THERMODYNAMIC MODELING

OF ADVANCED CYCLES FOR

HIGH SPEED PROPULSION
Author
Jose Tomás Sánchez Esparza

Directors
Bayindir H. Saracoglu
Pedro Piqueras Cabrera

Von Karman Institute for Fluid Dynamics

Universitat Politècnica de València
Escuela Técnica Superior de Ingeniería del Diseño

Valencia, July 2018

Thermodynamic modeling of advanced cycles
for high speed propulsion

Jose Tomás Sánchez Esparza

Von Karman Institute for Fluid Dynamics

Universitat Politècnica de València
Escuela Técnica Superior de Ingeniería del Diseño

July 2018
 Choose life. Choose a job. Choose a career. Choose a family. Choose a big television,
choose washing machines, cars, compact disc players, and electrical tin openers. Choose
good health, low cholesterol and dental insurance. Choose fixed-interest mortgage
repayments. Choose a starter home. Choose your friends. Choose leisure wear and
matching luggage. Choose a three piece suite on hire purchase in a range of fucking
fabrics. Choose DIY and wondering who you are on a Sunday morning. Choose sitting
on that couch watching mind-numbing spirit-crushing game shows, stuffing junk food into
your mouth. Choose rotting away at the end of it all, pishing your last in a miserable
home, nothing more than an embarrassment to the selfish brats you have spawned to
replace yourselves. Choose your future. Choose life . . . But why would I want to do a
thing like that? I chose not to choose life: I chose something else. And the reasons?
There are no reasons. Who needs reasons when you’ve got heroin?

Mark Renton - Trainspotting

4
 Acknowledgments

I would like to express my gratitude to my supervisor at the Von Karman Institute for
Fluid Dynamics for helping me so far with my project during my internship, for teaching
me and giving me pieces of advice in so many topics, even not related with science and
technology. To my Erasmus coordinator, Sergio Hoyas Calvo, whose dedication makes
students go abroad and become more competent, efficient and better people. To Francisco
Torres Herrador, for helping me in the VKI from the first minute. And finally, to my
thesis supervisor in UPV, Pedro Piqueras Cabrera, for helping me with my thesis from
my home university.

I would also like to thank very deeply my friends from Patrimmonia for being my
family during this incredible year. Thank you all, this project has been quite easier to
carry forward thanks to your hard emphasis to distract me and make me enjoy out of the
master’s life.

Finalmente, quiero agradecer a mi familia la oportunidad que me brindan cada día de

poder lograr todo lo que me propongo. Gracias por estar siempre ahí, por apoyarme en
cada decisión y en cada gesto. Sin vosotros nada de esto sería posible.

i
ii
 Abstract

Majority of the hypersonic and supersonic air vehicles utilizes complex propulsion systems
empowered by combined-cycle engines. The architecture of such engines is quite com-
plex to model with high fidelity methods. Transients within variety of the components
throughout the trajectory play an important role on the operability and consequently on
the design of the component system.

Hence, 0-Dimensional and 1-Dimensional approaches gain relevance for the concep-
tual design viability studies of the advanced propulsion cycles, such as air-turbo rockets,
scramjets or detonation engines. The project aims to construct a framework suitable to
design and simulate modern propulsion systems for high-speed transport or space access.

One of the combined-cycle engines is modeled and studied in this document. The Air
Turbo Rocket, ATR, is understood as the evolution of the turbojet and the rocket engine,
due to the fact that it combines components and characteristics of both of them. Moreover,
it provides a unique set of features because the shaft power to the fan is independent from
the flight regime, and hence from the fan performance. In addition, the ATR Expander,
ATR EXP, is equipped with a regeneration system composed by two heat exchangers, one
for the combustion chamber and one for the nozzle, to add heat to the working fluid, the
fuel, and increase the thermodinamic efficiency.

This report focuses its efforts in the analytic research of some features of the ATR
EXP in Matlab. Namely, the operational range of the engine will be studied, as well
as the minimum area nozzle required for the mission, the minimum turbine work and the
bleeding ratio that is necessary in supersonic regime. Apart from that, a brief introduction
to EcosimPro and its environment has been made, giving examples of high speed engines
studied in this numerical software.

iii
iv
 Resumen

La mayoría de los vehículos hipersónicos y supersónicos no autónomos utilizan complejos

sistemas de propulsión dotados de motores de ciclo combinado. La arquitectura de estos
motores es bastante compleja como para modelarla con un alto grado de precisión. Los
transitorios dentro de la variedad de los componentes que conforman el motor, a lo largo
de toda la trayectoria del vehículo, juegan un rol importante en cuanto a la operatividad
de la misión, y consecuentemente en el diseño del componente.

Por tanto, los métodos 0-dimensional y 1-dimensional tienen bastante relevancia en

este tipo de aplicaciones, para centrarse en un diseño conceptual y de viabilidad para cic-
los avanzados de propulsión, tales como los air turbo-rockets, los scramjets y los motores
de detonación. Este proyecto tiene como objetivo construir un marco factible para diseñar
y simular los modernos sistemas de propulsion para el transporte de alta velocidad o el
acceso al espacio.

Uno de los motores de ciclo combinado se ha modelado y estudiado en este documento.

El motor Air Turbo Rocket, ATR, está considerado una evolución del motor turbojet y el
motor cohete, debido al hecho de que combina tanto componentes como características de
ambos. Además, proporciona una serie de características únicas, ya que la potencia del
eje que se da al fan es independiente de la condición de vuelo, y por tanto, del rendimiento
del fan. Adicionalmente, el ATR Expander, ATR EXP, está equipado con un sistema de
regeneración compuesto por dos intercambiadores de calor, uno para la cámara de com-
bustión y otro para la tobera, que proporcionan calor al fluido de trabajo, el fuel, para así
incrementar la eficiencia termodinámica del ciclo.

Este trabajo focaliza sus esfuerzos en el estudio analítico de algunas características del
ATR EXP en Matlab. Concretamente, se estudiará el rango operacional del motor, así
como también el área mínima de la tobera requerida para la misión, el trabajo mínimo
que debe hacer la turbina, y el coeficiente de sangrado que debe existir para régimen
supersónico. Aparte de eso, se hará una introducción breve a EcosimPro y su entorno,
dando ejemplos de motores de alta velocidad estudiados en este software numérico.

v
 Nomenclature

A transversal area [m2 ]

a speed of sound [m s−1 ]

B bypass ratio [−]

Ch constant specific heat capacity [m2 s−2 K −1 ]

Cp heat capacity at constant pressure [m2 s−2 K −1 ]

D diameter [m2 ]

Dh hydraulic diameter [m2 ]

e total energy e = H − pν[m2 s−2 ]

G mass flux [kgm2 s−1 ]

g gravity constant [m2 s−2 ]

H total enthalpy H = h + 0.5ν 2 [m2 s−2 ]

h enthalpy[m2 s−2 ]

hc convective heat transfer coefficient [m2 s−2 ]

If fluid equivalent inertia [m−1 ]

Iv valve equivalent inertia [m]

Isp specific impulse [ms−1 ]

K thermal conductivity [kgms−3 K −1 ]

Kf pressure loss factor [−]

q heat [m2 s[−2]]

T temperature [K]

T thrust [N ]

Tsp specific thrust[N s−1 ]

v velocity [ms−1 ]

W Work [kgm2 s−2 ]

Ma Mach number

Nu Nusselt number

Pr Prandtl number

Re Reynolds number

vii
Acronyms

ATR Air Turbo-Rocket

ATREX Air Turbo-Ram-jet wit Expander engine

CFD Computational Fluid Dynamics

DAE Differential Algebraic Equation

DASSL Differential Algebraic System Solver Algorithm

DMR Dual-Mode Ramjet

ESA European Space Agency

ESPSS European Space Propulsion System Simulation

HAP Hypersonic Air-Breathing Propulsion

HSI High Speed Intake

LSI Low Speed Intake

MR Air-to-Fuel Ratio

RBCC Rocket Based Combined Cycles

SABRE Synergetic Air-Breathing Engine

TIT Turbine Inlet Temperature

TPR Total Pressure Recovery

TSFC Thrust Specific Fuel Consumption

TSTO Two Stage to Orbit

Greek Symbols

αc intake mass capture ratio

∆ increment

δp pressure loss

r wall rugosity [m]

ηn nozzle efficiency

ηo overall efficiency

ηp propulsive efficiency

ηth thermal efficiency

γ heat capacity ratio

viii
ν dimensionless velocity

π compression ratio

ρ density [kgm−3 ]

τ characteristic time [s]

ξ friction factor [m−1 ]

Superscripts

o stagnation quantity

Subscripts

∞ free stream

a air

amb ambient

cc combustion chamber

crit critical condition

f fan

h hydrogen

N net

n nozzle

p pump

t turbine

Other Symbols

∆H specific work [m2 s−2 ]

ṁ massflow [kgs−1 ]

q̇ heat flux [kgs−3 ]

F thrust [kgms−2 ]

ix
x
 Contents

Abstract iii

Resumen v

General Index x

Figures Index xii

Tables Index xiv

1 Introduction 1
1.1 High speed propulsion context . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 High speed propulsion and its environment . . . . . . . . . . . . . . 1
1.1.2 Hypersonic vehicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.3 Cycles designed for high speed propulsion . . . . . . . . . . . . . . . 3
1.1.4 Air Turbo-Rocket engines . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Project goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 The Numerical Tool: EcosimPro 6

2.1 A general introduction to the software . . . . . . . . . . . . . . . . . . . . . 6
2.2 The object oriented modeling . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3 Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.4 The ESPSS libraries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

3 The Air Turbo Rocket engine 16

3.1 Engine background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2 Thermodynamic cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.3 Operational functioning for a given mission . . . . . . . . . . . . . . . . . . 21

4 Engine modeling in EcosimPro 26

4.1 Turbojet engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.2 Ramjet engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.2.1 Single point calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.2.2 Parametric study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.3 High-speed propulsion engines modeling . . . . . . . . . . . . . . . . . . . . 39

5 Studies and Results 42

5.1 Influence study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
5.2 Operational range of the engine . . . . . . . . . . . . . . . . . . . . . . . . . 43
5.3 Engine performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.4 Nozzle throat area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.5 Turbine work vs πf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

6 Financial estimation 66

7 Conclusions and future work 68

7.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
7.2 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

xi
xii
 List of Figures

2.1 EcosimPro software - general view . . . . . . . . . . . . . . . . . . . . . . . 6

2.2 Example of library "Fluid Flow 1D" with all its components [1] . . . . . . . 8

3.1 The ATR-EXP cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

3.2 MR2 aircraft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.3 MR2, the civil transport aircraft. The numbers make reference to the sta-
tions: 1 low speed intake, 2 high speed intake, 3 nozzle, 4 ATR duct, 5
DMR duct. [2] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.4 MR2 trajectory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.5 MR2 installed thrust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.6 MR2 intake TPR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.7 MR2 Low Speed Intake ratio . . . . . . . . . . . . . . . . . . . . . . . . . . 23

4.1 Schematic of the turbojet engine . . . . . . . . . . . . . . . . . . . . . . . . 26

4.2 Turbojet engine model in EcosimPro . . . . . . . . . . . . . . . . . . . . . . 26
4.3 Boundary conditions wizard to create a partition in EcosimPro . . . . . . . 28
4.4 Thrust vs TIT for a turbojet engine, modeled in EcosimPro . . . . . . . . . 31
4.5 Thrust vs TIT for a turbojet engine, modeled in GasTurb . . . . . . . . . . 31
4.6 Nozzle area vs TIT for a turbojet engine, modeled in EcosimPro . . . . . . 32
4.7 Nozzle area vs TIT for a turbojet engine, modeled in GasTurb . . . . . . . 32
4.8 Ramjet schematic in EcosimPro . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.9 Valve position vs time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.10 Temperatures at nodes 1, 12 and 30 in the ramjet combustion chamber . . . 35
4.11 Air, fuel and total massflows . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.12 Mach number histogram inside the combustion chamber . . . . . . . . . . . 35
4.13 Ramjet engine thrust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.14 Combustion chamber temperature vs altitude and Mach number . . . . . . 37
4.15 Fuel-to-air ratio vs altitude and Mach number . . . . . . . . . . . . . . . . . 38
4.16 Ramjet engine thrust vs altitude and Mach number . . . . . . . . . . . . . . 38
4.17 Air Turbo-Rocket Expander engine schematic in EcosimPro . . . . . . . . . 39

5.1 Turbine expansion ratio vs heat addition to the working fluid, at M0 = 1.1 . 44
5.2 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 1.1 and πf = 2.4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.3 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 1.1 and πf = 3.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.4 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 1.1 and πf = 3.4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.5 Turbine expansion ratio vs heat addition to the working fluid, at M0 = 1.7 . 47
5.6 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 1.7 and πf = 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.7 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 1.7 and πf = 2.5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
5.8 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 1.7 and πf = 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
5.9 Turbine expansion ratio vs heat addition to the working fluid, at M0 = 2.1 . 49

xiii
 LIST OF FIGURES

5.10 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 2.1 and πf = 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
5.11 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 2.1 and πf = 2.5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
5.12 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 2.1 and πf = 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5.13 Turbine expansion ratio vs heat addition to the working fluid, at M0 = 3.2 . 51
5.14 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 3.2 and πf = 1.5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
5.15 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 3.2 and πf = 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
5.16 Turbine expansion ratio vs heat addition to the working fluid, at M0 = 4 . . 53
5.17 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 4 and πf = 1.25 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
5.18 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 4 and πf = 1.5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
5.19 Turbine expansion ratio vs heat addition to the working fluid, at M0 = 4.5 . 55
5.20 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 4.5 and πf = 1.2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.21 Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 4.5 and πf = 1.4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.22 Engine performance at M0 = 1.1 . . . . . . . . . . . . . . . . . . . . . . . . 57
5.23 Engine performance at take off and M0 = 1.7 . . . . . . . . . . . . . . . . . 58
5.24 Engine performance at M0 = 2.1 and M0 = 3.2 . . . . . . . . . . . . . . . . 58
5.25 Engine performance at M0 = 4 and M0 = 4.5 . . . . . . . . . . . . . . . . . 59
5.26 Nozzle throat area vs πf and MR in subsonic regime . . . . . . . . . . . . . 60
5.27 Nozzle throat area vs πf and MR in supersonic regime . . . . . . . . . . . . 60
5.28 Nozzle throat area vs πf and bleeding ratio . . . . . . . . . . . . . . . . . . 61
5.29 Nozzle throat area vs πf and turbine work in subsonic regime . . . . . . . . 62
5.30 Nozzle throat area vs πf and turbine work in supersonic regime . . . . . . . 62
5.31 Turbine power vs. fan pressure ratio and mixture ratio, at subsonic regime 63
5.32 Turbine power vs. fan pressure ratio and mixture ratio, at supersonic regime 64

xiv
 List of Tables

4.1 Input data for the turbojet calculus . . . . . . . . . . . . . . . . . . . . . . 29

4.2 Design point calculated in EcosimPro and GasTurb . . . . . . . . . . . . . . 30
4.3 Initial parameters for a single point calculus in the ramjet engine . . . . . . 34
4.4 Boundary conditions for the parametric study . . . . . . . . . . . . . . . . . 36

5.1 Input data for the meaningful parameters of the thermodinamic cycle . . . 42
5.2 Influence study over the turbomachinery cycle . . . . . . . . . . . . . . . . . 43
5.3 Trajectory points used for the operational range study . . . . . . . . . . . . 43

6.1 Project’s financial estimation . . . . . . . . . . . . . . . . . . . . . . . . . . 66

xv
xvi
 1
Introduction

1.1 High speed propulsion context

1.1.1 High speed propulsion and its environment
High speed propulsion is becoming nowadays a challenge in which thousands of engineers
are working on, making big progress and achieving goals to make possible high speed
propulsion as a fact in the long-term future. In order to accomplish this objective, some
programs are being carried out by the main agencies in aerospace investigation, such as the
European Space Agency, ESA, with its program LAPCAT, Long-Term Advanced Propul-
sion Concepts and Technologies.

