ (3)
𝛾 ⎢⎢ 𝛾𝑝 − 1
− Throughout the integration procedure, the local velocity compo-
2
𝑇 𝑒𝑇 − 1 ⎥ nents are determined by the solutions of a numerical integration
⎣ ⎦
scheme of Runge–Kutta 4 (RK4) [26]. The integration of the streamlines
Thus, the formulation for local velocity components is given in Eq. (3)
outlining a Busemann flow field is initiated by flow characteristics
as a function of temperature since 𝛾 = 𝛾(𝑇 ) and 𝑎 = 𝑎(𝑇 ) which are
influenced by the strength of the terminating conical shock. The in-
utilized in the Euler equations providing the relations of mass and
tegration direction spans from the exit of the intake, representing the
momentum conservation in conical coordinates given in Eq. (4) [25].
most downstream flow condition within the Busemann flow field. At
𝑑𝑉𝜃 𝑑𝜌 each step of angular integration, the velocity components are computed
2𝜌𝑉𝑟 + 𝜌𝑉𝜃 𝑐𝑜𝑡(𝜃) + 𝜌 + 𝑉𝜃 =0
𝑑𝜃 𝑑𝜃 utilizing the value of the upstream temperature to begin with. Then,
(4)
1 𝑑𝜌 −1 𝑑𝑉 𝑑𝑉
= (𝑉𝑟 𝑟 + 𝑉𝜃 𝜃 ) Eq. (8) is solved for temperature with known radial and angular
𝜌 𝑑𝜃 𝑎2 𝑑𝜃 𝑑𝜃 velocity components. Therefore, an iterative process of minimizing the
Hence, by substituting Eq. (3) into the momentum conservation equa- difference between temperature value inputted in Eq. (7) and the one
tion provided in Eq. (4), variation of density along the angular coor- computed from Eq. (8) is performed until the converge criteria of
dinate can be obtained in terms of local velocity components which in 𝜖𝛥𝑇 < 10−4 is met.

351
A.C. Ispir et al. Acta Astronautica 217 (2024) 349–362

Although the Busemann intakes are demonstrated to be highly

1 𝑑𝑝 1 𝑑𝜌 1 𝑑𝑇 ̄
1 𝑑 𝑀𝑊
efficient for compression of an incoming hypersonic stream owing to = + − (13)
high total pressure recovery [27,28], they are associated with inferior 𝑝 𝑑𝑥 𝜌 𝑑𝑥 𝑇 𝑑𝑥 ̄
𝑀𝑊 𝑑𝑥
starting characteristics with respect to the flow deflection based intake To solve numerically flow variables which are density, velocity,
designs [29]. Additionally, they tend to have excessive length scales and temperature the governing equations set, Eqs. (9)–(13) can be
in order to achieve perfect flow alignment with the incoming flow converted to the following forms:
at the leading edge [30,31]. Nonetheless, a significant portion of this 1 𝑑𝜌 1 𝑑 𝑚̇ 1 𝑑𝑈 1 𝑑𝐴
length does not contribute to the compression process while yielding an = − − (14)
𝜌 𝑑𝑥 𝑚̇ 𝑑𝑥 𝑈 𝑑𝑥 𝐴 𝑑𝑥
increase in viscous drag and weight. In order to resolve these issues and
2
establish an acceptable trade-off between airframe integration and per- 1 𝑑𝑈 −1 1 𝑑𝑝 2𝛾𝑀 𝐶𝑓 𝛾𝑀 2 (1 − 𝜖) 𝑑 𝑚̇
= ( + + ) (15)
formance, leading edge truncation is proven to be an effective tool [32]. 𝑈 𝑑𝑥 𝛾𝑀 2 𝑝 𝑑𝑥 𝐷 𝑚̇ 𝑑𝑥
While there are various approaches to perform leading edge truncation ( ( )
𝑑𝑇 1 ∑ 𝑑𝑌 1 ∑ ( 𝑑 𝑚̇ 𝑖,𝑎𝑑𝑑𝑒𝑑 ) ℎ 𝑑𝑚
and flow topology optimization [33], Tognelli et al. demonstrated that = ∗ − (ℎ𝑖 𝑖 ) + [ ℎ𝑖 − 0 ]−
𝑑𝑥 𝐶𝑝 𝑑𝑥 𝑚̇ 𝑑𝑥 𝑚̇ 𝑑𝑥
the intake design can be performed directly in consideration of a 𝑖 𝑖
) (16)
prescribed truncation angle to achieve compatibility of flow properties 2𝐶𝑓 𝐶𝑝 (𝑇𝑎𝑤 − 𝑇𝑤 ) 𝑑𝑈
−𝑈
and flow alignment a priori [34]. This prevents the performance degra- 𝑃 𝑟2∕3 𝐷𝐴 𝑑𝑥
dation due to disruption of the isentropic compression field induced by The pressure is solved via ideal gas relation given in Eq. (13).
a purely geometric truncation strategy [35].
The change of molecular weight of the mixture in differential form
In consideration of implementing the truncation methodology, the
is given as Eq. (17):
flow field behavior composing the intake shape can be divided into two
main regions as an oblique shock wave induced by a truncation angle 𝑑𝑀𝑊 2∑ 1 𝑑𝑌𝑖
= −𝑀𝑊 ( ) (17)
(𝛽) and the isentropic compression field entrained by the Busemann 𝑑𝑥 𝑖
𝑀𝑊 𝑖 𝑑𝑥
contours. Hence, while the variation of flow properties across the 𝑑 𝑚̇
𝑖,𝑎𝑑𝑑𝑒𝑑
oblique shock is computed from well known equations [36], impact The mass gradient term, 𝑑𝑥
, signifies mass addition, specifi-
of the compression field constructed via the Busemann contours is cally fuel injection into the system duct. However, given the consider-
quantified by means of isentropic relations [25]. Thus, depending on able length of the combustor duct in the investigated scramjet engine,
the truncation angle of choice, first the post-shock flow properties it is assumed that the fuel is perfectly mixed with the air stream at the
including the variation in flow direction is computed. Then, this is burner entrance. Consequently, this term is considered to be zero in the
used as a design constraint to be met by the Busemann contours as solutions.
the integration of the streamline is propagated upstream starting from Five equations of the main flow variables (Eqs. (14)–(17)) cou-
the intake exit. Every intake geometry represents a unique solution for pled with the species equation (Eq. (10)) of each species in the sys-
which the terminating shock intensity is the eigenvalue constrained by tem are solved numerically. The reaction rates of each species, 𝜔̇ 𝑘 in
the intake exit Mach number, intake inlet Mach number, and intake Eq. (10) are calculated by the finite-rate chemistry approach as given
inlet flow angle. in Eqs. (18)–(20) [37]. The progress rate for reaction j is denoted by
Q𝑗 where 𝜈𝑘𝑗 ′ and 𝜈 ′′ represent the molar stoichiometric coefficients
𝑘𝑗
2.2. Combustor and nozzle design module of species k in reaction j. The reaction rate 𝜔̇ 𝑘 for species k is the
summation of rates 𝜔̇ 𝑘𝑗 produced by all M reactions:
The intake design procedure detailed above is linked with the
combustor and nozzle design module outlined in this section. After the ∑
𝑀 ∑
𝑀
𝜔̇ 𝑘 = 𝜔̇ 𝑘𝑗 = 𝑀𝑊𝑘 𝜈𝑘𝑗 𝑄𝑗 (18)
compression along intake diffuser, high-speed air stream in a scramjet 𝑗=1 𝑗=1
engine is mixed with hydrogen in the combustion chamber and the
reactions is anticipated to take place when the temperature of mixture where
reaches auto-ignition limit. To numerically describe the flow in the ′′
𝜈𝑘𝑗 = 𝜈𝑘𝑗 ′
− 𝜈𝑘𝑗 (19)
combustor and nozzle, a comparable methodology is employed based
on a quasi-one-dimensional set of governing equations coupled with
∏
𝑁 ∏
𝑁
detailed-chemistry relations, as outlined by Zhang et al. [5]. The values ′ ′′
𝑄𝑗 = 𝐾𝑓 𝑗 [𝑋𝑘 ]𝜈𝑘𝑗 − 𝐾𝑟𝑗 [𝑋𝑘 ]𝜈𝑘𝑗 (20)
of all flow variables and the geometric area of the combustor and 𝑘=1 𝑘=1
nozzle sections vary as functions of the axial distance. The flow is K𝑓 𝑗 and K𝑟𝑗 represent the forward and reverse rates of reaction
assumed to be steady-state and behave as an ideal gas. The scramjet j, respectively. They are computed using the Arrhenius approach, as
combustor and nozzle walls are assumed to be adiabatic, thus there outlined in the Eqs. (21) and (22).
is no heat transfer between the flow and duct walls. The intake exit ( )
𝐸𝑗
flow conditions computed through intake design and analysis module 𝐾𝑓 ,𝑗 = 𝐴𝑓 ,𝑗 𝑇 𝛽𝑗 exp − (21)
detailed above are defined as the inlet conditions for the solutions 𝑅𝑇
of the combustor and nozzle components. Accordingly, the governing 𝐾𝑓 ,𝑗
set of equations which are for conservation mass, species, momentum, 𝐾𝑟,𝑗 =
∑𝑁
( 0 ) (22)
𝑝 𝛥𝑆 𝛥𝐻𝑗0
( 𝑅𝑇𝑎 ) 𝑘=1 𝜈𝑘𝑗 exp 𝑅𝑗 −
energy, and the ideal gas equation are given below, respectively: 𝑅𝑇
1 𝑑 𝑚̇ 1 𝑑𝜌 1 𝑑𝑈 1 𝑑𝐴
= + + (9) here 𝑃𝑎 is 1 bar. 𝛥𝑆𝑗0 and 𝛥𝐻𝑗0 are entropy and enthalpy changes of the
𝑚̇ 𝑑𝑥 𝜌 𝑑𝑥 𝑈 𝑑𝑥 𝐴 𝑑𝑥
reaction 𝑗.
𝑑𝑌𝑖 𝜔̇ 𝑀𝑊𝑖 𝐴 1 𝑑 𝑚̇ 𝑖,𝑎𝑑𝑑𝑒𝑑 𝑌 𝑑 𝑚̇ In the present study, the walls of the scramjet combustor and nozzle
= 𝑖 + − 𝑖 (10)
𝑑𝑥 𝑚̇ 𝑚̇ 𝑑𝑥 𝑚̇ 𝑑𝑥 were considered adiabatic. Consequently, the second-to-last term on the
2𝛾𝑀 2 𝐶𝑓 right side of Eq. (16) is set to zero. The following equations provide the
1 𝑑𝑝 𝛾𝑀 2 𝑑𝑈 𝛾𝑀 2 (1 − 𝜖) 𝑑 𝑚̇
+ + + =0 (11) specific stagnation enthalpy (ℎ0 ), the specific static enthalpy (ℎ), and
𝑝 𝑑𝑥 𝑈 𝑑𝑥 𝐷 𝑚̇ 𝑑𝑥
∑ the specific static enthalpy of the 𝑖𝑡ℎ species (ℎ𝑖 ).
𝑑ℎ 1 𝑑[ 𝑖 ℎ0,𝑖 𝑚̇ 𝑖 ]𝑎𝑑𝑑𝑒𝑑 ℎ 𝑑𝑚 𝑑𝑈
= − 0 −𝑈 (12) ℎ0 = ℎ + 𝑈 2 ∕2 (23)
𝑑𝑥 𝑚̇ 𝑑𝑥 𝑚̇ 𝑑𝑥 𝑑𝑥

