Bai Giang MayTau Michigan

Informal Course Notes for
INTRODUCTION TO
MARINE ENGINEERING
by
Professor Emeritus Michael G. Parsons, FSNAME
Version: September 2015
Department of Naval Architecture and Marine Engineering

University of Michigan
Ann Arbor, MI 48109 – 2145
Michael G. Parsons
© -------------------------- 2015
All Rights Reserved
– ii –
CONTENTS
pages
Foreword ……………………………………………….……………………………..… v
1. Introduction …………………………………………………………...……….. 1-1 to 1-8
2. Review of Thermodynamics ……………...…….…………………………..… 2-1 to 2-48
3. Marine Fuels …………………...……………………………………………… 3-1 to 3-16
4. Reciprocating Prime Movers – Diesel and Natural Gas…………………….… 4-1 to 4-36
5. Gas Turbine Prime Movers …………….……………………………...……… 5-1 to 5-20
6. Steam Turbine Plants ………….………………………………….…...……… 6-1 to 6-30
7. Turbine Theory ………………...…………………………………...………… 7-1 to 7-16
8. Electrical Systems and Integrated/Hybrid Plants ……………………………… 8-1 to 8-26
9. Shafting and Components ………………………..…………………………… 9-1 to 9-14
10. Gearing, Clutches and Flexible Elements ….……………………………… 10-1 to 10-20
11. Engine-Propulsor Matching ………………………………...…………….… 11-1 to 11-42
12. Machinery Vibration Excitation and Mitigation …………………...………. 12-1 to 12-32
13. Design and Analysis of Fluid Systems ………………………………………13-1 to 13-34
14. Heat Transfer and Heat Exchangers ……………………………………...… 14-1 to 14-22
APPENDICES
A. Extracts from Crane Company Technical Paper No. 410M …………….…………… 4 pp.
B. SI Units: Thermodynamic Tables ………………………………………...………… 18 pp.
– iii –
- iv -
Foreword
This set of notes was initially developed in 2003 to provide an introduction to marine
engineering for students in the Naval Architecture and Marine Engineering BSE program at the
University of Michigan. The course NA331 Marine Engineering I is a required, junior-level
course that forms a sequence with NA332 Marine Electrical Engineering. NA331 follows
prerequisite course ME 230 Thermodynamics I, which covers thermodynamics through the
Second Law and simple cycles. It has as a co-requisite NA320 Marine Hydrodynamics I, which
provides a first course in fluid mechanics. The goal is a reasonable understanding of practical
marine engineering and a foundation in the relevant fundamental mechanics.
With the extensive use of Integrated Electric Plants (IEP) in recent years and the evolution
since 2000 of LNG as a general ship fuel, not just boil-off from cargo, the notes were extended in
2013 to include treatment of these topics as well. To enable to notes to also be used as a text in a
Masters level introduction to marine engineering, brief introduction to the marine machinery
vibration excitation and mitigation was added for the same reason. These extensions were
motivated by the additional use of these notes in NAME578 Marine Engineering in the MEng
program at the University of British Columbia in the fall 2013 and 2014.
These courses use the auxiliary text Woud, H. K., and Stapersma, D., Design of Propulsion
and Electrical Power Generation Systems, The Institute of Marine Engineering, Science, and
Technology, London, 2002, and the supplemental reference text Harrington, R. L. (ed.), Marine
Engineering, SNAME, Jersey City, NJ, 1992. The courses are primarily defined, however, by
these course notes.
The recommended use of these notes is as follows: undergraduate introduction to
mechanical marine engineering (e.g. UM NA331): Chapters 1, 2, 3, 4, 5, 6, 7, 9, 10, 11, 13, 14.
graduate introduction to marine engineering (e.g. UBC NAME578): all Chapters.
Thanks are extended to the students who tolerated the real-time development and revision
of the course notes when they were first used in these courses.
Mike Parsons
Newport, Oregon
March 2015
-v-
-vi-
Introduction to Marine Engineering Notes
Chapter 1 Introduction
Marine engineering is the branch of engineering that deals with the synthesis, design,
analysis, and control of mechanical and electrical propulsion and auxiliary systems used onboard
ships, offshore platforms, and marine craft. This is the definition of marine engineering used in
relation to university engineering degrees accredited by the Accreditation Bureau for
Engineering and Technology (ABET) as required for licensing as a Professional Engineer (P.E.)
within the United States.
The naval architect or marine engineer today will generally not be involved in the detailed
design of the components utilized in these plants – that is usually the task of the mechanical
engineers and/or the electrical engineers employed by the equipment vendor. The primary
exception is when the marine engineer must get involved in maintenance planning and the
detailed assessment of problems and potential solutions. The marine engineer must, however,
understand the operating principles, specification, selection, operating limitations, and control
strategies for these components so that effective systems can be developed.
In some contexts such as academies and industry schools, “marine engineering” also
includes the preparation of officers and crew to operate marine systems. In parts of Europe in
particular, marine engineering can also include the detailed design of marine unique components,
such as low-speed diesel engines, as well as their use and integration into marine systems.
The following marine engineering systems must be provided onboard ships, offshore
platforms, and marine craft [1]:
• propulsion including the propulsor and emissions control
• maneuvering and motions control devices and systems
• ship service electrical power generation and distribution system
• cargo, ballast water and bilge water conditioning, treatment and handling systems
• fresh water, grey water and black waste systems
• heating, ventilation, and air conditioning (HVAC) systems
• other environmental control systems
• damage control and fire detection and fighting systems
• hydraulic systems
• compressed air systems
• anchoring and mooring systems
• mission related systems for combat, oil drilling, dredging, cable laying, etc.
1-1
A major part of the design of larger vessels involves the development and design of the main
propulsion system that must include a large number of components and auxiliary systems to
support to basic operation of the propulsion plant. These might include the:
• prime mover and its energy supply
• shafting and bearings
• power conditioning and transmission: electrical or mechanical
• propulsor
• propulsion control system
• fuel conditioning and transfer systems
• prime mover cooling systems
• lube oil conditioning and transfer systems
• control and starting air systems
• emissions control systems
• propulsion machinery monitoring systems
The marine engineer also leads or becomes involved in the related structural design, foundation
support, operational dynamics, vibration, shock, and control of these systems.
There are a wide variety of options for the propulsion of marine vessels and craft. These
options involve choices relative to the basic source of the energy, the type of prime mover, the
means used to transmit and condition the power, and the type of propulsor [2, 3]. The primary
options for main propulsion of larger vessels are outlined in Fig. 1.1. The more common choices
are shown in the bold lines; less common and rare choices today are shown in normal weight
lines. The most appropriate choice varies, of course, with vessel mission, size, speed and
business case. Most large commercial ships today utilize diesel prime movers with a low-speed
diesel prime mover directly connected to a fixed-pitch propeller the most common. Other
vessels utilize medium-speed diesel prime mover(s) with mechanical gearing and a controllable-
reversible pitch propeller or with reversible mechanical gearing and a fixed-pitch propeller.
Electrical transmission systems are now used extensively – particularly military vessels,
auxiliary/support vessels and cruise ships. Larger military ships are primarily either
nuclear/steam propelled with mechanical gearing and fixed-pitch propellers or gas turbine
powered with mechanical gearing and controllable-pitch propellers. Gas turbines are used
occasionally in large cruise ships. Smaller high-speed craft use high-speed diesels or gas
turbines with various types of transmission and propulsors. Diesel submarines, Remotely
Operated Vehicles (ROV’s) and Autonomous Underwater Vehicles (AUV’s) use battery power
for submerged operations.
1-2
energy prime transmission/

propulsor
source mover energy conversion
low-speed direct
diesel connection propeller
med/high- CRP
fossil speed I.C. mechanical propeller
fuels engine gearing
residual, gas water
intermediate, turbine jet
diesel, LNG,
marine gas oil electrical
boiler/ trans/motor steerable
steam plant thrusters
generator,
cabling,
fissionable reactor/ power electronics, vertical axis
material (U) steam plant motor, etc. propellers
air
screws
hydrogen fuel cell
Key:
more common
electrical less common, rare
source battery
submarines/ROV’s/AUV's/hybrid plants
wind sails
I.C. = internal combustion

AUV = autonomous underwater vehicle ROV = remotely operated vehicle
FPP = fixed-pitch propeller CRP = controllable-reversible-pitch propeller
Figure 1.1 Most Common Marine Propulsion Alternatives
Figure 1.1 shows the various single concept propulsion plant options that are in use. When
there is (1) an economic justification to recover waste energy to provide additional propulsion
1-3
power or propulsion efficiency or (2) the need for a high efficiency cruise operating mode and a
less efficient, high power sprint mode, a combined plant is often used. For example, marine gas
turbine plants are used with waste (exhaust) heat steam boilers and smaller steam turbines to
provide additional propulsion power from the same fuel consumption. These combined plants
are identified by acronyms beginning with CO for combined and then segments for the two types
of prime movers used for propulsion. Thus, a COmbined Gas turbine And Steam turbine plant,
as described above, would be identified as a COGAS plant. If both the prime movers are in
operation together at full power or over a significant operating range, the conjunction And is
used. If there is a handoff at some intermediate power and both prime movers are not normally
used together, the conjunction Or is used. A small military vessel might use a small diesel
prime mover, with its high efficiency, for the cruise condition (say less than 15 knots) or a large
gas turbine, with its high power to weight ratio, for the sprint condition (say up to 30 knots).
This design would be identified as a CODOG plant with either the diesel or the gas turbine used
to propel the ship at any particular time. A recent CODED (COmbined Diesel Electric and
Diesel) plant for a high-speed ferry is shown in Fig. 1.2. The generators supply the pod
propulsor, the bow thrusters, and the ship service electrical load.
Figure 1.2 Schematic of CODED Propulsion Plant for High-Speed Ferry
1-4
When studying marine engineering, it is important to keep in mind the factors that need to
be understood and often quantified in marine engineering design and tradeoff studies. These are
the factors that provide the basis for the many design choices required during the plant’s design.
The principal considerations often include the following:
• Life Cycle Costs. Commercial, public and military design decisions are heavily
affected by the life cycle cost of the options. This must consider the installed costs or
capital expenses (CAPEX), operating costs (OPEX) including maintenance, and the
contingency or risk costs related to breakdowns, etc. The time value of money must be
properly considered [4] in this assessment.
• Reliability. The sea is a very severe environment and vessels are often far from
external support. Propulsion and maneuvering are usually critical to survival. There is,
therefore, a need for a higher level of conservatism in marine plants and equipment
compared to land based use and factory testing.
• Maintainability. Given the importance of the marine engineering systems, they must
be able to be maintained efficiently throughout the life of the vessel. Much of this
equipment cannot be exchanged for a replacement for off-vessel maintenance.
Considerations include time between overhauls, preventative maintenance requirements,
the consequence of failures, etc.
• Weight and Centers. The weight of machinery, including the required endurance fuel
and other fluids, are important to the overall ship design. Some designs, such as ore
carriers, are weight limited and machinery weight directly affects mission and/or vessel
size.
• Volume and Arrangement. The required volume of the machinery spaces, including
their shape and location, is important to the overall ship design. Some designs, such as
cruise liners, military vessels, and containerships, are volume limited and machinery
volume and required arrangement directly affects mission and/or vessel size.
• Specific Fuel Rate. A specific fuel rate is an expression of thermodynamic efficiency.

Specific Fuel Rate (sfr) or Specific Fuel Consumption (sfc) is inversely related to
1-5
thermodynamic efficiency. It is the fuel used per unit of power times unit of time; e.g.
kg/kW h, t/kW h, or lbf/hp h. Low-speed diesels offer the best specific fuel rate available
today, but fuel cells will threaten this advantage. Nuclear steam plants and simple cycle
aviation gas turbine have the highest.
• Fuel Type. The type of fuel chosen affects the fuel unit cost, fuel volume, fuel
weight, and fuel treatment required. Total fuel cost per unit of power and time is the
specific fuel rate times the unit cost of the fuel; e.g. dollars/kW h. Steam boilers burn the
lowest grade, cheapest residual fuel, but this is being challenged by Liquefied Natural
Gas (LNG) today. Nuclear plants require the most expensive fuel, but one reactor core
load may last the lifetime of the ship. The range of cost of marine fuels for use in gas
turbines to boilers is more than a factor of 2 depending upon type [5].
• Potential for Air and Water Pollution. The choice of machinery type and model can
have a significant effect on the resulting air emissions and water pollution risk of a
vessel. As increasingly severe restrictions are applied by the Environmental Protection
Agency (EPA), state agencies, or International Maritime Organization (IMO), machinery
designs and/or choices will have to change. The recent move to natural gas fuel for air
emissions reduction has been the most dramatic change.
• Survivability. The type, redundancy, and arrangement of the ship’s machinery affect
the ability of the ship to survive damage from collisions, grounding, or battle.
• Reversing Capability. Some types of machinery are reversible by design; e.g., steam
turbines and low-speed diesels. Other types of prime movers are not reversible, thus,
requiring some form of ship reversing capability through electrical transmission,
reversing mechanical reduction gears, or controllable reversible pitch (thrust) propulsors.
• Ease of Automation. The type of machinery directly affects the ease of its automation
to achieve effective bridge control and reduced manning. In this regard, gas turbines are
probably the most attractive and nuclear steam plants the least.
• Compatibility with Auxiliary Requirements. The prime movers or prime mover waste
heat can also serve other ship needs for power, heat, etc. The need for large amounts of
1-6
cargo handling power, maneuvering or mission power may directly affect the choice of
machinery type. The use of an electric transmission (generators, power transmission and
conversion, motors, and gearing if needed) can support the move to an Integrated Electric
Plant (IEP) or “all electric ship” using central power generation for all uses [6]. This
approach also makes full installed prime mover power available for electric weapons or
other intermittent mission uses.
• Operating Personnel. Manning costs are one of the largest components of ship
operating costs. The need to reduced manning is increasingly important, particularly in
military ships that have much more opportunity to improve. Both the number and
required training level (cost) of machinery personnel are important.
• Rating Limitations. Marine machinery is usually available only in discrete sizes. It is

often not possible to obtain the maximum size desired, leading to the use of multiple
units. Often the exact size needed is not available, leading to the need to install oversized
machinery.
• Speed Compatibility with Propulsors. Directly connected low-speed diesels are often
not available with the rated rotation rate (rpm) needed to match the optimum efficiency
propulsor, leading to the use of a suboptimal propulsor. Systems with electrical
transmission and/or gearing with an rpm change between the prime mover and the
propulsor are more likely to support the use of an optimum propulsor.
• Fractional Power Needs and Operating Profile. The percent of time that a vessel will
operate at each speed, the operating or mission profile, can drive the design of the
machinery. A long distance commercial vessel will operate essentially all of its sea time
at its service speed. In contrast, a military vessel might only operate a few percent of the
time at its maximum speed. Some vessels have long periods of time in locks and
restricted channels or must wait for tidal conditions for port entry leading to significant
time at intermediate speeds. Some vessels spend a large part of their time loitering and
station keeping and only a small amount at full power.
1-7
• Inherent Safety. Some choices of machinery type have a direct effect on the inherent
safety of vessels. For example, gasoline engines are often excluded from consideration
for some small craft designs due the inherent explosion hazard from leaks.
• Management Preference. Some ship owners have specific desires relative to

machinery choice due to past experience and/or the skill set of their operating personnel.
• “Buy Domestic” Policy. Machinery choices are often influenced by the desire to have
equipment manufactured and supported from a given country. U.S. Navy procurements
and Jones Act buildings have traditionally been subject to “Buy American” provisions,
but today licensing arrangements and “branch” acquisitions usually work around this
issue.
Differences relative to these issues need to be considered when studying the various options
available to the marine engineer.
As noted above, this introduction to marine engineering is designed to provide a
fundamental introduction to the principles, analysis methods, and design choices involved in
developing the propulsion and auxiliary systems for use onboard ships, marine vehicles, and
offshore platforms. This will include the fundamentals of the prime movers and principal
components, as well as the principles of fluid system design, electrical distribution system
design, and heat transfer that are fundamental to the design, sizing, and selection of marine
propulsion and auxiliary systems and components.
References
1. Woud, H. K., and Stapersma, D., Design of Propulsion and Electrical Power Generation
Systems, The Institute of Marine Engineering, Science, and Technology, London, 2002.
2. Harrington, R. L. (ed.), Marine Engineering, SNAME, Jersey City, NJ, 1992.
3. Carlton, J. S., Marine Propellers and Propulsion, Butterworth Heinemann, UK, 1994.
4. Benford, H., “Engineering Economics,” Chapter 6 in Lamb, T. (ed.), Ship Design and
Construction, Vol. I, SNAME, Jersey City, NJ, 2003.
5. www.bunkerworld.com
6. Doerry, N. H. and Fireman, H., “Designing All Electric Ships,” Proceedings of the 9th
International Marine Design Conference, Ann Arbor, MI, May 2006.
1-8
Chapter 2 Review of Thermodynamics
This chapter is a review of those thermodynamics fundamentals that are used throughout
the remainder of the course. This is not intended to be a complete or an adequate presentation
for someone who is just beginning to study thermodynamics. The presentation draws heavily
upon Chapters 1 through 10 of the excellent Sonntag, Borgnakke, and Van Wylen text [1], but
any beginning thermodynamics text is a useful supplement if needed. In general the
International System of Units (SI) will be used, but special emphasis will also be placed on the
various other units that are encountered in marine engineering practice. The discussion of
availability in Section 2.6 may be new to some readers.
2.1. Pressure, Temperature, and Other Properties

Pressure. Pressure is a normal force per unit area. The definition is the limit of the
normal force per unit area as that area is taken to zero. The pressure, therefore, is applied to a
point. The pressure is the same in all directions at any point when a fluid is at rest. When a fluid
is in motion, the normal force per unit area varies with direction, but the mean value can be
defined as the thermodynamic pressure that we designate as P.
Pressure can be expressed as either an absolute or gage pressure or as a vacuum. Gage
pressure is that commonly read on ordinary gages; it is the pressure above atmospheric pressure.
Vacuum is commonly read on ordinary vacuum gages; it is measured downward from
atmospheric pressure. Absolute pressure is the pressure above a true zero pressure. These
definitions are illustrated in Fig. 2.1. Engineers in practice are often very sloppy in defining
which pressure is being used and only the context determines what pressure is intended.
pressure above atmospheric pressure

gage pressure
absolute atmospheric pressure
barometric vacuum
pressure pressure below atmospheric pressure
pressure absolute pressure
absolute zero pressure
Figure 2.1 Relationships among Absolute Pressure, Gage Pressure, and Vacuum
2-1
The standard atmospheric (barometric) pressure is 760 mm of mercury (Hg) in SI units

based upon the height a column of mercury will rise in a manometer. Recall that the SI system is
based upon the kilogram (kg) as a unit of mass, the meter (m) as the unit of length, and the
second (s) as the unit of time. The English Engineering System, still seen commonly in marine
engineering practice in the United States, uses the pound mass (lbm or lb) as the unit of mass, the
foot (ft) as the unit of length, and the second as the unit of time. The standard acceleration of
gravity g in these systems is,
g = 9.806 65 m/s2 in SI units
g = 32.174 ft/s2 in the English Engineering System
The prefixes most commonly used with the SI system are shown in Table 2.1. The standard
atmosphere in these systems is,
1 atm. = 760 mm Hg = 0.101 325 MPa absolute = 101.325 kPa absolute

1 atm. = 29.92 in Hg = 14.696 lbf/in2 absolute (psia)
The definition of these units will be reviewed below. Pressure is often quoted in the unit of
atmospheres (atm), multiples of the standard atmosphere.
Table 2.1 Most Common SI Unit Prefixes [2]
definition prefix symbol

109 giga G
106 mega M
103 kilo k (note lower case is used)
10–1 deci d
10–2 centi c
10–3 milli m
10–6 micro 
In SI units, the units of force is the newton (N) and the unit of pressure or stress is the
pascal (Pa) which are defined as,
1 N = 1 kg m/s2 (about ¼ lbf)

1 Pa = 1 N/m2
2-2
Common engineering pressures and stresses involve magnitudes of 105 or 106 and higher in
pascals. Thus, megapascals (MPa) or kilopascals (kPa) are commonly used. For convenience in
practice, the bar has also been defined as another unit of pressure. The bar is defined as,
1 bar = 0.1 MPa = 105 Pa = 14.50379 psi
This unit of pressure is in common use in marine engineering. Notice that this is only slightly
smaller (1.3%) than one atmosphere. (In common practice, it is often associated with 1 atm,
which it approximates.)
Earlier metric practice used a kilogram force (kg or kgf) or kilopond (kp) as the unit of
force so it is still common to see pressures expressed in units of kg/cm2, kgf/cm2, or kp/cm2. The
kilogram force is defined as 9.980 665 kg m/s2 so this yields,
1 kgf/cm2 = 0.980 665 bar = 98 066.5 Pa = 14.223 36 psi
Notice that this unit is also somewhat smaller (3.2%) than one atmosphere and it is slightly
smaller (1.9%) than the bar. This unit of pressure is still seen in marine engineering practice in
Japan.
Note in Fig. 2.1 that a pressure below atmospheric pressure can be expressed either as an
absolute pressure or a vacuum. Care must be exercised here to be precise and know which is
being used. Practice is unfortunately often sloppy in this regard. Often only the context and
experience can be used to tell what is intended. For example, a typical commercial marine
propulsion steam plant condenser is designed to operate at a very low 1 1/2” Hg absolute
pressure. Practice traditionally rounds the standard atmosphere to 30” Hg and the condenser is,
thus, designed to operate at a 28 1/2” Hg vacuum. In this example, the magnitude quickly
indicates whether absolute pressure or vacuum is being quoted. In other cases, however, the
context is not so obvious. As with quoting absolute or gage pressures, one should clearly state
whether a pressure or a vacuum is being stated.
Example 2.1 The peak pressure in the cylinder of a low-speed diesel at its Maximum
Continuous Rating (MCR) is 92 kp/cm2 abs. Express this pressure in atmospheres, bars,
pascals, and psia.
92 kp/cm2 x 0.980 665 bar/kp/cm2 = 90.22 bar absolute = 9.022 MPa absolute
92 kp/cm2 x 14.223 36 psi/kp/cm2 = 1308.55 psia
92 kp/cm2 x 14.223 36 psi/kp/cm2 /14.696 psi/atm = 89.04 atm.
2-3
Units of Force, Distance, and Velocity. As an aside here, it is useful to review the
definition of the various units of force (weight) that are used in marine engineering practice; i.e.,
1 lbf = 4.448 222 N = 0.453 59 kgf

l short ton (ST) = 2000 lbf
1 long ton (LT) = 2240 lbf
1 metric ton or tonne (t) = 1000 kgf = 9806.65 N = 2204.6 lbf
The general term ton usually means long ton in marine practice in the United States; it usually
means the metric ton or tonne elsewhere. Using the spelling tonne for the metric ton prevents
confusion. The metric ton or tonne (t) will be used here unless clearly stated otherwise. Again
precision in the clear use of units is important. The difference between the long ton and the
metric ton can be significant. For example, the difference in an annual fuel bill for a 30,000 hp
ship that is incorrectly calculated based upon long tons rather than metric tons would almost be
your annual salary (if you are lucky) or about $120,000. For completeness, the units of distance
and velocity commonly used in marine engineering practice are as follows:
1 ft = 0.3048 m
1 statute mile = 5280 ft = 1.60934 km
1 nautical mile = 6076.10 ft = 1.85200 km
1 knot = 1 nautical mile/h = 1.6878 ft/s = 0.51444 m/s
Temperature. Temperature can be expressed either on an absolute or common

temperature scale. In SI units, the common temperature scale is in degree Celsius (previously
Centigrade) with a symbol ˚C. Water boils at 100˚C and freezes at 0˚C at 1 atm. The associated
absolute temperature scale is the Kelvin scale (symbol K without the degree mark), which has
the same degree size but a different zero point. The two scales are related by,
TK = TC + 273.15
where TK indicates temperature on the Kelvin scale, etc. The common temperature scale in the
English Engineering System is the Fahrenheit scale with the symbol F. Water boils at 212 F and
freezes at 32 F at one atm. The associated absolute temperature scale is the Rankine scale that
has the same degree size and the symbol R. The two scales are related by
TR = TF + 459.67
2-4
The two common scales are related by
TF = 1.8 TC + 32
Properties. The thermodynamic state of a pure substance can be defined by any two
independent properties. The properties depend only on the existing state or condition of the
substance and are independent of the past history or the path through which the substance arrived
at that state. Temperature and pressure are independent thermodynamic properties of a fluid
when a single phase exists. In the change of phase of a substance, such as vaporization (boiling)
of water or condensation of steam, there is a unique combination of temperature and pressure at
which the phase change occurs. The temperature at which vaporization (or condensation) occurs
at constant pressure is called the saturation temperature (Tsat); the pressure at which vaporization
occurs at constant temperature is called the saturation pressure (P sat). The liquid and vapor
phases can exist together in equilibrium only at (Tsat, Psat). With this unique, one-to-one
correspondence between Tsat and Psat, temperature and pressure are not independent
thermodynamic properties when two phases exist together, thus, another property is needed to
define the state of the fluid. The vaporization line for water over the range of interest in marine
engineering practice is illustrated in Fig. 2.2. To the left or above the curve, water exists as a
subcooled fluid. To the right and below the curve, vapor exists in a superheated state. The
degrees of superheat are the number of degrees the vapor has been heated above T sat for the
existing pressure.
P vaporization line (Tsat, Psat)

60 bar
870.2 psia compressed or
subcooled liquid degrees of
superheat
1 atm
14.7 psia superheated
~ 1 bar vapor
100˚C 275.6˚C T
212 F 528.1 F
Figure 2.2 Pressure-Temperature Vaporization Line for Water
2-5
At saturation conditions (Tsat, Psat), the liquid and vapor phases can exist together in
equilibrium in any proportion. The thermodynamic properties quality and moisture are used to
define these proportions. The quality is the ratio of the mass of the vapor to the total mass
(weight percent) and is given the symbol x. A saturated liquid is 100% liquid at (T sat, Psat) and
has quality 0% or 0.00. A saturated vapor or dry steam is 100% vapor at (Tsat, Psat) and has
quality 100% or 1.00. (Moisture has the inverse definition; i.e., it is the ratio of the mass of the
liquid to the total mass. The symbol m, potentially confused with mass, is used with m = 1 – x).
A mixture of fluid (water) and vapor (dry steam) at saturation conditions (T sat, Psat) is called wet
steam.
Specific volume v is the inverse of density; i.e., m3/kg or ft3/lbm. Recall that total
(extensive) thermodynamic properties, such as volume, have capitalized symbols while mass
specific (intensive) thermodynamic properties, such as specific volume, have lower case
symbols. Thus, volume is V while specific volume is v = V/m. Figure 2.3 shows a schematic
temperature-specific volume (T – v) diagram for water. The character of this diagram is typical
for many combinations of properties for a substance with a change of phase from liquid to vapor.
It is, thus, worth careful review. The saturated liquid (x = 0%) and saturated vapor (x = 100%)
lines are shown. Above the critical point discernable liquid and vapor phases no longer exist in
equilibrium together. Typical lines of constant pressure and constant quality are shown.
The Thermodynamic Properties Tables (or Steam Tables [3]) are presented in
thermodynamics texts [1]. The SI units tables from Sonntag, Borgnakke, and Van Wylen [1] are
included for reference as Appendix B. Classroom experience indicates that it is usually
worthwhile to review how these properties are presented for water in the wet steam or saturated
water region. As an example using pressure as the independent variable, a portion of the table
from Sonntag, Borgnakke, and Van Wylen is shown in Fig. 2.4. At this point we will
concentrate only on the specific volume portion, but the method of presentation is the same for
the other properties as well. Looking at Fig. 2.4 at a saturation pressure of 2.25 MPa or 22.5 bar,
the saturation temperature is 218.45˚C. The specific volume for the saturated fluid v f is given as
0.001187 m3/kg and for the saturated vapor vg as 0.08875 m3/kg. The change in specific volume
from saturated liquid to saturated vapor vfg = vg – vf = 0.08756 m3/kg is also given for
2-6
convenience. The properties of wet steam at these conditions can be obtained by simply taking a
mass-weighted average of the properties of the saturated liquid and vapor that compose the wet
steam as illustrated in Ex. 2.2.
T
critical point 220.88 bar
saturated constant P line

fluid line
superheated
subcooled vapor
liquid
constant x saturated
vapor line
wet
steam
Figure 2.3 Schematic T-v Diagram for Water
Figure 2.4 A Portion of Tables of Thermodynamic Properties – Saturated Water Region [1]
2-7
Example 2.2 Obtain the specific volume of wet steam with a quality x = 0.80 at a
pressure of 22.5 bar absolute. These two independent properties define the state. Using a
mass weighted average of the saturated fluid and saturated vapor properties,
mtotal v = mfluid vf + mvapor vg
v = mfluid/mtotal vf + mvapor/mtotal vg = (1 – x) vf + x vg
or alternatively = vf + x vfg = vg – (1 – x) vfg
v = 0.2 (0.001187) + 0.8 (0.08875) = 0.07124 m3/kg
The same technique applies for other properties of wet steam. The alternative forms may be
more convenient in some computations.
Example 2.3 Steam has a specific volume of 0.0930 at a pressure of 22.5 bar. Determine
its quality.
v = vf + x vfg = 0.001187 + x (0.08756) = 0.0930 m3/kg; x = 1.0486, impossible
Recall that quality is only defined when 0 ≤ x ≤ 1.00. A quality greater than one implies a
superheated vapor. That is the case in Example 2.3 since v exceeds vg. A quality less than zero
is also undefined and similarly would imply a subcooled liquid.
Example portions of the superheated vapor water properties and the compressed liquid
table are shown in Fig. 2.5 and 2.6, respectively. In Figs. 2.5 and 2.6, the pressure and
temperature are the independent variables; the associated saturation temperature is shown in
parenthesis after the pressure. Linear interpolation is typically used within these tables.
Figure 2.5 A Portion of Tables of Thermodynamic Properties – Superheated Vapor Region [1]
2-8
Figure 2.6 A Portion of Tables of Thermodynamic Properties – Subcooled Fluid Region [1]
Example 2.4 Obtain the specific volume of a superheated vapor at 7.8 psig and 150˚C.
P = 7.8 psig = 22.496 psia = 0.1551 MPa abs.
Interpolating between P = 100 kPa = 0.100 MPa and P = 0.200 MPa in Fig. 2.5 yields,
v (0.100 MPa, 150˚C) = 1.936 36 m 3/kg
v (0.200 MPa, 150˚C) = 0.959 64 m 3/kg
v = 1.936 36 + (0.959 64 – 1.936 36) • 0.0551/0.100 = 1.398 19 m 3/kg
Example 2.5 Obtain the specific volume of a subcooled liquid at 50 bar and 208˚C.
Interpolating between 200˚C and 220˚C for P = 5000 kPa in Fig. 2.6 yields,
v = 0.001 153 + (0.001 187 – 0.001 153) • 8/20 = 0.001 167 m3/kg
2.2 Heat, Work, Energy, and Power
Energy can take a number of forms: heat, work, kinetic energy, potential energy, etc.
Heat and work are transient phenomena that involve the movement of energy across a system
boundary. Work is defined as the movement of energy that could have the sole effect of lifting a
weight; i.e. a force acting through a distance. By convention in thermodynamics, work is
defined to be positive if done by the system on the surroundings or environment. Heat is defined
2-9
as the movement or transfer of energy by virtue of a temperature difference. Heat is defined to

be positive if the transfer is to the system from the environment. Power is the energy rate.
A process that involves no heat transfer across the system boundary is called adiabatic.
A process that takes place at a constant temperature is called isothermal.
Energy units have evolved for both work and heat. The SI unit of energy is the joule (J)
and the corresponding unit of power is the watt (W) while the most useful unit of power is the
kilowatt (kW) defined as follows:
1J=1Nm
1 W = 1 J/s = 1 N m/s
1 kW = 1 kJ/s = 3600 kJ/h
The SI unit of energy joule is used as the unit of both work and heat. The calorie is a metric unit
of heat energy that is still in use, particularly in quantifying the heat content of fuels. The calorie
(cal) is defined as the amount of heat that will raise the temperature of 1 gm of water 1˚C or 1 K.
The calorie and joule are related by,
1 kcal = 4.1868 J
In the English Engineering System, the unit of heat energy is the British Thermal Unit
(Btu). This was originally defined as the amount of heat that will raise the temperature of 1 lbm
of water from 59.5 F to 60.5 F. It is now defined as,
1 Btu = 778.17 ft lbf = 1.0550 kJ.
There is no separate unit of work in the English system; the ft lbf is usually used. It is useful to
recall the unit conversion factors shown in Table 2.2. These are all expressed as a dimensionless
1 so that they can be used whenever needed in equations to resolve units. The second English
Engineering System factor is often designated as gc in thermodynamics texts. Fortunately, the
corresponding quantity in SI units, the second SI factor, is just 1.
Table 2.2 Unit Conversion Factors Used to Resolve Units
778.17 ft lbf/Btu = 1 32.174 lbm ft/(lbf s2) = gc = 1
1 Nm/J = 1 1 kg m/(N s2) = 1
2-10
The common unit of power in the English Engineering System is the horsepower that has
the symbol hp. Since there is also a metric horsepower, we will use the symbol with the
subscript Br. The English horsepower is defined as,
1 hpBr = 550 ft lbf/s
The older metric horsepower, also called the German Pferdestaerke (P.S.) or French Cheval
Vapeur (C.V.), is defined in terms of the kilogram force as follows:
1 hpM = 4500 kgf m/min
The two horsepowers relate to the kilowatt as follows:
1 hpBr = 0.74570 kW
1 hpM = 0.73550 kW = 0.98632 hpBr
The three units of power relate to the Btu as follows:
1 hpBr = 550 ft lbf/s • 3600 s/h /(778.17 ft lbf/Btu) = 2544.43 Btu/h

1 hpM = 2509.63 Btu/h
1 kW = 3412.14 Btu/h
The kilowatt will be used here as the principal unit of power. When the kilowatt is used to
quantify electrical power, such as the output of a generator, the prefix e is often used to
differentiate between electrical power (ekW) and mechanical power (kW, mkW or brake bkW).
2.3 The First Law of Thermodynamics
The First Law of Thermodynamics is a statement of the conservation of energy. It states

that as a system undergoes any cycle, the cyclic integral of the heat (the net heat transfer during
the cycle) is proportional to the cyclic integral of work. In SI units where the joule is the unit of
both work and heat, the constant of proportionality is 1 and the First Law becomes,
∮ ∂Q = ∮ ∂W (2.1)
This is an experimentally observed fundamental law of nature.
2-11
Equation 2.1 can be used to prove the existence of the thermodynamic property that we
call energy. By definition, a property of a substance depends only on its existing state or
condition. It is independent of past history; i.e., how the substance arrived at that state. If a
substance were to undergo a change from state 1 to state 2 by some process A and then from
state 2 back to state 1 by another process B, the path of its properties (e.g. P, v) might appear as
in Fig. 2.7. We could also move between states 1 and 2 by process A and then back to state 2 by
a different process C. Using paths A and B, the First Law requires,
2A 1B 2A 1B
Q + Q = W + W

1A 2B 1A 2B
Rearranging and changing the sign and direction of integration for path B this yields,
2A 2B
(Q – W) = Q – W)

1A 1B
Using paths A and C, we similarly have,
2A 2C
(Q – W) = Q – W)

1A 1C
Subtracting these last two equations and rearranging we obtain,
2B 2C
(Q – W) = Q – W) (2.2)

1B 1C
Equation 2.2 states that the quantity in parenthesis is independent of the path taken between
states 1 and 2. It is, therefore, the differential or change in a property that we call energy,
E = Q – W (2.3)
For convenience, we consider this energy to be composed of three identifiable parts; i.e., the
kinetic energy (KE) associated with the velocity V (note the bold notation) of the substance, the
potential energy (PE) associated with the elevation z of the substance in a gravitational field, and
2-12
the internal energy (U) that is reflected in the temperature of the substance. Using this
breakdown, eq. 2.3 can be rearranged and integrated between two states 1 and 2 to yield,
1Q2 = U2 – U1 + m (V22 – V12)/2 + mg (z2 – z1) + 1W2 (2.4)
This is another statement of the First Law of Thermodynamics. This is a statement of

conservation of energy for a fixed mass of a substance as it changes state from state 1 to state 2.
This equation can be divided by the mass m to yield,
1 q2 = 1Q2/m = u2 – u1 + (V22 – V12)/2 + g (z2 – z1) + 1w2 (2.5)
Notice that the mass specific thermodynamic property internal energy u = U/m is given in Tables
of Thermodynamic Properties in the same manner as v. As energy per unit mass, u has units
kJ/kg (or Btu/lbm).
P
2
A B
C
Figure 2.7 Demonstration of the Existence of Thermodynamic Property Energy
In processes that involve the movement of fluid across a boundary of the control volume,
it is convenient to define another thermodynamic property called enthalpy. It is defined as
follows:
H = U + PV
h = H/m = u + Pv
By definition, enthalpy is a combination of thermodynamic properties so it too is a property, it

depends only upon the state of a substance. The property h represents the energy carried by a
2-13
unit mass of the substance as it enters or leaves a control volume. The first part u is the internal
energy of that mass; the second part Pv is the flow work that must be done to cross the system
boundary. The mass specific thermodynamic property enthalpy h = H/m is given in the Tables
of Thermodynamic Properties in the same manner as v and u.
The thermodynamic evaluation of most marine engineering systems involves the use of
the idealized steady-state, steady-flow process. This model assumes (1) a control volume that
does not move with respect to the coordinate reference frame, (2) a fixed mass within the control
volume with properties that are constant at each point, and (3) constant rates of mass flow, heat
transfer, and work across the control volume boundary. This model is shown schematically in
Fig. 2.8. The mass flow rate crossing the control volume boundary at various points is obtained
by integrating the product of the fluid density and velocity over that area; i.e.,
m= VdA
A
In most applications, the density is constant over A and the integral yields the area average
velocity V times the area A,
m= VdA = V = V/v (2.6)

