Journal of Marine Engineering & Technology

ISSN: 2046-4177 (Print) 2056-8487 (Online) Journal homepage: http://www.tandfonline.com/loi/tmar20

Simulation of large marine two-stroke diesel

engine operation during fire in the scavenging air
receiver
A G Livanos Graduate Student, G Theotokatos Research Engineer &
Professor N P Kyrtatos
To cite this article: A G Livanos Graduate Student, G Theotokatos Research Engineer &
Professor N P Kyrtatos (2003) Simulation of large marine two-stroke diesel engine operation
during fire in the scavenging air receiver, Journal of Marine Engineering & Technology, 2:2, 9-16
To link to this article: http://dx.doi.org/10.1080/20464177.2003.11020170

Published online: 01 Dec 2014.

Submit your article to this journal

Article views: 95

View related articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=tmar20
Download by: [109.98.161.141]

Date: 04 November 2015, At: 12:16

Simulation of large marine two-stroke diesel engine operation during fire in the scavenging air receiver

Simulation of large marine two-stroke

diesel engine operation during fire in
the scavenging air receiver

Downloaded by [109.98.161.141] at 12:16 04 November 2015

AG Livanos, Graduate Student, G Theotokatos, Research Engineer, and

Professor NP Kyrtatos, all of the Laboratory of Marine Engineering, National Technical
University of Athens, Greece

The operation of a large two-stroke marine diesel engine during fire in the scavenging air
receiver was investigated through simulation. A detailed performance prediction code for
reciprocating engines was used, extended with a model capable of simulating fires in
scavenging spaces. Simulation runs of several cases of fire were performed, including the
case of a severe fire which caused turbocharger compressor surging and flow reversal. The
case of engine slow-down during a severe fire was also examined revealing that, under
certain conditions, the engine cannot be stopped. By analysing the simulation results, an
insight into the effects of fire in the scavenging space on the operation of the engine and
turbocharger was obtained.

INTRODUCTION

ngine simulation codes are widely used during the

design, development and optimisation of the reciprocating internal combustion engine. Especially in
recent years, the increased complexity of engine
configurations in conjunction with the introduction of electronic systems for controlling the various engine parameters
necessitates the extensive use of detailed engine simulation
AUTHORS' BIOGRAPHIES
George Livanos received a diploma in naval architecture and
marine engineering from the National Technical University of
Athens (NTUA) in 2002. He is studying towards a PhD degree
in the field of extreme operation parameter engines.
Gerasimos Theotokatos graduated from the Department of
Mechanical Engineering, (NTUA) in 1994. He obtained a Doctor
of Engineering degree from NTUA in 2001 in the discipline of
marine engineering. Since 1996, he has been working on research
and development projects with the Laboratory of Marine
Engineering of NTUA.
Professor Nikolaos P Kyrtatos is Head of Division and Director
of Laboratory of Marine Engineering, (NTUA), Greece. He is
President of CIMAC-International Council on Combustion Engines
(2001-2004), and his other qualifications include a BSc from the
University of Newcastle upon Tyne (1975), and DIC, PhD,
Imperial College, University of London (1979).

No. A3

A3 Livanos

codes during the engine design procedure, in order to reduce the

engine development time and cost.
Such engine simulation codes can be used, not only for
matching the engine with its turbocharger, testing various engine controllers and possible design options, but also for investigating complex phenomena which may occur during engine
operation. Thus, engine simulation tools are suitable alternatives for evaluating engine operating conditions in cases where
experimental measurements would be difficult due to high cost
and safety risk.
In the case of fire in scavenging spaces, the results from
simulation can contribute to the better understanding of the
behaviour of the engine and its turbocharger under such conditions, and can further assist in the design of appropriate measures for minimising the adverse effects.

