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

net/publication/327409702

Multidisciplinary optimisation of bipropellant rocket engines using H2O2 as

oxidiser

Article  in  Aerospace Science and Technology · September 2018

DOI: 10.1016/j.ast.2018.08.036

CITATIONS READS

3 223

3 authors:

Adam Okninski Jan Kindracki

Institute of Aviation Warsaw University of Technology
44 PUBLICATIONS   139 CITATIONS    47 PUBLICATIONS   630 CITATIONS   

SEE PROFILE SEE PROFILE

Piotr Wolanski
Warsaw University of Technology
20 PUBLICATIONS   39 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

System of satellite mutual navigation for servicing and formation flying applications View project

ILR-33 “Bursztyn” [Amber] sounding rocket View project

All content following this page was uploaded by Adam Okninski on 08 November 2020.

The user has requested enhancement of the downloaded file.

 Aerospace Science and Technology 82–83 (2018) 284–293

Contents lists available at ScienceDirect

Aerospace Science and Technology

www.elsevier.com/locate/aescte

Multidisciplinary optimisation of bipropellant rocket engines using

H2 O2 as oxidiser
Adam Okninski a,∗ , Jan Kindracki b , Piotr Wolanski a
a
Center of Space Technologies, Institute of Aviation, Al. Krakowska 110/114, 02-256 Warsaw, Poland
b
Institute of Heat Engineering, Warsaw University of Technology, Nowowiejska 21/25, 00-665 Warsaw, Poland

a r t i c l e i n f o a b s t r a c t

Article history: Today’s rocket chemical propulsion systems for in-orbit use rely on toxic liquid rocket propellants. The
Received 28 August 2018 need for environmentally-friendly spacecraft and upper-stage high-performance engines can be seen.
Accepted 29 August 2018 This paper covers the topic of optimisation of Geostationary Transfer Orbit apogee bipropellant rocket
Available online 3 September 2018
engines, using highly concentrated hydrogen peroxide as oxidiser and kerosene as fuel. Performance,
structural and heat transfer analyses are described. In particular a detailed mass model for hydrogen
peroxide/kerosene rocket propulsion systems is discussed. Special care is given to bipropellant rocket
systems using staged combustion configurations where a catalyst bed is utilised. The optimisation process
was conducted using Matlab software. The ultimate goal of this work was the development of a tool
for hydrogen peroxide/hydrocarbon bipropellant rocket propulsion system optimisation in terms of given
requirements and design constraints. The presented software may enable the development of advanced,
environmentally-friendly satellites and highly-efficient architectures utilising storable green propellants,
including small upper-stage propulsion systems.
© 2018 Elsevier Masson SAS. All rights reserved.

1. Introduction jectors [3,4]. Some work on green satellite thrusters, with the aim
of introducing them to the market, is done at Moscow Institute
Green bipropellant rocket engines using hydrogen peroxide (HP) of Aviation. Swirl injectors are utilised and thrust chamber as-
and hydrocarbons (HC) for spacecraft and missiles were under de- semblies, with thrusts in the range of 200–500 N, are tested [5].
velopment in the 60’, however they did not find extensive use Storable green propulsion bipropellant rocket engines of thrust up
in-orbit due to a wide spread of hydrazine derivatives and dini- to 2500 N using 90% HP and kerosene are under development in
trogen tetroxide propellant technology. State of the art large scale South Korea [6,7]. Laboratory research has been also undertaken
HP/HC rocket engines included inter alia British Gamma propul- in Austria [8], Italy [9,10], China, Malaysia, Taiwan and Japan [1].
sion systems, which evolved from World War 2 German designs, These activities focus on component development, such as catalyst
as well as the RD-161P, powering the third stage of some Soyuz chambers. HP bipropellant engines using different fuels were in-
launch vehicle variants [1]. Since the 70’ of the last century low vestigated during FP7 GRASP (GReen Advanced Space Propulsion)
activity in the field was present. HP gained interest again in the project and recently in PULCHER (Pulsed Chemical Rocket with
90’. Small thrusters were laboratory tested by the General Kinetics Green High Performance Propellants), both funded by the European
Company [2]. They were very compact and had high design perfor- Commission. Recent activities of the European Space Agency prove
mance. However, there is little information in literature concerning its increased interest in HP as a propellant for mono- and bipropel-
test results. All of up-to-date non-laboratory engines used catalytic lant applications (i.e. project GRACE – Green bi-propellant apogee
HP decomposition before chamber injection. American develop- rocket engine for future spacecraft). Current trends show the need
ment programs dedicated to HP/HC rocket thruster research have for storable green propulsion thrusters, especially for Geostationary
been realised at University of Purdue. Mainly pseudo-hypergolic Transfer Orbit applications, where thrusts of about 500 N are re-
ignition was studied and experimentation concerning combustion quired [11]. Therefore, an apogee-class bipropellant rocket engine
stability was done. The designs used pintle and transverse fuel in- development was initiated at the Institute of Aviation in Warsaw,
Poland [12–14]. With increased interest in sounding rockets and
launchers [15–18] propulsion technology is a development prior-
* Corresponding author. ity. Work on catalyst bed technology is conducted at the Insti-
E-mail address: adam.okninski@ilot.edu.pl (A. Okninski). tute of Aviation within European and national programs [19,20]

https://doi.org/10.1016/j.ast.2018.08.036
1270-9638/© 2018 Elsevier Masson SAS. All rights reserved.
 A. Okninski et al. / Aerospace Science and Technology 82–83 (2018) 284–293 285