A
The bar over the velocity will be dropped for simplification and the area average velocity will be
assumed.
control volume (c.v.)
inflow Vi , mi
constant mass m cv area Ae

area Ai
exiting flow Ve, me
Figure 2.8 Steady-State, Steady-Flow Model
2-14
Conservation of mass (continuity) applied to the control volume requires that,
mcv =  mi –  me = 0
so we have
 mi =  me = m (2.7)
The First Law can also be stated on a time-rate basis for the steady-state, steady-flow control
volume as follows:
QCV +  mi (hi + Vi2/2 + gzi) =  me (he + Ve2/2 + gze) + WCV (2.8)
This is an expression of conservation of energy per unit time of the steady-flow process. The
first and last terms are now the rates of heat transfer and work crossing the control volume
boundary in kJ/s (or Btu/s). The second term on the left is the sum of the rates at which energy
is being brought into the control volume by the various incoming flows m i. The first term on the
right is the sum of the rates at which energy is being removed from the control volume by the
various exiting flows me. If eq 2.8 is divided by the total flow entering or leaving m, it becomes,
q +  (mi/m) • (hi + Vi2/2 + gzi) =  (me/m) • (he + Ve2/2 + gze) + w (2.9)

where
q = QCV /m and w = WCV /m
Each term of eq. 2.9 will now have units kJ/kg (or Btu/lbm). If there is only one flow entering
and one flow exiting the control volume, eqs. 2.7 and 2.9 become,
m = m i = me (2.10)
q + hi + Vi2/2 + gzi = he + Ve2/2 + gze + w (2.11)
Notice that q and w are the heat transfer and work (excluding flow work), respectively, crossing
the control volume boundary per unit mass of the fluid passing through the control volume. In
2-15
eq. 2.5, q and w are the heat transfer and work crossing the control volume boundary per unit
mass of the fluid within the control volume.
Example 2.6 A steam turbine used to drive an electric generator on a diesel powered
ship has the inlet and exhaust conditions shown. The steam is generated in a waste heat
boiler that is heated by the 290˚C diesel exhaust. Determine the power output of this
turbine.
1000 kJ.h
Q = 1000 kJ/h heat
heat loss
loss
mi
w
turbine
control volume
me
Figure 2.9 Sketch for Example 2.6
mi = 4000 kg/h me = 4000 kg/h

zi = 7 m above baseline ze = 6 m above baseline
Pi = 18 bar = 1.8 MPa abs. Pe = 1 1/2” Hg abs.
Ti = 260˚C xe = 0.90
Vi = 200 ft/s ≈ 60.0 m/s Ve = 330 ft/s ≈ 100.0 m/s
These velocities are the maximum design velocities recommended for marine engineering
practice for steam and exhaust at high vacuum [4]. The exhaust pressure Pe is typical
condenser pressure for commercial practice, which in SI units becomes,
Pe = 1.5” Hg • 0.101322 MPa/atm/29.92” Hg/atm = 5.08 kPa = 0.05 bar
From the Tables of Thermodynamic Properties,
hi = h (1.8 MPa, 260˚C) = 2934.6 kJ/kg

hf (5 kPa) = 137.82 kJ/kg hg (5 kPa) = 2561.5 kJ/kg
he (5 kPa, x = 0.90) = 0.1 • 137.82 + 0.9 • 2561.5 = 2319.1 kJ/kg
Applying eq. 2.11 to the steam turbine control volume,
2-16
q + hi + Vi2/2 + gzi = he + Ve2/2 + gze + w
where q = Qcv/m = – 1000 kJ/h/4000 kg/h = – 0.25 kJ/kg

Vi2/2 = 60.02 m2/s2/2 • 1 J/N m • 1 N s2/kg m • kJ/1000 J = 1.80 kJ/kg
Ve2/2 = 100.02 m2/s2/2 • 0.001 kJs2/kgm2 = 5.00 kJ/kg
g(zi – ze) = 9.80665 m/s2 • 1 m • 0.001 kJ s2/kg m2 = 0.01 kJ/kg
Notice that even with these maximum velocities, the kinetic energy terms are small. The
change in potential energy and the heat transfer are also very small. Solving for the
work,
w = q + (hi – he) + (Vi2/2 – Ve2/2) + g(zi – ze)

= – 0.25 + (2934.6 – 2319.1) – 3.2 + 0.01 = 612.1 kJ/kg
W = mw = 4000 kg/h • 612.1 kJ/kg • (1/3600) kWh/kJ = 680.1 kW = 912.0 hpBr
The AC generator driven by this turbine might produce an output of 640 ekW and this
might be enough electrical power for the entire ship while at sea.
Note in this realistic example, that the heat transfer and changes in kinetic energy and
potential energy are small compared with the change in enthalpy. As a result, typical practice at
the first level of analysis is to treat an insulated component like the turbine as adiabatic and to
neglect the changes in kinetic energy and potential energy. The First Law applied to the turbine
work is then just,
w = hi – he = = 615.5 kJ/kg (+ 0.55% approximation error)
The incoming kinetic energy is actually utilized by the turbine. The heat loss and exhaust kinetic
energy represent losses in the turbine (negative quantities above) and are considered in the
detailed evaluation of the efficiency of the component. The exhaust (kinetic energy) loss can be
incorporated by raising he to the enthalpy of the steam after the flow is slowed to near zero
within the condenser; i.e., he’ = he + Ve2/2.
Example 2.7 In Example 2.6, calculate the kinetic energy of the exhaust flow in English
Engineering System units,
Ve2/2 = 3302 ft2/s2/(2 •32.174 lbm ft/lbf s2 • 778.17 ft lbf/Btu) = 2.17 Btu/lbm
2-17
Example 2.8 In Example 2.6, calculate the minimum area Ae of the exhaust trunk
(exhaust line) from the turbine. From the Tables of Thermodynamic Properties,
vf (5 kPa) = 0.001 m3/kg vg (5 kPa) = 28.192 m3/kg

ve = v (5 kPa, x = 0.90) = 25.373 m3/kg
and using eq. 2.6 with the maximum permissible velocity,
Ae = mve/Ve = 4000 kg/h • 25.373 m3/kg/(100 m/s • 3600 s/h) = 0.282 m2
Thus, a rectangular trunk somewhat larger than 0.7 m x 0.4 m would be adequate.
Example 2.9 In Example 2.6, calculate the area Ai of the inlet line for the turbine if this
maximum design velocity were to be used. From the Tables of Thermodynamic
Properties,
v(1.8 MPa, 260 ˚C) = 0.12812 m 3/kg

Ai = 4000 kg/h • 0.12812 m3/kg/(60 m/s • 3600 s/h) = 0.00237 m2
Thus, a pipe with only a 0.055 m = 2.16” internal diameter (ID) would be needed. A 2”
Schedule 80 pipe (ID = 2.323”) or even larger might actually be used.
Notice the very large change in the specific volume of the steam as it expands between turbine
throttle conditions and the exhaust conditions. The specific volume changes by a factor of 200 in
the example above. The change would be even greater across a propulsion turbine due to its
higher throttle inlet steam conditions.
2.4 The Second Law of Thermodynamics
The Second Law of Thermodynamics reflects the physical reality that certain processes
can only proceed in a single direction. For example, heat will be transferred only from a higher
temperature to a lower temperature. There are two classic statements of the Second Law [1]:
Kelvin-Planck Statement: It is impossible to construct a device that operates in a cycle

and produces no effect other than the raising of a weight and the exchange of heat with a
single reservoir.
Thus, exchange of heat with a second reservoir is needed in a heat engine.
Clausius Statement: It is impossible to construct a device that operates in a cycle and

produces no effect other that the transfer of heat from a cooler body to a hotter body.
2-18
Thus, work must be supplied to a refrigerator or heat pump.

A simple heat engine cycle and a simple refrigeration cycle composed of steady-state,
steady-flow processes are shown in Fig. 2.10. For the heat engine, the Kelvin-Planck statement
requires that the heat engine receive QH from a higher temperature reservoir and also exhaust
heat QL to a lower temperature reservoir. The First Law eq. 2.1 applied to the fixed mass of
fluid within the cycle system boundary yields,
QH – QL = W
The thermal efficiency of the cycle th is the work produced W divided by the amount of energy
that must be supplied QH; i. e.,
th = W/QH = (QH – QL)/QH = 1 – QL/QH (2.12)
By virtue of the Kelvin-Planck statement of the Second Law QL > 0 and, thus, the thermal
efficiency is always less than one.
QH QH
system boundaries
boiler condenser
pump turbine
W W
expansion
valve
compressor
condenser evaporator
QL QL
simple heat engine simple refrigeration cycle
Figure 2.10 Simple Heat Engine and Refrigeration Cycles
For the refrigeration cycle, the Clausius statement requires that W > 0. The First Law eq.
2.1 applied to the fixed mass of fluid within the refrigeration cycle system boundary yields,
2-19
– QH + QL = – W
The measure of the “efficiency” of this cycle is the coefficient of performance  which is the
amount of heat removed QL divided by the amount of work that must be supplied; i.e.,
 = QL/W = QL/(QH – QL) = 1/(QH/QL – 1) (2.13)
By virtue of the Clausius statement QH > QL and, thus, the coefficient of performance is always
positive.
Reversible Processes and Irreversibilities. A reversible process is an idealized process,
which once having taken place, could be reversed and have no effect on the system or
surroundings. It is an ideal, quasi-equilibrium process without losses. Irreversibilities are losses
that occur in any real process. Common examples in marine engineering applications are
friction, free expansion, heat transfer through a finite temperature difference, mixing of different
substances, electrical resistance, and combustion. Reversibility can be considered in terms of
internal reversibility or external reversibility. It is possible for a system to be internally
reversible (within the system boundary), but externally irreversible if the only irreversibility
occurs across the system boundary. The most common example is an internally reversible
system that exchanges heat across its boundary through a finite temperature difference.
The Inequality of Clausius and Entropy. A corollary of the Second Law of
Thermodynamics that can be shown to be true for all cycles is the Inequality of Clausius [1]. It
holds for all reversible and irreversible heat engines and refrigeration cycles. The inequality
states that the cyclic integral of the heat transferred in the cycle divided by the absolute
temperature is less than or equal to zero; i.e.,
∮ ∂Q/T ≤0 (2.14)
In this expression, the equality can be shown to hold for reversible cycles; the inequality holds
for irreversible cycles. If a cycle satisfies this inequality, it is consistent with the Second Law of
Thermodynamics. Note that the temperature used in this statement is the thermodynamic or
absolute temperature (K or R). Thus, any results derived using eq. 2.14 require the use of
absolute temperature. This is a common source of error.
2-20
The Inequality of Clausius can be used to demonstrate the existence of another

thermodynamic property called entropy in the same manner as we demonstrated the existence of
the property energy. Consider two reversible cycles as shown in Fig. 2.11. One cycle moves
between states 1 and 2 using paths A and B. The second moves between states 1 and 2 using
paths A and C.
P
2
A B
C
Figure 2.11 Demonstration of the Existence of the Thermodynamic Property S
Since the two cycles in Fig. 2.11 are reversible we have from eq. 2.14,
2A 1B 2A 2B
Q/T = 0 = Q/T + Q/T = Q/T – Q/T

AB 1A 2B 1A 1B
2A 1C 2A 2C
Q/T = 0 = Q/T + Q/T = Q/T – Q/T

AC 1A 2C 1A 1C
Subtracting the second equation from the first and rearranging yields,
2B 2C
Q/T = Q/T
1B 1C
Since this integral is path independent; i.e., it depends only on the states 1 and 2, the integrand
must be a thermodynamic property. It is possible then to define a property entropy as follows:
2-21
dS = (Q/T)rev. (2.15)
and integrating between two states yields,
S2 – S1 = Q/T)rev. (2.16)
1
This definition is for a reversible process since we have used the equality in the Inequality of
Clausius. Also, the temperature in eq. 2.15 and eq. 2.16 is the absolute temperature. The
entropy per unit mass s = S/m has units of kJ/kg K or Btu/lbm R. The entropy per unit mass s is
given in the Tables of Thermodynamic Properties in the same manner as v, u, and h.
If we assume that process C in Fig. 2.11 is irreversible and we use the inequality in eq.
2.14 for the cycle using paths A and C, we can follow the same approach as above to obtain the
following results for any real cycle,
dS ≥ Q/T (2.17)
and integrating between two states yields,
S2 – S1 ≥ Q/T (2.18)
1
If the process between states 1 and 2 is reversible, the equalities hold and eqs. 2.15 and 2.16
result. Equation 2.18 states that entropy must increase in any real (and therefore irreversible)
adiabatic (Q = 0) process. If a process is both reversible, so the equality holds, and also
adiabatic, so that Q = 0, eq. 2.18 states that the entropy is constant. Such a process is, therefore,
termed isentropic (reversible and adiabatic). If the process is both reversible and isothermal, so
the temperature can be brought outside the integral, eq. 2.18 yields.
S2 – S1 = Q/T (2.19)
If a process is internally reversible, eq. 2.15 yields,
Q = TdS
2-22
which can be integrated to yield,
1Q2 = TdS (2.20)

1
This indicates that the area under a T-S diagram of an internally reversible process between
states 1 and 2 (down to absolute zero) equals the amount of heat transferred in that process.
If we consider a compressible substance that undergoes a change of state with negligible
kinetic energy and potential energy change, the First Law eq. 2.3 yields,
Q = dU + W
If the process is a reversible quasi-equilibrium process, the work is given by,
W = PdV
Equation 2.15 also yields,
Q = TdS
Using these three equations we have
TdS = dU + PdV [called the First Gibbs Equation]
and dividing by the mass this yields
Tds = du + Pdv (2.21)
By definition, the enthalpy per unit mass is h = u + Pv which can be differentiated to yield,
dh = du + Pdv + vdP
This can be solved for du and substituted into eq. 2.21 to produce,
Tds = dh – vdP (2.22)

and
TdS = dH – VdP [called the Second Gibbs Equation]
2-23
Equations 2.21 and 2.22 and the Gibbs Equations involve only state properties and are, therefore,
true independent of path or the reversibility or irreversibility of the processes involved. Thus,
they can be used whenever convenient to relate changes in the various properties.
The rate form of the Second Law eq. 2.17 applied to a steady-state, steady-flow process
involving a control volume yields,
 me se –  mi si ≥ QCV/A)/T dA (2.23)
A
where A is the area of the control volume boundary surface. This corresponds to the eq. 2.8
statement of the First Law. If there is a single flow entering the control volume and a single flow
leaving the control volume, this becomes,
m (se – si) ≥ (QCV/A)/T dA (2.24)

A
Example 2.10 Evaluate the amount of power needed to pump 100 000 kg/h of water at
125˚C from 3 bar to 60 bar absolute in an isentropic (ideal) pump. These values are
typical of the main feed pump in a steam plant producing about 25 000 kW.
For an isentropic (reversible, adiabatic) steady-state, steady-flow process, the Second

Law eq. 2.24 becomes simply,
se = si and ds = 0
The First Law eq. 2.11 yields for an adiabatic (Q = 0) process,
ws = (hi – he) + (Vi2 – Ve2)/2 + g(zi – ze)
The subscript s is used to denote the isentropic (ideal) work. Using eq. 2.22 with ds = 0
yields dh = vdP which can also be used to avoid having to evaluate the enthalpy of the
compressed liquid. Neglecting the kinetic energy change (zero with incompressible flow
and equal diameter inlet and outlet lines) and neglecting the potential energy change, we
then have
e e
ws = – dh = – vdP (2.25)
i i
From the Tables of Thermodynamic Properties:

vi = v (0.3 MPa, 125˚C) = 0.001 065 m3/kg
ve ≈ v (6.0 MPa, 125˚C) = 0.001 062 m3/kg
2-24
vi ≈ v e
Since the flow is essentially incompressible, constant density  or v, the specific volume
v can be brought outside the integral in eq. 2.25 to yield
ws = – vi (Pe – Pi) (2.26)

This yields
ws = – 0.001065 m3/kg (6.0 – 0.3) x 106 N/m2 = – 6070.5 N m/kg
and the power is
mws = – 6070.5 N m/kg 100 000 kg/h • kJ/(1000 N m) • kW/(3600 kJ/h)

= – 168.6 kW = – 226.1 hpBr
The work is negative since power must be supplied to the system to drive the pump.
Example 2.11 Repeat Example 2.10 in English Engineering System units.

vi = v (43.51 psia, 257 F) = 0.017 06 ft3/lbm
Pi = 3 bar • 14.503 75 psi/bar = 43.51 psia
Pe = 870.22 psia
m = 220 462 lbm/h
ws = – 0.01706 ft3/lbm • (870.2 – 45.3) lbf/in2 • 144 in2/ft2/(778.17 ft lbf/Btu)

= – 2.6099 Btu/lbm
power = mws = – 2.6099 Btu/lbm • 220 462 lbm/h / (2544.43 Btu/hp Br h) = – 226.1 hpBr
Example 2.12 The condenser for the turbogenerator set turbine in Example 2.6
condenses the exhaust steam to saturated fluid at the condenser pressure of 1 1/2” Hg abs.
Use both the First and Second Laws to calculate the rate at which heat is transferred to
the sea water cooling the condenser. The process T-s diagram is shown in Fig. 2.12.
T
abs. saturated
fluid
1 1/2 " Hg
line
abs
2 saturated
e vapor line
x = 0.90
Figure 2.12 Property Sketch for Example 2.12
2-25
Using the First Law neglecting kinetic energy and potential energy changes with no
work in the condenser,
q + h e = h2 , q = h 2 – he
from Example 2.6 he = 2319.1 kJ/kg

from the Tables h2 = hf (0.005 MPa) = 137.82 kJ/kg
q = (137.82 – 2319.1) kJ/kg = – 2181.28 kJ/kg

eQ2 = mq = 4000 kg/h (–2181.28 kJ/kg) = – 8725.1 MJ/h
The result is negative because heat is transferred from the system. In the condenser, the
condensation process is isothermal at Tsat so the Second Law and eq. 2.19 yields
eQ2 = T(S2 – Se) = Tm (s2 – se)
From the Tables

T = Tsat (0.005 MPa) = 32.88˚C + 273.15 = 306.03 K
s2 = sf (0.005 MPa) = 0.4764 kJ/kg K
sg (0.005 MPa) = 8.3951 kJ/kg K
se = s (0.005 MPa, x = 0.9) = (1 – x)sf + x sg = 7.6032 kJ/kg K
eQ2 = 306.03 K • 4000 kg/h • (0.4764 – 7.6032) kJ/kg K = – 8724.1 MJ/h
Theoretically the same as in the calculations above, but slightly different (0.01%) due to
the number of significant figures used in the Tables of Thermodynamic Properties.
Mollier Chart. An extremely useful tool in thinking and analysis related to steam
turbines is the specialized h-s diagram called the Mollier chart. An SI unit version for steam is
shown in Fig. 2.13. An English Engineering System units version for steam is shown in Fig.
2.14. The chart has the basic characteristics of the T-v and P-s diagrams; i.e., saturated fluid
line, saturated vapor line, wet steam region, etc. The charts are especially prepared, however, to
include the region of properties associated with turbines. They are used to establish the
properties along the turbine expansion state line. The portion of the saturated fluid line that
would be applicable to the consideration of the pumps in steam cycles is off the chart to the
lower left. Thus, the typical steam Mollier charts are used primarily in dealing with steam
turbines. The charts are very handy to use in preliminary engineering calculations; large-scale
versions are available with the Steam Tables [3] if increased accuracy is needed. The use of the
chart is strongly encouraged as a check when the Tables of Thermodynamic Properties are used.
2-26
Obtaining the enthalpy of a wet steam from the Tables involves enough steps and calculations
that errors are common.
Figure 2.13 Mollier Chart for Steam – SI Units
2-27
Figure 2.14 Mollier Chart for Steam – English Engineering System Units
2-28
Notice in Fig. 2.13 that lines of constant temperature, constant pressure, and constant percent
quality are included. In Fig. 2.14, lines of constant temperature, constant pressure, constant
percent moisture (1 – x ), and constant superheat are included. In the wet steam region there is a
one-to-one correspondence between Tsat and Psat so only lines of constant pressure are included
in this region. The English unit chart includes lines of constant pressure in units of “Hg abs. in
the wet steam region where this is particularly useful in dealing with turbine exhaust (condenser)
conditions. Superheat is defined as the amount the temperature is above the saturation
temperature for the existing pressure; i.e., T – Tsat(P). Note that at high superheat the lines of
constant temperature become horizontal. Thus, enthalpy becomes a function of temperature only
and a second independent property is not needed to obtain enthalpy for steam at high superheat.
Example 2.13 Use the Mollier chart to obtain the enthalpy and entropy of a turbine
exhaust having conditions of 0.006 MPa abs. and 88% quality.
from Fig. 2.13

h (0.006 MPa, 88%) = ~ 2277 kJ/kg
s (0.006 MPa, 88%) = ~ 7.39 kJ/kg K
Example 2.14 Use the Mollier chart to obtain the enthalpy and quality of a turbine
exhaust at 1 1/2” Hg abs. if the expansion across the turbine is isentropic from throttle
inlet conditions of 850 psia and 950 F. Use English units.
from Fig. 2.14
si (850 psia, 950 F) = ~ 1.655 Btu/lbm R
moving straight down the Mollier chart for an isentropic (constant s) expansion to a
pressure of 1 1/2” Hg abs. yields,
he (1 1/2” Hg, s = 1.655 Btu/lbm R) = ~ 945 Btu/lbm
xe (1 1/2” Hg, s = 1.655 Btu/lbm R) = ~ 83.8%
Component Efficiencies. The thermal efficiency of the total cycle of a heat engine was
established in eq. 2.12. Efficiencies can also be defined for individual components or processes
within the cycle that are essentially adiabatic (insulated) and involve the production or input of
mechanical work. From the Second Law eq. 2.18 any real, adiabatic process will involve
irreversibilities that will produce an increase in entropy. Sketches of a simple turbine expansion
process and a simple pumping or compressor process are shown on the h-s diagram in Fig. 2.15.
2-29
h Pi Pe
h
Pe e
i es
actual
actual state line Pi
ideal e ideal
es i
s s
turbine pump or compressor
Figure 2.15 Turbine and Pump or Compressor Processes on an h-s Diagram
In each case in Fig. 2.15, the inlet condition is designated with the subscript i. The actual
exhaust conditions are designated with the subscript e; the ideal (reversible, adiabatic or
isentropic) exhaust conditions are designated with the subscript es. Note that in each case, by the
Second Law the actual adiabatic processes involve an increase in entropy. Using the First Law
neglecting kinetic and potential energy changes, the work in either case is just the change in
enthalpy. The isentropic turbine efficiency can be defined as the ratio of the work actually
obtained to the work that could be obtained from an ideal turbine that exhausts to the same
exhaust pressure. Thus,
turbine = wa/ws = (hi – he)/(hi – hes) (2.27)
The isentropic pump or compressor efficiency can be defined as the ratio of the work required by
the ideal pump/compressor to the work required by the actual pump/compressor when both
discharge to the same discharge pressure. Thus,
pump/compressor = ws/wa = (hi – hes)/(hi – he) = (hes – hi)/(he – hi) (2.28)
In dealing with pumps, it is usually convenient to deal with the work as a positive number and
not follow the sign convention for work that must be used when formally applying the First Law.
The sign is known from the physical application. The last form of eq. 2.28, thus, has a positive
numerator and denominator. Note that eq. 2.26 can be used to obtain the ideal pump work w s
without the need to evaluate enthalpies. Notice also that the key to these definitions of efficiency
2-30
is that the ideal, reference component exhausts (discharges) to the same pressure as the actual
component. Other component efficiency definitions are possible as will be seen in Section 2.6.
Example 2.15 Determine the enthalpy of a fluid leaving the pump considered in
Example 2.10 if it has an efficiency of 70%. Determine the enthalpy added to the fluid as
it passes through the pump.
from Example 2.10 for the ideal pump

ws = – 6070.5 J/kg = – 6.07 kJ/kg
from the Tables
hi (0.3 MPa, 125˚C) = 525.11 kJ/kg
from eq. 2.28
wa = hi – he = ws/pump = – 8.67 kJ/kg, the enthalpy added to the fluid
he = hi + 8.67 = 533.78 kJ/kg
The constant pressure specific heat is the amount of heat needed to change the temperature of
a unit mass one degree at constant pressure. For water this is approximately,
CP ≈ 1 Btu/lbm R ≈ 1 kcal/kg K = 4.1868 kJ/kg K
This result is exact for water at 1 atm. absolute and the temperature used to define the Btu or the
calorie. The Tables of Thermodynamic Properties have more exact results for other conditions if
required. The enthalpy added to the fluid by the pump is called the pump regain. In Example
2.15, the addition of 8.67 kJ/kg to the fluid will raise its temperature,
CP = (dh/dT)P ≈ h/T
T = h/CP = 8.67 kJ/kg / 4.1868 kJ/kg K = 2.1 K or 2.1˚C
The constant pressure specific heat can be used here since the water is essentially
incompressible. The work required to pump the fluid is, therefore, not lost, but reappears as an
enthalpy gain and an associated temperature increase of the fluid.
Carnot Cycle. From the Kelvin Planck statement of the Second Law, we know that a heat
engine (power cycle) must operate between two heat reservoirs. The resulting thermal efficiency
was established in eq. 2.12. The Carnot cycle is the ideal (reversible) cycle that possesses the
best efficiency that can be achieved when operating between any two particular reservoirs. This
ideal cycle is defined to consist of four processes as follows:
2-31
1–2 reversible isothermal heat addition

2–3 reversible adiabatic (isentropic) expansion
3–4 reversible isothermal heat rejection
4–1 reversible adiabatic compression or pumping
The T-s diagram for these processes is shown in Fig. 2.16. From eq. 2.19, the heat transfer in a
reversible, isothermal process is the area below the process curve on the T-s diagram using an
absolute temperature scale or,
QH = TH(S2 – S1)
QL = – TL(S4 – S3) = – TL(S1 – S2) = TL(S2 – S1)
The minus sign is used in the second equation to produce a positive number for Q L, which is
defined as a quantity of heat being transferred from the system. Using these results in eq. 2.12
yields the thermal efficiency for the Carnot cycle; i.e.,
Carnot = 1 – TL(S2 – S1)/[TH(S2 – S1)] = 1 – TL/TH (2.29)
where TL and TH are again absolute temperatures. Given two reservoirs at TL and TH, this
reversible cycle is the most efficient cycle that could be constructed between these reservoirs.
This ideal cycle is a direct consequence of the Kelvin-Planck statement of the Second Law of
Thermodynamics; any better efficiency would violate the Second Law.
T
QH
1 2
TH Carnot
w net = Q H - Q L
area within cycle
on T-s diagram
TL
4 3
QL
s
s1 s2
Figure 2.16 T-s Diagram for Carnot Cycle
2-32
Example 2.16 The typical marine steam cycle has a maximum temperature that is
limited by the impurities in residual fuel to about 1000 F. (For maximum steam
temperatures above about 950 F, the vanadium impurity in the boiler fuel causes slagging
deposits on the outside of the superheater tubes.) The condenser is cooled by the sea; the
design sea water temperature is usually taken as 75 F in commercial practice [5].
Determine the Carnot efficiency of an ideal cycle operating between reservoirs at these
two temperatures.
TH = 1000 + 459.67 = 1459.67 R
TL = 75 + 459.67 = 534.67 R
Using eq. 2.29
Carnot = 1 – 534.67/1459.67 = 63.4%
This is the most efficient cycle that could be developed. Anything better would violate the
Second Law. Actual marine steam cycles have thermal cycle efficiencies below 40% when the
change of state and all the realistic irreversibilities are included. For comparison, the largest
low-speed marine diesels have thermal efficiencies slightly above 50% today.
2.5 The Ideal Gas Model
The ideal gas model is a useful approximation in many applications in marine

engineering. It is observed experimentally that low-density gases obey the following
relationship on a per mole basis,
Pv = RT (2.30)
where R is the Universal Gas Constant, v is the volume per mole of gas, and T is the absolute
temperature. The Universal Gas constant is found experimentally to be,
R = 8.314 5 J/(kg mole K) = 1545 ft lbf/(lbm mole R)
For n moles of gas with molecular weight M, the mass m = nM and eq. 2.30 yields,
PV = nRT = m (R/M) T = mRT (2.31)
Pv = RT (2.32)
2-33
where R is now a material dependent gas constant with units kJ/kg K or Btu/lbm R. Values for
dry air and steam at particular temperatures are given in Table 2.3. This equation of state is
simply a model that is useful approximation in treating low density gases. It is especially useful
here in the study of the fundamental characteristics of internal combustion engine cycles since it
is a reasonable approximation and eliminates the need for “Steam Tables” for the mixture of air
and combustion gases. For a fixed mass of ideal gas, eq. 2.31 yields,
PV/T = mR = constant (2.33)
which is a combination of Boyle’s and Charles’ Laws.
Table 2.3 Gas Constants and Specific Heats for Dry Air and Steam
dry air steam

SI English SI English
R 0.2870 kJ/kg K 53.34 ft lbf/lbm R 0.4615 85.76

CP0 1.0035 kJ/kg K 0.240 Btu/ lbm R 1.8723 0.447
CV0 0.7165 kJ/kg K 0.171 Btu/lbm R 1.4108 0.337
k = CP0/CV0 1.400 1.400 1.327 1.327
Note: SI specific heats for 25˚C; English Engineering System specific heats for 77 F.
In the treatment of ideal gases, it is useful to utilize the constant pressure and constant
volume specific heats listed in Table 2.3. These are defined in general as follows:
CP = (∂h/∂T) P (2.34)
CV = (∂u/∂T)V (2.35)
The notation indicates that the partial derivative is taken at constant pressure in eq. 2.34 and at
constant volume in eq. 2.35. From the definitions, the specific heats have units kJ/kg K or
Btu/lbm R. For an ideal gas, the internal energy and enthalpy can both be shown to be functions
of temperature only so equations 2.34 and 2.35 become, respectively,
CP0 = dh/dT and dh = CP0dT (2.36)

CV0 = du/dT and du = CV0dT (2.37)
2-34
The subscript 0 is often added to indicate low density or “zero” pressure ideal gas. The ratio of
these specific heats is defined as,
k = CP0/CV0 (2.38)
Since CP0, CV0, and k are defined in terms of thermodynamic properties, they also are
thermodynamic properties.
For many engineering applications, the specific heats can be assumed to be constant.
Equation 2.36 can then be integrated to yield,
h2 – h1 = CP0(T2 – T1) (2.39)
This is an extremely useful result since it allows us to obtain enthalpy changes in an ideal gas
process from the absolute gas temperatures at the inlet and outlet of the process. Equation 2.39
involves only properties and, thus, it depends only upon the end states 1 and 2 and not on the
path through which the fluid moved between states 1 and 2. We can also derive an expression
for the change of entropy in a process involving an ideal gas with constant specific heats.
Substituting v from the ideal gas law eq. 2.32 and dh from eq. 2.36 into eq. 2.22 yields,
ds = CP0dT/T – RdP/P
This can be integrated between states 1 and 2 to give,
s2 – s1 = CP0 ln(T2/T1) – R ln(P2/P1) (2.40)
Another useful result is obtained by studying what happens to an ideal gas with constant
specific heats when it undergoes a reversible, adiabatic (isentropic ds = 0) process. From the
definition of enthalpy, we have
h = u + Pv
dh = du + d(Pv)
Using eq. 2.36 and eq. 2.37 and differentiating eq. 2.32, this becomes
CP0dT = CV0dT + RdT

CP0 = CV0 + R (2.41)
2-35
Using this in eq. 2.38, we obtain the preliminary result,
k = (CV0 + R)/ CV0 or CV0/R = 1/(1 – k) (2.42)
Now using eq. 2.21 and eq. 2.37 for an isentropic process we get,
Tds = 0 = du + Pdv = CV0dT + Pdv (2.43)
Differentiating the ideal gas law eq. 2.32 yields,
Pdv + vdP = RdT (2.44)
Equation 2.44 can be solved for dT and this result can be substituted into eq. 2.43 to give
(CV0/R)(Pdv + vdP) + Pdv = 0
Equation 2.42 can now be used in this equation to yield,
vdP + kPdv = 0
The solution to this differential equation can be seen by differentiation to be,
Pvk = constant (2.45)
when k is a constant. This important result applies to an ideal gas with constant specific heats
undergoing an isentropic process. A direct result of eq. 2.45 for an isentropic process between
states 1 and 2s (either expansion in an ideal turbine or compression in an ideal compressor) is
that,
(P2s/P1) = (v1/v2s)k (2.46)
Using the ideal gas law eq. 2.32 with eq. 2.46, it can also be shown that,
(T2s/T1) = (P2s/P1) (k – 1)/k = (v1/v2s) k – 1 (2.47)
Example 2.17 Evaluate the validity of using the ideal gas model with constant specific
heats to obtain the work produced by an auxiliary turbine that receives steam at 20 bar
and 370˚C and exhausts the steam at 1 bar and 120 ˚C.
2-36
Neglecting the heat losses and the kinetic and potential energy changes, the First Law eq.
2.11 gives
w = hi – h e
From the Tables of Thermodynamic Properties,

hi = h (2 MPa, 370˚C) = 3181.4 kJ/kg
he = h (0.1 MPa, 120˚C) = 2716.6 kJ/kg
giving w = 3181.4 – 2716.6 = 464.8 kJ/kg
Using the ideal gas model for steam, CP0 = 1.8723 kJ/kg K from Table 2.3 and the
change in temperature is 370 – 120 = 250˚C = 250 K so eq. 2.39 yields,
w = he – hi = CP0 (Te – Ti) = 1.8723 kJ/kg K • 250 K = 468.1 kJ/kg
The estimate is within 0.7%. The model happens to be very good in this case because the
steam is superheated and in the region of the Mollier chart Fig. 2.13 where enthalpy is
primarily a function of temperature only. In general, the accuracy of this model is not
this good and we will, therefore, not use the ideal gas model further for steam.
Example 2.18 The compressor portion of a diesel engine turbocharger compresses dry air
from 1 atm and 20˚C to 2.5 atm. If the compressor is 70% efficient, determine the work
required to drive the compressor and the exhaust temperature of the air as it leaves the
compressor. Use the dry air ideal gas, constant specific heat model and Table 2.3.
The compressor will have an h-s diagram as shown in Fig. 2.15. The ideal
compressor will be isentropic so eq. 2.47 can be used to obtain its exhaust temperature;
i.e., using
Ti = 20˚C + 273.15 = 293.15 K, the absolute temperature must be used
Pes/Pi = Pe/Pi = 2.5
(k – 1)/k = (1.4 – 1.0)/1.4 = 0.2857
Across the ideal compressor, equation 2.47 yields,
Tes = Ti (Pes/Pi) (k – 1)/k = 293.15(2.5)0.2857 = 380.88 K
Neglecting heat losses and kinetic and potential energy changes, the First Law eq. 2.11
and eq. 2.39 yield for the ideal compressor,
ws = hi – hes = CP0(Ti – Tes) = 1.0035 (293.15 – 380.88) = – 88.04 kJ/kg
The component efficiency definition eq. 2.28 can now be used to obtain the work
required by the actual compressor; i.e.,
wa = ws/c = – 88.04/0.70 = – 125.77 kJ/kg
2-37
Equation 2.39 can now be used again to obtain the exhaust temperature from the actual
compressor,
wa = hi – he = CP0 (Ti – Te)
Te = Ti – wa/CP0 = 2.93.15 + 125.77/1.0035 = 418.48 K = 145.3˚C = 293.6 F
The air mass flow rate is needed to obtain the total power required to drive the
compressor(s). The specific air flow rate for a large low-speed diesel is about 8 kg/kW h
so the turbocharger compressor(s) for a 20 000 kW (26 820 hp Br) low-speed propulsion
diesel would require,
power = 8 kg/kW h • 20 000 kW/3600 kJ/kW h • 125.77 kJ/kg = 5589.8 kW
This power is provided by the engine exhaust gas driven gas turbine part of the
turbocharger(s). The low-speed diesels are highly turbocharged; here the turbocharger
power is about 28% of the engine output.
2.6 Availability and Irreversibility
The thermodynamic concepts of availability and irreversibility are becoming increasingly

useful in marine engineering applications due to the attention being paid to improved efficiency
and to the development of waste heat recovery systems for diesels and gas turbines. Here we
will introduce the concepts for a non-cyclical, steady-flow, steady-state process in which kinetic
energy and potential energy changes can be neglected. In waste heat applications, we have a
source of exhaust steam, exhaust gas, or cooling water; and wish to utilize its energy content to
provide additional shaft power, generate electricity, heat cargo, distill water, etc. Such processes
will exchange heat only with a single reservoir, the environment. The treatment given here
draws upon presentations given by Sonntag, Borgnakke, and Van Wylen [1] and Haywood [6].
A corollary of the Second Law of Thermodynamics states that in a system that exchanges
heat with a single reservoir, the work done will be the same for all reversible processes between
the same fixed end states. This reversible work will be the maximum that any process could
achieve. If the final state is in equilibrium with the environment, this reversible work is the
maximum amount that is available in the incoming fluid for use. This reference amount of
available reversible work represents for an open system what the Carnot cycle efficiency
represents for a cycle operating between two reservoirs. In both cases, the condition of the
2-38
environment or the low temperature reservoir sets the maximum capability of the process or
cycle. Any real cycle or process will only be able to achieve an amount less than this ideal.
Consider the open system shown in Fig. 2.17. The system receives flow at an absolute
temperature T1, h1, and s1 and exhausts flow at T2, h2, and s2. The process is internally and
externally reversible. It produces work and exchanges heat with the environment that is at an
absolute temperature T0, h0, and s0. Neglecting the kinetic and potential energy changes, the
First Law eq. 2.11 yields,
qrev + h1 = h2 + wrev (2.48)
If the process is externally reversible, the heat transfer must take place at an infinitesimal
temperature difference; i.e., at a constant T0, so the Second Law for a reversible, isothermal
process eq. 2.19 yields,
qrev = T0 (s2 – s1) (2.49)
Combining eqs. 2.48 and 2.49, the reversible work obtained from this process is,
wrev = (h1 – T0s1) – (h2 – T0s2) (2.50)
This is the maximum amount of mechanical work that could be extracted in this process given
the existing environmental conditions T0. This is analogous to the Carnot cycle for a cycle.
wrev
T 1 , h1 , s T2 , h 2, s2
1
T0
m m
qrev
T0, h 0 , s0
Figure 2.17 Internally and Externally Reversible Open System
2-39
If the exhausting fluid in the above process is in equilibrium with the environment (state
2 equals to state 0), the process will extract all the available reversible work from the incoming
flow. Thus, rearranging eq. 2.50,
wrev = (h1 – h0) – T0 (s1 – s0) (2.51)
This is the available (or maximum possible) reversible work for the incoming flow given the
existing environmental conditions. This characteristic of the flow is called the availability of the
fluid per unit mass and given the symbol . If we drop the subscript 1, a fluid at any state will
have the availability,
= (h – h0) – T0 (s – s0) (2.52)
In Europe, particularly Germany, this quantity of maximum extractable work is called the exergy
of the fluid and given the symbol E or  = m. Note that this is not a thermodynamic property
because it depends upon the state of the environment as well as the state of the fluid. Of the total
enthalpy drop between that of the incoming fluid and equilibrium with the environment, only 
is actually available for work; the amount T0(s – s0) is unavailable by virtue of the Second Law
of Thermodynamics. In Europe, this unavailable energy is called anenergy or anergy. Figure
2.18 shows this situation on an h-s diagram. Using eq. 2.22, it can be seen for a constant
pressure process, dP = 0, dh/ds = T, where T is the absolute temperature. The slope of an isobar
(constant pressure line) on the h-s diagram is, therefore, given by,
(∂h/∂s) P = T
The tangent to the P0 isobar at state 0 has slope T0 and this tangent divides the total enthalpy
drop h – h0 at s into its available and unavailable parts.
The maximum available work from the open process in Fig. 2.17 can be seen from eqs.
2.50 and 2.52 to be equal to the change in the availability between states 1 and 2; i.e.,
wrev1-2 = 1 – 2 = [(h1 – h0) – T0(s1 – s0)] – [(h2 – h0) –T0(s2 – s0)]

