See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/366323539

An Automated Aerodynamic Analysis System in Missile Based on Open-

Source Software

Article in International Journal of Aeronautical and Space Sciences · November 2022

DOI: 10.1007/s42405-022-00558-0

CITATIONS READS

0 455

1 author:

Sang Chul Lee

m-g-r
6 PUBLICATIONS 30 CITATIONS

SEE PROFILE

All content following this page was uploaded by Sang Chul Lee on 16 December 2022.

The user has requested enhancement of the downloaded file.

International Journal of Aeronautical and Space Sciences
https://doi.org/10.1007/s42405-022-00558-0

ORIGINAL PAPER

An Automated Aerodynamic Analysis System in Missile Based

on Open-Source Software
Kang Kuk You1 · Jung Hyun Ha1 · Sang Chul Lee1

Received: 28 October 2021 / Revised: 19 August 2022 / Accepted: 16 November 2022

© The Author(s), under exclusive licence to The Korean Society for Aeronautical & Space Sciences 2022

Abstract
An automated aerodynamic analysis system that predicts aerodynamic characteristics of the missile configuration at the early
design stage is developed using open-source software and presented. This system consists of three modules for geometry
modeling, mesh generation, and flow analysis. The missile configuration defined by Missile DATCOM is created using Open-
CASCADE software. The unstructured tetrahedral mesh is generated by NETGEN software with minimal input parameters.
In particular, the density-based coupled solver, TSLAeroFoam, is used to predict aerodynamic coefficients accurately in the
compressible flow regime. The presented system is verified for three representative missile configurations, and the results
show good agreements in aerodynamic coefficients with the experiment.

Keywords Automated aerodynamic analysis system · Open-source · OpenCASCADE · NETGEN · OpenFOAM

1 Introduction configurations should be conducted repeatedly in MDAO.

Consequently, the automation of geometry representation
Multidisciplinary design analysis and optimization (MDAO) and mesh generation is essential for work efficiency.
is a technique used widely in the conceptual design stage. Several studies have been previously conducted to develop
The missile design includes analysis of various disciplines partially or fully automated processes. Marco et al. [5] devel-
such as aerodynamics, propulsion, trajectory, etc. MDAO oped an automated aircraft modeling API, named JPAD,
with design parameters can allow designers to understand the which uses geometric parametrization, but the automated
overall characteristics of missiles but requires a large num- processes for mesh generation and flow analysis were not
ber of analysis cases. For that reason, low-fidelity solvers included. Tomac et al. [6] developed a module for generating
like Missile DATCOM [1] and Aeroprediction [2], which are the mesh automatically as a part of the CEASIOM frame-
semi-empirical codes and do not require CAD, are mainly work. The module was developed using the open-source
adopted for MDAO. However, some studies showed that mesh generation software, but it had a low robustness in
the low-fidelity solvers cannot predict aerodynamic perfor- the creation process of prism layer for thin airfoil. Ordaz
mance accurately compared to the experiment results [3, 4]. et al. [7] introduced an automated process of generating a
Therefore, a computational fluid dynamics (CFD) can be an tetrahedral mesh for sonic boom analysis of supersonic con-
alternative option for aerodynamic analysis. figuration, but a regionally restricted license is applied to the
There were some challenges in applying CFD to MDAO. mesh generation software, AFLR3. Gu et al. [8] presented
Unlike semi-empirical solvers, CFD requires water-tight a fully automated analysis workflow that can be applied for
geometry representation from the design parameters, and it overall aircraft design, but the open-source software was used
is also necessary to create the volume mesh for flow analysis. only for the geometry generation while other processes used
These time-consuming processes are performed manually by the commercial software such as Pointwise and ANSYS Flu-
the user and a large number of simulation cases for similar ent.
This study aims to develop a fully automated aerodynamic
B Sang Chul Lee analysis system based on open-source software. Addition-
[email protected] ally, a special treatment is performed on the missile geometry
to solve a problem related with the robustness for gen-
1 Research Center, NEXTfoam Co., LTD., Seoul, Republic of
erating the mesh of high quality. For the flow analysis,
Korea