For developing an efficient propulsion plant that is able to travel at supersonic or

even hypersonic velocities and whose applications will be focused on commercial flights or
space access, there are some tasks to deal with: the airframe and the optimization of the
propulsion unit. When the trajectory analysis is being studied, its common to presume
the performances of both the airframe: drag, lift, thrust, fuel consumption, etc.; and the
propulsion plant, which are required to accomplish the mission. In turn, the airframe
and the propulsion unit are designed assuming performances too, and then the mission
is imposed on the performances calculated when overcoming this step. The separated
design of these areas from the general system is often performed in subsonic airplanes
and applications, which constitute problems in which installation and interaction effects
are less sensitive to the operational range and flight regime. Nevertheless, components
become more delicate when flying at supersonic regime, e. g. the intake is a critical
component, due to the fact that it has to deal with lots of different scenarios of velocity
and ambient conditions, so its range of operation has to be very wide in order not to
establish a mismatch between it and the downstream components like compressors, and
then provoke a huge loss in efficiency, specific thrust and consumption, etc. That’s why
it is very important, even more in supersonic and hypersonic regime, the address of the
total interaction between components and between the propulsion unit and the airframe
from the very early design stages.

The propulsion unit design include complex challenges that need to be overcome. De-
spite the effort employed in ramjet and rocket engines, the High Air-Breathing Propulsion,
HAP, technology presents disparities related to its flight regimes that have no easy solu-
tions. It may exists a solution for one flight speed corridor, but that is not good for the
other ones. Then, it is been suggested to combine different propulsion systems in order to
facilitate the engine operation at different flight speeds with different modes of propulsion.
For instance, the Turbine Based Combined Cycle, TBCC, join both the turbine and the
ramjet/scramjet systems. In this framework, the turbine cycle is used to drive the aircraft
to flight speeds in which the ramjet/scramjet unit can be ignited and then used to reach
and maintain the hypersonic regime. Nevertheless, TBCC cycles contain multiple difficul-
ties of implementation, as the transition through the different propulsion cycles and the
effective integration between them [3]. Advancements in this area are needed, and this is
in turn subdivided in several areas of investigation, such as:

1
 CHAPTER 1. Introduction

• Air breathing engines:

- Dual-mode ramjet/scramjets, DMR.

- High speed turbines
- Rocket based hypersonic flight; limited success with ramjet missiles at subsonic
to supersonic speeds.
- Operational hypersonic vehicle even with much work done in ramjets and scram-
jets.

• Rocket propulsion:

- Improvements to liquids, solids and hybrids used in RBCC or booster stages.

– Combustion inestability.
– Internal flowfield modeling.
– Components; nozzles and turbomachinery.
- Booster stages to make them more efficient and cost effective.
– Flight testing.
– TSTO system.

• Combined cycles:

- Turbine based combined cycles, TBCC,

– Variable geometry.
– High speed turbines and turbojets.
– Transitioning and integration.
– Materials.
– Thermal management.
– Engine components.
- Rocket based combined cycles, RBCC,
– Rocket-scramjet integration.
– Linear aerospike rockets and nozzles.

• Compressor and turbine blades.

• Engine components.

- Seals.
- Turbomachinery.
- Seals.

• Fuels and combustion.

• Heat management.

• Engine performance.

2
SECTION 1.1. High speed propulsion context

• Flow paths:

- Intake.
- Isolator.
- Combustion chamber.
- Nozzle.
- Structure.
- Acoustics.

As this list suggests, there is a huge amount of remaining work to make high speed
propulsion based on combined cycles a reality. This document is going to be focused just
on describing, modeling and studying one type of the combined cycle engines, the Air
Turbo-Rocket, ATR, but first a review of hypersonic vehicles is introduced.

1.1.2 Hypersonic vehicles

There are several ways to classify hypersonic vehicles, and one of the most interesting
ones is by their propulsion plants. Vehicles with a single type of power plant usually uses
either variable or multimodal cycle engines. Vehicles using a combination of power units
base its operational work on different engines, depending on the flight phase. A repre-
sentation of this concept could be a turbo-ramjet combined with several rocket engines
to thrust up the Sänger two stage to orbit, TSTO, vehicle [4], [5], and a cycle composed
by an ATR and a DMR to power up the LAPCAT II Mach 8 civil transport vehicle. On
its behalf, multimodal engines are capable of modifying their configuration to accomplish
the best efficiency at each stage of the flight. This is a concept represented by the Syner-
getic Air-Breathing Engine, SABRE, designed for launching the Skylon by a single stage to
orbit, SSTO, mission [6], [7], and the Scimitar, for the LAPCAT A2 Mach 5 vehicle [8], [9].

Designing a high speed air-breathing cycle engine is an assignment in which there

exists a trade-off to be followed between the specific impulse, thrust-to-weight ratio and
operational envelope. The specific impulse is a very useful parameter to measure the cycle
efficiency, due to the fact that is defined as the thrust generated by the engine divided by
the fuel mass flow it consumes. TBCC are very effective to maximize the specific impulse
in this kind of missions, so that’s the reason why the majority of engines related to this
technology are based on these components.

1.1.3 Cycles designed for high speed propulsion

The propulsion cycles thought to achieve the high speed regime are differentiated in three
main categories, depending on the implementation of TBCC, RBCC or a ramjet/scramjet
based cycle.

The TBCCs are composed of the turbojet and the ramjet cycles, particularly turbo-
ramjet (Griffon II engine [10] and Sänger) and turbofan-ramjet (RTA-GE57, evolution of
J58 engine from the SR71 Blackbird [11], [12]). Variants of this technology comprehend the
modifications of the classical turbojet configuration, the ATR [13], [14] and the ATREX
(Air Turbo-Rocket Expander) [15], a pre-cooled ATR. There are other configurations, such

3
 CHAPTER 1. Introduction

as the MIPCC engine [16] and the Rocket Augmented Turbine [17] which include some
modifications to the turbojet cycle to provide some specific performances.

The RBCCs are based on combinations of air-breathing and rocket elements. For in-
stance, SABRE, Scimitar, KLIN, RB-545 and GTX engines [7], [18], [6]. SABRE and
Scimitar engines are similar to the ATR in the core, but with a combustion chamber and
a nozzle characteristic of rocket systems.

Finally, the ramjet/scramjet engines have a particularity: they do not provide static
thrust, so they need to be accelerated by other propulsion plant. Consequently, the booster
and the engine are integrated to achieve a good configuration. If desired, more information
about this type of cycles is available in [7], [18], [6], [19].

1.1.4 Air Turbo-Rocket engines

Within the ATR technology, there are two distinguished categories depending on the
source of the hot gas driving the turbine. The configuration in which this document is
focused is called the EXP one: the combustion is carried out just in one single combustion
chamber, and the turbine is driven by regeneratively heated fuel extracted from this single
chamber and from the nozzle. The second category, the gas generator, GG, is based on
a pre-combustor burning a rich mixture of on-board oxidizer and fuel to generate the hot
gases which are going to discharge to the main combustion chamber that mixes these gases
with the air coming from the atmosphere.

NACA designed an ATR-GG in 1956 [13], based on gasoline as fuel and nitric acid
as oxidizer, with a driven fan through a gearbox and a partial admission turbine. The
optimum condition was established to Mach 2.3 and 14 km of altitude. When the super-
sonic cruise is taking place, the fan is windmilling and then the engine is behaving as a
ramjet engine. During the 60s other concepts of ATR-GG were patented, as well as EXP
configurations were starting to be developed [20], [21].

SNECMA developed in 1992 an ATR-GG combined with a ramjet engine. It was

based on two counter-rotating fans driven by two contra-rotating turbines, and there was
H2 cooling to the fan blades when the flight speed exceeded Mach 3. As the speed in-
creases, the blades change its position by pitch control to make the transition to ramjet
mode.

The ATR EXP started its development thanks to the investigations carried out by
ASTRIUM [22] and SNECMA [23], based on preliminary data extracted by ESTEC [24].

4
SECTION 1.2. Project goals

1.2 Project goals

There are several objectives that will be accomplished in the developing of this project:

• Documenting and studying high speed propulsion engines constitutes the first objec-
tive of the project, due to the extensive background that exists and that constitutes
the starting point of the current work. An exhaustive compilation of information
will be made, studying engines such as the SABRE, the Scimitar and the ATR-EXP.
Nevertheless, the main goal of the project is studying this last engine, so understand-
ing this engine, its meaningful parameters and its behavior when some conditions
have changed will be necessary.

• Learning and understanding the software EcosimPro will be the second objective.
As next sections will describe, this software constitutes the perfect tool for modeling
high speed propulsion engines.

• Being able to develop some studies about turbojet and ramjet engines in EcosimPro,
and its homologous in some other software such as Matlab and GasTurb for future
comparisons and validations will be the third objective, proving that the student has
acquired the sufficient knowledge for modeling high speed propulsion engines.

• The final objective of the project is focused on studying in Matlab the high speed
propulsion engine ATR-EXP, Air Turbo Rocket - Expander. The utilization of this
software is due to its capability of performing studies faster than EcosimPro, be-
ing able to obtain tendencies, behaviors and conclusions that will be used for more
specific studies in EcosimPro. Thus, this studies will be used for obtain the mean-
ingful parameters of the engine, focused on the turbomachinery cycle. Besides, the
operational range of the engine will be obtained, and once this framework is cal-
culated, several studies analyzing behaviors when some parameters have changed,
such as the bleeding, the nozzle area or the turbine specific work, will be carried out.
Similar studies in EcosimPro, in which the accuracy of the model will be increased
considerably, are proposed for future works.

5
6
 2
The Numerical Tool: EcosimPro

The methodology followed in this document will be presented below. For the realization
of the different studies that have been developed in this project, some numerical tools
have been used. These tools allows the user to analyze the engine in its several states of
working, such as design, off-design and transient conditions. In the design phase, the cal-
culus focuses on the determination of the geometry and the cycle for a specified mission.
Then, the off-design operation is analyzed to study the engine behavior throughout its
operational range. Finally, the transient stage is the most complex one, due to the fact
that it combines the off-design state of all components which integrate the engine, plus
the transient term of the governing equations.

In order to study these three states of operation, the simulation environment Ecosim-
Pro has been used, combined with the European Space Propulsion System Simulation
(ESPSS) libraries, which include all the components needed to build the engine models.

2.1 A general introduction to the software

As commented before, EcosimPro is a simulation tool developed by Empresarios Agru-
pados which counts on an user-friendly environment destined to model both simple and
complex physical processes that are declared either as systems of differential-algebraic
equations or as ordinary-differential equations, combined with discrete events.

Figure 2.1: EcosimPro software - general view

EcosimPro employs a specific programming language named EL, which is thought to

model physical systems based on discrete events and equations. This language is the
official one utilized by the European Space Agency (ESA) and it permits the mathemat-
ical modeling of physical components by representing them through differential-algebraic

7
 CHAPTER 2. The Numerical Tool: EcosimPro

equations. The goal of the EL developing is modeling combined continuous and discrete
systems, doing it in a clear and easy way.

EL is a very powerful tool that is used in applications such as preparing experiments,

calculating design points and off-design behaviors, transient processes, etc. It is not only
useful for developing models and computing them, but also for the post-processing task,
by generating plots, reports, and compatible data with other languages like FORTRAN,
C and C++.
This language is designed to solve systems of differential-algebraic equations automati-
cally. Nevertheless, the complexity is hidden internally, where the system is solved: the
user has to define the high-level equations by using a high-level object-oriented language
that is very similar to other object-oriented languages. Consequently, in EL language
components have the capability to inherit properties, equations and variables from other
object, and in turn can be established as a base for creating a more complex object or
system. As a matter of fact, these characteristics allow the user to be able to create big,
high complex models that can go from a simple electric circuit to a huge engine designed
for a space mission.

There are many concepts in EcosimPro, but the most important ones are:

• Workspaces: A workspace consists on a combination of libraries the user is going

to use in its simulations and experiments. The libraries can be included, deleted
or created, depending on the necessities of the project. Alternatively, the user can
create different workspaces with different configurations and libraries if desired.

• Libraries: Libraries are sets of elements related to the same area of study, stored
in one unique file. In a library, multiple kinds of elements can be included, such as
components, variables, constants, functions, ports or classes. An important property
of this type of units is that they can interact in between with other libraries, due to
their multi-disciplinary constitution.