= (h1 – h2) – T0(s1 – s2) (2.53)
2-40
In this result, the second term is positive since the Second Law requires that for any real process
s2 > s1. Equation 2.53 can be rearranged to give the actual or irreversible work from the process,
w = (h1 – h2) = (1 – 2) + T0(s1 – s2) (2.54)
Thus, the first term on the right is the available reversible work for the process between states 1
and 2 and the second term is the lost work due to the irreversibilities of the process. Figure 2.19
show this situation on an h-s diagram.
h
P
h
P0
availability  exergy
h – h0
slope T0
T0(s – s0) , anergy

h0 unavailable
s
s0 s
Figure 2.18 Availability, Exergy, and Anergy on h-s Diagram
Note that instead of using eq. 2.27; i.e.,
 = (h1 – h2)/(h1 – h2s) isentopic efficiency
to define the efficiency of the process shown in Fig. 2.19, we could also use a definition of
efficiency based upon the available reversible work in the presence of the existing environmental
conditions (T0, h0); i.e.,
a = (h1 – h2)/(h1 – h2a) Second Law efficiency
Either approach is valid as long as the basis for defining the reference process is used
consistently. We will use the eq. 2.27 efficiency definition unless specifically stated otherwise.
2-41
h
P1
1 P2
h1
available actual work

energy h1 – h2
2
h2  -  2a
 P0
2s
slope T0 slope T0
lost work due to
h0 Irreversibility
0 T0(s2 – s1)
s0 s1 s2 s
Figure 2.19 h-s Diagram for Process between States 1 and 2
We noted above that the second term on the right in eq. 2.54 was the work lost due to the
irreversibilities in the actual process. On a total basis, the quantity is called the irreversibility
with the symbol I. We can develop this more completely here by using eq. 2.11 for a general
steady-flow, steady-state process. The actual work output from a process between states 1 and 2
with negligible kinetic and potential energy changes will be,
1w2 = (h1 – h2) + 1q2
and the reversible work available from the process will be given by eq. 2.53; i.e.,
wrev = 1 – 2 = (h1 – h2) – T0 (s1 – s2)
The irreversibility will be the difference between these two quantities,
I/m = wrev – 1w2 = (h1 – h2) – T0 (s1 – s2) – (h1 – h2) – 1q2
= T0 (s2 – s1) – 1q2 (2.55)
2-42
If a steady-state, steady-flow process had more than one incoming and outgoing flow, the
total reversible work and irreversibility become, respectively,
Wrev =  mii –  mee (2.56)
I =  meT0se –  miT0 si – Qcv (2.57)
where Wrev and Qcv are the total reversible work and heat transfer, respectively, crossing the
control volume boundary.
Example 2.19 The concept of availability can be used to gain a first understanding of the
problem of recovering waste heat from a low-speed diesel engine. A particular 10
cylinder, low-speed diesel engine operating at the Maximum Continuous Rating (MCR)
produces 35,100 kW. A sketch of the engine is shown in Fig. 2.20. When the ambient
air temperature is 25˚C and the ambient sea water is 15˚C, the energy produced by the
combustion of the fuel appears approximately as follows:
component power form

50.00% shaft work 35,100 kW
24.17% exhaust gases 16,965 kW 225,800 kg/h exhaust @ 320˚C
9.66% jacket cooling 6,784 kW 430,000 kg/h fresh water @ 68.0˚C
2.87% piston cooling 2,015 kW 105,000 kg/h fresh water @ 60.9˚C
11.48% intercooler cooling 8,062 kW 565,000 kg/h sea water @ 44.3˚C
0.57% lube oil cooling 396 kW sea water from heat exchanger @ <38˚C
1.25% radiation losses 878 kW heating to engine room
100.00% total from fuel 70,200 kW
For maximum total system efficiency, it is desirable to recover and utilize as much as
possible of the energy in the exhaust gases and the cooling fluids for the engine. We will
neglect the lube oil cooling and radiation losses as small. The environment can be taken
as the seawater at 1 atm. and 15˚C so,
From the Tables T0 = 15˚C + 273.15 = 288.15 K

h0 = h (0.10135 MPa, 15˚C) ≈ hf (15˚C) = 62.98 kJ/kg
s0 = s (0.10135 MPa, 15˚C) ≈ sf (15˚C) = 0.2245 kJ/kg K
The four principal sources of waste energy contain the following energy per hour:
jacket cooling 6,784 kW • 3600 kJ/kW h = 24.422 GJ/h

piston cooling 2,015 kW • 3600 kJ/kW h = 7.254 GJ/h
intercooler cooling 8,062 kW • 3600 kJ/kW h = 29.023 GJ/h
exhaust gases 16,965 kW • 3600 kJ/kW h = 61.074 GJ/h
2-43
stack
exhaust gas boiler
exhaust intake air
turbocharger(s)
cylinder intercooler
jacket cooling T
cooling
piston 320˚C exhaust gases
280˚C
lube oil piston 160˚C
cooling cooling 5.4 bar steam
generation
kJ/h
exhaust gas boiler temperatures
low speed diesel engine
Figure 2.20 Sketches for Example 2.19
The exhaust gases contain the greatest amount of energy. The combustion of the
hydrogen in the fuel yields water vapor; the combustion of the sulfur impurity in the fuel
yields SO2. In the exhaust gases, which contain excess oxygen, some of the SO 2 is
converted to SO3. If the exhaust gases are cooled too far (below about 130-140˚C
depending upon the %S and %O2) sulfuric acid condenses out and corrodes the exhaust
path. This “cold end corrosion” consideration limits the minimum exhaust gas
temperature to about 160˚C. The specific heat of the exhaust gases is 1.05 kJ/kg K so
due to this limitation the recoverable content is not actually 61.074 GJ/h, but
oQi = meCP(To – Ti) = 225,800 kg/h • 1.05 kJ/kg K • 160 K = 37.934 GJ/h
= 10,537 kW
This consideration has reduced the potentially “available” energy by 38%. The exhaust
gases after the turbocharger are at too low a pressure to be useful in a second gas turbine,
so they are used in a waste heat boiler to generate steam as shown in Fig. 2.20. For
example, this heat could be used to preheat, evaporate, and then superheat 14,600 kg/h of
steam with outlet conditions of 5.4 bar and 280˚C. The finite temperature difference
necessary in the boiler between the exhaust gases and the water and steam introduces
additional irreversibilities and lost work. We can evaluate the total availability of the
2-44
steam produced in the waste heat boiler and the total availability of the cooling flows to
better quantify the ideal or maximum amount of reversible work that could be recovered.
Using saturated fluid properties as reasonable approximations for the sub-cooled liquids,
we get
jacket cooling hj ≈ hf (68˚C) = 284.61 kJ/kg

sj ≈ sf (68˚C) = 0.9304 kJ/kg K
j = (hj – h0) – T0(sj – s0) = (284.61 – 62.98) – 288.15 (0.9304 – 0.2245)

= 18.225 kJ/kg
mjj = 430,000 kg/h • 18.225 kJ/kg = 7.837 GJ/h = 2176.9 kW
piston cooling hp ≈ hf (60.9˚C) = 254.89 kJ/kg

sp ≈ sf (60.9˚C) = 0.8425 kJ/kg K
p = (hp – h0) – T0(sp – s0) = (254.89 – 62.98) – 288.15 (0.8425 – 0.2245)

= 13.833 kJ/kg
mpp = 105,000 kg/h • 13.833 kJ/kg = 1.452 GJ/h = 403.5 kW
intercooler cooling hi ≈ hf (44.3˚C) = 185.61 kJ/kg

si ≈ sf (44.3˚C) = 0.6297 kJ/kg K
i = (hi – h0) – T0(si – s0) = (185.61 – 62.98) – 288.15 (0.6297 – 0.2245)

= 5.872 kJ/kg
mii = 565,000 kg/h • 5.872 kJ/kg = 3.317 GJ/h = 921.5 kW
generated steam hs = h (0.54 MPa, 280˚C) = 3021.7 kJ/kg

ss = s (0.54 MPa, 280˚C) = 7.3494 kJ/kg K
s = (hs – h0) – T0(ss – s0) = (3021.7 – 62.98) – 288.15 (7.3494 – 0.2245)

= 905.68 kJ/kg
mss = 14,600 kg/h • 905.68 kJ/kg = 13.223 GJ/h = 3673.0 kW
These estimates represent the maximum reversible work that could be recovered using
open systems exchanging heat with the environment. Any real systems would, of course,
produce even less.
In practice, the fresh water cooling and steam systems are closed systems (cycles) to
eliminate the need to continuously generate water. Some energy is also recovered in the
2-45
form of heat rather than mechanical work. This analysis does reveal the fundamental
problem of recovering waste energy from a large diesel or gas turbine and yields the
theoretical upper bounds. In terms of energy rate (kW), an upper bound on the
recoverable energy in then
source energy content recoverable energy percent of fuel content

exhaust gases 16,965 kW 3,673.0 kW 5.23 %
jacket cooling 6,784 kW 2,176.9 kW 3.10 %
piston cooling 2,015 kW 403.5 kW 0.57 %
intercooler cooling 8,062 kW 921.5 kW 1.31 %
total recovery possible 7,174.9 kW 10.21 %
Thus, even though the diesel is very efficient at converting 50% of the fuel energy
content to mechanical work, the best that can be expected thermodynamically for waste
heat recovery from the exhaust gases and cooling fluids losses is another 10.21%. This is
often still a cost effective approach. As noted, additional irreversibilities reduce the
actual recoverable energy even further. The effects of the cold end corrosion limit,
exhaust gas boiler finite temperature difference, and Second Law implications, however,
can be seen.
The large amount of energy in the cooling flows is, by virtue of the Second Law,
difficult to recover due to the low temperatures of the flows compared to the
environment. Current designs are able to recover a total energy equal to only about 3 to
4% of the fuel content or 6 to 8% of the shaft work. This is, however, still cost effective.
Example 2.20 An insulated (externally adiabatic) low-pressure feed heater in a steam

cycle is shown in Fig. 2.21. Saturated vapor steam extracted from the turbine at state 1 is
condensed in the shell of the heat exchanger to heat condensate from state 3 to state 4.
The shell drain, state 2, is saturated fluid. Evaluate the irreversibility of this heat
exchange process per kg of condensate due to the finite temperature difference required
across the heat exchange surface. The environment is 1 atm. and 25˚C.
from the Tables: T0 = 25˚C + 273.15 = 298.15 K

h1 = hg (86˚C) = 2653.6 kJ/kg
h2 = hf (86˚C) = 360.10 kJ/kg
h3 = h (0.50 MPa, 33˚C) = 142 .81 kJ/kg
h4 = h (0.49 MPa, 80˚C) = 338.77 kJ/kg
s1 = sg (86˚C) = 7.5445 kJ/kg K
s2 = sf (86˚C) = 1.1460 kJ/kg K
s3 = s (0.50 MPa, 33˚C) = 0.4743 kJ/kg K
s4 = s (0.49 MPa, 80˚C) = 1.0721 kJ/kg K
Qcv = 0 and Wcv = 0
2-46
Using the First Law eq. 2.8 to find  = m12/m34, the amount of extraction flow needed per
kg of condensate to produce the desired conditions,
 Hin = m12 h1 + m34 h3 = m12 h2 + m34 h4 =  Hout
= m12/m34 = (h4 – h3)/(h1 – h2) = 0.08544
Using eq. 2.57 to evaluate the irreversibility per kg of condensate,
I/m34 = T0(s4 + s2) – T0(s3 + s1)

= 298.15 (1.0721 + 0.08544 • 1.1460 – 0.4743 – 0.08544 • 7.5445)
= 15.24 kJ/kg
Note that this irreversibility can be used to evaluate heat exchanger performance or as a
rational quantitative measure of merit (demerit) to be minimized in the optimization of
the heat exchanger design.
extracted heating steam
saturated vapor
1 86˚C
4.9 bar, 80˚C 5 bar, 33˚C
m34
4 3
heated condensate flow
2 saturated fluid
m12 = m34 86˚C
Figure 2.21 Sketch for Example 2.20
References
1. Sonntag, R. E., Borgnakke, C. and Van Wylen, G. J., Fundamentals of Thermodynamics, 5th
Edition, John Wiley & Sons, Inc., New York, 1998.
2. “Metric Editorial Guide,” SNAME Edition, American National Metric Council, January
1978.
3. Keenan, J. H., Keyes, F. G., Hill, P. G., and Moore, J. G., Steam Tables: Thermodynamic
Properties of Water including Vapor, Liquid, and Solid Phases (International System of
Units – SI), John Wiley & Sons, Inc., New York, 1978.
4. Harrington, R. L. (ed.), Marine Engineering, SNAME, New York, 1971, p. 679.
5. “Marine Steam Power Plant Heat Balance Practices,” SNAME Technical and Research
Bulletin #3-11, SNAME, New York, Feb. 1972.
6. Haywood, R. W., Analysis of Engineering Cycles, Pergamon Press, Oxford, UK, 1967.
2-47
2-48
Chapter 3 Marine Fuels
Marine propulsion fuel choice is again in transition having gone from sail to steam (coal)
in the 1800’s and then to liquid petroleum in the early 1900’s. Liquid fuels can range in
viscosity and density from gasoline (petrol) used in outboard motors and automotive type
inboards in small craft to Marine Gas Oil (MGO), Light Diesel Oil (LDO), Marine Diesel Oil
(MDO) through the various blended Intermediate Fuel Oils (IFO) to the refinery residual
Heavy Fuel Oil (HFO) or Bunker C.
Fuel cost and increasingly strict air emission requirements are now producing a transition
to natural gas or dual fuel (gas and diesel) in many applications. With the development of
Liquefied Natural Gas (LNG) cargo vessels in the early 1970’s, LNG cargo boil-off began to
be used in steam and then diesel plants. The first non-LNG cargo vessel to use LNG fuel was
the Norwegian ferry Glutra delivered in 2000. Today there are about 300 LNG cargo vessels
using natural gas cargo boil-off. In July 2014, there were 48 LNG powered non-LNG cargo
vessels, mostly ferries and offshore platform supply vessels, with another 53 on order.
Naval nuclear powered vessels (carriers and submarines) use metallic or oxide uranium
as their basic fuel supply. Fuel cells using hydrogen or hydrogen reformed from hydrocarbon
fuels are used in some submarines or hybrid vessels. Coal is still used in a few older vessels.
Waterfront Shipping of Vancouver, B.C., has recently signed a letter of intent to build
four methanol carriers that use low-speed dual-fuel diesel engines that will burn 95% methanol
and 5% diesel. Stena Line is also converting the diesel ferry Stena Germanica to operate on
methanol-MGO/diesel dual fuel. Ocean Yield ASA (Oslo) has ordered three ethylene gas
carriers that will use ethane-diesel dual fuel engines. Other proposals are to use LPG
(propane), dimethyl ether (DME) and bio-ethanol.
Classification of Liquid Fuels. One way to classify fossil fuels is based on the processing.
Crude Oil is unrefined petroleum. It is suitable for fuel in boilers and low-speed diesel
engines, but it is generally not used today because it contains lighter constituents that present
an explosion hazard. Crude oil contains all of the low flash-point constituents that are removed
early in the refining process to produce gasoline, etc. It, therefore, presents some of the same
hazards as gasoline.
3-1
After the refining process removes the gasoline and other valuable light distillate
products, the remaining oil is termed a residual oil [1, 2]. This oil is heavy, both in terms of
density and viscosity, and contains the concentrated impurities from the crude oil and
additional residual material introduced as part of catalytic cracking process, so-called “cat
fines” such as aluminum and silicon. The residual will also contains the concentrated sulfur -
that contributes to SOx air pollution; vanadium - that can lead to fouling of boiler and engine
surfaces; and other constituents that may be limited by fuel specifications and air emissions
requirements. Residual oil is attractive, of course, because of its low cost and it is the fuel of
choice in marine boilers – Bunker C, now RMG, is the universal fuel for steam ships. Low-
speed diesel engines are qualified to burn residual, but more likely they use one of the
intermediate fuels that are easier to handle and process shipboard.
The distillate fuels are oils that have been extracted from the crude oil by atmospheric
distillation, vacuum distillation and/or visbreaking, major steps in modern refining [2].
Aircraft jet fuels (JP-4 and JP-5, basically kerosene) are the lightest in density, viscosity, and
color and occasionally used in marine gas turbines. Marine Gas Oil (MGO) is a slightly
heavier distillate that is often used in gas turbines and high- and medium-speed diesels.
Marine Diesel Oil (MDO), another common diesel that is heavier than MGO, is a common fuel
for smaller medium-speed diesels. Both MGO and MDO are used in marine gas turbines, but
usually with tighter specifications on impurities that may exclude some of these fuels from use
on aero-derivative gas turbines.
Intermediate Fuel Oil (IFO) is a mixture of residual oil and one of the distillates designed
to provide a cheaper fuel, but one with a viscosity low enough for easier shipboard handling.
IFO is used in low-speed and larger medium-speed diesels. There are several common grades
of intermediates indexed by their viscosity. The intermediates are typically mixed dockside as
they are delivered to the ship using residual and distillate streams to provide the properties
desired. The cost is basically a weighted average of the cost of the constituents used to create
the intermediate.
Table 3.1 shows recent liquid fuel costs at selected ports in the world [3]. IFO380 has a
kinematic viscosity of 380 centistokes (1 cSt = 1 mm2/s) at 50˚C; IF180 has a viscosity of 180
cSt at 50˚C. Units are USD ($) per metric tonne (t).
3-2
The variability and volatility of fuel prices over location and time is significant and this
provides an opportunity for strategic and tactical optimization of fuel purchasing. For
comparison the Houston prices for IFO180 and MDO (~MGO) were $174.00/t and $246.50/t,
respectively, on October 14, 2003 and $334.00/t and $557.50/t, respectively, on June 12, 2006.
Table 3.1 Marine Fuel Prices March 18, 2015 [www.bunkerworld.com/prices/ 3]
Port IFO380 IFO180 MDO MGO
Fujairah, Persian Gulf 315.00 $/t 344.00 -- 745.50

Houston 286.50 443.50 -- 785.50
Rotterdam 274.00 296.00 -- 492.00
Singapore 306.00 322.50 514.50 524.50
Liquid Fuel Specifications: Various specification systems are in use world-wide for marine
fuels. The U.S. standards are either American Society of Testing and Materials (ASTM) or
military specifications (Mil-Specs). The ASTM system publishes specifications for diesel, gas
turbine, intermediate, and furnace fuels. Within each there are classes of fuels designated by a
number plus an abbreviation for the use. Diesel specifications are 1-D, 2-D, and 4-D. Gas
turbine specifications are 1-GT, 2-GT, 3-GT, and 4-GT. “Number 2 diesel”, the common
diesel used by cars and trucks, is a colloquialism for a “fuel that meets the ASTM 2-D
specification.” The number categories (2-D and 2-GT) are similar with respect to viscosity,
but differ with respect to the control of impurities. The number 6 furnace oil is basically the
residual oil or Bunker C.
The Mil-Specs specify requirements for the U. S. Navy. For the many years of steam
boiler use, Navy Special Fuel Oil (NSFO) was used as a relatively clean, blended residual with
optimum energy density on a volume basis. It had a viscosity that would place it today as an
intermediate fuel on the scale designated as Intermediate 4 (kinematic viscosity of 400 seconds
Redwood at 100 F). Navy Diesel is similar to No. 2 Diesel, but permits a somewhat higher
viscosity than ASTM 2-D. The current international naval fuel for both gas turbines and
diesels is the NATO F-76 distillate.
In Europe, fuel specifications have been issued by individual countries and
internationally by the Conseil International de Machines a Combustion (CIMAC). The current
3-3
specifications are published as International Standard ISO 8217. Table 3.2 adapted from
Woud and Stapersma [4] outlines these classifications. The ISO standards specify the density
at 15˚C, kinematic viscosity at 50˚C, flash point, pour point, carbon residual, and ash, water,
sulfur, vanadium, and aluminum limits for each fuel [4]. RMG 35 (35 cSt at 100˚C) is still
referred to as HFO 380 (380 cSt at 50˚C) in common practice.
Table 3.2 ISO Designations for Marine Fuels [adapted slightly from 4]
Product ISO Designations

Distillate products (DM = distillate marine)
Gaseous fuels
methane (when liquefied, LNG)
propane and butane, which make up Liquid Petroleum Gas (LPG)
Light fuels
gasoline
kerosene
gas oil
Diesel fuels
Marine Gas Oil (MGO) DMX
Light Diesel Fuel Oil (LDF or LDO) DMA
Marine Diesel Fuel Oil (MDF or MDO) DMB
Blended Marine Diesel Fuel Oil (BMDF) DMC
light distillate with up to 20% residual
Residual products (RM = residual marine)
Intermediate Fuel Oils (IFO) RMA to RMF
Residual blended with up to 40% distillate
Heavy Fuel Oil (HFO), also Marine Fuel Oil MMFO), RMG to RMK
Bunker Fuel Oil (BFO), Bunker C
Liquid Fuel Properties. The principal liquid fuel properties will be outlined here with
emphasis on those that affect the overall ship design and application.
Density. Fuel density can be specified as 15 (at 15˚C) in (kg/m3) or specific gravity SG
(non-dimensional or t/m3 = kg/m3/1000), API gravity, or various other sets of units. Generally,
light fuels are much lighter than water (e.g. gasoline SG = 0.729, JP-5 SG = 0.817, No. 2 diesel
SG = 0.852) while the heavier fuels approach the density of water (e.g. IFO 380 SG = 0.944
and Bunker C SG = 0.972). Density is important for shipboard storage requirements and fuel
treatment (centrifuging) effectiveness.
3-4
Petroleum crude (cargo) and fuel products often have their density specified in units of
API gravity. This is measured using a direct reading from an American Petroleum Institute
(API) standard hydrometer. The draft of this hydrometer in the fluid at 60 F determines the
value. Thus, the scale is inversely related to density. The modulus of the hydrometer is set so
that API gravity 10 corresponds to a specific gravity of 1.0 [t/m3]. The API gravity and
specific gravity are related by,
Specific gravity (SG) = 141.5/(131.5 + API gravity)
Viscosity. Viscosity is typically measured by the time it takes the fluid to pass through a
designated orifice at a particular temperature. The viscosity increases with density and
decreases with temperature. It is important for pumping, centrifuging, fuel injection and
atomization. There are numerous scales in use with the most common with respect to marine
–6
fuels the kinematic viscosity in centistokes (cSt = 10 m2/s = 1 mm2/s) at 50˚C or 100˚C and
in seconds Redwood 1 at 100 F. By the ISO standard, the viscosity for heavy fuels should be
specified in cSt at 100˚C. Woud and Stapersma list the following upper limits for shipboard
operations [4]:
pumping  ≤ 500 cSt
separation by centrifuging  ≤ 40 cSt
fuel injection or atomizing  ≤ 15 cSt
Fuels are heated to produce the viscosity needed for a particular task. Thus, a heavy fuel oil
might be heated to 100 – 150˚C for fuel injection in a large diesel engine.
A chart of the viscosity of various liquid fuels in the most common units as a function of
temperature is shown in Fig. 3.1. Some common marine fuels are shown. The intermediate
fuel designations shown are the ASTM designations where Int. 15 indicates an intermediate
with viscosity of 1500 second Redwood at 100 F; essentially an IFO 380. The ranges of
viscosity for atomization in marine boilers (15 cSt ≤  ≤ 60 cSt) and diesel fuel injection (8 cSt
≤  ≤ 27 cSt) are shown to reveal the temperature to which the various fuels have to be heated
for this purpose. Automatic viscosity control systems provide this required temperature fuel to
the engine. The maximum viscosity for pumping is shown in Fig. 3.1 to be about 1000 cSt.
3-5
Figure 3.1 Marine Fuel Viscosity – Temperature Relationship [Exxon]

3-6
Heating Value. The heating value for a marine fuel is measured in a bomb calorimeter or
calculated from its approximate hydrocarbon (C/H ratio) composition. It is specified as either
a Higher Heating Value (HHV or hU), which assumes that all H2O produced in the combustion
condenses to water, or the Lower Heating Value (LHV or hL), which assumes that the water
remains in vapor form. Since the chemical composition correlates with density, there is a
direct (inverse) correlation between density and heating value. Common heating values for
marine fuels are as follows:
LHV = hL HHV = hU
MGO 44,660 kJ/kg
MDO 42,700 kJ/kg
HFO 40,500 kJ/kg
Bunker C 43,380 kJ/kg
Units of kJ/kg and Btu/lbm are related by 2.326 kJ/kg = 1 Btu/lbm. The HHV can be
approximated by the following correlations if other information is lacking [5]:
HHV [kJ/kg] = 40,940 + 123.5 API gravity

= 24,698 + 17,477/SG
The heating value per unit volume increases with density even though the HHV declines with
density so the maximum energy per unit of volume of tankage on a ship can be obtained with
the heavier fuels.
Ignition Properties. The ignition properties of distillate fuels are typically specified by
the cetane number (CN). This is measured by comparing the ignition of the fuel to that of a
mixture of cetane and a low ignition quality constituent Heptamethylnonane (HMN);
specifically CN = % cetane + 0.15•% HMN. Thus, pure cetane has a value of 100; values of
CN as low as 40 to 60 can still have good ignition properties in diesel engines. The higher the
engine rpm, the higher the required cetane number.
Flash Point. The flash point is a measure of fuel safety on board ships. It is the
temperature at which it is possible to ignite the vapor above the fuel. The flash point for fuel
stored in bulk on ships must be higher than 60˚C.
3-7
Water Content. The water in the fuel is an impurity that must be centrifuged out before
use. It also devalues the fuel. Salt water contributes to sulfidation of hot parts within engines.
Carbon Residue. The Conradson test measures the carbon remaining after heating the
fuel. It measures the tendency of the fuel to form carbon deposits on exhaust valves and other
engine parts. It is low in distillates and may reach 22% in residual fuel [4].
Ash Content. The ash content measures the amount of inorganic materials such as
metals and metal oxides in the fuel. It can reach 0.2% in heavy fuels [4].
Aluminum plus Silicon Content. Aluminum and silicon (“cat fines”) can be present in
residual fuels as a remainder from the aluminum powder added as a catalytic agent during
cracking. Aluminum along with silicon particles can lead to engine wear.
Sulfur Content. Sulfur is a major contributor to air emissions through the production of
oxides of sulfur during combustion (SOx). It can also lead to sulfidation corrosion in hot
engine parts. As noted earlier, if exhaust gases are allowed to cool below 240 F, sulfuric acid
formation will occur from the water and the SO3. Since all of the sulfur in the emissions come
from that contained in the fuel, the International Maritime Organization (IMO) through the
Marine Pollution (MARPOL) agreement (and the U.S. EPA) now limit the sulfur content in all
marine fuels. This has been reduced in steps starting initially from 4.5% S as shown in Fig. 3.2
for the most recent period and near future. There is both a general (global) limit and a more
restrictive limit for Environmental Control Areas (ECAs) that currently include the North Sea,
Baltic, and the sub-Arctic coastal and inland regions of the U.S., Canada, and St. Pierre and
Miquelon within 200 nautical miles.
Vanadium Content. Vanadium can form vanadium pentoxide V2O5 that can lead to
deposits on boiler tubes and engine cylinders and exhaust valves below its melting point
675˚C. It is highly corrosive.
3-8
Figure 3.2 IMO MARPOL Marine Fuel Sulfur Limits [6]
The properties of fuels vary continuously over time and location. The user must be
cautious about what fuel is taken on board. Various organizations offer fuel analysis services
and some analyses are done shipboard. Figure 3.3 shows a sample of a table published
periodically by the ABS Oil Testing Service from its IFO180 and IFO380 testing. The ISO
8217 standards are given at the top for comparison. The density is in kg/m3 or SG•1000. The
water, ash, and sulfur are given in mass percent. The aluminum, silicon, and vanadium
impurities are given in mg/kg or parts per million by mass. Note the high variability with
location. Similar variability is seen over time.
Natural Gas. Natural gas is an attractive, economical, clean fuel, but in mobile applications it
has the disadvantage of large volumes required for storage. Since natural gas is not treated in
standard marine engineering texts (e.g. [1] and [4]), this section, mostly adapted from [6], will
be somewhat more involved than above.
On Liquefied Natural Gas (LNG) carriers the fuel is readily available as boil-off
produced by the warming of the liquefied gas cargo. This can be re-liquefied in a refrigeration
plant or it can be used in a dual-fuel diesel engine, gas turbine, or marine boiler. A number of
short sea/coastal/harbor vessels now also use natural gas for fuel cost and air emissions
reasons. Natural gas obtained with the crude oil stream on offshore platforms can be used in
power generation plants, in lieu of its being flared as a waste product, if there are insufficient
quantities for export.
3-9
Figure 3.3 Sample Bunker Quality Report by ABS Oil Testing Service
Natural gas is a relatively safe, non-toxic, lighter than air fuel. Its principal
characteristics are summarized in Table 3.4. Ignition temperature is 600˚C, relatively high
compared to diesel (250˚C). There is no visible smoke and no sludge deposits. It is flammable
only in the concentration range of 5-15% in air.
3-10
Table 3.4. Principal Characteristics of Natural Gas (adapted from [7])
Natural gas is primarily methane (CH4) and its content and properties vary with source.
The North American supply is higher in methane than most external sources. A greater
hydrogen to carbon ratio gives natural gas a 20 to 25% reduced production of Green House
Gases (primarily CO2) compared to oil. When it is liquefied to -162˚C for transport and
storage, the sulfur is removed which means that the SOX are eliminated as well. It is clean
burning with significantly reduced NOX naturally meeting IMO (MARPOL) requirements
(more later). Particulates are essentially eliminated compared with oil-based fuels. These
emissions improvements are summarized for a dual-fuel (99% natural gas with a 1% MDO
pilot) engine on LNG relative to a HFO burning engine in Fig. 3.4.
100
80
HFO Engine
Percent
60
40
Dual Fuel
20 Engine on LNG
0
Figure 3.4 Air Emissions Reduction for Wärtsilä Dual-Fuel Engine Compared to HFO
Engine (courtesy of Wärtsilä, [8])
3-11
Although all fuel price markets are volatile, the price of natural gas has declined in recent
years, primarily because of the rapid development of shale gas, and it has been relatively
steady. It is now cheaper than Heavy Fuel Oil (HFO) on a price per energy content basis;
significantly less than MDO. In early 2008, the relative price per energy content had a ratio of
about 10:13:22 USD/MBtu for LNG:HFO:MDO [8]. In North America, the natural gas market
spot price is indexed relative to the Henry Hub price which is used in gas futures trading on the
New York Mercantile Exchange. It is usually express in USD/million Btu [typically written as
$/mmBtu]. The Henry Hub is a physical location in the Sabine Pipe Line, LLC’s pipeline in
Erath, Louisiana, which is connected to 13 pipelines that cover the United States and North
America. The delivered price of LNG includes the Henry Hub commodity cost for the gas, the
cost of liquefaction, and the cost of the gas and LNG delivery; these can be roughly equal
thirds of the final price at the ship.
Natural gas can be stored onboard ships either as LNG or Compressed Natural Gas
(CNG). CNG is stored in high pressure gas bottles at about 200-248 barg (2900-3600 psig),
which requires about 6 times as much ship volume as storing LNG. LNG fuel is stored on
smaller vessels in IMO (ABS) Independent Type C pressurized cryogenic tanks at -162˚C.
Tanks located within the hull are limited to 10 barg (145 psig) maximum working pressure [9].
The storage is at 5 to 9 barg (72.5 to 132 psig) with this pressure used to move the re-gasified
natural gas into the engines without the need for feed pumps. The tanks are double-walled,
insulated, cylindrical tanks with dished ends. The inside shell is made of a cryogenic material
such as 304L stainless steel; the outer shell is either of stainless steel or carbon steel. The 5-9
in. space between the walls is, for example, insulated with perlite and vacuum [10]. The tanks
do not require cooling equipment since the LNG cools and maintains the tank temperature
through boil-off and/or pressure rise in the tank.
The lower energy content of LNG requires tankage with about 70% more net volume
than on a MDO vessel of equal range. The cylindrical LNG tanks must be supported separate
from surrounding structure with a clearance of at least 450 mm laterally and 150 mm below the
tank [9]. There is also a cryogenic material cold box (tank room) that contains heat exchangers
to control tank pressure and heat the LNG for its re-gasification and warming to about 15˚C as
needed for injection to an engine. These will result in the requirement for even more hull
volume than would be required for a typical prismatic HFO or MDO tank that is built into the
3-12
ship’s structure above the inner bottom. A Wärtsilä concept paper [11] states that the weight of
fuel storage in pressurized cryogenic tanks is about 1.5 times greater and the parallelepiped
volume typically required within a ship is about 3 to 4 times that of MDO storage above the
inner bottom.
Engine Room Safety. Due to the risk of ignition and explosion of trapped natural gas in
enclosed spaces within a ship, all recent marine LNG plants have been designed with an
inherently safe or double walled design philosophy [7, 9]. This requires the use of double
walled piping from the bunkering station to the LNG storage tank and then from the tank to
the cylinder head of the consuming engine(s). The double gas isolation valves required in the
engine room are also contained within a cryogenic material pressure vessel called a Gas Valve
Unit (GVU) or a separate room that forms the second boundary around these potential sources
of leakage. The space between the double boundaries must be inerted or ventilated at an
exchange rate of 30 times per hour. With this approach all other mechanical and electrical
equipment within the machinery spaces can be of conventional design without added
explosion-proof features.
Methane Slip. There is a small fraction of LNG that is unburned in LNG gas engines
and passes up the stack. This is called methane slip. The engine manufacturers are working to
minimize this fraction. The EPA regulates the release of non-methane hydrocarbons (HC), but
methane is not currently regulated. Methane, however, is a Green House Gas (GHG) that has
21 times greater negative impact on the environment per gram than CO2, which currently gets
most of the EPA attention. This is based upon the Global Warming Potential of CO2, as the
reference 1, and methane at 21. This is presented in the United Nations Framework
Convention on Climate Change (UNFCC) based upon each GHG’s ability, compared to CO2
per unit weight, to trap heat within the atmosphere and its decay rate within the atmosphere
over a 100 year time horizon [12].
Brittle Cracking. Any spill of LNG onto steel ship structure can result in brittle
cracking of the structure due to residual stresses since the LNG will rapidly cool the steel well
below its nil ductility temperature. This requires the installation of separate cryogenic material
drip trays large enough to hold likely spills below all bunkering stations and connections and
components that might develop a leak. If the LNG storage tank does not have a cryogenic
3-13
material outer boundary, the surrounding LNG tank room must also be made of a cryogenic
material in order to survive a leak in the internal tank boundary.
Coal. Coal was once the principal fuel for ships. The known world coal reserves are greater
than the oil reserves, and coal may reappear as a marine fuel at some point in the future. Both
diesels and gas turbines have been operated on pulverized coal and coal/fuel slurries, but this is
more problematic in internal combustion engines. Coal’s primary disadvantages are the
residual ash, difficulty in handling, and lower heat content per tonne or m3. Coal boilers also
require significantly larger furnaces. Liquid fuel has been made from coal using processes that
were developed to offset the scarcity of oil during World War II.
Nuclear Fuel. Nuclear fuel is radically different than chemical fuels. Its outstanding feature is
its great energy density – roughly 2.3 x 1010 kJ/kg or about 5 x 105 times that of petroleum fuel.
Metallic tubes containing uranium oxide pellets or bonded plates containing metallic uranium
are used in a shipboard core and then removed in a depleted condition with essentially no
visible physical change. Current U.S. Naval cores are able to last the life of the ship
eliminating the need for costly, radioactive refueling operations requiring the penetration of the
submarine or ship hull. The spent cores are, however, highly radioactive from long-lived
fission products requiring expensive reprocessing, retention, or disposal.
References
1. Winkler, M. F., “Fuels and Fuel Treatment”, Ch. XII in Harrington, R. L. (ed.), Marine
Engineering, SNAME, New York, 1992.
2. Vermeire, M. B., “Everything You Need to Know about Marine Fuels,” Chevron Global
Marine Products, July, 2007; www.chevronmarineproducts.com
3. www.bunkerworld.com/prices/
4. Woud, H. K. and Stapersma, D., Design of Propulsion and Electric Power Generation
Systems, IMarEST, London, 2002.
5. Johnston, R. M., Brockett, W. A. and Bock, A. E., Elements of Applied Thermodynamics,
U. S. Naval Institute, Annapolis, MD, 1958.
6. Parsons, M. G., O’Hern, P. J. and Denomy, S. J., “The Potential Conversion of the U.S.
Great Lakes Steam Bulk Carriers to Liquefied Natural Gas Propulsion – Initial Report,”
Journal of Ship Production and Design, 28-3: 97-111, August 2012.
7. Parsons, M. G., “Engine Room Safety for LNG Fueled Vessels: An examination of
propulsion plant design and operation,” Great Lakes/Seaway Review, 40-2: 49-51,
October-December, 2011.
3-14
8. Parsons, M. G., “LNG-Fueled Vessels – Looking at green designs for emission control
areas,” Great Lakes/Seaway Review, 38-4: 51-52, April-June 2009.
9. American Bureau of Shipping, “Guide for propulsion and auxiliary systems for gas fuelled
ships, American Bureau of Shipping, Houston, May, 2011.
10. Karlsson, S. and Sonzio, L., “Enabling the safe storage of gas onboard ships with the
Wärtsilä LNGPac”, Wärtsilä Technical Journal in detail, 1: 52-56, 2010.
11. Hannula, S., Levander, O. and Sipilä, T., “LNG cruise ferry – a truly environmentally
sound ship” (www.wartsila.com), 2009.
12. UNFCC, United Nations framework convention on climate change,
http://unfcc.int/ghg_data/items/3825, 2011.
3-15
3-16
Chapter 4 Reciprocating Prime Movers – Diesel and Natural Gas
A prime mover is a device that converts primary chemical (or nuclear) energy into
mechanical energy [1]. The principal prime movers used on ships are the diesel, gasoline or natural
gas reciprocating engine, gas turbine, and fossil fuel/natural gas fired boiler or nuclear reactor
heated steam turbine plant. The predominant prime mover on commercial ships today is the diesel
engine. They first appeared on ships with the launch of the MS Selandia by Burmeister & Wain
(B&W) of Copenhagen, Denmark, in 1912. The diesel currently dominates because of its high
efficiency, high reliability, relatively simple maintenance, tolerance for low quality (cheap) fuels,
and its low capital and maintenance costs. The diesel’s primary disadvantages are its air emissions
(oxides of nitrogen = NOx, oxides of sulfur = SOx, and particulates) relative to natural gas engines
and low power density [kW/t] relative to gas turbines. Marine diesels currently range in power
from a few kW to 87.22 MW (116,964 hpBr). We will study the classification, operating principle,
thermodynamics, and operating constraints of the diesel and natural gas engines used on board
ships.
Before discussing diesel engines specifically, it is useful to review some of the classifications
that are applied to prime movers in general. They can be internal combustion (IC) devices, where
the combustion of the fuel occurs within the working fluid (diesels, gasoline engines, natural gas
engines, and gas turbines), or external combustion devices, where the combustion occurs outside
of the working fluid (steam boilers and Sterling cycle engines, which are used today for Air
Independent Propulsion (AIP) on some Swedish built non-nuclear submarines). Prime movers can
be reciprocating engines using pistons (diesels, gasoline engines, natural gas engines, or Sterling
cycle engines) or rotating engines (gas turbines or steam turbines). Diesels and natural gas engines
are reciprocating, internal combustion prime movers. Natural gas can also be used in gas turbines
or in marine propulsion or auxiliary boilers.
Classification. Reciprocating engines can be classified a number of ways. There are two primary
types of engines based upon their basic mechanical construction. They can be either crosshead or
trunk piston engines. These arrangements are shown in Fig. 4.1. All low-speed marine engines
are of a crosshead arrangement in which the piston is connected to a piston rod that only moves
vertically. The piston rod is connected to a vertically sliding crosshead at its lower end. The
4-1
crosshead is connected to the crank pin (crank web) by a connecting rod. These engines are
currently only built in an in-line (L) arrangement. All medium-speed and high-speed marine
engines are of a trunk piston design typical of the common automotive engines. In this case, there
is no crosshead and no piston rod. The piston is connected directly to the crankshaft pin by a
connecting rod. These engines can be either in-line (L) or V-arrangements (V) in which two
pistons, one for each bank, are connected to each crank pin. Some trunk piston engines are
constructed with two pistons in each cylinder operating in opposite directions with the combustion
region between them – the opposed piston engine configuration. These pistons are connected to
upper and lower crankshafts that are geared together at the end of the engine.
Figure 4.1 Crosshead and Trunk Piston Engine Arrangements [1]
The crosshead engines have the advantage of (1) less piston-cylinder liner wear because the
crosshead rather than the piston skirt carries the lateral loads, (2) a separation of the piston region
from the crankshaft region by the crosshead seal which protects the crankshaft oil allowing the
safe use of poorer grades of fuels, and (3) this separation allows the use of cheaper crankshaft
lubricating oil since it is not exposed to fuel combustion products as is the more expensive cylinder
4-2
oil. In trunk piston engines, only one type of the oil is used and it must have the more expensive
cylinder oil additives. The main disadvantage of the crosshead engine is a much greater height
and, thus, an overall larger and heavier engine even at the same engine rpm.
Internal combustion reciprocating engines can also be classified on the basis of the way the
fuel is ignited within the cylinder. In some engines, the air is compressed in the cylinder to a high
temperature without the fuel and then the relatively slow burning fuel is injected into the hot gas
which ignites the fuel. These engines are called compression ignition engines. Some engines have
the fuel and air supplied together to the cylinder after mixing in a carburetor or the cylinder head.
The fuel is then ignited by an electric spark from a spark plug – the spark ignition engine.
Internal combustion reciprocating engines can also be classified on the basis of the number
of strokes (up or down) of the piston needed to complete one thermodynamic cycle of the engine.
Low-speed diesels are all two-stroke engines in which one revolution or two strokes complete one
cycle of the combustion within the cylinder. In these engines, intake and compression are
completed on the up stroke while power and exhausting are completed on the down stroke. These
engines require a scavenging blower, supercharger, or turbocharger to facilitate the exhaust/intake
process. Today they usually have exhaust valves in the cylinder head to support the exhaust
process and use intake ports in the side of the cylinder for the intake. Medium-speed and high-
speed engines can also be two-stroke. They are, however, more commonly made as four-stroke
engines in which two revolutions or four strokes complete one cycle of the engine. In this case,
compression and then exhaust occur on the two up strokes while power and then intake occur on
the two down strokes giving the cycle of compression, power, exhaust, and intake. These engines
must have both intake and exhaust valves to control the intake and exhaust. Most automotive
engines and non-low-speed marine engines are four-stroke.
We have already used the classification terms: low-speed, medium-speed, and high-speed
diesels. These terms refer to the revolutions per minute (rpm) of the engine, but the classification
is somewhat arbitrary and vague. Table 4.1 adapted from Woud and Stapersma [1] separates the
three types at 300 and 1000 rpm. The term low-speed really applies only to the crosshead engines,
regardless of rpm, which have become increasingly high rpm in recent years and, thus, low-speed
and medium-speed (trunk piston) engines can actually overlap in their rpm ranges. The separation
between medium- and high-speed engines is fairly arbitrary and Woud and Stapersma’s 1000 rpm
value is reasonable.
4-3
Table 4.1 Typical Performance Parameters of Marine Diesel Engines (2006) [1]
Data Low-speed Medium-speed High-speed
Process 2-stroke 4-stroke 4-stroke