352
A.C. Ispir et al. Acta Astronautica 217 (2024) 349–362

∑
ℎ= ℎ𝑖 𝑌𝑖 (24) According to this curve, the supersonic flow along the scramjet engine
𝑖 path is expected to decelerate and approach the sonic limit, where
𝑇 the Mach number reaches unity with the presence of the reaction (in
ℎ𝑖 = ℎ0𝑓 ,𝑖 + 𝐶𝑝,𝑖 𝑑𝑇 (25) case of thermal choking). However, transitioning between the regimes
∫298 K
(supersonic and subsonic) would require entropy reduction which is
ℎ0𝑓 ,𝑖 represents the heat of formation of the 𝑖𝑡ℎ species, and the non-physical.
heat capacity of all species in the gas mixture is defined as provided
in Eq. (26). 2.3. Validation
∑
𝐶𝑝 = 𝐶𝑝,𝑖 𝑌𝑖 (26)
𝑖 The verification of the intake design and analysis module of the full
In each numerical step, the flow Mach number is updated with propulsive path investigation methodology introduced in this study is
the local speed of sound which is a function of the temperature and performed by means of comparative numerical simulations of higher
molecular weight and given in Eq. (27): fidelity and the capability of the design approach to represent physi-
cally realizable intakes is demonstrated in comparison to experimental
𝑈
𝑀= √ (27) studies [8]. The design parameters for the intake contours are provided
𝛾𝑅𝑇 as boundary conditions (𝑀∞ = 5 at 25 km of altitude with 𝑀𝐼,𝐸 =
where R is gas constant and 𝛾 3.5) to the numerical simulations directly to assess the validity of the
𝐶𝑝 streamtraced contours in terms of providing the desired flow field prop-
𝛾= (28) erties (Fig. 3, top). The numerical investigations were composed of a
𝐶𝑣
mesh convergence analysis over which a mesh convergence history not
𝑅 = 𝑅𝑢 ∕𝑀𝑊 (29) only by means of relative errors between the simulations but also with
respect to the performance parameters indicated by the reduced order
A system of ordinary differential equations (ODEs) is formulated analysis is extracted. All simulations yielded results with a maximum
based of the transport equations (mass, momentum, energy and of ∼1% deviation from the reduced order predicted values in terms of
species), detailed chemistry described through the combustion mech- intake exit Mach number (𝑀𝐶,𝐼 ), mass flow rate (𝑚̇ 𝑎𝑖𝑟 ), total pressure
anism of Jachimowski (involving 33 reactions among 13 species) [38], recovery (TPR), compression pressure ratio (CPR), and temperature
and Arrhenius temperature dependent reaction rates. JANAF tables are ratio (TR) (Fig. 3).
utilized for extraction of thermochemical properties for each specie [39]. The validation of the combustor and nozzle analysis module is
A stiff ODE solver capable of operating at various time and/or spatial conducted by comparing the predicted flow properties with the experi-
scales is needed to solve the set of equations in the current work. mental and numerical data provided by Torrez et al. [41]. Throughout
Following the suggestion of Zhang et al. [5], the SUNDIALS (Suite the experiments conducted on a dual-mode ramjet combustor, the
of Nonlinear and Differential/Algebraic Equation Solvers) code [40], facility is operated in scramjet (supersonic combustion) mode during
developed and published by Lawrence Livermore National Laboratory, which pressure measurements along the combustor and isolator sec-
was preferred for an iterative solution. Among the six solvers provided tions are acquired. The corresponding experimental data is utilized
in the code, the CVODE solver, accessible in the MATLAB environment for the validation studies of the MASIV code, developed for solving
for Sundials version 2.6.2, was selected for its capability to handle flow within ramjet and scramjet engines using various reduced-order
stiff or non-stiff ODE systems in initial value problems. To address approaches. Additionally, higher order numerical investigations are
the mathematical stiffness problem, primarily attributed to the reactive performed using the Ansys Fluent commercial code for comparison with
part of the ODE solution, a nonlinear Newtonian solver with backward the MASIV results. The results obtained with MASIV code, Fluent and
differencing approximations was employed. In this time-independent the experimental campaigns are compared against the ones provided
problem, the numerical steps along the axial axis of the duct vary by the methodology of the present study under similar boundary and
depending on the local stiffness. In other words, the numerical step can operating conditions.
even be set to 10−10 m during the combustion to handle the solutions As illustrated in Fig. 4, the decrease in temperature during the mix-
of the detail chemistry relations. On the other hand, 0.01 meter mesh ing process (between 0.35 and 0.36 m) is attributed to the injection of
size is sufficient to solve the expansion along the nozzle as displayed hydrogen fuel into the air stream with a lower static temperature. The
in Fig. 2. heat exchange between the two streams results in a significant reduc-
For the initial value problem addressed in the current study and tion in the pre-combustion flow temperature. In solving the combustor
presented in the 𝑓 (𝑥, 𝑍) form as described above, a solution always and nozzle modules for this scenario, the final mixture temperature
exists, unless there is a mathematical singularity. Here, 𝑥 represents before combustion initiation was calculated through 0D heat transfer
axial distance, while 𝑍 represents the flow variables for which we are computations, considering specific heats and mass flow rates. The sim-
seeking solutions at each spatial step. The equation system is solved ulations in this study operated under the assumption that the hydrogen
within a continuous axial interval extending from the combustor inlet fuel is heated to the temperature of the incoming air stream before
to the nozzle outlet section in the scramjet engine. Moreover, for each mixing. Consequently, there is no heat transfer occurring between the
simulation in the design space of the scramjet engine, there is only fuel and the incoming air stream into the combustor. Although any
one solution that satisfies the ODE set which is unique to the initial fuel–air mixing model is not defined in the combustor module, the
conditions provided by an intake geometry and operational parameters results of pressure drop, Mach number increase and temperature rise
(such as ER). In the case of thermal choking, there is a singularity were perfectly matched with the ones of the MASIV and significant
point achieved when the sonic condition is reached. However, until agreement is denoted with the experimental data [8]. The comparisons
the axial location where local Mach number equals unity the ODE of flow pressure evolution in the nozzle also yielded good agreement
set is solved with the given initial conditions. Thus, combination of with the experimental results.
design and operational conditions that yield thermal choking within the
combustor duct are characterized for determination of the net positive 3. Results
thrust generation envelope.
In addition, the Rayleigh flow curve characterizes the temperature The simulations for investigating the design space of the scramjet
change and entropy generation of a non-adiabatic flow through a engine including combustor and nozzle parts can be divided into three
constant area duct where heat addition or extraction is considered. distinctive groups according to the observed engine phenomenon. First

