Rocket engines:

Turbomachinery
Hans Mårtensson
Sonny Andersson
Stefan Trollheden
Staffan Brodin

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Where is the meat ?
• No factor of 5-10 on Isp…
• Systems are proven in operation
– Flying commercial load to orbit
– Have flown to the moon
• ”Industrial” objectives optimization
– Cost/Resource consumption per payload delivery
– Safety
– Operability
– Time lines

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Contents
• Introduction to how the turbopumps relate
to engine type
• Example of a first pass preliminary design
of turbopumps for a rocket engine
• Discussion on detailed design issues, and
how modern tools affect design

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Cycle objectives
Fn is the thrust
Fn
Isp = g is the gravitation constant
g ⋅ m& m& is propellant mass flow
v jet is nozzle exit velocity
Fn = m& v jet
• Deliver a maximum of thrust using a minimum of fuel
– A rockets may take 10 tons to orbit at a 1000 ton take-off weight,
the absolute majority is fuel.
• The thrust is dependent on pressure ratio to increase jet
velocity and allow high nozzle area ratio
– Higher pressures give better Isp but more complex machines

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Turbomachine design goals
• Objective 1: Deliver the reactants to the
thrust chamber at specified pressure.
• Optimize the efficiency
• Lower weight
• Allow robust operation
• Allow stable manufacturing process
• Minimize cost

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 3 principal engine types

NASA SP-8107

• GG entry pressure is above chamber pressure, sufficient to overcome ”plumbing”,

and injection losses
• Stage combustion and expander exit pressures are above chamber pressure, leading
to high pressures in the turbine.
• The expander is simplified by not needing pre-burner, but adapted to hydrogen due to
cooling properties. Also very sensitive to turbomachery efficiency due to limited
heating capacity.

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Range of applicability of different
cycles
1000 psia=70bar

• Ranges of applicability of different cycles

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Examples of rocket engines
Engine Turbine inlet Pressure Flow
pressure ratio
Gas Vulcain 100 bar high low flow
generator F1,Atlas, pressure
Titan ratio

Stage SSME 400 bar low pressure high flow

combustion RD-170 ratio

Expander Vinci 200 bar low pressure high flow

RL-10 ratio turbine

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 SNECMA Vulcain2 engine
Gas generator cycle

• Thrust weight ratio

– Vulcain 2 T/W=60
– Modern fighter engine T/W=10
(Source: fichier technique available at • Note fuel flow 320 kg/s to payload ~10 tons ->
www.snecma.com) requirement for rapid start in seconds
– Fighter engines go to full power in minutes (in the
extreme)

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 VINCI LH2 turbopump
(SNECMA/VOLVO)

Low blade height in

turbine
Real gas effects
2.8 MW D=120mm

• 2 stage pump + inducer

• 1 stage turbine with low blade height and no OGV
– Note thick turbine disk indicating high tip speed
– 240 K at turbine entry
– No exit guide vane from turbine
Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,
 Vulcain 2 LOX turbo pump
(AVIO/VAC)

• Single stage pump w. inducer

• 2 stage overhung turbine
• Separation of L2-Hot gas
Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,
 SNECMA VINCI engine
expander cycle

• Turbines in series use H2 heated at combustor jacket

• Chamber equilibrium temp. ~3500K
• No pre-burner needed
Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,
 Fuels
• The most important difference is density
– Density – pump head and power: example at 100 bar
Density * Pump head at 100 bar
[kg/m3] [m]
LH2 75 13600
RP-1 (Kerosene) 810 1260
LO2 1200 849

• Other important properties

– Lubricating and cooling properties
– Soot/Particles – erosion, wear, clogging
• Hydrocarbons (Methane, Kerosene…)
– Material compatibility,
• H2 embrittlement,
• O2 Oxidization
Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,
 Sample Turbine gas compositions
LH2/LO2 Fuel Rich RP-1/LO2 Fuel lean
Mole fractions at eq. Mole fractions at eq.
O/F=1.023, 100bar O/F=37, 1000K, 100bar
*H2 0.87110 *CO2 0.06007
H2O 0.12890 H2O 0.05857
*O2 0.88137

• The LH2/LO2 is most benign

RP-1/LO2 Fuel rich
O/F=0.13
although with concerns for
MOLE FRACTIONS hydrogen embrittlement
100bar 500bar • High partial pressures O2 may
CH4 0.29388 0.33322 cause oxiization of materials
*CO 0.00736 0.00600
*CO2 0.00610 0.00664 • Soot/graphite carbon for fuel rich
C2H6 0.00012 0.00040 RP-1 may clog/erode the turbine
*H2 0.12417 0.07918
H2O 0.07177 0.07663
C(gr) 0.49660 0.49792
Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,
 Cavitation-NPSH
• Erosion and loss of pump head may result
• NPSH – Net Positive Suction Head
– Measures Pump inlet pressure margin to vapor
pressure
– The necessary NPSH depends on purity of the liquid
and inducer
• Cavitation is a dimensioning criterion on the
pump
– Tank pressure, separate boosters, inducer design
and design speed are controlling devices
– The trade goes to weight and complexity of the entire
system

