Study and design of a low cost cooled nozzle

through metal additive manufacturing

Document
Report

Author
Héctor Juan Marı́

Director
Jaume Solé Bosquet
Co-director
Manel Soria Guerrero

Degree
Bachelor’s degree in Aerospace Vehicle Engineering

Examination session
Spring 2021
It is difficult to say what is impossible, for the
dream of yesterday is the hope of today and
the reality of tomorrow.

Robert H. Goddard
 Acknowledgements

To my parents, grandmother, brother and sister in law for bringing me up with the values of love and
hard work. Their unconditional support has been and always will be the driving force behind my
growth as a person, a student and from now on, an engineer.

To my nephews India and Mateo, for whom I wake up every day wanting to be a better man.

To the many friends that my time in Barcelona have given me and that have made these the best years
of my life.

To Xavi and Claudia, for always being by my side and believing in me.

And finally, to the teachers of the ESEIAAT whose knowledge has paved the way for our development
as engineers, especially to Jaume Solé for his unequivocal support as my tutor during the development
of this project.

A mis padres, abuela, hermano y cuñada por educarme en los valores del amor y el trabajo duro. Su
apoyo incondicional me ha dado y siempre me dará fuerza para crecer como persona, estudiante y de
ahora en adelante, como ingeniero.

A mis sobrinos India y Mateo, por los cuales me levanto cada dı́a queriendo ser mejor.

A los muchos amigos que me ha dado este tiempo en Barcelona y que han hecho de estos los mejores
años de mi vida.

A Xavi y Claudia, por estar siempre a mi lado y nunca dejar de creer en mı́.

Y finalmente, a los profesores de la ESEIAAT, cuya sabidurı́a ha sido clave en nuestro desarrollo
como ingenieros, especialmente a Jaume Solé por su apoyo inequı́voco como mi tutor durante el
desarrollo de este proyecto.
 Abstract

State of the art rocket engines are reaching performance levels that have become limited by the
available manufacturing techniques. With a great focus on these engines’ efficiency, rocket engine
producers are focusing on increasing combustion chamber pressures and temperatures, which in turn
produces increasingly greater structural and thermal loads. Complex cooling systems are usually used
to tackle this issue, being the most predominant formula to use the stored cryogenic propellant as a
coolant before being injected in the combustion pressure, more commonly known as regenerative
cooling. However, these cooling systems are limited by the state of the art manufacturing techniques.
To deal with this issue, efforts have been made to implement modern metal Additive Manufacturing
(AM) techniques to rocket engine production.

In order to show the extent of the benefits of such innovations, a rocket engine nozzle has been
designed following the characteristics of SpaceX’s latest Raptor engines. This design has been tested
through Computational Fluid Dynamics (CFD) and Finite Element Method (FEM) simulations so as
to validate its performance parameters and study how the implementation of metal AM techniques
would improve the engine’s performance. The economic feasibility of such implementation has been
studied through the comparison to former techniques, as well as their impact to the environment.

Los motores de cohete actuales están alcanzando niveles de rendimiento que han sido frenados
por las técnicas de producción disponibles. Con la eficiencia de estos motores en mente, sus
productores se centran en incrementar la presión y temperatura de sus cámaras de combustión,
hecho que causa considerables cargas estructurales y térmicas. Complejos sistemas de refrigeración
son usados para abordar este problema, siendo la fórmula predominante usar el combustible
almacenado a temperaturas criogénicas como refrigerante antes de ser inyectado en la cámara de
combustión, técnica que se denomina como refrigeración regenerativa. Sin embargo, estos sistemas
de refrigeración están limitados por las actuales técnicas de manufacturación. Para abordar este
problema, se están realizando esfuerzos para implementar las modernas técnicas de fabricación aditiva
a la producción de motores de cohete.

Con el fin de mostrar hasta qué punto pueden estas técnicas suponer un beneficio, se ha diseñado
una tobera de motor de cohete siguiendo las caracterı́sticas de los nuevos motores Raptor de SpaceX.
Este diseño ha sido probado a través de simulaciones de Dinámica de Fluidos Computacional (CFD)
y el Método de los Elementos Finitos (FEM) para validar su rendimiento y estudiar cómo puede
la implementación de nuevas técnicas de fabricación mejorar el rendimiento de estos motores. La
viabilidad económica de esta implementación ha sido estudiada a través de su comparación con
técnicas anteriores, ası́ como su impacto medioambiental.
Contents

Contents i

List of Figures iii

List of Tables iii

1 Introduction 1
1.1 Aim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.4 Justification and utility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2 State of the art 3

2.1 Liquid Propellant Rocket Engines . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Rocket Engine and Nozzle Theory Fundamentals . . . . . . . . . . . . . . . . . . . 6
2.3 Optimal Nozzle Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4 Cooling system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.5 Manufacturing techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.6 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.7 Approach and selection of alternatives . . . . . . . . . . . . . . . . . . . . . . . . . 15

3 Development of the solution 17

3.1 Nozzle design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.1.1 Initial calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.1.2 Nozzle geometry determination . . . . . . . . . . . . . . . . . . . . . . . . 18
3.1.3 Nozzle and cooling system design . . . . . . . . . . . . . . . . . . . . . . . 20
3.1.4 Drawings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.2 Nozzle CFD simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2.1 2D simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2.2 3D simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.3 Nozzle structural FEM simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.3.1 Structural analysis meshing . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.3.2 Structural analysis setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.3.3 Structural analysis results . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.4 Nozzle thermal FEM analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.4.1 Thermal analysis setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.4.2 Thermal analysis results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

i
4 Analysis of the results 45
4.1 Economic Feasibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.1.1 Market study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.1.2 Project cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.1.3 Fabrication costs through metal additive manufacturing . . . . . . . . . . . . 48
4.1.4 Profitability analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.2 Environmental considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.2.1 Atmospheric contamination . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.2.2 Effect of additive manufacturing technologies on environment . . . . . . . . 51
4.2.3 Space debris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.2.4 Environmental impact of the project . . . . . . . . . . . . . . . . . . . . . . 52
4.3 Planning and scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.3.1 Task identification and description . . . . . . . . . . . . . . . . . . . . . . . 52
4.3.2 Task length and dependencies . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.3.3 Gantt diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.3.4 Discussion on planning and schedule . . . . . . . . . . . . . . . . . . . . . 57
4.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.5 Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.6 Proposed work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

5 Bibliography 61

ii
List of Figures
1 Rocket engine diagram [2] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2 Nozzle geometry diagram [2] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3 Compressed nozzle design [8] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4 Schematic view of cooling channels [12] . . . . . . . . . . . . . . . . . . . . . . . . 10
5 Thrust chamber cooling designs [14] . . . . . . . . . . . . . . . . . . . . . . . . . . 11
6 Calculated nozzle geometry before truncation. Obtained with [22] . . . . . . . . . . 19
7 Final geometry for the nozzle outline . . . . . . . . . . . . . . . . . . . . . . . . . . 20
8 Schematic of the cooling system. Courtesy of Elisei Maslov . . . . . . . . . . . . . 21
10 Final nozzle design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
11 Final nozzle technical drawing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
12 2D simulation mesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
13 2D simulation results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
14 Simulation flow bodies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
15 Fluid mesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
16 Solution of flow parameters after simulation . . . . . . . . . . . . . . . . . . . . . . 32
17 Solution of flow parameters in nozzle region . . . . . . . . . . . . . . . . . . . . . . 33
18 Detail of the nozzle exit wall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
19 Heat flux through the nozzle wall . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
20 Fluid parameters along center-line of the nozzle . . . . . . . . . . . . . . . . . . . . 35
22 Thrust output through mesh refinement . . . . . . . . . . . . . . . . . . . . . . . . . 37
24 Nozzle FEM mesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
25 FEM pressure condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
26 FEM structural analysis results before and after thermal strain . . . . . . . . . . . . 40
27 Detail of the deformed nozzle and the undeformed nozzle wireframe . . . . . . . . . 42
28 Detail of the true scale of the deformation on the cooling channel . . . . . . . . . . . 42
29 Thermal analysis results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
30 Comparison of traditional and additive manufacturing evolution [28] . . . . . . . . . 48
31 Gantt diagram of the project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

List of Tables
1 State of the art LPRE propellant comparison [3] . . . . . . . . . . . . . . . . . . . . 3
2 State of the art rocket engine comparison [3] . . . . . . . . . . . . . . . . . . . . . . 5
3 Inconel 718 composition by mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4 Additively manufactured Inconel 718 tensile properties [18] . . . . . . . . . . . . . 14
5 C-18150 composition by mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

iii
6 C-18150 properties [19] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
7 Initial design parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
8 Obtained parameters comparison with avaliable data . . . . . . . . . . . . . . . . . 18
9 Mass imbalances for the 2D case . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
10 Mass imbalance through mesh refinement . . . . . . . . . . . . . . . . . . . . . . . 28
11 Mesh quality metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
12 Mass flow rate balance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
13 State of the art rocket engine production cost comparison [3] . . . . . . . . . . . . . 45
14 Project budget . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
15 Software and PC project cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
16 L-PBF design specifications (Steel) . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
17 Task dependencies and lengths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

iv
1 INTRODUCTION

1 Introduction

1.1 Aim

A liquid propellant cooled rocket nozzle will be designed through CAD software taking SpaceX’s
latest Raptor engines as reference, with an emphasis on its regenerative cooling system. This design
will be tested through FEM and CFD simulations so as to ensure the required performance parameters
are met and it can withstand its extreme operation conditions. The effects of the implementation of
metal additive manufacturing techniques for its production will be studied, taking into consideration
its economic feasibility so as to provide a more competitive solution.

1.2 Scope

For the complete realization of the project, it will include the following tasks:

• Study of the state of the art of liquid propellant rocket engines

• Calculation of optimal geometry for the nozzle

• Study of an appropriate cooling system

• Design of the nozzle and its cooling system through CAD software

• Study and selection of appropriate material and manufacturing techniques

• Analysis of the nozzle flow through CFD simulations

• Structural and thermal study through FEM simulations

• Production of manufacturing drawings

• Economic feasibility study and comparison with other manufacturing methods

• Environmental impact study

• Drafting of a final report with the explanation of the developed tasks and conclusions

1.3 Requirements

The requirements of the project will consist of:

• 3D design software: SolidWorks

• 2D CFD software: OpenFOAM

• 3D CFD and FEM software: ANSYS

• Thrust for full-scale model: 2 MN

• Combustion chamber pressure: 300 bar

1
1 INTRODUCTION

• Combustion chamber temperature: 3550 K

• Geometry optimized for a height of 3000 m

• Designed for the use of methane and liquid oxygen as propellants

• Cooling system able to withstand operation conditions

1.4 Justification and utility

With the significant amount of research and development that metal additive manufacturing has been
receiving during the last decade, prior manufacturing techniques are becoming obsolete. Especially
in the aerospace sector, the high precision, automatizing and low production times that these new
technologies bring with them could provide the next step in the production of aeroplane components
or rocket engines, among others.

While these technologies are in earlier stages of development, they will replace most of alternative
production systems in the near future. As these Additive Manufacturing (AM) technologies will be
able to provide better precision and even the ability to manufacture multi-material structures, the
components that have been designed with the restrictions of obsolete technologies in mind can and
will be updated to the possibilities that the emerging technologies provide.

In rocket engine nozzle design, the introduction of these technologies will mean that more complex
geometries can be attained, which will have a direct effect on their performance and efficiency. On top
of that, the cooling systems of these nozzles, which are limited by actual manufacturing techniques,
could be re-designed thanks to the lower tolerances that these AM technologies provide, which would
lead to a higher re-usability of these engines.

On that last note, the emergence of private space companies (e.g.: SpaceX) which have put great
efforts into the re-usability of rockets in order to maximize benefits will lead to a swifter transition
to more efficient manufacturing techniques. The automation that these AM technologies provide will
lead to a faster and more efficient production of components with a more optimal use of resources.

In order to provide more depth to the study of the emergence of metal AM technologies, this project
will study to what extent they can have an impact on LPRE design and production.

2
2 STATE OF THE ART

2 State of the art

2.1 Liquid Propellant Rocket Engines

The reference rocket engine is the SpaceX Raptor, which is a Liquid Propellant Rocket Engine
(LPRE) that runs on methane and Liquid Oxygen (LOX), as these can be produced at the surface
of Mars [1] and one of the objectives of such engines is to power the first human-manned journey to
the Red Planet.