Abbreviations

CR Combustion Chamber Contraction Ratio LRE Liquid Rocket Engine

GEO Geostationary Earth Orbit MDO Multidisciplinary Optimisation
GTO Geostationary Transfer Orbit
MMH Monomethylhydrazine
HC Hydrocarbon
HP Hydrogen Peroxide MON Mixed Oxides of Nitrogen
LAE Liquid Apogee Engine TCA Thrust Chamber Assembly

and small-scale research on thermal decomposition of HP is done activities using this program have been done, mainly for satellite
at Warsaw University of Technology [21]. The introduction of a propulsion systems [40,41]. Airbus Space & Defence uses ROCFLAM
bipropellant thruster using HP as oxidiser and an environmentally (ROcket Combustion Flow Analysis Module) for simulating storable
friendly fuel will enable gradual withdrawal of toxic systems from toxic rocket engines utilising film cooling. Combustion is modelled
the market. HP/HC combinations have high average densities, es- and flow fields and heat transfer can be determined. Separate pro-
pecially for high oxidiser-to-fuel mass ratios. Very good density grams are used for cryogenic engines and additional tools enable
specific impulses can be obtained, because optimal specific im- propulsion system life prediction [42]. NASA Chemical Equilibrium
pulse occurs for relatively high O/F. Density specific impulses above with Applications is extensively used worldwide for propulsion
typical for MON/MMH combinations are achievable. Initial research system thermodynamic calculations [43]. Ponomarenko developed
shows that overall propulsion systems mass may practically not
a commercial code for LRE analysis – RPA (Rocket Propulsion Anal-
suffer from changing propellants from toxic combinations to HP/HC
ysis), which also uses a chemical equilibrium approach in order
and in some situations may lead to mass savings [22]. Naturally
to calculate TCA performance. It also enables conducting analyses
spacecraft mechanical configuration may pose a challenge, since
of chamber and nozzle cooling and considering effects of using
two tanks of the same size would not be possible for HP/HC,
different engine cycles [44]. Another program worth mentioning
unlike for MON/MMH, and large centre of gravity shifts would oc-
cur. Nevertheless, relatively high performance and compliance with is SCORES (SpaceCraft Object-oriented Rocket Engine Simulation)
REACH (Registration, Evaluation, Authorisation and Restriction of developed by Way and Olds at Georgia Institute of Technology.
Chemicals) regulation make HP/HC rocket engines one of the most It enables predicting rocket engine performance (specific impulse
serious candidates for MON/MMH replacement. and thrust) for given design variables. However, for some engines
it gives large relative errors [45]. An overview of some of the key
1.1. Design and optimisation software computer programs dedicated to LRE design and analysis was de-
scribed by Ernst [46]. Nearly every space propulsion development
The development of most liquid rocket engines is based on project consists of design optimisation. This is however usually
existing designs and lead engineers’ experience. However, using al- not global optimisation and focuses rather on element modifica-
ready flown and tested configurations causes significant limitations tion. Performing multidisciplinary considerations requires accurate
[23]. Even most advanced empirical development programs present modelling of aero-thermodynamic conditions and their impact on
scaling complications and some failures [24]. What is notable, up component performance and mass. A detailed review by Shelton
to date no universal method giving fully predictable scaling results [47] presents key American developments in the field of propul-
has been developed. Importantly, the use of existing engines as a sion system mass modelling. However, it mainly considers launcher
baseline does not enable easy optimisation. During the Space Race, configurations. Use of algorithms based on empirical sizing equa-
optimal control theory gained focus and rapid advances in the field
tions is insufficient apart from early project phases. For HP/HC LRE
of optimisation were made using computers. Numerous state-of-
these methods are not practical, due to the lack of operational in-
the-art methodologies and case studies used in aerospace projects
orbit propulsion systems of this type. Within this study, a limited
have been presented by Vanderplaats since the 1970’ [25–27].
range of configurations is analysed, since the development of an
Notable research was also conducted by Sobieski-Sobieszanski at
all-purpose model would limit mass modelling accuracy.
NASA [28]. A detailed review of optimisation in aerospace engi-
neering using parallel evolutionary algorithms was presented by
Wang et al. [29]. Various publications concerning rocket vehicles
2. Subject of optimisation
and their components have been written during the last 20 years.
Research dedicated to LRE optimisation and sizing is on-going for a
number of decades and propulsion system optimisation efforts are Bipropellant thrusters using 98% HP and kerosene are the sub-
presented in literature: [30–38]. However, up to 2014 no detailed ject of this optimisation. RP-1 is modelled due to its extensive use
HP/HC LRE optimisation data was commonly available. One of the for rocket propulsion and normalised properties. Although exten-
first attempts concerning space applications was done by Moon et
sive research is done in the field of hypergolic green bipropellants
al. [39]. Use of a simple mass model and given spacecraft data en-
engines [48], a traditional catalyst bed configuration is considered
abled determining the thrust level, chamber pressure, nozzle area
due to its higher technology readiness level. A simplified illustra-
ratio and oxidiser-to-fuel ratio in order to fulfil mission require-
tion of the architecture analysed is presented in Fig. 1. A thermal
ments. A Matlab code for optimising TCA parameters was used.
ignition system utilising a catalyst bed is assumed. Fuel is injected
1.2. Rocket engine modelling into a hot flow of decomposed 98% H2 O2 . A transverse liquid-gas
injection method is considered with a stepped combustion cham-
Various computer programs have been developed to ease en- ber geometry. Recent research in the field of HP/HC rocket engines
gine design and modelling. The ESA-supported EcoSimPro is a valu- in Poland utilised such a configuration [1,12,49,50]. The stepped
able tool for aerospace propulsion development. Apart from sub- combustion chamber approach, enabling flame stabilisation, mix-
system performance calculations, it enables obtaining flow pres- ing enhancement limiting chamber wall temperature was tested
sure drops, heat fluxes etc. Numerous design and optimisation extensively worldwide [3,6,9].
286 A. Okninski et al. / Aerospace Science and Technology 82–83 (2018) 284–293

Fig. 1. Engine configuration.