353
A.C. Ispir et al. Acta Astronautica 217 (2024) 349–362

Fig. 2. Varying spatial mesh resolution along the combustor and nozzle depending on the local mathematical stiffness of the governing set of equations.

Table 1
The design variable ranges for the scramjet engine design space analysis.
Variables Operational range
Cruise Mach number (𝜙1 ) 6, 7, 8
Cruise altitude [km] (𝜙2 ) 25, 30, 35
Variables Design range
Equivalence ratio (𝜙3 ) 0.1 ≤ 𝐸𝑅 ≤ 0.7
Truncation angle (𝜙4 ) 0≤𝛽≤8
Pre-combustion Mach number (𝜙5 ) 1.1 ≤ 𝑀𝐼,𝐸 ≤ 3

(𝑚̇ 𝑓 𝑢𝑒𝑙 ), and unburnt hydrogen mass fraction are computed by means
of the modules detailed above with given design inputs (Fig. 5). The
generated scramjet engine data including 9650 design scenarios paves
the way for comprehensive discussions regarding the impacts of the
input parameters on the intake and engine performance (Section 3.3).

3.1. Feature importance study

This section presents a feature importance study which is conducted

using the Shapley Additive Explanations (SHAP) framework [18]. This
analysis is valuable to enhance the interpretation of the conclusions
drawn from the scramjet flow path and gain a deeper understanding of
the scramjet propulsive performance concerning specific design vari-
Fig. 3. The computational setup of CFD simulations for validation of the intake design ables by quantifying the individual impact of design variables on each
module (1st row). Comparison of Mach number (2nd and 3rd rows) and compression
objective function. The average marginal contribution, also known as
pressure ratio (4th and 5th rows) evolution over the contours of the Busemann intake
obtained from the design module in comparison to the CFD simulations. the SHAP value, for each design variable (including flight altitude,
flight Mach number, intake exit Mach number, intake truncation angle,
and fuel–air equivalence ratio (ER)) is calculated across all potential
design configurations using Eq. (30).
group is where thermal choking is detected especially because of a
fuel injection with an extreme equivalence ratio (ER), whilst the other ∑ |𝑆|!(|𝛷| − |𝑆| − 1)!
𝛹𝑖 = [𝜑𝑆∪𝑖 (𝑎𝑆∪𝑖 ) − 𝜑𝑆 (𝑎𝑆 )] (30)
group is able to produce significant amounts of thrust, and ones where 𝑆⊆𝛷∕{𝑖}
|𝛷|!
no reaction occurs. One simulation in the second group takes approxi-
mately 0.1 h with a single core of a stand-alone desktop computer with In Eq. (30), 𝜑𝑆∪𝑖 represents the trained model. To compute the
specifications of AMD Ryzen 7 2700X Eight-Core Processor 3.70 GHz impact of the 𝑖th design variable on the investigated objective function,
64 GB RAM whereas other types of simulations generally take less another model, 𝜑𝑆 , is defined by excluding the 𝑖th design variable.
computational time. Predictions coming from these two trained models are compared with
The design space of the scramjet engine under investigation was the input, 𝜑𝑆∪𝑖 (𝑎𝑆∪𝑖 ) − 𝜑𝑆 (𝑎𝑆 ), where 𝑎𝑆 represents the inputs of a
generated with input parameters shown in Table 1. The intake trun- feature in the subset, 𝑆. As each design variable’s contribution is con-
cation angle 𝛽, flight properties (altitude and 𝑀∞ , combustor inlet tingent on the presence of the remaining design variables, the variances
Mach number (𝑀𝐶,𝐼 ), and equivalence ratio (ER) are selected as design are computed across all conceivable subsets, excluding the 𝑖th design
variables. The objective functions which are total pressure recovery variable, 𝑆 ⊆ 𝛷∕{𝑖}.
(TPR) at the intake, captured air mass flow rate (𝑚̇ 𝑎𝑖𝑟 ), startability Within the SHAP framework, the contributions denoted by 𝛹𝑖 quan-
index, uninstalled thrust (𝐹𝑢 ), specific impulse (𝐼𝑠𝑝 ), fuel consumption tify the effect of a particular design variable on the objective functions.

354
A.C. Ispir et al. Acta Astronautica 217 (2024) 349–362

Fig. 4. Details regarding the test facility employed for validation, along with pressure, temperature, and Mach number comparisons with experimental and numerical results in
both the combustor and nozzle components, are illustrated in frames C.1 and C.2, respectively.

Fig. 5. A generic schematic of the scramjet flow path analysis including flight conditions (denoted by blue) and design variables (denoted by green) and performance parameters
(denoted by red) of the scramjet engines under investigation. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this
article.)

355
A.C. Ispir et al. Acta Astronautica 217 (2024) 349–362

Fig. 6. Results of the sensitivity analysis performed by Shapley Additive Explanations (SHAP) framework, gauging the influence of individual design variables on the objective
functions concerning scramjet engine intake performance. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this
article.)