These engines use chemical combustion energy to expand the combustion gases in a supersonic
nozzle. The internal energy of the propellants is used in a high-pressure chamber to heat the reaction
gases to temperatures up to 4100 ◦C. These gases are then expanded in a supersonic nozzle and
accelerated to high velocities (1800-4300 m/s), thereby imparting momentum to the vehicle. These
nozzles consist of a converging section to accelerate the gas to the Mach number and a diverging
section to continue to accelerate the fluid once it reaches that point. As the gas temperatures are high
above the melting point of the build materials, it is necessary to cool these elements in most cases. [2]

While it is a recent alternative, methane as a propellant provides a good trade-off in many aspects,
while not being the most optimal option in all of them, as it can be seen in Table 1.

RP-1 Methane Hydrogen

Density (g/L) 813 422 70

Oxidizer:Fuel Ratio 2.7:1 3.7:1 6:1
1 Liter of LOX:
1:0.52 1:0.73 1:2.7
X Liter of Fuel
Efficiency
370s 459s 532s
(Max. theoretical ISP)
Combustion Temp. (K) 3670 3550 3070
Boiling Point (K) 490 111 20
CO2 &H2 0
Combustion Byproducts CO2 &H2 0 CO2 &H2 0
& Unburnt Carbon
Manufacturable on Mars No Yes Yes

Table 1: State of the art LPRE propellant comparison [3]

As it can be seen in Table 1, methane provides a mid-point alternative in all the various aspects to take
into consideration when designing a rocket. First of all, a denser fuel will need smaller and lighter
tanks for a given mass of propellant, which will result in a lighter rocket, which is optimal. If we
take into consideration the oxidizer-to-fuel ratio, for one litre of LOX, one would need a little over

3
2 STATE OF THE ART

half a litre of RP-1, 0.73 litres of liquid methane and 2.7 litres of hydrogen. That way, fuel tanks
for methane need to be a bout 40% bigger than they need to be for RP-1. Compared to hydrogen,
methane tanks would need to be 3.7 times smaller.

Another variable to take into consideration is the rocket’s efficiency, which is measured by the specific
impulse or ISP. This number relates the amount of fuel necessary to produce the same amount or thrust
and it can be thought about as the fuel economy of a rocket. The numbers displayed in Table 1 are
calculated considering ideal conditions and are much lower than they would be on a real rocket engine
(e.g., 350s for the Merlin 1D vacuum which runs on RP-1). Here, we can see that a hydrogen is the
best option while methane-powered engines provide a mid-range alternative. If one takes a look at the
combustion temperature, lower temperatures will be easier on the engine and will potentially make
for a longer lifespan. Here, hydrogen is the best option. In terms of storage, it is useful to examine
the propellant’s boiling temperature, as a higher boiling temperature will make it easier to store the
fluid. Here, hydrogen has a near absolute zero boiling point, while methane has a similar boiling point
as LOX, which means that similar storage systems can be applied to both methane and LOX. If one
considers the combustion byproducts, the use of RP-1 is the only option that pollutes with unburnt
carbon, even though that water vapor pollution is a serious issue in upper levels of the atmosphere
[4].

In conclusion, it can be seen that hydrogen would be the most efficient and powerful propellant, were
it not for its many drawbacks. Hydrogen is difficult to store (i.e., low density, low boiling point)
which leads to leaks and an over-complicated storage system while also being relatively expensive to
refine, store and transport. While engines such as the Space Shuttle Main Engine (RS-25) have opted
for hydrogen as its propellant, more recent trends in the industry point towards methane as being
the next big implementation in LPRE manufacturing. While providing good storage properties, it is
more efficient than the mainstream RP-1. On top of that, its production can be easily implemented in
other planets, and it is the main reason companies like SpaceX, which are leading the Mars race, are
starting to implement methane-propelled rocket engines (e.g., Raptor Series).

Having taken a look at the different propellant options, a comparison of the state-of-the-art LPREs
will be performed so as to compare them to SpaceX’s Raptor. Some of the most important data to
take into consideration is displayed in Table 2.

Raptor Merlin RD-180 F-1 BE-4 RS-25

Closed Closed Closed Closed

Cycle Open Open
(Full Flow) (LOX rich) (LOX Rich) (Fuel Rich)
Fuel Type Methane RP-1 RP-1 RP-1 Methane Hydrogen
Total Trust
2.00 0.84 3.83 6.77 2.40 1.86
(MN)

4
2 STATE OF THE ART

Raptor Merlin RD-180 F-1 BE-4 RS-25

Specific Impulse
330 282 311 263 310 366
at SL
Chamber
300 97 257 70 135 206
Pressure (bar)
Price ∼$2M <$1M $25M $30M ∼$8M >$50M
Reusability
50 10 No No 25 19
(n. flights)
Potential cost
(per flight $20 $117 $6527 $4431 $133 $1414
per kN)

Table 2: State of the art rocket engine comparison [3]

Being one of the most recent LPREs, the SpaceX Raptor is aimed at being the most efficient,
modern and cost-effective option to date. First of all, it is one of the first rocket engines to
effectively implement ’full flow’ pump systems. This cycle consists of two preburners which use
non-stoichiometric fuel-to-oxygen ratios to produce controlled combustions that drive the fuel and
liquid oxygen into the combustion chamber at extremely high pressures, which result in higher thrust
and efficiency. Furthermore, as it can be seen in Table 2, the Raptor has the highest chamber pressure
in any other LPRE, which is possible thanks to its pump system. On top of that, while not being
able to output the most thrust, the Raptor has the highest efficiency. This fact relates to the trend of
opting for several smaller rocket engines instead of one (or a few) bigger units. Another important
point is the cost of these engines. While most of them were developed by public agencies like the
NASA or the Roscosmos, the introduction of private companies to the space sector has resulted in
a considerable decrease in expenses. This has been achieved mostly through the re-utilization of
launchers and rocket engines. As Table 2 shows, modern rocket engines like the Raptor or Merlin
engines are designed so as to be able to operate several times with little to no refurbishment, while
older engines were designed only to last during a specific mission. This has driven down the cost
per kN of thrust from $4431 for the F-1 to a merely $20 for the Raptor, which equates to a 99.5%
decrease in price per unit of thrust.

Various other aspects have made the increment in efficiency and decrease in cost possible. On the
one hand, higher chamber pressures and temperatures, which in turn have provided modern LPREs
with higher efficiencies, have been achieved through the implementation of modern pump systems,
high thermal conductivity [5] and high-strength materials and complex cooling systems, which will
be dealt with in Section 2.4. On the other hand, the implementation of modern additive manufacturing
techniques and construction techniques which take into consideration the re-utilizable nature of state

5
2 STATE OF THE ART

of the art rocket engines are the main reasons of the reduction in the cost of space travel [6]. These
AM techniques provide a better utilization of material, reduced production times and costs and the
possibility to automate every step of the production, which provides a more optimal solution than
actual methods [7].

2.2 Rocket Engine and Nozzle Theory Fundamentals

The basic principles of rocket propulsion are those of mechanics, thermodynamics and chemistry.
These will be reviewed so as to provide a general view on the matter and enumerate the expressions
which will be used later in the development of the project.

In the following image, the pressure distribution as well as the velocity of exhaust gases is plotted in
a diagram of a rocket engine 1.

Figure 1: Rocket engine diagram [2]

Using the control volume approach, one can reach the following expression for the thrust of a rocket
engine:

F = ṁv2 + (p2 − p3 )A2 (1)

The effective exhaust velocity c represents an average at which propellant is being ejected from the
rocket vehicle, and it is defined as:

c = v2 + (p2 − p3 )A2 /ṁ (2)

Which is in turn related with the total impulse It , found by integrating the thrust force over the time
of its application:

Z t
It = Fdt = c/g0 (3)
0

6
2 STATE OF THE ART

with g0 being the acceleration of gravity

In order to evaluate the performance of a rocket system and determine its size and generic shape, it is
necessary to use the principles of gas dynamics and thermodynamics. To do that, some assumptions
have to be valid:

• The working fluid is homogeneous.

• The working fluid obeys the perfect gas law.

• The flow is adiabatic.

• There are no friction losses, shock waves or discontinuities within the flow.

• The propellant flow rate is steady and constant.

• All exhaust gases leave the nozzle with a velocity parallel to the nozzle axis

• The gas velocity, pressure, temperature and density are all uniform across any section normal
to the nozzle axis.

• Chemical equilibrium is established in the combustion chamber.

• Propellants are stores at ambient temperatures.

These assumptions will permit the derivation of a quasi-one-dimensional set of equations which will
be used in the initial calculations of the project.

The minimum nozzle cross-sectional area is commonly referred to as the throat area At . The ratio of
the nozzle exit area A2 to the throat area At is called the nozzle expansion area ratio ε:

ε = A2 /A1 (4)

The maximum gas flow rate per unit area occurs at the throat, and as this is where the flow reaches
its Mach number, the pressure, temperature, velocity and optimal area at this point are completely
defined:

pt /p1 = [2/(k + 1)]k/(k−1) (5)

Tt = 2T1 /(k + 1) (6)

r
2k
vt = RT1 (7)
k+1
s
ṁ rT1
At = (8)
p1 k[2/(k + 1)](k+1)/(k−1)

7
2 STATE OF THE ART

Using the throat area, the pressure at a certain point in the nozzle and the ratio of specific heats k, the
following expression can be used to find the area of that particular point:

v "
 1/k−1  1/k u  (k−1)/k #
At k+1 py uk +1 py
= t 1− (9)
Ay 2 p1 k−1 p1

Using this expression for a certain exhaust pressure will yield the required initial geometry traits for
an optimal nozzle.

Finally, the exhaust velocity will be needed, which is also completely defined:

v "
u
u 2k  (k−1)/k #
p2
v2 = t RT1 1 − (10)
k−1 p1

With this set of equations and the operation altitude, required levels of thrust, combustion pressure
and temperatures and exhaust gas composition one is able to define the initial parameters of an ideal
nozzle.

2.3 Optimal Nozzle Geometry

Even though there are many available nozzle configurations, the bell-shaped nozzles are the most
commonly used today. They are characterized by a high expansion angle immediately downstream
of the nozzle throat followed by a gradual reversal of the slope to exit divergence angles of less than
10º as it can be seen in Figure 2. These are more efficient than rudimentary conical-shaped nozzles,
as they provide the same amount of expansion using far less longitude, and therefore less mass.

Figure 2: Nozzle geometry diagram [2]

8
2 STATE OF THE ART

Regardinf the diverging section, changing the direction of supersonic flows with an expanding wall
geometry requires expansion waves. As the flow is turned in the opposite direction between points I
and E, oblique compression waves occur, which result in an energy loss.

To tackle this issue, the method of characteristics is used, in which the nozzle geometry is calculated
so that the wall coincides with points in space where the expansion and compression waves balance
each other out. That way, the flow is effectively turned near axial without strong shock waves and
consequent loss of energy.

The problem with the geometries produced by the method of characteristics is that they yield
excessively long nozzles. However, different methods of nozzle truncation have been studied which
only slightly compromise nozzle performance, with losses in performance that range from 0.02 to
0.34 percent [8].

A compressed truncated nozzle contour is obtained by truncating a perfect nozzle contour (obtained
from the method of characteristics) and linearly compressing the truncated perfect nozzle contour in
the axial direction to obtain the desired nozzle length as seen in Figure 3.

Figure 3: Compressed nozzle design [8]

The procedure to obtain a perfect compressed nozzle is as follows. Firstly, one obtains point E’ where
the required exit diameter is met. Then, one compresses every point of the resulting contour with the
compression factor C, given by

xe“ − xa
C= (11)
xe‘ − xa

And because of the fact that this procedure creates a discontinuity in the nozzle wall slope in the
attachment point A, this points are shifted in the x direction until the compressed contour is tangent
to the throat contour [8].

9
2 STATE OF THE ART

It also has to be taken into consideration that the converging nozzle section, as well as the throat
contour itself, do not significantly affect nozzle performance [2], as the flow can easily be accelerated
to Mach regimes with basic conical sections. Therefore, no real method will be used to define these
sections other than approximation.

2.4 Cooling system

As rocket nozzles have to endure increasingly harsher conditions in order to improve performance,
improvements in cooling systems are required. While cooling of the diverging section of nozzles can
be achieved by passive methods such as radiative cooling (e.g., SpaceX Merlin), more critical parts
of the engine, like the combustion chamber and throat have to be actively cooled [9].

The regenerative cooling method has arguably been, for the last decades, the preferred option by
manufacturers. This method consists in driving the propellant (stored at cryogenic temperatures)
through channels in the walls of the nozzle, providing an effective cooling strategy while raising the
propellant’s energy before being used in the combustion, which is convenient in terms of combustion
[10].