3. GLARE-O-1.0 – Green Liquid Apogee Rocket Engine showing that for some cases grids consisting below 1000 points
Optimisation were sufficient [22].
In order to ensure ignition of kerosene injected into decom-
The developed algorithm enables the optimisation of HP/kero- posed 98% HP, a minimal chamber pressure of about 9 bar is
sene bipropellant rocket engine chamber pressure, nozzle exit acceptable [53]. Therefore, chamber pressures investigated are in
pressure, oxidiser-to-fuel ratio and combustion chamber contrac- the range of 9–20 bar, where the upper limit significantly exceeds
tion ratio. As input, rocket engine nominal thrust level and total operational pressures for in-orbit systems.
impulse should be given. No orbital mechanics calculations are
conducted. The duration of the flight is assumed not to influ- p c ∈ (9, 20) [bar]
ence the total impulse required. Before initialisation, additional
Also the nozzle exit pressure is limited:
constraints and requirements should be given. This includes lim-
itations such as maximum nozzle exit diameter etc. Different ob-
p e ∈ (1000, 10000) [Pa]
jective functions can be used – i.e. the total mass of the propul-  
sion system. The goal of the algorithm is to minimise the objec- and p 0 = 0 [Pa]
tive function value. Four different variables are to be determined,
The lower boundary of 1000 Pa is set due to nozzle exit diameter
therefore algorithm optimisation was not a key concern and is not
limitations. To ensure proper ignition, combustion and relatively
discussed within this paper. Its data processing diagram is given
high performance, the O/F ratio is constrained:
in Fig. 2. Several aspects of engine design and operation are taken
into account during optimisation:
O/F ∈ (4, 11) [ ]
• Engine sizing On basis of existing practical designs and taking into account large
• Detailed mass modelling losses for small CR, CR values are also limited:
• Determination of flow thermodynamic properties and perfor-
mance characterisation CR ∈ (2, 20) [ ]
• Combustion chamber heat transfer and stress analysis
Input values are also constrained, narrowing the class of systems
considered, in order to maintain possibly high calculation accuracy.
Similarly as in [39], the Matlab fmincon() function is used. It
The total impulse value is limited:
deals with nonlinear smooth problems with constraints. Simple
inequality constraints are used. Due to the fact that fmincon()  
I t ∈ 0.1 × 106 , 20 × 106 [N s]
is a local optimisation method, using the Byrd algorithm [51,52],
a global search of starting points is done. In the developed tool, The lower limit corresponds to small monopropellant thrusters
a straightforward search using a uniform point distribution in the used for attitude control, while the larger value is well beyond the
R 4 ∈ [ p c , p e , O/F, CR] space was used. This space was limited due total impulse of the largest propulsion systems used for GTO trans-
to the optimisation variables constraints. One of the issues ad- fers [54]. Similarly, the thrust input constraints are as following:
dressed during the development of the program, was defining the
initial R 4 search grid. The influence of the grid was investigated, T ∈ (20, 1500) [N]
 A. Okninski et al. / Aerospace Science and Technology 82–83 (2018) 284–293 287

Fig. 2. Green Liquid Apogee Rocket Engine Optimisation tool schematic.

As can be seen, a slightly wider range of thrusts than for typical This formula is based on data of historical and operational space-
in-orbit bipropellant applications is considered. This allows using craft with bipropellant pressurisation subsystems including regu-
the algorithm for design activities concerning exploration missions. lators, for which values of p c / p tank f between 1.6 and 2.4 were
found. For modern pressurising systems with regulators the value
3.1. Modelling of 2.0 is a valid assumption. For the oxidiser tank, higher pressure
is needed in order to allow a significant pressure drop on the cat-
While obtaining component properties, several loops are be- alyst pack. A simplified approach for the pressure drop from [56]
ing used. These include obtaining convergence of heat fluxes and is used giving:
mass flow rates, while not exceeding acceptable stress levels. Due
to ground handling loads, it is assumed that all TCA components [bar]
p tankox = 2p c + 80 L cat
have wall thicknesses above 1 mm. A detailed mass model was [m]
developed with over 30 elements of the propulsion system being
modelled. TCA components are outlined in Fig. 3. Element geo- where L cat is the length of the catalyst bed. It has to be noted, that
metrical proportions are not restricted in order to enable their in reality the pressure drop is dependent on the type of catalyst
clear graphical presentation. Feeding subsystem components, such support material and the flow regime and influences the cata-
as tanks and valves, are also considered. lyst’s pack efficiency. These effects were not considered within this
study. All tanks were modelled as spherical pressure vessels with
3.1.1. Mass modelling safety factors of 1.5. However, there are some tank elements that
Most element masses are calculated via the determination of have to be added, like expulsion devices, venting etc. NASA report
their volume and knowing particular construction material den- by Myers et al. [57] gives a following correlation for lightweight
sity. During each optimisation loop, mass modelling takes place. tank masses:
Elements are modelled as axisymmetric plates and thin-walled
17.5V propellant
shells. It is assumed that each element is manufactured only from mtank = 0.6 + [kg]
[m3 ]
one specific construction material. The relative mass of propellant
residuals is assumed to be 3%. An ullage volume factor of 1.05 is where V propellant is the given liquid propellant volume. However,
taken into account [55]. The fuel tank pressure is assumed to be: this formula does not include the effect of tank pressure on its to-
tal mass. Therefore a membrane-theory-based correlation for wall
p tank f = 2p c thickness of pressure-loaded spherical vessels is used instead:
288 A. Okninski et al. / Aerospace Science and Technology 82–83 (2018) 284–293

Fig. 3. TCA divided into separate elements to ease modelling.