Figs. 6 and 7 demonstrate the SHAP values for the objective func- flow properties extracted at the exit of the intake. The intake exit
tions. The colorbars labeled ‘‘Feature value’’ correspond to each design Mach number (𝑀𝐼,𝐸 ) and the leading edge truncation angle (𝛽) are
variable, ranging from the minimum (labeled ‘‘Low’’) to the maximum the two design parameters dictate the shape of the intakes for a given
(labeled ‘‘High’’). A positive SHAP value means that the increasing cruise Mach number. Accordingly, for constant leading edge truncation
magnitude of a design variable positively impacts the objective func- angle, decreasing 𝑀𝐼,𝐸 increases compression pressure ratio. Although
tion. For example, higher values of the intake truncation angle (𝛿) the total pressure recovery (TPR) reduces with decreasing 𝑀𝐼,𝐸 due
negatively affect the thrust per unit mass flow rate (𝑇 𝑃 𝑅), as evidenced to the stronger terminating conical shock, the unique geometry of the
by the negative SHAP values in Fig. 6. Configuring the scramjet intake Busemann diffuser enables high efficiency compression of the incoming
with a higher truncation angle leads to increased total pressure losses air. This is achieved by minimizing the compression performed through
along the intake duct. Conversely, the startability index is directly the terminating shockwave which is the main source of total pressure
proportional to 𝑀𝐼,𝐸 ; hence, the SHAP value of this design variable loss (Fig. 8). On the other hand, increasing 𝛽 at constant intake exit
is positive for higher startability indexes. Furthermore, this method Mach number decreases CPR due to increasing total pressure loss
allows for the assessment and quantification of the significance of induced by the incidence shockwave attached to the leading edge. The
design variables. It is evident that some design variables do not affect geometric modifications introduced on the design procedure, by means
certain objective functions. For instance, flight altitude only impacts of varying the truncation angle is rather minimal. Thus, the terminal
the determination of the captured air mass flow rate among intake shock intensity is preserved while increasing 𝛽 rises the intensity of
performance outputs. Therefore, the SHAP analysis assigns it a zero the incidence shock which in return reduces TPR. Nevertheless, as a
impact for 𝐶𝑃 𝑅, 𝑇 𝑃 𝑅, and the startability index. Similar arguments portion of the compression is performed at the leading edge already, the
can be made for the equivalence ratio (ER), as it primarily influences deceleration required to slow down the incoming air to the desired pre-
flow and combustion phenomena downstream of the intake. combustion properties is relaxed for the Busemann diffuser component
which improves the startability characteristics of the intakes.
3.2. Intake performance In terms of the influence of varying design parameters on mass flow
rate (𝑚̇ 𝑎𝑖𝑟 ), a strict dominance of CPR is observed. The aforementioned
Performance specifications of the intakes operating at different effect of changing the exit flow properties to meet the demands of com-
cruise conditions and design parameters are analyzed by means of the bustion chamber on the static pressure development over the contours

356
A.C. Ispir et al. Acta Astronautica 217 (2024) 349–362

Fig. 7. Results of the sensitivity analysis performed by Shapley Additive Explanations (SHAP) framework, gauging the influence of individual design variables on the objective
functions concerning the overall performance of the scramjet engine. (For interpretation of the references to color in this figure legend, the reader is referred to the web version
of this article.)

of the intake directly correlates with the resultant 𝑚̇ 𝑎𝑖𝑟 . Elevating the conditions. Additionally, the effect on temperature is purely adiabatic
exit flow velocity in terms of 𝑀𝐼,𝐸 lowers the induced deceleration for varying operating conditions as the total temperature is preserved.
which reduces the level of compression performed by the intake. Even Therefore, the temperature ratio is kept constant as the cruise altitude is
though, static temperature rises with decreasing 𝑀𝐼,𝐸 at the exit of varied. Nonetheless, as the static temperature is altered with changing
the intake, the effect of temperature variation on 𝑚̇ 𝑎𝑖𝑟 is overpowered altitude, the corresponding pre-combustion temperature is modulated
by the static pressure elevation. Hence, decreasing 𝑀𝐼,𝐸 at constant in relation to 𝑇∞ . On the other hand, changing flight Mach number
leading edge truncation angle, yields the 𝑚̇ 𝑎𝑖𝑟 values to rise. Whereas and keeping the intake exit Mach number range constant, alters the
with increasing 𝛽 at constant 𝑀𝐼,𝐸 degrades the mass flow capturing temperature ratio. Hence, the flow at the inlet of the intake (freestream)
capabilities of the intakes which is quantified by a corresponding entails a greater stagnation temperature which is converted to a higher
decrease in 𝑚̇ 𝑎𝑖𝑟 (Fig. 9). static temperature value at the combustion inlet (Fig. 8).
Moving on to the effect of operating conditions in terms of altitude In accordance with the variations of CPR and TR induced by the
and flight Mach number, the intake performance is observed to change changing altitude, the mass flow rate of air entrained by the intake
drastically even with constant design parameters over the designated is drastically reduced as the altitude is increased. There are two main
operational range. Accordingly, increasing altitude is reducing the dy- components affecting the value of the 𝑚̇ 𝑎𝑖𝑟 at constant contraction ratio
namic pressure the intake is exposed to. The ratios of flow properties of the intake operating at constant intake exit Mach number which are
(TPR, CPR, temperature ratio (TR)) at the exit of the intake and at the the local flow density and flow speed. Increasing the altitude lowers
inlet of the combustor do not change with altitude since the intakes the static temperature which decreases the local speed of sound, thus
are designed with a Mach number dependent methodology. Hence, the flow velocity at constant Mach number increases. Moreover, the
variations in the cruise altitude do not introduce any modifications density variations with altitude are mainly governed by the static
on the CPR range achieved by the intake. However, it reduces the pressure as modulations in the range of CPR overweights the ones
mass flow rate through the intake which has a significant impact in temperature ratio. Hence, the resulting effect on 𝑚̇ 𝑎𝑖𝑟 is a rather
on startability. Thus, with increasing mass flow rate while the CPR significant reduction in comparison to lower altitudes. Varying the
range is kept constant, the startability is hindered which confines the flight Mach number considerably alters the startability characteristics,
operable design space of the intakes to provide feasible pre-combustion as the operable range within the investigated design space is shifted

357
A.C. Ispir et al. Acta Astronautica 217 (2024) 349–362

Fig. 8. Performance maps of intakes under investigation illustrating mass flow rate (𝑚̇ 𝑎𝑖𝑟 ) (color contoured), compression pressure ratio (𝐶𝑃 𝑅)(solid lines) and temperature ratio
(𝑇 𝑅)(dashed lines) values for varying leading edge truncation angle (𝛽) and Mach number at combustor inlet (𝑀𝐶,𝐼 ) at different flight Mach number (6 ≤ 𝑀∞ ≤ 8) (from left to
right) and altitude 35–25 km (from top to bottom). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