These cooling channels are usually rectangular-shaped, as it provides the best surface-to-area ratio, as
it can be seen in Figure 4. As one might imagine, increasing the number of channels will result
in better cooling performance. It has also been shown that the increase in the channels’ aspect
ratio increases the coolant’s velocity and therefore its heat transfer capabilities. Thus, increasing
the channels’ aspect ratio from 1 to 8 can lead to up to a 50% increase of the coolant heat transfer
coefficient in transcritical fluids [11]. Accordingly, the most efficient aspect ratios have been shown
to be between 4 and 8.

Figure 4: Schematic view of cooling channels [12]

10
2 STATE OF THE ART

However, it has also been shown that these techniques greatly increase the pressure drop across the
system, which in turn needs to be powered by a sufficiently powerful pump system. In order to tackle
this issue, variable cross-sectional area channels are used, using a lower aspect ratio for less critical
regions of the nozzle. This way, the pressure drops are greatly reduced without compromising the
cooling system efficiency [12].

Recent studies show that hot-gas-side wall heat flux and temperature are non-uniform along the length
of the nozzle [13], being substantially higher in the throat region. This means that higher coolant heat
flux is needed in these critical areas, which can be achieved by increasing the channel aspect ratio or
by increasing the number of cooling channels by the bifurcation of the aforementioned.

2.5 Manufacturing techniques

Nozzle manufacturing was once based around tubular wall designs (Figure 5 a), made of singularly
formed tapered tubes that were bonded together (e.g., Space Shuttle Main Engine). Nevertheless,
this technique imposes the use of stainless steel or nickel alloys (e.g., Inconel) with low thermal
conductivity, which limits the maximum sustainable heat flux and thus the maximum chamber
pressure.

Nowadays, the channel-wall design is the most predominant practice (Figure 5 b). This type of
cooling system is achieved by machining rectangular slots from a relatively thick chamber. An outer
wall is bonded to the ribs by processes like brazing or electro-deposition. With this technique, the
inner and outer walls can be made of different materials. Normally, a higher thermal conductivity
material like a copper alloy is used for the inner wall, while mechanical stiffness is achieved by using
high-strength materials for the outer wall (e.g., nickel alloys) [14].

(a) Tubular wall (b) Channel wall

Figure 5: Thrust chamber cooling designs [14]

However, this technique has its limitations, as the milling machines that are used to machine the
coolant channels limit their dimensions to thicknesses greater than 1mm. As it has been explained,

11
2 STATE OF THE ART

the objective is to decrease channel thickness to increase aspect ratio, and therefore cooling efficiency.
Moreover, these production techniques often prove labor-intensive, economically costly and result in
a high part count. These areas are where modern additive manufacturing techniques provide a better
solution than previous approaches.

These novel approach provides opportunities for the use of new materials as well as weight
optimization and the ability to design complex shapes and geometries that would otherwise be
impossible to create. On top of that, AM designs may benefit from a considerable reduction
of production time when compared to more traditional methods. This would explain why many
companies are starting to invest in the development of such technologies, as they are referred to as
”key technology approach”. Among the many different metal additive manufacturing techniques that
have surged in the recent years, the most predominant techniques have been through laser powder bed
fusion (L-PBF) and directed energy deposition (DED) [7].

Laser powder bed fusion, or also known as ”Selective Laser Melting” or ”Direct Metal Laser
Sintering” is one of the most characteristic approaches to metal AM techniques. The metal feedstock
is provided in powder form, which is deposited and melted layer-by-layer on a build plate. A laser
beam provides necessary energy at micro-scale focus, melting the powder, and after one layer is
finished, the build plate is lowered, re-coated with powder and the cycle restarts. This is done in
an inert environment, with layer heights ranging from 20-100 µm. As this type of manufacturing is
enclosed in the powder deposition area, it is clear that the main limitation of this techniques is its
build size. One of the key parameters is the minimum wall thickness (which will in turn limit the
number of cooling channels), which in the case of L-PBF has been studied (e.g., Patel et al. [15]) to
be around 0.6 mm for Inconel 718. Below that size, thicknesses will not often be repeatable or will
result in excessive porosity [7].

Another limitation of this technique is that it can only use one type of metal for a particular part.
Given the needs of the elements at hand, this technology (and other AM techniques) has been
demonstrated in materials with high strength (e.g., Inconel) and high conductivity (e.g., C-18150).
While superalloys are in use (e.g., SpaceX SuperDraco), they are not an optimal material for high
performance combustion chambers as they usually lack the required thermal conductivity. On the
other hand, copper alloys such as the C-18000 and GRCop series, while having less specific strength,
provide much better thermal conductivity.

Since 2014, NASA has successfully implemented L-PBF techniques using GRCop-84 (Cu-8 at.%
Cr-4 at.% Nb), GRCop-42 (Cu-4 at.% Cr-2 at.% Nb), and C-18150 copper thrust chambers. These
alloys are high conductivity, high-strength, dispersion strengthened copper alloys for use in high
temperatures and high heat flux applications, which make them suitable for the job at hand. The only
real drawback is a low cycle fatigue life of the additive manufactured specimens, which will likely
be resolved as the technologies advance. This series of copper alloys provides several advantages

12
2 STATE OF THE ART

compared to previous copper alloys, such as:

• Resistance to oxidation and blanching during thermal and oxidation-reduction cycling

• High maximum use temperature of 800ºC

• Good mechanical properties at high temperatures.

• Lower thermal expansion to reduce thermally induced stresses and low cycle fatigue

On the other hand, in Directed Energy Deposition , the feedstock is deposited from a deposition head
through powder or wire feedstock: while the former provides better tolerances at the cost of being
time-consuming, the latter has a higher deposition rate but cannot reach the same level of resolution.
In comparison with L-PBF, the feedstock is deposited locally to create a freeform part, which solves
the size manufacturing limitations of the latter. That is why DED processes are more suitable for
large thrust chambers and channel-cooled nozzles. Furthermore, this technology provides the ability
to manufacture bimetallic parts, which in turn solves the strength to thermal conductivity compromise,
as well as monolithic [7].

While bimetallic component manufacturing through DED is still in initial stages, several instances
have been studied in which successful bonds have been created between the different metals. On
the other hand, monolithic structures have been researched further and have consequently had more
significant results, which are limited by the robotic system being used in the deposition head. On
this note, NASA has demonstrated the use of the DED process in various large scale integral channel
nozzle structures using nickel and iron-based super-alloys. Once printing and post-processing was
concluded, these demonstrator units showed less than 0.5 mm deviations from nominal geometry
using 3D scanning [7]. Moreover, these units have been shown feasible through repetitive hot-fire
testing, through which no manufacturing-related faults were perceived.

However great the benefits of this technology over traditional manufacturing techniques may be, it
has its limitations, as it is still in development. It is estimated that thrust chambers created using
laser powder directed energy deposition have a technology readiness level ≥ 6. One of the most
important limitations of DED technologies is the inability to control the surface roughness of the
interior of cooling channels, which can result in unacceptable pressure drops along them. On top of
that, minimum wall thickness will also be a major constraint, as the goal is to achieve the thinnest
channels possible and thin walls could result in porosity and lack of repeatability. In order to solve
this issues, companies like SpaceX are using these AM techniques in the production of monolithic
structures that are then traditionally machined to form the cooling channels. This is the case for the
Merlin and most recent Raptor engines, which use a DED manufactured copper alloy main chamber
on which cooling channels are machined. Finally, an Inconel jacket is electro-blazed on top of the
main chamber [16].

13
2 STATE OF THE ART

2.6 Materials

As it has been seen in the previous section, the main materials that are being used in LPRE nozzle
designs are the Inconel and the C18150 series. These two materials will be studied in depth
so as to gather an understanding of their benefits over other materials. Other materials cannot
be additively manufactured because the melting and solidification dynamics during the printing
process cause intolerable microstructures with periodic cracks. However, efforts are being made
by introducing nanoparticles of nucleants that control solidification during manufacturing of high-
strength aluminium alloys [17]. Consequently, while the aforementioned metals are the state-of-the-
art build materials, other types of better performing alloys could be available in the near future.

On the one hand, the Inconel series is a family of austenitic nickel-chromium-based super-alloys. This
family is designed to resist a wide range of corrosive environments while displaying exceptionally
high yield, tensile and creep-rupture properties at high temperatures. That is why it is widely used
in gas turbine engine parts, LPRE components and other applications in the aerospace sector. The
specific composition of Inconel 718 can be seen in Table 3.

Ni Cr Fe Mo Nb & Ta Co Mn

50.0–55.0 17.0–21.0 Balance 2.8–3.3 4.75–5.5 1.0 0.35

Cu Al Ti Si C S P

0.3 0.2–0.8 0.65–1.15 0.35 0.08 0.015 0.015

Table 3: Inconel 718 composition by mass

The tensile and thermophysical properties of additively manufactured Inconel 718 which will be used
in later stages of the project can be seen in Table 4. These have been extracted from Reference [18].

Ultimate
Yield Coef. of Thermal
Density tensile Poisson’s
Alloy E (GPa) strength thermal conductivity
(g/cm3 ) strength ratio
(MPa) expansion (W/M·K)
(MPa)

IN718 8.192 192 998 1162 0.3 1.2e-5 11.2

Table 4: Additively manufactured Inconel 718 tensile properties [18]

While Inconel could provide the necessary tensile properties for its use in the harsh conditions of a
rocket engine nozzle, it lacks the required heat transfer capabilities. This way, it is often combined
with copper-based alloys in bimetallic structures. Most importantly, the C-18150 series are copper

14
2 STATE OF THE ART

alloys with an attractive balance of high temperature strength, creep resistance, low cycle fatigue life
and thermal conductivity. More specifically, the C-18150 alloy will be studied, the composition of
which can be seen in Table 5.

Alloy Cu Cr Zr

C-18150 Balance 1 0.15

Table 5: C-18150 composition by mass

The following table contains the tensile and thermophysical properties of C-18150, which have been
extracted from [19].

Ultimate
Yield Coef. of Thermal
Density E tensile Poisson’s
Alloy strength thermal conductivity
(g/cm3 ) (GPa) strength ratio
(MPa) expansion (W/M·K)
(MPa)

C-18150 8.89 117 >379 >379 - 1.71e-5 323.4

Table 6: C-18150 properties [19]

When comparing the two options, a good understanding of why Inconel and C-18150 alloys are used
in combination can be achieved. While having similar densities, C-18150 alloys provide much higher
thermal conductivity with an elastic modulus that is comparable to that of Inconel 718 alloys.

2.7 Approach and selection of alternatives

After thorough research on the matter of Liquid Propelled Rocket Engines, several decisions have
been taken that will define the characteristics of the project.

First of all, the design approach of the nozzle geometry will be that of a De Laval nozzle, with a
cylindrical section for the combustion chamber and its respective converging, throat and diverging
sections. This last section will be designed following the Bell shape approach, and its geometry will
be calculated through the ’method of characteristics’ so as to obtain the most efficient nozzle possible.

Given the trend towards the use of methane as the new state of the art LPRE propellant, this nozzle
will be designed with its use in mind. This fact will shape all the calculations that will be made, as
the thermodynamic characteristics of the exhaust gases will be major variables on the development of
the flow.

15
2 STATE OF THE ART

The cooling system of the nozzle will be designed following the regenerative cooling approach. This
will mean that the cryogenic methane will be forced through channels in the nozzle walls, helping
to reduce the temperature of the nozzle while increasing its energy and also help achieve a more
efficient combustion. Using the AM techniques studied in Section 2.5, the number and geometry of
these cooling channels can be designed for a more optimised cooling to take place, which can be used
to opt for more radical design concepts that can lead to better performance (e.g., higher combustion
chamber pressure and therefore higher thrust, among others).

While bimetallic structures are being developed at the moment through AM technologies, these are
still in preliminary stages. Consequently, the nozzle will be a monolithic structure, which will have
to be strong enough to sustain the high pressures but also be conductive enough so as to evacuate
enough heat so that the high temperatures don’t lead to catastrophic failure. For this reason, C-18150
has been chosen as the design material, as the other options (i.e., Inconel series), while having a
higher specific strength, would lack the thermal conductivity required in order to operate under the
maximum temperatures.