123
 International Journal of Aeronautical and Space Sciences

TSLAeroFoam [9], a density-based coupled solver which the geomerty-related procedures are carried out such as the
was developed previously by our research group for com- creation of the computational domain with the boolean oper-
pressible flow, is used. The verification of presented system ation and the identifying process of the specific faces to set
is conducted for three different missile configurations. the mesh sizes. In the SMESH module, the processes of
This paper is organized as follows: Sect. 2 describes the setting mesh sizes and generating the volume mesh are per-
modules in an automated aerodynamic analysis system. The formed. The meshing algorithms for unstructured tetrahedral
mathematical expression for TSLAeroFoam is briefly intro- mesh are adopted from the NETGEN [13] software, which
duced in Sect. 3. Section 4 shows the results of aerodynamic is embedded in SALOME, then the volume mesh is exported
analyses for test cases, and the conclusion of this study is as the UNV file format.
given in Sect. 5. Finally, the flow analysis is conducted using OpenFOAM
[14], which is an open-source CFD software package. Open-
FOAM has over 100 standard solvers that can be applied
2 Automated Aerodynamic Analysis System to a wide range of flow simulations and over 150 standard
utilities for pre-processing, mesh generation and conversion,
The overall automated system for the missile aerodynamic and post-processing. The pre-processing for flow analysis is
analysis depicted in Fig. 1. The system consists of three indi- conducted using the mesh created from the previous module
vidual processes: geometry modeling, mesh generation, and and the input parameters.
flow analysis. For fully automated analysis without any user’s
intervention, the required input parameters for each module
are defined in the XML (eXtensible Markup Language) file 2.1 Geometry Modeling
that can store hierarchical data.
First of all, for geometry modeling, the geometric param- The geometric parameters for the body geometry are shown
eters based on the definition method in Missile DATCOM in Fig. 2. There are two different options to define the body
are required, then OpenCASCADE (OCC) [10], an open- in Missile DATCOM. The geometric parameters for each
source geometry modeling kernel, uses those parameters to option are summarized in Table 1. Option 1 defines the body
create geometry. OCC is a fully object-oriented C++ API by dividing it into the nose, centerbody, and afterbody, and
and supports a full-scale BRep (Boundary Representation) five nose shapes (conical, tangent ogive, power series, and
format. OCC has a large number of modeling functions and Haack/Karman series) are available. Option 2 defines it using
can generate sophisticated geometry rapidly. In this study, the longitudinal stations and corresponding width of each
pythonOCC [11], an OCC which is wrapped into python, is station. The cross-section of the body can be defined as
used for geometry generation, and the created geometry is circle or ellipse in both options. The three different types
exported as the STEP file format. of nose are shown in Fig. 3, and all nose types are avail-
Secondly, SALOME [12] software, an open-source plat- able for both options. The geometric equations for various
form for pre-and post-processing of numerical simulation, nose shapes and types are presented in Ref. [15]. In this
is used to generate volume mesh for flow analysis. Among study, the sharp nose is modified into a blunted nose with a
the modules of SALOME, the GEOM and SMESH mod- small radius to facilitate the mesh generation and flow anal-
ules are used for mesh generation. In the GEOM module, ysis.

Fig. 1 Automated aerodynamic analysis system

123
International Journal of Aeronautical and Space Sciences

Fig. 2 Body geometric parameters

Table 1 Summary of body geometric parameters

Option Parameters

Option1 Nose length lN

Nose width wN
Blunted/truncated nose radius rB
Centerbody length lC
Centerbody width wC
Afterbody length lA
Afterbody width wA
Ellipticity E
Option2 n-th longitudinal coordinates x(n)
n-th station width w(n)
Blunted/truncated nose radius rB
Ellipticity E Fig. 4 Process of generating body solid

edges, and the body solid is generated by rotating the face

The process of generating the body solid is described in to the x-axis. Additionally, the scale operation for ellipticity
Fig. 4. The coordinates of the points representing the out- can be performed when the cross-section is an ellipse.
lines are calculated with the given parameters, and the outline The process of generating finset geometry consists of
edges are created using those points as the vertices. After that, defining planform, selecting airfoil for each station and
a face is created using a set of closed wires that consist of attaching method to the body. Table 2 summarizes the param-