& 7) and the engine operating maps are extracted to correlate the
response of scramjet engines to the selected design and operational
parameters (Fig. 10).
The design space of the scramjet engines can be described in three
distinctive zones based on the flow phenomena taking place through
the combustor duct. First zone is identified as the non-ignition zone
where the compression at the intake is not strong enough to heat up
the fuel–air mixture to the auto-ignition conditions. Second zone is in
which thermal choking occurs inside the combustion chamber where
the flow decelerates to sonic limit due to the extreme heat addition.
Finally, the third zone is the region of productive performance where
Fig. 9. The range of mass flow rate (𝑚̇ 𝑎𝑖𝑟 ) values for different flight Mach number and
altitude conditions.
the propulsion system can produce favorable thrust with the given
intake and combustor design conditions. The borders of these regions
which have direct connection to the operating characteristics of the
scramjet engine may change depending on the design variables. A
towards higher intake exit Mach numbers with the increasing the flight
careful investigation is essential in order to comprehend the impact
Mach number. Accordingly, the increasing exit Mach number of the
of these independent variables on the system characterization and
intake allows higher CPR design points to appear since the higher
understand the engine capabilities. All three regions could be observed
velocity of the exiting flow allows to overcome larger adverse pressure
in the solutions obtained for Mach 6 and 7 cruise speed. However,
gradients. Referring to the aforementioned discussion on the domi-
thermal choking zone of the engine at Mach 8 flight cruise was not
nance of CPR for determining the 𝑚̇ 𝑎𝑖𝑟 through the intake, increasing
observed within the defined range of the design variables (Fig. 10 and
allowable CPR (by self-startability constraints) enables a greater 𝑚̇ 𝑎𝑖𝑟
Fig. 11).
to be accommodated by the intakes (Fig. 9).
Uninstalled thrust noted by 𝐹𝑢 is one of the most important perfor-
mance outputs which is an indication of the chemical energy conversion
3.3. Propulsive performance scale of the hydrogen fuel into the kinetic energy and pressure forces.
In feasibility analysis and conceptual design exercises, 𝐹𝑢 is the first
The propulsive performance specification of the scramjet engine op- parameter to evaluate the requirements of a flight trajectory. The
eration is assessed by means of uninstalled thrust (𝐹𝑢 ), specific impulse present analysis visualized in Fig. 10 shows that it is a function of the
(𝐼𝑠𝑝 ), fuel consumption (𝑚̇ 𝑓 𝑢𝑒𝑙 ), overall efficiency (𝜂𝑜 ), and unburnt design variables under investigation. The main factor affecting 𝐹𝑢 is the
hydrogen mass fraction (𝑌H2 ). The impact of the design variables on captured 𝑚̇ 𝑎𝑖𝑟 , hence intake design is significantly influences the overall
these parameters is analyzed using the feature important study (Figs. 6 system performance. The 𝑚̇ 𝑎𝑖𝑟 captured into the propulsive duct varies

358
A.C. Ispir et al. Acta Astronautica 217 (2024) 349–362

Fig. 10. Performance maps of scramjet engines under investigation illustrating Uninstalled thrust (𝐹𝑢 )(color contoured), fuel consumption (𝑚̇ 𝑓 𝑢𝑒𝑙 )(solid lines) and specific impulse
(𝐼𝑠𝑝 )(dashed lines) values for varying leading edge truncation angle (𝛽) and Mach number at combustor inlet (𝑀𝐶,𝐼 ) at different flight Mach number (6 ≤ 𝑀∞ ≤ 8) (from left to
right) and altitude 35–25 km (from top to bottom). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

unburnt and consequently the system is not able to produce thrust.

This phenomenon addresses the non-ignition zone in the performance
maps. The investigated flight Mach number has a significant impact on
the boundaries of these non-ignition regions with respect to 𝑀𝐶,𝐼 . In
this regard, it is observed that the system is not capable of providing
favorable initial conditions to the combustion chamber, resulting in
positive thrust generation for combustor inlet Mach numbers beyond
𝑀𝐶,𝐼 = 2.9, 2.5 and 4.2 at the cruise speeds of 𝑀∞ = 6, 7 and 8
respectively.
As depicted in Fig. 10, area of the non-ignition zones in the de-
sign space is strongly correlated with the captured air mass flow rate
Fig. 11. The range of uninstalled thrust (𝐹𝑢 ) values for different flight Mach number
and altitude conditions. (𝑚̇ 𝑎𝑖𝑟 ) which is a function of intake design parameters as well as the
operational conditions such as altitude and flight Mach number. The
engine operating maps derived for Mach 6 indicates that 𝑀𝐶,𝐼 = 2.9
(for 25 and 30 km cruise altitude) results in the mixture exiting the
with the flight altitude and Mach number which causes variation in the
combustor without reacting. Contrarily, the intake designed with the
uninstalled thrust production as observed in Fig. 10.
same intake exit Mach number could provide favorable flow conditions
Although the captured 𝑚̇ 𝑎𝑖𝑟 and accordingly cumulative effects
to the combustor 35 km of cruise altitude. This observation can be
of the flight altitude and Mach number dominate uninstalled thrust
linked to the captured 𝑚̇ 𝑎𝑖𝑟 . In Eq. (10) that is solved for each species in
(Fig. 7), intake exit Mach number (𝑀𝐼,𝐸 ) has great effect on the thrust
the reactor, the decrease of the total mass flow rate in the denominator
calculations as clearly seen in Fig. 10. The direct impact of 𝑀𝐼,𝐸 on part of the first term on the right side of the equation grows the gradient
uninstalled thrust pertains to pre-combustion flow conditions. On the of species 𝑖 and consequently warms up the mixture to the auto-ignition
other hand, the captured air mass flow rate by the supersonic intake levels. Another parameter regarding the size of the non-ignition zone
strongly depends on the 𝑀𝐶,𝐼 as illustrated in Fig. 6. Hence, it is worth is the fuel mass flow rate, 𝑚̇ 𝑓 𝑢𝑒𝑙 . The greater values of ER for the same
to note that the 𝑀𝐶,𝐼 influences the performance by altering the mass intake design parameters results in the increase of the injected fuel flow
flow rate of the air traveling along the propulsive path. The investigated rate linearly as the flow properties at the combustion chamber inlet
𝑀𝐶,𝐼 ranges differently for three flight Mach numbers because the remain unchanged. Moreover, the enthalpy of the mixture increases
deceleration rate along the supersonic intake varies with the flight with addition of the hydrogen. Even if a compression of the air-stream
speed as mentioned in Section 3.2. If the compression throughout the along the intake does not provide sufficient temperature rise, fuel
intake is insufficient to raise the temperature of the air–fuel mixture injection with a larger ER value promises a high internal energy and
above the auto-ignition limit, the flow leaves the combustion chamber thereby the mixture can reach auto-ignition limit before it exits from