Once the nozzle has been designed taking into consideration the previous basis, a study of the flow
through the nozzle, as well as its effect on the aforementioned will be developed. This study will be
firstly developed as an axisymmetric 2D case which will then be brought to 3D for a more realistic
study of the behaviour of the flow. The results of these simulations will then be further used in
structural and thermal analysis on the nozzle. The results of these simulations will finally be used in
order to asses the suitability of the previous decisions on the problem at hand.

16
3 DEVELOPMENT OF THE SOLUTION

3 Development of the solution

3.1 Nozzle design

First of all, an efficient geometry will be calculated so as to accelerate the fluid from the combustion
chamber to the exit of the nozzle. Using this geometry and once the cooling system has been defined,
the complete design of the nozzle will be produced, including its regenerative cooling system.

3.1.1 Initial calculations

Through thorough research on the functioning and characteristics of SpaceX’s latest Raptor engines
[16], the data in Table 7 has been found. These parameters will be used in order to calculate the
initial dimensions of the geometry of the nozzle. Moreover, these will also be used as the boundary
conditions of our CFD simulations in later sections.

Initial Parameters

T 2 MN
p1 300·105 Pa
T1 3550 K
k 1.3
R 370 J/kgK
p2 60·103 Pa

Table 7: Initial design parameters

It should be noted that the thermodynamic constants have been calculated for typical methane
combustion gases [20] and the exit pressure has been calculated for an altitude of 4000m through
the International Standard Atmosphere [21]1 .

Using the thermodynamic analytical relations provided in Section 2.2, a MATLAB code which can be
found in the Appendix document has been created in order to obtain the following key dimensional
parameters:

1 With subscripts 1, p and 2 denoting parameters at combustion chamber, throat, and exit sections respectively.

17
3 DEVELOPMENT OF THE SOLUTION

Obtained parameters Official SpaceX data [16]

Ar 30.61 34.34
ṁ 645.59 kg/s -
Tt 3087 K -
Dt 0.2169 m 0.2216 m
v2 2943 m/s -
D2 1.2 m 1.3 m
Is 300.1 330

Table 8: Obtained parameters comparison with avaliable data

When compared to the official SpaceX data, one can see that these first approximations are
relatively accurate. One should note that the equations that have been used are derived from several
assumptions, and do not exactly represent the most efficient design. However, they will be used as
initial design requirements that will be complemented through the use of other techniques.

3.1.2 Nozzle geometry determination

With the acquired data, a program developed by K. Palaveev et al. [22] has beed used so as to obtain
the divergent region of the nozzle geometry. This code uses the method of characteristics mentioned
in Section 2.3 to find the wall region which will minimize the creation of shock waves and therefore
provide the most optimal flow possible. The setup parameters can be seen in the Appendix document.
The number of characteristic lines provides more resolution on the geometry when increased, and
the number of iterations has been studied to converge before the selected number [22]. The initial
expansion ratio and characteristic net length have been selected through an iterative process so as to
find an optimal nozzle length and exit diameter, the geometry of which can be seen in Figure 6. That
way, the set of setup parameters has been chosen so that the exit diameter is slightly greater than the
one obtained in the previous section, as the truncation method will result in the reduction of these
dimensions.

18
3 DEVELOPMENT OF THE SOLUTION

Figure 6: Calculated nozzle geometry before truncation. Obtained with [22]

This method provides an excessively long nozzle, which has been truncated using the method
mentioned in Section 2.3 by J.D. Hoffman [8]. The truncation coefficient has been selected so that
the divergent part of the nozzle has a longitude and exit diameter similar to that of the actual Raptor,
which are approximately 1.5 m and 1.3 m respectively. This truncation coefficient can be seen in
Equation 12.

1.58 − x0
C= (12)
2.598 − x0

Note that 2.598 is the x dimension at which the required exit diameter of 1.3m is reached, as explained
in [8]. Every point of the geometry has then been truncated with the truncation coefficient and a factor
accounting for the displacement of the attachment point equal to 0.01m.

The rest of the dimensions have been extracted from SpaceX official data [16] or reverse-engineered
in the regions where geometry is not critical for flow development. The resulting geometry for the
nozzle outline can be seen in Figure 7.

19
3 DEVELOPMENT OF THE SOLUTION

Figure 7: Final geometry for the nozzle outline

3.1.3 Nozzle and cooling system design

Before using the calculated geometry to implement a final design, the cooling system characteristics
shall be defined, which will in turn affect the final nozzle wall width.

The general disposition of the cooling system will follow that of the Raptor engine. In essence, the
propellant (in this case, liquid methane) will be introduced into the walls at an intermediate point of
the nozzle. This fluid will be driven by a pump in two different directions: upward into the combustion
chamber region and downward into the divergent part of the nozzle. The two flows will have two
different outlets. The upward outlet will be located at a midpoint of the combustion chamber, while
the downward outlet will be placed in an intermediate point of the diverging section. This means that
the downward flow will circulate to the end of the nozzle to then flow in the opposite direction on a
second level of the nozzle. This schematic is shown in Figure 8

20
3 DEVELOPMENT OF THE SOLUTION

Figure 8: Schematic of the cooling system. Courtesy of Elisei Maslov

Taking into consideration the limitations of modern AM technologies provided in Section 2.5, the
dimensions of such cooling system have been chosen. Two different sections have been designed for
the cooling channels. A higher aspect ratio of 6 has been chosen for the channel in the throat, as it
is the region with the greatest heat flux and therefore higher wall temperature [13]. A higher aspect
ratio will increase the velocity in a channel with constant pressure difference, which will increase heat
transfer to the coolant. For regions with lower wall temperatures, channels with an aspect ratio of 4
have been chosen. This duality of aspect ratios will lead to reduced required pressure differences and
therefore a smaller pump system, and can be seen in Figure 9a.

After studying the available studies and comparing with previously designed rocket engine cooling
systems, a channel height of 3 mm has been chosen across all the channels, with a 1.5 mm wall
between the channels and the hot gas side wall and exterior wall. Therefore, the nozzle wall width
is 6 mm, and double that in the region of downward flow of the system. Moreover, the inlet and
outlet pipes have been designed according to the manufacturer’s specifications. The geometry and
arrangement of these channels can be seen in Figure 9b and 9c.

21
3 DEVELOPMENT OF THE SOLUTION

(a) Detail of designed variable aspect ratio channel

(b) Detail of designed regenerative cooling channel inlet arrangement

(c) Detail of cooling channel outlet and channel u-turn

22
3 DEVELOPMENT OF THE SOLUTION

In order to be coherent with the limitations mentioned in Section 2.5, a number of 550 cooling
channels keeps all the dimensions below the required tolerances. This amount of cooling channels
would translate to a 40% increase in the number of cooling channels when compared to a milled
channel wall design nozzle like the one on the Space Shuttle Main Engine [14].

The resulting nozzle can be seen in Figure 10.

Figure 10: Final nozzle design

Using the SolidWorks mass properties feature and the parameters for the chosen build material (C-
18150), the design occupies 0.05 cubic metres and weighs 473.01 kg.

3.1.4 Drawings

Given the actual tolerances of the DED technologies, the technical drawing of the designed nozzle has
been manufactured and can be seen in figure 11. This technical drawing won’t serve as direct means
for nozzle production, as AM technologies would only need an .STL file in order to manufacture the
component. However, the dimensions of the most important geometries are clearly stated so as to
provide a graphical understanding of the different geometries.

23
3 DEVELOPMENT OF THE SOLUTION

Figure 11: Final nozzle technical drawing

24
3 DEVELOPMENT OF THE SOLUTION

3.2 Nozzle CFD simulations

In order to assess the performance of the designed nozzle, Computational Fluid Dynamics simulations
will be used in order to simulate the operating conditions of the fluid. The resulting flow pressure,
temperature and velocity will be used in subsequent structural and thermal FEM simulations. Firstly,
a 2D axisymmetric analysis will be performed using OpenFOAM with which a first approximation of
the flow will be obtained. Following this first results, the case will be brought to 3D using ANSYS,
which will provide a more accurate representation of the flow as well as the necessary tools for a more
in-depth analysis of the results.

3.2.1 2D simulations

Meshing

In this 2D case, the fluid domain is comprised of the interior of the nozzle and an outer
section for the development of the flow. This domain has been designed and meshed using the
’ofblockmeshdicthelper’ python library [23], creating a wedge of 5º of the nozzle geometry and after-
nozzle region through splines and several blocks for region refinement. This mesh can be seen in
Figure 13, with a total count of 7935 elements. As it can be seen, a finer mesh has been chosen for
the most critical parts of the flow (i.e., the region directly behind the nozzle).

Figure 12: 2D simulation mesh

25
3 DEVELOPMENT OF THE SOLUTION

Simulation setup

This simulation has been performed using a laminar model so as to ease the attainment of results.
Taking into consideration this is a compressible case, the ’rhoCentralFoam’ solver has been used. The
flow has been simulated as a transient case up to 0.25s, time in which the flow was fully developed
and no major changes could be observed in the fluid domain.

Regarding the boundary conditions, the following have been imposed:

• Inlet: pressure inlet at the combustion chamber of 300 bar of pressure and a temperature of
3500 K

• Outlet: atmospheric pressure (101325 Pa) and temperature (300 K)

• Walls: slip wall condition

• Axis: symmetry condition

Regarding the simulation control parameters, adjustable run time has been used with a maximum
Courant number of 0.5 so that the program could decrease the time-step in case divergence occurred.
An initial time-step of 10− 7 s has been chosen. Other system and constant parameters can be found
in the Appendix document.

Simulation results

After the simulations were completed, the following contour plots have been retrieved:

26
3 DEVELOPMENT OF THE SOLUTION

Figure 13: 2D simulation results

27
3 DEVELOPMENT OF THE SOLUTION

In these Figures, the scales have been modified so as to provide a better understanding of the flow
conditions. While the velocity scale is the global scale, the pressure scale should go up to 3 · 107 Pa
and the temperature, up to 3500 K.

In these contours, the over-expansion of the nozzle can clearly be observed, which produces a
concertina effect that dissipates throughout the length of the flow. These ’mach diamonds’ can also be
seen in the pressure contours taking the form of pressure maximum and minimum regions that occur
one after the other. In the nozzle area, it can be seen how the flow rapidly accelerates to the exhaust
velocity of ∼ 2600 m/s, which is while the pressure (which is imposed at the inlet with a value of
3 · 107 Pa) goes down to near 0. The regions of the flow with higher pressures are also the ones with
higher temperature, which is coherent with the thermodynamic relations explained in Section 2.2.

Simulation verification and validation

In order to verify these simulations, several checks will be performed.

First of all the mass imbalance has been calculated, and can be seen in Table 9.

Surface Mass rate (kg/s) Velocity (m/s)

Inlet 21.1366 186.26
Outlet 21.4915 2415.14

Mass Imbalance 0.3549 kg/s

Table 9: Mass imbalances for the 2D case

As it can be seen, the mass imbalance represents 1.65% of the outlet mass flow rate, which can be
considered sufficiently accurate for this type of initial simulation.

To further verify this case, two additional simulations were performed with a 20% finer mesh at
each step. The mass imbalances of these simulations, together with the first one can be seen in the
following table:

Simulation element count Mass imbalance (kg/s)

7935 0.3549
9126 0.2893
10495 0.2237

Table 10: Mass imbalance through mesh refinement

In this table, one can see how refining the mesh provides much more accurate depictions of the flow,

28
3 DEVELOPMENT OF THE SOLUTION

which will likely converge to 0 as the number of elements is increased.

In order to validate this simulation, the produced thrust will be calculated and compared to SpaceX’s
official data. In order to do that, the following equation will be used, extracted from Section 2.2.

T = ṁ · voutlet · nwedges = 3.73MN (13)

This thrust is in the order of magnitude of the expected performance of the real Raptor engines
(∼ 2 MN). Nonetheless, this is too great of a thrust output, which can be explained by the fact
that this is a 2D case and substantial error can be obtained and the fact that the slip wall condition
induces a far greater mass flow rate, which in turn would provide excessive levels of thrust compared
to the real engine.

3.2.2 3D simulations

As this is an axisymmetric problem, the flow conditions will be simulated on half of the model.
Therefore, a denser mesh has been used given the node number limitation, which stands at 500.000
for the Student version of ANSYS, which has result in more consistent results.

Meshing

Given that the simulations are to take place in sea level conditions, mach diamonds are expected to
occur, as the nozzle is over-expanded. This means that the main flow will occur right behind the exit
of the nozzle. A denser mesh, therefore, shall be used in this region. An outer flow region is also
needed for the outlet conditions (mainly atmospheric pressure) to be reached. Three different fluid
bodies have been created so as to be able to give different refinements to the different sections of the
flow. Therefore, different bodies for the flow inside the nozzle, the flow directly behind the nozzle
and an outer layer around this last section. The fluid body which is out of the nozzle extends from
−1 m from the end of the nozzle to 20 m from that point in the X-direction. These three fluid bodies
can be seen in Figure 14.