Fig. 3 Nose options

123
 International Journal of Aeronautical and Space Sciences

Table 2 Summary of Finset geometric parameters

Category Parameters

Planform n-th station chord length c(n)

n-th station semi-span location s(n)
n-th station distance from nose to X LE (n)
leading edge
Airfoil Thickness-to-chord ratio of upper zu
surface
Thickness-to-chord ratio of lower zl
surface
Fraction of chord from leading edge L max
to maximum thickness section
Fraction of chord of constant L flat
thickness section
Fig. 6 Airfoil geometric parameters
Leading edge radius r
Attaching Roll angle of each panel φ(n)
Method Dihedral angle of each panel γ (n)

Fig. 5 Planform geometric parameters

Fig. 7 Geometric parameters for attaching method

eters that define the finset geometry, and the parameters of

the planform are described in Fig. 5. The planform is defined
by multiple spanwise stations that consist of chord length, increases the roll angle. The roll angle and dihedral angle can
semi-span length, and the distance from the nose to the lead- be set individually for each panel.
ing edge of root chord at each station. The process of creating the finset solid is described in
There are four types of airfoil in Missile DATCOM: hexag- Fig. 8. First of all, the airfoil wire at each spanwise station is
onal, circular arc, NACA, and the shape as defined by x, created with the given parameters. After that, the planform
y coordinates. The parameters of each airfoil are shown in solid is constructed by lofting the airfoil wires, and the roll
Fig. 6, and the diamond-shaped airfoil is a special case of the and dihedral angles are applied by rotating each panel.
hexagonal airfoil which sets L FLAT to zero. The sharp edges As described above, the missile geometry is generated
for hexagonal and arc airfoils are modified to have a small using fuse operation of body and finset solids which were
radius on the leading and trailing edges to facilitate mesh created from vertex to solid, so it has the water-tight char-
generation and flow analysis. acteristic which is applicable to mesh generation and flow
The parameters for the attaching method to the body are analysis.
shown in Fig. 7. The roll angle (∅) is measured clockwise XML input files applied for generating body and fin-
from the top center line, and the positive dihedral angle (γ ) set solids are shown Figs. 9 and 10, respectively, and the

123
International Journal of Aeronautical and Space Sciences

Fig. 8 Process of generating finset solid

Fig. 10 XML example for geometry modeling (Finset)

generate the whole missile geometry takes only a few sec-

onds (1–2 s).

2.2 Mesh Generation

An automated mesh generation begins from the CAD geome-

try created in the previous process. The input parameters are
required minimally, and they are about the computational
domain, the mesh size, and the prism layer. The parameters
of the computational domain are shown in Fig. 12 and sum-
marized in Table 3. The computational domain consists of
the spherical far-field region and two cylindrical refinement
Fig. 9 XML example for geometry modeling (Body)
regions. The refinement regions are used to accurately cap-
ture the shock wave and the wake flow generated around the
parameter of list (i.e., array) type is indicated by ‘param’ missile. The sizes of the computational domain and refine-
tag. Figure 11 illustrates missile solids created automatically ment regions are adjusted by given parameters.
with different parameters. The example solids show that the The parameters related to the mesh size are composed
present module for geometry modeling can create various of the maximum size, minimum size, and growth rate. The
configurations according to the different input parameters mesh size can be set individually for missile surfaces, each
while the same algorithm is executed. A series of tasks to refinement region, and far-field region. Also, to predict the

123
 International Journal of Aeronautical and Space Sciences

Table 3 Summary of computational domain parameters

Parameters

Far-field region radius R

Inner refinement region radius Ri
Inner refinement region height Hi
Distance from nose for inner refinement region Xi
Outer refinement region radius Ro
Outer refinement region height Ho
Distance from nose for outer refinement region Xo

Fig. 11 Various missile configurations

Fig. 13 XML example for mesh generation

aerodynamic coefficients accurately, the local refinement is

applied to the nose and the base face of the body. The cell
size for face refinement is automatically calculated using
the diameter of the missile from geometric parameters. The
parameters for the generation of the prism layer consist of
the number of layers, the stretch factor, and the desired y + .
The total thickness of the prism layer is calculated using
freestream conditions from the flow analysis parameters.
Figures 13 and 14 show the example of XML input files
Fig. 12 Computational domain parameters used for mesh generation and the automatically generated
volume mesh in the core refinement region, respectively. The
local face refinements applied on the blunted nose and base
face and the prism layers created on the junction between
the body and finset and the tip face of each finset panel are
shown clearly.