• Components: These concepts are the most important ones in the whole modeling
task. A component represents a physical unit, like a compressor, a turbine, a valve,
etc., and it models as well its discrete, continuous or sequential behavior, depending
on the case study. A component can be connected to other components to create
general systems, which are going to be computed for different goals and purposes.
For instance, an inlet could be connected to a compressor, a combustion chamber, a
turbine and a nozzle to build up a turbojet engine model.
There are two categories: concrete and abstract components. The last ones are
used just to join sets of variables, discrete events and equations, and have no real
significance. Abstract components are inherited by others, establishing a parent-
and-children structure in which children use the data from a general parent. It
provides simplicity to the code, since it is well structured and avoid repetition.

• Ports: These units represent the connection between components. There are several
categories of ports, and the user can create them and choose its category depending
on the necessity of the model [25].

8
SECTION 2.1. A general introduction to the software

Figure 2.2: Example of library "Fluid Flow 1D" with all its components [1]

Actually, ports exchange variables between the components they join. The main ad-
vantage of working with this type of element is that no connection between variables
is needed, because the port is the one in charge to transmit data from one component
to the following one. Ports have several properties, such as the type of data they
work with (Real, boolean, etc.), the type of connection (equal, sum, etc.), and the
direction (in or out). In turn, there are several categories of ports, depending on
what is being transported; fluid, thermal energy, control data, etc.

• Partitions: Once the system is made by choosing and connecting the components,
the mathematical model has to be created by ordering the equations, choosing the
boundary and initial conditions, and proposing a set of residue equations and tearing
variables. Partitions are exactly this process, and for a single model several of
them can be created, by choosing different boundary conditions, different unknown
variables, etc.

• Experiments: Once the partition is created, experiments are the next step in the
pre-process stage of the simulation. They are created to define the correct use of
the model, by specifying the initial state of the system, the boundary conditions,
integrating the model between an initial and a final time and choosing the step,
choosing whether doing a steady or a transient calculus, etc.

Regarding the mathematical capabilities of EcosimPro, the software works with sym-
bolic equations (derivatives, reduction, etc.) and numerical solvers for differential-algebraic
equations (DAE) such as Runge-Kutta and DASSL, Differential-Algebraic System Solver
algorithm, and non-linear equations (Newton-Raphson) [26]. The program automatically
controls the interaction between these solvers, and uses sparse and dense matrix formats
which change in function of the Jacobian matrix, in order to deal with thousands of vari-

9
 CHAPTER 2. The Numerical Tool: EcosimPro

ables that take part in the mathematical model [27].

When creating a partition, the numerical tool is provided with wizards for defining de-
sign problems, establishing boundary conditions, overcoming algebraic loops and reducing
DAE problems.

Given a non-linear system of equations as the following one

(
F (ẋ, x, t; u) = 0
(2.1)
x(t0 ) = 0

For the resolution of this system of equations, the Differential-Algebraic System Solver
algorithm (DASSL) for implicit systems is used. DASSL results very useful when two types
of problems need to be solved, and cannot be handled by other ODE solvers. Contrary
to Runge Kutta method, for instance, it is not necessary that the derivative is in explicit
form. Another type of DASSL can be used in EcosimPro, which is the DASS-SPARSE
solver. It is used when the Jacobian is very large, with a high number of zeros. By us-
ing the SPARSE option, the calculation time is reduced considerably. The advantage of
using DAEs is that when modeling a physical system, the originated equations are DAEs
themselves, and not ODEs. ODEs are special cases of DAEs, so an ODE can be written
explicitly as a DAE.

The solver used by EcosimPro (DASSL) consists on a method in which the time deriva-
tive y 0 (t) is replaced with an approximation by differences, and the resulting equations
for time are solved by using an implicit Powell’s Hybrid Method, by iterations. Then, the
algorithm replaces the time derivative not only using first order approximations, but also
by using a backward differentiation of order, where k is between 1 and 5. The value of k
and the time step are parameters chosen automatically by the solver at each step, based
on the behavior of the solution. As mentioned before, the implicit iteration technique
is a Powell’s Hybrid Method. As in any other method, the vector of unknown variables
requires a set of starting values. The convergence is obtained faster when these values are
close to the solution ones. The DASSL algorithm estimate the starting values, evaluating
the interpolation polynomial for the solution calculated the last k + 1 times at the current
time tn .
If the initial conditions are inconsistent, the solver may diverge. This difficulty is observed
when creating large algebraic systems by introducing algebraic compatibility equations be-
tween different components. To solve this problem, the order of the algebraic constraints
has to be reduced through the implementation of the perturbed equation. In the case of
the components that take part in an ATR engine, this perturbed equation corresponds to
the compatibility of mass between two connected components, named by 1 and 2.

ṁ1 − ṁ2 = τ m
¨2 (2.2)

This equation replaces the original stiff equation which was obtained when the char-
acteristic time was approaching to zero. By this measure, the system of equations will be
split in two separated parts, and will be solved independently.

10
SECTION 2.2. The object oriented modeling

2.2 The object oriented modeling

When building up models composed by a set of components, object-oriented languages
constitute a powerful tool to avoid complexity and leverage the code reuse, in order to
create models which can be independent and whose maintainability is considerably easy.
The development based on modular procedures makes possible the creation of complex
models by combining two methods:

• Using existing components and inherit them by extension.

• Aggregating and instantiating existing components.

These methods are thought to create components that are able to represent real and
entire physical systems. Nevertheless, single or intermediate components could be simu-
lated as well (a turbine, for instance, could be simulated by aggregating some components
to study its performance, it is not needed to simulate big systems as turbojet engines to
study the behavior of one single component). The development of the model and the main-
tenance time that are needed for consistent results are critical tasks that are optimized
with some important tool

• Encapsulation: Encapsulation of complexity is one of the goals of an object-oriented

language and modeling. Both the abstraction capability and the encapsulation of
data are very efficient ways to make the process much simpler. The main elements
to carry out this task are components and libraries.
For instance, a certain user could build a model in which there are hundreds of vari-
ables and equations that may work properly, but when other user wanted to read
it, it could become difficult to follow and understand. The libraries include all the
elements which are part of one subject, and the components constitute the funda-
mental unit for expressing a dynamic behavior by modeling its physical state and
operation. Their public interface lie in its data, ports and construction parameters,
and are visible for the user when the modeling phase is being carried out. The port
includes variables for connections to the components, and they transport also the
behavior and restrictions between ports. The definition of different classes of ports
abide in the system is capable of insert connection equations for the user not to be
concerned about them. If desired, the user could add some new components without
being worried about the connection equations but their internal behavior, because
the connections between ports are the ones in charge to determine whether the union
is possible or not.

• Inheritance: This is the tool employed for provide EL language of power for sharing
behavior and interfaces. When many components are used in the same model, they
may share a quite similar behavior. In order to simplify the mathematical model,
the language could deliver equations and data from the parent component to the
children ones, in order not to extend the model with similar equations and variables.
As a matter of fact, libraries are created by defining a parent component and then
inheriting its properties to other children, which will have the same properties than
its father, and then changing the code and including new features to specify the
behavior of this last component. Moreover, the software allows multiple inheritance;

11
 CHAPTER 2. The Numerical Tool: EcosimPro

a component could be created by taking properties from several parents, so the

creation of new components by using different parts of other ones is allowed.

• Aggregation: As commented before, a component can be created by inheriting prop-

erties from other one or ones. As well, they can have some instances of other com-
ponents, in order to increase the feature of reusing code and properties, since the
new component is based on others already built up.
If the language apply this principle iteratively and without any limit, the final compo-
nent could have a huge complexity, due to the large quantity of equations it includes.
Nevertheless, this complexity is internally hidden by the aggregation of previously
verified instances of other classes.

• Data abstraction: The language provides numerical data such as INTEGER, FLOAT
or REAL, and other convenient data as STRING, for instance.. As well, it includes
the possibility of defining user’s own data types [28].

2.3 Classes
In EL language, classes are quite similar to the ones present in other object-orientated
languages, such as Java or C++, but with one fundamental difference: EL is more restric-
tive, because it is simpler. It must be taken into account that users of this program are
engineers, not programmers.

This language gives the user the possibility to encompass data and behavior in the
same item, and then call it in a certain model to use it. The difference between a class
and a component is that the second one is equipped with dynamic equations and discrete
events that will be solved by the numerical tool. On its behalf, a class just depicts a
certain behavior, and only permits the declaration of methods and variables.

Classes are usually employed when the user want to model complex systems where
exists the possibility of improving the use of functions if they are referred to a similar
utility, so they are stored together and share memory by using similar variables.

2.4 The ESPSS libraries

EcosimPro is a software destined to model and calculate simple and complex physical
systems modeled by mathematical equations. In order to do this, an object-oriented pro-
gramming language is created, and its base, the components, are stored in libraries which
include equations, variables, ports, classes, etc., as commented before. The European
Space Propulsion System Simulation is an ESA initiative focused on the creation of a
platform for the spacecraft and launch vehicle propulsion systems simulation. Namely, it
is formed by a set of libraries designed for taking part in the numerical software Ecosim-
Pro. In order to provide real, consistent results, the libraries have to be updated with the
state-of-art technology at the time they are used for analyzing phenomena mainly focused
in propulsion systems.
The group in charge to create ESPSS and EcosimPro was composed by four companies:
Empresarios Agrupados, which are the ones under the care of coordinating the project

12
SECTION 2.4. The ESPSS libraries

and modeling ESPSS; CENAERO, responsible for developing the 1D algorithms, zooming
with CFD codes and system optimization, since it is a company specialized in CFD and
Optimization methods; EADS Astrium Space Transportation, in charge of creating the
User Requirements and Validation cases; and finally KOPOOS, responsible for the Tank
filling and Priming validation cases, since it is an enterprise with experience in satellite
propulsion systems, both with chemical and electrical power [29].

These libraries count on several tools designed to model propulsion systems features
such as:

• Standard properties for pressurant, propellant and some other fluids and materials,
taking into account the possibility of creating new ones by the user.

• A general fluid 1D model for two phase flows, destined to systems where heat transfer
takes place or controllers are needed.

• Libraries to model 1D turbo machinery and tanks behaviors.

• Combustion chambers, nozzles and cooling systems libraries.

It is remarkable that ESPSS includes a library focused just on steady components;

obviously, it is not the same to model a compressor actuating in its steady behavior or in
its real one, taking into account parameters like inertia, time delays, etc. Then, ESPSS
contemplate that possibility and provides libraries not only for real behavior, but for steady
one too. Having said this, the main applications for these libraries are the modeling of:

• Spacecraft systems, including processes of prime, heat transfer, tank behavior, pres-
sure regulators, etc.

• Ground pressurization systems.

• Priming phenomena, calculated in pipes or in other subsystems that are not the
main problem in spacecraft general systems.

• Liquid rocket engine cycles with one or several burners, including turbo machinery
and cooling systems. This is the application in which the present document is going
to be focused.

The circulating flow across the linear components, doing it in a one-dimensional way,
is resolved by the governing equations of mass, momentum and energy as:

∂ω ∂f (ω)
+ = Ω(ω) (2.3)
∂t ∂x

where the terms ω, f (ω) and Ω(ω) are the conservative variables, flux and source terms,
respectively.
 
ρ
 
ω = A ρν

 ρe 


13
 CHAPTER 2. The Numerical Tool: EcosimPro

 

 ρν 

f (ω) = A ρν 2 + p

ρν(e + p/ρ)

 

 0 

Ω(ω) = 21 ξρ|v|A + p(dA/dx) (2.4)
 
 q̇ 

The source term of the equation of energy, q̇ is related to the heat transfer coefficient,
hc , related to convective effects, and it can be rewritten as:

q̇ = hc (Tw − T ) (2.5)
where Tw and T corresponds to the wall and fluid temperature, respectively. As well,
a correlation to represent the heat transfer mechanism is needed, and it is modeled by
the Nusselt number, which is in turn utilized to compute the convective heat transfer
coefficient:

hc = N uk/Dh (2.6)
N u = f (Re, P r) (2.7)

where Dh is the hydraulic diameter of the section. The friction factor, which will model
the pressure loss per unit of length in pipes and ducts, is computed using the correlation
by [20Churchill]

ξ = Kf /Dh f (Re, r /Dh ) (2.8)

f (Re, ξr /Dh ) = 8((8/Re)12 + (K1 + K2 )−3/2 )1/12

K1 = [−2.457 ln (7/Re)0.9 + 0.27 r /Dh ]16 (2.9)

K2 = (37530/Re)16
where r is the wall absolute rugosity, and Kf is the loss factor, which is used for
scaling the pressure loss to satisfy design specifications.
Then, the Equation 2.3 is computed with a centered strategy in a grid space, where
ω, the conservation variables, and Ω(ω), the source terms, will be evaluated in the center
of every cell, that is, the nodes, and f (ω), the flux terms, at the interfaces, that is, the
boundary between one cell and the next/previous one [2].

The components of ESPSS are classified in resistive or capacitive components, similar

to the electrical circuits. A resistive component takes the state variables (density, pressure,
enthalpy, chemical composition and velocity) as input data, and then calculate the flow
variables, as volumetric, mass and enthalpy flows. The capacitive component makes the
inverse process, so when the user is building a model up, he has to combine one resistive
component with one capacitive, successively. The integral form of the Equation 2.3 is
computed in a resistive component, for example a valve, as:

14
SECTION 2.4. The ESPSS libraries


G|G|
 1 − ξ G2

 (G s 2
Gc ) ≤ ζ
s
If (AĠ + ȦG) + Iv Ġ = ∆po (2.10)
G|G|
(G

s 2
1− Gc ) >ζ


G2c

where Gs and Gc are the steady and critical mass fluxes, If and Iv represent the fluid
and valve inertia and ζ the pressure loss coefficient, and the pressure drop can be expressed
as:

1 1
∆po = (p + ρν 2 )in − (p + ρν 2 )out (2.11)
2 2
√
The steady solution to the upper branch for Equation 2.10 is G = Gs ζ, while for
the bottom branch, the chocked conditions, it is G = Gc . The fluxes Gs and Gc can be
computed as:
p
Gs = 2ρ∆po
!(γ+1)/2(γ−1)
2 + (γ − 1)M a2
Gc = ρa
γ+1
where the calculation of ρ, Ma and γ is calculated before entering in the valve. It is
known that the pressure losses in the laminar range are proportional to the mass flow, so
the source term can be linearized: G|G| → kG [36 thesis victor]. The pressure drop
when there’s turbulent flow, assuming constant area and a valve inertia close to zero, gives
the family of solutions as
√ !
Gs ∆po ζ
G(t) = √ tanh t (2.12)
ζ If AGs
which solution is limited by the supercritical brand
!
∆po
G(t) = Gc tanh t (2.13)
If AGs
On its behalf, the capacitive components, volumes and mainfolds, solve the conserva-
tion equations of mass and energy:
X X
ρ̇V = ṁin − ṁout (2.14)
in out
X X
(ρ̇e + ρė)V = (ṁin H)in − (ṁout H)out (2.15)
in out

As shown above, both resistive and capacitive components are employed together to
solve the conservation equations that govern the problem. The resistive components are
designed to solve the flow quantities (ṁ,ṁH), that are characterized by the flow area,
whereas both the velocity and the state properties (h,T ,ρ,p), are calculated by the capac-
itive components, characterized by its volume.