Construction Crosshead Trunk piston Trunk piston
Output power range [kW] 1600-97300 500-23280 15-9000
Output speed [rpm] 80-300 300-1000 1000-3500
Fuel type HFO or IFO HFO, IFO or MDO MDO
Specific fuel rate [g/kW h]* 160-180 170-210 200-220
Specific air rate [kg/kW h] 7.0-9.0 6.0-9.0 5.5-7.5
Specific NOx Emission [g/kW h] 14-22 10-18 7-13
Specific mass [kg/kW] 17-60 5-20 2.3-6
Specific volume [dm3/kW] 12-55 4-28 2.8-8
Specific cost [$/kW] L: 440-460 L: 240-360
* ISO standard on MDO V: 190-310 V: 200-260
All low-speed engines and the larger medium-speed engines typically burn Heavy Fuel Oil
(HFO) or Intermediate Fuel Oil (IFO). Operation in Environmental Control Areas (ECA) is
changing this. Smaller medium-speed engines burn only MDO or MGO, which can cost 45-80%
more than a blended fuel. For example, Houston prices on June 11, 2013, showed IF180 (4-8%
MDO depending upon the residual stock) at $650.50/t and MGO at $950.00/t (+46%) [2].
A typical crosshead low-speed marine propulsion diesel is shown in Fig. 4.2. This engine is
a MAN B&W K98-MC engine that produces 80,080 kW at 94 rpm in the 14 cylinder version.
Note the size of the crew person on the upper level. The two-stroke intake ports are indicated as
6. The receiver for the constant pressure air supply to the turbocharger (3) is indicated as 2. A
typical V-arrangement (V) medium-speed marine propulsion diesel is shown in Fig. 4.3. This
engine is a Wärtsilä V46 that produces 19,200 kW at 600 rpm in the 16 cylinder version.
4-4
Figure 4.2 Typical Low-Speed Diesel Engine – MAN B&W K98-MC
4-5
Figure 4.3 Typical Medium-Speed Diesel Engine – Wärtsilä V46
4-6
Diesels can also be classified by their aspiration scheme – the way the air is supplied to the
cylinders. A few marine engines today are still naturally aspirated meaning that there is no high
pressure mechanism other than the cylinder motion itself to force (pull) air into the cylinders. Two-
stroke naturally aspirated engines must still have a low-pressure scavenging blower to assist the
exhaust process.
Most marine diesels today have some type of supercharging to ensure that the intake air is
at a higher pressure than the exhaust backpressure in order to force compressed, high-pressure air
into the cylinders. Superchargers, specifically, are mechanical driven compressors. The most
common supercharging approach on marine diesels is to use a turbocharger in which an exhaust
gas waste heat gas turbine drives a rotating air compressor to supply high-pressure air (2 to 6
atmospheres) into the cylinders. The high temperature air is also often cooled by seawater in an
aftercooler to further increase its density before injection into the cylinder so that even more fuel
can be burned. If two-stages of compression are used, the air can also be cooled between the stages
by an intercooler. The turbocharged engines (essentially all large marine diesels since 1950) have
a much higher specific power, but they are generally slower to respond to transients since the
turbocharger must wait for an increase in exhaust flow to fully respond. The turbocharger turbines
can be supplied directly from the cylinders (pulse type turbocharging) leading to high peak
pressures, but variable pressure and lower turbine efficiencies. They can also be supplied from
exhaust receivers that smooth out the pressure from the individual cylinders (constant pressure
turbocharging) giving less energy, but better turbine efficiencies.
Operating Principle. Reciprocating engines operate on either a two-stroke or four-stroke cycle

as noted. They produce power through a torque QB applied to the output shaft at some rotating
frequency e to give a brake power,
PB [kW] = QB e QB [kN m]
= QB 2ne e [rad/s]
= QB 2 Ne/60 ne [rev/s]; Ne [rev/m = rpm]
Note that Woud and Stapersma use MB (moment) for the torque QB [1]. The torque is produced
on the crankshaft by the combustion pressure on the piston being transmitted to the crank pin
through the connecting rod.
4-7
The idealized thermodynamic model for the compression ignited engine is the Diesel cycle
that consists of the following four processes:
1-2 isentropic compression with fuel injection at about 2
2-3 constant pressure expansion with heat addition from the fuel
3-4 isentropic expansion
4-1 constant volume exhaust and intake modeled as heat rejection to the atmosphere
These processes model the real cycle whether it is two-stroke or four-stroke engine. The P-V
indicator diagram for a two-stroke diesel engine and P-v and T-s diagrams for the idealized Diesel
cycle are shown in Fig. 4.4. The fuel combustion is assumed to be in the constant pressure
expansion stage since both MDO or HFO are fairly slow burning. This idealized cycle can be used
to study the effects of the primary design parameters on the performance of a diesel engine.
Figure 4.4 Indicator Diagram for Two-Stoke Diesel Engine

and P-v and T-s Diagrams for Ideal Diesel Cycle [1]
(TDC = top dead center; BDC = bottom dead center)
The idealized thermodynamic model for the spark ignited engine is the Otto cycle that
consists of the following four processes:
1-2 isentropic compression with spark ignition at about 2
2-3 constant volume heat addition from the fuel
The P-V indicator diagram for a four-stroke gasoline engine and P-v and T-s diagrams for the
idealized Otto cycle are shown in Fig. 4.5. The fuel combustion is assumed to be in the constant
volume expansion stage since gasoline is a fast burning fuel.
4-8
Figure 4.5 Indicator Diagram for Four-Stoke Gasoline Engine

and P-v and T-s Diagrams for Ideal Otto Cycle [1]
Real diesel and natural gas engines today actually operate somewhere between these two
idealized models; i.e., the peak pressure is reached somewhere between state 2 of the Otto cycle
and state 3 of the Diesel cycle. The idealized model that reflects this situation and provides the
best model for modern marine engines is the Seiliger cycle [1], also called the Dual, Limited
Pressure, Mixed or Sabathe cycle [3, 4]. It consists of the following five processes:
1-2 isentropic compression with fuel injection at about 2
2-3 constant volume heat addition from the fuel
3-4 constant pressure expansion with additional heat addition from the fuel
The P-V indicator diagram for a typical modern two-stroke marine diesel and P-v and T-s diagrams
for the idealized Seiliger cycle or Dual cycle are shown in Fig. 4.6. In this model the pressure rises
quickly as the fuel starts to burn and then combustion continues as the piston begins to move
downward.
The cylinder in a reciprocating engine can be described by its internal diameter or bore B
and its stroke Ls from top dead center (TDC) to bottom dead center (BDC), which is also the
diameter of the path swept by the crank pin. The swept volume of the cylinder between the
maximum volume at BDC and the compressed volume at TDC is Vs = VBDC – VTDC = B2Ls/4.
The stroke-bore ratio is expressed as s = Ls/B. The geometric compression ratio is  =
VBDC/VTDC. The engine displacement of an i cylinder engine is the swept volume of all the
cylinders iVs = iALs, using cylinder area A = B2/4. The piston speed is defined as the average
linear speed per cycle or 2LsNe, not the rotational speed Ne.
4-9
Figure 4.6 Indicator Diagram for a Modern Two-Stoke Marine Diesel Engine
and P-v and T-s Diagrams for Ideal Seiliger or Dual Cycle [1]
The work per cycle from a reciprocating engine can be analyzed from the P-V diagram for
each cylinder. The work per cycle in the expansion part is the integral under the P-V diagram; the
work per cycle in the compression part is also the integral under the P-V diagram. Therefore, the
net work per cycle is the area within these processes on the P-V diagram given by the cyclic
integral,
WI = ∮ PdV
This is usually taken as the work in just the expansion and compression strokes. The net work in
the exhaust and intake processes are usually included within the mechanical efficiency of the
engine; it is positive with turbocharged engines where to turbocharger pressure is higher than the
exhaust backpressure and negative in naturally aspirated engines where the cylinder must do the
air pumping work.
Traditionally, the pressure history within the cylinder was measured mechanically by a
device called an indicator and the experimental P-V plot was, therefore, called an indicator card
or indicator diagram. The net area within this diagram was called the indicated work – the
thermodynamic work done by the cycle. This same work can be modeled as a hypothetical
constant average pressure difference between the expansion and the compression parts of the cycle
operating between the same volume limits VBDC and VTDC. This hypothetical average pressure
difference is called the indicated mean effective pressure (IMEP or PI) defined such that,
4-10
IMEP = PI = WI/Vs = ∮ PdV/Vs = ∮ PdV/(VBDC – VTDC)
The IMEP is typically expressed in bar = 105 Pa (recall that 1 bar as only slightly smaller than 1
atmosphere, about 14.5 psi).
If we consider this constant pressure as acting on i pistons, the total force on the pistons will
be given by,
F = iPIA where A = piston area = B2/4
The total thermodynamic power produced by these i cylinders as they move through their stroke
Ls, called the indicated power PI, will be given by,
PI = iPILsAN’/(60 s/min • )
Note that we use the bold here for power to differentiate it from pressure. Here N’ is the power
strokes per minute or N’ = N [rpm] for a two-stroke engine or N’ = N/2 [rpm/2] for a four-stroke
engine. The unit conversion constant  is 1000 Nm/kWs for SI units with the pressure in Pa, and
dimensions in m; or 550 ft lbf/hpBrs for English units with the pressure in psi, area in in2, and stroke
in ft. This formula has classically been called the “PLAN formula” expressing the thermodynamic
power produced in terms of engine parameters and mean indicated pressure PI.
The actual power produced by an engine will be less than the thermodynamic power due to
friction losses, intake/exhaust pumping losses/gain, valve operation, and losses to mechanically
connected lube oil, fuel, and cooling water pumps. These losses are reflected in an engine
mechanical efficiency m. The actual power delivered to the engine output connection flange is
called the brake power PB given by,
PB = 2QBN/(60 • ) = mPI = imPILsAN’/(60 s/min • ) = iPBLsAN’/(60 s/min • )
The pressure difference in the last expression is the brake mean effective pressure PB = mPI, also
indicated as BMEP. This is a useful measure of engine loading; it is a hypothetical constant
average pressure difference defined by this formula. The BMEP is proportional to the torque
produced by the engine and also the throttle (fuel) setting of the engine since all the other
parameters relating these quantities are constant for any particular engine at a particular speed
(rpm). The BMEP is limited by the thermal loading of the engine and the mass of air that can be
4-11
charged into the cylinder to burn fuel effectively. The BMEP is, therefore, heavily affected by the
level of turbocharging of the engine.
The mechanical efficiency of a reciprocating engine is about 80% for a naturally aspirated
engine (large exhaust/intake pumping work losses); about 90% for smaller turbocharged engines;
and as high as 95% for the largest two-stroke, low-speed diesels. The BMEP is in the range of 4.8
to 6.0 bar for naturally aspirated engines and in the range of 15 to 30 bar for turbocharged engines.
Since the BMEP is proportional to the power produced per unit volume of the engine, one can
readily see the dramatic impact of the introduction and development of turbocharging during the
past 50 years. For two specific examples today, the medium-speed, four-stroke Wärtsilä 9 cylinder
in-line 9L64 diesel engine that produces 18,090 kW at 327 rpm has a BMEP of 25.5 bar; the low-
speed, two-stroke Wärtsilä-Sulzer 14 cylinder 14RTA96C diesel engine that produces 80,080 kW
at 102 rpm has a BMEP of 18.6 bar.
As noted, the power output of the engine (and BMEP) is limited by the thermal loading of
the engine. This refers to the thermal stresses in the cylinder walls, pistons, exhaust valves, etc.;
oil deterioration; and loss of material strength at the high operating temperatures. The thermal
loading is more controlling with size so it limits the BMEP in low-speed diesels to lower values
than used in medium-speed engines, as seen above. A major part of the continuing development
of large diesels has been focused on improving cooling and mechanical design so that greater
thermal loading can be tolerated safely. The typical cooling of marine diesels is as follows:
cylinder walls – cooled by water jackets within the walls
pistons – where maximum temperature Tmax  B • BMEP
small size engines – pistons not cooled
medium size engines – pistons oil cooled
large engines – pistons water or oil cooled
exhaust valves – cool head near seats or eliminate valves altogether by using exhaust ports
in the cylinder walls
The crankcase lubricating oil is water cooled either directly or by recirculating the oil.

Similarity Relations. Reciprocating engine similarity relations are idealized, but they can be
useful in making engineering estimates with respect to engines with the same strokes/cycle and
aspiration type. For such engines, it can be observed that the following are approximately constant,
4-12
1. stroke-bore ratio Ls/B – set by compression ratio and combustion volume shape
2. weight-displ. ratio W/iALs – set by cylinder geometry
3. piston speed 2LsN – set by wear considerations and inertial loading
4. BMEP = PB – set by aspiration scheme and level of thermal design
As an example, suppose we want to compare two “similar” engines with respect to specific weight
W/P. We have for two engines with the same number of cylinders i = io,
PB/PBo = PBLsA N’/PBoLsoAN’o
With equal BMEP, PB = PBo and equal piston speed 2LsN’ = 2LsoN’o, this ratio becomes PB/PBo =
A/A. With a constant stroke-bore ratio and weight displacement ratio,
Ls/Lso = B/Bo = A1/2/Ao1/2
W/Wo = iALs/ioAoLso = A3/2/Ao3/2
so (W/PB)/(Wo/PBo) = (PB/PBo)1/2 or W/PB  PB1/2
This relationship is reasonably good for making estimates even though the BMEP = PB useable
actually decreases with engine size.

Torque and Power Characteristics. Reciprocating engines are essentially constant brake torque
QB machines at any constant fuel setting. We saw above for the “PLAN formula” that the engine
torque,
QB  PB = mPI
Also, the mechanical efficiency is essentially constant for all operating conditions and P I is
determined by the amount of fuel injected into the cylinders. The amount of fuel is set by the fuel
rack that controls positive displacement injection pumps with output independent of N’ so we also
have,
QB  PB   (fractional throttle, as a decimal or percent)
The ideal reciprocating engine will then have the characteristics as shown in Fig. 4.7 where here
the subscript zero (…)o indicates the rated condition of the engine. All engines have a minimum
idle speed N’min (25-60% rated) below which they cannot run because there is insufficient heat for
proper ignition. Inertial forces also limit the maximum speed N’max to which the engine can be
operated (103-110% rated).
Actual naturally aspirated diesels deviate slightly from these characteristics because,
4-13
m tends to drop off at low rpm as all the losses are not proportional to rpm
2. aspiration and fuel pumping may be less effective at higher rpm
Thus, the torque vs. rpm curve usually curves downward slightly and the PB vs. rpm curve may
drop slightly below linear with rpm at the higher rpm. For a wide range of preliminary design
work, particularly if specific vendor data is unavailable, the ideal model described above is
reasonably valid for a naturally aspirated engine and extremely valuable.
III < II < I throttle settings  or BMEP values

Torque QB/QBo =  Power PB/PBo =  N/No
Figure 4.7 Ideal Reciprocating Engine Torque and Power Characteristics
Turbocharged engines exhibit an even greater curvature in the torque vs. rpm curve than do
typical naturally aspirated engines. The turbocharger design is usually not optimized for maximum
power leading to a decrease in engine torque at higher and lower speeds and a greater curvature in
the characteristics. Further, when a turbocharged engine slows down there is not enough exhaust
gas energy to run the turbocharger at the required level resulting in a low output of air and
incomplete combustion and smoke from a diesel burning engine. It is also possible for the
turbocharger rotor rpm to become unstable at low speeds leading to an unstable situation called
compressor surge. This can damage the turbocharger and must be avoided. Diesel engines also
should not be operated continuously at too low a load (below 25 to 40% throttle) because the
cylinder walls collect unburned fuel leading to engine fouling and high maintenance. Maximum
exhaust temperature limits, maximum cylinder temperature limits, and turbocharger rpm limits
4-14
may also influence the acceptable operating region [5]. The net effect of these limits yield
operating envelopes as shown in Fig. 4.8 for naturally aspirated and turbocharged diesels.
Key: 1 = min. idle rpm limit, 2 = max. rpm limit, 3 = min. load fouling limit, 4 = maximum throttle or PB limit
(a) naturally aspirated diesel engines
Key: 1 = min. rpm limit, 2 = max. rpm limit, 3 = min. load fouling limit, 4 = maximum throttle or P B limit,
5 = turbocharger smoke and/or surge limits
(b) turbocharged diesel engines
Figure 4.8 Operating Envelopes for Typical Diesel Engines [1]
The loss of operating region for highly turbocharged modern engines can present problems
in some marine applications because the envelope does not provide enough margin with respect to
the propulsor load. Some engines are almost constant rpm engines with very severely limited
operating envelopes, necessitating the use of variable pitch propellers so the load curve can always
be kept below the compressor limit with a margin for acceleration. When used in planing hulls
4-15
that exhibit a high “hump load” as the craft comes up on a plane, the engine may not have enough
power capability at these low speeds to bring the vessel up to speed. Also when two engines are
used on one shaft, the compressor limits may unacceptably restrict the capability of the engines
when they are used individually. A number of design features can be used to expand the operating
envelope limitation provided by the turbocharger [1]:
1) pulse type turbochargers can be use in lieu of constant pressure turbochargers,
2) parallel turbochargers can be used with one cut-out at partial loads (sequential
turbocharging),
3) a waste gate can used where the turbocharger turbine is sized for less than full load and
excess exhaust can be dumped to the stack at high loads,
4) blow off can be used where excess compressor outlet at low loads is dumped to the stack,
5) a variable stator blade geometry turbine can be used to allow it to be adapted to part load
performance.
The effect of some of these approaches on the feasible torque-rpm (QB-Ne) envelope is illustrated
in Fig. 4.9 [1]. The variable geometry approach has had little use on diesel engines due to the low
grades of fuel used and the potential for fouling of movable parts within the turbine.
Figure 4.9 Effect of Turbocharging Changes on Torque-RPM Envelope [adapted from 1]
4-16
Also shown in Fig. 4.9 is the Combinator (control) curve for the Caterpillar MaK M43C 500
rpm medium speed diesel series (up to 9000 kW) illustrating one of the most restrictive torque-
rpm capabilities due to its turbocharging and current emissions requirements. This engine
obviously cannot be used with a fixed-pitch propeller. The pitch must be reduced during startup
to reduce the load until the propeller gets up to 90% rated rpm. From a practical standpoint, this
engine is almost a constant-rpm engine.
Specific Fuel Consumption. The overall efficiency of a diesel engine depends upon many factors.
The combustion of the fuel will not be completely effective. A fuel has a heat content expressed
as a Higher Heating Value (HHV or hU) and a Lower Heating Value (LHV or hL). The HHV
assumes that all the water created in the combustion is converted to liquid water; the LHV assumes
that it remains a vapor as it goes out the exhaust. In marine plants, the exhaust is intentionally
kept above the saturation temperature to prevent condensation in the exhaust path that can lead to
the formation of corrosive sulfuric acid (from the water and SO3). Thus, the LHV is appropriate
for overall diesel engine efficiency predictions. For diesel applications,
LHVMDO = 42,700 kJ/kg and LHVHFO ≥ 40,500 kJ/kg
The ISO standard used for rating diesel engines is based upon the use of MDO, so it may be
necessary to correct data for HFO use. The overall efficiency of a diesel engine in converting the
fuel’s chemical energy to useful brake mechanical energy can be express as follows [1]:
e = m • I = m • comb • q • th
where m = BMEP/IMEP is the mechanical efficiency and I is the indicated efficiency associated
with the indicator diagram; comb is the combustion efficiency that accounts for any incomplete
combustion of the fuel (nearly 100%), q accounts for the heat losses through the cylinder liner
and head into the engine room, the cooling water, and the lubricating oil; and th is the
thermodynamic efficiency of the cycle that we will consider below.
The overall engine efficiency for current marine diesels is in the range 0.38 ≤ e ≤ 0.52. This
implies a specific fuel consumption (sfc) or specific fuel rate (sfr) given by,
sfc = mf/PB = (3600 s/h 1000 g/kg mf)/PB = 3,600,000/(e • hL) g/kWh
4-17
Using the expected range of e and the lower heating values for MDO and HFO this yields the
ranges,
160 ≤ sfc ≤ 220 g/kWh for MDO with = 42,700 kJ/kg
160 ≤ sfc ≤ 235 g/kWh for HFO with = 40,500 kJ/kg
The specific fuel rate varies over the entire engine operating envelope. It tends to be the best
on the propeller operating curve at about 90% rated rpm, but specific vendor data should be sought
for each engine. The best form of this data is a contour plot of specific fuel rate over the entire
(PB, Ne) operating region. A typical specific fuel rate map for a medium-speed diesel expressed
as changes in g/kWh from the rated sfc is shown in Fig. 4.10 (Wärtsilä V46).
Figure 4.10 Typical Specific Fuel Rate Map for Medium-Speed Diesel
4-18
Note in Fig. 4.10, that the optimum specific fuel rate is at about 92% rated rpm. Some
vendors just supply curves for specific fuel rate along the propeller operating curve (assumed to
be cubic in rpm or speed, the so-called Propeller Law) and possibly also along a constant rpm line
at rated or optimum rpm. The latter is of interest for diesel generator applications where the prime
mover must operate at constant rpm in order to produce constant frequency AC power.
Annual Fuel Cost. When a plant with a reciprocating engine(s) is to be analyzed, the specific
fuel consumption sfc in g/kWh or g/ekWh for the prime movers can be obtained from vendor data
or similar engines. This should be obtained for three, four, or more power levels and the data can
be represented by a quadratic or cubic polynomial in per cent rated power % using the
Regression add-on in Excel or with a similar tool. A quadratic fit requires three or more data
points; a cubic requires at least four data points. Do not be tempted to use a higher polynomial
because while it can give a closer fit to five or more data points, it will introduce oscillations,
particularly between the data points, that will not represent the engine performance well.
When an entire plant has I prime movers and J mission profile operating modes the total
annual fuel cost can be obtained by developing a spreadsheet to calculate the following:
$
=∑ ∑ % % % /10 (4.1)
% =
% =
% [ ]
ℎ ℎ
% =
= [ ]
$
= [ ]
=ℎ [ℎ]
The mission profile modes j can be in the columns and the various prime movers i can be in the
rows of the spreadsheet. If a prime mover, such as a dual-fuel generator, operates on more than
one fuel, it can be given a row for each fuel.
4-19
Engine Air Consumption. The amount of air utilized by a diesel engine can be described by the
air fuel ratio defined as
afr = mair/mfuel
When the composition of the marine fuel (HFO or MDO) used by a diesel engine is analyzed
chemically for complete, equilibrium combustion one can establish the amount of air needed to
completely utilize the fuel. This air-fuel ratio is called the stoichiometric air-fuel ratio
 = mair,min/mf ≈ 14.5 kg air/kg fuel
In real engine operation, additional air must be provided to enable complete combustion, to
eliminate smoking, and to cool the engine. The ratio of this air to the stoichiometric value is called
the excess air ratio,
= mair/mair,min = afr/
The excess air ratio for marine diesels is in the range 1.8 ≤ ≤ 2.2 depending upon the fuel quality
with MDO at the lower end and HFO at the upper end.
In real engines, additional air may also be required to accomplish the scavenging of the
exhaust gases from the cylinders. On small high-speed and medium-speed four-stroke diesels
without intake and exhaust valve overlap this requirement is eliminated. On low-speed two- stoke
diesels this can add significant air to the requirement. The total air supplied to the engine can be
express by a total excess air ratio,
total = mair in/mair,min and afrtotal = total 


The total excess air ratio for marine diesels is in the range 1.8 ≤ total ≤ 2.8 with HFO burning, low-
speed diesels at the upper end. Using the stoichiometric air fuel ratio, this gives 26 ≤ afr total ≤ 40
kg air/kg fuel, with four-stroke MDO engines at the lower end and two-stroke HFO engines at the
upper end. Expressing this requirement in terms of the power output of the engine as a specific
air consumption gives,
specific air consumption = sac = mair in/PB = afr • sfc
4-20
where sfc is the specific fuel consumption. This characteristic lies in the range 5.5 ≤ sac ≤ 9.0 kg
air/kW h, with high-speed diesels at the low end and large two-stroke diesels at the high end.
Engine Ratings. When working with marine engines, it is important to understand the rating
system used by the vendors. The issue of rating an engine for a particular application (project) is
a complicated issue that is ultimately the joint responsibility of the vendor and the project marine
engineer. It is complicated because (1) the power output is dependent upon atmospheric
conditions, cooling water temperatures, etc. so there may be a need to correct for special specific
conditions; (2) the vendor quoted results are based upon a standard MDO fuel and may not include
all auxiliaries and factors such as mechanically driven cooling pumps, thrust bearing loading,
intake duct losses, silencers, exhaust duct losses, etc.; and (3) fuel consumption, engine wear and
time between overhauls depends on engine loading. The marine engineer must sort all this out and
arrive at the best engine rating for each application, termed here a “Project MCR.”
In the absence of specific vendor data, Woodward [6, 7] recommended derating a diesel
engine as follows:
derate power 1% for each 10% increase in relative humidity
derate power 2 1/2% for each 10 F increase in ambient air temperature
derate power 10% for each 4”Hg reduction in barometric pressure
derate power 1% for each 4”H2O increase in engine backpressure
derate power 2% for an aftercooled/intercooled engine for each 10 F increase in
sea water temperature, and also
derate power 1% for each 10 F increase in fuel temperature [8].
Similarly, in preliminary design without specific vendor’s data increase the specific fuel rate 1/3%
for each 1% derating of the engine power.
To appreciate the effect of loading on engine wear, Woud and Stapersma [1] state that engine
cylinder wear maintenance costs of an i cylinder engine are proportional to a wear index defined
as follows:
Wear index = i B PB Cm
where PB is Brake Mean Effective Pressure and Cm is the mean piston speed 2NeLs. Thus by this
model, the cylinder maintenance costs vary linearly with rpm (N e) and throttle level (  PB).
Therefore, it might make sense in a particular application to reduce the project rated point to be
4-21
used for the engine selection to extend the engine rebuild time. This might also place the “Project
MCR” in the region of the highest specific fuel rate achieving fuel economy advantages as well.
Remember, however, that when this is done the engine cost, weight, and volume are still for the
“fully rated” engine and you will purchase more engine than absolutely needed.
The approach used to rate diesel engines varies with the type of engine and the vendor. In
general, for large engines the vendor states a Maximum Continuous Rating (MCR) that is
permitted for continuous operation. They may also state additional rating points for fuel economy
or reduced maintenance reasons. For example, MAN B&W low-speed diesels are rated at four
operating points defined as:
L1: Nominal MCR, rated rpm and rated PB or BMEP