123
International Journal of Aeronautical and Space Sciences

Fig. 14 Automatically generated

volume mesh

2.3 Flow Analysis Table 4 Summary of changeable variables for flow analysis

Category Parameters
OpenFOAM does not provide GUI, unlike commercial CFD
programs, and uses text-based user interface. The simulation Flow conditions Mach number M
case directory consists of three subdirectories: 0, constant, Static pressure (Pa) p
and system. In the 0 directory, the information related to ini-
Static temperature (K) T
tial and boundary conditions is stored. A separate file should
Gas constant Specific heat ratio γ
be written for each field variable. In the constant directory,
Heat capacity at constant pressure Cp
the information that does not change during the analysis (e.g.
the volume mesh, physical properties, turbulence model, and Prandtl number Pr
other constants) is stored. Finally, the system directory con-
tains the information related to the simulation control. The
parameters for conducting simulation such as the start/end
time, CFL number, discretization schemes, and linear solvers
are written in this directory. next chapter. Consequently, a few parameters summarized
Because there are a large number of options that can be in Table 4 can be selected as changeable variables for flow
selected for setting the case, the options for automated case analysis. Figure 15 shows an example XML file for flow
setting should be limited. For instance, k −ω SST model [16] analysis.
is fixed for the turbulence model in this study, and TSLAero- The automated case setting process starts with creating
Foam which was previously developed by our research group a ‘Default’ directory. In the analyses for the same geom-
as a density-based coupled solver for compressible flow is etry, constant and system directories contain identical data
fixed also for the numerical solver. The governing equations regardless of the change of flow condition. Therefore, con-
and algorithm of this solver are briefly introduced in the stant and system directories are created within the ‘Default’
directory. Then, the volume mesh generated in the previous

123
 International Journal of Aeronautical and Space Sciences

The previous equation is discretized in space as

−
→  −
∂Wi → −→ 
Vi + F c, i j − F v, i j Si j  0, (5)
∂t
j∈N (i)

−→
where Vi is volume of the cell, W i is average value at each
cell center. N (i) is the set of neighbor cells, Si j is the face
component oriented from cell i to j.
For temporal discretization in Eq. (5), the first order back-
ward Euler is applied as follows,


Vi −→ n+1 − →n
Wi − Wi
ti
 − → n+1 − → n+1
+ F c, i j − F v, i j Si j  0,
Fig. 15 XML example for flow analysis j ∈N (i) (6)

and both convective and viscous fluxes are linearized using

process is converted to OpenFOAM mesh using ideasUn- Taylor’s series expansion as follows
vToFoam utility. Since all boundary types are converted as
‘patch’ type by default, the boundaries for the missile surface Vi − →n  −→ n
W i + Ac, i j − Av, i j  W i j S i j  −Resi n ,
must be changed to ‘wall’ type after the mesh is converted. ti
j∈N (i)
When the simulation directories are created for each flow (7)
condition, constant and system directories are copied from
the ‘Default’ directory. After that, 0 directory which reflects →n −
− → n+1 − →n −
→ − →
the corresponding flow condition is created. where  W i  W i − W i , Ac  ∂ Fc /∂ W is convec-
−
→ − →
tive flux Jacobian, Av ∂ Fv /∂ W is viscous flux Jacobian,
−
→ n →n 
−
and Resi n  j∈N (i) F c, i j − F v, i j S i j is residual
3 Numerical Solver term. Since a large amount of memory is required to store
these Jacobian, the lower–upper symmetric Gauss–Seidel
The conservation of mass, momentum, and energy equations (LU-SGS) algorithm [17] is used.
of compressible flow are as follows First of all, the Jacobian matrices are simplified using
Steger–Warming’s flux vector splitting method [18] for con-
∂ρ  −→
+ ∇ · ρ U  0, (1) vective flux and thin shear layer approximation (TSL) for
∂t viscous flux as below
 − →
∂ ρU  −
→ − → 
Vi − →n −→n
+ ∇ · ρU ⊗ U + ∇p  ∇ · τf, (2) W i + Ac, i + − Av, i ∗  W i S i j
∂t ti
j ∈N (i)
∂(ρ E)  −→  →
−  −→ n
+ ∇ · ρ H U  ∇ · τ f · U + ∇ · (k∇T ), (3) + Ac, j + − Av, j ∗  W j S i j  −Resi n .
∂t j ∈N (i) (8)
−
→
where ρ is the density, U is the velocity vector, p is the pres- Secondly, the diagonal terms are split into lower(owner)
sure, τ f is the stress tensor, E is the energy, H is the enthalpy, and upper(neighbor) triangular terms.
k is the thermal conductivity, and T is the temperature. The
integral form of the system of equations is as follows Vi − →n  −→n
W i + Ac, i + − Av, i ∗  W i S i j
ti
 −
→   j∈N (i)
∂W −
→ − →
dV + Fc − Fv dS  0, (4) 
V ∂t −→ n
S
+ Ac, j − − Av, j ∗  W j S i j
→ 
− −
→  j ∈L(i)
where W  ρ, ρ U , ρ E is conservative variable vector,  −→ n
−
→ −
→ + Ac, j − − Av, j ∗  W j S i j  −Resi n .
Fc is convective flux vector, and Fv is viscous flux vector. j ∈U (i) (9)