29
3 DEVELOPMENT OF THE SOLUTION

Figure 14: Simulation flow bodies

After some initial tests with coarser (and therefore more unstable) meshes, a final mesh has been
created with 334.059 elements. As it can be seen in Figure 15, the nozzle and after-nozzle regions
are composed of tetrahedral elements, while the outer body is composed of hexahedral elements.
Moreover, element body sizing of 0.04, 0.1 and 0.14 m have been imposed in the nozzle, after-nozzle
and outer fluid bodies respectively, with a finer mesh used for more critical parts of the flow. An edge
sizing has also been used in the nozzle and after-nozzle interface so as to provide better skewness.
Finally, an inflation method has been used in the interior of the nozzle wall in order to achieve a better
resolution of the boundary layer.

Figure 15: Fluid mesh

30
3 DEVELOPMENT OF THE SOLUTION

As Table 11 shows, the maximum skewness is under 0.9 and the minimum orthogonality is over
0.1, with which it can be concluded that the mesh is inherently optimal for its use in subsequent
simulations.

Maximum Minimum Average

Skewness 8.17e-7 0.858 0.141

Orthogonal Quality 0.142 1 0.868
Aspect Ratio 1.036 11.876 1.85

Table 11: Mesh quality metrics

Simulation setup

The flow has been simulated in a density-based solver, as it calculates a compressible flow. These
simulations have been performed as a steady state case, given the fact that a transient case would
require out-of-scope computation time and power.

Firstly, several laminar simulations were performed so as to evaluate different parameter combinations
that led to convergence. After that, the turbulent SST K-Omega model was chosen, so that a k-omega
model is used near the walls and a k-epsilon model is used in the free stream. Firstly, a first order
solution was calculated to convergence of residuals to 10−6 , and this simulation was then used as
initial parameters for the second order simulation with the same order of convergence.

For the boundary conditions, the following have been selected: an inlet pressure of 300 bar and a
temperature of 3000 K, an outlet pressure of 1 bar and a temperature of 300 K, no-slip condition for
the walls and a symmetry condition for the faces of the section of the domain.

Simulation results

After the simulation has converged, the obtained results for velocity magnitude, pressure, temperature
and mach number of the fluid body can be seen in Figure 16.

31
3 DEVELOPMENT OF THE SOLUTION

Figure 16: Solution of flow parameters after simulation

32
3 DEVELOPMENT OF THE SOLUTION

In these contour plots, the characteristic mach diamonds can be clearly seen, which form due to the
over-expanded nature of the nozzle and the boundary conditions. As it can be gathered from the
pressure contours, the exhaust pressure is lower than atmospheric pressure (105 bar), which results in
a contraction of the flow. This produces a cascading effect downstream, which is dissipated along the
X-axis. If these results are compared to the 2D case, it can be seen how the dissipation of the velocity
and other flow variables is much larger in the 3D case. This fact could be caused by several reasons:
the fact that a turbulent model and a Sutherland viscosity have been used in this case, among others.

Taking a closer look at the nozzle flow region, the obtained results for velocity magnitude, pressure,
mach number and temperature of the fluid body can be seen in Figure 17.

Figure 17: Solution of flow parameters in nozzle region

Taking a closer look at the flow in the nozzle region, Figure 17 shows how the flow accelerates
substantially in the region near the throat, while the aft region of the diverging section is designed
so as to straighten the flow. Also, one can see how Mach = 1 is achieved at the throat of the nozzle,
which is optimal. It can also be seen how the flow accelerates all throughout the length of the nozzle,
as the pressure decreases. If the region near the nozzle exit wall is inspected (Figure 18), the boundary
layer can be seen, as well as the flow separation that occurs in the nozzle exit due to its over-expanded
nature.

33
3 DEVELOPMENT OF THE SOLUTION

Figure 18: Detail of the nozzle exit wall

Taking a look at the heat transfer characteristics of the flow (Figure 19), one can see how the
turbulence in the flow produces patches where heat transfer is higher. However, these higher heat
transfer regions are more abundant at the start of the converging part of the nozzle and the throat of
the nozzle. This fact is consistent with the conclusions of the study by Naraghi et al. [24], which
shows that the conditions for the greatest heat flux are met at the throat of the nozzle.

Figure 19: Heat flux through the nozzle wall

In the following plots, the values of velocity, pressure, temperature and Mach number can be seen
along the center-line of the nozzle.

34
3 DEVELOPMENT OF THE SOLUTION

Figure 20: Fluid parameters along center-line of the nozzle

The data in these plots will be used in the posterior structural analysis of the nozzle, and serve to
validate the simulations when compared with the nozzle theory explained in the Bibliography (See
[2], [25]).

Finally, using the average value of the nozzle outlet speed and the mass flow rate obtained from
ANSYS CFD-Post with the thrust equation from Section 2.2, one obtains:

T = ṁ · v2 = 356.98 · 2 · 2386.85 = 1.704MN (14)

Which is a rather similar to the official SpaceX reports, which rate the Raptor engine at 1.9 MN of
thrust. However, the obtained result does not totally correspond with this data, which can be caused
by several factors that will be discussed in the simulation verification (Section 3.2.2).

Simulation verification

In order to verify these simulations, several checks have been made so as to ensure the simulations
correctly solved the model.

As it can be seen in Figure 21a, the simulation has been performed until stabilization of the residuals,
so that no further computation time would provide a more exact solution without changing the mesh.

35
3 DEVELOPMENT OF THE SOLUTION

The outlet mass flow has also been monitored 21b and can be used as a convergence criterion.

(a) Scaled residuals throughout the simulation

(b) Mass flow rate throughout the simulation

Next, the continuity of the solution has been verified through the calculation of the mass imbalance.
As it can be seen in Table 12, mass imbalance is in the order of 10−3 , which only represents 0.0733%
of inlet mass flow rate and is therefore negligible.

Inlet 356.7202
Outlet -356.7176
Balance 0.002615073

Table 12: Mass flow rate balance

On another note, it is crucial to perform a mesh refinement study in order to provide a more precise
solution of the model. To do that, the mesh has been refined two times by increasing the number
of cells in the nozzle and after-nozzle regions by 20%. The thrust at nozzle exit has been plotted in
Figure 22.

36
3 DEVELOPMENT OF THE SOLUTION

Figure 22: Thrust output through mesh refinement

While an asymptotic behaviour was expected, it is not the case with these simulations. This conduct
can be explained when comparing the flow at the exit of the nozzle in the coarse and fine mesh
simulations. In Figure 23a, one can see how the refinement of the mesh produces a more accurate
depiction of the flow separation in the nozzle exit, which becomes greater as more elements are
introduced in the region. This explains the loss in mass flow rate and exhaust velocity that causes the
reduction in thrust.

(a) Flow separation in 1st refinement (b) Flow separation in 3rd refinement

37
3 DEVELOPMENT OF THE SOLUTION

Another check that could be performed (but is considered out of scope) would be to check if the
designed domain is sufficiently large. As boundary conditions are being imposed that may not be
achieved in reality, this may be a source of error. In order to asses this fact, the domain of flow out of
the nozzle would have to be extended and the results would have to be compared with the ones that
have been extracted from this first model.

3.3 Nozzle structural FEM simulations

Once the various CFD simulations have been performed, the extracted data will be used together
with the geometry of the nozzle and the properties of the chosen build material so as to provide an
insight into the structural and thermal behaviour of the element. These Finite Element Method (FEM)
simulations will be performed through a static structural simulation through ANSYS Mechanical.

3.3.1 Structural analysis meshing

Given the 32.000 element upper limit for the ANSYS Mechanical Student Version, the geometry that
has been used for these simulations is a 1º slice of the nozzle. This geometry includes one regenerative
cooling channel, and while this would result in a geometry with far less channels than the proposed
solution, it will give great insight into the structural behaviour of the nozzle regardless.

Figure 24 shows the resulting mesh, made up entirely of tetrahedral elements with a size of 0.01m,
which results in a total of 15.941 elements. This figure shows 10 repetitions of the used geometry so
as to provide a better visual understanding of the structure of the mesh.

Figure 24: Nozzle FEM mesh

38
3 DEVELOPMENT OF THE SOLUTION

3.3.2 Structural analysis setup

As this is a static structural test, the main conditions that will be studied will be the effect of thermal
expansion and the effect of the pressure inside the nozzle created by the combustion of the propellants.
For these simulations, C-18150 has been chosen as the construction material for the nozzle, as
explained in Section 2.7.

Regarding the chosen boundary conditions, the following have been used:

• A fixed support in the combustion chamber end of the nozzle

• A thermal condition of 700K on the whole body

• The imported pressure from the CFD simulations along the length of the nozzle, which can be
seen in Figure 25

Figure 25: FEM pressure condition

The imported pressure has been used in the initial step while the thermal expansion has been
introduced in an intermediate step so that the behaviour of the nozzle for the two conditions can
be studied.

3.3.3 Structural analysis results

After the simulation has converged, the following total deformation and Von Mises equivalent stress
have been retrieved, and can be seen in Figure 26

39
3 DEVELOPMENT OF THE SOLUTION

Figure 26: FEM structural analysis results before and after thermal strain

40
3 DEVELOPMENT OF THE SOLUTION

At first glance, the simulations show how the thermal expansion causes deformations nearly an order
of magnitude higher than the deformation caused by the pressure of the hot-gas side. In the initial
step, the largest deformation can be seen where the pressure is higher, as one would expect. However,
when the thermal expansion is introduced, the highest deformation can be seen at the exit of the
nozzle, as this expansion builds up throughout the length of the nozzle. This is caused by the fact that
while the hot-gas pressure only acts perpendicular to the surface of the nozzle, the thermal expansion
is multi-directional.

On the other hand, if one was to take a look at the equivalent stresses, their distribution doesn’t
endure major changes when the thermal expansion is applied. The region of higher stresses is the
combustion chamber and the start of the converging section of the nozzle. Here, the pressure and
thermal expansion induced stresses are of the same order of magnitude.

If the undeformed model is compared to the true-scale model, the geometric differences can be seen.
While these are visible (Figure 27), they do not involve a great difference from the undeformed model
and won’t produce major differences in the development of the previously studied flow. However, the
extent of these differences can only be inferred. As Figure 27 shows, the nozzle will also increase
in length, which will in turn give the fluid more room to expand and therefore accelerate. This will
result in a slight increase of the nozzle thrust.

41
3 DEVELOPMENT OF THE SOLUTION

Figure 27: Detail of the deformed nozzle and the undeformed nozzle wireframe

The interior of the cooling channels will also be inspected in order to assess if the variation in the
channel geometry is of importance. As Figure 28 shows, no major variations in the channel geometry
can be seen.

Figure 28: Detail of the true scale of the deformation on the cooling channel

42
3 DEVELOPMENT OF THE SOLUTION

3.4 Nozzle thermal FEM analysis

In order to assess the thermal capabilities of the nozzle and its refrigeration system, a set of steady
thermal FEM simulations will be performed. The mesh will be identical to the one that has been used
in the previous Section.

It should be noted that a complete thermal analysis would need a coupled CFD and FEM study in
order to iteratively solve the wall temperature of the nozzle and the heat flux produced by the flow
through the nozzle. This has been studied in an article by Jiawen Song et al. [13], so its results will
be used to provide indicative results.

3.4.1 Thermal analysis setup

First of all, a heat flux distribution has been modeled along the longitude of the nozzle through the
results of the aforementioned study. The study shows that the particular flow parameters at the throat
of the nozzle (e.g., velocity, temperature, pressure) lead to this region being where the highest heat
flux occurs, so the model shows a peak in heat flux in that region and then stabilizes.

On top of that, a convection boundary condition has been established in the surfaces of the cooling
channels, iterating the film coefficient until the results showed acceptable wall temperatures. Higher
film coefficients have been chosen for the inner channel wall (i.e., the one closest to the hot-gas), as
the cooling fluid tends to be faster in those regions and therefore more effective in heat transfer, as
explained in Reference [12]. The parameters for this convection heat flux are the following:

• Coolant temperature: 100 K

• Film coefficient for channels with AR = 6:

1. For the inner wall: 80000 W /m2 · K

2. For the outer walls: 50000 W /m2 · K

• Film coefficient for channels with AR = 4:

1. For the inner wall: 50000 W /m2 · K

2. For the outer walls: 30000 W /m2 · K

3.4.2 Thermal analysis results

Figure 29 shows the results of the thermal simulations. As it can be seen in the Figure, the throat,
which is the region with the highest hot-gas heat flux, is kept below 30% of the melting temperature
of C-18150, which would be structurally acceptable.