3 FStank p tanki ρtank The traditional L ∗ parameter approach is used. The L ∗ value was
mtanki = 0.6 + V tanki [kg]
2 σtank selected after analysis of HP/Kerosene laboratory thruster efficien-
This way tank mass is a function of its volume, safety factor, cies. A value of 1.2 m is utilised. For flame stabilisation, a stepped
interior pressure, average shell density and material strength. It chamber configuration is assumed and the suggested optimised
was decided to implement the constant of 0.6 kg from the pre- flameholder height (hstep ), based on vortex size prediction [3], is:
vious equation [57] enabling modelling the fact of worse mass 1
ratios achievable for very small tanks. The material assumed is a hstep = dc
6
composite shell with a density of 1800 kg/m3 . Despite its non-
isotropic properties an all-direction yield strength of 500 MPa is where dc is the combustion chamber inner diameter. For off-the-
considered. This is a serious simplification, but allows for rapid shelf components, such as valves, existing space-qualified compo-
tank sizing, while maintaining a conservative approach. A no-load- nents properties were implemented. It is assumed that material
bearing chemically-compatible liner is assumed. What is notable, chemical compatibility does not influence their mass. Moreover,
this method proves to be in very good agreement with existing for in-space chemical propulsion systems feeding pressures are in
commercially available tanks for in-orbit use. Using a simplified a relatively narrow range and pressure has little impact on flowline
analysis on the basis of the energy conservation principle [55], component mass. From analysis of space-rated propulsion system
pressurant mass for each propellant is calculated. It is assumed valves of leading companies a following relation for flow control
that both tanks are pressurised with the same type of gas (helium) valves was found:
and the pressurant has the same storage temperature. A factor
mTCA valve = f (ṁ f ) = 1.5437 ṁ0.4226
of 1.05 is introduced, for safety reasons, due to the requirement
for the possibility of excess gas occurrence [55]. Given a specified Mass of necessary interfaces, pipes, sensors, control equipment (in-
pressurant mass requirement, it can be proved, that the spherical cluding regulators), fittings and additional vales can be estimated
pressure vessel mass is constant, regardless of the designed storage as 20% of the propulsion system’s dry mass. A similar, however
pressure. We can obtain: slightly more optimistic, correlation was used by researchers from
KAIST [39].
mHe tank ∼
= 2 mpress
3.1.2. Combustion and thermodynamics
This matches accessible NASA data for lightweight components Combustion products thermodynamic parameters were deter-
reasonably well [57]: mined using NASA CEA 2 software. Direct automatic use of CEA
output can be possible in a more advanced, future version, thanks
mHe tank = 0.3 [kg] + 2.174 mHe
to the code developed by Bartkowiak [58]. In this version of the
This correlation takes into account the fact that smaller tanks have developed tool, thermodynamic parameters are based on CEA data
worse fluid mass fractions and that a larger number of smaller interpolation in the following form:
spherical pressure vessels instead of one is not optimal in terms of    
O O
mass minimisation. The following correlation is finally used: Property , p c = Property8 bar ∗ f Property ( p c )
F F
mHe tank = 0.3 [kg] + 2 mHe Maximum relative errors were below 0.002 for properties within
an O/F range of 4–12 and p c range of 8–20 bar. Examples of prop-
It is more optimistic than the NASA formula, but this is due to the
erties obtained during HP/kerosene engine modelling are given in
advances in the field of materials in recent years. For both propel-
Fig. 4.
lants separate pressurant tanks are assumed. The pressurant tank
Performance calculations were implemented using 1D and
pressure is set to:
quazi-1D correlations from state-of-the-art literature [55,59]. Cal-
culations were performed for a wide range of O/F and p c values
p He tank = 15 p f tank
assuming chemical equilibrium. However, the specific heat ratio
This was set using data of apogee propulsion systems of historical and molar mass of combustion products are stated as constant
European satellites including DFS Kopernikus, Eureca and Eurostar. along the nozzle. Losses in chamber pressure due to non-infinite
In order to determine the minimal chamber length, good knowl- contraction ratios are taken into account. A pressure loss factor
edge of combustion phenomena and their time scales is essential. for fully-developed turbulent flow in pipes gives an estimation of
 A. Okninski et al. / Aerospace Science and Technology 82–83 (2018) 284–293 289

Fig. 4. Selected properties of combustion products of HP/kerosene.

friction loss increase for large chamber lengths. This is however transfer between nozzle elements would be small. For typical LRE
negligible. Non-ideal correction factors were used for mass flow about 70–95% of heat transfer occurs due to convection, while the
and specific impulse calculation: rest due to radiation. In literature, a methodology considering only
radiation of H2 O and CO2 can be found [60]. For RP-1 and 98%
ṁ = ξd ṁideal HP with an O/F of 7 and p c of 10 bar, their mass fraction in the
Isp = ξ v Ispideal combustion products mixture is over 85%, what shows that this
methodology can be assumed as appropriate, especially that radi-
ξ v takes into consideration a lower than ideal specific impulse and
ation of CO is negligible [61]. Therefore, a method described by
ξd a slightly higher mass flow than planned what is explained by
Sutton [55]. ξd = 1.05 and ξ v = 0.91. These values correspond to Sinajew, Dobrowolski is used, where radiative heat fluxes can be
engines of the considered size [1]. determined [62]:
  3.5   3 .5 
3.1.3. Heat transfer 4184 3 10 TG T GW
qCO2 = 3.5 p CO2 L rad − [W]
It is assumed that only forced convection and radiation effect 3600 9.81 100 100
heat transfer through combustion chamber walls. No heat accumu-
 0.8  3  3 
lation is considered because the aim is to operate in steady-state 4184 10 TG T GW
mode. Both radial and axial temperature gradients should present,
qH2 O = 3.5 p H2 O L 0rad
.6
− [W]
3600 9.81 100 100
but no circumferential component, due to axisymmetric configura-
tions analysed, occurs. Yet it is assumed that heat transfer occurs where T G is the combustion temperature, T GW is the flow-side
only radially and does not decrease performance [58]. This is a wall temperature and p i is the combustion product partial pres-
simplified approach and it is proposed since the full design is not sure obtained using Dalton’s Law. L rad is the characteristic radi-
known. Relatively low chamber pressures are envisaged for in-orbit ation dimension, which represents the average beam path length
propulsion systems and therefore axial pressure- and tempera- inside the combustion chamber. For chamber elongations above 4
ture gradients are lower than in small LRE with high operational it can be assumed as 0.9 dc where dc is the combustion chamber
pressures. Nozzle contour elements are very thin in reference to internal diameter. For lower elongations the following correlation,
their length and wall temperature gradients along LRE main axis based on data from [62], is used.
are relatively low in comparison to radial temperature gradients.
 4  3
Moreover, ceramics, which have low thermal conductivity, are con- L rad Lc Lc
sidered for the nozzle throat region, thus axial conductive heat = 4.2424 · 10−3 + 6.242 · 10−2
dc dc dc
290 A. Okninski et al. / Aerospace Science and Technology 82–83 (2018) 284–293

 2  
−1 Lc −1 Lc
− 3.456 · 10 + 8.727 · 10
dc dc

+ 2.960 · 10−2

L c is the length of the cylindrical part of the combustion chamber.