123
International Journal of Aeronautical and Space Sciences

Fig. 16 NASA TP 1078 (Sparrow) [19]

The previous equation is represented as block matrix sys- Table 5 Flight conditions for Sparrow
tem
Flow conditions Value

(D + L + U )W n  −R n . (10) Mach number 1.5

Static pressure (Pa) 18114.8
Finally, factorization approximation is applied to previ- Static temperature (K) 233.793
ous equation and represented using two sweep symmetric Angle of attack (°) 0, 5, 11, 17, 23, 27, 32
Gauss–Seidel method as follows

Forward : (D + L)W ∗  −R n , (11)

Backward : (D + U )W n  DW ∗ . (12)

4 Test Cases

Three representative missile configurations which have wind

tunnel data are chosen as the validation cases for the auto-
mated system. The model geometry, flow condition, and
analysis result for each missile configuration are presented
in the following sections.

4.1 Sparrow Missile

The first missile configuration, which is called the Spar-

row [19], is depicted in Fig. 16. The Sparrow consists of
an ogive-cylindrical body, cruciform wings, and in-line tails.
A trapezoidal planform with diamond airfoil is used for the
wings and a delta planform with the hexagonal airfoil is used
for the tails. In this study, the configuration with a roll ori- Fig. 17 Automatically generated mesh for Sparrow
entation angle of 45◦ is used for analysis. The selected flight
conditions are reported in Table 5.
Figure 17. shows the automatically generated surface
mesh and volume mesh in the core refinement region. The are created on the missile surface. The total number of sur-
spherical far-field region is created with a radius of 20 missile face mesh is 348,516 and the volume mesh is approximately
lengths. The average y + is set to be 1.0, and 20 prism layers 23.3 million.

123
 International Journal of Aeronautical and Space Sciences

Fig. 19 Pressure contours for Sparrow

angles of attack, CA values are little bit over-predicted but

match with the overall trend line well. The normal force coef-
ficient (CN ) matches very well with the experiment at all
angles of attack, and the pitching moment coefficient (CM )
matches with the overall trend line compared to the wind
tunnel test.
Figure 19. shows the static pressure contours for two
angles of attack. The shock wave and wake flow are clearly
captured for both angles of attack with the help of the refine-
ment region. The pressure distribution and shock wave are
symmetric at the zero-degree angle of attack. At a high angle
of attack, the shock wave is generated asymmetrically due
to flow direction, and higher pressure distribution occurs on
the lower surface of the missile.

4.2 NASA Tandem Control Missile

NASA Tandem Control Missile (TCM) [20] is selected as

the second test case. The missile configuration is shown in
Fig. 20. The model has a tangent ogive nose and cylindrical
body, cruciform canards, and tail fins. The canards have a
Fig. 18 Longitudinal aerodynamic coefficients for Sparrow
trapezoidal planform with diamond airfoil, and the tail fins
have a trapezoidal planform with the modified double-wedge
airfoil. The flow analysis is performed on the configuration
Figure 18. shows the results of longitudinal aerodynamic with bt /bc of 1.25 in Ref. [20]. The flight conditions used for
coefficients versus the angle of attack. The axial force coeffi- the flow analyses are reported in Table 6.
cient (CA ) predicted by TSLAeroFoam is in good agreement Figure 21 shows the automatically generated surface mesh
with experimental data at low angles of attack, but at high and volume mesh in the core refinement region. Like the first