L2: at reduced PB (e.g. ~80% rated) and rated rpm
L3: at rated PB and reduced rpm (e.g. ~93% rated)
L4: at reduced PB (~80%) and reduced rpm (~93%)
Data is given for all four rating points; the L1 point is the engine MCR. The reduced P B, reduced
rpm L4 rating is sometimes called an economy rating. Wärtsilä-Sulzer low-speed diesels use a
similar approach, but their version of the L2 and L4 points are defined at a reduced power rather
than a reduced PB. A “Project MCR” to be used in any particular project might be selected
anywhere in these regions.
On smaller engines, the rating issue becomes more complicated. Taking the smaller
Caterpillar engines as an example, Caterpillar uses the following ratings:
Continuous A: load factor 80-100%, up to 100% time at rated rpm; 5000-8000 h/yr; heavy
duty applications
Medium Duty B: load factor 40-80%, up to 80% time at rated rpm; full load 10 h out of
each 12 h; 3000 to 5000 h/yr; e.g. fishing vessels
Intermittent C: load factor 20-80%, up to 50% time at rated rpm; full load 6 h out of 12 h;
2000-4000 h/yr; e.g. ferries and harbor tugs
Patrol Craft D: load factor up to 50%, up to 16% time at rated rpm; full load 2 h out of
12 h; 1000-3000 h/yr; e.g. working planing hulls and bow thrusters
High Performance E: load factor up to 30%, up to 8% time at rated rpm; full load 1/2 h
out of 12 h; 250-1000 h/yr; e.g. pleasure planing yachts
It is important that the marine engineer, in consultation with the vendor, select the appropriate
rating for each application. For their larger 3600 series engines (1730-5420 kW), Caterpillar uses
4-22
just a Continuous Service Rating for vessels with continuous duty and a Maximum Continuous
Rating for vessels with varying loads.
Natural Gas Engines. Natural gas engines are similar to diesel engines, but they have some
unique features that will be treated here [adapted from 9]. The natural gas burning marine engines
available today use a lean burn concept to achieve both high thermal efficiency and low NOx
emissions at the same time. This characteristic is illustrated in Fig. 4.11 which shows the BMEP,
thermal efficiency, and NOx emissions as a function of the air-fuel ratio. The air-fuel ratio is
expressed here as the excess air ratio , the ratio of the air-fuel ratio to the stoichiometric air-fuel
ratio. Thus, the stoichiometric is 1 on this graph.
Figure 4.11 Thermal Efficiency and NOx Production versus Non-dimensional Excess
Air/fuel Ratio (courtesy Wärtsilä)
Engine operation is limited in two regions in Fig. 4.11. If the BMEP is too high near
stoichiometric conditions there is early uncontrolled ignition of the fuel called knocking
(detonation). This limits the practical compression ratio and power output. If the BMEP is too
high at higher  the cylinder can misfire, lose power and run unevenly. When the  is in the range
2.0 to 2.3, over twice the amount of oxygen needed for complete combustion under equilibrium
conditions, the natural gas engine can achieve high power, high thermal efficiency, and low NOx
emissions simultaneously. NOx is produced by high cylinder temperatures and long residence
time at these temperatures. By using considerable excess air (lean burn) the cylinder temperatures
are kept low and the NOx production is reduced.
4-23
There are three principal types of natural gas engines offered today: high-pressure dual fuel
Diesel cycle engines; low-pressure, spark ignited Otto cycle engines; and low-pressure dual-fuel,
micro pilot ignited engines. The high-pressure engines use a gas compressor to inject the gas at
300-350 bar after compression so the engines can operate using a Diesel cycle compression
ignition. This gas compression can require up to 5% of engine power. There are three main
approaches,
 Large low-speed engines burning a variable mixture of liquid fuel (HFO) and LNG cargo
boil-off
 Low-speed engines that use mostly gas with up to 5% liquid fuel pilot for more stable
ignition
 Medium and high speed engines using a variable mixture of liquid fuel (MDO) and up to
about 70% gas
The spark ignited engines use two Otto cycle configurations, where the fuel and air are compressed
together in the cylinder and then ignited by a spark plug:
 Engines with the gas and air premixed before the turbocharger
 Engines with the gas supply in each cylinder head
The operating cycle for the latter engines is illustrated in Fig. 4.12. These are four-stroke engines
and the exhaust stroke is not shown. Most of the gas (about 90%) is injected into the intake air
within the cylinder head. This lean mixture fills the cylinder during the intake stroke. The
remaining gas is injected into a pre-chamber within the head that contains a spark plug. This
separation is used to allow a richer, more stoichiometric, mixture in the pre-chamber for reliable,
high energy ignition when the plug fires.
Figure 4.12 Operating Cycle for Low Pressure, Lean Burn, Spark Ignited Natural Gas
Engines (courtesy Wärtsilä)
4-24
The dual-fuel, micro pilot ignited engines also use two Otto cycle configurations:
 Engines with the gas and air premixed before the turbocharger
 Engines with the gas supply in each cylinder head
The operating cycle for the latter engines is illustrated in Fig. 4.13. These are four-stroke engines
and the exhaust stroke is again not shown. The gas is injected into the intake air within the cylinder
head. This lean mixture fills the cylinder during the intake stroke. The gas is ignited by injecting
a small amount of diesel into the cylinder where it ignites from the high temperature at the end of
compression (a Diesel cycle principle). This ignition source produces a much higher energy
(~1000x) ignition source than the approach using spark plug ignition of gas within a pre-chamber.
These engines produce about 99% of the power from gas and about 1% from the diesel fuel. These
dual-fuel engines offer a valuable fuel source reliability advantage since they can also operate on
diesel fuel only using a normal Diesel cycle.
The older style engines that inject the gas before the turbocharger do not meet the
Classification Society inherently safe or double-wall piping classification and must, therefore, be
used in an Emergency Shutdown (ESD) safety design. They have not been used on vessels since
the Norwegian ferry Glutra, the first LNG powered non-LNG cargo vessel delivered in 2000 and
other early designs.
Figure 4.13 Operating Cycle for Low Pressure, Dual-fuel Micro-pilot Ignited Natural Gas
Engines (courtesy Wärtsilä)
All of the low-pressure engines considered above, use low pressure (4-5 bar) natural gas
injection with the fuel pressurized by the pressure created within the LNG storage tank due to its
4-25
natural or controlled warming. All LNG engines used on non-LNG cargo vessels to date (2014)
have been of this low-pressure type. The spark-ignited engines provide the advantages of slightly
better fuel rate and NOx emissions, a simpler single fuel storage/supply system, and improved
dynamic load response. Acceleration load response on these engines can be made similar to
conventional diesel engines by allowing the mixture in the cylinder to become temporarily richer
during the acceleration by the use of a variable geometry turbocharger made possible by the clean
burning fuel. The micro pilot dual-fuel engines provide the important advantage of providing a
second, alternative fuel capability which will be very attractive to some owners, particularly during
the early years of LNG fuel infrastructure development.
Because the heat content of natural gas can vary by source, the fuel use of these engines is
expressed as a specific energy consumption or heat rate in kJ/kWh. The Rolls-Royce B series
engines have a heat rate of about 7,475 kJ/kWh at MCR so that with a natural gas with a Lower
Heating Value of LHVLNG = hL = 45, 300 kJ/kg from Table 3.4, this represents a specific fuel
consumption of 165 g/kWh, comparable to the most efficient low-speed diesel engines and better
than similar sized medium-speed diesels. If equation 4.1 is being used to evaluate the annual fuel
cost of a plant involving LNG engines, the sfci(%Pij) can be replaced by,
sfci(%Pij) [g/kWh] = heat ratei(%Pij) [kJ/kWh]*1000 [g/kg]/LHVLNG[kJ/kg].
Air Emissions Requirements. Marine reciprocating engines produce emissions that are harmful
to the atmosphere and the environment. The International Maritime Organization (IMO),
Environmental Protection Agency (EPA), etc. requirements are evolving. The harmful emissions
are CO, CO2, NOx (NO, NO2, and N2O) that are produced by the high temperatures within the
cylinders, SOx that are a direct result of the sulfur content in the fuel, soot and other particulate
matter (PM), methane and unburned hydrocarbons (HC) including volatile organic compounds
(VOC). Figure 3.2 showed the evolving marine fuel sulfur limits with a 10 fold decrease in the
ECA sulfur limit coming in January 2015. Operators can still burn higher sulfur content fuels if
an Exhaust Gas Cleaning System that will reduce the exhaust SOx to equivalent levels is installed.
The current international NOx standard for marine engines is given by IMO in Annex VI to
the MARPOL 73/78 convention as shown in Fig. 4.14. NOx is produced in the high temperatures
within the cylinder and production depends upon the engine technology so NOx control involves
4-26
engine design and exhaust after treatment. These requirements involve staged reductions in the
allowable NOx emissions. They only apply to new engines or conversions, when installed. The
existing Tier II requirements for the large Category 3 engines are met with on-engine
modifications. It is generally considered that the Tier III requirements, which come into effect in
2016, will require the use of Selective Catalytic Reduction (SCR) systems or some other after
treatment on marine diesels.
Figure 4.14 MARPOL Annex VI (EPA) ECA NOx limits
Emissions Factors – Specific Pollutant Emissions. The specific emission of each category is
usually expressed as an Emission Factor (EF) as g pollutant/kWh. Note that Woud and Stapersma
[1] use the term specific pollutant emission (spe) for this quantity and also use a pollutant emission
ratio [g pollutant/kg fuel burned] so that spe = per • sfc, when they use their definition of sfc [kg
fuel/kWh]. Typical values for the principal specific emissions for various types of propulsion are
shown in Table 4.2 [10]. The EPA Category 3, Tier III rates shown are the lower NOx requirements
for 500-900 rpm medium speed diesels that come into effect in January 1, 2016, and so presume
the use of after treatment. Note that the LNG engine eliminates SOx and PM and easily meets the
Tier III NOx standard without any after treatment.
Table 4.2 Typical Specific Emissions from Plants at Rated Power [10]
4-27
The Emissions Factors (spe) [g pollutant/kWh] for engines can often be obtained from the
vendor at a few points for different per cent power loading or general values can be assumed from
similar engines. For calculations, it is helpful to express these as quadratic or cubic regression
equations as a function of per cent power as done for the specific fuel consumption sfc [g/kWh].
For diesel engines, the SOx, PM, fine particulates PM2.5 and CO2 depend only on the fuel
consumption (sfc) and the sulfur content of the fuel (%S as a decimal), so general formulas can
be used if not otherwise available [11]:
EFSO2 = sfc*2*0.97753*%S
EFPM = 0.25 + sfc*7*0.02247*(%S – 0.001)
EFPM2.5 = 0.97EFPM
EFCO2 = sfc*0.87*(44/12)
Annual Air Emissions. When an entire plant has I prime movers and J mission profile operating
modes the total annual emission [t/yr] for a particular pollutant can be obtained by developing a
spreadsheet, similar to that needed for annual fuel cost, to calculate the following:
=∑ ∑ % % % /10 (4.2)
% =
% =
% [ ]
ℎ ℎ
% =
= [ ]
=ℎ [ℎ]
4-28
The mission profile modes j can be in the columns and the various prime movers i can be in the
rows of the spreadsheet. If a prime mover, such as a dual-fuel generator, operates on more than
one fuel, it can be given a row for each fuel.
Exhaust Gas Cleaning Systems (Scrubbers). For fuel cost reasons, many diesel vessels that
operate outside the ECAs most of their voyage will continue to burn HFO. When they enter an
ECA they will have to meet the tighter exhaust standards. The options are to switch to MDO/MGO
within 200 nm or install an Exhaust Gas Cleaning System (EGCS) or scrubber to continue to burn
the higher sulfur content HFO. Of course, other vessels may operate most of their time within an
ECA and still choose to burn HFO, in which case they will also need to install an EGCS. The
recent study “Exhaust Gas Cleaning Systems Selection Guide” completed by The Glosten
Associates for the Ship Operations Cooperative Program (SOCP) provides a good summary of the
evolving technologies and options [12]. This section is adapted from Parsons et al. [13].
Currently four EGCS options are available [12], three wet systems and a dry system, as
follows:
Wet Systems
open-loop systems – circulate sea water through the scrubber to remove SOx and
discharge the effluent overboard.
closed-loop systems – circulate fresh water containing a metered amount of caustic
soda (NaOH) through the scrubber to remove SOx; a small bleed-off fraction
is removed for water treatment before discharge overboard.
hybrid system – a combination of the above two using both sea water and fresh water
with NaOH.
Dry System – passes the exhaust gas over a dry bed containing Calcium Hydroxide
pellets to remove SOx; can also inject urea/ammonia to remove NOx.
A simplified concept schematic for the wet closed-loop scrubber offered by Wärtsilä is shown in
Fig. 4.15. In the closed loop, fresh water is used as the scrubbing water as the exhaust gases pass
through a droplet separator before entering the stack. The scrubbing water is circulated in two
circuits. In the first circuit (not shown) scrubbing water is circulated from the scrubber wet sump
through pumps back to spray nozzles below the packing bed inside the scrubber unit.
In the second circuit, scrubbing water is circulated from the wet sump through pumps to a
cooling heat exchanger to reduce the scrubbing water temperature. From the heat exchanger, the
scrubbing water is routed to spray nozzles near the top of scrubber. Water is also supplied to the
4-29
mid part of the scrubber to further improve the SOx removal efficiency. Scrubbing water passes
through the packing bed and is collected and removed at the bottom.
A small quantity (~1%) of the scrubber water is led to the bleed-off water treatment unit
where flocculation and coagulation chemicals are added to enhance the flotation of solids in the
tank. The treatment effluent is monitored for pH, polycyclic aromatic hydrocarbons (PAH) and
turbidity.
Figure 4.15 Schematic for Closed-loop Exhaust Gas Scrubber System (from Wärtsilä)
NaOH (50% caustic soda solution) is added to the circulating scrubbing water to maintain
the process pH and consequently the SOx removal efficiency. NaOH is fed to the process as
required by process chemistry. The main input data for NaOH feed control are the sulfur content
of the fuel and engine load. The main control is automatically adjusted based on the scrubbing
water pH to compensate for variations and inaccuracy in the fuel sulfur data. Sea/lake water in an
external loop is used for cooling of the scrubbing water. An existing cooling water system can be
used or an additional pump can be provided, if needed. Handling of NaOH requires impervious
clothing, chemical boots, safety helmet, safety goggles or face shield, emergency wash down
stations, etc.
4-30
The scrubber unit for a 5400 kW unit would be about 26 tall, 12.5 in diameter and weigh
about 10 LT. The scrubber would also function as an exhaust silencer so a separate silencer would
not have to be included.
Since the wet systems cool and moisten the exhaust gases, there is some concern that they
may not be compatible with series Selective Catalytic Reduction needed to meet NOx requirements
on vessels built after January 1, 2016. The use of both may also require an exhaust fan to prevent
excessive engine back pressure [12].
Selective Catalytic Reduction (SCR). When the tighter ECA NOx requirements come into effect
in January 2016, large EPA Category 3 marine diesel engines will need to install exhaust gas after
treatment to meet the NOx standard. The common solution is Selective Catalytic Reduction using
liquid urea as a source of ammonia. These systems are already in use in vessels that visit
California, which has had tighter NOx standards within the 20 mile limit for years. This is another
large component in the engine exhaust path. When aqueous urea is injected into the exhaust, the
aqueous urea vaporizes and decomposes to form ammonia and carbon dioxide. Within the SCR
catalyst, the NOx are catalytically reduced by the ammonia (NH3) into water (H2O) and nitrogen
(N2), which are both harmless; and these are then released through the exhaust.
The smaller marine diesel engines are using similar systems that are typically installed
directly on the engine and use Diesel Exhaust Fluid (DEF) containing aqueous urea as the source
of ammonia. The lone exception to date is the GE 12-cylinder V250 (3,500 kW) EPA Category 2
diesel that has met its EPA Tier 4 qualification using Exhaust Gas Recirculation (EGR) as opposed
to SCR. The NOx is produced by the high temperatures within the cylinder; EGR reduces this
temperature allowing the NOx requirement to be met with some loss of combustion efficiency. GE
compensates for this loss by using a high-pressure common rail fuel injection system and two-
stage turbocharging.
An example SCR installation is shown in Fig. 4.16. The aqueous urea is sprayed into the
exhaust line by compressed air from a dosing unit supplied from a urea tank in the engine room.
When the exhaust passes through the catalyst unit the NOx are converted to water and nitrogen. A
NOx analyzer monitors the level of NOx remaining after the unit. This particular catalyst unit
requires a volume 2.0 m3/MW and weighs 0.76 t/MW. The minimum temperature at the inlet to
the SCR is a function of the sulfur content of the exhaust coming from the fuel. It increases with
4-31
greater sulfur content ranging from 260 to 350˚C. This temperature is needed to prevent the
formation of ammonia sulphate and the loss of the ammonia for the reaction with the NOx. This
leads to the concern about combining a SCR in series with an EGCS wet scrubber, which cools
the exhaust gas through its water injection.
Figure 4.16 Typical Selective Catalytic Reduction Installation (adapted from MAN)
4-32
Thermodynamics. The working fluid in internal combustion engines is a complex mixture of air,
fuel, and combustion products. Considerable insight and preliminary design information can be
obtained, however, by analyzing the working fluid using the Air Standard Ideal Gas model. This
treats the working fluid as ideal gas air with constant specific heats C V0 and CP0.
Air Standard Otto Cycle. Recall that the Otto cycle is the idealized model for the spark
ignition (gasoline or natural gas) engine as shown in Fig. 4.5. The thermodynamic efficiency of
this cycle will be given by,
th = (QH – QL)/QH = 1 – QL/QH
where the combustion process (2-3) heat addition QH and the exhaust-intake processes (4-1) heat
rejection QL are both constant volume processes with V1 = V4 and V2 = V3.
Applying the First Law to these ideal gas constant volume processes yields,
inqout = uout – uin = CV0(Tout – Tin) or specifically here
2 q3 = CV0 (T3 – T2) = QH/m
4 q1 = CV0(T1 – T4) = – QL/m

or,
th = 1 – mCV0(T4 – T1)/mCV0(T3 – T2) = 1 – T1(T4/T1 – 1)/T2(T3/T2 – 1)
Now the power process (3-4) and the compression process (1-2) are isentropic processes. For
an ideal gas, the absolute temperatures at the end of these isentropic processes are related by,
(T2/T1) = (V1/V2)k–1 and (V4/V3)k–1 = (T3/T4)
so with V1 = V4 and V2 = V3 (T3/T2) = (T4/T1)
and th = 1 – T1/T2 = 1 – r1–k
where r is called the compression ratio V1/V2 = V4/V3. Thus, the thermodynamic efficiency of a
gasoline engine is an increasing function of the compression ratio since k > 1. The practical
efficiency of the gasoline engine is limited, however, by the tendency of the fuel for pre-ignition,
called knock, at high compression ratios. Because the fuel is not present during the compression
process, diesels are able to utilize higher compression ratios than can gasoline engines.
4-33
Air Standard Diesel Cycle. Recall that the Diesel cycle is the idealized model for the
compression ignition (diesel) engine as shown in Fig. 4.4. The thermodynamic efficiency of this
cycle will also be given by,
th = (QH – QL)/QH = 1 – QL/QH
where the combustion process (2-3) heat addition QH is now a constant pressure process with P2 =
P3 and the exhaust-intake processes (4-1) heat rejection QL remains a constant volume process
with V1 = V4.
Applying the First Law to the ideal gas constant pressure heat addition yields,
2 q3 = u3 – u2 + 2w3 = u3 – u2 + ∫ Pdv = u3 – u2 + P3v3 – P2v2 = h3 – h2
2 q3 = CP0 (T3 – T2) = QH/m
And applying the First Law to the ideal gas constant volume heat rejection yields,
4 q1 = u1 – u4 = CV0(T1 – T4) = – QL/m
giving for the thermodynamic efficiency,
th = 1 – mCV0(T4 – T1)/mCP0(T3 – T2) = 1 – (T4 – T1)/k(T3 – T2)
where recall that k is the ratio of the ideal gas specific heats k = CP0/CV0. The only difference from
the Otto cycle result is the k in the denominator.
To move further with this development, recall that the compression ratio r = V1/V2 and now
define a fuel cutoff ratio rL = V3/V2, the ratio of the volume at the end of the fuel combustion to
the volume at the beginning of fuel injection. The power process (3-4) and the compression
process (1-2) are again isentropic processes. For an ideal gas, the absolute temperatures at the
ends of these isentropic processes can be related by,
(T2/T1) = (V1/V2)k–1 and (T3/T4) = (V4/V3)k–1
Using the first yields, T2 = T1r k–1
4-34
The heat addition process is a constant pressure process and using the Ideal Gas equation of state
gives,
PV/T = constant so that with P2 = P3 V3/T3 = V2/T2
which can be solved to give T3 = T2(V3/V2) = T1r k–1rL
Now returning to the second isentropic result and using V1 = V4, it can be solved for T4
T4 = T3(V3/V4)k–1 = T1r k–1rL (rL/r) k–1 = T1rLk
Substituting these three temperatures into the thermodynamic efficiency expression yields,
th = 1 – (T1rLk – T1)/k(T1r k–1rL – T1r k–1) = 1 – r 1–k[(rLk – 1)/k(rL – 1)]
The term in the square brackets is always greater than 1 and in the limit as rL goes to one it becomes
1. Therefore, this result appropriately reduces to the Otto cycle result when rL = 1.
Air Standard Seiliger Cycle. Recall that we noted that real diesel engines reach their peak
pressure somewhere between the beginning of combustion and the end of combustion. The ideal
cycle that best applies to this case is the Seiliger or the dual cycle as shown in Fig. 4.6. The
thermodynamic efficiency of this cycle will be given by,
th = (QH1 + QH2 – QL)/(QH1 + QH2) = 1 – QL/(QH1 + QH2)
where the combustion process (2-3) heat addition QH1 is a constant volume process with V2 = V3,
combustion process (3-4) heat addition QH2 is a constant pressure process with P3 = P4, and the
exhaust-intake processes (5-1) heat rejection QL is a constant volume process with V1 = V5.
The Seiliger cycle can be analyzed using the same methods as used above [1, 4] to give the
following result for the thermodynamic efficiency,
th = 1 – r 1–k{(rPrLk – 1)/[(rP – 1) + krP(rL – 1)]}
where r is the compression ratio as before, r = V1/V2

rL is the fuel cutoff ratio similar to before, rL = V4/V3, and
rP is the pressure rise ratio in the first heat addition, rP = P3/P2
4-35
Note that when rP = 1, this thermodynamic efficiency becomes the Diesel cycle result and when rL
= 1 this becomes the Otto cycle result as it should. Note that Woud and Stapersma [1] denote the
pressure rise ratio rP as parameter a and the fuel cutoff ratio rL as b.
References
2. www.bunkerworld.com
3. Stone, R., Introduction to Internal Combustion Engines, 2nd Edition, SAE, Warrendale, PA,
1985.
4. Johnston, R. M., Brockett, W. A., and Bock, A. E., Elements of Applied Thermodynamics,
U. S. Naval Institute, 1951.
5. Harrington, R. L. (ed.), Marine Engineering, SNAME, New York, 1992.
6. Woodward, J. B., Low Speed Marine Diesel, Wiley-Interscience Publication, John Wiley and
Sons, Inc., New York, 1981.
7. “Marine Diesel Power Plant Performance Practices,” SNAME Technical and Research
Bulletin 3-27, 1975.
8. Brady, R. N., Diesel Fuel Systems, Reston Publishing Co., Reston, VA, 1981
9. Parsons, M. G., “Engines for Natural Gas Fueled Vessels: A Look at the Technologies used
in Available Gas Engines,” Great Lakes/Seaway Review, 40-4: 59 & 61, April-June, 2012.
10. Parsons, M. G., O’Hern, P. J. and Denomy, S. J., “The Potential Conversion of the U.S.
Great Lakes Steam Bulk Carriers to LNG Propulsion – Initial Report,” Journal of Ship
Production and Design, 28-3: 97-111, August, 2012.
11. Harkins, R. W., “Great Lakes Marine Air Emissions – We’re Different up Here!”, Marine
Technology, 44-3: 151-174, July, 2007.
12. Reynolds, K. J., “Exhaust Gas Cleaning Systems: Selection Guide”, Glosten File No.
10047.01, 22 Feb. 2011, Rev. A, prepared for the Ship Operations Cooperative Program
(SOCP).
13. Parsons, M. G., O’Hern, P. J., Harkins, R.W. and Denomy, S. J., “The Potential Conversion
of the U.S. Great Lakes Steam Bulk Carriers to LNG Propulsion – Final Report”, Journal
of Ship Production and Design, 29-4: 162-182, November 2013.
4-36
Chapter 5 Gas Turbine Prime Movers
The dominant prime mover on non-nuclear naval vessels in the United States and much of
the world is the gas turbine. They are in increasing use on cruise ships were they offer vibration
and arrangements advantages. They are also being use in new, larger LNG carriers where they
offer arrangements advantages and can burn LNG boiloff. Gas turbines also compete with high-
speed diesels for application on high-speed ferries (the Incat gas turbine dual fuel LNG burning
catamaran ferry Francisco recently made 58.1 knots during light sea trials) and other high-speed
craft. Gas turbines compete primarily due to their modular construction, low emissions, and high
power density (low weight and volume per kW).
Gas turbines have the following advantages relative to steam plants:
• fast startup – minutes instead of many hours;

• modular construction – repair is typically by replacement;
• reduced manning and easy automation;
• reduce air emissions; and
• high reliability and maintainability.
Gas turbines have an advantage relative to diesels based upon their higher power density and
reduced air emissions, but have the following important disadvantages:
• lower efficiency;
• higher quality and cost fuel if not burning natural gas; and
• difficulty to repair in place.
Highly volume limited and high-speed applications can take best advantage of the high power
density of gas turbines.
Classification. Gas turbines can be classified in a number of ways. First, they are described as
aircraft or aero-derivative gas turbines or industrial (heavy-duty) gas turbines. The aero-
derivative engines are “marinized” versions of aircraft engines and are, thus, characterized by
very light-weight and high power density. The aircraft jet engine is adapted for sea level
operation and used as the front end or gas generator. The jet nozzle is replaced by additional
power turbine stages to convert the remaining hot gas energy to mechanical work. Turbo-prop
and helicopter engines are a more direct application. There are also some gas turbines that are
marine adaptations of engines that were initially developed for land-based or offshore platform
5-1
power generation and pipeline pumping applications. These engines are somewhat heavier and
take up slightly more volume. A section drawing of a typical aero-derivative gas turbine, the
25,000 kW General Electric LM2500 that is used on most U.S. Navy applications, is shown in
Fig. 5.1. A section drawing of a light industrial gas turbine, the 17,000 kW ABB-Stal GT35 used
on offshore platforms and some high-speed ferry applications, is shown in Fig. 5.2.
Figure 5.1 Cutaway View of Aero-derivative LM2500 Gas Turbine
Figure 5.2 Cutaway View of Light Industrial ABB-Stal GT35 Gas Turbine
Gas turbines can also be classified by the type of thermodynamic cycle and the number of
separate rotating shafts. The gas turbine, in general, must consist of a large air intake; a
5-2
compressor that might use 8-20 stages; a combustor where the fuel is injected into the air flow
and burned; a turbine that might use 5-15 stages; and a finally a large exhaust path. Gas turbines
with just these five basic components are said to have a simple cycle. Additional components
such as recuperators (or regenerators), intercoolers, and additional reheat combustors are also
used to improve efficiency and such gas turbines are said to have complex or advanced cycles.
The final stages of the turbine(s) where the output mechanical work is generated can be
mechanically connected to the remainder of the turbine and compressor or it can be a completely
separate power turbine connected only by the gas flow. When they are mechanically connected,
the gas turbine is called a single-shaft machine. When there is a gas generator followed by a
mechanically separate power turbine, the gas turbine is called a two-shaft or dual shaft machine.
The difference is very important for the characteristics and area of application of the gas turbine.
In some gas turbines, multiple concentric shafts can be used to link portions of the turbine to
portions of the compressor. These machines might still be described as “single-shaft” if the
output part of the turbine is mechanically connected to any part of the gas generator. As we will
see below when we look at the operating characteristics of gas turbines, the single-shaft
machines have a very narrow rpm range of operation and they are primarily suited for electrical
generation applications where the operation will be at a single rated rpm. They can also be used
successfully in some race boat applications where the mechanical connection can prevent
turbine/propeller overspeed in a seaway. The two-shaft machines, with the separate power
turbine, generally behave like a steam boiler (the gas generator) and a separate steam turbine (the
free power turbine) and have ideal characteristics for ship mechanical propulsion. Schematics
of single-shaft and two-shaft simple-cycle gas turbines are shown in Fig. 5.3 [1].
power turbine
a. single shaft b. two-shaft

Figure 5.3 Schematics of Single-Shaft and Two-Shaft Simple-Cycle Gas Turbines [1]
5-3
Typical characteristics of current aero-derivative marine gas turbines are presented in Table
5.1 [adapted from 1]. The similar data for medium-speed marine diesels is repeated from Table
4.1 for comparison. The industrial gas turbine example shown in Table 5.1 is the ABB-Stal GT35
used on recent high-speed ferries. These gas turbines utilize larger combustors giving more
combustion time permitting the use of the 50% cheaper IF30 fuel. The advanced cycle example
shown in Table 5.1 is the principal example available today, the Rolls-Royce WR21 ICR
(intercooled recuperative or regenerative) gas turbine being developed for the U.S. Navy.
Table 5.1 Typical Performance Parameters of Medium-Speed Diesel Engines

and Marine Gas Turbines (2013) [adapted from 1]
Data Medium-speed Aero-derivative Industrial Rolls-Royce

Diesels Gas Turbines ABB-Stal ICR
GT35 WR21
Process/cycle 4-stroke simple cycle simple cycle advanced cycle

Construction trunk piston two-shaft two-shaft two-shaft
Output power range [kW] 500-23280 6000-41000 17000 24000
Output speed [rpm] 300-1000 3300-7000 3300 3600
Fuel type HFO or MDO MGO or JP5 MDO or IF30 MGO
Specific fuel rate [g/kW h]* 170-210 210-280 260 200
Specific air rate [kg/kW h] 6-9 10-15 10.5
Specific NOx
emission [g/kW h] 2-18 2-5 2 3
Specific mass [kg/kW] 5-20 0.6-1.4 1.5-2.0 1.8
Specific volume [dm3/kW] 4-28 2.5-4.5 6.0 4.1
Specific cost [$/kW] L: 240-360 200-310 515
* ISO standard on MDO V: 190-310
MGO = marine gas oil, a light diesel oil ~NATO F-76
Analysis of Gas Turbine Cycles. The cycles of gas turbines can be analyzed to establish their
thermodynamic efficiency, air flow, etc. and to understand the effect of engine and ship system
design parameters on their operation. These analyses can be made using the Air Standard Ideal
Gas – Constant Specific Heats model.
5-4
Air Standard Brayton Cycle. The idealized model for the simple cycle gas turbine is the
Brayton or the Joule cycle. This ideal model is composed of the following four processes:
1-2 isentropic compression in the compressor

2-3 constant pressure heat addition in the combustor
3-4 isentropic expansion in the turbine
4-1 constant pressure heat rejection in the exhaust and intake of the cycle.
This can be seen to model the five components of the simple cycle gas turbine. The associated
P-v and T-s diagrams are shown in Fig. 5.4.
Figure 5.4 P-v Diagram and T-s Diagram for the Ideal Brayton Gas Turbine Cycle
The thermodynamic efficiency of this gas turbine cycle will be given by the generally valid
expression,
th = wnet/QH = [wturbine – wcompressor]/heat added = [(h3 – h4) – (h2 – h1)]/(h3 – h2)
Specializing now to the Ideal Gas, Constant Specific Heats model, this can also be written as,
th = 1 – QL/QH = 1 – mCP0(T4 – T1)/mCP0(T3 – T2) = 1 – T1(T4/T1 – 1)/T2(T3/T2 – 1)
Since the ideal turbine and ideal compressor are isentropic and the two heat exchange processes
are at constant pressure, we also have when using an absolute temperature scale,
constant pressures
(T2s/T1)k/(k – 1) = (P2/P1) = (P3/P4) = (T3/T4s)k/(k – 1) so (T4/T1) = (T3/T2)

isentropic isentropic
If we define an isentropic pressure ratio as rps = P2/P1, the thermodynamic efficiency is then,
5-5
(1 – k)/k
th = 1 – T1/T2 = 1 – rps where (1 – k)/k = – 0.2857
Thus, the ideal Brayton cycle thermodynamic efficiency depends only upon the isentropic
pressure ratio – the pressure change across the compressor. Note that the first part of this result
in terms of T1/T2 is the same as obtained for the Otto cycle.
In terms of the system enthalpies for any gas turbine cycle, the specific air consumption sac
[kg air/kWh], heat rate [kJ/kWh], and the specific fuel consumption sfc or specific fuel rate sfr
[g fuel/kWh] are as follows with work in [kJ/kg = kWs/kg]:
sac = (3600 s/h)/wnet = 3600/[(h3 – h4) – (h2 – h1)]

heat rate = (3600 s/h)/th
sfr = sfc = (3600 s/h •1000 g/kg)/thhL = 3,600,000(h3 – h2)/{[( h3 – h4) – (h2 – h1)]hL}
where recall that hL is the Lower Heating Value (LHV) of the fuel. In the second formula, the
combustor is assumed to be 100% efficient in converting hL to heat within the process gas. In
English units, the sac and sfr are in lbm/hpBRh and the heat rate is in Btu/hpBrh so there is an
additional units resolution constant 2544.43 Btu/hpBRh in the numerator of each expression. The
h can be expressed as changes in absolute T using the constant specific heats, h = CP0T.
The Effect of Component Efficiencies. We can now add real component efficiencies to
the compressor and turbine. The resulting T-s diagram is now as shown in Fig. 5.5. Recall that
these component isentropic efficiencies are defined as:
turbine = (h3 – h4)/(h3 – h4s) and compressor = (h2s – h1)/(h2 – h1)
or using the Ideal Gas – Constant Specific Heats model results these become,
turbine = (T3 – T4)/(T3 – T4s) and compressor = (T2s – T1)/(T2 – T1)
Now consider the results for a series of gas turbines with rps = (P2/P1) = 18 with various
component efficiencies turbine = compressor and a peak cycle temperature T3 = 1200 K. This
analysis, which neglects the pressure drop in the combustor as small (P 2 = P3), is shown in Table
5.2. Notice how important the component efficiency are to the cycle feasibility and efficiency;
5-6
the 70% case will not even work because the compressor requires more work than produced by
the turbine leaving nothing for output. Even the 75% case has a cycle efficiency of only 1%.
The development of suitably efficient compressors delayed the development of the jet engine in
World War II.
actual
isentropic, ideal
Figure 5.5 T-s Diagram for Simple Gas Turbine Cycle with Real Components
Table 5.2 Analysis of the Effect of Component Efficiencies on Cycle Efficiency
ideal
ηturb = ηcomp. 100% 95% 90% 85% 80% 75% 70% equation used
T1 [K] 293.15 293.15 293.15 293.15 293.15 293.15 293.15 specified
T2s [K] 669.45 669.45 669.45 669.45 669.45 669.45 669.45 T1(P2/P1=18)(k - 1)/k
T2 [K] 669.45 689.26 711.27 735.86 763.53 794.89 830.73 T1 + (T2s - T1)/ηcomp.
T3 [K] 1200.00 1200.00 1200.00 1200.00 1200.00 1200.00 1200.00 specified
T4s [K] 525.47 525.47 525.47 525.47 525.47 525.47 525.47 T3/(P3/P4=18)(k - 1)/k
T4 [K] 525.47 559.20 592.93 626.65 660.38 694.10 727.83 T3 - (T3 - T4s)ηturb.
qH [kJ/kg] 532.40 512.53 490.44 465.76 438.00 406.53 370.56 Cp0(T3 - T2)
wturb. [kJ/kg] 676.89 643.04 609.20 575.36 541.51 507.67 473.82 Cp0(T3 - T4)
wcompr. [kJ/kg] 377.62 397.50 419.58 444.26 472.03 503.50 539.46 Cp0(T2 - T1)
wnet [kJ/kg] 299.27 245.55 189.62 131.09 69.48 4.17 -65.64 wturb. - wcomp.
ηth 0.562 0.479 0.387 0.281 0.159 0.010 wnet/qH
sac [kg/kWh] 12.0 14.7 19.0 27.5 51.8 863.2 3600/wnet
sfc [g/kWh] 150 176 218 300 531 8218 3600000/(ηthhL)
hL = 42,700 kJ/kg for MDO k = 1.4 Cp0=1.0035 kJ/kgK
When real components (efficiencies less than one) are used, the overall cycle efficiency
exhibits an optimum isentropic pressure ratio rps rather than continuously increasing with
pressure ratio as found for the ideal Brayton cycle. This result can be seen by considering the
5-7
analysis shown in Table 5.3. In this analysis, cycles with various pressure ratios with peak
temperature T3 = 1200 K and component efficiencies compressor = turbine = 0.90 are analyzed. For
these parameters, the optimum pressure ratio rps* = 19.18 with the results essentially flat from 18
< rps < 20. Economical, practical design for these conditions would then use r ps = 18.
Table 5.3 Analysis for Optimum Pressure Ratio with T3 = 1200K

and 90% Component Efficiencies
Pressure rps
Ratio 12.0 14.0 16.0 18.0 20.0 22.0 24.0 equation used
T1 [K] 293.15 293.15 293.15 293.15 293.15 293.15 293.15 specified
T2s [K] 596.23 623.07 647.30 669.45 689.91 708.96 726.80 T1rps(k - 1)/k
T2 [K] 629.90 659.73 686.65 711.27 734.00 755.16 774.99 T1 + (T2s - T1)/comp.
T3 [K] 1200.00 1200.00 1200.00 1200.00 1200.00 1200.00 1200.00 specified
T4s [K] 590.01 564.59 543.46 525.47 509.89 496.19 484.01 T3/rps(k - 1)/k
T4 [K] 651.01 628.13 609.11 592.93 578.90 566.57 555.61 T3 - (T3 - T4s)turb.
qH [kJ/kg] 572.09 542.16 515.14 490.44 467.63 446.40 426.50 Cp0(T3 - T2)
wturb. [kJ/kg] 550.91 573.87 592.96 609.20 623.27 635.64 646.65 Cp0(T3 - T4)
wcompr. [kJ/kg] 337.93 367.86 394.88 419.58 442.39 463.63 483.52 Cp0(T2 - T1)
wnet [kJ/kg] 212.98 206.01 198.08 189.62 180.88 172.02 163.12 wturb. - wcomp.
th 0.372 0.380 0.385 0.387 0.387 0.385 0.382 wnet/qH
sac [kg/kWh] 16.9 17.5 18.2 19.0 19.9 20.9 22.1 3600/wnet
L
sfc [g/kWh] 226 222 219 218 218 219 220 3600000/(thh )
The Effect of Peak Temperature T3. With real ( < 1) components, the specific air
consumption, cycle efficiency and the specific fuel rate are also strongly dependent upon the
value of the peak temperature and not just on the isentropic pressure ratio r ps as found for the
Brayton cycle. Table 5.4 shows the result of the analyses of cycles with r ps = 18 and both the
compressor and turbine component efficiencies equal to 85%. The cycle efficiency can be seen
to be strongly dependent on the peak temperature. Peak temperatures are currently limited to
about 1500 K. Further, increases will be made possible by the use of improved turbine blade
cooling, improved blade mechanical design, ceramic blades or coatings, etc. Note also that the
air rate reduces rapidly with increases in the peak temperature.
The Effect of Pressure Losses. Both the efficiency and power output of gas turbines are
sensitive to intake and exhaust pressure losses. The intake and exhaust air flows are large so it is
necessary to use large ducts in order to keep the pressure drop low. The U.S. Navy standard
5-8
rating is for a 4” H2O pressure drop in the inlet duct and 6” H 2O pressure drop in the exhaust
duct. Note that,
1” H2O = 0.0361 lbf/in2 = 0.249 kPa and 1 atm. = 407.09” H2O
These are very small pressure losses. The intake and exhaust ducting through the ship must,
therefore, have a large cross-sectional area and be developed carefully to ensure that the pressure
drops are acceptable. Figure 5.6 shows the effect of the intake and exhaust pressure drop on the
heat rate [kJ/kWh], fuel rate, and power of a simple cycle gas turbine. Most vendors typically
quote gas turbine characteristics for zero intake and exhaust losses so the marine engineer must
correct the power output and fuel rate for the actual conditions expected in the application.
Table 5.4 Analysis of the Effect of Peak Temperature with rps = 18

and 85% Component Efficiencies
Peak T3 [K]
Temperature 1200 1300 1400 1500 1600 1700 1800 equation used
T1 [K] 293.15 293.15 293.15 293.15 293.15 293.15 293.15 specified
T2s [K] 669.45 669.45 669.45 669.45 669.45 669.45 669.45 T1(P2/P1=18)(k - 1)/k
T2 [K] 735.86 735.86 735.86 735.86 735.86 735.86 735.86 T1 + (T2s - T1)/comp.
T4s [K] 525.47 569.26 613.05 656.84 700.63 744.42 788.21 T3/(P3/P4=18)(k - 1)/k
T4 [K] 626.65 678.87 731.09 783.31 835.54 887.76 939.98 T3 - (T3 - T4s)turb.
qH [kJ/kg] 465.76 566.11 666.46 766.81 867.16 967.51 1067.86 Cp0(T3 - T2)
wturb. [kJ/kg] 575.36 623.30 671.25 719.19 767.14 815.09 863.03 Cp0(T3 - T4)
wcompr. [kJ/kg] 444.26 444.26 444.26 444.26 444.26 444.26 444.26 Cp0(T2 - T1)
th 0.281 0.316 0.341 0.359 0.372 0.383 0.392 wnet/qH
sac [kg/kWh] 27.5 20.1 15.9 13.1 11.1 9.7 8.6 3600/wnet
L
sfc [g/kWh] 300 267 248 235 226 220 215 3600000/(thh )
The Effect of Ambient Temperature. The specific fuel rate (cycle efficiency), power
output and air flow of a gas turbine are usually rated at 59 F (15˚C) 60% relative humidity. The
marine engineer must correct the rating data for the design operating conditions expected. Gas
turbines develop much greater power in cold conditions, but may reach torque limits requiring
reduced throttle so the increased power may not be fully usable. With colder, denser air there is
more kg air per unit volume and the fuel rate improves. Figure 5.7 shows the typical effect of
ambient temperature on the fuel rate and power of a gas turbine. The effect of torque limits and
the actual realizable power is also shown.
5-9
Figure 5.6 Effect of Intake and Exhaust Losses on Gas Turbine Fuel Rate and Power [2]
Figure 5.7 Effect of Ambient Temperature on Gas Turbine Fuel Rate and Power [2]
5-10
The Regenerative Gas Turbine Cycle. The advanced gas turbine cycles utilize waste
heat recovery using regenerators (also called recuperators) and possibly also intercooling and
reheat combustors to improve the cycle efficiency. We will look first at the regenerative cycle
that uses exhaust gas waste heat to preheat the compressed air in a counter-flow heat exchanger
before it is heated further in the combustor. Since the turbine exhaust is used to heat the air flow,
less fuel is needed to reach the peak temperature of the cycle. A schematic and T-s diagram for
an ideal regenerative cycle is shown in Figure 5.8.
Figure 5.8 Schematic and T-s Diagram for Ideal Regenerative Cycle
The ideal regenerator could heat the compressor output T2 all the way to the turbine outlet
temperature T4, and since the same flow is on both sides of the heat exchanger, this would also
cool the exhaust flow from T4 down to T2. A finite temperature difference is needed to drive the
heat transfer, however, so the incoming air can only heated to some T5 < T4 and likewise the
exhaust is only cooled to some T6 > T2. A regenerator effectiveness R is defined as,
R = (T5 – T2)/(T4 – T2) < 1, the actual T related to the ideal T
Because the heat removed from the exhaust will equal the heat added to the compressor outlet
(for an insulated, adiabatic regenerator) and the flows are the same, (T5 – T2) = (T4 – T6).
The ideal regenerator is shown in Fig. 5.8 and this will be considered first here. The
thermodynamic efficiency of any regenerative cycle will be given by,
th = (qH – qL)/qH = [CP0(T3 – T5) – CP0(T6 – T1)]/CP0(T3 – T5)
and the efficiency of the ideal R = 1 regenerative cycle with T2 = T6 and T4 = T5 can be written,
th = 1 – (T2 – T1)/(T3 – T4) = 1 – T1(T2/T1 – 1)/[T3(1 – T4/T3)]
5-11
Now applying the Ideal Gas results to the isentropic compression and expansion processes,
constant pressures
(T2/T1) = (P2/P1)(k – 1)/k = (P3/P4)(k – 1)/k = (T3/T4) so (T3/T4) = (T2/T1)

isentropic isentropic
Using the isentropic pressure ratio rps = P2/P1, the thermodynamic efficiency is then,
th = 1 – T1/T3 • T2/T1= 1 – T1/T3 • rps(k – 1)/k
The ideal regenerative cycle thermodynamic efficiency yields an improved th compared to the
ideal Brayton cycle for the same rps, T3, and T1, however, increasing rps now hurts the efficiency.
The rps must be low enough that T4 > T2 to permit heat transfer in the regenerator; and rps must be
high enough to keep air flow and component volumes reasonable. In practice, the industrial gas
turbines as used in a very successful class of Chevron product tankers used regeneration to get an
acceptable (slightly higher) th with a lower peak temperature so that lower grade, cheaper fuels
could be used. A series of analyses of regenerative gas turbine cycles with a modest peak
temperature T3 = 1000 K and 90% compressor and turbine efficiencies, and a 90% regenerator
effectiveness are shown in Table 5.5 to reveal this tradeoff. The minimum sac is at about rps = 6.
Table 5.5 Analysis of Regenerative Cycles with Peak Temperature 1000 K,

90% Component Efficiencies, and 90% Regenerator Effectiveness
Pressure rps
Ratio 2.0 4.0 6.0 8.0 10.0 12.0 14.0 equation used
T1 [K] 293.15 293.15 293.15 293.15 293.15 293.15 293.15 specified
(k – 1)/k
T2s [K] 357.35 435.61 489.11 531.01 565.97 596.23 623.07 T1rps
T2 [K] 364.48 451.44 510.88 557.44 596.28 629.90 659.73 T1 + (T2s - T1)/comp.
T5 [K] 790.93 680.24 626.56 592.91 569.18 551.25 537.07 T2 + R(T4 - T2)
T3 [K] 1000.00 1000.00 1000.00 1000.00 1000.00 1000.00 1000.00 specified
(k – 1)/k
T4s [K] 820.34 672.96 599.35 552.06 517.96 491.67 470.49 T3/rps
T4 [K] 838.31 705.67 639.42 596.86 566.17 542.51 523.44 T3 - (T3 - T4s)turb.
T6 [K] 411.87 476.86 523.74 561.38 593.27 621.16 646.10 T4 – R(T4 – T2)
qH [kJ/kg] 209.81 320.87 374.74 408.51 432.33 450.32 464.55 Cp0(T3 - T5)
qL [kJ/kg] 119.13 184.36 231.40 269.17 301.17 329.16 354.19 Cp0(T6 - T1)
wturb. [kJ/kg] 162.26 295.36 361.84 404.56 435.35 459.09 478.23 Cp0(T3 - T4)
wcompr. [kJ/kg] 71.58 158.84 218.50 265.21 304.19 337.93 367.86 Cp0(T2 - T1)
th 0.432 0.425 0.383 0.341 0.303 0.269 0.238 wnet/qH = 1 - qL/qH
sac [kg/kWh] 39.7 26.4 25.1 25.8 27.4 29.7 32.6 3600/wnet
L
sfc [g/kWh] 195 198 220 247 278 313 355 3600000/(thh )
5-12
More Advanced Gas Turbine Cycles. More advanced gas turbine cycles (as used in a
few Soviet marine applications) utilize waste heat recovery using recuperators (regenerators),
add intercooling with seawater, and add reheat combustors to further improve the cycle
efficiency. When multiple stages of compression are used, intercooling can be used to cool the
air between the stages of compression. When multiple stages of gas generator turbines are used,
reheat combustors can be used between the stages of turbines. The schematic of such a gas
turbine is shown in Fig. 5.9. The resulting P-v and T-s diagrams are shown in Fig. 5.10.
concentric shafts
Figure 5.9 Schematic of Gas Turbine with Intercooling, Regeneration, and Reheat [1]
Figure 5.10 P-v and T-s Diagram for Advanced Gas Turbine Cycle with
Intercooling, Regeneration, and Reheat
The thermodynamic cycle efficiency of the cycle shown in Fig. 5.10 will be as follows
using the Ideal Gas - Constant Specific Heats model,
th = wnet/qH LP turbine HP compressor Combustor