The effect of soot deposition on walls is not considered. Convec-
tion heat transfer was calculated using the Bartz equation, which
enables determining the heat transfer coefficient between the gas
side and the chamber wall [63]. Convective and radiative heat
fluxes are combined and a classic iterative approach, based on
Fourier’s Law, is used to determine wall temperatures. Checks if Fig. 5. Example problem search function value during algorithm run.
these values are within an acceptable range are done.
4. Example problem and results
3.1.4. Stress
The developed algorithm considers basic thermal and pressure A single example is given in this paragraph. The Eutelsat-2
loads. Non-uniform temperatures of the combustion chamber ele- satellite is considered due to commonly available data concerning
ments cause significant thermal stresses. They can be easily de- its mission [54]. Using a GTO manoeuvre, it had a total impulse
fined for cylindrical geometries, where heat is transferred only of 3.1 · 106 N s utilising a storable toxic bipropellant system. An
radially [64]. Radial, hoop and longitudinal stress components are apogee velocity increase of approximately 1800 m/s was required.
considered. Their values can be written, as a function of distance The satellite mass was 1878 kg including 1069 kg of propellants.
from chamber major axis, as following. The propulsion system dry mass was 101 kg. Since for a given
launch opportunity one would like to maximise the payload mass
 b2  in comparison to the total launched mass, the following is investi-
− E α ( T a − T b ) ln( br ) r2
−1
σr therm (r ) = − gated: What is the maximum allowable dry mass, excluding propulsion
2(1 − ν ) ln( ba ) b2
−1
a2 system elements, for the GEO Eutelsat-2 satellite if the toxic propulsion
system was substituted by a HP/HC green bipropellant rocket system?
 b2 
− E α ( T a − T b ) ln( br ) − 1 r2
+1 Maximising this value is envisaged to enable launching a heavier
σθ therm (r ) = +
2(1 − ν ) ln( ba ) b2
−1 payload to GEO. The satellite total launch mass is defined as the
a2 same as in the original design, so that the analysis does not impact
σz therm (r ) = ν (σr therm + σθ therm ) − λ E T z the launcher architecture. A constant thrust of 500 N is assumed
and the original satellite bipropellant total impulse is given as a
T a and T b are inner and outer wall temperatures and a and b are requirement. In order to answer the question stated, the following
radii of the inner and outer wall surface. ν is Poisson’s ratio, α is optimisation function is used.
the coefficient of thermal expansion, E is the material’s Young’s
modulus and λ represents thermal conductivity. No creep defor- 1
F obj (x) =
mation is analysed. For the purpose of the developed algorithm, msat
the “−λ E T z ” term of σz therm (r ) is not taken into account,
where msat is the dry mass of the satellite excluding its propulsion
since ceramics have low conductivity and axial heat transfer is not
system. Algorithm optimisation constraints are unchanged accept
considered. Stress due to chamber internal pressure is determined
the minimum value of p e , which is modified to 3000 Pa in or-
basing on handbook expressions derived from the membrane the-
der to limit nozzle dimensions. In Fig. 5 the optimisation function
ory. Thermal and pressure-related values obtained for each direc-
value is shown during optimisation. The final step size, which is
tion are added and total stress is calculated from the Huber–Mises
the value of x displacement within the R 4 ∈ [ p c , p e , O/F, CR] space,
Theorem, knowing that principal axis stresses are considered. Max-
is 1.58979 · 10−7 .
imum stress is compared with material strength at the tempera-
This simple exercise shows that a chamber pressure of 10 bar,
ture obtained from thermal calculations.
high contraction ratio of 11 and a maximum allowed nozzle ex-
pansion ratio (exit pressure reaching 3000 Pa) are optimal values
3.1.5. Construction materials for the problem considered with an O/F just over 6. A comparison
Material properties were implemented into the algorithm as of results concerning Liquid Apogee Engines (LAE) from this op-
functions of temperature. It was decided to baseline radiation as timisation exercise with original Eutelsat-2 data and results from
the main cooling mechanism. Some disadvantages of this config- reference [39] are given in Table 1.
uration are relatively large costs of elements’ manufacturing since A reasonable expansion ratio of 335 was obtained using the
the use of mostly refractory materials and advanced coatings or developed algorithm, while reference [39] gives a result of 580,
composites are required [65]. Although molybdenum configura- what is a surprisingly high value in comparison with state of the
tions were also studied, thermal barrier and anti-oxidation coat- art systems. In the present study a high contraction ratio was ob-
ings for this material were not defined. Therefore, for the purposes tained. This is due to high heat transfer loads occurring for low CR
of most analyses rhenium is set as the combustion chamber mate- and pressure losses. Moreover, for higher diameters, slightly bet-
rial. Rhenium mechanical properties were obtained using literature ter autoignition characteristics parameters can be observed [69].
data [66,67]. ANSI 316L is considered for the catalyst bed housing. Although the effect of the oxidiser-to-fuel mass flow ratio on all
C/SiC is used for certain nozzle elements. Relatively low combus- chamber flow properties is modelled, the nozzle outflow assumes
tion temperatures, in comparison to other bipropellants, make it a frozen chemical composition as in [55]. This leads to some ap-
perfectly suitable and recent investigations at German Aerospace proximations and influences performance prediction. For a more
Centre (DLR) showed that with an extra liner, it is resistant to ero- practical O/F of 7 [39] calculations were made and the impact of
sion [68]. the considered variables is shown in Fig. 6.
 A. Okninski et al. / Aerospace Science and Technology 82–83 (2018) 284–293 291

Table 1
Comparison of LAE – this study, Eutelsat-2, KAIST.

Parameter This study Eutelsat-2 KAIST [39]

LAE S400 LAE LAE
Nominal p c [bar] 10.1 10.0 11.7
Propellants, O/F [ ] 98% HP/RP-1, 6.0 MON-1/MMH, 1.65 90% HP/Jet-A, 7.36
Nozzle area ratio 335 330 580
Nominal specific impulse [s] 315 318 312
Density specific impulse [s] 406 369 399

Fig. 6. Available mass maximisation results (for an O/F of 7).