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Cycle effects - turbine design

Cycle
Turbopump Turbine Components

Efficiency Manufacturing
Isp- what system ?
Vibration control processes
Merging all data Loading

• Each level propagates a requirement to the

lower system level.
• The requirement must be made based on
estimates of what is possible
– Technology demonstrators
– Results of earlier experience
Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,
 Development issues
• Concept/preliminary design – assess potential
– Trade Cost-Risk-Performance (TRL-5/6 )
– With engine designer
• Draw on experience
• Anticipate technology advances
• Design – Realize solution
– Analytical predictability
– Draw on experience
• Verification (Analytical – Test)
– Satisfaction of spec
– Few hardwares
• Margin and experience, introduce robust design concepts

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 First pass design
• Consider a GG cycle
Propellant mass flow rate 300 kg/s at O/F=6, Pchamber 100 bar, Pinlet 1 bar

Pump delivery data

Pexit Massflow Pexit Mass flow
Pinlet LH2 LH2 LO2 LO2

1 bar 132 bar 42.9 kg/s 120 257.1 kg/s

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Efficiency potential and overall
sizing
• Yellow dot indicates
solutiond
ns-ds diagram with from NASA SP-8109

Lower rim Specific Ω Qin

speed
speed ns =
(gH ) 4
3

Specific D(gH )
1
4
diameter ds =
Qin

Tip speed limitation forces lower efficiency potential if

single stage H2 is used
Technology development may off-set this

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 First pass candidate solutions
for the pump
(No NPSH concession made, No Bearing DN or rotordynamics consideration)

Tip Tip speed Efficiency Power

Shaft diameter Utip potential needed
speed D [m/s] η P
N [mm] [-] [kW]
[RPM]
LO2 22980 132 158 0.8 3500
LH2 1-stage 122100 104 665 0.8 10200
LH2 1-stage reduced 106140 83 462 0.7 11700
ns,ds
LH2 2-stage 72600 124 469 0.8 10200

• Select best solution within experience base for

product development
• Or push technology by selecting aggressive
single stage solution
Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,
 Turbine
• Deliver power to pump at given speed
– Generally over op. range
• For a GG cycle the exit pressure is low
– Trade design pressure ratio with thrust that can be achieved at exit
pressure. Minimum is what is needed not to recirculate hot gas in
engine bay.
• Use 1000K inlet temperature as upper limit for uncooled blades and
vanes
– All current machines are uncooled
• Optimize with respect to
– Weight
– Efficiency
– Cost
– Industrial issues
• Quality, Safety
• Commonality, Experience, Tooling
– Structural

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 First glance
Green LH2
Red LO2
U/C0 U/C0
0.8 0.8
Ptin Ttin Shaft Power Pout Massflow* Efficiency
[bar] [K] speed [MW] [bar] [kg/s] *
[RPM] [-]

LH2 100 1000 72600 10.2 10 5.27 55

LO2 100 1000 22980 3.5 10 2.6 38

γ −1
 
 P γ 
C 0 = 2C p (T00 − Tis ) = 2C p T00 1 − out  =2655m/s -> LOX U/C0 ~0.12, LH2~0.2
 P00
 

• The LO2 needs to deliver 1/3 of the LH2 power, savings

on cost to be done at the LO2 hence pick 1-stage
• The LH2 will be based on 2 stage estimates
Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,
 Picking out the stages in ns-ds
diagram
• Use diagram
again to start
off iteration

2 stg 1 stg
LO2 1 stg 2 stg
LH2

Turbine Balje
Figure 11 ns-ds diagram for turbines with selected points for the LO2 red and LH2 green
options

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Define a turbine layout
• Select and work in pitchline mode
– Using simple loss model + experience
– Evaluate limits
Tip clearances critical due to
blade height
Axial clearance-Chord-Length-
weight
Blade count – loading - OGV-
cost
Flow areas and angles – LE/TE
thickness - sensitivity to
tolerances
Reaction-Axial thrust-Efficiency
Off-design trades

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Early structural allowances
• Select INCO718
– Stage 2 runs a bit cooler
than stage 1
”Specific root stress”

AN = π r − r
2
( 2
O I
2
)⋅ Ω 2

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 2 stage LH2 Layout view

AN2 on stg 2 is marginal at best

Options: Select new material, Lower
speed or push the limit by good
detailed design
Angles and areas within bounds avoiding later problems with tolerances
Margin on efficiency
Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,
 Ready to start detailed design
• Trends
– Aspect ratio: classical h/c>2 in order to limit sec loss.
Now we have allowed 0.8.
– Blade loading Zweifel nmber up >1, 0.8 typical
number in litterature. Saves cost.
– Blisk designs avoid expensive firtree and dead rim