= [(T6 – T7) + (T8 – T9) ) + (T9 – T10) – (T4 – T3) – (T2 – T1)]/[(T6 – T5) + (T8 – T7)]
HP turbine Power turbine LP compressor Reheat Combustor
5-13
As usual, temperatures T1, T6, T8 and P4 are typically specified. Intermediate pressures P2 = P3,
P7 = P8, and P9 are obtained using the isentropic relationships and component efficiencies and
noting that all of the work of the HP turbine is utilized by the HP compressor; i.e. (T6 – T7) = (T4
– T3), and all of the work of the LP turbine is utilized by the LP compressor; i.e. (T8 – T9) = (T2 –
T1). The temperatures T5 and T11 are obtained using the regenerator effectiveness R;
temperature T3 is obtained using an intercooler effectiveness, which is defined as I = (T2 –
T3)/(T2 – T1). Note that unlike the regenerator effectiveness R, I can be greater than 1 since the
seawater may be at a lower temperature than the inlet air.
The intercooling lowers T4 and reheat raises T8, T9 and T10 so more regeneration can be
done (allowing more degrees of regeneration = T5 – T4). Greater regeneration raises the cycle
efficiency. If reheat and intercooling are used without regeneration (as with a simple Brayton
cycle), the efficiency may actually be reduced. They might still be used, however, to reduce
compressor work or reduce the air rate. These tradeoffs are illustrated in the example shown in
Fig. 5.11. In this Figure, temperatures are in ˚R and the work and heat transfer are also
expressed in the same units; i.e, h/CP0  T with the Ideal Gas – Constant Specific Heats
model. In the limit, as the number of reheat and intercooling stages approach infinity, the
regenerative cycle would approach the ideal Ericcson cycle that achieves the Carnot efficiency.
The use of a regenerator also lowers the final exhaust temperature, thus easing the problem of
exhaust cooling to minimize IR detectability of the ship.
Since air-to-air heat transfer is relatively ineffective (compared to water) the regenerators
add considerable volume and weight to the gas turbine. The Roll-Royce WR21 ICR (intercooled
recuperative) gas turbine is a combination of an aero-derivative gas turbine and regeneration
with intercooling between the two stages of compression. The WR21 uses three shafts: one for
the low pressure turbine driving the low pressure compressor, another concentric shaft for the
high pressure turbine driving the high pressure compressor, and a separate shaft transmitting the
power turbine output to the load. A cutaway of this engine is shown in Fig. 5.12.
Torque and Power Characteristics. The basic torque and power characteristics of a gas
turbine depends upon whether the gas turbine is a single-shaft or a two-shaft (gas generator and
separate power turbine) design. A single shaft gas turbine has a rapidly falling torque output with
5-14
Cycle Description W’Turb W’Comp W’net Q’H Q’L degrees th wA wF

Regeneration sac sfc
A Simple cycle 825 357 468 1168 700 0 40.0 22.7 0.346
B +Intercooling 825 311 514 1330 816 0 38.7 20.7 0.359
C +Regeneration 825 357 468 825 357 343˚ 56.8 22.7 0.244
D +Reheat 901 357 544 1398 854 0 39.2 19.5 0.356
E +all three 901 311 590 901 311 590 ˚ 65.5 18.0 0.212
wA [lbm air/hp h]; wF [lbm fuel/hp h]; W’ as W/Cp0 [˚R]; Q’ as Q/C p0 [˚R]
Figure 5.11 Comparison of Ideal Gas Turbine Cycles with T1 = 75 F, Overall Pressure
Ratio = 6, and Peak Temperature 1600 F
5-15
Figure 5.12 Rolls-Royce WR 21 ICR Regenerative Gas Turbine (note regenerator size)
decreasing rpm. As the rpm drops, the compressor gas flow output decreases and then so does
the turbine power output. These characteristics can be seen in Fig. 5.13 adapted from Woud and
Stapersma [1].
Figure 5.13 Comparison of Torque-Speed and Power-Speed Characteristics of Single-shaft

and Two-shaft Gas Turbines
In contrast, the ideal two-shaft gas turbine has the similar characteristics to a steam
turbine where there is also mechanical separation from the source of the working fluid (boiler or
reactor plant steam generator) and the turbine. In this case, the normalized torque and power are
given by the following if  is the decimal percent throttle or fuel rate,
5-16
QB/QBo = (2 – N/No) and PB/PB0 = QB/QBo•N/No = (2 – N/No)N/No
This gives the very desirable rising torque characteristics with decreasing rpm as shown in Fig.
5.13. The power output is essentially constant for moderate changes in rpm below rated. This
can be especially valuable in icebreaking operations.
Two-shaft gas turbines are ideally suited for ship power applications. The gas generator
can be started with a brake on the load shaft. The torque-rpm characteristics are ideal. The free
power turbine does present problems, however, on small boats where the propeller can
frequently emerge for the water. Two-shaft gas turbines usually use power control where fuel is
provided to maintain gas generator rpm without regard for propeller rpm in order to prevent fuel
oscillations in a seaway [3]. An overspeed trip protects the power turbine in the case of propeller
loss. This is in contrasted to speed control where the propeller rpm is controlled and an override
is provided to protect against excessive compressor rpm. Single-shaft gas turbines are used for
generator applications, since they can be started at no load and can operate at constant rpm.
They are also used in small and racing boats where the load can be applied by a clutch and the
mechanical link between the propeller and the compressor protects the overall system against
overspeed during frequent propeller emersions.
Problems in Marine Applications. The marine application presents some special problems for
gas turbines associated with water and salt in the air, impurities in the fuels, and dirt and sludge
coming from shipboard tanks.
Effects of Salt. Salt comes from the incoming air and from contamination in shipboard
tanks. The salt can cause three types of problems.
1. Salt can coat the compressor blades causing a reduction in compressor and a serious
loss of power output in only a few hours, especially on small craft and hovercraft.
Inertial and mesh filters are used in intakes; intakes are placed in high, inboard and aft
facing locations; the compressors are washed with fresh water at low rpm; and the
compressors are cleaned with walnut shells, rice, or other materials at high rpm to
reduce the salt buildup.
2. Salt can lead to corrosion of cold compressor metals, which is usually controlled by
proper selection of materials and the use of coatings in the marinization process.
5-17
3. Salt can also result in corrosion of hot combustor and turbine parts. At higher peak
temperatures (T ≥ 1700 F) sulfur impurities in the fuel go from a trioxide in the
presence of salt to a sulfate and then sulfuric acid (sulfidation). Control is by the
selection of materials, the use of low sulfur fuels, intake air location and treatment,
the treatment of fuel for sulfur, and/or the centrifuging of the fuel to eliminate water.
Effects of Dirt and Sludge. An important problem in using gas turbines onboard ships is
the need to keep the fuel clean, especially when normal shipboard tanks are used for fuel storage.
Fuel nozzles can easily plug leading to hot spots and combustor wall burnout. Contaminants can
be water from condensation and/or contamination; dirt and scale from tanks, especially in new
ships and heavy weather; and bacterial sludge that grows at the fuel/water interface. The same
concerns can be present, to a lesser extent, in diesel installations. The fuel systems include
features to deal with these contaminants. For example, one early GT ship filtered out about 40 lb
(55% iron oxides; 40% sea salts) from the fuel per trip during winter North Atlantic operations.
A typical fuel system might include centrifugal purifiers between the shipboard tanks and the
fuel day tank, a strainer prior to the fuel pumps, and then a series of 75 m filters, 20 m
coalescent filters, and an 8 m filter prior to the high pressure fuel pump located on the engine.
Specific Fuel Rate. We have already looked at the design condition thermodynamic efficiency
and the related specific fuel rate. One important characteristic of gas turbines is how this
specific fuel rate varies with engine operating point. Figure 5.14 shows the specific fuel rate
(lbm/hpBrh) map for the GE LM2500 gas turbine at standard Navy conditions. A primary
disadvantage of most gas turbines is the rapid increase in specific fuel rate at lower loads. This
occurs because at low loads the compressor produces greater excess air (relative to
stoichiometric conditions) through the combustors causing a significant decrease in peak
temperature and, thus, cycle efficiency. Figure 5.15 taken from Woud and Stapersma [1] also
shows a typical specific fuel rate characteristic along the propeller load curve. To address this
problem, some gas turbines, such as the Rolls-Royce WR21, use variable nozzles in the turbine
to reduce the excess flow, maintain the peak temperature, and produce a flatter sfc curve with
load. Figure 5.16 shows a typical result.
5-18
Figure 5.14 Typical Gas Turbine Specific Fuel Rate Map
5-19
Figure 5.15 Typical Gas Turbine Specific Fuel Rate with Propeller Load [1]
Figure 5.16 Typical Gas Turbine Specific Fuel Rate with Propeller Load
with Variable Nozzle Geometry (Rolls-Royce ICR WR21) [1]
References
2. Harrington, R. L. (ed.), Marine Engineering, SNAME, New York, 1971 and 1992.
3. Woodward, J. B., Marine Gas Turbines, Wiley-Interscience Publication, John Wiley and
4. “Marine Gas Turbine Power Plant Performance Practices,” SNAME Technical and Research
Bulletin 3-28, 1975.
5-20
Chapter 6 Steam Turbine Plants
The dominant propulsion plant for large naval combatants; e.g., aircraft carriers,
and nuclear submarines is the steam plant. These are nuclear reactor powered on United
States vessels. Steam plants are also used today on many Liquid Natural Gas (LNG)
carriers where there is a ready supply of LNG boil-off for use in conventional boilers.
Mitsubishi recently completed the 7th of a new class of 155,300 m3 LNG carriers that use
reheat steam plants for increased efficiency. Many steam plants are still in operation
from earlier construction. Simpler steam plants are also used for waste heat recovery and
cogeneration with diesel and gas turbine prime movers. For example, a recent Royal
Caribbean cruise ship used a COGES, combined gas turbine and steam electric drive,
system with two LM2500+ GE gas turbines for 50 eMW and a waste heat steam plant
producing another 9 eMW. Bleed steam from the steam turbine supplies steam to the
distillers, air conditioning reheat, laundry, galley, etc. for an efficient integrated system.
Analysis of Steam Cycles. We will consider various steam cycles in order to understand
their basic design parameters and tradeoffs and the function of principal components.
The simplest ideal steam cycle would consist of a steam source (residual fuel fired boiler,
waste heat boiler, or nuclear steam generator), a turbine, a main condenser, and a high
pressure feed pump.
A propulsion-size system will utilize a “turbine” consisting of many (up to 30 or so)
nozzle/blade stages that might be arranged in up to three separate casings. Most steam
plants today utilize a high pressure (HP) turbine (turning at perhaps 6000 rpm) and a
separate low pressure (LP) turbine (turning at about 3000 rpm). These turbines are
geared together through a two-input, double reduction gear. Some plants used an HP
turbine, an intermediate pressure (IP), and an LP turbine. Propulsion steam plants also
have a more compact and less efficient astern turbine built into the low-pressure end of
the LP turbine so that it can exhaust to the same main condenser. Smaller auxiliary steam
turbines used for electrical generation turbo-generator sets have fewer stages and are
constructed as a single casing or single “cylinder” unit. Even smaller steam turbines such
as a main feed pump turbine might use a single nozzle/blade stage.
6-1
Essentially all large power steam cycles, as shown schematically in Fig. 6.1, include
some form of feed heating in which the feed water leaving the condenser is preheated
using bleed steam taken from between stages of the turbine. The first feed heater that
would be used is the deaerating feed tank (DFT) because the DFT has three primary
functions. First, it is a saturated condition tank with a water-steam interface that
provides the opportunity for fluid level control within the plant. Second, it uses bleed
steam that is sprayed into its steam volume to preheat the feed water. Finally, the DFT,
along with the air ejector connected to the condenser, serves to remove any
noncondensible gases that enter the system in the portions that operate below atmospheric
pressure (deaeration). The air ejector is a steam driven venturi jet pump that pulls steam
and air from the condenser to help establish and maintain a vacuum in the condenser. The
vent from the top of the DFT is directed to a drain cooler. With the DFT having a
variable fluid level, it is necessary to add an additional, low-pressure condensate pump to
pump the process water from the condensate well at the bottom of the condenser into the
DFT. Simpler systems without a DFT must perform water level control at the condenser
condensate well.
Figure 6.1 Typical Ship Boiler-Fired Propulsion Steam Plant Schematic
6-2
Larger power steam plants will also have additional stages of feed heating that are
supplied from bleed points on the steam turbine below or above the supply to the DFT.
Feed heaters located between the condenser and the DFT are called low-pressure feed
heaters. Their saturated fluid drains are typically drained to an upstream (lower pressure)
feed heater or pumped back into the condensate path with a small drain pump. The feed
heaters located in the high pressure piping between the feed pump at the outlet of the
DFT and the boiler are called high-pressure feed heaters. Their saturated fluid drains are
usually directed to the next upstream feed heater or DFT. A large ship propulsion boiler-
fired steam cycle might have five stages of feed heating: a DFT, a low-pressure feed
heater, and three high-pressure feed heaters. A lower pressure nuclear steam cycle might
have only two-stages of feed heating: a DFT and a low-pressure feed heater.
Large propulsion boilers consist of a saturated section that produces saturated steam
and then a separate series pass back through the furnace that superheats the steam to
higher steam conditions, e.g. 1200 psig, 975 F. Nuclear steam generators only produce
saturated steam leading to much lower cycle efficiencies. Advanced cycle steam plants
might also use a second return to the boiler to perform reheat between the HP and IP
turbines or the HP turbine and the LP turbine. A schematic for a typical steam cycle for a
boiler-supplied ship propulsion steam cycle is shown in Fig. 6.1 Additional cycles can
be seen in Marine Engineering [2].
Rankine Cycle. The idealized cycle for the simple, practical steam cycle is the
Rankine Cycle. When the water-steam phase change is considered, the Carnot cycle is
impractical because it would require pumping wet steam and/or heating at constant
temperature and decreasing pressure. If the cycle were completely in the wet steam
region, the cycle would have excessive moisture in the turbine exhaust resulting in
excessive turbine blade erosion. The Rankine cycle allows the pumping of a subcooled
fluid and heating at constant pressure including the possibility of superheating the steam.
The ideal Rankine cycle is defined as follows:
1-2 reversible, adiabatic (isentropic) pumping in the main feed pump

2-3 constant pressure heating (QH) in the boiler
3-4 reversible, adiabatic (isentropic) expansion in the turbine
4-1 constant pressure condensing/cooling (QL) in the condenser
6-3
The cycle schematic and T-s diagram are shown in Figure 6.2.
Figure 6.2 Schematic and T-s Diagram for Ideal Rankine Cycle
Using the First Law neglecting kinetic energy and potential energy changes gives,
boiler qH = h3 – h2
turbine wt = h3 – h4S (ideal turbine, h4 = h4S)
condenser qL = h4 – h1
feed pump wp = h2S – h1 = v1(P2 – P1)  (ideal pump, h2 = h2S)
In SI units with pressure in kPa, is just 1. In English units with work in Btu/lbm and
pressure in psi, is (144 in2/ft2)/778.17 ft lbf/Btu. The thermodynamic efficiency of the
cycle will then be,
boiler condenser
th = (qH – qL)/qH = [(h3 – h2) – (h4 – h1)]/(h3 – h2)
= wnet/qH = [(h3 – h4) – (h2 – h1)]/(h3 – h2)

turbine pump boiler
The main parameters of the basic cycle are the steam temperature T3, the steam pressure
P3 ( ≈ P2 in simplified cycle analyses), and the condenser pressure P4 ( ≈ P1 in simplified
cycle analyses). We can consider how each of these parameters affects the
thermodynamic cycle efficiency th and what considerations limit their practical values.
1) Decreasing the condenser pressure P4 (increasing vacuum) improves th.
standard: for sea water at 75 F design condition
commercial practice – 1 1/2 “Hg abs. or 28 1/2 “Hg vacuum [1]
Navy practice – 5 “Hg abs. or 25 “Hg vacuum at rated power, which
yields smaller condensers and sea water circulating pumps; at cruise
6-4
the system will operate near 1 1/2 “Hg abs.

limitations: 1. sea conditions, may reach 85 F in Gulf of Mexico, etc.
2. condenser and circulating system size, cost, weight, etc.
3. moisture content of steam at exhaust which is limited to
x ≈ 88% due to concern for erosion of last stages of the
turbine and the condenser tubes
4. volume of exhaust flow since
vg (28 “Hg vacuum) = 339.2 ft3/lbm
vg (29 “Hg vacuum) = 652.3 ft3/lbm
2) Increasing steam pressure P3 (at constant T3) improves th.
standard: commercial – 600 psig, 900 F (ANSI standard 600 psi fittings)
later 850 psig, 950 F (ANSI 900 fittings)
with reheat 1450 psig, 950 F (ANSI 1500 fittings)
Navy practice – last non-nuclear 1200 psig, 975 F (ANSI 1500)
typical nuclear 700 psig, 550 F
limitations: 1. turbine component efficiency (not included yet) decreases
with P3 for a given power since the flow volume is smaller
2. need to utilize standard rating valves, fittings, and flanges
3. higher cost and weight of higher pressure boundary
4. exhaust moisture gets worse at constant T3 as illustrated in
Fig. 6.3. Increasing T3 to T3’ at constant P3 will help.
Figure 6.3 Effect of Increasing P3 and T3 on Exhaust Quality
6-5
3) Increasing steam temperature T3 (at constant P3) improves th.

limitations: 1. vanadium pentoxide (from fuel impurity Va) causes
slagging on superheater tubes, which may lead to
burnout, if Tmax ≥ 950-1000 F.
2. need to utilize standard rating valves, fittings, and flanges
The sensitivity of the basic cycle efficiency to P3, T3, and P4 at a base design of 850
psig/950 F/1 1/2 “Hg is shown in Fig. 6.4. Note that the base cycle has a thermodynamic
efficiency of only 37.2% with four feed heaters while diesels now reach 52% and gas
turbines can reach 40%. A nuclear steam cycle efficiency will be about 28% due to the
lower (saturated) steam conditions.
Figure 6.4 Sensitivity of 850 psig/950 F/1 1/2 “Hg Steam Cycle Efficiency
6-6
Steam Cycles with Real Components. The real steam cycle deviates from the
ideal due to the component efficiencies. Also there are losses within the boiler in the
combustion and heat transfer processes that we will consider later in B. The turbine
component efficiency definition and analysis (isentropic to get h4s and then application of
the component efficiency to obtain h4) is the same as used for gas turbines except that
enthalpies and the Steam Tables have to be used here. The feed pump component
efficiency definition and analysis are similar to that used for the gas turbine compressor.
The T-s diagram for the steam cycle with real components is shown in Fig. 6.5.
actual exhaust conditions

isentropic, ideal
Figure 6.5 T-s Diagram for Steam Cycle with Real Components
The thermodynamic efficiency for this cycle and related terms used to describe the
cycle are as follows in SI and English units, respectively:
turbine pump boiler
efficiency th = wnet/qH = [(h3 – h4) – (h2 – h1)]/(h3 – h2)
cycle steam rate = 3600/wnet [kg steam/net kWh] = 2544.43/wnet [lbm steam/net hpBRh]
turbine steam rate = 3600/wturbine [kg steam/turbine kWh] = 2544.43/wturbine
heat rate = 3600/th [kJ to steam/kWh] = 2544.43/th [Btu/hpBrh]
boiler efficiency B = heat to steam/heat potential of fuel
= heat rate/(fuel rate • hU) , where hU = Higher Heating Value
specific fuel rate = 3 600 000/(thBhU) [g/kWh]= 2544.43/( thBhU) [lbm/hpBrh]
Note that for a good (low) specific fuel rate, a steam plant must have a high cycle
thermodynamic efficiency t and also a good boiler efficiency B. As we will note
below, there are conflicts between having both these high that the total system design
must balance.
6-7
Additional losses in real cycles include piping temperature and pressure losses that
are considered in detailed cycle Heat Balance analyses by using the enthalpy appropriate
for each point in the system. There are also parasitic losses present in any real, complete
system. These include fluid losses to soot blowing (to clean soot from boiler tubes using
steam jets), steam atomization (to facilitate the fuel atomization in the boiler burners),
shaft seal gland leakage, etc. There are also non-propulsion components that represent
parasitic losses to the main steam cycle. These might include turbogenerators for electric
power generation, distilling plants for fresh water production, low pressure steam
generators for service steam, etc. These parasitic losses are also included in detailed Heat
Balance analyses.
The Reheat Cycle. One way to improve cycle efficiency is to utilize reheating.
This is actually used in conjunction with regeneration or feed heating to be considered
below. Reheating results in an improved efficiency at an increase in equipment
complexity and cost. Reheating involves removal of the steam from the turbine(s) after
partial expansion and its return to the boiler to be reheated at essentially constant pressure
to usually its initial temperature. We, thus, have a simple reheat cycle as shown
schematically and on the T-s diagram shown in Fig. 6.6.
Figure 6.6 Schematic and T-s Diagram for a Simple, Ideal Reheat Cycle
The thermodynamic efficiency for the simple reheat cycle shown in Fig. 6.6 is as
follows:
HP turbine LP turbine pump main boiler reheater
th = wnet/qH = [(h3 – h4) + (h5 – h6) – (h2 – h1)]/[(h3 – h2) + (h5 – h4)]
6-8
Now in general, the reheat improves the cycle efficiency for constant P3, T3, and P6; the
larger wnet allows smaller steam flow for the same power; and there is an optimum
pressure at which to perform the reheat for maximum efficiency. The crossover point
between the normal HP and LP turbines is often lower than optimum so most reheat plant
designs utilize HP, IP, and LP turbines and reheat between the HP and IP turbines. The
use of reheat allows the use of a higher steam pressure P3 without having exhaust
moisture problems so typical marine reheat cycles have used 1450 psig (100 bar) steam.
The principal problems with reheat cycles are that they are more complex and, thus,
potentially less reliable. They cost more; e.g., one quote was 8.5% more expensive to
buy, 4.6% more to install, and 16% more to maintain than an equivalent non-reheat
installation. These added costs can, of course, be offset by the fuel savings of the higher
efficiency. Special provisions must also be provided to protect the reheat tubes during
backing since there is no flow through the reheat path during astern operation. Reheat
cycles (with regeneration) represent the most advanced steam cycles used on ships.
The Ideal Regenerative Cycle. The second major way to improve steam cycle
efficiency is to use regeneration or feed heating. We will first motivate the use of feed
heating by recalling that the Carnot cycle efficiency for typical marine steam conditions
(p. 2-33) would be about 63%. A Rankine cycle with ideal components and operating
between the same temperature extremes would have an efficiency of only about 33%.
The main difference is that the average temperature of heat addition is higher with the
Carnot cycle since all heat addition is done at the maximum cycle temperature. With
feed heating we can preheat the feed water and, thus, raise the average temperature of
heat addition in the Rankine cycle. This will improve th towards the Carnot efficiency.
An idealized regenerative cycle would have T = 0 heat transfer between the steam
expanding in the turbine and the feed water returning to the boiler. Schematically this is
sketched along with the associated T-s diagram in Fig. 6.7. Using the fact that the area
under the T-s diagram down to absolute zero equals the amount of heat transfer and the
fact that the areas 1-2-2’-1’ and 3-4-5 are equal, it can be shown that the idealized cycle
in Figure 6.7 has, without superheating, the same qH and qL as the Carnot cycle and, thus,
6-9
the same efficiency. With feed heating from state 2 to state 2’, all heat addition is at T3
and th idealize regenerative cycle = Carnot. While this should motivate the use of feed heating,
this is not a practical solution due to the T=0 heat transfer requirement, low peak
temperature (no superheat) and the excessive exhaust moisture that would erode the
turbine and condenser.
Figure 6.7 Schematic and T-s Diagram for an Idealized Regenerative Cycle
The Real Regenerative Cycle. A practical alternative to the Ideal Regenerative

Cycle is to extract 5-10% of the steam flow at various points in the turbine expansion and
use this extraction or bleed steam to heat the feed water. The quantity of steam
continuing through the rest of the turbine is reduced, but its thermodynamic state is
unchanged. A modern marine turbine has N stages (typically 15 to 25) of expansion
through nozzles that would allow extraction at N – 1 intermediate points to feed N – 1
feed heaters. This would yield the closest approach to the Carnot efficiency. Each
additional feed heater, however, produces a decreasing improvement in efficiency. The
design tradeoff then becomes the determination if the fuel savings resulting from the
addition of another feed heater offsets the cost of buying, installing, and maintaining the
additional heat exchangers, valves, piping, etc. plus offset the added weight and volume
penalty to the overall machinery plant. For each additional heater, the optimum final feed
temperature increases further toward the boiler saturation temperature. This situation is
illustrated in Fig. 6.8. The most modern commercial marine steam plants typically have
4 or 5 stages of feed heating without an economizer (a waste heat feed water preheater
6-10
within the boiler exhaust path) or 2 stages of feed heating with an economizer. Nuclear
steam plants typically have 2 stages of feed heating.
Figure 6.8 Effect of Stages of Feed Heating on Cycle Efficiency and Optimum Boiler
Feed Temperature [3]
We will consider here the analysis of regenerative cycles in a simplified form. Real
cycle analyses require an iterative Heat Balance solution that is usually performed with
specialize computer software [4]. We will analyze an example with zero to four feed
heaters. As noted, the first feed heater to be used would be a deaerating feed tank (DFT)
as shown in Fig. 6.9. The condensate and the extraction steam are sprayed into the steam
volume to achieve deaeration (O2 reduction) needed to minimize corrosion of the high
temperature, high pressure carbon steel parts of the system. The DFT is a direct contact
feed heater with complete mixing of all fluids and a steam volume at Tsat, Psat (= 30 to
100 psia). The steam/water interface allows volume monitoring and control within the
system. The DFT is a large component and it is usually placed high in the engine room
on commercial ships in order to use gravity to help provide sufficient inlet pressure at the
main feed pump to prevent pump cavitation. On military ships, the DFT cannot be
placed as high as needed so a main feed booster pump is used to apply the main feed
pump inlet pressure required to prevent cavitation.
The second feed heater to be added could be either a low-pressure feed heater (in
the low-pressure condensate system before the DFT) or a high-pressure feed heater (in
the high-pressure feed system after the feed pump). Both are surface type shell-tube heat
exchangers with (1) no mixing of fluids except possibly in the shell, (2) the fluid to be
6-11
heated in the tubes, and (3) the extraction steam being condensed in the shell. At typical
high pressure feed heater is shown in Fig. 6.10.
Figure 6.9 Typical Deaerating Feedtank (DFT) Arrangement [2]
Figure 6.10 Typical Horizontal High-Pressure Feed Heater [2]
6-12
The extraction steam for high pressure feed heaters is usually well superheated so
the heat exchanger includes a desuperheating zone (circle 11 in Fig. 6.10), a series of
baffles to create a pressure drop, and also a similarly baffled subcooling zone (circle 12).
The resulting temperatures of the heating and heated fluids are shown schematically in
Fig. 6.11. In preliminary design, the marine engineer picks the temperature he or she
wants from the feed heater so temperatures at points a and b in Fig. 6.11 are known. The
details of the feed heater are not known at this stage so reasonable values are assumed for
the following:
TTD = terminal temperature difference = Tsat on shell side – Tb

TD = temperature difference = Tdrain from subcooling zone – Ta
With these assumptions, Tsat (therefore Psat) in the shell can be found, the appropriate
extraction point can be determined, and the drain conditions can be found. The Mollier
chart can be used to obtain the extraction steam conditions at the desired Pextraction = shell
Psat.
Figure 6.11 Temperature Change within High-Pressure Feed Heater
The low-pressure feed heater can be a single shell-tube heat exchanger, but often it
is combined within a single shell with the gland exhaust condenser that cools exhaust
from the turbine shaft seal glands, the DFT vent flow, and possibly also the air ejector
condenser outlet flow; and a separate cooler for its own drains. A cutaway of a typical
low-pressure feed water heater is shown in Fig. 6.12. The drain from the low-pressure
feed water heater is usually saturated. Due to its lower shell pressure, the bleed steam for
the low-pressure heater usually has negligible, if any, superheat. Again, the designer
usually chooses the desired heater outlet temperature Tb and with a reasonable
6-13
assumption of a TTD (usually about 10 F), the required shell side Tsat, Psat = Pextraction can
be found. The Mollier chart can again be used to obtain the extraction steam conditions.
Figure 6.12 Typical Low-Pressure Feed Heater [2]
Regenerative Cycle Analysis. Consider a cycle with initially just a DFT and a
low-pressure feed heater. We can approach this using the following general procedure:
1) draw the cycle schematic and T-s diagram showing all states;
2) indicate the flows in each branch using a normalized basis of 1 kg/h at the throttle;
3) establish all enthalpies using desired heater outlet conditions, heat exchanger
assumptions (TTD and TD), component efficiencies, turbine state line on the
Mollier chart, etc.
4) use the First Law at each feed heater to obtain the necessary extraction fractions
xi  kg or lbm extracted for heater i/kg or lbm entering the turbine throttle
(do not confuse this xi with the thermodynamic property quality)
5) determine wnet and qH using the correct flows in each part of the system;
6) determine th, steam rate, heat rate, etc. as required.
Consider now the schematic shown in Fig. 6.13.
6-14
Figure 6.13 Schematic and Data for Cycle with Two Stages of Feed Heating
Here we will neglect the gland exhaust condenser and separate drain cooler within
the low-pressure feed heater. The T-s diagram will appear as shown in Fig. 6.14. In this
sketch, it is only a coincidence that state 6’ is saturated. The low-pressure heater heats
the condensate from state 2 to state 2’; the DFT heats the condensate from state 2’ to state
3. Here, we assume that the crossover, state 6, is at 80 psia; actually it would probably be
higher and a reducing valve would be used to ensure a constant DFT pressure.
Figure 6.14 T-s Diagram for Two Feed Heater Example
6-15
Next we need some design logic to select T2’. Thermodynamically we will get the
best th (and reasonable components) if each stage of feed heating heats the condensate
the same amount so let T3 – T2’ = T2’ – T1. The condensate pump heats the fluid very
little due to its low P and this can be neglected in this calculation so we have,
T3 – T1 = 312 – 91.7 = 220 F so T per heater ≈ 110 F

let T2’ = 92 + 110 = 202 F
and we want to extract steam at a pressure to yield,
T sat(shell) = T6” = T2’ + TTD = 212 F

giving Pshell = P6” = P6’ = 14.7 psia
as the desired extraction pressure neglecting the small pressure drop in the piping. At this
stage, we can neglect piping pressure drops throughout so we can obtain all the needed
enthalpies. We need the turbine state line to get h6 and h6’. We can obtain this from the
vendor or approximate the state line. The crudest acceptable approach is to apply the
stated turbine efficiency across the entire turbine and then draw a straight line between
states 5 and 7 on the Mollier chart. Enthalpies h6 and h6’ can then be obtained at the
points where 80 psia and 14.7 psia lines cross the state line. This preliminary design
approach will yield the situation depicted in Fig. 6.15. If we were to apply turbine from h5
to h6 or from h6 to h6’, rather than across the whole turbine in one step, we would get a
different result since the constant pressure lines are not parallel and equally spaced.
Figure 6.15 Sketch of h-s Turbine State Line for Example
6-16
Next we need to obtain the extraction fractions x1 and x2 by using the First Law to
write heat balances for each feed heater assuming that they are adiabatic with the outside
due to their insulation. These heat balances are shown in Fig. 6.16 where the First Law
is applied on a total H = mh basis.
continuity of mass with mixing
no mixing
Figure 6.16 Heat Balances for Feed Heaters to Obtain Extraction Fractions
Finally, to obtain wnet and qH we only need to carefully keep track of the actual flow
through each branch for 1 lbm flow at the throttle.
HP turbine 1st part of LP turbine 2nd part of LP turbine feed pump condensate pump
wnet = 1• (h5 – h6) + (1– x2)(h6 – h6’) + (1 – x2 – x1)(h6’ – h7) – (h4 – h3) – (1 – x2)(h2 – h1)
qH = 1• (h5 – h4)
Now we can complete the numerical example for the cycle shown in Fig. 6. 13. Using the
data, we get the following results:
h1 = hf (1 1/2 “Hg abs.) = 59.7 Btu/lbm

h2 = h1 + hcondensate pump = 60.1 Btu/lbm
hcond. = hs/p = [v1(P2 – P1) 144 in2/ft2/778 ftlbm/Btu]/0.60 = 0.39 Btu/lbm
h2’ = h(P2’, T2’) = 170.3 Btu/lbm
h3 = hf (80 psia) = 282.2 Btu/lbm, saturated conditions in the DFT
h4 = h3 + hfeed pump = 286.4 Btu/lbm
hfeed pump = hs/p = [v3(P4 – P3) 144/778]/0.60 = 4.2 Btu/lbm
h5 = h(P5, T5) = 1482.2 Btu/lbm
6-17
h6 = h(state line, P6) = 1288.0 Btu/lbm

h6’ = h(state line, P6’) = 1177.0 Btu/lbm
h6” = hf (P6”) = 180.2 Btu/lbm
h7 = h5 – turb.(h5 – h7s) = 1022.8 Btu/lbm h7s = h(s5 = s7s, P7), Mollier chart
x2 = (h3 – h2’)/(h6 – h2’) = 0.100

x1 = (1 – x2)•(h2’ – h2)/(h6 – h6”) = 0.0995
wnet = 413.0 Btu/lbm

qH = 1195.8 Btu/lbm
th = wnet/qH = 34.6%

cycle steam rate = 2544.43/wnet = 6.16 lbm/hpBrh
heat rate = 2544.43/th = 7350 Btu/hpBrh
We can also compare the results for the system shown in Fig. 6.13 using this simple
analysis for a regenerative cycle with 0, 1, 3, and 4 feed heaters. Using B = 0.88 and a
Higher Heating Value hU = 18,650 Btu/lbm, this yields the English unit results shown in
Table 6.1. You can see the diminishing returns from adding each additional feed heater. As
noted above, the most modern marine steam cycles usually use 4 or 5 stages of feed heating
without an economizer in the boiler. Modern marine reheat cycles usually use 5 stages of
feed heating.
Table 6.1 Numerical Results for Feed Heating for Cycle Shown in Figure 6.13
stages of th x1 x2 x3 x4 m* specific fuel

feed heating [lbm/hr] rate [lbm/hpBrh]
0 32.1 140,000 0.483

1 34.0 0.1810 157,000 0.456
2 34.6 0.0995 0.1000 154,000 0.448
3 34.8 0.0890 0.0980 0.0887 165,925 0.445
4 35.1 0.0830 0.0777 0.0814 0.0870 179,175 0.442
* for 25,000 shp net from cycle; turbine assumed to drive feed pump, too.
The Heat Balance Process. Design calculations for real steam cycles including the
actual turbine state line, turbogenerators, distilling plants, fluid losses, service steam, etc.
are performed by computer [4] or by hand as a Heat Balance process. Industry standards
and guidelines are included in SNAME T&R Bulletin #3-11 [1]. For a complex cycle,
6-18
the solution must be iterative and this is suited to computer solution. The Solver
capability within Excel could also be used effectively.
Preliminary heat balances are performed early in design to accomplish the
following:
1. perform design tradeoff studies on number of feed heaters, cycle conditions, etc.
2. establish the 24 hour average fuel rate for economic studies,
3. provide the basis for sizing and specifying the various cycle components that
must be ordered very early in design.
Additional heat balances are performed later to evaluate partial power and off-design
conditions, provide the basis for evaluation of plant performance on sea trials, etc. These
later heat balances are somewhat different because each of the major components has
already been specified and you are working with “known” characteristics rather than
being free to specify them; e.g., you know the actual feed heater characteristics (heat
transfer surface area, materials, etc.) rather than assume typical TTD and TD values.
The complex cycle cannot, in general, be solved directly for the xi using
simultaneous equations as used in our sample problem for a regenerative cycle, but the
basic approach carries over. A brief outline of the classic iterative procedure for a
preliminary design heat balance is as follows:
1. establish steam conditions, condenser vacuum, number of feed heaters and basic
cycle arrangement for the calculation;
2. establish the desired outlet temperature from each heater using knowledge of
optimal final feed temperature, approximately equal temperature rise across
each heater, etc.;
3. establish necessary extraction pressures using desired heater outlet
temperatures, TTD and TD assumptions, pressure drop estimates, etc.;
4. obtain or estimate the turbine state line and steam demand for steam auxiliaries,
losses, etc.;
5. assume an initial value for the throttle flow and each extraction flow, perhaps
starting from the results of a simplified analytical solution as shown in Table 6.1
as the first guess.
6-19
6. move around the cycle in the direction of the flow beginning at the superheater
outlet and update the flows and conditions based upon the current inputs at each
component, junction, etc. using continuity and the First Law;
7. using the current inlet flows at each feed heater, calculate a new extraction flow
needed to yield the desired heater outlet temperature using a First Law heat
balance;
8. when the cycle is completed back to the boiler, update the throttle flow to the
main turbine to get the correct power output with the latest extraction flows;
9. return to step 6 and iterate until the flows converge.
With reasonable initial guesses 5-10 iterations are usually enough.

