Available online at www.sciencedirect.

com

ScienceDirect
Energy Procedia 101 (2016) 614 – 621

71st Conference of the Italian Thermal Machines Engineering Association, ATI2016, 14-16
September 2016, Turin, Italy

Preliminary Design Method of a Turbopump Feed System

for Liquid Rocket Engine Expander Cycle
Angelo Letoa,*, Raffaele Vottaa, Aldo Bonfigliolib
a
CIRA Italian Aerospace Research Center, via Maiosie, 81043 Capua(CE), Italy
b
University of Basilicata, V.le dell’Ateneo Lucano, 10, 85100 Potenza, Italy

Abstract

The present research effort deals with simplified theoretical models for the preliminary design and performances assessment of
centrifugal pumps for liquid rocket propulsion. These models have been developed within the Concurrent Design Facility, under
development at the Italian Aerospace Research Centre (CIRA), in the framework of the HYPROB program. In particular, this work
is aimed at developing a theoretical model, via the implementation of a MatLab code, capable to predict the geometry and
performance of centrifugal turbopumps, thus providing useful indications for the preliminary design of the turbopump feed system.

© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
© 2016 The Authors. Published by Elsevier Ltd.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the Scientific Committee of ATI 2016.
Peer-review under responsibility of the Scientific Committee of ATI 2016.
Keywords: Centrifugal Pump; Methane; Liquid Rocket Engine;

1. Introduction

In the development of space propulsion systems, the study of the turbopumps used in liquid rocket engines is of
great interest, since its proper design allows to increase the generated thrust. Due to the complexity in analysing the
flow behaviour inside the centrifugal pump, accurate Computational Fluid Dynamics (CFD) design analyses are

* Corresponding author. Tel.: +39-3396582340.

E-mail address: a.leto@cira.it

1876-6102 © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the Scientific Committee of ATI 2016.
doi:10.1016/j.egypro.2016.11.078
 Angelo Leto et al. / Energy Procedia 101 (2016) 614 – 621 615

required. However, in the framework of a preliminary design phase, such as those used in the Concurrent Design
Facility (CDF), fast predicting methods are mandatory. Therefore, CFD analyses are not a viable option within CDF
investigations, being very complex and time consuming and the turbopump design must be afforded mainly by means
of engineering tools.
A MatLab tool, called PGD (Pump Global Design), which relies upon the CoolProp [1] libraries to obtain the
thermodynamic data, has been developed at CIRA within the HYPROB programme. The goal of the HYPROB project
is to contribute to the evolution and consolidation of the national technology and system development capabilities on
rocket propulsion for future space applications. Overall the project pursues two strategic goals: i) the ability to design
and build rocket engines with liquid or hybrid propellants, verified by building system and technological
demonstrators; and ii) the development of ground or flight qualified enabling technologies and the related instrumental
capabilities, to support the design of future space. Liquid oxygen and methane (LOX/LCH4) rocket engine
technologies represent the major effort within the project. In fact, the development of hydrocarbon, such as methane
(LOX/CH4), or hybrid is becoming a technological asset for future generation launchers and space transport systems.
Today methane appears to be the most interesting avenue for liquid propellant rockets because it brings together – in
addition to good operating performance – easy storage, limited danger and no toxicity, as well as relatively low
extraction and conditioning costs.
In this framework, this paper describes the simplified theoretical models implemented in the PGD tool that are
currently being used for the preliminary design and performance assessment of the centrifugal pumps used in a rocket
propulsion system. PGD is able to determine the global geometry and performances of the turbomachines, as well as
the thermodynamic properties of the fluid at discharge.
PGD has been preliminary validated using the available data of two engines, namely the RL10A-3-3A and VINCI;
then it has been use for the preliminary design of the centrifugal pump of a Methane expander rocket engine.

Nomenclature
ΔH Total head ρin Inlet density
pd Discharge pressure Ns Specific speed
pS Suction pressure ω Pump rotating speed
Ys - Yd Friction losses in the channels before Q pump flow rate
and after the pump ηH Hydraulic efficiency
ρd Discharge density

2. Pump Global Design

Starting from engineering relationships of centrifugal turbopumps (CTP) global parameters a MatLab Tool, namely
Pump Global Design (PGD) has been developed. This tool allows preliminary design of CTP at an engineering-based
level.
The PGD tool requires different input parameters. They can be distinct in primary and secondary inputs, according
to the scheme provided in Figure 1.
Inputs

Primay Inputs

Cool Prop

Input from
designer
Outputs

Fig. 1. PGD tool scheme

The PGD tool requires different input parameters. The primary input parameters are inlet temperature, mass flow
rate, inlet pressure and discharge pressure. The PGD code makes use of CoolProp libraries through which the obtained
secondary inputs, that are inlet density, vapor pressure, discharge density and temperature.
616 Angelo Leto et al. / Energy Procedia 101 (2016) 614 – 621

3. Theoretical models for specific speed and head-density relationship

The theoretical models for head-density relationship and specific speed, developed in the PGD tool are summarized
hereinafter.