ENGINE SIMULATION MODELS

The unit modelled in this work is a large, slow-speed, ninecylinder, two-stroke, uniflow-scavenged marine diesel engine
that is directly coupled to the ships propeller. The engine
employs the constant pressure turbocharging system and is
equipped with three turbochargers. One air cooler is connected
after each compressor. In addition, electrically-driven blowers
are mounted on either end of the scavenging air receiver to allow
adequate scavenging at part load engine operation, (Fig 1).1
The scavenging air receiver is bolted on to the cylinder
blocks and communicates with the cylinders through circular
openings. Scavenge air exiting the air coolers enters the receiver
through openings located at the lower part of the receiver. Valve

Journal of Marine Engineering and Technology

9/29/03, 12:19 PM

Simulation of large marine two-stroke diesel engine operation during fire in the scavenging air receiver
housings are placed at the bottom of the receiver above the air
inlets from the coolers. Each valve housing has a number of
non-return valves (flap valves), to prevent backflow to the
cooler when the auxiliary blower is in operation, (Fig 2). 1

Downloaded by [109.98.161.141] at 12:16 04 November 2015

Fig 1:The motion of scavenge air (white arrows) and axhaust

gases (black arrows) through the engine1

Fig 2: The scavenge air and the valve housing, mounted in the
bottom of the receiver1
The basic characteristics of the engine are given in Table I,
whereas the engine configuration is shown in Fig 3.
For the prediction of engine performance under steady state
and transient conditions, a very detailed code has been developed and used for a number of years.2 The code is flexible,
allowing the simulation of a variety of engine configurations
including four-stroke, two-stroke, diesel, gasoline, natural gas,
turbocompound engines, etc. The code is of control volume
(filling and emptying) type but it can also estimate the onedimensional flow effects inside the engine manifolds, based on
a pseudo one-dimensional pipe model. An engine configuration
is built up using a number of basic engineering elements such

as flow receivers (cylinders, plenums), flow controllers (valves,

compressors, turbines, heat exchangers, pipes), mechanical
elements (crankshaft, shafts, gearboxes, clutches, shaft loads),
and controller elements (speed governor, PID controllers). A
turbocharged engine can be modelled as several flow receiver
elements (control volumes) interconnected by flow controller
elements. The outside environment is regarded as a fixed fluid
element (constant pressure, temperature and chemical composition). In addition, mechanical connections are set between the
cylinders and the crankshaft, which can be connected to an
external load via a gearbox and/or clutch. A speed governor
element can be used for the control of the fuel rack position
which, in turn, determines the amount of fuel injected into the
engine cylinders, in order to regulate the engine speed around a
desired set-point.
The overall engine configuration used for simulation is
shown in Fig 3. Specifically, the scavenging air receiver
is modelled as a constant volume plenum and the spaces
beneath the cylinders are modelled as variable volume
plenums.
The flapper valves, which are fitted on the inlet side of the
receiver, are modelled as variable-orifice, non-return valves.
The cylinder ports are modelled as variable-opening valves
controlled by the crank/piston geometry. Each air cooler is
modelled using a plenum to account for its volume as well as a
heat exchanger element to account for its thermal exchange
function.
For examining the behaviour of the engine and its turbocharger in cases where compressor surging occurs, a compressor model capable of predicting the compressor dynamic behaviour including cases of turbocharger compressor operation
beyond the surge line had previously been incorporated into
the code.3 By using this model, the code also provides morerealistic predictions for the compressor dynamics behaviour
No. of
cylinders
Bore
Stroke

Nine

Turbochargers

Three ABB VTR 714

900mm
2550mm

MCR
Boost pressure
at MCR
BMEP at MCR

41 162kW, 94rev/min
3.6bar

Compression 16.8
ratio

18bar

Table I : MAN B&W 9K90MC characteristics

Fig 3: Engine configuration for overall

simulation

10

A3 Livanos

Journal of Marine Engineering and Technology

10

9/29/03, 12:19 PM

No. A3

Simulation of large marine two-stroke diesel engine operation during fire in the scavenging air receiver
because the inertia of the air contained within the compressor
passages, and also the change of the compressor characteristics
from their steady-state form during transients, are taken into
account.4,5