It can be again seen that low chamber pressure and high expan- to MON/MMH. Despite a fully new mass model, similar operat-
sion ratios are beneficial, enabling mass minimisation. For higher ing parameters of the proposed LAE, as in state of the art storable
mixture ratios increased contraction ratios are recommended. propulsion, were obtained. The modelling method proposed does
The maximum mass available for the satellite platform (without not fully take into account the fact that off-the-shelf components
propulsion system) and payload for an O/F of 7 is 711 kg, while are used. Heritage solutions enable lowering overall development
the original EUTELSAT-2 achieved a value of 708 kg. This proves time and costs and this was not assessed via an optimisation vari-
that very similar systems in terms of mass and performance can able. However, using existing elements, such as tanks etc. causes
be developed using HP/kerosene propellants. Lower thermal loads some components to be capable of withstanding unnecessarily
and high storage density make HP/kerosene a serious candidate high loads, increasing propulsion system mass. Future improve-
for LAE applications and replacing existing MON/MMH architec- ments of the presented MDO software should consist of imple-
tures. menting a wider choice of chamber cooling methods and materi-
als, as well a variety of fuel injection configurations. A chemical
5. Conclusions kinetics module would improve performance prediction. System
reliability analyses must also be considered. Moreover, in terms
The developed algorithm enables conducting significant ad- of spacecraft or satellite optimisation, different propellants apart
vances in the field of HP/HC rocket engines without the need to from HP should be analysed as possible oxidisers for in-orbit and
focus on long-lasting conceptual project phases. A swift transition exploration missions. A wider construction material database with
to Computational Fluid Dynamics and Finite Element Method anal- chemical compatibility data would also be an asset. In order to use
yses can be done using the “baseline” design obtained through the developed algorithm for product development, more advanced
the Matlab-based tool. Most formulas used were not presented thermal models should be implemented. Taking engine control and
herein, yet an overview of the methodology was given. The de- power supply constraints, as well as plume interaction with the
veloped algorithm allows obtaining reasonably accurate mass pre- engine and spacecraft elements, into account would be highly ap-
dictions for HP/HC LAE. The software can be also used for more preciable. A separate module should be dedicated to engine de-
detailed project phases if 98% HP and kerosene are used as pro- velopment and operation costs. Adding these features would be a
pellants. The lack of operational spacecraft with HP/HC rocket next step toward the introduction of green bipropellant systems
engines may lead to some uncertainty concerning some project for in-space applications.
assumptions, however the work described presents one of the
most detailed attempts, to optimise this class of rocket propulsion Conflict of interest statement
systems, noted. Results of an optimisation problem were given,
showing that HP/kerosene may serve as an attractive alternative There is no conflict of interest.
292 A. Okninski et al. / Aerospace Science and Technology 82–83 (2018) 284–293