3.1. Specific speed

The specific speed was first introduced by Camerer in 1914 and further developed by Stepanoff in 1948 [2]. The
pump specific speed is a characteristic value typically defined at the point of maximum efficiency, which is usually
the design point [3]. It is defined as:

ωඥQ
Ns ൌ ሺgΔHሻ3Ȁ4 (1)

In the PGD tool the following empirical correlation for the impeller hydraulic efficiency,Ʉୌ is implemented [5]:

Ʉୌ ൌ ͲǤͶͳͻͺͻ ൅ ʹǤͳͷʹͶୱ െ ͵ǤͳͶ͵Ͷୱଶ ൅ ͳǤͷ͸͹͵ୱଷ (2)

3.2. Relationship between head and pressure

The manometric head ΔH is the total head developed by the pump. This head is slightly less than the head generated
by the impeller due to frictional losses in the pump. It represents the pressure increase generated by the pump between
the discharge and suction sections and can be expressed as:
୮ౚ ି୮౩
ο ൌ ൅ ୱ ൅ ୢ (3)
஡‫כ‬୥

The expression for discharge pressure [3] is:

୮
’ୢ ൌ ቀ஡ ౩ ൅ οቁɏୢ (4)
౟౤

Equation (4) shows that the pump discharge pressure pd equals the propellant pump suction pressure ps plus the
pressure rise across the pump, Δppump.
The peripheral velocity, u, the tangential component of absolute velocity, c u, and the relative velocity, w, are
evaluated in PGD by the following relationships:

୥οୌ ୥୼ୌ ୳ୡ

౫
—ൌට …୳ ൌ ™ ൌ  ୘ୟ୬ሾஒሿ (5)
ந ୳஗ౄ

4. Test case

The reliability of the present theoretical model has been evaluated by considering two test cases, namely the
RL10A-3-3A and VINCI engines, as summarized hereinafter.

4.1. RL10A-3-3A Engine

The RL10A rocket engine is an important component of the United States space infrastructure [4]. Two RL10
engines form the main propulsion system for the Centaur upper stage vehicle, which boosts commercial, scientific,
and military payloads from a high altitude into Earth orbit and beyond (planetary missions). The Centaur upper stage
is used on both Atlas and Titan launch vehicles. The RL10A-3-3A engine design is based on an expander cycle.
Hydrogen fuel is used to cool the thrust chamber and nozzle, and the thermal energy transferred to the coolant is
used to drive the turbopumps. Engine data available in literature for the RL10A-3-3A are summarized in Table 1 [5].
 Angelo Leto et al. / Energy Procedia 101 (2016) 614 – 621 617

Table 1. Engine Data RL10A-3-3A

Engine Data
Thrust 73.0042 [KN] Pressure discharge 36.694 [bar]
Fuel Flow Rate (LH2) 2.7945 [Kg/s] Temperature discharge 26.47 [K]
ΔHead (Centrifugal Pump 1 stage) 5138.3184 [m] Mass flow discharge 2.7714 [Kg/s]
Speed n 31494 [rpm] Density discharge 68.639 [Kg/m3]
Torque (1 stage) 72.93 [Nm] Pressure drop cooling system 17 [bar]
Efficiency (1stage) 0.5854 Turbine pressure inlet 56.10 [bar]
Pressure inlet 1.847 [bar] Temperature inlet gas hydrogen 213 [K]
Temperature inlet 21.44 [K] Turbine pressure outlet 38.76 [bar]
Density inlet 69.47 [Kg/m3] Pressure Chamber 32.75 [bar]

The PGD tool design results and their comparison with those obtained in [5] using PUMPA code developed by
NASA are provided in Table 2.

Table 2. Comparison between PGD and PUMPA results for the RL10A-3-3A engine.
Symbol PGD tool PUMPA
n [rpm] 31452 31494
Torque [Nm] 65.2 72.93
Power [KW] 2171 2405.2
Head ΔH [m] 5119 5138.32
Discharge density ρd 68.686 68.639
Outlet Diameter D2 [m] 0.173 0.1796
Exit Blade Height [m] 0.0062 0.0058
Discharge Temperature Td [K] 26.52 26.47

Moreover, the results provided by the PGD tool for the velocity triangles at both the inlet and outlet sections are
summarized in Table 3 and Fig.2.