43
3 DEVELOPMENT OF THE SOLUTION

Figure 29: Thermal analysis results

An unacceptably high temperature is reached in the combustion chamber, but this fact had been
previously accounted for, as this chamber was made longer than necessary in the first place so as to
ease the flow development in the posterior CFD simulations. However, one can see that the inlets
of the cooling channels in the intake are too far apart, resulting in an excessive temperature which
could cause a catastrophic failure. In order to solve this issue, the inlets of the downward and upward
cooling channels would have to be brought closer together or even merged into one, so that the coolant
could reach that particular region of the nozzle.

Regarding the other regions of the nozzle, the temperatures don’t exceed the operational limits of the
build material, so that the calculated level of coolant heat flux could be able, in theory, to provide
enough heat flux to the structure when operating under normal conditions.

Using the film coefficients that have been estimated, one would be able to calculate the flow velocity
needed. With this value, the difference of pressure in the cooling channels could be found, which
would establish a lower limit for the pump system of the propellant.

44
4 ANALYSIS OF THE RESULTS

4 Analysis of the results

4.1 Economic Feasibility

4.1.1 Market study

Despite having received an increase in funding in recent years [26], Spain’s space industry’s
development is quite far from other European neighboring countries. However, efforts are being
made and research is being outputted so as to bring Spain’s space industry closer to the forefront.

Several companies have benefited from ESA contracts, such as PLD Space, which is developing two
partially-reusable launch vehicles, the Miura 1 and Miura 5. These launchers will mainly be used to
launch CubeSats or microsatellites into orbit, and are expected to perform their first flight in 2024.
These launchers will use TEPREL engines, which use RP-1 as a propellant to produce 30 kN of
thrust at sea level [27]. These engine’s combustion chamber is regeneratively cooled and follows the
channel wall design, manufactured through a milling process.

Another company which uses similar manufacturing techniques and materials to the ones that the
studied nozzle would use is Pangea Aerospace. However, their propulsion systems use the aerospike
philosophy instead of the typical bell-shaped used in the design of this project. These engines use
methane and liquid oxygen as propellants, and use regeneratively cooled and additively manufactured
thrust chambers. Nevertheless, these engines are designed to produce considerably lower levels of
thrust than the designed nozzle and are not representative enough.

To get a clear grasp of the direct competition in similar performance rocket engines, one should look
beyond Spain’s borders. Taking into account the options that steer away from the use of RP-1, we
have the SpaceX Raptor, Blue Origin’s BE-4, NASA’s RS-25 and ESA’s Vulcain. As Table 13 shows,
these engines move in the same range of thrust, yet they show great disparity in their prices.

Raptor BE-4 RS-25 Vulcain

Total thrust (MN) 2 ∼2.4 1.86 1.14

Price ∼$2M ∼$8M >$50M ∼$8.2M
Cost per kN $1000 $3333 $26881 $7192
Potential cost per flight $20 $133 $1414 -

Table 13: State of the art rocket engine production cost comparison [3]

The two engines on the right represent the rocket engine of choice from the mid 2000s onward which
were characterized by using hydrogen as a propellant. The two engines on the right, however, have
been developed with a completely different philosophy, as they will be used in privatized space flight.
This means that they have been developed with profit in mind in comparison with the two former

45
4 ANALYSIS OF THE RESULTS

engines, which were funded by governmental institutions. This fact leads to much cheaper engines
which are built to be re-utilised, drastically driving down the cost per flight.

4.1.2 Project cost

Once the market has been studied, the cost of the whole project will be estimated. This will encompass
the cost of every step of the project up until the manufacturing of the full-scale nozzle, which will be
dealt with in the following section. Table 14 shows the budget for the design and prototyping of the
project.

Group Object Product cost (C) Hours (h) Cost (C)

State of the art LPRE research 15 225

Rocket engine theory
15 225
funadamentals research
Rocket engine cooling
15 225
systems research
Manufacturing techniques
20 300
research
Subtotal 65 975
Nozzle design Initial calculations 10 150
Geometry determination 20 300
Nozzle design 25 375
Cooling system design 40 600
Subtotal 95 1425
CFD Simulations Introduction course 15 225
CFD simulations 2D 45 675
CFD meshing 3D 15 225
CFD simulation setup 3D 40 600
CFD simulation calculations 3D 80 1200
CFD result analysis 3D 40 600
Subtotal 235 3525
FEM Simulations FEM meshing 10 150
FEM simulation setup 15 225
FEM simulation calculations 20 300
FEM result analysis 10 150
Subtotal 55 825
Gadgets & software Personal computer 160.00
SolidWorks license 88.89
ANSYS license 6666.67
Microsoft Office license 2.50
Subtotal 6918.06
Project drafting Project charter 15 225

46
4 ANALYSIS OF THE RESULTS

Group Object Product cost (C) Hours (h) Cost (C)

Project report 100 1500

Oral presentation 35 525
Subtotal 150 2250

TOTAL 6918.06 600 15918.0556

Table 14: Project budget

In this table, a cost of 15C/hour has been used for the project engineer, which is an average price for
the design, simulations, drawing production and drafting.

Another aspect to take into consideration is the cost of the software licenses and the amortization of
the personal computer which has been used for the development of the project. In order to provide
a precise cost for the software licenses, an annual workday count of 1800 hours has been chosen. In
the following tables, these costs are calculated:

SolidWorks work time (h) 40.00

Cost of annual license per working hour (C/h) 2.22
Cost of project license (C) 88.89

ANSYS work time (h) 400.00

Cost of annual license per working hour (C/h) 16.67
Cost of project license (C) 6666.67

Microsoft Office work time (h) 30.00

Cost of annual license per working hour (C/h) 0.08
Cost of project license (C) 2.50

PC initial cost (C) 1800.00

Amortization years 5.00
Amortization per hour (C/h) 0.20
Project PC work time (h) 800.00
PC amortization during project (C) 160.00

Table 15: Software and PC project cost

47
4 ANALYSIS OF THE RESULTS

4.1.3 Fabrication costs through metal additive manufacturing

In an era in which private space companies are becoming increasingly prominent, the cost-
effectiveness of new thrust chambers is an essential part in the economic feasibility of their projects.
Traditional manufacturing techniques (e.g., forging, machining, electroplating, etc.) often prove to be
labor-intensive, costly and result in a high part count, which directly affect profitability. On the other
hand, additive manufacturing, provides considerable improvements in decreased manufacturing time,
automation of manual operations, the use of a wide range of materials and the use of integral chamber
designs. The decrease in manufacturing time and associated cost that these AM technologies provide
can be seen in Figure 30. As the figure shows, a decrease of 83% in production time from worst to
best scenario could be achieved in the near future. Reports from NASA point to a large scale integral
channel HR-1 alloy nozzle measuring 152 cm in diameter and 178 cm in height being manufactured
using LP-DED technologies in 90 days deposition time [7].

Figure 30: Comparison of traditional and additive manufacturing evolution [28]

Given the fact that production time could be reduced by up to ∼ 80%, if the power consumption of
the DED machines is estimated equal to that of actual manufacturing techniques, that could lead up
to the same reduction in production cost. On top of that, less personnel would be required, as these
technologies are presumed to be fully automatic when full development is achieved. Being able to
manufacture a monolithic structure in site would also mean that transport time and cost would be
reduced. While the exact reduction in production cost that these novel metal AM technologies will
have in the long run cannot be exactly determined, it is clear that the reduction in production time and
required personnel will make for a more economically optimal alternative [29].

In order to analyze the production costs of the nozzle, a scaled 1:10 model has been designed so as
to be manufactured by a private company. This model has been designed taking into account the
limitations of the technology which the manufacturing company of choice (i.materialise) uses. This
company uses Laser Powder Bed Fusion (L-PBF) technology with the specifications shown in Table
16.

48
4 ANALYSIS OF THE RESULTS

Minimum wall thickness (mm) 1 to 3

Minimum details (mm) 0.8 to 1
Accuracy +/- 1%
Maximum size (mm3 ) 762 x 393 x 393

Table 16: L-PBF design specifications (Steel)

The budgeted cost for this model made out of stainless steel is 305 C, which would be out of reach for
this project. As no broken down budget is provided by the manufacturer, it has not been possible to
find out what percentage of the cost is due to the cost of the material or the cost of production.

4.1.4 Profitability analysis

In order for the commercialization of the designed nozzle to be fruitful, several key points have to be
taken into consideration:

• In order to be able to mass produce the designed nozzle, a DED 3D printer would have to be
acquired, which can cost up to $2M [30]. This is a considerable initial investment and could be
non-profitable in the long term.

• The re-usability of the produced nozzles could be crucial in the economical success of this
endeavour. If these nozzles could be used multiple times in several rocket launches, the break
even point would go drastically down. While initial structural and thermal tests have been
developed, these are not sufficient in order to label them as re-usable. To do that, vibration and
fatigue simulations would have to be developed as well as physical tests on scaled nozzles.

• It should be reinstated that most companies in the space sector operate partly through
governmental contracts (in Europe, mainly through ESA contracts). This way, being able to
benefit from public resources could be an efficient way to solve the initial investment problem
or further economical issues that may arise.

4.2 Environmental considerations

Being one of the most pressing problems of today’s world, it is essential to take the environmental
impact of space travel into consideration. As it is a complex and nuanced matter, these environmental
considerations will focus on the impact of the technologies of the designed nozzle and its engine
to the atmosphere, the potential influence that emerging AM production techniques can have on the
issue and the space debris affair.

49
4 ANALYSIS OF THE RESULTS

4.2.1 Atmospheric contamination

The designed nozzle will use methane and liquid oxygen as propellants, following the trend created by
SpaceX’s Starship and Blue Origin’s New Glenn’s first stage. This propellant combination produces
CO2 and H2 O as combustion products, as well as different amounts of NOx depending on altitude as
an after-burning effect of the flame coming in contact with the air [31]. These gases are known as
greenhouse gases, as they absorb more heat than the standard composition of Earth’s atmosphere, and
have been demonstrated to be one of the main causes of global warming.

However, when compared with other alternatives, this propellant combination is the most
environmentally friendly after hydrogen and liquid oxygen. Other propellants, when used in rocket
propulsion, produce other types of substances that are far more detrimental to the atmosphere: solid
rocket propellants produce aluminum oxide, soot, hydrogen chloride, nitrogen oxides among others;
RP-1, which is a highly refined form of kerosene, produces nitrogen oxides, carbon soot, carbon
monoxide and sulfur compounds; and hypergolic fuels produce soot, sulfur containing compounds
and nitrogen oxides [4]. Out of these gases, the US’s Environmental Protection Agency considers
nitrogen oxides, sulfur oxides and carbon monoxides as heavy pollutants [32]. Moreover, chlorine,
alumina and nitrogen oxides can destroy ozone and are considered as ozone-depleting substances,
which have been heavily monitored and restricted since 1996. On top of that, chlorine, sulfuric
compounds and nitrogen oxides can cause acid rain which is acutely harmful to marine life, vegetation
and life in general. To sum it up, it has been shown that rockets that use solid rocket boosters or RP-1
generate about 30 times more atmospheric heating than a hydrolox rocket [31].

It should also be noted that these emissions, due to the nature of rocket flight, are produced in all layers
of the atmosphere. While water vapor and CO2 are cycled swiftly in lower layers of the atmosphere,
these and other substances remain longer in the upper layers, meaning the radiative forcing they
produce is more severe [31]. This fact should be object of study, as SpaceX’s Starship and Super
Heavy are expected to produce around 2683 and 2199 metric tonnes of CO2 and H2 O respectively
[16]. Nevertheless, if one compares the CO2 and H2 O emissions to metric tonne of payload for
different launchers, it can be observed that many rockets have much worse ratios than that of the
methane-propelled rockets. For example, the Space Shuttle produces 16 and 24 metric tonnes of CO2
and H2 O respectively for every tonne of payload, while the same ratios for SpaceX’s Starship are 27
and 22. On top of that, the Space Shuttle produces 21.7 metric tonnes of ozone-depleting substances
for every metric tonne of payload, while Starship produces near-zero emissions of those substances
[33].