The only new development needed is in step 8. Replacement factors are derived to
facilitate this step. When you extract steam from the turbine, you will get less power
from the turbine unless you increase the throttle flow since the flow through the final part
of the turbine is reduced. The replacement factor is, therefore, defined as follows:
Ri  the fraction of the extraction flow to heater i, mi, that must be added to the
throttle flow without extraction flow to heater i in order to maintain the
same turbine output.
With no extraction, we have a non-extraction steam rate for the turbine which is
estimated with the state line. The non-extraction flow at the throttle in English units
will be,
mnon-extraction = 2544.43 • desired P[hpBr]/(hthrottle – hexhaust)
This flow and the central part of the state line are basically unchanged by extraction
(a small correction to hexhaust due to a lower flow rate to the condenser is
incorporated). At the end of step 7, we have new values for the extraction flows, mi,
so from the definition of Ri step 8 can be performed using:
n
mextraction = mnon-extraction +  Ri mi
i=1
The mnon-extraction and the Ri can be updated at each iteration for the new hexhaust if it is
corrected. The replacement factor can be defined to be,
6-20
Ri enthalpy drop lost due to extaction i/total enthalpy drop through whole turbine
(hi – hexhaust)/(hthrottle – hexhaust)
where hi and hthrottle are constant and hexhaust is usually updated at each iteration. The
expression for Ri would have to be adapted appropriately in a reheat cycle. This
definition for the replacement factor can be demonstrated using our example regenerative
cycle results: n
m extraction = mnon-extraction +  Ri xi mi extraction since mi = xi mi extraction

i=1 n
mextraction = 2544.43 P/[(hthrottle – hexhaust) • (1 – Rixi)]
i=1
and using the net turbine work from our earlier two feed heater example,
mextraction. = 2544.43 P/wturbine

= 2544.43 P/[(h5 – h6) + (1 – x2)(h6 – h6’) + (1 – x1 – x2)(h6’ – h7)]
= 2544.43 P/[(h5 – h7) – x2(h6 – h7) – x1(h6’ – h7)]
= 2544.43 P/{(h5 – h7)[1 – x1(h6’ – h7)/(h5 – h7) – x2(h6 – h7)/( h5 – h7)]}
R1 R2
An example heat balance for a 30,000 shp non-reheat, 5 feed heater marine steam
propulsion plant is shown in Fig. 6.17. In this heat balance, the extraction fractions EF
are the replacement factors Ri defined here.
Boilers. The combustion in boilers is analyzed similar to our discussion of the air
requirements for diesel engines except that there different losses. The stoichiometric air
fuel ratio requirement for residual fuel can be determined chemically to be about,
 ≈ 13.75 lbm or kg dry air/lbm or kg fuel

A real boiler cannot achieve perfect, equilibrium conditions reflected in the
stoichiometric results so depending upon the burner arrangement, atomization
effectiveness, furnace volume, furnace shape, air preheating, etc. excess air needs to be
provided above the stoichiometric result. A typical amount with air preheating is about
15% excess air and typical without air preheating is about 20%. Some highly refined
burner arrangements can get this down to 3%. There must also be allowances for air
leakage: usually 2.5% for casing leakage plus another 15% if a rotary regenerative air
preheater is used. This yields the real requirement,
6-21
Figure 6.17 Example Heat Balance for a 30,000 shp Steam Propulsion Plant
6-22
excess air leakage

afr = 13.75 lbm dry air/lbm fuel • 1.0165 lbm wet air/lbm dry air • 1.15 • 1.175
= 18.9 lbm wet air/lbm fuel = air fuel ratio
which compares to 26-40 for diesels and 30-60 for gas turbines.
Boiler efficiency B depends primarily on the final stack temperature and to a lesser
extent on the incoming air temperature as shown in Fig. 6.18. Modern commercial boilers are
able to reach 88-90% efficiency. This value is limited by the need to maintain the flue gases
above the dew point to prevent the formation sulfuric acid from the SO3 combustion products.
The final temperature is, thus, limited to 240-280 F or 116-138˚C. Military boilers were more
weight and space limited and most did not have efficiencies above 75 or 80%. The efficiency
can be calculated by an output method that compares the output steam enthalpy increases to the
fuel potential or by a losses method that details the losses from the boiler. See SNAME’s
Marine Engineering for a detailed treatment [2].
Figure 6.18 Boiler Efficiency Dependence in Stack and Inlet Temperatures [2]
A sketch of a typical natural circulation D-type marine boiler is shown in Fig. 6.19. This
boiler has an economizer in the stack to preheat the feed water, superheater tubes in the center,
6-23
and five burners on the front of the boiler. Other designs are top fired or have the burners
oriented tangentially at the corners of the furnace region to create a swirling motion for more
complete combustion – and less excess air requirement.
Figure 6.19 Cross Section of a Natural Circulation Marine D-type Boiler
A naturally circulating boiler producing 850 psig steam with saturated fluid in the coolest
tubes (Tsat = 528 F) could not cool the stack gases below about 650 F (B = 80% in Fig. 6.18)
without some auxiliary heat exchangers to cool the gases closer to the 280 F needed for
maximum efficiency. Thus, economizers or air preheaters or both are needed to get B to 88-
90%. As noted, an economizer is a flue gas feed water preheater. This can present a conflict
to regeneration within the steam cycle. To get the flue gases down to 320 F (for B = 88%),
the feed water must be 270-280 F so this does not allow much feed heating in the steam cycle
6-24
if the economizer is the last heat exchanger in the stack. In this case, steam air preheaters
would be used to improve combustion. More feed heating is possible in the steam cycle if an
air preheater is used after or in place of the economizer. The most effective air preheater is the
rotary regenerative air preheater in which replaceable baskets of heat transfer surfaces are
rotated continuously between the hot flue gases and the incoming air. This air preheater can
get the stack gases down to 240 F, permit more steam cycle feed heating, and get the boiler
efficiency up to 90%. The seal leakage between the two halves requires more excess air,
however.
Because some auxiliary steam applications cannot effectively utilize highly superheated
steam, boilers are equipped with desuperheaters to supply auxiliary steam with only 50-100 F
superheat. This approach is used rather than just using the saturated steam from the steam
drum to ensure that there is a cooling flow through the superheater tubes during warmup and
standby when the main turbines are secured. In Fig. 6.19, desuperheater tubes in the water
drum at the bottom left cool the superheated steam. In some designs, a control desuperheater
or atemporator is used to remix desuperheated steam with the superheated steam to provide
precise temperature control needed to ensure that maximum system design temperatures are
not exceeded in transients.
Ultra Steam Cycle (UST) Reheat Plant. The most advanced marine steam plant produced
today is the Ultra Steam Cycle (UST) reheat plants provided by Mitsubishi Heavy Industries
(MHI) for recent LNG carriers. The schematic of this cycle is shown in Fig. 6.20. It utilizes
reheat between a high pressure turbine and an intermediate pressure turbine to be able to reheat
at the optimum pressure. Steam conditions from the boiler are 10 MPa (1450 psi) and 560˚C
(1040 F), compared to 6 MPa/515˚C typical of the most advanced non-reheat plants. The
stated efficiency is “equal to the new diesel plants” [5]. It has two stages of feed heating (a
low pressure feed heater and a DFT) and a large economizer to get the stack temperatures
down. A shaft generator/motor is used to permit full power with the reheater down and
bypassed. The boiler, reheater and economizer are arranged in series up the stack casing with
the turbines and reduction gear below on the lowest level of the engineroom. The plant
produces equivalent CO2, but much lower levels of NOx, than either a diesel plant or a
combined cycle gas turbine/waste heat steam turbine electric plant for an LNG carrier.
6-25
Figure 6.20 Ultra Steam Turbine (UST) Reheat Cycle Schematic [5]
Reactor Steam Plants. Most steam plants produced in the U.S. today are associated with
naval nuclear plants on carriers and submarines. All naval reactors today are pressurized water
reactors (PWR) where the primary coolant that passes through the core is normal water and this
water also provides the moderation that slows fast neutrons to the thermal speeds needed to
sustain the chain reaction. These reactor plants have two separate circulating systems. The
primary coolant passes through the core and is entirely contained within the shielded reactor
compartment, because it becomes short-term highly and long-term slightly radioactive as it
passes through the core. The primary coolant heat is transferred to heat steam within a steam
generator that is also located within the reactor compartment shielding. Thus, the steam that
enters the essentially conventional (but low steam conditions) steam plant engine room does
not present a radiation exposure potential to the crew under normal conditions.
The reactor compartment component arrangement of a typical pressurized water reactor
plant is shown in Fig. 6.21. This reactor compartment contains a central reactor pressure
vessel that contains the Uranium core and the control rods, four primary coolant loops each
with a reactor coolant pump, four steam generators, and a pressurizer to provide primary
6-26
system water level and pressure control. All the primary coolant system is located within a
shielded containment volume surrounded by heavy lead and steel (for gamma ray shielding)
and light water and/or polyethylene (for neutron shielding). A cross section of a typical reactor
pressure vessel is also shown in Fig. 6.21. The coolant enters at 550 F and leaves at 600 F so
with this low T a high flow is needed. This example reactor vessel shows instrumentation
nozzles in the bottom of the pressure vessel. This would not be allowed in a military reactor
pressure vessel for reliability reasons.
Figure 6.21 Reactor Compartment Schematic Arrangement and

Reactor Pressure Vessel [6, Westinghouse]
Cross sections of the associated steam generators and pressurizer are shown in Fig. 6.22.
The steam generator has a large U-tube bundle through which the primary coolant passes. The
feed water is supplied to the shell of the steam generator and this boils to produce saturated
steam. There are complex moisture separator features at the top to remove as much moisture
from the steam as possible. This design presents an important cycle limitation since the
resulting steam is essentially saturated without the superheat needed for highest cycle
efficiencies.
6-27
The safe operation of the core requires that the primary coolant does not boil (remains
subcooled) under all conditions so an excess pressure must be imposed upon the system. This
pressure is imposed by the pressurizer. The pressurizer provides pressure control by having a
higher temperature saturated steam volume that is heated by electric heaters when the pressure
falls and cooled by a water spray when the pressure rises too high. The steam-water interface
also allows the coolant volume monitoring and control needed for the primary coolant system.
Figure 6.22 Steam Generator and Pressurizer [6, Westinghouse]
The basic design of PWR systems and their steam generators lead to very modest steam
conditions and thermodynamic efficiencies of only about th = 28%. The nuclear power
provides such compelling other advantages, however, that the low efficiency is offset. If the
pressurizer were designed to operate at 1850 psia with Tsat = 625 F and there is a temperature
margin to ensure that the core does not develop a steam bubble, the core primary coolant outlet
6-28
temperature might then be limited to about 575 F. With a normal T across the counter-flow
steam generator, the steam outlet temperature might be limited to about 510 F. With a
saturated outflow, the steam pressure would then be only about 700 psig. This compares to
850 psig, 950 F steam conditions in modern non-nuclear systems. With saturated steam
entering the HP turbine, the moisture content of the crossover to the LP turbine will be too high
to prevent erosion within the LP turbine. To deal with this, a large mechanical moisture
separator is installed between the HP and LP turbines to reduce the crossover steam moisture
content to about 2% (x = 0.98) at approximately constant pressure before it enters the LP
turbine. The resulting Mollier chart stateline is shown in Fig. 6.23. This is a large additional
component that must be accommodated within the arrangements and design of a nuclear steam
plant.
Figure 6.23 Stateline through a Nuclear Steam Turbine with a Moisture Separator
6-29
References
1. “Marine Steam Power Plant Performance Practices,” SNAME T&R Bulletin #3-11,
SNAME, New York, 1972.
3. Haywood, R. W., Analysis of Engineering Cycles, Pergamon Press, Oxford, UK, 1967.
4. Parsons, M. G. and LaGuardia, T. L., "A Preliminary Heat Balance Computer Program
Conforming to SNAME T&R Bulletin #3-11", Marine Technology, Vol. 17, No. 2, April
1980.
5. Ito, M., Hiraoka, K., Matsumoto, S. and Tsumura, K., “Development of High Efficiency
Marine Propulsion Plant (Ultra Steam Turbine),” MHI Technical Review, Vol. 44, No. 3,
Sept. 2007.
6. Lamarsh, J. R. and Baratta, A. J., Introduction to Nuclear Engineering, Third Edition,
Prentice Hall, Upper Saddle River, NJ, 2001.
6-30
Chapter 7 Turbine Theory
Steam turbines and gas turbines and axial compressors operate on similar principles. Here
we will consider the primary thermodynamics and the principal design features of these turbines.
Turbines consist of a series of nozzles that convert thermodynamic energy, h, to velocity
followed by blades or “buckets” that convert the velocity to rotational mechanical work. A
marine propulsion steam turbine might consist of as many as 25 or 30 stages each consisting of
a ring of nozzles followed by a ring of blades around the circumference of a rotor. An LM2500
gas turbine has a total of 8 stages of nozzles and blades, 2 in the gas generator turbine followed
by 6 in the free power turbine (see Fig. 5.1).
Nozzles. A nozzle is a passage with a decreasing or decreasing and then increasing flow area.
It can be convergent – always decreasing, or convergent-divergent – decreasing to a minimum
area called the throat and then increasing. Consider the flow through an ideal (isentropic) nozzle
with inlet at state 1 and exit at state 2s as shown in Fig. 7.1.
Figure 7.1 Schematic and h-s Diagram for Nozzle
The governing equations are the First Law, the Second Law, and continuity so,
First Law: q + h2s + V2s2/2 + z2g = h1 + V12/2 + z1g + w

zero negl. negl. zero
Second Law: s2s = s1
Continuity: m = AiVi/vi = constant
7-1
The velocity created by this ideal nozzle will then be V2s2 = V12 +2(h1 – h2s) and with the inlet
velocity small V12 << V2s2,
V2s ≈ [2(h1 – h2s)] ½
For a real nozzle, the expansion will be to state 2 rather than 2s which will produce,
V2 ≈ CVV2s = N½ V2s ≈ [2N(h1 – h2s)] ½ ≈ [2(h1 – h2)] ½
where the nozzle coefficient (efficiency) N = (h1 – h2)/(h1 – h2s) ≈ V22/V2s2, the ratio of kinetic
energies; and the velocity coefficient CV is the ratio of the exit velocity to the exit velocity that
would exist with isentropic expansion to the same exhaust pressure. Typically 0.92 ≤ N ≤ 0.96
and 0.96 ≤ CV ≤ 0.98 in reasonable designs.
Convergent nozzles will accelerate the flow up to the point where the speed of sound in the
fluid is reached at the outlet of the nozzle. Beyond this point the flow rate cannot increase further
and the nozzle is said to have choked flow. Convergent-divergent nozzles will accelerate the
flow and then decelerate the flow in the divergent section unless the speed of sound is reached
in the throat (minimum area) of the nozzle. In this condition, the maximum flow rate, or choked
flow, is reached and the flow can continue to accelerate to supersonic speeds in the divergent
part of the nozzle if the discharge pressure is low enough. If the downstream pressure is between
the value that will first produce sonic speed in the throat and the pressure that will first produce
acceleration through the entire divergent part of the nozzle, a normal shock (jump in
thermodynamic properties and velocity from supersonic to subsonic) will occur in the divergent
part of the nozzle producing a subsonic discharge velocity.
The choked flow condition will occur when the pressure at the minimum area point in the
nozzle reaches 0.5283 of the inlet pressure in ideal gas dry air. Experimentally this ratio is about
0.545 for superheated steam and about 0.577 for saturated steam. If the nozzle discharge
pressure is reduced further below the value that yields this pressure ratio, choked flow exists in
the nozzle and the flow rate cannot be increased further.
Work from a Turbine Blade. Consider a moving turbine blade with a tangential velocity VB.
If we write the steady-flow, adiabatic First Law in absolute coordinates where the work is w and
7-2
subtract the First Law in relative coordinates where there is no blade motion and, thus, no work;
we can obtain an expression for the work from a single blade. Consider a blade vector diagram
as shown in Fig. 7.2.
Figure 7.2 Blade Vector Diagram for Blade Moving at VB
The First Law in the two coordinate systems will be as follows:
Absolute coordinates h2 + V22/2 = h3 + V32/2 + w

Relative coordinates h2 + V2r2/2 = h3 + V3r2/2 + 0
and subtracting we can solve for the work w in terms of the four velocities,
w = [(V22 – V2r2) – (V32 – V3r2)]/2
A second approach for finding the work is through the use of the Momentum Equation
statement of Newton’s Second Law which for an open system becomes,
F = d(mV)/dt or with constant vector velocities Vi and Ve
F = d(mV)/dt|CV + miVi – meVe which for steady-state, steady-flow yields
F = m(Vi – Ve)
This vector expression reduces to just the tangential direction since there is only motion in this
direction. Thus, the tangential force Rt produced on the moving blade with be given by,
Rt = m(Vit – Vet) and the work will be
7-3
w = W/m = Rtdx/dt/m = RtVB/m = VB(V2rcos2 + V3rcos3) or,

= VB(V2cos + V3cos)
The Simple Impulse Turbine or Rateau Stage. The most common marine steam turbine type
is the impulse turbine. In this turbine, the entire pressure drop in the ideal stage occurs in the
nozzle and the blade receives a tangential force solely by redirecting the relative flow through
the blade angle 2 + 3. In the impulse turbine 2 ≈ 3 and the incoming relative velocity V2r is
slowed a small amount by friction to CBV2r = V3r where CB is the blade velocity coefficient (0.85
≤ CB ≤ 0.92). The nozzle angle is 5 ≤  ≤ 25˚ and usually less than 15˚ on large turbines. Using
this geometry in the work expression above in terms of the relative velocities and expressing the
velocity in terms of the blade speed/steam speed ratio VB/V2 yields,
w = V22[(VB/V2) cos – (VB/V2)2] (1 + CBcos3/cos2)
Note that the final term in parenthesis is close to 2 since CB ≈ 0.90 and 2 ≈ 3. The energy
available to the blade is V22/2 so we can express a blade diagram efficiency as,
B’ = w/ V22/2 = 2[(VB/V2) cos – (VB/V2)2] (1 + CBcos3/cos2)
Notice from this result that B’ is quadratic in the blade speed ratio VB/V2. Most turbines are
designed so that the blade speed ratio VB/V2 at the design point will yield maximum B’ or,
VB/V2| max B’ = (cos)/2 nearly 1/2 for small , giving
B’ max = cos2(1 + CBcos3/cos2)/2 ≈ cos2 nearly one for small 
Recall that the steam speed V2 is set by the enthalpy drop (or pressure drop) across the nozzle
and the nozzle efficiency N.
The blade speed VB is given by the rotor N [rpm] and the local radius r of the blade element,
VB = 2rN/60
This is limited by available radius, centrifugal stresses in the blade element, and the moisture
content of the steam. The blade material and surface treatment determines the limiting velocity
with respect to erosion from the moisture.
The complete turbine stage now has an efficiency given by,
7-4
s = NB’M
Where N is the nozzle efficiency, B’ is the blade diagram efficiency, and M is the mechanical
efficiency. The mechanical losses in M include windage losses due to friction of the non-
working parts of the rotor; diaphram packing or leakage losses due to leakage around the nozzles
and blades; partial admission losses due to some blades churning dead steam and having to
accelerate the steam when they enter the nozzle region if the nozzles do not extend 360˚ around
the turbine; moisture losses due to the moisture droplets hitting the back of the blade due to their
slower absolute velocity; radiation losses due to heat loss from the turbine; and friction losses
due to bearing friction.
Velocity Compounded Impulse Stages or the Curtis Stage. For acceptable values of
blade speed VB and the enthalpy drop across most turbines, V2 will be too large for optimum
blade diagram efficiency B’ with a simple single stage impulse turbine. This leads to the use of
multiple stages so that the h and, therefore, the V2 will be closer to optimum for each stage.
There are two approaches to staging with the impulse turbine: velocity compounding and
pressure compounding.
In velocity compounding, the nozzle still has the total enthalpy or pressure drop (ideally)
for the entire stage, but two or three moving blade rows are used following the nozzle. Between
the moving blades is a fixed blade which does not have a nozzling effect but simply redirects the
flow into the following moving blade. This appears as shown in Fig. 7.3 for a Curtis stage or
two-row velocity compounded impulse stage. If we were to analyze the work from this stage as
we did above and establish the blade speed/steam speed ratio for the maximum blade diagram
efficiency, we would get the following neglecting friction:
VB/V2|max B’ = (cos)/4 and B’ Max = cos2

With the optimum ratio one-half that for the simple impulse stage, the Curtis stage can utilize
twice the V2 as can the Rateau stage for the same VB and still have optimum blade diagram
efficiency. Note that without friction, the optimum efficiencies are the same; with friction the
Curtis stage will be somewhat less due to more overall blade surface area subjected to friction.
7-5
Figure 7.3 Arrangement and Conditions through Curtis Stage
There are also three row velocity compounded impulse turbines used in some small
auxiliary turbine applications. They have VB/V2| max B’ = (cos)/6 and again the optimum
efficiency B’ max = cos2without friction that will be even larger in this case.
Pressure Compounded Impulse Stages. Impulse turbines can also be staged by simply
stacking a series of simple Rateau stages (nozzle with all the P followed by a moving blade
with only a small P due to just friction). This can allow the h for each stage to yield the
optimum blade speed/steam speed ratio and yield optimum or near optimum overall results. This
is the approach used in many marine steam power turbines.
The Reaction Turbine. A completely different concept is to have the moving blades have the
same nozzle effect as the fixed nozzles. There will, therefore, be significant pressure drop
(enthalpy drop) across the entire ideal stage. The degree of reaction is defined as follows:
R = h across the moving blade/h across the whole stage
By this definition, the impulse turbine could just be considered a reaction turbine with R = 0.
Some marine propulsion steam turbines and all axial marine gas turbines are reaction turbines.
Some steam turbines include stages with R varying from 0 at the high pressure end increasing to
R = ½ at the low pressure end. The R = ½ or pure reaction stage, used in most gas turbines,
7-6
receives about half of its tangential force impulsively by redirecting the flow and half by reaction
or jet action by converting enthalpy drop to more velocity within the moving blade.
If we were to analyze the work from the R = ½ reaction stage and establish the blade
speed/steam speed ratio for maximum efficiency as we did above for the simple impulse turbine,
we would obtain,
w = V22[2(VB/V2)cos – (VB/V2)2]
Using NB here for the efficiency of both the nozzle and blade diagram, which corresponds to
NB’ for the impulse stage, this can be differentiated to obtain the blade speed/steam speed ratio
for maximum efficiency giving,
VB/V2| max B = cos and NB max = Ncos2

Considering both the fixed nozzle and the moving blade, we have an optimum blade speed/steam
speed ratio that is twice that of the simple impulse turbine with the same maximum ideal
efficiency. The entire reaction stage will now have s = NBM.
The comparison of the stage efficiency versus velocity ratio VB/V2 for the various types of
turbines with friction included is shown in Fig. 7.4. This result is extremely important because
it is the basis for the choice of turbine types for various applications. Note that the peak
efficiencies are at blade speed/steam speed ratios (cos)/4, (cos)/2, and cos for the Curtis
stage, Rateau stage, and pure Reaction stage, respectively. With friction, the Curtis stage (higher
V2 and more parts) has lower efficiency than either the Rateau or reaction stage.
Figure 7.4 Comparison of Stage Efficiencies for Turbine Types [1]
7-7
A physical comparison and pressure and velocity plots of the various turbine types are
shown in Fig. 7.5. Note the basic difference in construction. The impulse turbine does not have
significant pressure drop across the rotor (it may even have pressure equalization holes in the
disk as shown in Fig. 7.3) so it is constructed as a disk on a shaft with the moving blades at the
outer rim. The reaction turbine has a large pressure drop across the moving blades so it is
constructed as a drum rotor with just the blades extending from the drum to facilitate the
mechanical design. This construction difference has implications for the mechanical efficiency
M. The seal areas between the casing and the rotor lead to the leakage losses. On a pressure
compounded impulse turbine the annular nozzle diaphragms that extend inward from the casing
and experience the nozzle pressure drop, extend all the way down to the shaft so that the seal
circumference is small. On a reaction turbine both the nozzles and the blades have pressure drop
and, thus, need seals. Further, the circumference of each of these seals is large due to the large
rotor diameter. The impulse turbine will have larger windage losses since there is a large surface
area of non-working disk spinning in the steam. In contrast, the reaction turbine has almost no
non-working rotor in contact with the steam or air to experience frictional drag.
Figure 7.5 Comparison of Pressure, Velocity, and Construction of Turbine Types [1]
7-8
The Reaction (R = 1/2), Rateau, Curtis Stage Tradeoff. Designing or selecting a turbine
for maximum efficiency, we have the following since VB is generally fixed by the application
rpm and the radius available:
V2 for max  ideal hs per stage (approx.)
R = 1/2 reaction V2 = VB/cos hs = 2hs blade = 2V22/2N = VB2/(Ncos2

Rateau impulse V2 = 2VB/cos hs = V22/2N = 2VB2/(Ncos2

Curtis impulse V2 = 4VB/cos hs = V22/2N = 8VB2/(Ncos2
Here we have neglected the nozzle inlet velocity V1 as small compared to V2. Thus, a reaction
turbine needs twice as many stages as a Rateau turbine for the same total h, same VB, and
optimum efficiency. Further, it is desirable to use partial admission at the first stage of a steam
turbine so that just a few, sequenced high pressure throttle valves need to be used. In partial
admission, the nozzles extend only about 60˚ of the 360˚ around the first stage and this is only
practical with an impulse turbine that has no pressure drop across the moving blades. Steam
practice typically uses an all Rateau turbine or a Curtis first stage followed by Rateau or reaction
stages in the HP turbine. The Curtis first stage replaces 8 reaction stages and shortens the rotor.
Since all of the large pressure drop for the Curtis stage is taken in the nozzle, this also
significantly lowers the pressure and temperature for which the HP turbine casing needs to be
designed. Gas turbines have much lower total h and with annular combustors partial admission
is not an issue, so gas turbines typically use pure reaction stages.
In many marine applications, compactness is more important (with reasonable efficiency)
than achieving optimum efficiency. As a result, the Curtis stage is usually used in small auxiliary
turbines, such as main feed pump turbines. In astern turbines, a Curtis stage followed by a single
Rateau stage might be used.
Torque Versus RPM Characteristic of Turbines. The torque characteristic results that can be
derived for the simple impulse turbine are applicable to all turbines, except single shaft gas
turbines. At the optimum efficiency, VB/V2 = cos/2 and V3 is minimum ( = 90˚) so using the
development for the work from the impulse turbine we have,
7-9
Q = Rtr = mr(V2cos + V3cos) = mrV2cos


At VB = V2cos, the blade is moving as fast as the steam tangentially so there can be no torque
on the blade and B’ = 0. Finally, at stall, VB = 0, so that V2 = V2r ≈ V3r and with 2 ≈ 3 this
yields,
Q = Rtr = mr(V2rcos2 + V3rcos3) ≈ 2mrV2cos


if we assume that the turbine were actually designed for the stall condition so that  ≈ 2. This
crude result is very close for actual turbines (usually the stall torque is more like 1.9 than 2 times
the torque at maximum efficiency).
Expressing the torque and rpm (proportional to VB) normalized by their values at the rated
point (Q0, N0), which is usually the point of maximum efficiency, this yields the result shown in
Fig. 7.6 where  is the percent throttle (expressed as a decimal) and where we have used Rt 
m which implies that Q  . A plot of P/P0 would be quadratic in N/N0 similar to turbine. This
shows that power is very flat near N0 and, thus, turbines are often described as “constant power”
machines, which is reasonably true for modest deviations from N0.
Figure 7.6 Efficiency and Torque Characteristic of Ideal Turbines
Turbine State Lines. The stage efficiencies developed above include a loss due to the exiting
velocity V3 not being zero. In a multi-stage turbine this velocity is utilized by the following
stages so only the final stage will experience an exit loss or exhaust loss Ve2/2. With this in
mind, the turbine vendors carefully analyze the whole turbine stage by stage as above. What
marine engineers need, however, for multistage turbines is the overall result and also the state
line if extraction is to be used. The overall result corresponds roughly to the following:
7-10
HP and LP cross-compound propulsion turbines 80 ≤ turbine ≤ 85%

Single casing propulsion turbines (≤ 20000 hpBr) 78 ≤ turbine ≤ 83%
Turbogenerator set multistage turbines 65 ≤ turbine ≤ 75%
Small auxiliary turbines such as for feed pump drive 45 ≤ turbine ≤ 60%
The state line can be approximated by various methods. Marine Engineering [2] and
SNAME T&R Bulletin #3-11[3] include empirical “cookbook” schemes to estimate the
efficiency and state line based upon the principal parameters (inlet temperature and pressure,
exhaust pressure, power, etc.). As an example, the SNAME Bulletin #3-11 method for
propulsion turbines includes a way to establish the state line end point (SLEP) based upon the
theoretically available energy and principal parameters. The state line is then approximated as
a straight line on the Mollier chart from the top point (inlet h0, and 0.9P1) down to the SLEP. In
the region where the extractions occur, this model is acceptably close for use in preliminary
design. A real turbine usually has a fairly inefficient first stage with pressure drop to about 2/3
P1 (perhaps a Curtis stage) followed by a gradually curving state line to the SLEP. The exhaust
loss is added to the SLEP to get the used energy that drives the load, overcomes bearing losses,
and overcomes the astern turbine rotation loss. This model is depicted in Fig. 7.7 [2].
Figure 7.7 Model for Non-reheat Turbine State Line [2]
7-11
At partial powers, the state line shifts significantly to the right as the throttling and control
reduces the outlet pressure from the first stage and reduces its efficiency. The pressure at each
of the extraction points will also be reduced as a direct result. A typical partial power state line
is also depicted in Fig. 7.7.
Turbine Examples. We will close by looking at three example marine turbines. Figure 7.8
shows a typical HP steam turbine. Study of this figure shows that the turbine consists of a Curtis
first stage that allows the partial admission and significantly reduces the design pressure and
temperature for the casing since it takes a large pressure drop across its nozzle. Note that there
is a region following this first stage for the steam to spread out to the full 360˚ following the
partial admission. The Curtis stage is followed by 14 reaction stages. The characteristic drum-
blade construction of the reaction turbine can be seen. There are two extraction points provided;
after the Curtis stage and then after the first 6 reaction stages.
Figure 7.8 Example HP Steam Turbine – Curtis Stage plus 14 Reaction Stages [2]
7-12
Figure 7.9 shows a typical LP steam turbine used in a different application (vendor). This
turbine has nine pressure compounded Rateau impulse stages in the ahead LP portion. Note the
long final stage blades needed to accommodate the high specific volume of the steam at the
exhaust conditions. Also note that the cross-over from the HP outlet to the LP inlet is below the
turbine shaft line in this case. The characteristic disk-shaft construction of the impulse turbine
can be seen. The LP turbine casing also includes the astern turbine that consists of a Curtis stage
followed by a single Rateau stage. By placing the astern turbine at the low pressure end of the
LP turbine, the two turbines can share the same exhaust path to the main condenser. Note the
large size of the exhaust flange.
Figure 7.9 Example LP Steam Turbine – Nine Rateau Impulse Stages [2]
In the Mitsubishi Heavy Industries (MHI) Ultra Steam Turbine (UST) reheat plant, the HP
and the IP turbine are installed on the same shaft within a single casing with their flow paths in
opposite directions as shown in Fig. 7.10. Having opposite flow directions places the high
pressure/high temperature boiler and reheat outlet steam in the center rather than having to seal
these to the atmosphere at the ends of the combined turbine. A dummy ring is used to create the
7-13
shaft seal between these two inlet steam pressures. A thermal shield is used to protect the high
pressure casing from the highest steam temperatures.
IP turbine HP turbine
Figure 7.10 Ultra Steam Plant (UST) Combined HP Turbine (1 Curtis and 5 Rateau
Impulse Stages) and IP Turbine (6 Rateau Impulse Stages) [4]
Figure 7.11 shows the turbine end of the LM2500+ marine gas turbine. This turbine
consists of two parts since this is a two-shaft gas turbine. The turbine consists of eight reaction
stages. The first two reaction stages form the high pressure turbine (HPT) that drives the
compressor. The remaining six reaction stages form the free power turbine that drives the load.
The characteristic drum-blade construction of the reaction turbine can be seen. Diaphrams
extending inward from the outer shell contain the nozzles. The axial compressor portion of the
LM2500+ has a similar appearance.
7-14
Figure 7.11 Turbine End of LM2500+ Gas Turbine (2+6 Reaction Stages) [GE]
References
1. Saarlas, M., Steam and Gas Turbines for Marine Propulsion, Naval Institute Press,
Annapolis, MD, 1978.
3. “Marine Steam Power Plant Performance Practices,” SNAME T&R Bulletin #3-11,
4. Ito, M., Hiraoka, K., Matsumoto, S. and Tsumura, K., “Development of High Efficiency
Marine Propulsion Plant (Ultra Steam Turbine),” MHI Technical Review, Vol. 44, No. 3,
Sept. 2007.
7-15
7-16
Chapter 8 Electrical Systems
Electrical systems are provided onboard ship for ship service power and typically
maneuvering devices. Increasingly they are also used for the propulsion power of the vessel.
8.1 Ship Service Electrical System
Ship service power includes most propulsion auxiliaries, auxiliary systems, and all
domestic needs of the vessel including all communications and control. Vessels in North
America and the offshore industry typically use 60 Hz alternating current (AC) power, while 50
Hz is common in the rest of the world. Ship service systems here are most commonly 450 V 3
phase () today for generation and most large loads and 120 V single phase for lighting and hotel
loads with the voltage change accomplished by transformer. Loads that require uninterruptable
power, such as control computers and sensors, have special provisions such as back up with
batteries. See Woud and Stapersma [1], Marine Engineering [2, 3] or Patel [4] for a more
detailed discussion of the various voltages and details used on vessels.
Generators are typically rated in kVA (kilovolts*amps) and ekW power output, with AC
power output
ekW = power factor *V*A with P.F. = 0.80 typically used in initial design
Thus as an example, the Wärtsilä 8R22 ship service generator producing 60 Hz 3 power at 720
rpm has engine output 1080 kW, generator output 1300 kVA, and generator electrical output
1040 ekW. Thus, P.F. = 1040/1300 = 0.80 and gen = 1040/1080 = 0.963. The generators need
to be selected so that the normal at-sea load of the vessel can be supplied with one set down. A
small emergency generator is provided to assume the critical loads needed for ship maneuvering,
communications and navigation in the event of a loss of main power. On larger vessels, the
emergency generator needs to be located above the damage control deck (protection from
flooding) and outside of the engine room (protection from fire). Shaft generators are sometimes
used, but to provide constant frequency ship service power they require constant rpm shaft
operation necessitating a controllable pitch propeller or an output frequency converter. They are
typically driven from a Power-Take-Off (PTO) output on the reduction gear. Diesel generators
8-1
should not be operated for long periods of time below about 35% of rated load or high
maintenance can result.
The one-line diagram is a schematic of the switchboards, circuit breakers, load centers,
components and interconnection features that make up the ship service or any electrical power
system on a vessel. It shows the fundamental protections and redundancy provided. A typical ship
service power system one-line diagram is shown in Fig. 8.1. The main switchboard supplies 450 V
larger loads, a 230 V supply to refer containers, and 120 V for smaller loads and lighting.
Transformers are used to supply the lower voltages. The emergency switchboard is normally
supplied from the main switchboard, with an interlock that will start the emergency generator and
close its breaker on loss of power. Concentrations of loads are supplied from distributed load
centers rather than connecting all loads to the main switchboard. A shore power connection is
provided for use in port, if available. This is becoming more common with the move to go “cold
iron” while in port to minimize emissions. Larger and more complicated vessels have multiple
main switchboards with part of the generators supplying each board. The boards can be operated
cross-connected or “split plant” for reliability reasons. Critical loads often have feeds from either
main switchboard. Uninterruptable Power Supply (UPS) loads are supplied with battery backup.
Traditional Electric Load Analysis (ELA). The ELA is performed to establish the load
on the ship service electrical system in the various operating conditions or modes j of the vessel:
at sea, maneuvering, battle, loading, unloaded, locking/docking, in port, etc. Computer software
or spreadsheets can be used to evaluate the average load in mode j:
Lj =  fij i for i = 1….N
where i is the rated power of load i and fij is the load factor (LF), the fraction of the time load i
will operate in mode j times the ratio of its operating ekW in mode j to its rated ekW. It is
unrealistic and wasteful to just sum up the connected rated loads, i. The rated power is usually
used for i because in preliminary design you can usually obtain the typical rated size for
particular load in a similar type of plant, but you do not know the actual operating load in a
particular mode until later in design. The load factors are defined as,
fij = operating ekWi/rated ekWi * hours i is used per day in mode j/24
8-2
and rated ekWi = rated output kWi/m, with typical AC induction motor efficiency m ≈ 0.85 -
0.92 for smaller to larger kW motors. The load factors are usually recommended based upon
typical plants, but you should not use these blindly if your application is atypical for some reason.
For the largest loads, a detailed evaluation of the operating ekW in each mode j is recommended
rather than just using the general load factors. A good source for load factors for naval ships is
provided in Design Data Sheet 310-1, Rev. 1 [5].
Figure 8.1 Typical Containership Ship Service Power System One-Line Diagram [3]
The generators would be sized for the design load,
Ldesign = max {Lj} + margin for growth and non-average conditions
On military ships the margin might be 25-30%; on commercial ships there is usually enough
conservatism in the load factors that a specific margin can be omitted or at most 3-5%.
8-3
The traditional ELA is satisfactory if the number of loads N is large, but it can be
ineffective for smaller vessels. A truncated portion of ELA from the older edition of Marine
Engineering [3] is shown in Fig. 8.2. The newer edition [2] is misleading in that it shows all fij =
1.0; legend has it that the SNAME lawyers did not want to show real values for liability reasons!
In this example, ekW input = (rated hpBr output *0.7457 kW/hpBr)/(m = 90%) for the large main
condenser circulating pump.
Figure 8.2 Portion of Typical Commercial Vessel Electric Load Analysis [3]
Monte Carlo Electric Load Analysis. On smaller vessels with fewer loads, you need more
margin for non-average conditions and the traditional ELA can be misleading. Woodward and
Vibrams [6] recommended a rational basis for picking this margin using Monte Carlo simulation.
Consider first the traditional ELA applied (1) to a large ship with 100 loads of 10 ekW each,
where each operates 50% of the time; and (2) to a small craft with 2 loads of 5 ekW each, where
each operates 50% of the time (assuming here that the operating and rated kW are the same).
Using probability and assuming that all N loads are independent:
8-4
ship (N = 100) small craft (N = 2)

ave. L = 0.5 * 10 = 500 ekW ave. L = 0.5 * 5 = 5 ekW
a 20% overload 600 ekW would but both loads will be running
occur only 2.5% of the time, so ½ * ½ = ¼ or 25% of the time, so
a small margin or “slop” in the a 100% margin for non-average
fij will be adequate. conditions would be justified.
This situation can be addressed rationally by using Monte Carlo simulation. A random number
generator can be used to simulate a large number of random operating condition load samples Lk
for 1 ≤ k ≤ K ≈ 10,000. For each load i, (assuming only one mode j here) draw a random
number 0 ≤ rik ≤ 1 and set fik = 0 if rik > the fraction of time the load is running (off); if not set fik
= 1 (on). Sum the total load Lk = fiki and repeat for all k. After K samples, generate the
statistics for the load that will follow a Normal Probability Distribution for large N with,
mean load L = (1/K) Lk and variance or standard deviation 2 = (1/K) (Lk – L)2
Recall that for a Normal Distribution the load L will be within L ± , 68.3% of the time; within
L ± 2, 95.5% of the time; and within L ± 3, 99.7% of the time, giving a rational way to select
the margin for non-average conditions. A selection like 2 would be reasonable.
8.2 Electric Motor Torque versus RPM Characteristics
The primary shipboard motors can either be (1) synchronous or asynchronous induction
AC motors or (2) shunt or series wound DC motors. This brief section is taken mostly from
Patel [4]. Electrical motors have an armature coil that creates a rotating magnetic field and a
field coil that interacts. They are composed of a stator which surrounds a rotor that produces the
output torque at some speed. AC motors have the armature coil in the stator and the field coil in
the rotor. DC motors have the reverse; the armature coil is in the rotor and the field coil is in the
stator. The torque versus rpm characteristics of these motors have implications for their
selection and matching with their loads.
AC induction motors are the most common shipboard motors. They are simple,
brushless, low-cost and rugged. At all significant powers 3-phase AC power is used. This feeds
belts of coils spaced around the stator. The number of poles in this winding determines the
synchronous speed that the resulting magnetic field will rotate relative to the input frequency,
synchronous rpm = Ns = 120 x frequency / number of poles