Table 3. Velocity triangle

Symbol Outlet Inlet Section
Head Coefficient ψ 0.6
Flow Coefficient φ 0.076
Pheriferical velocity u [m/s] 289 133.78
Tangential component of absolute velocity cu [m/s] 236.5 89
Relative velocity w [m/s] 57 38.3
Tip blade angle β [°] 22.7 39.6
Flow angle α [°] 5.3 15.3
Diameter D [m] 0.176

Fig. 2. (a) Outlet velocity triangle; (b) Inlet velocity triangle.

618 Angelo Leto et al. / Energy Procedia 101 (2016) 614 – 621

4.2. VINCI engine

Vinci is a new-generation upper-stage cryogenic rocket engine for launch vehicles [6]. It is being developed by
Snecma and other European partners as part of a European Space Agency (ESA) program. Firing tests started in April
2005 on a test stand run by the German Aerospace Center (DLR).
The Vinci engine is a cryogenic expander cycle rocket engine, is bi-propellant, fed with liquid hydrogen and liquid
oxygen. Engine data available in literature for the VINCI rocket are summarized in Table 4 [7].

Table 4. Engine Data VINCI

Engine Data
Thrust 180 [KN] Turbine pressure inlet 180 [bar]
Fuel Flow Rate (LH2) 5.8 [Kg/s] Temperature inlet turbine gas hydrogen 240 [K]
Speed 90000 [rpm] Turbine pressure outlet 90 [bar]
Power 2800 [KW] Pressure chamber 60.8 [bar]
Centrifugal Pump Pressure discharge 225 [bar] Expansion ratio 240
Pressure drop cooling system 45 [bar] Nozzle exit diameter 2.2 [m]

The PGD tool design results and their comparison with the experimental data are provided in Table 5; once more,
fairly good agreement can be observed between the available data and the PGD results.

Table 5. Comparison between PGD results and experimental data

Symbol PGD tool VINCI
n [rpm] 89980 90000
Torque [Nm] 297.3 296
Power [KW] 2802 2800
Outlet Diameter D2 [m] 0.155 1.16

The PGD tool design results for velocity triangle at the inlet and outlet sections are summarized in Table 6.

Table 6. Velocity triangle

Symbol Outlet Inlet Section
Head Coefficient ψ 0.6
Flow Coefficient φ 0.077
Pheriferical velocity u [m/s] 730.4 299.5
Tangential component of absolute velocity cu [m/s] 589.4 222
Relative velocity w [m/s] 152 101
Tip blade angle β [°] 21.77 39.6
Flow angle α [°] 5.46 16.12
Diameter D [m] 0.155

5. Preliminary Design of a Centrifugal Pump for Methane Expander Rocket Engine

One further application of the PGD tool is given in this paragraph: it consists in the preliminary design of a turbo-
pump for the Methane Expander Cycle Rocket Engine (MECRE), similar to the LM10-MIRA, currently under
development within an international collaboration between the Italian company AVIO and the Russian company
KBKHA [8].
Since the available information for this methane - oxygen rocket is limited to the thrust, NASA's Rocket Propulsion
Analysis (RPA) [9] software was used to guess the chamber data, which are summarized in Table 7.
The turbopump designed for the LM10 is characterized by a low volumetric flow pump with a high pressure rise,
achieved with only one centrifugal stage.
 Angelo Leto et al. / Energy Procedia 101 (2016) 614 – 621 619

Table 7. Engine Methane Data RPA

Engine Data
Thrust 70 [kN]
Specific impulse 371.34 [s]
Total mass flow rate 19.21 [Kg/s]
Oxidizer mass flow rate 15.21 [Kg/s]
Pressure chamber 55 [bar]

The PGD tool design results for the MECRE engine are provided in both Table 8 and Table 9 and the design results
for the velocity triangle at the inlet and outlet sections are also shown in Fig. 3.

Table 8. PGD tool results

Result
ΔHead 3392 [m]
Speed n 44780 [rpm]
Torque 62.26 [Nm]
Pressure inlet 1.69 [bar]
Temperature inlet 114 [K]
Density inlet 418.93 [Kg/m3]
Pressure discharge 162 [bar]
Temperature discharge 121.6 [K]

Table 9. Velocity triangle

Symbol Outlet Inlet Section
Head Coefficient ψ 0.607
Flow Coefficient φ 0.074
Pheriferical velocity u [m/s] 236.36 106.36
Tangential component of absolute velocity cu [m/s] 197.5 78.12
Relative velocity w [m/s] 42.66 32.2
Tip blade angle β [°] 24.39 28.72
Flow angle α [°] 5.098 11.20
Diameter D [m] 0.097

Fig. 3. (a) Outlet velocity triangle; (b) Inlet velocity triangle.