All in all, one can conclude that the new generation of methane and hydrogen propelled rockets are
considerably less detrimental to the atmosphere than former alternatives. It should be taken into
consideration that the global CO2 output of rockets represented solely the 5.9 10−6 % of all CO2
emissions on Earth [4]. While there are far more polluting industries, this fact should not fend society

50
4 ANALYSIS OF THE RESULTS

off from trying to prevent possible future environmental problems due to rocket travel.

4.2.2 Effect of additive manufacturing technologies on environment

As seen in Section 4.1.3, the utilization of AM technologies provides a faster and more efficient
solution to manufacturing than previous techniques. The reduction in production time has been
estimated to be of a maximum of 80%. However, the gained production time might be invested
in other activities that may be harmful to the environment, so that the effect that this reduction in
production time cannot be directly related to a reduction in power usage.

If the pollution resulting from the extraction and processing of the used materials is analogous in
traditional and AM manufacturing techniques, and assuming both techniques consume the same
amount of power and are used for the same amount of workload, this would lead to a reduction
of 80% in the manufacturing CO2 emissions and radioactive waste produced. It should be noted that
scarce information is yet available on the real power usage of the DED processes that would be needed
to additively manufacture a nozzle of such size. However, some studies support the fact that these
technologies are more energy efficient than previous manufacturing methods [34].

4.2.3 Space debris

The space debris issue has become one of the most pressing matters of modern space travel. The
unregulated disposal of non-functional man-made objects in Earth’s orbits during the last decades,
combined with the rise of satellite constellation businesses could result in catastrophic scenarios such
as an orbital cage, which could indefinitely put a stop to human space travel [35].

This issue is mainly being caused by non-functional satellites, intermediate rocket stages that because
of their high speeds are trapped orbiting earth and the collision between existing space debris. As
these objects are orbiting Earth at tens of thousands of kilometres per hour, collisions between these
objects produce great quantities of high-velocity particles. These process could result in a chain
reaction that could be detrimental to satellite technology and space travel [35].

At the moment, these space debris objects are being monitored by space agencies like the ESA, that
currently knows the existence of approximately 1950 discarded rocket stages as well as millions of
debris fragments ranging from defunct satellites to fragments of 1 cm in size [36]. This high number
of intermediate rocket stages orbiting Earth could come in contact with operational satellites or be part
of a collision with a smaller object, therefore worsening the matter. It is essential for space agencies
then to eliminate the existing space debris and provide clear instructions for their disposal.

Notwithstanding the surge of the satellite constellation businesses (e.g., Starlink), which will in turn
significantly increase the number of rocket launches, the reusable approach that rocket technology
has taken will mean that the first stages of these launchers will be able to return to earth so that they
can be used again after certain refurbishment. This fact will surely stabilize the number of rocket

51
4 ANALYSIS OF THE RESULTS

stages on orbit, and if the existing ones are properly deorbited, that could mean the problem could be
controlled.

There are several ways in which space agencies are dealing with these issues. In summary, NASA
has developed several action plans in order to mitigate the issue and protect the existing operational
satellites [37]. On top of that, efforts are being made to develop active space debris removal
technologies, which would allow humanity to dispose of debris that is now indefinitely orbiting the
Earth [38].

4.2.4 Environmental impact of the project

Finally, the effect of the whole project on the environment will be studied using the documentation
provided by the ’Oficina Catalana del Canvi Climàtic’ [39].

First of all, the total CO2 emissions will be calculated. The power consumption of the personal
computer used for the project is estimated to be around 200W, and the project has been completed
in 600 work-hours, which add up to a consumption of 120kWh. Using the national average of
0.31kgCO2,eq /kW h, the amount of emitted CO2 is found to be 37.2 kg. This number is comparable
with the amount of CO2 emissions of the average car in Spain over a period of 10 days.

Moreover, the radioactive waste created through the development of the project will be calculated. As
the national average of radioactive waste was 0.54mg/kW h in 2020, the previously calculated energy
consumption can be used to compute the produced waste. This way, the produced radioactive waste
adds up to 64.8mg.

4.3 Planning and scheduling

Once the technical aspect of the project has been achieved, an organizational outlook will be
performed. In order to do that, the project charter key points will be analysed and compared with
the real schedule.

4.3.1 Task identification and description

The complete list of project tasks and their description can be found in the following pages.

1. Study of state of the art

1.1. Study of bibliography: Study recent articles as well as relevant books on the matter
(e.g., ‘Rocket propulsion elements’, George P. Sutton) for a better understanding of the
functioning of nozzles and the modern approach to their design.

1.2. Determination of necessary equations: Extract the necessary equations for the first
approximations from the bibliography.

52
4 ANALYSIS OF THE RESULTS

1.3. Choose a rocket nozzle style: Choose a relevant nozzle design for its study.

1.4. Research engine specifications: Research the chosen engine specifications (e.g., sizing,
expected chamber pressure and temperature, thrust).

1.5. First design approximations: Use the equations and parameters that have been found in
order to calculate a first approximation of the design parameters (e.g., gas exhaust speed,
pressure and temperature).

2. Nozzle design

2.1. Design of optimal geometry for the converging-diverging section: Calculate an efficient
geometry for the nozzle using the most modern design techniques and the first
calculations.

2.2. Importing geometry to SolidWorks: Using geometry calculations to create a sketch on

which to initialize the 3D design.

2.3. Design of nozzle outline: Use SolidWorks to design the entire nozzle and its auxiliar parts
(e.g., reinforcements if need be).

2.4. Design of the cooling system: Calculate the cooling parameters that will ensure a safe
operation for the nozzle and implement it into the design.

2.5. Design refinement: Refine the design using the information gathered from the
simulations.

3. CFD and FEA simulations

3.1. Structural analysis: Use ANSYS to create a suitable mesh and perform structural
simulations. Analyse the results.

3.2. Thermal analysis: Use ANSYS and the previously designed mesh to perform thermal
simulations. Analyse the results.

3.3. CFD simulations: Use ANSYS to create a fluid mesh and perform compressible CFD
simulations so as to have an understanding of its functioning.

3.4. Comparison with projected results: Use the results obtained in each simulation and
compare them to the first approximations.

4. Study of manufacturing

4.1. Study of the actual manufacturing techniques for rocket nozzles: Study the state of the art
of modern rocket nozzle production and analyse its pros and cons.

4.2. Study of the additive manufacturing techniques: Study the state of the art of actual metal
additive manufacturing techniques and select the most suitable one for the purpose at

53
4 ANALYSIS OF THE RESULTS

hand.

4.3. Study and selection of appropriate materials: Knowing what materials can be used
through the selected method, select an appropriate material for the nozzle.

4.4. Production of a scaled nozzle unit: Select a local manufacturer with which a scaled unit
can be produced and send the data for the designed model.

4.5. Verification of measurements and tolerances: Once the nozzle has been manufactured,
examine and study the measurements and tolerances.

5. Economic feasibility study

5.1. Study of the cost of design and production of a modern nozzle: Analyse the cost of the
methods that have been used before the appearance of additive manufacturing.

5.2. Study of the cost of design and production of a nozzle through additive manufacturing:
Analyse the cost of production through metal additive manufacturing techniques.

5.3. Comparison of results: Compare the results of both additive and non-additive
manufacturing and study the economic feasibility of the former.

6. Environmental considerations

6.1. Study of the environmental impact of selected propellant: Analyse the environmental
aspects of the production of the selected propellant as well as its resulting combustion
gases.

6.2. Study of the environmental impact of actual manufacturing technologies: Analyse the
environmental impact of additive manufacturing techniques.

6.3. Study of the possible environmental improvements with additive manufacturing

technologies: Analyse the environmental impact of former techniques and explain the
implications of the use of the new technologies.

7. Drafting of required documents

7.1. Drafting of thesis report: Draft the important aspects of the project with LaTeX software.

7.2. Production of manufacturing drawings: Design manufacturing drawings of the designed

nozzle.

54
4 ANALYSIS OF THE RESULTS

4.3.2 Task length and dependencies

In the following table, each task has been related to an approximated workload and the required
preceding task to be completed for each one of them to be initiated.

Code of task Task workload (days) Preceding task(s)

1.1 14 -
1.2 1 1.1
1.3 2 1.2
1.4 3 1.3
1.5 3 1.4
2.1 5 1.5
2.2 1 2.1
2.3 10 2.2
2.4 10 2.3
2.5 10 2.4 + 3.4
3.1 10 2.4 + 4.3
3.2 10 3.1
3.3 10 3.2
3.4 5 3.3
4.1 10 1.5
4.2 10 4.1
4.3 5 4.2
4.4 10 2.5 + 4.3
4.5 2 4.4
5.1 10 2.4 + 4.3
5.2 5 5.1
5.3 5 5.2
6.1 5 2.4 + 4.3 + 5.3
6.2 5 6.1
6.3 5 6.2
7.1 93 -
7.2 5 2.5

Table 17: Task dependencies and lengths

55
4 ANALYSIS OF THE RESULTS

4.3.3 Gantt diagram

Figure 31: Gantt diagram of the project

56
4 ANALYSIS OF THE RESULTS

4.3.4 Discussion on planning and schedule

Given that the previous Gannt Diagram was created in earlier stages of the project and its purpose
was mainly indicative, there have been some changes to the tasks and their duration.

On the one hand, given the actual Covid-19 pandemic situation, it has been unfeasible to produce a
metallic scaled unit of the designed nozzle. In the first place, the agreement between Hewlett Packard
and the Polytechnic University of Catalonia was going to be summoned for this purpose, but pure
online interaction with the manufacturers hindered communications and therefore the realisation of
the task. Other alternatives were taken into consideration, one of them being for a private company to
print the nozzle (in this case, i.materialise), but the cost was completely out of reach for the project’s
actual budget.

On the other hand, there have been minor changes in the development of the several tasks and their
length. Firstly, the CFD simulations were performed before the structural and thermal tests, as the
flow parameters were used in these calculations so as to provide more realistic results. Secondly, the
work-time for the CFD simulations was far greater than what it was first estimated. Getting used
to the ANSYS environment and performing considerably difficult-to-converge simulations (due to
compressible flow, turbulent models, etc) meant that more than a month and a half was spent until
reasonable results were attained. Finally, all other tasks have been performed in the expected time
frame with no considerable deviations.

In summary, while there were certain variations to the planning of the project’s tasks, the project
schedule has been successfully completed in the required time frame.

4.4 Conclusions

After the completion of the proposed solution and the analysis of the subsequent results, the
fundamental conclusions of the project will be drawn:

• The use of thermodynamic analytical expressions together with the method of characteristics
and the ideal nozzle trucation method implemented by Hoffman [8] have been found to provide
effective means to perform the calculations of the geometry of a LPRE nozzle. These methods
have been used together with the performance parameters of SpaceX’ latest Raptor engine
series in order to design a LPRE nozzle that uses the regenerative cooling approach. Extensive
2D and 3D simulations have been performed using the combustion chamber parameters of these
engines so as to solve the required flow parameters. The results show that the nozzle is able to
produce effectively the same levels of thrust as the real engines, therefore validating the design.

57
4 ANALYSIS OF THE RESULTS

• The improvement in tolerances that the modern metal AM technologies provide will lead to
more efficient regenerative cooling systems. On the one hand, this fact has allowed for the
design of higher aspect ratio rectangular channels, which will result in a higher heat transfer
to the coolant from the hot nozzle wall. Additionally, more cooling channels can be designed
into the nozzle, which will also increase the cooling system efficiency. On the other hand,
the automatized nature of DED technologies have allowed the channels to be designed using
complex geometries. In this case, a variable aspect ratio design has been chosen, so that the
most critical regions of the nozzle (i.e., the throat) have a higher aspect ratio than other regions,
allowing for the coolant to accelerate, therefore increasing its heat transfer capabilities while
limiting the required pressure loss along the channels.

• The choice of a high strength and high thermal conductivity material such as the C-18150 has
been proven to be a suitable choice. First of all, it has already been used in DED additively
manufactured rocket nozzles with satisfactory results. On top of that, the flow parameters of the
aforementioned CFD simulations have been used in thermal and structural FEM simulations in
order to assess the suitability of the physical properties of this build material during normal
operation conditions. The results have been satisfactory, as the design has not shown signs of
structural or thermal failure along the various analysis.

• The implementation of DED technologies to rocket engine production will have a direct impact
in the cost of their manufacturing. It has been shown that the level of automatizing that these
techniques possess will bring production times down to 80% when compared to previous
manufacturing techniques. This fact will not only drive the cost of production down, but
will also provide high launch rate companies with the ability to mass produce these type of
components and therefore accelerate their progress. While the acquisition of a DED machine
would mean a considerable investment, it could fit into the budget of a company that aimed to
benefit from governmental contracts or private investors.