Downloaded by [109.98.161.141] at 12:16 04 November 2015

MODEL OF FIRE IN SCAVENGING AIR

RECEIVER
Under normal conditions, the used lubricating oil and any
unburned fuel collecting in the under-piston scavenge space is
drained continuously through a drain pipe to a dedicated scavenge drain tank. Blockage of this pipe increases the risk of
scavenge fire due to accumulation of used lubricating oil and
sludge in the scavenge space. If flakes of burning or glowing
carbon deposits drop into the oil sludge at the bottom of the
scavenge air box, the sludge can be ignited and serious damage
can be done to the piston rod and the walls of the scavenge air
box.
Ignition of carbon deposits in the scavenge air box can be
caused by:1,6
1. Prolonged blow-by.
2. Slow combustion in the cylinder, owing to incorrect
atomisation, or misaligned fuel jets.
3. Blow-back through the scavenge air ports, owing to an
incorrectly adjusted exhaust cam disc or large resistance in the
exhaust system (back pressure).
As the scavenge fire propagates, the pressure and temperature of the scavenging air increase. The compressor is forced to
operate at higher pressure ratios and lower flow rates. Thus, the
compressor may not be able to support the high pressure field,
developed downstream, and compressor-surging with severe
flow fluctuations and reversals may occur. A detailed description of the compressor surging phenomenon can be found in
Greitzer,7 and Theotokatos and Kyrtatos.4
To evaluate the physical properties of the drained oil,
samples of scavenging air receiver sludge obtained from two
ships with engines similar to the one considered here were
analysed. The flashpoint was found to range from 120oC to
160oC and the heating value was an average of 10 090cal/g.
Since the composition of sludge in the scavenging air receiver
is variable, the burning of the sludge was considered using an
equivalent diesel-oil burning rate:

[ ]

t = T / 1000, where T is the temperature of fuel K

In the present work, a model was developed based on the description

of scavenge fire phenomenon presented in Wharton.6 Thus the fire in the
scavenge spaces was considered to be a heat release-rate function, which
is defined using three independent parameters (Fig 4):

eq.max ).
The maximum equivalent burning rate of sludge ( m
The duraction ( t1 ) of the initial phase of fire, where the
burning rate increases linearly.
The duration ( t 2 ) of the steady-state final phase of fire,
where the burning rate is constant.

Fig 4: Function of equivalent burning rate of oil sludge used in

the calculations

h sludge kJ is the heat of formation of oil sludge

kg

The burning rate profile observed in pool fires9 was assumed to

apply in the present case. The function of equivalent burning rate
of sludge used in the present study is shown diagrammatically in
Fig 4. An estimate of the overall sludge quantity which may
accumulate in the scavenging spaces of engines of such type and
size, was obtained from the engine manufacturer and ship operators. Thus the quantity of sludge was assumed to be around 150kg.
An estimation of fire duration was also made (seven to 10min),
based on discussion with ship operators who had experienced such
incidents. An average burning rate of oil sludge (0.25kg/s-0.50
kg/s) was calculated from the above assumptions and the value is
in good agreement with measured data from pool fires.9

equivalent kg is the equivalent burning rate of sludge

m
s

INITIAL CALIBRATION

sludge h sludge = m
equivalent h diesel , where:
m
sludge kg is the burning rate of oil sludge
m
s

h diesel kJ is the heat of formation of diesel fuel

kg
The heat of formation of the diesel fuel was taken from
Heywood:8

No. A3

A3 Livanos

4.184 1000

148.6

t2
t3
t4
1

t
+

+
0.0518 26.614
9
.
1063
246
.
97
143
.
74
32
.
329

t
2
3
4

h diesel =

Initially, steady-state simulation runs at various loads were

performed. The required input and validation data, including
the engine geometric data, the fuelling and injection timing
data, the heat release rates, the compressor and turbines maps
and the blower curve, were available from the engine and
turbocharger manufacturers.10
For initial validation of the simulation predictions for steady
state operation, the engine shop trial data were used. Engine general
performance data are also published by MAN B&W.11 Some
additional performance parameters for a similar engine were also

11

Journal of Marine Engineering and Technology

11

9/29/03, 12:20 PM

Downloaded by [109.98.161.141] at 12:16 04 November 2015

Simulation of large marine two-stroke diesel engine operation during fire in the scavenging air receiver

Fig 5: Response of engine during fire

available from measurements onboard a containership.12
After the steady-state runs, a case of engine transient operation
was simulated. The engine was considered to be operating at full
speed and, for this case, the ordered engine speed, the engine speed,
the propeller shaft torque, the fuel rack position and the engine boost
pressure were available.10 The measured ordered speed was given as
input in the simulation code. In addition, the engine load torque was
considered to be the propeller shaft torque measured in Kyrtatos et
al.10 The engine speed governor model was calibrated based on the

available measured data. The results that were derived using these
values in the simulation were found to be in good agreement with the
respective measured data.4
In previous studies of the authors,3 a model capable of
predicting the compressor dynamic behaviour including cases
of turbocharger compressor surging was presented. This model
was an extension of previous formulations,7,13 and was extensively validated using published experimental data13 for a
centrifugal compression system.