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Now start 3D CFD

3D multi-stage is standard
Is unsteady important to
performance ?
Modest changes to the
geometry is possible
Pick up major defects of the
design

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Secondary flow system
• Ingress
– from main gas path may change the
state of the cavity
• Axial thrust balance
– Bearing stiffness and life vary with axial
load. Total axial load must not change
sign
– Governed by reaction this varies with
op. point.
• Heat loading
– Hot start followed by adaption to steady
state at diffeent rates causes strain
– Shut down processes causes a cold
shock
– Hot geometry causes shift in
performance - clearances
• Purge
– Oxygen pumps may inverts purge
pushing hot gas through cavity
– LH2 is extremely cold with very small
sealing gaps

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Blade and vane active cooling
• 25% efficiency gives ~1300K Tin
– Save 30% of the flow used to drive the turbine
• 50% efficiency could allow ~2000K
– But thermal strains will increase (30K-2000K)
• Hotter and heavier but definitely possible on the
H2 turbine
– Cooling with Oxygen is off for fuel rich pre-burne, so
we need plumbing mixing in hydrogen for the LOX
turbine.
– Complexity and expense in blades will make this
difficult

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Design optimisation
• Push CFD and FE upstream allowing them to
affect earlier decisions
– Adapt to new design points rapidly
– Automated area ruling variation is key to matching

At 10% reaction
layout,
optimised with
nozzle

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Robust design multi-point
• Changes to the cycle and operability requires
off-design operation, often difficult in early
design

Refence design point

Worst case off design point

Solutions with almost flat efficiency over
the envelope can be found

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Robust design
• Impact on cost in manufacturing
• Low scatter in preformance parameters
• Avoids traditional inherited conservatism

parameter dimension tolerance

leading edge 0.209 mm ± 0.1

radius
trailing edge radius 0.157 mm ± 0.1

chord length 11.3 mm ± 0.2

stagger angle 1.786 ° ±2
fillet radius 0.55 mm ± 0.15
blade height 12.5 mm ± 0.2
blade thickness 4.8 mm ± 0.2
Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,
 Mechanics
• Many problems in engine development relate to turbopumps
– "History of Liquid Rocket Engine Development in the United
States1955-1980"
• Transients
– Thick walled components designed for burst (pressure vessels,disk …)
have large thermal inertia.

Order of magnitude comparison of cycles for turbomachines

Operating time
Number of cycles

Power generation gas turbine 1E5-1E6

/ Steam turbine

Commercial jet aero engine 10000-50000 30000-100000 hours

Fighter engine 1000-10000 1000-10000 hours

Rocket 4-10 1-10 hours

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Vibration
• A large share of late hard and expensive
developments problems, all programs
– Forced response, flutter,rotor dynamic or
unidentified
• Testing
– Performed in scale rig, TP rig and at engine
level.
– Red lines monitored on each individual

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Unsteady flow forcing
• Animation The unsteady force can
be >30% of the steady
loading
In our example ~100N
oscillating per blade

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Blade Vibration
Forced response
• Avoid known and 1st order crossings
imperative
• Higher pressures require greater
damping
– Test
Campbell diagram Disk modes Campbell diagram Blade modes
7000 12 EO (S2-S1)
40000
1 EO 19 EO (IS1)
6000
2 EO
35000 31 EO (S2)
5000 3 EO
38 EO (2S1)
4 EO 30000
4000 57 EO (S1)
1ND
Hz

3000 2ND
Hz
25000 60 EO (R1)
3ND 1F
2000
4ND 20000 1T
1000 60 % 1E
80% 15000
0 2F
800 1000 1200 1400 1600 1800 10000 2T
RPS 800 1000 1200 1400 1600 1800 2E
60 %
RPS
80%

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Verify HCF margin
Amplitude
stress

load

Haigh
• In the Haigh diagram HCF diagram
limits the low AN2 (P/A)
will pay off again by
allowing higher vibration
stress amplitude
Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,
 Flutter - Aeroelastic instability
<=Unstable
3.5

Axial
Tangential
Aerodynamic amplification factor

2.5 Torsion

1.5

Critical reduced
frequencies
0.5

0.0 0.2 0.4 0.6 0.8 1.0 1.2

-0.5
Stable =>

π ⋅ Chord ⋅ f
Reduced frequency

In our example: k=
LO2 1F k∼0.08 -> unstable Vrel
LH2 1F k∼0.2 -> marginally stable
1F, 1T probably subcritical
NEED TO ENSURE SUFFICIENT
DAMPING
Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,
 Machine hardware
• Manufacture
– Looks easier on drawing !

10 mm !

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,

 Conclusion
• The turbomachines are key components
that will keep improving gradually.
• New automated methods come into use
improve the designs
• Analytical verification

Hans Mårtensson, VKI-RTO March 2007 Rocket engines: turbomachinery,