• Following the modern trend of selecting methane as the new state of the art propellant will lead
to a reduction of rocket engine greenhouse effect gases. On top of that, the reusable nature of
these engines will stagnate the amount of discarded rocket stages that populate Earth’s orbit.
It should also be noted that the aforementioned reduction in production times that come with
new metal AM technologies will contribute to the reduction in power usage and therefore in
greenhouse effect gas emissions.

4.5 Recommendations

To whom the design and simulations that have been developed may concern, the following
recommendations have been drafted:

• Regarding the calculation of the nozzle geometry, several References (e.g., [2], [25]) point to

58
4 ANALYSIS OF THE RESULTS

the diverging nozzle being the most critical in the development of the flow. Consequently,
the converging section can be calculated as a conical section. The method of characteristics
used to calculate the diverging section geometry can be time-consuming to implement, so it is
recommended to look online for a code that fits to the needs of the project. In this case, the
code developed by K. Palaveev et al [22] was used.

• With respect to the design of the cooling system of the nozzle, rectangular section channels
have been found to be the most effective in many occasions and is now standard practice. With
the dimensions of the designed nozzle (and therefore the Raptor engine nozzle), an upper limit
of 550 channels has been found with the channel geometry that has been used when taking into
consideration the tolerances of state of the art additive manufacturing techniques. As time goes
on these techniques will become more efficient and precise and this upper limit will be able to
be extended.

• When performing 2D flow simulations in OpenFOAM, the example library should be

researched deeply, as most of the simulation parameters are not considerably intuitive and some
examples can help to work out the correct setup for the simulation.

• In relation to the 3D simulations in ANSYS, one should build the simulation step by step, as
these are quite difficult to get to converge. These simulations should be tackled through trial and
error, as every attempt one is destined to be closer to the combination of geometry, mesh and
setup parameters that will lead to coherent results. As such, one should start with simple fluid
bodies, lower boundary condition parameters than the final ones and even non-compressible
simulations and build his/her way up until the required parameters can be used.

4.6 Proposed work

Given the workload limitations of this project, there are many aspects that could be studied in order
to take it one step further. These further tasks will be dealt with in the following list:

• Considering the limitations of the ANSYS Software Student version, the CFD and FEM
simulations that have been performed are not as exact as they could be. In the interest of
further verification, these simulations would have to be repeated using meshes in the order of
> 106 elements for the flow simulations and > 105 elements for the structural analysis.

• In order to fully evaluate the performance of the designed nozzle, a coupled analysis would have
to be performed. To do this, the flow inside the cooling channels would have to be simulated and
the result parameters would then have to be used in the structural and thermal FEM simulations
on the nozzle.

• The geometry of the nozzle is analytically optimal, as it has been calculated through the method
of characteristics. However, there are many parameters in this method and the applied method

59
4 ANALYSIS OF THE RESULTS

of truncation that could result in more efficient geometries for the nozzle. In order to find
the best possible geometry, the performed simulations could be repeated on a variable set of
geometries, so that the obtained results could be compared.

• As it has been explained in Section 4.3.1, the production and testing of a scaled model have
been circumstantially non-viable. While simulations can give considerably realistic results to
our problems and convergence parameters might point to the solution being legitimate, it is not
until the design is built and tested that one can be sure of its efficient functioning. Consequently,
a metal additive manufactured scaled prototype of the nozzle would have to be produced and
tested.

60
5 BIBLIOGRAPHY

5 Bibliography
[1] Ash, R., Dowler, W., & Varsi, G. (1978). Feasibility of rocket propellant production on mars.
Acta Astronautica, 5(9), 705–724. https://doi.org/https://doi.org/10.1016/0094-
5765(78)90049-8
[2] Sutton, G. P. (2017). Rocket propulsion elements. John Wiley & Sons.
[3] Dodd, T. (2019). Is spacex’s raptor engine the king of rocket engines? https : / /
everydayastronaut.com/raptor-engine/ (accessed: 10.05.2021)
[4] Dodd, T. (2020). What is the environmental impact rockets have on our air? https : / /
everydayastronaut.com/rocket-pollution/ (accessed: 12.05.2021)
[5] Ellis, D., Gray, H., & Nathel, M. (2001). Aerospace structural materials handbook supplement
grcop-84.
[6] Jones, H. W. (2018). The recent large reduction in space launch cost.
[7] Kerstens, F., Cervone, A., & Gradl, P. (2021). End to end process evaluation for additively
manufactured liquid rocket engine thrust chambers. Acta Astronautica, 182, 454–465. https:
//doi.org/https://doi.org/10.1016/j.actaastro.2021.02.034
[8] Hoffman, J. D. (1987). Design of compressed truncated perfect nozzles. Journal of Propulsion
and Power, 3(2), 150–156. https://doi.org/10.2514/3.22967
[9] Naraghi, M., Dunn, S., & Coats, D. (n.d.). A model for design and analysis of regeneratively
cooled rocket engines. 40th AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit.
https://doi.org/10.2514/6.2004-3852
[10] Wang, T.-S., & Luong, V. (1994). Hot-gas-side and coolant-side heat transfer in liquid rocket
engine combustors. Journal of Thermophysics and Heat Transfer, 8(3), 524–530. https://
doi.org/10.2514/3.574
[11] Pizzarelli, M., Nasuti, F., & Onofri, M. (2012). Cfd analysis of transcritical methane in rocket
engine cooling channels. The Journal of Supercritical Fluids, 62, 79–87. https://doi.org/
https://doi.org/10.1016/j.supflu.2011.10.014
[12] Ulas, A., & Boysan, E. (2013). Numerical analysis of regenerative cooling in liquid propellant
rocket engines [VFE-2]. Aerospace Science and Technology, 24(1), 187–197. https://doi.
org/https://doi.org/10.1016/j.ast.2011.11.006
[13] Song, J., & Sun, B. (2016). Coupled numerical simulation of combustion and regenerative
cooling in lox/methane rocket engines. Applied Thermal Engineering, 106, 762–773. https:
//doi.org/https://doi.org/10.1016/j.applthermaleng.2016.05.130
[14] Pizzarelli, M. (2017). Regenerative cooling of liquid rocket engine thrust chambers. https:
//doi.org/10.13140/RG.2.2.30668.92804

61
5 BIBLIOGRAPHY

[15] Patel, N., Standbridge, S., Berghe, M. V. D., & Devalaraju, V. (n.d.). Design and additive
manufacturing considerations for liquid rocket engine development. Aiaa propulsion and
energy 2019 forum. https://doi.org/10.2514/6.2019-4392
[16] SpaceX. (2019). Draft environmental assessment for the SpaceX Starship and Super Heavy
launch vehicle at Kennedy Space Center (KSC). https://netspublic.grc.nasa.gov/
main/20190801_Final_DRAFT_EA_SpaceX_Starship.pdf
[17] Martin, J. H., Yahata, B. D., Hundley, J. M., Mayer, J. A., Schaedler, T. A., & Pollock, T. M.
(2017). 3d printing of high-strength aluminium alloys. Nature, 549(7672), 365–369.
[18] Lewandowski, J. J., & Seifi, M. (2016). Metal additive manufacturing: A review of mechanical
properties. Annual Review of Materials Research, 46(1), 151–186. https://doi.org/10.
1146/annurev-matsci-070115-032024
[19] Copper Development Association Inc. (2020). C-18150 properties. https : / / alloys .
copper.org/alloy/C18150 (accessed: 01.06.2021)
[20] Kestin, J. (1948). Thermodynamic properties of combustion gases. Aircraft Engineering and
Aerospace Technology, 20(8), 234–242. https://doi.org/10.1108/eb031660
[21] Engineering Toolbox. (2021). International standard atmosphere. https : / / www .
engineeringtoolbox . com / international - standard - atmosphere - d _ 985 . html
(accessed: 28/04/2021)
[22] Kyril Palaveev, T. S. (2020). A detailed presentation and implementation procedure of
axisymmetric method of characteristics for rocket nozzle design. https : / / github . com /
KyrilPalaveev/Axisymmetric- Method- of- Characteristics- Nozzle- Designer-
with-GUI. (accessed 19.02.2021)
[23] Aoki, T. (2020). Helper utilities for openfoam blockmeshdict generation. https://github.
com/takaakiaoki/ofblockmeshdicthelper (accessed: 29.05.2021)
[24] Naraghi, M., & Foulon, M. (2008). A simple approach for thermal analysis of regenerative
cooling of rocket engines. ASME International Mechanical Engineering Congress and
Exposition, Proceedings, 10. https://doi.org/10.1115/IMECE2008-67988
[25] Hill, P. G., & Peterson, C. R. (1992). Mechanics and thermodynamics of propulsion. Addison-
Wesley.
[26] El Paı́s. (2019). España aumenta 700 millones su contribución a la exploración del espacio.
https : / / elpais . com / elpais / 2019 / 04 / 26 / ciencia / 1556280067 _ 704804 . html
(accessed: 27.05.2021)
[27] PLD Space. (2021). Miura 5. https : / / www . pldspace . com / es / miura - 5 (accessed:
27.05.2021)

62
5 BIBLIOGRAPHY

[28] Gradl, P., Mireles, O., & Andrews, N. (2020). Introduction to additive manufacturing for
propulsion and energy systems. https://doi.org/10.13140/RG.2.2.23228.05761
[29] Zhang, T., & Miyamoto, C. (2014). 3d printing: A cost effective and timely approach to
manufacturing of low-thrust engines. https://doi.org/10.2514/6.2014-3502
[30] Sciaky Inc. (2021). Industrial metal 3d printers. https://www.sciaky.com/additive-
manufacturing/industrial-metal-3d-printers (accessed: 01.06.2021)
[31] Ross, M. N., & Sheaffer, P. M. (2014). Radiative forcing caused by rocket engine emissions.
Earth’s Future, 2(4), 177–196. https : / / doi . org / https : / / doi . org / 10 . 1002 /
2013EF000160
[32] United States Environmental Protection Agency. (2020). Air pollution: Current and future
challenges. https : / / www . epa . gov / clean - air - act - overview / air - pollution -
current-and-future-challenges (accessed: 16.05.2021)
[33] Stevens, M. H., Lossow, S., Fiedler, J., Baumgarten, G., Lübken, F.-J., Hallgren, K., Hartogh,
P., Randall, C. E., Lumpe, J., Bailey, S. M., Niciejewski, R., Meier, R. R., Plane, J. M. C.,
Kochenash, A. J., Murtagh, D. P., & Englert, C. R. (2012). Bright polar mesospheric clouds
formed by main engine exhaust from the space shuttle’s final launch. Journal of Geophysical
Research: Atmospheres, 117(D19). https : / / doi . org / https : / / doi . org / 10 . 1029 /
2012JD017638
[34] Baumers, M., Tuck, C., Hague, R., Ashcroft, I., & Wildman, R. (2010). A comparative study of
metallic additive manufacturing power consumption. 21st Annual International Solid Freeform
Fabrication Symposium - An Additive Manufacturing Conference, SFF 2010, 278–288.
[35] Kessler, D. J., & Cour-Palais, B. G. (1978). Collision frequency of artificial satellites: The
creation of a debris belt. Journal of Geophysical Research: Space Physics, 83(A6), 2637–
2646. https://doi.org/https://doi.org/10.1029/JA083iA06p02637
[36] European Space Agency. (2021). Esa and unoosa illustrate space debris problem. https :
//www.esa.int/Safety_Security/Space_Debris/ESA_and_UNOOSA_illustrate_
space_debris_problem (accessed: 17.05.2021)
[37] Wen, K., Chen, X.-w., & Lu, Y.-g. (2020). Research and development on hypervelocity impact
protection using whipple shield: An overview. Defence Technology. https : / / doi . org /
https://doi.org/10.1016/j.dt.2020.11.005
[38] Noble, B., Almanee, Y., Shakir, A., & Park, S. (2016). Design and evaluation of an orbital
debris remediation system, 159–164. https://doi.org/10.1109/SIEDS.2016.7489290
[39] Oficina Catalana del Canvi Climàtic. (2021). Guia pràctica per al càlcul d’emissions de gasos
amb efecte d’hivernacle (geh). https://canviclimatic.gencat.cat/web/.content/
04_ACTUA/Com_calcular_emissions_GEH/guia_de_calcul_demissions_de_co2/
190301_Guia-practica-calcul-emissions_CA.pdf (accessed: 10.06.2021)

63