Fig 6: Predicted effect of maximum equivalent burning rate of fire

12

A3 Livanos

Journal of Marine Engineering and Technology

12

9/29/03, 12:20 PM

No. A3

Simulation of large marine two-stroke diesel engine operation during fire in the scavenging air receiver

Downloaded by [109.98.161.141] at 12:16 04 November 2015

FIRE IN THE SCAVENGING AIR RECEIVER:

PARAMETRIC STUDY
Having validated the simulation code against both steadystate and transient data, several cases of engine operation, with
fires in the engine scavenging air receiver, were examined. In
these cases, it was considered that initially the engine was
operating at 74.6rev/min and at 50% load.
Fire in only one cylinder was simulated, but the results showed
that it had only little effect on engine operation. In the results
presented here, fire was assumed to have propagated to all cylinders.
Firstly, a fire in the scavenging air receiver was considered,
with maximum equivalent burning rate value 0.25kg/s and t1 =5s,
t 2 =30s. In the initial phase of fire, the compressor operating
point approaches the surge limit (Fig 6). This is due to the fact that
fire increases the pressure of air in the scavenging air receiver
located downstream of the compressor. This forces the compressor to operate at higher pressure ratios, lower flow rates.
In the final phase of fire, the compressor operating point is
seen to move away from the surge limit, to higher rotational
speeds, pressure ratios and mass flow rates. This is due to
increased exhaust gas pressure and temperature, which are
caused by the fire. This increase of pressure and temperature,
which occurs two seconds after the fire, increases the energy
content of exhaust gases and thus the available turbine energy.
With respect to engine performance, the fire causes a reduction in engine rotational speed and torque, see Fig 5. This is due
to the higher temperature and changed composition of the
scavenging air, and reduced mass flow rate of fresh air at the
start of the fire. All these factors contributed to reduced trapped
air/fuel ratio and engine-developed torque which, in turn, resulted in the engine propeller speed drop, since the propeller
load remained constant. The governor detected the engine
rotational speed reduction, and increased the fuel mass rate.
After the quenching of the fire, there was a small crankshaft
speed overshoot.
The effect of maximum equivalent burning rate of sludge
in engine operation was further studied. The fire was
assumed to be developed in the overall space of the inlet
receiver with t1 =5s, t 2 =30s. As can be clearly seen in
Fig 6, an increase in the maximum equivalent burning rate
eq.max ) causes the movement of the compressor operating
(m
point to higher rotational speeds and pressure ratios. After
the quenching of the fire, the turbocharger returns to its
initial operating point. Thus, the operating point of compressor moves within a loop, which is expanding towards
higher rotational speeds in the compressor operating area
eq.max ) increases
as the maximum equivalent burning rate ( m
from 0.10kg/s to 0.30kg/s. Assuming a value of 0.30kg/s
for the maximum equivalent burning rate, the compressor
operating point approaches the compressor surge line and
the compressor speed reaches 10 200rev/min, which is the
turbocharger shaft speed at the maximum continuous rating
eq.max greater
(MCR) point of the engine. For values of m
than 0.30kg/s, compressor surging occurs. As shown in
Figs 5 and 6, the temperature and pressure of air in the
scavenging air receiver increase. This forces the operating
point of compressor to approach the surge limits, as was
previously explained. The temperature of gases in the
exhaust gas receiver also increases.

No. A3

A3 Livanos

Fig 7: Predicted effect of time parameter t1 on compressor

response
The effect of time parameters t1 , t 2 of the fire was also studied.
Time parameters are very significant and greatly influence the response of the turbocharger. When the duration of initial phase of fire
t1 decreases, the phenomenon of fire develops more rapidly. The
severity of the fire and the rapidly-developed pressure field in the
scavenge space force the compressor to suddenly approach its surging
limits (Fig 7). When the duration of the final stage t 2 decreases, the
locus of compressor operation point shrinks to a smaller trajectory
around the initial operating point (Fig 8).
Fig 8: Predicted effect of time parameter t2 on compressor
response

13

Journal of Marine Engineering and Technology

13

9/29/03, 12:20 PM

Simulation of large marine two-stroke diesel engine operation during fire in the scavenging air receiver

EFFECTS OF SEVERE FIRES

Downloaded by [109.98.161.141] at 12:16 04 November 2015

It was shown above that, if the maximum burning rate is high

and/or the duration of the initial phase of the fire is very small,
a severe fire commences and as a result compressor surging may
occur. It is important to note that a burning rate which will not
cause compressor surging with a certain duration of initial
phase, can nevertheless cause surging with a shorter duration of
the initial phase. Thus, in the engine configuration considered,
if a fire of burning rate 0.25kg/s commences in the scavenging
air receiver, compressor surging will not occur, unless the
duration of initial phase ( t1 ) is smaller than 2.5s (Fig 7).