359
A.C. Ispir et al. Acta Astronautica 217 (2024) 349–362

the combustion chamber. Hence, the borders of the propulsion zones

for cruise altitudes of 25–30 km (for the flight Mach number of M∞ =
6) are situated within the range of 2.7 < 𝑀𝐶,𝐼 < 2.8 while the design
conditions of 2.8 < 𝑀𝐶,𝐼 < 2.9 at higher altitudes could produce more
engine thrust.
In the propulsion regions, low limit of the 𝑀𝐶,𝐼 increases with
the higher values of ER. However, beyond a certain value of the ER,
especially for M∞ = 6 and 7, thermal chocking and corresponding
restriction of positive thrust production occurs highlighting the im-
portance of other selected design variables. For instance, beyond ER
= 0.45 and 0.5 (for M∞ = 6 at 25 and 30 km altitudes respectively)
no net thrust is achieved because of either thermal choking or non-
ignition in the burner. Even though favorable conditions are obtained
Fig. 12. Pressure rise in the scramjet combustion chamber with varying intake exit
for the combustion with suitable 𝑀𝐶,𝐼 i.e. sufficient compression at the Mach number (𝑀𝐶,𝐼 ) and equivalence ratio (ER) at 𝑀∞ = 8 and 25 km of altitude.
supersonic intake, high amount of fuel injection that yields extreme
heat addition into the air flow which can force the entire mixture to
of uninstalled thrust with varying ER values differs based on the intake
decelerate under the sonic limit although it elevates the temperature
exit Mach number (𝑀𝐼,𝐸 ) of choice. There is a non-linear increase of the
of the mixture to avoid leaving the combustion chamber unburnt. The
𝐹𝑢 with the greater values of the ER and decreasing levels of the 𝑀𝐶,𝐼 .
border line between non-blowout and thermal choking regions respects
The effects of the design variables which are ER and 𝑀𝐶,𝐼 on pressure
to this phenomenon and can be seen in all operating maps of the
rise are larger than temperature rise in the combustor. The effects of
scramjet engine given for Mach 6 and 7 design speeds in Fig. 10.
𝑚̇ 𝑎𝑖𝑟 on the engine performance are significant and additionally it is
The regime of the simulations performed at the same flight Mach,
worth to mention that the pressure forces having a great impact on the
𝑀𝐶,𝐼 , and ER may vary with the change of the altitude. For instance,
overall thrust production. Thus, the thrust trend follows that of pressure
the design scenario created with 2.8 of 𝑀𝐶,𝐼 , 0.58 of ER for the
Mach 6 cruise speed and 30 km altitude flight results the in thermal as shown in Figs. 10 and 12. On the other hand, the rate of temperature
choking region. However, these conditions can be in the propulsive change decreases at lower values of 𝑀𝐶,𝐼 while the pressure rise at the
zone provided that the altitude is chosen as 35 km by keeping the same direction increases non-linearly and uninstalled thrust follows this
other variables same. It can be stated that the importance of flight trend. Therefore, the pressure forces become the dictating factor for
altitude is linked to the fuel mass flow rate. Due to the higher air the propulsive performance, especially in the design data exploration
density, the 𝑚̇ 𝑎𝑖𝑟 is greater at lower altitudes for constant flight Mach performed for 𝑀∞ = 8.
number, hence fuel consumption is larger for the same range of ER For flight Mach numbers of 6 and 7, the thrust generation is
values. At higher altitude, it is less possible to observe thermal choking. observed to be amplified towards the boundaries of thermal choking
For the same design conditions at different altitudes, almost all the fuel where the excessive fuel injection beyond this limit yields unproductive
is consumed at 35 km altitude before the flow within the combustion propulsive solutions. Although thermal choking is not seen in the
chamber decelerates to the sonic condition. On the other hand, there database generated for 𝑀∞ = 8 since the flow exits the combustor
is a remarkable amount of fuel left unburnt flying at 25 km altitude before decelerating to sonic conditions within the chosen range of 𝑀𝐶,𝐼
when the flow Mach numbers to unity. Hence, the significance of the (2.7–4.5) and ER (0.1–0.7), the 𝐹𝑢 on the performance maps has an
altitude on the propulsive performance is emphasized based on these increasing trend to the potential thermal choking regions. For 𝑀∞ =
observations in engine design space definition with respect to 𝑚̇ 𝑎𝑖𝑟 and 7, the thermal choking region occurs at design scenarios created with
𝑚̇ 𝑓 𝑢𝑒𝑙 (See. 7). high values of the ERs and low values of the 𝑀𝐶,𝐼 s. On the other
Regardless of flight Mach number, the uninstalled thrust (𝐹𝑢 ) values hand, increasing of the flight altitude shrinks the thermal choking zone
are observed to be increasing at higher values of ER and lower values because of the similar reasons given above for Mach 6. The effects of
of 𝑀𝐶,𝐼 . The first component of this relation refers to the increasing the fuel consumption on the thermal choking region size change which
consumption of fuel which in case of thermal choking is absent. Hence, relates to the flight altitude and 𝑚̇ 𝑎𝑖𝑟 are also observed in 𝑀∞ = 7.
this yields higher thermal energy generation that is converted into The increasing altitude dwindles the thermal choking region on the
kinetic energy to elevate the generated thrust. The latter element rep- operating map of the 𝑀∞ = 7 (See Fig. 10). Thus, we can express that
resents the influence of combustion chamber initial conditions induced the design of a supersonic intake with 𝑀𝐶,𝐼 = 2 can yield significant
by the flow deceleration established by means of the intake contours. thrust production only if the fuel is injected with less than 0.5 of ER
Lower speeds at the combustion chamber inlet suggests that a higher for 𝑀∞ = 7 operation.
compression pressure of the incoming air is obtained without suffering The design space exploration of the scramjet engine under inves-
from startability concerns which yields a more efficient combustion tigation in the current work shows that the intake truncation angle
promoted by the higher initial temperature of the air–fuel mixture and 𝛽 has a slight impact on the thrust generation as demonstrated by
consequently greater uninstalled thrust. Nevertheless, the deceleration comparing the performance maps given by Figs. 10 and 13 for 𝑀∞ = 8
at the intake is restricted by the thermal choking phenomenon also as a at 25 km altitude. However, it does not yield significant changes on the
function of the amount of injected fuel (inputted with the ER) as clearly characteristics of non-ignition and thermal choking regions although
seen for the flight Mach number of 6 in the performance maps. In this we expect an improvement on the pre-combustion flow conditions by
regard, some unfavorable design configurations for 𝑀∞ = 6 and 7 are means of amplified deceleration. The selection of a large truncation
identified based on the detection of thermal choking. Nonetheless, these angle limits only the investigated pre-combustion Mach number range.
configurations are not eliminated from the investigated design space for In addition, pre-combustion flow conditions such as temperature and
𝑀∞ = 8 as no thermal choking is observed within the range of the given pressure do not vary notably with the 𝛽. One of the main reasons
variables. leading to the decrease of 𝐹𝑢 is associated with the mass flow rate of
The uninstalled thrust 𝐹𝑢 follows similar trend with pressure and captured air. Smaller values of air mass flow rate due to the high intake
temperature rises along the combustion chamber. The combustion pro- truncation angle reflects on the performance in terms of a lower thrust
cess in scramjet engines is highly influenced by the pre-combustion production.
pressure conditions i.e. the selected 𝑀𝐶,𝐼 and flight conditions. De- For the specific impulse (𝐼𝑠𝑝 ), importance of the design parameters
creasing 𝑀𝐶,𝐼 raises initial pressure and temperature and while reduc- are pronounced differently. In this regard, the flight altitude has a
ing the air velocity entering the combustion chamber. The variation great role on its determination whilst its effect is not as much as other