15
 CHAPTER 2. The Numerical Tool: EcosimPro

This document focuses on the thermodynamic modeling of air-breathing engines. Three

different engines have been calculated; two of them in EcosimPro, the turbojet and the
ramjet engine; and the Air Turbo-Rocket engine, ATR, which have been modeled in Mat-
lab. The studies that have been made in the numerical software EcosimPro are exposed
in Chapter 4.

16
 3
The Air Turbo Rocket engine

3.1 Engine background

As commented previously, this document is totally focused in the modeling and study of
the ATR engine. Designing a high-speed propulsion plant requires the compromise be-
tween the size of the engine, strictly linked with its weight, and the efficient utilization of
fuel. As pointed below, the TBCCs have a very interesting feature that makes them good
candidates to take part in the designing of the engine: their high specific impulse. The
ATR engine is a particular case in which the turbine-based cycle is in charge of the acceler-
ation stage of the trajectory, until a determined speed is reached, and the ramjet/scramjet
module is able to be ignited. Thanks to this configuration, which combines components
and characteristics from turbojets and rocket motors, the set of performances this engine
offers is unique; high specific thrust and thrust-to-weight ratio over a wide operational
range, talking about speed and altitude. These are the reasons why TBCCs constitute an
excellent solution for these types of missions.

Figure 3.1: The ATR-EXP cycle

The Figure 3.1 shows the schematic representing the ATR-EXP cycle. There, it can
be shown the flow path that both the air and the fuel (hydrogen) follow. First of all,
the air is ram compressed in the intake, which is designed to work at a wide range of
operation. From this station, it goes to the fan, that is driven by a generatively-heated
fuel-powered turbine. Afterwards, the air advances to the preburner, in which is mixed
and burned with the fuel. On its behalf, the fuel circuit is quite different. After going
out of the tank, a pump leads the hydrogen to the heat exchangers, which constitute the
regenerative cooling circuit for the burner and the nozzle. The heat pick-up is used to heat
the fuel up, and then it goes through the turbine to the preburner, where it is mixed with
the air, and straightaway burned. Then, the combustion gases enter in the final station of
the circuit, the nozzle, to produce thrust and go out from the engine.

17
 CHAPTER 3. The Air Turbo Rocket engine

The ATR cycle presents some characteristics that make the engine one of the most
powerful models for missions with wide ranges of operation: the shaft power delivered to
the fan is independent from both the fan performance and the flight regime. Conversely,
this feature is unsuitable talking about the engine stability if it is compared with the
turbojet, since in this last cycle the fan and the turbine are coupled, so the surge or stall
phenomenon is prevented by the own power delivered by the turbine [30]. As well, the
altitude the ATR can reach is higher due to the higher combustion chamber pressure ob-
tained, that is produced due to the combination of the fan work and the ram compression
in the inlet. In addition, the turbine inlet temperature is no longer a limiting parameter
which restricts the fuel injection and the flight speed, so the fan inlet temperature can be
reached also when the compression in the intake reaches very high total temperatures due
to the high velocity of the aircraft, and then this operational speed is increased by large
compared with the turbojet [31]. Currently, the ATREX engine investigation is totally
focused on the application to the Mach 8 cruise aircraft MR2 [6].

Figure 3.2: MR2 aircraft

The ATR developing had its first steps in the decade of 1960, when some patents started
to come out [31]. Apart from the works of Luidens and Weber [13], in 1992 Lardellier and
Thetiot [32] patented an ATR-GG combined to a ramjet. Later, in 1995, Bussi et al. [14]
calculated the operating lines over the characteristic map of a two counter-rotating stages
fan with variable pitch [33] at various flight conditions. Four years later, Christensen [31]
made a research about the design and performance of an ATR working with a solid fuel
gas generator. The turbomachinery employed for that study was specifically designed, and
so were the maps.

Currently, the majority of the existing literature makes reference to the air turbo-
rocket engine with gas generator. In terms of the expander type, ATREX is a pre-cooled
ATR created in the 1980s, thought as a fly-back booster to accelerate a reusable TSTO
space plane to Mach 6. For that mission, liquid hydrogen was employed to cool down the
air stream that circulated through the inlet, and then the burner walls. Originally, the
model was thought to have a configuration in which the turbine was located very close
to the fan in order to reduce weight and size of the engine. In fact, carbon composites
were utilized as well for that purpose. Later on, a workability study was carried out for a

18
SECTION 3.2. Thermodynamic cycle

modified ATREX with a conventional aft-turbine [34]. A more recent study was focused
on a subscaled precooled turbojet, just designed for ground and flight tests, in order to
study the work and efficiency of the precooling circuit [35].

At this moment the ATR engine is thought to be used in the MR2 civil transport air-
craft, in which it would act as the booster to accelerate the airplane up to Mach 4-4.5 for
having good conditions for igniting the DMR engines. Namely, this transition is carried
out at Mach 4.5 and 24 km of altitude. The propulsion unit consists on one ATREX and
two DMR engines, installated in parallel in the airframe shown in the Figure 3.3.

Figure 3.3: MR2, the civil transport aircraft. The numbers make reference to the stations:
1 low speed intake, 2 high speed intake, 3 nozzle, 4 ATR duct, 5 DMR duct. [2]

The air turbo-rocket engine has some significant parameters that, once identified, are
the starting point to study the engine behavior in every point of its trajectory and working
regime. As well, it is possible to carry out a parametric study and establish which are the
optimum points where the turbine work is minimum and thus the engine size and weight,
etc. Then, the following section is focused on identifying, analytically, which are the main
parameters that will govern the system in the whole analysis.

3.2 Thermodynamic cycle

For the analytic study that is going to be carried out next, some assumptions have been
taken in order to simplify the calculus and the formulation.

• The intake is assumed as ideal, so the TPR is equal to the unity.

• Both the air and the fuel are considered as perfect gases.

• The flow is analyzed in zero-dimensional conditions.

• Every component but the heat exchangers and the combustion chambers are assumed
adiabatic.

• Steady regime is considered along the whole engine.

• The nozzle is adapted in all the trajectory.

19
 CHAPTER 3. The Air Turbo Rocket engine

• The liquid hydrogen that pass across the pump is considered as an incompressible
flow.

To start the analysis, the specific work that is taken place in the turbine, pump and
fan are expressed as

T 0 Cpa γa −1 
∆Hf = ∞ πf γa − 1 (3.1)
ηf

(πp − 1) p03
∆Hp = (3.2)
ηp ρ03
 γh −1
 
1

γh
∆Ht = T50 Cph ηT 1 −  (3.3)
πt

where the subindexes f , p and t are referred to the fan, pump and turbine stations,
respectively, and a and h to the air and the hydrogen flows.

Going on with the development, and taking into account that both the fan and the tur-
bine discharge at the combustion chamber, it is possible to express the turbine expansion
ratio, πt , as a function of the pressure loss, δp , that takes place in the heat exchangers, the
fan and pump compression ratios and the relation between fuel and the ambient pressures,

p04 − p05
δp = (3.4)
p04

πp p03
πt = (1 − δp ) (3.5)
πf p0∞
So, the Turbine Inlet Temperature, TIT, can be expressed as
!
1 p0 q
T50 = T30 + − 1 (πp − 1) 3 0 + MR (3.6)
ηp Ch ρ3 Cph
Notice that the second term of the expression at the right hand side makes reference
to the temperature which is reached across the pump, and the third one is the increase
of temperature that is produced when the flow goes through the heat exchangers. In
addition, the reader can appreciate how the mixture ratio, M R, is influencing this heat
addition, and this parameter can be written as
ṁ31
MR = (3.7)
ṁ3
If an energy balance across the shaft is applied, a power equilibrium is established
between the turbomachinery of the engine.

M R ∆Hf + ∆Hp = ∆HT (3.8)

The Equation 3.8 indicates that the turbine is driving both the fan and the pump. As
it can be appreciated in the Figure 3.1, there are two turbines in the cycle, one driving the
fan and one for the pump. Nevertheless, a single turbine is being consider. This is just for

20
SECTION 3.2. Thermodynamic cycle

analytic and modeling purposes, as it will be appreciated as well in the model developed
in EcosimPro in the Chapter 5; one turbine can be considered to be responsible for the
supply of the power which two of them would do in the reality.

Both the cycle behavior and performance, determined by the system composed by
the equations 3.1 to 3.8, are modeled as a function of four parameters: the fan pressure
ratio, the mixture ratio, the heat addition in the heat exchangers and the pressure loss
throughout them.
!
q
Y = f πf , M R, , δp (3.9)
Cph T30
As well, the following demonstration will prove that the performance of the engine
is independent of the heat addition to the hydrogen flow, q, and the pressure drop that
takes place in the heat exchangers, δp . It will be exposed that the performance parameters
(specific thrust, specific impulse) and independent from these parameters, and thus from
the engine size (nozzle area), and dependent just of the fan pressure ratio and the mixture
ratio (for a given flight condition).

To start with, it is necessary to obtain the fan exhaust total temperature and total
pressure,
γ−1
 
0 0  π γ − 1
T31 = T21 1+ (3.10)
ηcc

Pc0 = P21
0
πf (3.11)
where P210 is the total pressure at the inlet of the fan that, once gone through the

intake, is the ambient total pressure multiplied by the inlet total pressure recovery,

0
P21 = P00 T P R (3.12)
Now, it is possible to obtain the chamber total temperature that is reached when the
combustion takes place,
0 +η L
M R Cp T31 cc
Tc0 = (3.13)
(1 + M R)Cp
The nozzle is always adapted, so it is imposed that the nozzle area ratio is changing
constantly while the trajectory is being followed, satisfying the design condition in which
P9 = Pamb . As a result, the exhaust velocity of the engine can be expressed as
v
u
u 2γRγ T 0  γ−1 !
pamb

γ
c
Vn = t 1− (3.14)
γ−1 poc

So Vn is function of poc and mixture ratio. The net thrust is calculated by Eq. 3.15,
which expresses that, for a given flight condition and fuel and air massflow, it is function
of the nozzle exhaust velocity and nozzle efficiency.

21
 CHAPTER 3. The Air Turbo Rocket engine

1
  
TN = ma 1+ Vn ηn − V0 (3.15)
MR
The specific thrust, then, can be obtained from

1
 
Tsp = 1+ Vn ηn − V0 (3.16)
MR
Efficiencies, specific impulse and specific fuel consumption are obtained from the fol-
lowing expressions.
h  i
1
TN V0 MR 1+ MR Vn ηn − V0 V0 M RTsp V0
ηo = = = (3.17)
ṁH2 Hp Hp Hp
h  i
1
2TN V0 2 1+ MR Vn ηn − V0 V0
ηp = = (3.18)
ṁth Vn2 − ṁa V02
 
1
1+ MR Vn2 ηn − V02
h  i
1
ṁth Vn2 − ṁa V02 MR 1+ MR Vn2 ηn − V02
ηth = = (3.19)
2ṁH2 Hp 2Hp

ηo = ηth ηp (3.20)

TN ηo Hp Tsp M R
Isp = = = (3.21)
ṁH2 g gV0 g

ṁH2 V0
T SF C = = (3.22)
TN ηo Hp
As a conclusion, it can be state that the performance parameters can be calculated
independently of the size, and then re-scaled (Ath ) to satisfy the necessity of the mission.

3.3 Operational functioning for a given mission

As stated in section 3.1, the MR2 follows a trajectory clearly defined, in which not only
the flight condition (altitude and flight velocity) is already planned, but also the installed
thrust, F, that is required to make the mission possible [2]. Figures 3.4 and 3.5 show both
the mission trajectory and the installed thrust.

As well, Figures 3.6 and 3.7 show the intake total pressure recovery along the flight
Mach number, and the air massflow ratio between the total massflow that enters in the
engine and the one that goes through the ATR engine, calculated by [36]. It is necessary
to mention that the engine is composed by two modules, ATR and DMR ones, and both
of them share the first intake. Then, there is a control surface that controls whether the
ATR intake is opened or not, as well as the DMR one, and logically it depends of the
flight condition. The section designed for the ATR is called LSI, Low Speed Intake, and
the one for the DMR, HSI, High Speed Intake. Both capture ratios are calculated as

22
SECTION 3.3. Operational functioning for a given mission

αcLSI = m˙21 /ṁ0

αcHSI = m˙20 /ṁ0

where ṁ21 , ṁ20 and ṁ0 are the air massflows going into the ATR module, DMR
module and main intake for both ones, respectively.