Fig 9: Compressor surge during severe fire

eq.max =0.4kg/s, t1 =2.5s,
For the case of fire with m
t 2 =10s, the results obtained are presented in Figs 9, 10, 11
and 12. In Fig 9, the trajectory of the compressor operating
point on the compressor map shows that the compressor
exhibited surging. This was expected because, as was already explained, the excessive pressure rise in scavenging
air receiver, due to the fire, urges the compressor to operate
under higher pressure ratios. The compressor is not able to
support the increased pressure field which is developed
downstream, and a sudden backflow occurs (negative value
of air mass flow in Fig 10). The compressor mass flow rate,
scavenging receiver pressure, turbocharger shaft speed and
torque are also shown in Fig 10. Initially, due to the compressor surging, the scavenging receiver pressure was reduced.
This caused a reduction in the trapped air/fuel ratio and
hence in the engine-developed torque, which, in turn, resulted in the engine propeller speed drop since the propeller
load remained almost constant. In addition, due to the fire
and lack of air, the exhaust receiver temperature increased
(Fig 11). The engine governor stabilised the engine after the
transient (Fig 12). Due to the rapid changes of the compres-

14

A3 Livanos

sor impeller absorbed torque during surge, the net turbocharger shaft torque was instantly altered during the flow
reversals. Successive compressor surging introduces severe
transient torsional loading into the turbocharger shaft, which
could eventually result in the turbocharger shaft failure.

Fig 10: Engine parameters during severe fire

Some results of simulation of engine slow-down during
severe fires in the scavenging spaces are also presented here.
It was considered that a severe fire (meq.max = 0.50kg/s,
t1 =10s, t 2 =90s) commences in all scavenging spaces. It
was further considered that an amount of the overall oil
flammable sludge, which accumulates in scavenging spaces,
is carried with the scavenge air and burns together with the
injected fuel inside the cylinders. Such conditions may arise,
for example, if there is a liberal supply of oil, say due to the
loss of a piston skirt.

Fig 11: Increase exhaust gas receiver temperature

during severe fire (m = 0.40kg/s, t1 = 2.5s, t2 = 10s)

Journal of Marine Engineering and Technology

14

9/29/03, 12:21 PM

No. A3

Simulation of large marine two-stroke diesel engine operation during fire in the scavenging air receiver
of the accumulated oil sludge, the engine cannot stop, but
continues to operate at very low rotational speed. An increase of
carried amount of flammable material, increases the no-fuel
engine rotational speed. It is obvious that under such conditions,
although the fuel rack position may be set to zero, the engine
will continue to operate by burning the flammable oil sludge
carried with the scavenge air. The loss of engine controllability
is evident.

Downloaded by [109.98.161.141] at 12:16 04 November 2015

Fig 12: Crankshaft torque and rotational speed

during severe fire (m = 0.40kg/s, t1 = 2.5s, t2 = 10s)
During the fire, the engine rack position was ordered to be
reduced linearly from 50% to 0% in a period of 20s. The
schedule of rack position reduction was properly chosen in
order to avoid compressor surging during engine slow-down
from 50% load under normal conditions (without fire). Several
cases were considered with 5%, 15%, 25% and 50% of the
overall oil sludge carried over and burning in the cylinders.