Acknowledgements [20] G. Rarata, K. Rokicka, The manganese oxides decomposition catalysts for highly
concentrated hydrogen peroxide, Trans. Inst. Aviat. 240 (3) (2015), https://doi.
org/10.5604/05096669.1194985.
This work is based on results of the project of the Space Tech- [21] Z. Gut, L. Mezyk, P. Wolanski, Research on thermal decomposition of hydro-
nology Department of Center of Space Technologies of the Insti- gen peroxide, in: 6th European Conference for Aeronautics and Space Sciences
tute of Aviation. It was developed thanks to the funding of the (EUCASS), Cracow, Poland, 2015.
Institute of Aviation and supports activities connected to bipropel- [22] A. Okninski, Green Bipropellant Rocket Engine Multidisciplinary Optimization
and Sizing Tool Development, MSc. Thesis, Supervisor: Jan Kindracki, Warsaw
lant rocket engine development in European programs. Some of
University of Technology, Warsaw, 2014.
the work was done during the preliminary phase of the European [23] S. Hyde, A. Okninski, Additive manufacturing design optimised bipropellant
Space Agency GRACE (Green bi-propellant apogee rocket engine for injector, in: Space Propulsion 2016, Rome, Association Aéronautique et Astro-
future spacecraft) – Contract No: 4000112705/14/NL/CBi – project nautique de France, 2016.
funded within ESA Polish Industry Incentive Scheme. The authors [24] J. Hulka, Scaling of Performance in Liquid Propellant Rocket Engine Combustors,
Jacobs Engineering, ESTS Group, NASA MSFC, Huntsville, AL, 2007.
would like to thank all the members of the project who com- [25] G. Vanderplaats, Structural optimization—past, present, and future, AIAA J.
mitted to the development of green LRE technology. In particular, 20 (7) (1982) 992–1000, https://doi.org/10.2514/3.51158.
great support was provided by M.Sc. Eng. B. Bartkowiak, M.Sc. Eng. [26] G. Vanderplaats, Numerical Optimization Techniques for Engineering Design:
K. Sobczak, Ph.D. Eng. P. Surmacz and Ph.D. Eng. D. Kublik. Initial With Applications, vol. 1, McGraw-Hill, New York, 1984.
[27] G. Vanderplaats, F. Moses, Structural optimization by methods of feasible di-
work was carried out by the authors of this paper in the Institute
rections, Comput. Struct. 3 (4) (1973) 739–755.
of Heat Engineering at Warsaw University of Technology. [28] J. Sobieszczanski-Sobieski, R.T. Haftka, Multidisciplinary aerospace design op-
timization: survey of recent developments, Struct. Optim. 14 (1) (1997) 1–23,
References https://doi.org/10.1007/BF01197554.
[29] J.F. Wanga, J. Periaux, M. Sefrioui, Parallel evolutionary algorithms for optimiza-
tion problems in aerospace engineering, J. Comput. Appl. Math. 149 (2002)
[1] A. Okninski, Design of a Small Green Bipropellant Rocket Engine Using Hydro-
155–169, https://doi.org/10.1016/S0377-0427(02)00527-7.
gen Peroxide as Oxidizer, Thesis, Warsaw University of Technology, February
[30] G. Cai, J. Fang, Y. Zheng, X. Tong, J. Chen, J. Wang, Optimization of system pa-
2014.
rameters for liquid rocket engines with gas-generator cycles, J. Propuls. Power
[2] M. Ventura, 300 lbf 90% H2O2/RP-1 Rocket Engine, General Kinetics Inc., 2004. 26 (1) (2010) 113–119, https://doi.org/10.2514/1.40649.
[3] J. Sisco, B.L. Austin, J.S. Mok, W.E. Anderson, Autoignition of kerosene by de- [31] A. Oyama, M.S. Liour, Multiobjective optimization of rocket engine pumps using
composed hydrogen peroxide in a dump-combustor configuration, J. Propuls. evolutionary algorithm, J. Propuls. Power 18 (3) (2002) 528–535, https://doi.
Power 21 (May–June 2005) 3. org/10.2514/2.5993.
[4] K. Miller, J. Sisco, B. Austin, T. Martin, W.E. Anderson, Design and ground test- [32] W. Shyy, P.K. Tucker, R. Vaidyanathan, Response surface and neural network
ing of a hydrogen peroxide/kerosene combustor for RBCC application, in: 39th techniques for rocket engine injector optimization, J. Propuls. Power 17 (2)
AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, Huntsville, Al- (2001) 391–401, https://doi.org/10.2514/2.5755.
abama, 2003. [33] W. Shyy, N. Papila, R. Vaidyanathan, K. Tucker, Global design optimization
[5] Seong-Up Ha, The Use of Hydrogen Peroxide, Moscow State Aviation Institute, for aerodynamics and rocket propulsion components, Prog. Aerosp. Sci. 37 (1)
Moscow, 2007. (2001) 59–118.
[6] S. Jo, Design and Validation of Bipropellant Rocket Engine with Storable Pro- [34] K. Jayashree, A. Ramamurthi, Optimization of mixture ratio distribution in liq-
pellant, Ph.D. Thesis, KAIST, 2012. uid propellant rocket thrust chamber, J. Propuls. Power 8 (3) (1992) 605–608.
[7] H. Kang, D. Jang, S. Kwon, Demonstration of 500 N scale bipropellant thruster [35] Pratt & Whitney, Orbit Transfer Vehicle Engine Study, Volume 2: Technical Re-
using non-toxic hypergolic fuel and hydrogen peroxide, Aerosp. Sci. Technol. 49 port, NASA MSFC, Alabama, 1980, N80-32432.
(2016) 209–214, https://doi.org/10.1016/j.ast.2015.11.038. [36] A. Martinez, H.G. Diem, Orbital Transfer Vehicle Advance Expander Cycle En-
[8] D. Krejci, A. Woschnak, C. Scharlemann, K. Ponweiser, Performance assessment gine Point Design Study, Rocketdyne, Alabama, 1979.
of 1N bipropellant thruster using green propellants H2 O2 /kerosene, J. Propuls. [37] D. Manski, G. Hagemann, D. Sahick, Dual-expander rocket engines in single
Power 29 (1) (2013) 285–289, https://doi.org/10.2514/1.B34633. stage-to-orbit vehicles, Acta Astronaut. 40 (2–8) (1997) 151–163.
[9] M. Volpi, Analisi dell’acessione di un propulsore bipropellente H2 O2 –C2 H6 , [38] E.K. Bair, Space Transportation Main Engine Configuration Study, NASA, Al-
Ph.D. Thesis, University of Pisa, Pisa, 2009. abama, 1986.
[10] A. Cervone, L. Torre, A. Pasini, L. Romeo, A. d’Agostino, Development of Hy- [39] Y. Moon, C. Park, S. Jo, S. Kwon, Design specifications of H2 O2 /kerosene bipro-
drogen Peroxide Rockets at ALTA S.P.A.: The Past, the Present and the Future, pellant rocket system for space missions, Aerosp. Sci. Technol. 33 (1) (February
International Astronautical Federation, Naples, 2012, IAC-12-C4.3.5. 2014) 118–121, https://doi.org/10.1016/j.ast.2014.01.006.
[11] The Tauri Group, State of the Satellite Industry Report, Satellite Industry Asso- [40] P. Cobas-Herrero, B. García-Gutiérrez, R. Pérez-Vara, R. Avezuela-Rodríguez, C.
ciation, 2013. Gregori-de la Malla, An ESA state-of-the-art simulation tool for space applica-
[12] A. Okninski, B. Bartkowiak, K. Sobczak, D. Kublik, P. Surmacz, G. Rarata, B. tions, in: 9th International Workshop on Simulation for European Space Pro-
Marciniak, P. Wolanski, Development of a small green bipropellant rocket en- grammes (SESP), ESTEC, Noordwijk, the Netherlands, November 2006.
gine using hydrogen peroxide as oxidizer, in: 50th AIAA/ASME/SAE/ASEE Joint [41] J. Amer, J. Moral, J. Salvá, Adaption of the ESPSS/ECOSIMPRO platform for the
Propulsion Conference, Cleveland, OH, 2014, AIAA-2014-3592. design and analysis of liquid propellant rocket engines, in: PEGASUS–AIAA Stu-
[13] W. Wisniowski, P. Wolanski, The role of the Institute of Aviation in Space dent Conference, Poitiers, 2012.
Technology Research, Trans. Inst. Aviat. 234 (1) (2014) 9–16, https://doi.org/ [42] Astrium, Rocket engine design and analysis, [Online] Astrium, http://cs.astrium.
10.5604/05096669.1106721 (in Polish). eads.net/sp/launcher-propulsion/analytical-engineering/rocket-engines.html.
[14] L. Loroch, L. Rams, Scientific infrastructure and organisation of the Center of [43] B.J. McBride, S. Gordon, Computer Program for Calculation of Complex Chemi-
Space Technologies of the Institute of Aviation, Trans. Inst. Aviat. 234 (1) (2014) cal Equilibrium Compositions and Applications: II. Users Manual and Program
17–24, https://doi.org/10.5604/05096669.1106737. Description, NASA, 1996.
[15] B. Marciniak, A. Okninski, B. Bartkowiak, M. Pakosz, K. Sobczak, W. Florczuk, [44] A. Ponomarenko, RPA – tool for rocket propulsion analysis, in: Space Propulsion
D. Kaniewski, J. Matyszewski, P. Nowakowski, D. Cieslinski, G. Rarata, P. Sur- Conference, Cologne, 2014.
macz, D. Kublik, D. Rysak, J. Smetek, P. Wolanski, Development of the ILR-33 [45] D. Way, J. Olds, SCORES: web-based rocket propulsion analysis for space trans-
“Amber” sounding rocket for microgravity experimentation, Aerosp. Sci. Tech- portation system design, in: 35th AIAA/ASME/SAE/ASEE Joint Propulsion Con-
nol. 73 (2018) 19–31, https://doi.org/10.1016/j.ast.2017.11.034. ference, Los Angeles, CA, 1999, AIAA 99-2353.
[16] A. Okninski, J. Kindracki, P. Wolanski, Rocket rotating detonation engine flight [46] R. Ernst, Liquid Rocket Analysis (LiRA), Development of a Liquid Bi-Propellant
demonstrator, Aircr. Eng. Aerosp. Technol. 88 (4) (2016) 480–491, https://doi. Rocket Engine Design, Analysis and Optimization Tool, Master of Science Thesis,
org/10.1108/AEAT-07-2014-0106. Delft University of Technology, 2014.
[17] P. Wolanski, M. Folusiak, D. Kublik, K. Swiderski, G. Rarata, K. Sobczak, P. Sur- [47] J.D. Shelton, Launch Vehicle Propulsion Parameter Design Multiple Selection
macz, W. Florczuk, Techniques and Technologies of Space Rockets, Institute of Criteria, PhD Thesis, University of Alabama in Huntsville, Huntsville, AL, 2004.
Aviation, Warsaw, 2009. [48] W. Florczuk, G. Rarata, Assessment of various fuel additives for reliable hyper-
[18] A. Okninski, D. Sors Raurell, A. Mitre, Feasibility of a low-cost sounding rock- golic ignition with 98%+ HTP, in: 66th International Astronautical Congress,
oon platform, Acta Astronaut. 127 (2016) 335–344, https://doi.org/10.1016/j. International Astronautical Federation, Jerusalem, 2015.
actaastro.2016.06.010. [49] A. Okninski, B. Bartkowiak, Green liquid rocket engine design, Trans. Inst. Aviat.
[19] P. Surmacz, Influence of various types of Al2 O3 /Mx Oy catalysts on performance 234 (1) (2014) 73–81, https://doi.org/10.5604/05096669.1106782 (in Polish).
of a 100 mm chamber for decomposition of 98%+ hydrogen peroxide, Trans. [50] K. Sobczak, P. Surmacz, B. Bartkowiak, A. Okninski, G. Rarata, P. Wolanski, D.
Inst. Aviat. 240 (3) (2015) 58–68, https://doi.org/10.5604/05096669.1194986. Kublik, F. Valencia Bel, Test campaign of a green liquid bi-propellant rocket
 A. Okninski et al. / Aerospace Science and Technology 82–83 (2018) 284–293 293