Figure 3.4: MR2 trajectory

Figure 3.5: MR2 installed thrust

For a mission in which the installed thrust is given, and taking into account that the
specific thrust is already obtained, the air massflow needed for the intake is calculated in
Eq. 3.23,

23
 CHAPTER 3. The Air Turbo Rocket engine

Figure 3.6: MR2 intake TPR

Figure 3.7: MR2 Low Speed Intake ratio

F
Tsp = (3.23)
ṁ21
For the supersonic regime, and depending on the working point, there is a limit in
which the air massflow that goes through the LSI is higher than ṁ21 . As a result, some
bleeding is necessary to respect the design condition. Consequently, a minimum Tsp is
defined,

F
Tsp > (3.24)
(ρν)∞ A0 αcLSI

where (ρν)∞ A0 αcLSI is the air massflow that enters in the ATR intake, ṁ21 . Taking
into account this restriction, a bleeding ratio is defined as,

24
SECTION 3.3. Operational functioning for a given mission

ṁbypass ṁbypass
B= LSI
= (3.25)
(ρν)∞ A0 αc ṁ21
So the bleeding ratio, which is as well a quantification of the excess of thrust if no
bleeding is applied, can be expressed also as a function of the specific thrust,
min
Tsp
F
B =1− =1− (3.26)
(ρν)∞ A0 αcLSI Tsp Tsp
min corresponds to the limit value established by Eq. 3.24.
where the Tsp

The minimum nozzle area is calculated differently for both the subsonic and the super-
sonic regime. For both regimes, the nozzle area is function of Vn , MR and the conditions
in the combustion chamber, and in the supersonic one, there is also one more parameter;
the bleeding ratio. The area in subsonic conditions is obtained from Eq. 3.27, and the
one for supersonic conditions, from Eq. 3.28.

1 Rc T9
 
A9 = ṁ21 1 + (3.27)
MR P9 V n
1 Rc T9
 
A9 = (ρν)∞ A0 αcLSI (1 − B) 1 + (3.28)
MR P9 V n
being P9 the pressure at the exhaust of the nozzle, equal to Pamb , due to the nozzle is
considered as adapted.
 γ−1 !!
P9

γ
T9 = Tc0 1 − ηn 1 − (3.29)
Pc0
For the throat area calculation, the following procedure is followed:
γ
1 γ−1
 
γ−1
Pcrit = Pc0 1− (3.30)
ηn γ − 1
γ
γ−1 2
 
γ−1
0
Pth = Pcrit 1+ Mth (3.31)
2
Vn
M9 = √ (3.32)
γ R T9
γ
γ−1 2
 
γ−1
P90 = P9 1+ M9 (3.33)
2
γ+1
A9 P0 1 2 γ−1 2
  
2(γ−1)
= th0 1+ M9 (3.34)
Ath P 9 M9 γ+1 2
Finally, the turbine power is calculated from the specific turbine work, which is inde-
pendent from the engine size, Eq. 3.3.
(
∆Ht ṁ21 /M R M a∞ < 1
Ẇt = LSI (3.35)
∆Ht (ρν)∞ A0 αc (1 − B)/M R M a∞ > 1

25
 CHAPTER 3. The Air Turbo Rocket engine

Once shown the whole procedure to compute and calculate the cycle that models the
ATR-EXP behavior, several studies have been carried out, and results are going to be
proposed in the Chapter 5. The studies that have been made are:

• Influence study of each meaningful parameter over the cycle.

• Operational range of the engine, focusing on the turbomachinery cycle, varying

its meaningful parameters to study the work made by the turbine and the ones
consumed by the pump and the fan.

• Performance (Tsp and Isp) of the calculated operational range.

• Minimum throat area needed for the correct functioning of the engine versus the fan
pressure ratio, and showing it in function of MR, bleed and turbine work.

• Turbine work versus fan pressure ratio.

26
 4
Engine modeling in EcosimPro

4.1 Turbojet engine

In this section, an easy example of using EcosimPro is going to be carried out, in order to
show the way to proceed in this software. First, a schematic of the engine is going to be
introduced in the Figure 4.1, as well as the stations numbering.

Figure 4.1: Schematic of the turbojet engine

Then, the model has to be created by choosing the components and connecting them.
For that purpose, the library TURBOJET is going to be used. There, all the components
needed for this type of engine can be found. So, it is necessary to select and drag the
components to the workspace and connect them by means of their ports.

Figure 4.2: Turbojet engine model in EcosimPro

Figure 4.2 shows that the engine is composed by eleven types of components. First, the
intake receives the air and decelerates it to discharge the flow to the compressor. Then,
there are two splits that are modeled as the cooling circuit that is going to be discharged
in the turbine and in the nozzle. If the air is going in its natural path, it will be driven to

27
 CHAPTER 4. Engine modeling in EcosimPro

the combustion chamber, where will be mixed with the fuel. Notice that there is a valve
controlling the fuel flow that enters in the burner, by means of a law that is coming from
a monitor located in the upper part of the workspace. After the burner, the turbine and
nozzle stations take place, and there the model is closed and finished.
The points that are next to the components represent the ports, and there is the start or
the end of the line that connects one component with the following one. Notice that there
are different lines depending on its color, which represents the type of port they are linked
to. The blue ones are referred to the fluid type, and it is understood as the flow circula-
tion throughout the components. The brown one represents the mechanical type. In this
model, it is representing the shaft that connects the compressor with the turbine. On its
behalf, the red line is representing the flow circulation. In general, this is not happening
in other libraries, but only in the TURBOJET one. In the others, all flows are represented
by fluid type ports, but in this library it seems that the air and the fuel are differentiated.
The orange one corresponds to the control library, which is in charge of establishing all
the control laws that governs the engine behavior. With this laws, the user could choose
to change, for example, the beta parameter of the compressor in function of the entering
mass flow, or the nozzle area in function of the fuel massflow added to the burner. Finally,
the other brown line, which is going from the inlet to the nozzle is a nozzle type, which
is for monitoring the thrust generated by this last component. Naturally, there are more
types of ports, but their explanation exceeds the goals of this document if they are not
included in the model utilized here.

Going on with the turbojet modeling procedure, the next step is to create the partition;
the mathematical model defining the tearing variables, the initial and boundary conditions,
the unknown data and the algebraic variables. This is one of the most critical steps when
modeling a system with EcosimPro, due to the fact that the user cannot define a system
he wants if the variables chosen are not correct, that is, that do not close the system
of equations. This procedure can be done by default by EcosimPro, which will choose
the mathematical model and create the partition automatically, or by the user, for whose
procedure some wizards will be used. The Figure 4.3 shows one of the wizards used to
create the partition. Namely, it is the one focused on choosing the boundary conditions,
which is the critical step. The followed procedure to create the partition is the following
one:

• Select the unknown variables of the problem. If this step is skipped out, the math-
ematical model could assume a value for variables that, in other way, should be
unknowns. In the first phase of designing, there are variables that are totally un-
knowns and need to be defined for subsequent calculus, such as the geometry (pipe
and valve areas, length of the ducts, etc.), the nominal speed of the moving parts of
the engine, the initial mass flow or the pressure rise of pumps.
• Remove dynamic variables from the equation system by deriving some equations, if
it is possible. This is due to that some dynamic variables depend on other dynamic
variables and thus cannot be integrated independently, so it is logical to think that
this step is not always needed.
• Establish the boundary conditions. This step is not only deciding which are going to
be the external conditions of the engine (temperature, pressure, velocity...) but also

28
SECTION 4.1. Turbojet engine

Figure 4.3: Boundary conditions wizard to create a partition in EcosimPro

the variables that the user do know. For instance, in the turbojet model, studying
its stationary behavior, the user could specify the compressor pressure ratio and the
efficiency of all the components as known variables.

• The last step is the algebraic wizard. When a non-linear algebraic equation system is
detected, the wizard assists the user to select the minimum set of algebraic variables
to solve the non algebraic box. EcosimPro is going to suggest a set of algebraic
variables, but of course is the user the one who chooses the final set. Nevertheless,
the software checks that selection and validates it or generates an error, asking a
change of variables to close the system correctly (physically speaking).

Finally, the partition is created, so does the mathematical model. In this stage of
the process, the experiment is going to be created, which constitutes the final step before
running the calculus and analyze results. As commented before, the experiment is the
tool needed to establish the numerical values to the algebraic variables, the boundary
conditions and other variables whose default value could be modified. For doing all these
tasks, the experiment is divided in several blocks, where the code is organized by categories
in function of their function in the code:

• The DECLS block is used to store and declare all the variables utilized in the exper-
iment. If the user is going to calculate a mass flow and name it as ṁ, then he has
to declare it in this block, specifying its type, which in this case will be REAL. As
described above, there are lots of types, such as TABLE1D, STRING, INTEGER,
BOOLEAN, etc.

29
 CHAPTER 4. Engine modeling in EcosimPro

• The OBJECT block is employed to use instances of C++ classes.

• The INIT block is designed for initialize dynamic and algebraic variables whose
derivatives appear in the formulation, as well as the algebraic unknowns. However,
this block is only needed for the variables that are not previously initialized, not
for all of them. An important feature of the system is that the dynamic variables
are fixed in the problem, but not the algebraic unknowns. The model could contain
implicit equations, so the number of algebraic variables that are treated in this
block has to be equal to the number of implicit equations. As well, it is important to
establish a coherent numeric value for the variables, due to the fact that the closer
the number is to the solution, the less time the simulation takes. Moreover, if the
values are so far from the final ones, results could converge into an unreal solution.

• The BOUNDS block is the one which sets the expressions for the boundaries of the
model. They can be not only numerical values, but also expressions depending on
time. During the experiment, the user can modify any boundary variable, and the
software will detect by its own when it is needed to reinitialize and modify these
variables.

• The last block is called BODY. The difference with other blocks is that this is
used for several tasks, not just one, like integrate models, calculate steady states,
generate reports, etc. In this unit, the integration time is selected, the integration
step, the absolute and relative error, the solver used (Runge-Kutta, DASSL adn
DASSL-SPARSE), and hundreds of other tasks.

Following the turbojet example, and providing the engine with correct input data,
which some of them are indicated in the Table 4.1, it is possible to calculate the cycle and
make an study of its behavior. First, the cycle will be analyzed choosing a single point,
and comparing its data with other software, GasTurb. Results are shown in Table 4.2

Parameter Value Units

Flight altitude 4000 M
M0 1.5 -
ṁa 70 kg s−1
Burner Air Loading 2.5 kg s−1 m3
πcc 0.99 -
πc 12 -
Compressor β parameter 0.7 -
ηT 0.9 -
TIT 1300 K

Table 4.1: Input data for the turbojet calculus

Once the design point is calculated, it is time to study the influence of one or several
parameters in the engine performance. As this is being done just for illustrative purposes,
no conclusions related to the thermodynamic cycle are going to be obtained from these
results.

30
SECTION 4.1. Turbojet engine

EcosimPro GasTurb
Parameter Units
value value
T2 380.1 379.9 K
P2 2.24 2.26 bar
T3 798.4 784.9 K
P3 26.9 27.2 bar
T4 1300 1300 K
P4 26.6 26.9 bar
T5 936.1 947.7 K
P5 5.9 6.1 bar
Nozzle outlet Ts 800.6 784.9 K
Net Thrust 29.02 29.64 kN
ṁf 1.02 1.02 kg s−1
FAR 0.0146 0.0146 -
TSFC 35.18 34.43 g kN −1 s−1
Aexit 0.093 0.090 m2
ηo 0.4 0.37 -

Table 4.2: Design point calculated in EcosimPro and GasTurb

For the current study, a single parameter has been selected to be changing in the cy-
cle; the Turbine Inlet Temperature, TIT. As mentioned before, there is no interest in the
thermodynamic cycle of the engine, so just two parameters are going to be shown, in order
to compare and validate the model of both programs. These parameters will be the nozzle
throat area and the gross thrust. First, gross thrust will be showed for both programs,
and then the nozzle area.

31
 CHAPTER 4. Engine modeling in EcosimPro

Figure 4.4: Thrust vs TIT for a turbojet engine, modeled in EcosimPro

Figure 4.5: Thrust vs TIT for a turbojet engine, modeled in GasTurb

32
SECTION 4.1. Turbojet engine

Figure 4.6: Nozzle area vs TIT for a turbojet engine, modeled in EcosimPro

Figure 4.7: Nozzle area vs TIT for a turbojet engine, modeled in GasTurb

The reader can appreciate that results extracted from both programs are quite simi-
lar. The discrepancies shown in Table 4.2 and in the exposed Figures are related to the
different models GasTurb and EcosimPro use for calculating the cycle. EcosimPro is a

33
 CHAPTER 4. Engine modeling in EcosimPro

more complex software that takes into account processes that are not present in GasTurb,
such as losses in pipes, in valves, more complete chemical interactions resolution, etc.
Nevertheless, both models present similar results, so the one that the user has done in
EcosimPro can be cataloged as validated.

4.2 Ramjet engine

In this section, a ramjet cycle model is going to be introduced. There will be two studies
which will be presented. The first one consists on a design point calculus, in which the
boundary conditions (flight altitude and Mach) will be imposed. The second one will
calculate all the operational points included in a huge range of altitude and Mach. In the
first stage of the problem, the schematic which contains the model has to be defined in
EcosimPro, which can be observed in Figure 4.8.

Figure 4.8: Ramjet schematic in EcosimPro

As Figure 4.8 shows, the ramjet engine is composed by a single component that models
the whole behavior of the engine. It actuates as a "black box" in which the user introduces
certain input parameters and obtain the appropriate results. If the user desires to know
the functioning of this black box, he has to consult the code directly, as it has been done in
this document to understand the functioning of valves, thrust calculation, pressure losses
in injectors, etc. Apart from the ramjet component, the schematic counts with a valve
that can regulate the fuel massflow entering in the engine. The orange point that appears
in the component is a control port, so a control law can be assigned to the valve to set the
amount of flow going through the valve, according to certain parameters such as altitude
or flight velocity.

4.2.1 Single point calculus

This study is focused in the calculus of a single point of the trajectory, in which some
input parameters are going to be introduced and then an analysis of the engine behavior
will be performed. Then, using the schematic presented in Figure 4.8, a partition and an
experiment is created, following the same procedure than the one indicated in Section 4.1.

34
SECTION 4.2. Ramjet engine

Parameter Value Units

Altitude 8000 m
Flight Mach 5 -
ηcc 0.98 -

Table 4.3: Initial parameters for a single point calculus in the ramjet engine

Table 4.3 shows some relevant parameters. However, some others must be indicated,
such as the combustor starter temperature, the pressure the H2 (the fuel used in this
calculus) tank will have, or the length of the ramjet, but there is no need to detail them
in the document because this study is centered in a qualitative analysis, not quantitative.
Moreover, this section will not cover a steady calculus of the ramjet cycle. Instead, an
experiment in which the fuel valve and the burner igniter are switched off and suddenly
connected on will be made. The signal related to the valve, employed to this purpose, is
shown in Figure 2.