123
International Journal of Aeronautical and Space Sciences

Fig. 20 NASA TP 2157 (TCM) [20]

Table 6 Flight conditions for NASA TCM

Flow conditions Value

Mach number 3.5

Static pressure (Pa) 1696.55
Static temperature (K) 94.2029
Angle of attack (°) −4, 0, 4, 10, 14, 18

configuration, the spherical far-field region has a radius of

20 missile lengths. The average y + is set to be 1.0, and 20
prism layers are created on the missile surface. The overall
surface mesh count is 353,508 and the volume mesh count is
approximately 25.5 million.
The longitudinal aerodynamic coefficients of NASA TCM
are presented in Fig. 22. C A values from the experiment show
some oscillation on the data. TSLAeroFoam little bit over-
predicts the axial force compared to the test but matches
well in the overall trend line. For normal force and pitching
moments, it is in very good agreement at all angles of attack.
Figure 23 shows static pressure contours for two angles Fig. 21 Automatically generated mesh for NASA TCM
of attack. The shock wave and wake flow are accurately
captured again for both angles of attack. The flow field is
predicted symmetrically for the zero-degree angle of attack.
At a high angle of attack, the shock wave is generated asym- of the body with small nose and tail fin of hexagonal airfoil.
metrically due to flow direction. Also, the large pressure The selected flight conditions are summarized in Table 7.
distribution occurs on the lower surface of the missile. Figure 25 shows automatically generated surface mesh
and volume mesh in the core refinement region. The infor-
mation of far-field region and prism layer is same as other
4.3 NASA TM 2005 – Configuration 3 test cases. The overall surface mesh count is 334,382 and the
volume mesh count is approximately 22.8 million.
The missile configuration 3 of Ref. [21] is chosen as the last The results of longitudinal aerodynamic coefficients ver-
test case. As shown in Fig. 24, Configuration 3 is composed sus the angles of attack are shown in Fig. 26. As mentioned in

123
 International Journal of Aeronautical and Space Sciences

Fig. 23 Pressure contours for NASA TCM

Fig. 24 NASA TP 2005—configuration 3 [21]

Table 7 Flight conditions for NASA TM 2005—configuration 3

Flow conditions Value

Mach number 2.36

Reynolds number (per foot) 1.5 × 106
Angle of attack (°) −1, 0, 2, 4, 6, 17

Ref. [4], there seems to be an error in the position of the cen-

ter of moment provided in Ref. [21], so the pitching moment
is calculated by moving the center of moment 1.75 inches
forward compared to the original value. The flow solver over-
predicts the axial force for all angles of attack, but predicts
Fig. 22 Longitudinal aerodynamic coefficients for NASA TCM the normal force and pitching moment very accurately.

123
International Journal of Aeronautical and Space Sciences

Fig. 25 Automatically generated mesh for NASA TM 2005—configu-

ration 3

Figure 27 shows static pressure contours for two angles of

attack. The shock wave and wake flow are clearly captured
like other test cases. At a high angle of attack, the large pres-
sure distribution occurs on the lower surface of the geometry.

5 Conclusions

This paper presents the fully automated aerodynamic analy-

sis system to predict aerodynamic characteristics for arbitrary
missile configuration in the early design stage. The proposed
system includes the modules of geometry modeling, mesh
generation, and flow analysis. The required parameters for
each module are defined in the XML file format, and all
modules are developed using the open-source software such
as OpenCASCADE, SALOME, and OpenFOAM.
The automated system is verified with three representative
missile configurations. The flow analysis results show that the
predicted aerodynamic characteristics are in good agreement
with the wind tunnel tests.
In the following studies, an automated analysis system
for predicting the control performance of the finset will be
developed.
Fig. 26 Longitudinal aerodynamic coefficients for NASA TM
2005—configuration 3