360
A.C. Ispir et al. Acta Astronautica 217 (2024) 349–362

is repetitively proposed at each design point by using axisymmetric

flow templates and flow variables were computed by solving one-
dimensional flow governing equations coupled with detailed-chemistry
approach and thermophysical tables along the combustor and noz-
zle components. Thereby, an engine database of ∼10,000 points is
generated. The overall and component-based performance (which are
compression ratio, total pressure recovery factor, and captured mass
flow rate of air (for intake), leftover hydrogen, pressure and tem-
perature rises (for combustor)) are discussed via the demonstrated
propulsive database maps. Design boundaries are defined by thermal
choking where the air flow is decelerated to sonic conditions and non-
ignition where the system produces poor thrust due to the unfeasible
pre-combustion flow conditions are determined. A feature importance
Fig. 13. Performance map of the scramjet engine for 𝑀∞ = 8 and 25 km altitude
study using Shapley Additive Explanations framework is also conducted
flight with 6◦ truncation angle (𝛽), given with uninstalled thrust (𝐹𝑢 ), fuel consumption
(𝑚̇ 𝑓 𝑢𝑒𝑙 ) (solid lines) and specific impulse (𝐼𝑠𝑝 ) (dashed lines). to facilitate the understanding of engine characteristics and depict
importance of the design variables by quantifying their impact on the
performance.
design configurations except truncation angle (See Fig. 7). Contrarily
The calculations of the system performance show that the air mass
to the thrust, the 𝐼𝑠𝑝 follows a decreasing trend with the increase of
flow rate captured by the supersonic intake dictates the uninstalled
the ER for the same 𝑀𝐶,𝐼 , flight altitude, flight Mach number and 𝛽.
thrust significantly. Therefore, it is worth concluding that flight prop-
Apparently, the produced thrust per consumed fuel reduces with the
erties are considered as the most important design parameters of a
higher values of ER and this results in a declining curve of the 𝐼𝑠𝑝
scramjet engine. On the other hand, intake exit Mach number also
towards the increasing of ER. The change of the uninstalled thrust is
determines how much air flow will be in-taken into the propulsive duct
not at the same rate with that of the 𝑚̇ 𝑓 𝑢𝑒𝑙 when the performance is
and directly related to the pre-combustion flow conditions. Therefore,
boosted by injecting more fuel. Therefore, beyond a certain ER value,
as depicted in the feature importance study, these design parameters
𝐼𝑠𝑝 starts to decrease (see the solutions of Mach 6 and 30 km flight in
are as important as the mission trajectory requirements in terms of
Fig. 10). It is possible to observe a parabolic curve in the map of 𝐼𝑠𝑝 . We
examine that the specific impulse generally takes its stationary point at uninstalled thrust.
around at 0.22 of the ER value for 𝑀∞ = 6. This can be attributed to The more compression attained along the intake results in better
non-linear increase of the uninstalled thrust until ER = 0.22 after which propulsive performance overall. However, combustion can end up the
the increase rate by the higher value of the ER declines dramatically. reaching thermal choking limit inside the burner when the intake exit
Furthermore, the system response to the increasing fuel consumption Mach number is chosen below a certain limit which is also a function
is observed to be quicker before ER = 0.22 is reached which is also of the injected amount of fuel. Thermal choking limit is only detected
noted to be similar for 𝑀∞ = 7 and 𝑀∞ = 8 with a value of ER = for cruise speeds of Mach 6 and 7 flights, whereas the flow did not
0.26. Hence, the most efficient operating condition in terms of specific decelerate to sonic conditions for Mach 8 simulations even if the fuel
impulse for such type of scramjet engine varies between 0.22 < ER < is injected with upper limit of the equivalence ratio.
0.26 for the three cruise speeds under investigation within the selected Intake truncation angle 𝛽 provides higher initial temperature and
design envelope. pressure for the combustion. Even though they provide more favorable
Moreover, for constant cruise altitude and leading edge truncation pre-combustion flow conditions with the promise of increased burner
angle (𝛽), increasing the flight Mach number causes more air mass flow performance, the losses on the total pressure cannot be compensated
to be captured into the engine duct which results in an increase in the in terms of propulsive performance as database and feature importance
generated uninstalled thrust values. On the other hand, the increase study demonstrate.
of the 𝑚̇ 𝑎𝑖𝑟 causes more fuel to be injected for the same ER. Since the The uninstalled thrust has an increasing trend with the greater
system cannot respond to this increment, the specific impulse decreases cruise Mach number and lower altitude. As stated, the captured mass
with the increase of the flight speed at the same altitude. Although the flow rate of the air and air density (related to the altitude) play very sig-
rate of pressure rise during the combustion increases at higher 𝑀∞ and nificant roles in this phenomenon. There is a direct proportion between
accordingly the uninstalled thrust (in this sense, a better performance is the uninstalled thrust and fuel consumption. Hence, while the thrust
attained via the higher values of 𝑚̇ 𝑎𝑖𝑟 at the same altitude flight), the 𝐼𝑠𝑝 increases, specific impulse remains still in many of the simulations
is dominated by the fuel consumption and it reduces since the system performed with varying equivalence ratio and intake exit Mach number
cannot compensate the increase of the consumed fuel with the higher for the same cruise altitude and Mach number.
thrust values. Finally, the intake truncation angle effect on the thrust In the design space of the investigation, a non-ignition zone is ob-
production is detailed above. Since the change on the fuel consumption served for all flight altitude and Mach numbers. In case the compression
is quite similar, the selection of 𝛽 does not seem to have any direct at the intake is not strong enough, there exists a possibility for the air–
impact on the specific impulse. However, the higher values of 𝛽 refers fuel mixture leaving the combustion chamber without being reacted.
more losses as quantified by the TPR (See Section 3.2) as discussed in This is completely linked to the chosen intake exit Mach number in
the intake analysis, it affects the overall system thrust production and the intake design. High intake exit Mach number values not only cause
accordingly the specific impulse. insufficient amount of captured air mass flow rate but also cannot
yield favorable pre-combustion flow conditions. Therefore, in scramjet
4. Conclusions engine design, intake exit Mach number must be selected within an
optimal range to avoid the issues regarding auto-ignition and thermal
In the present study, a comprehensive design space exploration choking.
of a scramjet engine is conducted to assess effects of several de-
sign configuration parameters and flight conditions (flight altitude Declaration of competing interest
and Mach number, intake truncation angle, equivalence ratio and
intake exit Mach number) on performance outputs which are identified The authors declare that they have no known competing finan-
as uninstalled thrust, specific impulse, fuel consumption and over- cial interests or personal relationships that could have appeared to
all thermodynamic efficiency. A new geometry for intake component influence the work reported in this paper.