Figure 4.9: Valve position vs time

As it has been modeled, the valve will let the fuel flow go into the core when the
simulation is at the instant TIME = 10 s. Nevertheless, the combustion will start when
TIME = 15 s, due to the fact that the igniter will be activated at this instant. This type
of inputs brings the possibility to study the engine behavior before and after the fuel is
added. Then, the results are presented in Figures 3 to 6. In addition, this technique helps
the numerical model gain some stability and velocity at the beginning of the calculus,
rather than initialize all the variables at the same time.

Once the boundary variables have been established, running the model is the next
step. Figures 4.10 to 4.13 show the results obtained in this study.

35
 CHAPTER 4. Engine modeling in EcosimPro

Figure 4.10: Temperatures at nodes 1, 12 and 30 in the ramjet combustion chamber

Figure 4.11: Air, fuel and total massflows

Figure 4.12: Mach number histogram inside the combustion chamber

As the results show, the influence of the fuel flow addition and the igniter activation
can be easily noticed looking at the previous figures. Figure 4.10 shows the temperature
that some nodes of the combustion chamber reach. This chamber is discretized in 30 nodes
to guarantee some accuracy. A prove of the total freedom that EcosimPro offers the user
to model thermodynamic processes such as combustion is the possibility to choose where

36
SECTION 4.2. Ramjet engine

Figure 4.13: Ramjet engine thrust

the fuel is going to be injected. In this example, the fuel is introduced in the chamber in
the nodes 20 to 24, 20 % each node, so as the results indicate, just the nodes above the
twentieth are going to increase its temperature. According to this, the node 30 is the only
one that increment its temperature in this graph. On its behalf, the massflow shown in
Figure 4.11 is increased in TIME = 10s, due to the valve opening.

As Figure 4.12 indicates, EcosimPro allows the user to plot histograms of variables
along the chamber discretized in nodes. This is very useful to examine how a variable is
evolving across both the time and the length. In this case, the Mach number is computed,
having its maximum at node 9, and reaching stability around the twentieth node.

Finally, thrust is shown in Figure 4.13. As it shows, there is a huge instability in

the beginning, reaching negative values due to the fact that there is no fuel injection nor
ignition. When this last phenomenon takes place, stability is reached and the software is
able to converge in a reasonable value, which is the final result of the experiment.

4.2.2 Parametric study

In a second phase of the ramjet analysis, a parametric study will be performed in order
to observe the influence of the flight conditions, as well as the operational range of the
engine. Therefore, a wide interval has been selected, both for altitude and flight Mach, as
it is shown in Table 4.4.

Parameter Interval Step Units

Altitude 2000 - 20.000 2000 m
Flight Mach 3-6 0.1 -

Table 4.4: Boundary conditions for the parametric study

Then, the results are introduced in the Figures 4.14, 4.15 and 4.16. Clearly, there is
a range of operation for each altitude and Mach. Regarding fuel-to-air ratio, FAR, and
temperature, there exist a value of this two variables for almost every operational point,
despite of the fact that there are many illogical values, such as extremely low burner ex-

37
 CHAPTER 4. Engine modeling in EcosimPro

haust temperatures or high fuel-to-air ratio values, so that is because there are some parts
of the lines that are “dashed”; because these illogical values are NaN values in the matrixes.

Analyzing the tendencies of the variables exposed, the FAR is increasing with the alti-
tude because the fuel massflow is determined by the difference of pressure between the fuel
tank and the combustion chamber. While the tank pressure is constant, the combustion
chamber one is decreasing with altitude, so there is an increasing pressure gradient, which
motivates the entering of larges amount of H2 fuel in the chamber. Nevertheless, the
Mach number makes the FAR decreases when it increases. That is because the amount of
air massflow that enters in the engine is higher, so the proportion makes the FAR lower.
Regarding the temperature in the combustion chamber, it is clear that the more fuel in-
jected, the more temperature is increased, reaching its maximum in the stoichiometric
value. That is why the temperature follows a similar tendency than the FAR. On its
behalf, the thrust is following similar tendencies than the other ones, establishing a clear
operational range for values of temperature and velocity. If the FAR could be maintained
constant, the thrust could maybe increase due to the increase of air massflow when the
Mach number is increased, and as well being increased (the thrust) when the altitude is
higher, due to the lower air density. Nevertheless, as the fuel-to-air ratio is decreasing and
so is the chamber temperature, so it motivates a decrease in the nozzle exhaust velocity,
and consequently in the net thrust the engine produces.

Figure 4.14: Combustion chamber temperature vs altitude and Mach number

38
SECTION 4.2. Ramjet engine

Figure 4.15: Fuel-to-air ratio vs altitude and Mach number

Figure 4.16: Ramjet engine thrust vs altitude and Mach number

39
 CHAPTER 4. Engine modeling in EcosimPro

4.3 High-speed propulsion engines modeling

In this document, the ATR engine has been studied analytically in Matlab, but it has been
developed as well in EcosimPro. The schematic of the engine has been created, and the
partition development and the configuration of experiments is proposed for future works.

Figure 4.17: Air Turbo-Rocket Expander engine schematic in EcosimPro

Figure 4.17 represents the schematic that is going to be used for the ATR cycle model-
ing. It follows the same structure than the Figure 3.1 There, it can be shown the flow path
that both the air and the fuel (hydrogen) follow. First of all, the air is ram compressed in
the intake, which is designed to work at a wide range of operation. From this station, it
goes to the fan, that is driven by a generatively-heated fuel-powered turbine. Afterwards,
the air advances to the preburner, in which is mixed and burned with the fuel. On its
behalf, the fuel circuit is quite different. After going out of the tank, a pump leads the
hydrogen to the heat exchangers, which constitute the regenerative cooling circuit for the
burner and the nozzle. The heat pick-up is used to heat the fuel up, and then it goes
through the turbine to the preburner, where it is mixed with the air, and straightaway
burned. Then, the combustion gases enter in the final station of the circuit, the nozzle, to
produce thrust and go out from the engine.

This schematic has some particularities that need to be commented. First, the bleeding
is modeled by a Tee-component that conducts part of the air out of the engine, and it is
necessary not to have an excess of air in the supersonic regime. Second, notice that there
is a Heater that models the heat exchange that the cooling jackets components, named
as HEX1 and HEX2, extract of both the PreBurner and the Nozzle to heat up the H2

40
SECTION 4.3. High-speed propulsion engines modeling

massflow. Finally, a single turbine is used for driving both the pump and the fan, just for
simplifying purposes in the modeling task.

The ATR schematic has been introduced in this document. As stated before, the
partition development and the configuration of experiments are proposed as future tasks,
in order to compare them with the analytic studies that are going to be exposed in the
next chapter.

41
42
 5
Studies and Results

This chapter is focused on showing the different studies that have been made on the ATR
engine, which have been introduced in the Chapter 3, and its discussion and analysis. It is
necessary to remark that these studies have been carried out analytically in Matlab. The
studies that have been made are:

• Influence study of each meaningful parameter over the cycle.

• Operational range of the engine, focusing on the turbomachinery cycle, varying

its meaningful parameters to study the work made by the turbine and the ones
consumed by the pump and the fan.

• Performance (Tsp and Isp) of the calculated operational range.

• Minimum throat area needed for the correct functioning of the engine versus the fan
pressure ratio, and showing it as a function of MR, bleed and turbine work.

• Turbine work versus fan pressure ratio. and mixture ratio.

5.1 Influence study

As there are four meaningful parameters of the ATR cycle, it is interesting to know the
sensitivity of each of these parameters in the general behavior of the engine. Consequently,
an influence study has been made: first of all, a reference case has been calculated, whose
meaningful parameters values are exposed in Table 5.1.

Variable Value Units

πf 2.5 [-]
MR 35 [-]
q̇ 500 [kJ/kg]
δp 0.1 [-]

Table 5.1: Input data for the meaningful parameters of the thermodinamic cycle

Then, four studies in which one parameter varied its value a 10% have been made.That
is the reason why Table 5.2 have five columns: the first one shows the results of the
reference case, and the following ones are referred to the four studies performed. There,
the specific works of the turbomachinery have been specified, as well as its variation
with the reference case. As well, this procedure has been followed for the performance
parameters.
As the Table 5.2 shows, the parameters that have more influence over the cycle are
the πf and the MR. Logically, the πf is the only parameter that affects to the fan specific
work, Eq. 3.1. As well, it has a strong influence in the turbomachinery performance,
so this is one of the most important meaningful parameters to study. On its behalf, the
mixture ratio has less importance in the turbomachinery work, but it is important in the
performance parameters. It can be shown that the πf and the MR are the only ones that

43
 CHAPTER 5. Studies and Results

Cases/Variables Reference ∆πf ∆M R ∆q ∆δp

∆Hf value [kJ] 166.2 144.8 166.2 166.2 166.2
∆Hf variation [%] - 12.87 0 0 0
∆Hp value [kJ] 814.7 331.8 765.2 465.7 835.1
∆Hp variation [%] - 59.3 6.1 42.83 2.5
∆Ht value [kJ] 6621.3 5389.6 7153.6 6272.4 6641.7
∆Ht variation [%] - 18.6 8.03 5.27 0.3
Tsp value [kJ] 1.57 1.54 1.48 1.57 1.57
Tsp variation [%] - 1.6 5.9 0 0
Isp value [kJ] 55.1 54.2 57 55.1 55.1
Tsp value [%] - 1.6 3.5 0 0

Table 5.2: Influence study over the turbomachinery cycle

affect to the specific thrust and specific impulse, as it was stated in Chapter 3.2. The heat
addition is a parameter that strongly affects to the turbine and the pump. As the heat
is added to the fuel, it only affects to its cycle, so that is the reason why this parameter
has no influence neither to the fan specific work nor to the performance parameters. Fi-
nally, the pressure drop on the heat exchangers is the last parameter to analyze. As it is
exposed in the table, just a 2.5 % of variation in the pump specific work takes place when
varying a 10% the value of the δp . So, it is correct to state that this variable has much
less importance over the thermodynamic cycle than the other ones.

As a result of this study, the following ones are going to be developed by varying the
fan compression ratio, the mixture ratio, and the heat addition. The pressure drop is not
going to be part of the parametric study due to its little influence over the thermodynamic
cycle.

5.2 Operational range of the engine

This study is focused on finding a combination of the meaningful parameters to get to
achieve the establishment of the operational line along the trajectory of the MR2 while
the ATR engine is working. As the trajectory may be discretized in thousands of points,
and this is an analytical and qualitative study in which the computational cost would be
extremely high to be performed in Matlab, just several points will be calculated, and they
are shown in Table 5.3.

1 2 3 4 5 6
M0 [-] 1.1 1.7 2.1 3.2 4 4.5
h [km] 13.9 17 18.9 22 24 25.6

Table 5.3: Trajectory points used for the operational range study

Then, the study consists on varying the value of the πf , the mixture ratio and the heat
extracted from the preburner and nozzle by the heat exchangers to add it to the fuel flow,

44
SECTION 5.2. Operational range of the engine

and compare it with the turbine expansion ratio, πt , that would be necessary to drive
both fan and pump under those conditions. As well, the pump and turbine specific work
divided by the mixture ratio vs the heat addition is going to be shown to expose the clear
influence of this parameter with the turbomachinery performance.

To start with, the study related with the first point, M0 = 1.1 and h = 13.9 km will
be exposed. The Figure 5.1 shows the turbine expansion ratio vs the heat addition, with
three different fan pressure ratios and a range of mixture ratio from 20 to 50.

Figure 5.1: Turbine expansion ratio vs heat addition to the working fluid, at M0 = 1.1

As it is exposed in Figure 5.1, the higher the πf , the more heat addition is necessary
to decrease the πt . It has to be taken into account that the purpose of adding heat to
the fuel is decreasing the turbine pressure ratio (downsize the turbine) so the engine size
is lower, and thus the weight. The mixture ratio does not influence very much the tur-
bine expansion ratio, and just in the range from 300 to 400 kJ its influence is noticeable:
when the fuel massflow is decreased (more mixture ratio) the turbine inlet temperature
increases, so it compensates the decrements of the fuel flow rate that is going through the
turbine, considering constant πt .

As it can be observed in this figure, there is a trade-off between the heat extracted by
the heat exchangers and the turbine expansion ratio: the attempt to develop a cycle with
low turbine expansion ratio is conditioned by the amount of heat that the heat exchangers
are able to extract. As stated before, it would be the optimum condition, in order to

45
 CHAPTER 5. Studies and Results

reduce size and weight. Nevertheless, doing a big turbine would make possible to operate
without such an efficient and powerful heat exchanger.

In the end, what it is intended is to increase both specific thrust and specific impulse,
so this study is focused on choosing the correct combination of parameters, but trying to
accomplish this condition. This analysis will be analyzed deeply in the Section 5.3.

Continuing with the results exposition, the specific works vs the heat addition to the
fuel is going to be exposed.

Figure 5.2: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 1.1 and πf = 2.4

Paying attention to Figures 5.2, 5.3 and 5.4, it can be observed that both the pump
and turbine specific works are decreasing with the increase of heat addition. That is pre-
dictable, taking into account the turbine expansion ratio in the Figure 5.1, and the Eqs.
3.2 and 3.3, where it is shown that the pump work is proportional to the turbine one.
As well, it is noticeable that the pump power is two orders of magnitude lower than the
turbine one, so this is the cause why the engine performance is insensitive to the pressure
loss in the heat exchangers and the pump.

Talking about the engine behavior when the πf is increased, it is shown that the specific
work on the turbine and the pump has to be higher. On its behalf, the mixture ratio is
influencing the works in a narrow range of heat addition, just similar to the Figure 5.1.

46
SECTION 5.2. Operational range of the engine

Figure 5.3: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 1.1 and πf = 3.1

Figure 5.4: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 1.1 and πf = 3.4

Now, results from the point 2, M0 = 1.7 and h = 17 km are going to be shown.