engine using catalytically decomposed 98% hydrogen peroxide as oxidizer, [61] M. Martinez-Sanchez, Radiation Heat Transfer and Cooling, [Online] http://
in: 53rd AIAA/SAE/ASEE Joint Propulsion Conference, Atlanta, GA, 2017, AIAA ocw.mit.edu/courses/aeronautics-and-astronautics/16-512-rocket-propulsion-
2017-4926. fall-2005/lecture-notes/lecture_11.pdf, May 2014.
[51] R.H. Byrd, Mary E. Hribar, Jorge Nocedal, An interior point algorithm for large- [62] G. Siniariew, M. Dobrowolski, Liquid Rocket Engine Design, Polish Ministry of
scale nonlinear programming, SIAM J. Optim. IX (4) (1999) 877–900, https:// National Defence, Warsaw, 1957 (Polish translation).
doi.org/10.1137/S1052623497325107. [63] D.R. Bartz, A simple equation for rapid estimation of rocket nozzle convective
[52] R.H. Byrd, J.C. Gilbert, J. Nocedal, A trust region method based on interior point heat transfer coefficients, Jet Propuls. 27 (1) (1957) 1235–1245.
techniques for nonlinear programming, Math. Program. 89 (1) (2000) 149–185. [64] M. Bjak-Zochowski, A. Jaworski, G. Krzesinski, T. Zagrajek, Mechanics of Struc-
[53] V.N. Sadov, Ignition of kerosene and hydrogen peroxide in combustion cham- tures, Warsaw University of Technology, Warsaw, 2004 (in Polish).
ber by fire jet, in: 9th International Hydrogen Peroxide Propulsion Conference, [65] S. Hyde, A. Okninski, Silicon oil thermal barrier system for in space applica-
Poitiers, September 2006. tions, in: Space Propulsion 2016, Rome, Association Aéronautique et Astronau-
[54] P. Erichsen, Performance Evaluation of Spacecraft Propulsion Systems in Rela- tique de France, 2016.
tion to Mission Impulse Requirements, Swedish Space Corporation, Solna/Swe- [66] J. Lambert, Refractory Metals and Alloys ASM Handbook, Volume 2: Properties
den, 1997. and Selection: Nonferrous Alloys and Special-Purpose Materials, ASM Hand-
[55] G.P. Sutton, O. Biblarz, Rocket Propulsion Elements, John Wiley & Sons, New book Committee, 1990.
York, 2001. [67] J.R. Wooten, P.T. Lansaw, High Temperature Oxidation-Resistant Thruster Re-
[56] Peroxide propulsion, decomposition catalysts, peroxide propulsion, [Online] search, NASA, Cleveland, Ohio, February 1990, NASA Contractor Report 185233.
http://www.peroxidepropulsion.com/article/38. [68] F. Breede, M. Friess, Development of Advanced CMC Materials for Dual-
[57] R.M. Myers, S.R. Oleson, F.M. Currant, S.J. Schneider, Chemical and Electric Bell Rocket Nozzles, Sonderforschungsbereich Transregio 40 – Annual Reports,
Propulsion Options for Small Satellites, NASA Lewis Research Center, 1994. 2009.
[58] B. Bartkowiak, Laboratory Liquid Rocket Engine – Heat Transfer Analysis, Pre- [69] J. Sisco, B.L. Austin, J.S. Mok, W.E. Anderson, Autoignition of kerosene by de-
MSc. Thesis, Warsaw University of Technology, February 2014. composed hydrogen peroxide in a dump-combustor configuration, J. Propuls.
[59] A. Huang, D. Huzel, Modern Engineering for Design of Liquid Propellant Rocket Power 21 (3) (May–June 2005).
Engines, Progress in Astronautics and Aeronautics, AIAA, 1992.
[60] S. Torecki, Rocket Engines, Wydawnictwo Komunikacji i Lacznosci, 1984 (in
Polish).

View publication stats