361
A.C. Ispir et al. Acta Astronautica 217 (2024) 349–362

Acknowledgments [19] M.K. Smart, Design of three-dimensional hypersonic inlets with rectangular-
to-elliptical shape transition, J. Propuls. Power 15 (3) (1999) 408–416, http:
//dx.doi.org/10.2514/2.5459.
This research study is financially supported by the European Union’s
[20] S. Molder, E.J. Szpiro, Busemann inlet for hypersonic speeds, J. Spacecr. Rockets
Horizon 2020 research and innovation program through the MORE 3 (8) (1966) 1303–1304, http://dx.doi.org/10.2514/3.28649.
& LESS (MDO and REgulations for Low-boom and Environmentally [21] R.J. Gollan, M.K. Smart, Design of modular shape-transition inlets for a conical
Sustainable Supersonic aviation) project, under grant agreement No hypersonic vehicle, J. Propuls. Power 29 (4) (2013) 832–838, http://dx.doi.org/
101006856. 10.2514/1.B34672.
[22] X. dong Wang, J. feng Wang, Z. jun Lyu, A new integration method based on
the coupling of mutistage osculating cones waverider and Busemann inlet for
References hypersonic airbreathing vehicles, Acta Astronaut. 126 (2016) 424–438, http:
//dx.doi.org/10.1016/j.actaastro.2016.06.022, URL https://www.sciencedirect.
[1] E.T. Curran, S. Murthy, Scramjet Propulsion, 189, American Institute of com/science/article/pii/S0094576516303460, Space Flight Safety.
Aeronautics and Astronautics, 2001, ISBN: 13: 9781563473227. [23] R. Courant, K. Friedrichs, Supersonic Flow and Shock Waves, in: Applied
[2] A.C. Ispir, K. Zdybał, B.H. Saracoglu, T. Magin, A. Parente, A. Coussement, Mathematical Sciences, Springer New York, 1999, URL https://books.google.be/
Reduced-order modeling of supersonic fuel–air mixing in a multi-strut injection books?id=Qsxec0QfYw8C.
scramjet engine using machine learning techniques, Acta Astronaut. 202 (2023) [24] G. Emanuel, M. Gad-El-Hak, Analytical fluid dynamics, Appl. Mech. Rev. 54 (4)
564–584, http://dx.doi.org/10.1016/j.actaastro.2022.11.013. (2001) B68.
[3] J. Ma, J. Chang, J. Zhang, W. Bao, D. Yu, Control-oriented unsteady one- [25] J. Anderson, Fundamentals of Aerodynamics, McGraw-Hill Education, 2010.
dimensional model for a hydrocarbon regeneratively-cooled scramjet engine, [26] H. Munthe-Kaas, High order Runge-Kutta methods on manifolds, Appl. Numer.
Aerosp. Sci. Technol. 85 (2019) 158–170, http://dx.doi.org/10.1016/j.ast.2018. Math. 29 (1) (1999) 115–127, http://dx.doi.org/10.1016/S0168-9274(98)00030-
12.012. 0, Proceedings of the NSF/CBMS Regional Conference on Numerical Analysis of
[4] L. Tian, L. Chen, Q. Chen, F. Li, X. Chang, Quasi-one-dimensional multimodes Hamiltonian Differential Equations.
analysis for dual-mode scramjet, J. Propuls. Power 30 (6) (2014) 1559–1567, [27] F.J. Malo-Molina, D.V. Gaitonde, H.B. Ebrahimi, S.M. Ruffin, Three-dimensional
http://dx.doi.org/10.2514/1.B35177. analysis of a supersonic combustor coupled to innovative inward-turning inlets,
[5] D. Zhang, Y. Feng, S. Zhang, J. Qin, K. Cheng, W. Bao, D. Yu, Quasi-one- AIAA J. 48 (3) (2010) 572–582, http://dx.doi.org/10.2514/1.43646.
dimensional model of scramjet combustor coupled with regenerative cooling, [28] F. Zuo, G. Huang, C. Xia, Investigation of internal-waverider-inlet flow pattern
J. Propuls. Power 32 (3) (2016) 687–697, http://dx.doi.org/10.2514/1.B35887. integrated with variable-geometry for TBCC, Aerosp. Sci. Technol. 59 (2016)
[6] S.M. Torrez, J.F. Driscoll, M. Ihme, M.L. Fotia, Reduced-order modeling of 69–77, http://dx.doi.org/10.1016/j.ast.2016.10.009.
turbulent reacting flows with application to ramjets and scramjets, J. Propuls. [29] F. Zuo, S. Mölder, G. Chen, Performance of wavecatcher intakes at angles of
Power 27 (2) (2011) 371–382, http://dx.doi.org/10.2514/1.50272. attack and sideslip, Chin. J. Aeronaut. 34 (7) (2021) 244–256, http://dx.doi.org/
[7] C.H. Birzer, C.J. Doolan, Quasi-one-dimensional model of hydrogen-fueled scram- 10.1016/j.cja.2020.04.004, URL https://www.sciencedirect.com/science/article/
jet combustors, J. Propuls. Power 25 (6) (2009) 1220–1225, http://dx.doi.org/ pii/S1000936120301278.
10.2514/1.43716. [30] V. Babu, K. Sinha, U. Pathak, Comparative analysis of ramp-type and Busemann
[8] B.O. Cakir, A.C. Ispir, B.H. Saracoglu, Reduced order design and investigation of intakes for hypersonic air-breathing engine, 2017.
intakes for high speed propulsion systems, Acta Astronaut. 199 (2022) 259–276, [31] Y. Xiao, L. Yue, P. Gong, X. Chang, Investigation on a truncated streamline-
http://dx.doi.org/10.1016/j.actaastro.2022.07.037. traced hypersonic busemann inlet, in: 15th AIAA International Space Planes and
[9] N.O.P. Raj, K. Venkatasubbaiah, A new approach for the design of hypersonic Hypersonic Systems and Technologies Conference, http://dx.doi.org/10.2514/6.
scramjet inlets, Phys. Fluids 24 (8) (2012) 086103, http://dx.doi.org/10.1063/ 2008-2634.
1.4748130. [32] S. Molder, Curved Aerodynamic Shock Waves (Ph.D. thesis), Department of
[10] X. Hao, J. Chang, W. Bao, Z. Zhang, A model of mode transition logic in Mechanical Engineering McGill University Montreal, 2012.
dual-mode scramjet engines, Aerosp. Sci. Technol. 49 (2016) 173–184, http: [33] A.K. Flock, A. Gülhan, Viscous effects and truncation effects in axisymmetric
//dx.doi.org/10.1016/j.ast.2015.12.001. busemann scramjet intakes, AIAA J. 54 (6) (2016) 1881–1891, http://dx.doi.
[11] S.M. Torrez, D.J. Dalle, J.F. Driscoll, New method for computing performance org/10.2514/1.J054287.
of choked reacting flows and ram-to-scram transition, J. Propuls. Power 29 (2) [34] A. Tognelli, B.O. Cakir, B.H. Saracoglu, Design and investigation of busemann
(2013) 433–445, http://dx.doi.org/10.2514/1.B34496. intakes for high-speed propulsion systems, in: AIAA SCITECH 2023 Forum, 2023,
[12] V. Fernández-Villace, G. Paniagua, J. Steelant, Installed performance evaluation http://dx.doi.org/10.2514/6.2023-2353.
of an air turbo-rocket expander engine, Aerosp. Sci. Technol. 35 (2014) 63–79, [35] T.F. O’brien, J.R. Colville, Analytical computation of leading-edge truncation
http://dx.doi.org/10.1016/j.ast.2014.03.005. effects on inviscid Busemann-Inlet performance, J. Propuls. Power 24 (4) (2008)
[13] B.O. Cakir, A.C. Ispir, F. Civerra, B.H. Saracoglu, Reduced order design space 655–661, http://dx.doi.org/10.2514/1.30178.
analysis of for ramjet engines with data mining techniques, in: AIAA SCITECH [36] V. Babu, Oblique shock waves, in: Fundamentals of Gas Dynamics, Springer
2023 Forum, 2023, p. 2017, http://dx.doi.org/10.2514/6.2023-2017. International Publishing, Cham, 2021, pp. 103–119, http://dx.doi.org/10.1007/
[14] M.G. Owens, S. Mullagiri, C. Segal, P.J. Ortwerth, A.B. Mathur, Thermal choking 978-3-030-60819-4_7.
analyses in a supersonic combustor, J. Propuls. Power 17 (3) (2001) 611–616, [37] T. Poinsot, D. Veynante, Theoretical and Numerical Combustion, RT Edwards,
http://dx.doi.org/10.2514/2.5785. Inc., 2005.
[15] S. Brahmachary, H. Ogawa, Multipoint design optimization of busemann-based [38] C.J. Jachimowski, An Analytical Study of the Hydrogen-Air Reaction Mechanism
intakes for scramjet-powered ascent flight, J. Propuls. Power 37 (6) (2021) with Application to Scramjet Combustion, Tech. Rep., NASA TP-2791, 1988.
850–867, http://dx.doi.org/10.2514/1.B38383. [39] B.J. McBride, Coefficients for Calculating Thermodynamic and Transport Prop-
[16] C. Fujio, H. Ogawa, Physical insights into multi-point global optimum design erties of Individual Species, Vol. 4513, NASA Langley Research Center,
of scramjet intakes for ascent flight, Acta Astronaut. 194 (2022) 59–75, http: 1993.
//dx.doi.org/10.1016/j.actaastro.2022.01.036. [40] A.C. Hindmarsh, P.N. Brown, K.E. Grant, S.L. Lee, R. Serban, D.E. Shumaker, C.S.
[17] F.-Y. Zuo, S. Mölder, Flow quality in an M-Busemann wavecatcher intake, Aerosp. Woodward, SUNDIALS: Suite of nonlinear and differential/algebraic equation
Sci. Technol. 121 (2022) 107376, http://dx.doi.org/10.1016/j.ast.2022.107376. solvers, ACM Trans. Math. Softw. 31 (3) (2005) 363–396, http://dx.doi.org/10.
[18] S.M. Lundberg, S.-I. Lee, A unified approach to interpreting model predictions, 1145/1089014.1089020.
Adv. Neural Inf. Process. Syst. 30 (2017) http://dx.doi.org/10.48550/arXiv.1705. [41] S.M. Torrez, J.F. Driscoll, M. Ihme, M.L. Fotia, Reduced-order modeling of
07874. turbulent reacting flows with application to ramjets and scramjets, J. Propuls.
Power 27 (2) (2011) 371–382, http://dx.doi.org/10.2514/1.50272.

362