47
 CHAPTER 5. Studies and Results

Figure 5.5: Turbine expansion ratio vs heat addition to the working fluid, at M0 = 1.7

Figure 5.6: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 1.7 and πf = 2

Showing the results from the point 2, M0 = 1.7 and h = 17 km, it is clear that the
tendency is quite similar, but with some particularities. Comparing Figures 5.1 and 5.5,
it is clear that the more velocity the engine is flying at, the more influence has the fan
pressure ratio. It can be observed that there is a higher separation between curves in 5.5
than in the previous one. Then, it means that the more πf needed, the more πt has to
be overcome. On its behalf, the heat addition is very important to low down the turbine
expansion ratio, so at these velocities this parameter is even more important than before.

48
SECTION 5.2. Operational range of the engine

Separately, the mixture ratio has some influence at low heat addition, when the more fuel
flow in the mixture, the more heat the engine needs, in order to be able to make a good
combustion process.

Figure 5.7: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 1.7 and πf = 2.5

Figure 5.8: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 1.7 and πf = 3

Figures 5.6, 5.7 and 5.8 show both the pump and the turbine specific works divided by
the mixture ratio, vs the heat addition. It can be noticed that the engine power demand
is maximum between this regime and the previous one, M0 = 1.7. The maximum levels

49
 CHAPTER 5. Studies and Results

of specific power take place along these two points, so they establish the maximum power
that the turbine has to provide as a design constraint.

Now, results from the point 3, M0 = 2.1 and h = 18.9 km are going to be shown.

Figure 5.9: Turbine expansion ratio vs heat addition to the working fluid, at M0 = 2.1

Figure 5.9 shows the turbine expansion ratio the engine needs at this flight condition.
Following the tendency of the previous flight regimes, it can be observed that the fan com-
pression ratio affects considerably to the πt demanding. The curves are more separated
than the previous regimes, so operating with πf = 3 would be logical if there were a heat
exchanger capable of providing such amount of heat. If that is not the case, the turbine
expansion ratio would have to be extremely high to satisfy this operation, so it is better
off to operate at low fan compression ratios to decrease the turbine size and weight, and
have less heat demand from the heat exchangers.

Figures 5.10, 5.11 and 5.12 follow an interesting tendency that needs to be commented.
The power demand of the engine at this flight condition is close to the maximum that the
cycle is going to need along the whole trajectory. That means that this area is the critical
one, talking about design terms. Then, analyzing the specific works, it is noticeable that
the curves admit just a certain range of heat addition to operate with, so here it is clear
the importance of the heat addition. Not only a big turbine would be needed, but also
a heat exchanger that is able to provide the enough amount of heat to satisfy the cycle
demand. If there is no heat exchanger, there is no cycle, as it is observed in the ranges of
300 to 370 in the Figure 5.11. There are no values there because there exists an extremely
high value of turbine expansion ratio, impossible to achieve in the reality, but possibly

50
SECTION 5.2. Operational range of the engine

as a mathematical solution of the general system of equations. The mixture ratio is an

important parameter under this flight condition, in which it can be observed that there
is a big difference of heat needed for the same turbine expansion ratio when changing the
MR from 20 to 50. This is clear in Figure 5.12.

Figure 5.10: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 2.1 and πf = 2

Figure 5.11: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 2.1 and πf = 2.5

51
 CHAPTER 5. Studies and Results

Figure 5.12: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 2.1 and πf = 3

Thereupon, results of point 4, M0 = 3.2 and h = 22 km will be exposed.

Figure 5.13: Turbine expansion ratio vs heat addition to the working fluid, at M0 = 3.2

Figure 5.13 show clearly what it is already commented above. When the Mach number
is increasing, the curves representing different fan compression ratios are further between
them, establishing a clear range of fan operation in the cycle, reducing it to values close

52
SECTION 5.2. Operational range of the engine

to the unit. It has to be taken into account that the ATR cycle is designed to operate
up to Mach 4.5, where the DMR is ready to be used. At Mach = 3.2, the DMR is
starting to work, but with αcHSI very lows. Progressively, this value will be increasing,
and consequently αcLSI will be decreasing, so this transition will be complete at Mach 4.5,
when the πf is equal to 1.

Figure 5.14: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 3.2 and πf = 1.5

Figure 5.15: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 3.2 and πf = 2

53
 CHAPTER 5. Studies and Results

Figures 5.14 and 5.15 show the specific works for this flight regime. As stated before,
the fan pressure ratio is governing these regimes, in which the stable behavior is near to
the unity. The tendencies that have been constant along all the study can be observed in
Figure 5.14. Nevertheless, these tendencies are similar to the previous ones when decreas-
ing the πf . If it is not reduced and maintained constant, as shows Figure 5.15, a huge
mixture ratio is needed to satisfy a reasonable turbine specific work, while decreasing this
parameter conducts the cycle to provide a big amount of heat to the fuel flow. So, it
is more interesting to continue decreasing the πf until the ATR engine is not working
anymore (πf = 1).

Results from point 5, M0 = 4 and h = 24 km, are exposed below.

Figure 5.16: Turbine expansion ratio vs heat addition to the working fluid, at M0 = 4

Following the tendency of reducing the πf , it is clear that the lower value this param-
eter has, the lower turbine expansion ratio. It can be observed in Figure 5.16 that the
πt are very low compared with the previous ones, logically. Moreover, if the fan pressure
ratio would be 1.1, the πt would be very close to the unity, as well. Take into account
that the only purpose of the turbine is driving both the pump and the fan, and the pump
does not consume much work, so it is clear that the turbine will be stopping its task very
soon (when reaching the point M0 = 4.5). Then, the operational point for this condition
will be around πf = 1.25, because the curve representing this point is able to operate in
the whole range of heat addition, so there is no great dependency of the heat exchangers,
case that is actually happening in the point πf = 1.5.

Figures 5.17 and 5.18 show the influence of the heat addition on the turbine and
pump specific works, varying the mixture ratio. As commented before, the curve that

54
SECTION 5.2. Operational range of the engine

Figure 5.17: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 4 and πf = 1.25

corresponds to πf = 1.25 is similar to the previous regimes, and allows more freedom
when designing the ATR cycle. Nevertheless, increasing the fan pressure ratio would lead
to need more turbine expansion ratio and heat addition, so the tendency of decreasing the
πf when the flight velocity increases is perfectly justified. Moreover, this is the desirable
condition; the transition between the ATR and the DMR is being carried out progressively,
with no discontinuities, and end up switching off the ATR engine softly, with no abrupt
changes.

Figure 5.18: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 4 and πf = 1.5

55
 CHAPTER 5. Studies and Results

Finally, results from point 6, M0 = 4.5 and h = 25.6 km, are exposed below.

Figure 5.19: Turbine expansion ratio vs heat addition to the working fluid, at M0 = 4.5

Figure 5.19 shows the πt needed at M0 = 4.5 with a certain amount of heat extracted
by the heat exchangers and added to the flow. Following the tendency of the previous
cases, it can be observed that the fan pressure ratio that is admited in this part of the
trajectory is very low. In fact, this point corresponds to πf = 1, but some more cases are
modeled in case of need of certain compression to the air. As the figure show, the correct
value of fan compression ratio for this point would go from 1 to 1.2, in which the heat
addition is not very big, and the πt has a very low value. Nevertheless, it is important to
remark that the engine is thought to establish a πf = 1 at this point, so this is the final
working point of the ATR engine.

Figures 5.20 and 5.21 show the specific works versus the heat addition and the mixture
ratio. It is clear that the curves corresponding to πf = 1.4 are not a good working point,
due to the large disparity of results when modifying the mixture ratio. Nevertheless, if the
fan pressure ratio were not equal to the unity, the point of πf = 1.2 is a stable point to
work with. Tendecies are similar than in previous cases, but with a considerable reduction
of turbine specific work.

As a conclusion for this section, it can be stated that the turbine specific work is
maximum within values from M0 = 1.7 to M0 = 2.1, and then it starts to decrease, along
with the fan pressure ratio, which will be decreasing little by little until it becomes the
unity, in the point M0 = 4.5. The heat addition is interesting because this allows the
engine reduce its size and weight, due to the less turbine expansion ratio that is needed
to satisfy the cycle demand. As well, increasing the mixture ratio leads to need less πt ,
due to the reduce of fuel massflow increases the turbine inlet temperature and leads to a

56
SECTION 5.2. Operational range of the engine

better expansion.

Figure 5.20: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 4.5 and πf = 1.2

Figure 5.21: Turbine and pump specific work vs heat addition to the working fluid, at
M0 = 4.5 and πf = 1.4

The operational range study has come to the end, and now it is time to comment the
engine performance along the trajectory computed in this last study (Table 5.3).

57
 CHAPTER 5. Studies and Results

5.3 Engine performance

This section is focused on the study of the specific impulse and specific thrust, depend-
ing on the mixture ratio and the fan pressure ratio. Maximizing these two performance
parameters is what it is intended when designing the ATR engine. That is why its study
is very important for the developing of this engine, so a study along the trajectory of the
MR2 will be exposed, following the same points the Table 5.3 showed.

To start with, the influence of each parameter will be analyzed using the Figure 5.22.

Figure 5.22: Engine performance at M0 = 1.1

Figure 5.22 shows the performance parameters specific thrust versus specific impulse
with different mixture ratios and fan compression ratios. As it is shown in the image, both
performance parameters show a trade-off between maximizing either the specific thrust
or the specific impulse. The curve, at a constant πf , is decreasing the specific impulse
when increasing the specific thrust. As well, at low mixture ratios the specific thrust is
maximum, due to the fact that the less air massflow there is in the engine (less MR), with
constant thrust, the higher value for the specific thrust. At the same time, if the MR is
high, it means that the fuel massflow is low, so it provokes that, at constant thrust, the
specific impulse increases its value. The only way to make both parameters higher at the
same time is increasing the fan pressure ratio. The Figure 5.22 shows an increase in both
variables when passing from πf = 1 to πf = 3. Nevertheless, it is proven in the previous
chapter that the increase in the fan pressure ratio could lead to a very high heat addi-
tion need, maybe impossible to satisfy by the heat exchangers. Then, as stated before, a
trade-off between these two parameters and with the turbomachinery variables is needed
when designing the operational points of the engine.

Next, the same results but for the rest of the points of Table 5.3 are going to be shown.

58
SECTION 5.3. Engine performance

Figure 5.23: Engine performance at take off and M0 = 1.7

Figure 5.24: Engine performance at M0 = 2.1 and M0 = 3.2

59
 CHAPTER 5. Studies and Results

Figure 5.25: Engine performance at M0 = 4 and M0 = 4.5

Figures 5.23, 5.24 and 5.25 show, along with the Figure 5.22, that both specific thrust
and specific impulse present the same order of magnitude, even quite similar values, when
changing the flight condition. Although there is a minimum variation, every operational
point shows the same tendency and similar values. This is a very important characteristic
of the ATR engine; it is able to present more or less the same behavior over the whole
trajectory, maintaining the performance parameters almost constant, if the mixture ratio
and the fan pressure ratio are chosen correctly. It is necessary to remind that the perfor-
mance parameters are independent from the pressure drop and the heat addition by the
heat exchangers, but dependent just of mixture ratio and πf (translated to the chamber
total pressure). As a result, if these two parameters are maintained more or less constant,
depending of the flight regime, the performance will be quite constant.

5.4 Nozzle throat area

This study is focused on the determination of the minimum nozzle area the engine needs to
have, in order to satisfy the mission specifications. Until this moment, the cycle has been
established without taking into account the amount of massflow that is going to circulate
throughout the engine. By calculating the performance parameters, the engineer can have
a general idea of the needs of the engine, but then, sizing the engine and calculating the
net thrust, the air and fuel massflows and the nozzle area are the next step, taking into
account the mission requirements.

This section will be focused on the nozzle area in subsonic and supersonic regime as a
function of the mixture ratio and fan pressure ratio, and then it will be shown the bleeding

60
SECTION 5.4. Nozzle throat area

influence on the area selection. Finally, another graph will be shown as a function of the
turbine work.

First, the nozzle area as a function of fan pressure ratio and mixture ratio, for both
subsonic and supersonic regimes are shown in Figures 5.26 and 5.27.

Figure 5.26: Nozzle throat area vs πf and MR in subsonic regime

Figure 5.27: Nozzle throat area vs πf and MR in supersonic regime

As shown in Figures 5.26 and 5.27, the nozzle area increase its value along the trajec-
tory (higher flight Mach, higher altitude) because the ambient static pressure decreases.
As well, it is clear that the mixture ratio has an important role when the flight velocity
increases, being independent from the nozzle area at low velocities, such as take off. That

61
 CHAPTER 5. Studies and Results

is because the nozzle exhaust velocity increases with the mixture ratio, so the area needed
for that massflow is higher than at lower velocities. On its behalf, the fan pressure ratio
provokes the area to reduce its value because the higher pressure there is in the combustion
chamber, the higher the density is; for a constant massflow, the area has to be reduced to
make the process stationary.

As proven in Chapter 3.3, some bleeding is needed for the correct functioning of the
engine, due to the fact that the ATR produces more thrust than the required in the MR2
mission. As a result, a study has been performed to investigate how much bleeding is
necessary as a function of the fan pressure ratio, and the area needed for that mission.

Figure 5.28: Nozzle throat area vs πf and bleeding ratio

As Figure 5.28 shows, the bleeding is taking place in points where the mixture ratio
is low, as Figure 5.27 showed. This is because the specific thrust is maximum when the
mixture ratio is low, as it was exposed in Section 5.3. The amount of bleeding is a measure
of excess of thrust, and it can be written as

min
Tsp
B =1− (5.1)
Tsp
Then, when the specific thrust is adopting high values, more bleeding is necessary to
obtain the desired thrust for the mission.

Finally, the last study of this section is dedicated to study the turbine work that is
necessary to drive the turbomachinery, as a function of the fan compression ratio and the
nozzle area, both in regime subsonic and supersonic.

62
SECTION 5.4. Nozzle throat area

Figure 5.29: Nozzle throat area vs πf and turbine work in subsonic regime

Figure 5.30: Nozzle throat area vs πf and turbine work in supersonic regime

Figures 5.29 and 5.30 show the nozzle throat area versus the πf , as a function of the
turbine work, for the subsonic and supersonic regime, respectively.