Fig 14: Compressor response during emergency engine shut

down with fire

Fig 13: Compressor response during emergency engine shut

down without fire
The trajectory of compressor operating point without fire is
presented in Fig 13, and with fire in Fig 14. As can be seen in Fig
14 with the fire, although the rack position reduces, the compressor operating point approaches the surge line and extends
almost parallel to higher pressure ratios and rotational speeds.
At the end of the phenomenon, the compressor operating point
doubles back parallel to the surge line towards lower rotational
speeds. As shown in Fig 15, if the amount of flammable material
carried-over into the combustion chamber is greater than 15%

No. A3

A3 Livanos

Fig 15: Crankshaft rotational speed during emergency shut

down for several percentages of carried-over inlet sludge
deposits

CONCLUSIONS
Fire in the scavenging air receiver affects greatly the response of both engine and turbocharger. An increase in the rate
of burning of oil sludge leads the compressor to operate at
higher rotational speeds and pressure ratios. If the sludge
burning rate is too high, compressor surging may occur. If the
duration of the initial phase of fire is short, the fire is more severe
and this may also cause surge. The duration of the final phase of

15

Journal of Marine Engineering and Technology

15

9/29/03, 12:21 PM

Simulation of large marine two-stroke diesel engine operation during fire in the scavenging air receiver
fire mainly affects the compressor rotational speed. As the
duration increases, the compressor rotational speed increases,
too. During the fire, the temperature in the exhaust gas receiver
is increasing. Sometimes, severe prolonged fires in scavenging
spaces, with sludge oil carried over into the combustion chamber, may result in loss of engine controllability since the fuel
cut-off cannot lead to engine stop.
The use of detailed simulation proves to be an indispensable
tool to gain insight into such complex engine operation phenomena.

ACKNOWLEDGEMENTS
The authors wish to thank MAN B&W Diesel A/S (Copenhagen) for useful discussions and the provision of engine test
data for initial code-validation purposes. The authors also thank
Kristen Navigation Inc (Athens) and Minerva Maritime Inc
(Athens) for discussion and for providing samples of scavenge
box sludge for analysis.

Downloaded by [109.98.161.141] at 12:16 04 November 2015

REFERENCES
1. MAN B&W. Instructions for 50-90MC Engines, Edition
8A, Air system of engine, pp 360-363.
2. Kyrtatos NP, (Ed), 2000. MOtor THERmodynamics Ver
1.4 USER MANUAL, NTUA/LME.
3. Theotokatos G, and Kyrtatos NP, Diesel Engine Transient Operation with Turbocharger Compressor Surging, SAE
Technical Paper No. 2001-01-1241, 2001.
4. Theotokatos G, and Kyrtatos NP. Analysis of a large twostroke marine diesel engine transient behavior during compressor surging, 7th International Conference on Turbochargers and
Turbocharging, London, 14-15 May 2002, IMechE Conf Trans

16

A3 Livanos

C602/005/2002
5. Theotokatos G., and Kyrtatos NP, 2002. Investigation of
a Large High Speed Diesel Engine Transient Behavior including Compressor Surging and Emergency Shutdown, ASME
Transactions Journal for Engineering for Gas Turbine and
Power, Vol 125 No 2, April 2003, pp 580-589.
6. Wharton AJ, 1994. Diesel Engines, ButterworthHeinemann Ltd, Oxford, UK
7. Greitzer EM. Surge and Rotating Stall in Axial Flow
Compressors. Part I: Theoretical Compression System Model,
ASME Journal of Eng. for power, Vol 98, pp 190-198, 1976
8. Heywood JB, 1988. Internal Combustion Engines Fundamentals. McGraw Hill
9. Chatris JM, Quintela J, Folch J, Planas E, Arnaldos J, and
Casal J. Experimental Study of Burning Rate in Hydrocarbon
Pool Fires, Combustion and Flame Journal, Vol 126, pp 13731383, 2001.
10. Kyrtatos NP, Theotokatos G, Xiros N, Marec K, and
Duge R. Transient Operation of Large-bore Two-stroke Marine Diesel Engine Powerplants: Measurements and
Simulations, 23rd CIMAC World Congress, Hamburg, Vol 4,
pp 1237-1250, 2001.
11. MAN B&W. K90MC Mk6 Project guide, 4th edition,
November 1997.
12. Kyrtatos _P, Theotokatos G, and Xiros N. Main Engine
Control for Heavy Weather Conditions: The ACME Project,
ISME 6th International Symposium on Marine Engineering,
Tokyo, pp 457-464, 2000.
13. Fink DA, Cumpsty NA, and Greitzer EM. Surge Dynamics in a Free Spool centrifugal Compressor System, ASME
Journal of Turbomachinery, Vol 114, pp 321-332, 1992.

Journal of Marine Engineering and Technology

16

9/29/03, 12:21 PM

No. A3