6. Sensitivity Analysis

In this section we show that the PGD tool is capable of performing parametric studies, such as those needed for its
future inclusion in an optimization process. This capability is demonstrated in Fig. 4, which shows the variation of
head, efficiency and power required as a function of the mass flow rate for the first stage of the centrifugal pump of
the RL10A-3-3A engine. The rotational speed has been kept constant at 31452 [rpm].
620 Angelo Leto et al. / Energy Procedia 101 (2016) 614 – 621

In this figure, the Head is in [m], the mass flow rate in [Kg/s]; while power in [kW]. In particular, H was scaled by
a factor of 10.

Fig. 4. Characteristic Curve RL10A-3-3A

In the similar way, the characteristic curve of the MECRE engine is shown in Fig 5.

Fig. 5. Characteristic Curve MECRE

 Angelo Leto et al. / Energy Procedia 101 (2016) 614 – 621 621

7. Conclusion

Liquid hydrogen is the fuel that provides the best performance, so it is often used in launchers. It has the
disadvantage of a very low density, also has a very low boiling temperature, due to what are needed large tanks for
the storage and centrifugal pumps with a high number of revolutions. The sizing of a pump for hydrogen is very
complicated because the high number of revolutions creates a suction pressure drop that triggers the cavitation
phenomenon. Regarding the methane represents an innovative and alternative fuel to hydrogen, and is currently in
phase of study for LRE.
The PGD MatLab code developed in this work is a fast predicting methodology that is able to determine the global
geometry and performances of the turbo-pumps used in Liquid Rocket Engines. Model predictions for the RL10A-3-
3A and VINCI engines have been validated using the available experimental data and, for the RL10-3-3A engine, also
using the simulation results provided by the PUMPA code developed at NASA. For both engines, good agreement
has been found between the data available in the literature and the PGD simulations. The tool was then used to predict
the main geometrical parameters and performance of a Methane turbopump feed system.
Future developments include the design of the blade profile and the performance prediction of the gas turbines used
to drive the pumps.

References

[1] [Online] www.coolprop.org.

[2] Alexey J. Stepanoff. Centrifugal and Axial Flow Pumps. Wiley, 1948.
[3] Michael E Binder. RL10A-3-3A Rocket Engine Modelling Project. NASA Technical Memorandum 107318. 1997.
[4] Jorge R. Santiago. Evolution of the RL10 Liquid Rocket Engine for a new upperstage application. AIAA, ASME, SAE, and ASEE, Joint
Propulsion Conference and Exhibit, 32nd, Lake Buena Vista, FL.1996.
[5] Michael E Binder. A Transient Model of the RL 10A-3-3A Rocket Engine. NASA CR 195478. AIAA Conference Paper 95-2968. 1995.
[6] SAFRAN, Snecma. Space Propulsion VINCI. www.snecm.com
[7] Volvo Aero Internal Documentation. Utveckling av endimensionellt beräkningsprogram för turbinanalys. Program identification number: A5
NT 132 1310E VOLV 2044.
[8] http://www.caeconference.com/proceedings/abstract/abstract_08/avio_cira_cfx.html. [Online]
[9] Alexander Ponomarenko. RPA: Tool for Rocket Propulsion Analysis, 2015.
[10] H. Ohashi. Analytical and experimental study of dynamic characteristics of turbo pump. NASA TN D-4298. 1968.
[11] Wojciech Rostafinski. An analytical method for predicting the performance of centrifugal pumps during pressurized startup. NASA TN D-
4967. 1969.
[12] R. K. Hoshide, C. E. Nielson. Study of blade clearance effects on centrifugal pumps. NASA. CR-120815. 1972.
[13] Russell B. Keller, Jr. of Lewis. Liquid Rocket Engine Centrifugal Flow Turbopumps. NASASP-8109. 1973.
[14] Joseph P. Veres. Centrifugal and Axial Pump Design and Off-Design Performance Prediction. NASA Technical Memorandum 106745. 1994.
[15] Johann Friedrich Gülich. Centrifugal Pump. Springer-Verlag Berlin Heidelberg. 2010.
[16] Christopher E. Brennen. Hydrodynamics of Pumps. Concepts ETI, Inc. And Oxford University Press. 1994.
[17] A. E. Krach, A. M. Suttont. Another look at the practical and Theoretical Limits of an Expander Cycle, LOX/H2 Engine. AIAA 99-2473.
1999.