63
 CHAPTER 5. Studies and Results

When the throat area and the compression ratio are increasing, the turbine load is
increasing as well. The power required by the turbomachinery is quite sensitive to the
throat area when the engine is flying at low Mach numbers. Then, when the M0 = 3
point is reached, the load is almost independent of the throat area, and the fan pressure
ratio is the dominant parameter, talking about the supersonic regime. Consequently, the
optimum configuration leads to the minimum turbine work, so these points will be the
ones chosen for the operational line, taking into account the heat addition that would be
possible to achieve by the heat exchangers.

5.5 Turbine work vs πf

Finally, the last study of this document focuses its interest in the turbine load, as a function
of fan pressure ratio and mixture ratio, both in subsonic and supersonic regimes.

Figure 5.31: Turbine power vs. fan pressure ratio and mixture ratio, at subsonic regime

Figures 5.31 and 5.32 show the turbine work vs. fan pressure ratio and mixture ratio.
As it is exposed in the images, the turbine load increases with the mixture ratio, due to
the fact that the air massflow through the fan is increasing, so this component needs more
work to drive all that flow. On its behalf, increasing the fan pressure ratio leads to an
augment of the power provided by the turbine. This is something logical, if the reader
takes into account that the more compression power the fan requires, the more power the
turbine has to produce to satisfy that power balance. Finally, it can be observed that the
maximum power demand corresponds to the areas close to M0 = 2, as it was predicted in
the last studies. When the velocity keeps increasing, the power demand is lower due to
the fan pressure ratio is lower as well.

64
SECTION 5.5. Turbine work vs πf

Figure 5.32: Turbine power vs. fan pressure ratio and mixture ratio, at supersonic regime

Five studies have been exposed in this document, and now, being conscious of the
behavior of the engine, the next step is to choose the operational points of the engine
along the whole trajectory. These operational points are chosen by a numerical algorithm
of optimization, made by a computer, but here it can be predicted which points they will
be. To start with, the maximum turbine power and the maximum nozzle area have to be
chosen. Then, it will impose a restriction in which many points that are exposed in these
figures will be out-of-range. Then, it has to be taken into account that it is preferable to
choose a mixture ratio close to the stoichiometric (M R = 34.33), in order to maximize
the heat transfer made to the walls and then to the heat exchangers. This is an important
restriction too: it has to be specified how much heat the heat exchangers can extract and
add to the fuel, due to this is a very important condition to minimize the turbine power.
Considering all these restrictions, the operational line of the engine has to be established,
but this is proposed for future works.

65
66
 6
Financial estimation

This chapter is dedicated to detail the economic estimation about the total budget of
the project. There are several fields that contribute to the final amount that have made
possible the realization of this work:

• The equipment used consists on the user computer, valuated in 2.000 e, and the
supercomputers that are present in the Von Karman Institute, which have a cost of
10.000 e.

• This project has been developed in the context of an internship, whose fellowship
has consisted on 350 e per 5 months, so the total sum rose to 1.750 e.

• The programs used for this project have been Matlab and EcosimPro. These are not
open software, they require a license for being able to work with them, so these are
also included in the budget. The license for EcosimPro is 10.000 e, and the Matlab
one, 800 e, according to their official webpage.

• Finally, the project has been supervised by a research engineer. The project has
taken 5 months, and the supervisor has dedicated 2 hours per week to the study,
so he has employed 40 hours in it. If the engineer is paid 80 e per hour, it can be
estimated that the derived cost of receiving help of such a qualified person is 3.200
e.

Table 6.1 sums up all the costs that contribute to the final sum.

Concept Quantity
Equipment 12.000 e
Fellowship 1.750 e
Licenses 10.800 e
Supervisor 3.200 e

Total 27.750 e

Table 6.1: Project’s financial estimation

67
68
 7
Conclusions and future work

7.1 Conclusions
The Air Turbo-Rocket Expander engine has been introduced in this document as a com-
bined cycle for high-speed propulsion, destined to commercial flights up to Mach 8 in the
aircraft designed by LAPCAT, MR2. This combined cycle is characterized by four mean-
ingful parameters: fan pressure ratio, mixture ratio, heat extracted and pressure loss in
the heat exchangers, while the turbomachinery efficiencies and the flight condition remain
constant. In addition, it has been proven that the performance parameters are only de-
termined by two variables: fan pressure ratio and mixture ratio.

Some studies related with supersonic engines such as turbojet and ramjet have been
exposed. These studies have been made in the numerical software EcosimPro, which is a
powerful tool to develop engine models by an object-orientated language, in which every
component has its own code and functions, and interact with other components by the
ports connection. In addition, this software is valid for steady, transient and optimization
calculations.

The Air Turbo-Rocket Expander engine has been modeled analytically and several
studies have been proposed in this document along the whole trajectory this engine has,
until it is switched off and the DMR operates at 100%. This trajectory goes from take off
to Mach 4.5, and there are several changes along this trajectory. First, the fan compres-
sion ratio increases up to 3, at M0 = 2, approximately, and then starts to lower its value
until it reaches the unity. As it is proportional to the turbine power, this component has
the same behavior. On another note, the supersonic regime needs a bypass duct not to
overcome the specific thrust the engine needs. There is a balance between the bleeding
and other meaningful parameters, such as mixture ratio or fan pressure ratio, in order to
choose the correct value in each point of the trajectory.

The ATR engine is designed within two main characteristics: maximize both specific
thrust and specific impulse. This goal represents a trade-off between both of the two pa-
rameters, and they depend solely on the fan pressure ratio and the mixture ratio. It has
been proven that increasing the πf both of them increase. However, increasing the fan
compression ratio would cause other problems, such as lack of turbine power to supply to
the fan or lack of heat addition to the fuel flow.

The operational points that will be set for the trajectory of the MR2 have to be
chosen taking into account the restrictions the engine presents. First of all, the mixture
ratio should be good enough to guarantee an efficient combustion in which a considerably
amount of heat is liberated, in order the heat exchangers to extract it and transfer it to the
fuel flow. According to this, the heat addition will determine the maximum turbine power
the cycle needs. One of the challenges of the ATR is to develop some good heat exchangers
to reduce the turbine size and weight and make the engine lighter, more compact, and
more efficient. In addition, two more restrictions are the maximum area nozzle and the
maximum turbine work possible for the engine, so these will determine the frontier between

69
 CHAPTER 7. Conclusions and future work

real possible operational working points and just mathematical solutions that are not able
to be implemented in real engines.

7.2 Future work

This proyect is so huge that many organizations are working in make the combined cycles
for high speed propulsion a realty. As a conclusion, there is so much work to do for the
future 20-30 years until this kind of engine becomes a reality. Nevertheless, it can be
established as future work:

• The selection of the operational line of the engine along the whole trajectory, and
its optimization.

• The developing of reliable models for the fan, the turbine, the pump, the heat
exchangers and the nozzle.

• The developing of the complete model in EcosimPro and its simulation.

• The specification of how the transition between the ATR engine and the DMR engine
is going to be performed.

70
71
72
 Bibliography

[1] Empresarios Agrupados, Astrium, Cenaero, Kopoos, and VKI. ESPSS - European
Space Propulsion System Simulation - EcosimPro Libraries User Manual. September
2012.

[2] Victor Fernandez Villace. Simulation, Design and Analysis of Air–Breathing Com-
bined–Cycle Engines for High Speed Propulsion.

[3] Timothy Barber, Brian Maicke, and Joseph Majdalani. Current State of High Speed
Propulsion: Gaps, Obstacles and Technological Challenges in Hypersonic Applica-
tions. American Institute of Aeronautics and Astronautics, August 2009.

[4] H. Sänger Kuczera Koelle, J. E. Space Transportation System Status, 1998.

[5] D. T Pratt Heiser, W. H. Hypersonic airbreathing propulsion. AIAA Education Series.

1994.

[6] J Steelant. LAPCAT: High-Speed Propulsion Technology.

[7] John Beyer, Julian Cooper, Gerald Holden, François Nectoux, Nancy Ramsey, David
Schorr, Tony Thompson, Andrew White, and Scilla McLean. North Atlantic Treaty
Organisation. In Scilla McLean, editor, High Speed Propulsion: Engine Design - Inte-
gration and Thermal Management, pages 204–233. Palgrave Macmillan UK, London,
1986.

[8] Victor Fernandez-Villace and Guillermo Paniagua. Numerical Model of a Variable-

Combined-Cycle Engine for Dual Subsonic and Supersonic Cruise. Energies, 6(2):839–
870, February 2013.

[9] F Jivraj, R Varvill, A Bond, and G Paniagua. The Scimitar Precooled Mach 5 Engine.

[10] History of Aviation, 2010.

[11] Nancy McNelis and Paul Bartolotta. Revolutionary Turbine Accelerator (RTA)
Demonstrator. In AIAA/CIRA 13th International Space Planes and Hypersonics
Systems and Technologies Conference. American Institute of Aeronautics and Astro-
nautics.

[12] Jinho Lee, Ralph Winslow, and Robert J Buehrle. The GE-NASA RTA Hyperburner
Design and Development. page 22, 2005.

[13] Roger W Luidens. An Analysis of Air-Turbo Rocket Engine Performance including ef-
fects of component changes. Technical Report RM E55H04a, Lewis Flight Propulsion
Laboratory, NACA, Cleveland, April 1956.

[14] Giuseppe Bussi, Guido Colasurdo, and Dario Pastrone. Analysis of air-turborocket
performance. Journal of Propulsion and Power, 11(5):950–954, September 1995.

[15] Nobuhiro Tanatusgu, Tetsuya Sato, Yoshihiro Naruo, Takeshi Kashiwagi, Tomoaki
Mizutani, Toshiyuki Monji, and Kenji Hamabe. Development study on ATREX en-
gine. Acta Astronautica, 40(2-8):165–170, January 1997.

73
 BIBLIOGRAPHY

[16] Preston Carter, Vladimir Balepin, Terry Spath, and Christopher Ossello. MIPCC
Technology Development. In 12th AIAA International Space Planes and Hypersonic
Systems and Technologies, International Space Planes and Hypersonic Systems and
Technologies Conferences. American Institute of Aeronautics and Astronautics, De-
cember 2003.

[17] R Freese Balepin et al, V. Rocket Augmentation for Combined Cycle Turboaccelerator
Jet Engine. December 2007.

[18] Dr Vladimir Balepin. High Speed Propulsion Cycles.

[19] Ed Envia. Engine Intake Aerothermal Design: Subsonic to High Speed Applications.
Lectures Series - Von Karman Institute, National Aeronautics and Space Adminis-
tration, 2011.

[20] W. A. Kuhrt. Turborocket Engine.

[21] G. T. Peters N. C. Newington and N. C. Rice. Regenerative Expander Engine, March

1966.

[22] N. Rackemann T. Fuhrmann and S. Krügen. EcosimPro ATR-EXP and ATR-GG

modeling and feasibility study for the ESTEC Mach 8 concept. Cologne - Germany,
2011.

[23] COSAP Sizing Tool Application for ATR Cycle. LAPCAT II Derivable D3.3.3, 2009.

[24] Preliminary Mach 8 Turbo-Based /DMR Vehicle Analysis: Conceptual Design of a

Dorsal-Type Vehicle: LAPCAT-MR2. LAPCAT II Derivable D3.1.4, 2009.

[25] Modelling Language.

[26] L. R. Petzold. DASSL: Differential Algebraic System Solver. Technical Report, Sandia
National Laboratories, Livermore, CA, 1983.

[27] Math. Algorithms & Simulation Guide.

[28] Alberto Jorrín. EcosimPro and its EL Object-Oriented Modeling Language. page 10.

[29] Romain Gérard. Waterhammer simulations in an EcosimPro/ESPSS environment.

PhD thesis, von Karman Institute for Fluid Dynamics, Brussels, Belgium, June 2013.

[30] M. E Thomas and J. A. Bossard. The influence of turbomachinery characteristics on

Air Turbo Rocket engine operation. Huntsville, ALA, 2000.

[31] Kirk Christensen. Air Turborocket/Vehicle Performance Comparison. Journal of

Propulsion and Power, 15(5):706–712, September 1999.

[32] A. M. J Lardellier and R. P. M. Thetiot. Combined Turborocket and Ramjet Propul-

sion Unit.

[33] M. Albers, D. Eckardt, P. A. Kramer, and N. H. Voss. Technology preparation for

hypersonic air–breathing combined cycle engines. American Institute of Aeronautics
and Astronautics, 1989.

74
BIBLIOGRAPHY

[34] Kousuke Isomura, Junsuke Omi, Nobuhiro Tanatsugu, Tetsuya Sato, and Hiroaki
Kobayashi. A feasibility study of a new ATREX engine system of aft-turbine config-
uration. Acta Astronautica, 51(1):153 – 160, 2002.

[35] Tetsuya Sato, Hideyuki Taguchi, Hiroaki Kobayashi, Takayuki Kojima, Katsuyoshi
Fukiba, Daisaku Masaki Okai, Keiichi, Kazuhisa Fujita, Motoyuki Hongo, and Shu-
jiro Sawai. Development study of a precooled turbojet engine. Acta Astronautica,
66(7):1169–1176, April 2010.

[36] C Meerts and J Steelant. Air Intake Design for the Acceleration Propulsion Unit of
the LAPCAT-MR2 Hypersonic Aircraft. page 18, 2013.

[37] I Rodríguez Miranda. Modeling, Analysis and Optimization of the Air Turbo Rocket
Expander Engine.

[38] Johan Steelant. Sustained Hypersonic Flight in Europe: First Technology Achieve-
ments within LAPCAT II. In 17th AIAA International Space Planes and Hypersonic
Systems and Technologies Conference. American Institute of Aeronautics and Astro-
nautics.

[39] Empresarios Agrupados, Cenaero, Kopoos, Astrium, and VKI. ESPSS - European
Space Propulsion System Simulation - Software Verification and Validation Plan.
2013.

[40] GasTurb GmbH . GasTurb 12. 2015.

[41] Victor Fernandez Villace, G Paniagua, and I. Rodriguez Miranda. Thermal evaluation
and cooling channels design - ATR cycle design. Theme 7: Transport. VKI, August
2013.

[42] Victor Fernandez Villace. Refined engine cycle analysis, results and preliminary di-
mensions - Integral performance of the MR2 propulsive plant in the Mach range from
1.5 to 4.5. Theme 7: Transport. VKI, August 2013.

75