123
 International Journal of Aeronautical and Space Sciences

6. Tomac M, Eller D (2011) From geometry to CFD grids—an

automated approach for conceptual design. Prog Aerosp Sci
47(8):589–596. https://doi.org/10.1016/j.paerosci.2011.08.005
7. Ordaz I, Li W, Campbell RL (2014) Automated tetrahedral mesh
generation for CFD analysis of aircraft in conceptual design. In:
52nd AIAA aerospace sciences meeting. https://doi.org/10.2514/
6.2014-0118
8. Gu X, Ciampa PD, Nagel B (2018) An automated CFD analysis
workflow in overall aircraft design applications. CEAS Aeronaut J
9:3–13. https://doi.org/10.1007/s13272-017-0264-1
9. Gill JH, Kim BY, Kim JH et al (2013) Development and validation
of a density-based implicit solver using LU-SGS algorithm. In: 8th
international OpenFOAM workshop
10. Open CASCADE. https://www.opencascade.com. Accessed 27
Aug 2021
11. pythonOCC. http://www.pythonocc.org. Accessed 27 Aug 2021
12. SALOME. https://www.salome-platform.org. Accessed 27 Aug
2021
13. Schöberl J (1997) NETGEN An advancing front 2D/3D-mesh gen-
erator based on abstract rules. Comput Vis Sci 1:41–52. https://doi.
org/10.1007/s007910050004
14. OpenFOAM. https://openfoam.org. Accessed 27 Aug 2021
15. Underwood ML, Rosema CC, Wilks BL et al (2007) Recent
improvements to Missile DATCOM. In: 25th AIAA applied aero-
dynamics conference. https://doi.org/10.2514/6.2007-3936
16. Menter FR, Kuntz M, Langtry R (2003) Ten years of industrial
experience with the SST turbulence model. In: 4th international
Fig. 27 Pressure contours for NASA TM 2005—configuration 3 symposium on turbulence, heat and mass transfer
17. Yoon S, Jameson A (1988) Lower-upper symmetric-Gauss–Seidel
method for the Euler and Navier–Stokes equations. AIAA J
Acknowledgements This work was supported by Grant 26(9):1025–1026. https://doi.org/10.2514/3.10007
(UE191079CD) from Agency for Defense Development (ADD) 18. Steger JL, Warming RF (1981) Flux vector splitting of the inviscid
and Defense Acquisition Program Administration (DAPA). gas dynamic equations with application to finite-difference meth-
ods. J Comput Phys 40(2):263–293. https://doi.org/10.1016/0021-
9991(81)90210-2
19. Monta WJ (1977) Supersonic aerodynamic characteristics of a
References Sparrow III type missile model with wing controls and compar-
1. Vukelich SR, Jenkins JE (1984) Missile DATCOM: aerodynamic ison with existing tail-control results. NASA TP 1078
prediction of conventional missiles using component build-up tech- 20. Blair AB, Allen JM, Hernandez G (1983) Effect of tail-fin span on
niques. In: 22nd AIAA aerospace sciences meeting. https://doi.org/ stability and control characteristics of a canard-controlled missile
10.2514/6.1984-388 at supersonic Mach numbers. NASA TP 2157
2. Moore FG, McInville RM, Hymer TC (1999) Application of 21. Allen JM (2005) Aerodynamics of an axisymmetric missile concept
the 1998 version of the aeroprediction code. J Spacecr Rocket having cruciform strakes and in-line tail fins from Mach 0.60 to
36(5):633–645. https://doi.org/10.2514/2.3495 4.63. NASA TM 2005-213541
3. Atik H, Erdem B, Ilgaz M et al (2008) Prediction capabilities
and comparison of panel, semi-empiric and CFD codes for mis-
sile aerodynamic analyses. In: 26th AIAA applied aerodynamics Publisher’s Note Springer Nature remains neutral with regard to juris-
conference. https://doi.org/10.2514/6.2008-6224 dictional claims in published maps and institutional affiliations.
4. Doyle JB, Paul J, DeSpirito J et al (2020) Results of the missile and
projectile aeroprediction discussion group case study. In: AIAA Springer Nature or its licensor (e.g. a society or other partner) holds
SciTech 2020 forum. https://doi.org/10.2514/6.2020-1992 exclusive rights to this article under a publishing agreement with the
5. Marco AD, Stasio MD, Vecchia PD et al (2020) Automatic author(s) or other rightsholder(s); author self-archiving of the accepted
modeling of aircraft external geometries for preliminary design manuscript version of this article is solely governed by the terms of such
workflows. Aerosp Sci Technol 98:105667. https://doi.org/10. publishing agreement and applicable law.
1016/j.ast.2019.105667

123

View publication stats