8-5
Thus, a four pole stator winding will produce a 120 x 60 / 4 = 1800 rpm synchronous speed with
a 60 Hz supply. The rotor consists of axial conductor bars with two end rings that short out all
bars (a squirrel cage configuration). Induction motors produce torque when there is a slip speed
= Ns – Nr producing a slip s = (Ns – Nr)/Ns, where Nr is the rotor speed. This slip induces a
current in the rotor by Faradays Law and its magnetic field then follows the rotating stator field.
The details of the motor design can produce a range of motor torque versus rpm
characteristics as shown in Fig. 8.3 for the various National Electrical Manufacturers Association
(NEMA) rated induction motors for low power applications. Note that the slip is the fraction or
percent of speed loss below synchronous speed. The NEMA B design with high maximum
(pullout) torque, low startup torque and low slip and thus high efficiency at rated torque is used
for general purpose use in fans, blowers and centrifugal pumps. The NEMA D design has high
(300% rated) startup torque and is used in cranes, hoists, winches and capstans. Propulsion
induction motors would emphasize the high efficiency similar to a NEMA A design. The
operating slip is essentially linear with the operating torque near the rated point so there is a
modest speed droop with load torque as shown in Fig. 8.4. Note that this plot is speed versus
load torque instead of the more common load torque versus speed as seen in Fig. 8.3.
Figure 8.3 Torque-RPM Characteristics of Various NEMA Induction Motors [4]

8-6
Figure 8.4 Speed Droop Characteristics of Electrical Motors [4]
AC synchronous motors have the highest efficiency and are typically only used for ship
propulsion. These motor feed DC to the rotor through slip rings and the resulting magnetic field
follows the rotating magnetic field of the stator at a torque angle with motors designed to
operate with 25˚ <  < 35˚ for transient stability since the motor will become unstable when
˚. It can only produce torque at close to synchronous speed so it is typically started as an
induction motor until synchronization is approached and then the rotor excitation is applied. The
slip rings add cost and maintenance. Motor efficiencies at high as 98% can be reached. The
synchronous motors can produce any load at its synchronous speed up to its rating as shown in
Fig. 8.4.
DC motors use a commutator and brushes to act as a mechanical switch to accomplish the
required current reversal in the rotor. This produces a varying magnetic field that responds to the
fixed stator field to achieve rotation and torque. The commutator adds maintenance and limits
the size (kW) and speed of DC motors. In the DC shunt type motors the stator field coils and
the rotor are wired in parallel electrically. These motors produce essentially constant RPM for
any constant input voltage. This gives good speed control and higher efficiency. They are thus
used for ship propulsion and general purpose use such as centrifugal pumps, fans, and blowers.
Their speed – torque characteristics are similar to induction AC motors as shown in Fig. 8.4. DC
series wound motors arrange the stator field coils and the rotor in series. These motors have
8-7
high torque at startup and are, therefore, used for cranes, hoists, capstans, winches, etc. They
also have fast speed at low load as shown in Fig. 8.4.
8.3 Integrated Electric Plants
Integrated Electric Plants (IEP), the all-electric ship, utilize electrical propulsion motors
and central station power generation that powers all propulsion, thruster, and mission related
equipment as well as general ship service needs. Integrated Electric Plants have become the
plant of choice in many recent naval vessels, cruise ships, high technology cargo vessels and
special purpose vessels, such as offshore supply and service vessels and icebreakers. This section
draws heavily on Parsons et al. [7].
George Stewart noted at the recent ASNE/SNAME Electric Ship Design Symposium that
“electric drive should be considered for any application where the ship spends much of its time at
loitering speeds, when high ship service power requirements exist, or where special mission
power requirements dictate its use” and “all modern electric drive installations are of the
integrated type with the same generators used to provide both propulsion and ship service
power” [8]. A typical system one-line diagram of an AC/AC Integrated Electric Plant (IEP)
from that presentation is shown in Fig. 8.5. The main switchboards supply the propulsion motors
through transformers and then AC-AC frequency converters for propeller speed control. For
reliability, a single screw design can utilize either two separate tandem motors on the main shaft
or two higher rpm, thus smaller and lighter, motors geared together through a single-reduction
combining gear as shown. A single direct-connected propulsion motor is also an option.
An interesting recent design example is the design of the ORCA class RO/RO vessels (MV
Midnight Sun and MV Northern Lights) for TOTE’s Tacoma to Anchorage service [9]. The
ORCA class ships are characterized by a relatively short sea time versus port time and a varying
speed profile due to weather, tides and a fixed schedule. In the initial study, an IEP using five
identical diesel generators was found to be the best choice. The final ORCA design used four
larger generators and two smaller generators so that the generators in use at any time could more
closely match the speed-power profile of the vessel, as shown in Fig. 8.6. With this design, most
of the operating generators are near 90% of its rating, where its specific fuel rate is optimum,
regardless of propeller loading.
8-8
Figure 8.5 One-Line Diagram for Typical Integrated Electric Plant [8]
Figure 8.6 Comparison of Power Available with Five Generators or Four Main plus Two
Smaller Generators with Ship Propulsion Requirement [9]
8-9
Kanerva and Nurmi noted [10] that in order to achieve all of the available IEP benefits
“the complete configuration of the vessel must be considered” which should include “machinery
location, machinery modularization, cargo tank location, cargo volume capacity, loading
capabilities, hull form, power requirement, structural principles, machinery and operational
availability, operation under special service conditions (restricted waterways, ice, etc.),
maintenance, building procedure, installation and timing.”
At first glance an IEP must overcome a significant propulsion efficiency disadvantage
relative to a traditional mechanical geared system. The required propulsion brake power PB is
related to the effective power P E (defined in Chapter 11), which reflects the hydrodynamic
resistance and speed of the vessel, by the following:
PB = PE/(H R O T S B)
where H is the hull efficiency, R is the relative rotative efficiency, O is the open water
propeller efficiency, T is the prime mover to shaft transmission efficiency and SB is the shaft
bearing and stern tube efficiencies. A vessel with mechanical drive and a single-stage reduction
gear would have efficiency T = G ≈ 0.985. In a conventional IEP, the transmission is electrical
with the prime mover driving a generator (Gen ≈ 0.96) with the output passing through a
switchboard (Sw ≈ 1.00), propulsion transformer (Tr ≈ 0.99), AC to AC frequency converter
(Fc ≈ 0.98) to an electric motor (M ≈ 0.96) that might drive the shaft through a propulsion
reduction gear (G ≈ 0.985), giving T =Gen Sw Tr Fc M G. If the electric motor turns at
propeller speed, the reduction gear can be eliminated. A compensating factor is that with shaft
reversal possible electrically, there is no need for a Controllable-Reversible Pitch (CRP) propeller
as is commonly used with today’s highly turbocharged reciprocating engines. The change in the
propeller efficiency P = RO from a CRP to a lighter, simpler, more efficient Fixed-Pitch
Propeller (FPP) will be from about 0.65 to 0.66 or 1.5% for example. Using all these
efficiencies, the IEP would have an initial disadvantage of about,
(ROG)Mech./(ROGenSwTrFcM)IEP ≈ 1.085
without gearing and ≈ 1.102 if a reduction gear is required. This apparent fuel use disadvantage
must be offset by system design and the operation of all generating sets close to their optimum
8-10
efficiency in all operating modes. The effect of power level on the specific fuel rate from a
typical diesel generator operating at constant rpm is shown in Fig. 8.7.
Figure 8.7 Change in Diesel Generator Specific Fuel Rate with Load at Constant RPM
A study of using an IEP in future Great Lakes ore carriers by Parsons et al. [7] showed that
with proper plant design, these vessels could overcome this apparent disadvantage due to the
various modes of ship operation as shown in Table 8.1. This provides a small fuel advantage
with additional advantages of fewer total prime movers onboard, fewer cylinders to maintain,
similarly improved air emissions, etc. for the vessel.
Table 8.1 Operating Modes for Duluth to Lower Lake Michigan Round-trip [7]
The resulting one-line diagram for a twin 5400 kW shaft 1000 Great Lakes self-unloader is
shown in Fig. 8.8. The five generators (four diesel sets at 2810 ekW; one dual-fuel LNG/diesel
set at 1014 ekW) feed the 4160V main bus, which is in sections for reliability. Any generator
can supply any load. There are 30˚ phase shift propulsion transformers to supply 12 pulse Pulse
Width Modulated (PMW) frequency converters [1] that supply variable voltage, variable
8-11
frequency power to the two direct connected 0-120 rpm induction propulsion motors that drive
fixed-pitch propellers. The large self-unloading conveyor and lateral thruster loads are also
supplied from the main 4160V switchboard. Two 4160V/450V transformers supply the two
sections of the main ship service switchboard, which supplies the ballast pumps and other ship
service loads. The emergency generator and emergency switchboard are typical. At this low
total power, a 450V system could be proposed by a vendor instead of a 4160V system.
Figure 8.8 One-line Diagram for 1000 Great Lakes Self-Unloader [7]
The machinery arrangement follows the observation [10] that a vertical arrangement is
usually optimum with a power flow down from the generators to the main switchboards, to the
8-12
power transformer/frequency conversion, to the motors for short main power cabling. The
inboard profile of the aft section of this vessel is shown in Fig. 8.9.
Figure 8.9 Inboard Profile of Aft Portion of IEP 1000 Great Lakes Self-Unloader [7]
Solid State Frequency Conversion. The primary technical development that has enabled
the significant move to Integrated Electric Plants (IEPs) in many recent larger ships has been the
development of solid state power electronics, particularly for high power alternating current
(AC) frequency conversion. This section is taken mostly from Parsons [11].
The goal in an IEP is to generate AC power at constant voltage and frequency so that the
power source(s) can be used for all the ship needs. To use part of this power for ship propulsion,
it is then necessary to convert the AC frequency efficiently while transmitting high propulsion-
level power. This eliminates the need for an expensive and less efficient Controllable Reversible
Pitch (CRP) propeller; the shaft speed is controlled by changing the frequency supplied to the
AC propulsion motor(s). Today this conversion can be done using solid state frequency
conversion devices with an efficiency of about 98%. See Woud and Stapersma for excellent
additional detail [1].
There are three types of AC frequency converters that the various vendors have used in
marine AC generator-AC motor propulsion systems: Pulse Width Modulation (PWM)
converters, Synchroconverters and Cycloconverters. The Pulse Width Modulation converter and
8-13
the Synchroconverter consist of an initial rectifier block that converts the constant frequency AC
power to DC power, a DC link and then a DC to AC inverter block that produces variable
frequency AC as its output. Therefore, these are classified as AC-DC-AC devices.
Cycloconverters convert the constant frequency AC directly to variable frequency AC so they
are classified as AC-AC devices. PWM converters can be used to control either synchronous or
asynchronous induction propulsion motors. They can be used for ahead (first quadrant, ship
speed V > 0, shaft rotation speed n > 0) and astern (third quadrant, V < 0, n < 0) operation, but
they cannot be used for regenerative braking in the other two quadrants. PWM converters are
usually used with induction motors since they are simpler and cheaper to buy and maintain.
Synchroconverters can be used only with synchronous AC propulsion motors, which can be used
in all four quadrants for propulsion (acting as a motor) or braking (acting as a generator).
Cycloconverters can also be used with either synchronous or asynchronous induction AC
propulsion motors and in all four quadrants.
The highest level schematic of a Pulse Width Modulation converter, which are the most
common today, is shown in Fig. 8.10. Its AC-DC rectification is performed by an uncontrolled
rectifier than consists of two diodes connected to each phase of the 3 phase AC power. The
diode is essentially a check valve that only passes current when it flows in one direction. One
diode of each pair passes the positive part of the sinusoidal AC current; the other has reverse
polarity so it passes the negative part of the current. This produces two positive lobes of current
in sequence. When all three phases, with their 120˚ phase shifts, are passed through three pairs
of diodes the total output has six lobes that can create an acceptably smooth “constant” voltage
DC current. If a smoother DC output is needed each phase can also be passed through a 30˚
phase shift transformer and this can go through six more diodes to produce a smoother final DC
consisting of 12 lobes. The DC-AC inverter part of the PWM converter uses either transistors or
Gate Turn-Off (GTO) thyristors. These act as controllable switches. The transistor acts like a
gate valve; it will transmit whenever a small (base current) control signal is on. An ordinary
thyrister acts like a gate valve and a check valve; it will transmit only a positive current when a
small (gate on) control pulse has been received. It automatically turns off when the current sign
reverses. A GTO thyrister has an additional control signal which allows it to be turned on and
off; it transmits whenever the current is positive and its gate-on pulse has been given. It turns off
whenever the current sign reverses or the gate-off pulse has been received. The PWM uses
8-14
either transistors or GTO thyristors to control the timing and width of output power pulses to
create a variable voltage and variable frequency AC output to control the attached motor.
Figure 8.10 Pulse Width Modulation (PWM) Frequency Controller Schematic [11]
The highest level schematic of a Synchroconverter is shown in Fig. 8.11. Its AC-DC
rectification is performed by a controlled rectifier that consists of two thyristers connected to
each phase of the power. By controlling the gate-on control pulse timing, the resulting six-lobe
DC power in the DC link has variable voltage. The DC-AC inverter part also uses thyristors
with their gate-on pulse linked to the rotation speed of the motor so that the inverter follows the
speed of the motor. The speed of the motor is then controlled by changing the voltage of the DC
link. The combination of the inverter and the connected synchronous AC motor effectively
behaves like a brush-less DC motor controlled by changing the DC voltage.
Figure 8.11 Synchroconverter Schematic [11]
The highest level schematic of a Cycloconverter is shown in Fig. 8.12. The Cycloconverter
controls the output AC frequency over the range of 0 to about 35% of the input AC frequency.
Thus, with 60 Hz input the maximum output frequency is only about 20 Hz, which is compatible
with using high motor pole numbers for large slow turning motors (20 Hz  20 poles  120
rpm) suitable for ship propulsion without the need for gearing. Cycloconverters use six ordinary
thyristers for each half of the converter supplying each of the three motor windings; so a total of
3x12 = 36 thyristers are used. The speed is controlled by changing the output frequency by
8-15
controlling the thyrister gate-on timing. The output voltage is constructed by switching the
motor winding to the phase of the input that has the correct voltage at any particular time. Three
isolated converter sections supply the three sets of motor windings.
Figure 8.12 Three Phase Cycloconverter Schematic [11]
The development of these efficient, high-power AC frequency conversion devices has

enabled the development of AC-AC Integrated Electric Plants for large ships with overall
transmission efficiencies of about 90% from the prime mover to the propeller. These can now
compete successfully with geared diesel plants and provide additional operational, economic and
emissions advantages.
Wärtsilä Low Loss Concept. A one-line diagram for a conventional IEP on a Platform
Supply Vessel (PSV) with two 5000 kW azimuthing propulsion thrusters, two bow tunnel
thrusters, two stern tunnel thrusters and an azimuthing bow thruster is shown in Fig. 8.13. Note
the use of a transformer for each of the major loads. These are used to create a 30˚ phase shift to
provide 12 lobe DC within the frequency converter speed controllers for these large motors.
These are large heavy components typically located in their own room or in the engine room near
the motors. The Low Loss Concept (LLC) patented and offered by Wärtsilä makes an interesting
change in this situation. The main buses are doubled within the switchgear and a single phase
shift transformer is use to supply the second part of each section. This replaces all of the
transformers used for each individual load, by providing a second bus supply with the needed
phase shift. A one-line diagram for a LLC design for a PSV is shown in Fig. 8.14. The phase
shift transformers can also be used as the step down transformers to the low voltage parts of the
system. The savings in transformer losses, component cost, weight and space are significant.
8-16
Figure 8.13 One Line Diagram for Traditional IEP for Platform Supply Vessel [Wärtsilä]
Figure 8.14 One Line-Diagram for Wärtsilä Low Loss Concept for PSV [Wärtsilä]
8-17
ABB DC Link Concept. Another recent innovation is the introduction of the ABB (ASEA
Brown Bovari) DC link concept, which brings to commercial application an approach used by
the U.S. Navy. This innovation introduced in 2011 was honored at the 2013 Offshore
Technology Conference as one of the outstanding technical innovations in the offshore industry.
Since most IEP applications today use Pulse Width Modulation (PWM) frequency
converters, as shown schematically in Fig. 8.10, for the frequency control of all AC motors, it is
reasonable to introduce a constant voltage DC link by merging all of the DC links within these
converters. The AC-DC rectifier part of the controllers (r ≈ 0.99) are moved to the generator
output. The DC-AC inverter part of the controllers (i ≈ 0.99) remain at the input to the loads.
This concept is illustrated in Fig. 8.15 [www.abb]. The complex AC switchboard and power
transformers (swtr ≈ 0.98 – 0.99) can be eliminated and the generators are then free to operate
at their optimum specific fuel rate rpm, as illustrated for a medium-speed generator in Fig. 8.16,
which can provide substantial fuel, emissions, and noise improvements.
Figure 8.15 Comparison of AC and DC Distribution Concepts [ABB]
8-18
Figure 8.16 Typical Diesel Generator SFR at Constant Frequency and Optimum RPM [12]
This concept offers many benefits [12]. The challenge is to provide equivalent electrical
safety with a DC power distribution system. The system concept developed provides equivalent
safety and utilizes a 1000 VDC distribution that is suitable for total powers up to 20 MW. Fault
protection is provided by a combination of fuses, isolator switches and controller turn-off
capability of the semiconductor power conversion devices. “Any fault can be cleared within a
maximum of 40 ms … a drastic reduction in fault energy levels compared to traditional AC
generation circuits where fault durations can reach up to 1 s” [12]. Further the complex controls
for synchronization of parallel AC generators is eliminated since all generators are producing DC
power through their rectifiers. The DC link is also directly compatible with the introduction of
DC devices such as super capacitors, batteries, fuel cells and solar cells. Shore power of 60 Hz
vessels in ports using 50 Hz power will also be facilitated. Batteries and super capacitors can be
needed to improve system dynamic response, particularly in heavy weather dynamic positioning
where thruster demands often exceed the speed at which the generators can be varied. Since the
generators can be operated at their optimum variable rpm, instead of constant rpm, fuel savings
of as much as 20% are possible. Improved system efficiency and a reduction in space, weight
and cost is possible by the elimination of the switchboards and power transformers. A reduction
in equipment footprint (30%) and generator noise have resulted.
8-19
The first vessel with an ABB Onboard DC Link was the 93 m multipurpose oil field supply
and construction vessel Dina Star complete in Norway in April, 2013. This vessel has four Cat
3516 2,300 ekW and one Cat C32 920 ekW variable speed generators supplying two 2,200 kW
main propulsion units and three additional DP thrusters. Specific fuel rate savings up to 27%
have been achieved by running the generators at variable instead of constant rpm.
8.4 Hybrid Electric Plants
Hybrid electric plants are beginning to appear in vessels with long periods of station
keeping, port time and loitering. We will look at recent examples.
Twenty Anchor Handling, Tug, Supply (AHTS) vessels were built recently in Norway with
hybrid diesel and electric plants [13]. The one-line diagram for this 690 V, 60 Hz LLC system is
shown in Fig. 8.17.
Figure 8.17 One-Line Diagram for Hybrid Anchor Handling, Tug, Supply (AHTS) Vessel [13]
8-20
The operating mode mission profile for these vessels is steaming at 12 knots 40% of the
time, steaming at 15 knots 10%, dynamic positioning alongside platforms 40% and moving
offshore platform anchors and towing using maximum power in the bollard pull mode only 10%
of the time. The vessels are twin screw with each CRP Kort nozzle propeller powered by either
or both an 8000 kW diesel engine and a 2400 kW electric motor through a Power-Take-In (PTI)
combining gear. The engines also power a Power-Take-Off (PTO) 3000 ekW generator on their
front end. Clutches are provided to both diesel loads. Two 2180 ekW auxiliary generators
supply either ship service power or propulsion power.
Batteries for Hybrid Plants. There has been major advances in batteries and battery
management in recent years. The most advanced available today for ship level powers appear to
be the Lithium Nickel Manganese Cobalt (NMC) batteries offered by Corvus Energy of
Richmond, BC. The characteristics of the principal battery types are compared in Table 8.2 [14].
Table 8.2 Comparison of Primary Battery Types [14]
The Lithum NMC batteries have 88% more energy and can produce 76% more power in
power applications than lead-acid batteries. They have a 99% charge/discharge efficiency
(compared to 60% for the lead-acid battery). They can operate (hold output voltage) all the way
to the end of the battery charge. They are completely sealed and emit no vapors or by products
that are explosive or flammable. Corvus’ AT6500 modules can hold 6.5 kWh and have Lloyd’s
8-21
Register Type Approval. They can be assembled in packs of up to 1,000 V. They are
maintenance free and have a working life of 10 to 20 years. They can be safely discharged in as
fast as 6 min. or safely charged in as fast as 30 min. They are 99% recyclable. These batteries
are used on the Boeing 787 Dreamliner.
Perhaps the first advanced diesel-battery hybrid was installed in the converted 78 x 34
5,100 hpBr Foss Maritime Co. ship assist tug Campbell Foss now operating in Los Angeles and
Long Beach [15]. This type of tug needs full power only 7 % of the time and has extensive
periods of idle or station keeping. This conversion followed their new construction of the hybrid
76 x 34 tug Carolyn Dorothy that was built with 126 lead-acid batteries. The Campbell Foss
plant replaced these with 10 Corvus Lithium NMC battery packs. She has two azimuthing stern
drives each powered by a 1902 kW diesel or a 500 ekW AC motor/generator. The batteries are
tied to the switchboard to provide an alternative source of energy. In all hybrid modes, the
motor/generators operate as motors on electricity from the batteries and the diesels are de-
clutched. In the ship assist mode, the main diesels provide the energy and the motor/generators
operate as generators to provide ship power and charge the batteries. The batteries can be
completely charged in all modes and by shore power in 30 minutes. The tug has two auxiliary
diesels: a 125 ekW generator and a 350 ekW generator. Neither operates when the main diesels
are used. Either can charge the batteries. The tug appears to have five operating modes:
Idle/stop 125 ekW generator or batteries

Transit one (up to 7 knots) 350 ekW generator
Transit two (up to 9 knots) 125 and 350 ekW generators
Battery transit batteries only
Ship assist main diesels
A comparison of the hybrid tug with a conventional version, showed it saves 20% on fuel cost
and produced 73% less PM, 51% less NOx and 27% less CO2. The main diesels operate only half
the hours compared to the conventional installations greatly extending the engine maintenance
cycle. The installation cost was about $2M more than a conventional version, but Foss received
$1M from the Ports of Long Beach and Los Angeles for the emissions reduction.
This is a heavily Canadian product: designer Robert Allen, Ltd., Vancouver, BC; batteries
Corvus Energy, Richmond, BC; hybrid design and control Aspin Kemp & Assoc., Owen Sound,
ON, and XeroPoint Energy, Dartmouth, NS.
8-22
The number of hybrid vessels is growing. With EU support, Scandlines is converting four
of its Germany to Denmark RO-RO ferries (MF Prinsesse Benedikte, MV Prins Richard, MV
Deutschland, and MV Schleswig-Holstein) to hybrid systems. The later three use 299 Corvus
AT6500 lithium polymer batteries to provide 2.7 MWh of power for these systems. Damen has
delivered the Bernardus, a hybrid version of their ASD 2810 tug. This vessel can run on
batteries in-port and at speeds up to 5 knots on 100% battery power. It has achieved savings of
up to 30% on fuel and up to 40% on emissions. Four more are under construction (2014). The
Scottish ferry operator Caledonian Maritime Assets has acquired two small (43.5 m, 150 PAX,
23 car, 2 truck) hybrid ferries to operate to its western islands. The MV Hallaig has experienced
fuel savings of up 38%. Caterpillar in cooperation with Kotug and Aspin Kemp & Associates
have delivered the hybrid tug RT Adriaan which has experienced a 25% fuel savings, 73%
reduction in PM, 50% reduction in NOx and 26% reduction in CO2. Two more are under
construction at both Damen and Choy Lee in Hong Kong.
8.5 Fuel Cells
Fuel cells combine hydrogen and oxygen electrochemically without combustion and
produced DC electricity [16]. They were invented in 1839 by Sir William Grove in England.
They were first used in the Apollo and Gemini space flights in the 1960’s to produce DC power
and drinking water. They are now in use at sea for Air Independent Propulsion (AIP) on German
submarines, producing up to 2 weeks at 5 knots submerged.
A fuel cell consists of two electrodes with an electrolyte in the middle. The anode reaction
is H2  2H+ + 2e- and the cathode reaction is O2 + 4H+ + 4e-  2H2O. They have the
advantages of potentially high efficiency reaching 65% (the best diesel generators approach
50%), lower emissions, shock hardness, and modularity. The cell itself is silent with no moving
parts, but that is not true for the entire plant. The disadvantages today are the volume of the
whole system (fuel source, processing, reforming to H2, and fuel cells), poor transient response,
difficult control, and cost.
There are many types of fuel cells based on the type of electrolyte used. The primary
types are:
 Alkaline Fuel Cell (AKC) - used in NASA missions; low power; efficiency to 70%; high
cost.
8-23
 Proton Exchange Membrane (PEM) - used in automotive applications and submarines;

low temperature 90˚C; light weight polymer construction; rapid load change possible.
 Phosphoric Acid Fuel Cell (PAFC) - used in premier power applications (hospitals and
air traffic control); up to 200 kW units; modest temperatures 210˚C; modest efficiency to
40%.
 Solid Oxide Fuel Cell (SOFC) - used for high power applications; very high temperatures
980˚C; planar or tubular ceramic electrolyte; slow startup and transient response issues;
system efficiencies to 60%; use of hot gases to run a gas turbine can raise the overall
plant efficiency even further.
 Molten Carbonate Fuel Cell (MCFC) - high temperatures 650˚C; efficiencies to 40%.
The PEM, PAFC, and SOFC cells can be used with fuel reformers that take natural gas, methane,
ethanol, petroleum distillates, etc. and produce H2, water and impurities. Diesel is actually one
of the best, most efficient storage media for H2 in terms of H2/m3 and H2/kg. Some types of fuel
cells such as the Direct Methanol Fuel Cell (DMFC) can use non-H2 fuels directly without
reforming. The MCFC internally reforms H2 rich gases or diesel to H2. Table 8.3 gives a
comparison of the entire plant required to produce electricity using the principal fuel cell types
that are candidates for shipboard use.
Table 8.3 Comparison of Electrical Power System Characteristics [16, 17]
The U.S. Coast Guard studied the feasibility of repowering the cutter Vindicator by
removing its four diesel generator sets and their sound isolation bedplate and replacing them with
commercially available MCFC fuel cells and the balance of the plant required to provide at least
the same power level as installed (approx. 2.5MW) [17]. The block diagram of this plant is
shown in Fig. 8.18. Note that the plant is much more than just the fuel cell stack. The concept
uses liquid NATO F-76 diesel as its input. The configuration consisted of 12 modules each
8-24
composed of 216 individual MCFC with the balance of the plant shown in Fig. 8.18 to produce
215 kW at 162 V DC. Each complete module was 3.5 x 8 x 7high and weighed 6.65 LT. To
obtain the required voltage and power for propulsion, 5 modules were used in series to produce
810 V DC needed to drive the existing 750 V DC propulsion motors and then two of these
groups were used in parallel to give the 2,150 kW needed for propulsion power. The two
remaining modules were then used in series to produce 430 kW at 324 V DC that was inverted to
220 V AC for ship service power. The concept would fit in the existing engine room and with
removal of the stack save 14.7 LT with an efficiency of 54% and extended range.
Figure 8.18 MCFC (DFC) Marine Power Plant Block Diagram [16, 18]
References
2. Schmid, W. E., “Electrical Systems,” Ch. XIX in Harrington, R. L. (ed.), Marine
Engineering, SNAME, Jersey City, NY, 1992.
8-25
3. Burr, M., “Electric Plants,” Ch. XVII in Harrington, R. L. (ed.), Marine Engineering,
SNAME, New York, NY, 1971.
4. Patel, M. R., Shipboard Electrical Power Systems, CRC Press, Boca Raton, FL, 2012.
5. Naval Sea Systems Command, “Electrical Power Load Analyses (EPLA) for Surface Ships,”
DDS 310-1, Rev. 1, 17 Sept. 2012. See www.
6. Woodward, J. B and Vibrams, F., “Electric Power for Small Commercial Vessels,”
Proceedings of SNAME Spring Meeting, 1973.
7. Parsons, M. G., Singer, D. J. and Denomy, S., “Integrated Electric Plants in Future Great
Lakes Self-Unloaders,” Journal of Ship Production and Design, Vol. 27, No. 4,
November, 2011, pp. 169-185.
8. Stewart, G., “Considerations in the Development of Concept Designs for Integrated Electric
Power Plants,” ASNE/SNAME Electric Ship Symposium, National Harbor, MD, Feb. 12-
13, 2009.
9. Boylston, J. and Brooks, M., “The Selection of the Power Plant for an Alaskan RO/RO
Ship,” Marine Technology, Vol. 38, No. 3, July 2001.
10. Kanerva, M. and Nurmi, J., “Innovative Diesel-Electric Powered Tanker Experience and
Future,” Seatrade Tanker Industry Convention, London, UK, February 11-12, 1998.
11. Parsons, M. G., “How Do They Do That? High power alternating current frequency
conversion,” Great Lakes/Seaway Review, Vol. 39, No. 2, Oct.-Dec. 2010, pp. 51-52.
12. Hansen, J. F., Lindtjorn, J. O. and Vanska, K., “Onboard DC Grid for Enhanced DP
Operation in Ships,” MTS Dynamic Positioning Conference, October 11-12, 2011.
13. Sortland, S., “Hybrid Propulsion System for Anchor Handling Tug Supply Vessels,”
Wärtsilä Technical Journal, 01.2008, pp 45-48.
14. Perry, B. and Simmonds, N., “Lithium Ion Energy Systems for the Next Generation of
Vessels: Power for a Green Future,” Corvus Energy, Richmond, BC.
15. Tyler, D. A., “Foss’s Second Hybrid Tugboat Employs New, More Powerful Lithium
Polymer Batteries,” Professional Mariner, September 2012.
16. Allen, S., Ashey, E., Gore, D. Woerner, J. and Cervi, M., “Marine Application of Fuel Cells:
A Multi-Agency Research Program,” Naval Engineers Journal, Vol. 110, No. 1, Jan.
1998.
17. Goubault, P., Greenberg, M., Heidenreich, T. and Woerner, J., “Fuel Cell Plants for Surface
Fleet Applications,” Naval Engineers Journal, May, 1994.
18. Kumm, W. H., “Feasibility Study of Repowering the USCGC VINDICATOR (WMEC-3)
with Modular Diesel Fueled Direct Fuel Cells,” Final Report, U.S. Coast Guard Research
and Development Center, 1997.
8-26
Chapter 9 Shafting and Components
The transmission/shafting system between the prime mover and the propulsor may be
purely mechanical or electrical/mechanical. The functions and objectives of the transmission
system are the following in a conventional propeller drive application:
1. transmit torque Q from the prime mover to the propeller,

2. support the propeller,
3. transmit thrust T from the propeller to the hull,
4. withstand transient loads; e.g. maneuvering, stopping, etc.,
5. be free from unacceptable vibrations
6. be very reliable,
7. allow the various operating modes required by the application,
8. be acceptably efficient.
Consult the sections in Woud and Stapersma [1] and Long’s chapter in Marine Engineering [2]
concentrating on the nomenclature, arrangement, and components of the shafting systems.
Figure 9.1 shows the arrangement of a typical single screw shafting system with mechanical
reduction gearing and a centerline stern tube. Figure 9.2 shows the arrangement of a typical
multiple screw shafting system with external water lubricated strut bearings.
Figure 9.1 Typical Single Screw Shafting System [2]
Figure 9.2 Typical Multiple Screw Shafting System [2]
9-1
The Shafting Design Problem. As with most design problems, shafting design involves a series
of iterations from conceptual through preliminary to final design. This could be outlined as
follows:
1. state design performance requirements number of shafts, type of service, etc.
2. establish design criteria ABS, Lloyds, Navy standards, etc.
3. develop a tentative arrangement prime mover location, propeller location,
bearing locations, etc. considering space
restrictions, rake angle limitations, clearances,
overhaul and maintenance access, etc.
4. establish sizing by criteria diameter(s), lengths of sections, flange
sizes, etc.
The diameter is usually established for the steady state (time average) loads with
allowance for the added dynamic loads in the safety margin in the coefficients in the
formulae. Consider the primary loads:
Torsion is the primary loading that sets shaft diameter. The shafts are typically mild
steel (tensile 60,000 psi or 415 N/mm2, yield 30,000 psi) except for inboard shafts on
naval and high powered commercial vessels. The shafting is usually sized for
strength and only rarely modified later based upon the vibration analyses. The
transmitted torque is,
Q = PB 60/(2N) kNm PB [kW], N [rpm]
Maneuvers may increase this steady torque 10-40%. There will also be a 2-15%
variable torque about the mean due to the unsteady wake field in which the propeller
operates. Typically,
single screw 0.06-0.15 more with fewer blades and even number of blades
twin screw 0.02-0.08 more with fewer blades, bossings, and skegs
Thrust is typically a small loading in terms of stress. For a single propeller case,
T = Rt/(1 – t) kN Rt, total resistance [kN], t thrust deduction
9-2
Submarines also have a thrust due to the hydrostatic loading on the end of the shaft,
Ts = whDs2/4 Ds seal diameter [m], h test depth [m]

w water weight density [t/m3]
There will also be a variable thrust about the mean due to the unsteady wake field in
which the propeller operates. Typically,
single screw 0.03-0.12 more with fewer blades and even number of blades
twin screw 0.02-0.13 more with fewer blades, bossings, and skegs
Thrust only sets the shaft diameter when the thrust bearing is forward of the gearing.
Bending due to gravity, shock, lateral (bending) vibrations, and the off-center thrust
(thrust may have an eccentricity of 5-8% Dp). Between uniformly spaced bearings
and with a cantilevered propeller, respectively, the moments due to gravity would be,
Mb = wL2/12 w weight per unit length, L span length (pinned-

pinned assumption)
and Mb = WpLp Wp propeller weight, Lp distance from assembly

center to effective support point in stern bearing
In water lubricated stern bearings, the length-diameter ratio L/D ~ 4.0 and the
effective support point is usually taken as 1 D forward of the aft end of the bearing.
In oil lubricated stern bearings, 1.5 ≤ L/D ≤ 2.0 so the effective support point is
typically taken as 0.5 D forward of the aft end of the bearing.
With combined torsion and bending,
Max. shear stress  = 16(Mt2 + Mb2)1/2/D3 Mt = Q torsional moment

Mb bending moment
or solving for the diameter of a solid round shaft,
D = [16(Mt2 + Mb2)1/2/allowable]1/3
The American Bureau of Shipping (ABS) classification rules do not specifically limit
the stress due to bending Mb [3]. The Navy design criteria limit the alternating
bending stress to 6000 psi [4]. The ABS rules are based upon past experience and
must, therefore, be critically evaluated for abnormal designs. ABS will accept a
9-3
rational analysis considering thrust, torque, and bending loadings showing a safety
factor of 2 on fatigue.
The basic ABS sizing rule (ABS 4-3-2/5.1) appears as follows [3]:
D = 100 K[(H/N)(C1/(U + C2)]1/3 mm for a solid round shaft
where H = rated power PB [kW], thus based only on the torque loading
K = shaft design factor
N = rated rpm
C1 = 560, C2 = 160
U = minimum specified ultimate tensile strength, but
≤ 415 N/mm2 tail and stern tube shafts with salt water filled
bearings and no continuous liner
≤ 600 N/mm2 for tail and stern tube shafts with oil filled
bearings or salt water filled bearings with
a continuous liner
2
≤ 800 N/mm for all other shafts
K = 0.95 – 1.20 depending upon type of prime mover and the

presence of a thrust collar; e.g.,
straight at thrust collars

diesels, without flexible coupling K = 1.000 1.100
turbines, electric drive, or diesels 0.950 1.045
with flexible coupling
= 1.22 tail shafts in oil filled bearings or salt water

filled bearings with a continuous liner
= 1.29 tail shafts in salt water filled bearings without
a continuous liner
= 1.15 – 1.18 stern tube shafts
Evaluating this formula from first principles shows that  ~ 7,200 psi compared to a
mild steel tensile yield of 30,000 psi and an allowable shear stress of 15,000 psi. The
shafting is usually manufactured with about a 3% (0.25-0.50”) margin on diameter.
Solid shafts are used in most commercial, fixed-pitch propeller applications. Hollow
shafts are used in CRP propeller applications to provide for the pitching mechanism
and in most high powered naval applications to save weight (a hollow shaft with 20%
greater diameter saves about 50% on weight). Torque tubes, rolled plate or
composite large diameter shafts, are used in some applications. The ABS rule for
hollow shafts where the bore Di exceeds 40% of the outside diameter Do is as follows:
9-4
Do min = D/[1 – (Di/Do)4]1/3
where the D is for the equivalent solid shaft and the solution it iterated for Do.
5. establish bearing number and placement

The bearing number is usually established by the weight per unit projected area
that each bearing can support. The placement includes the longitudinal locations plus
the vertical (alignment) and occasionally the lateral location as well. The long stern
tube bearings are typically “slope bored” to align the bearing with the shaft deflection
curve in its normal ahead load condition. The weight supported by a bearing is given
by,
Wcapability = PLD = PD2(L/D) P = bearing projected area design pressure,

typically maximum permissible – 5 to 10 psi
D = shaft journal diameter within the bearing,
typically shaft diameter + 3 to 6 mm
L = bearing length
The placement of the bearings considers (1) the ship structure and arrangement, (2)
the equality of line bearing loads and equality of gear (motor) bearing loads, (3) the
bearing unit design loads P and L/D ratios, (4) the need for shafting flexibility, and
(5) lateral vibration (whirling) natural frequencies of the shafting. Typical bearing
capabilities are as follows:
L/D design P
ring lubricated line bearings ~1.0 45 psi
disk lubricated line bearings ~1.0 75 psi
water lubricated stern or stern tube bearings –
rubber or wood ~4.0 40 psi
resins or plastics ~4.0 40 psi
oil lubricated white metal stern tube bearings ~2.0 75 psi
oil lubricated synthetic stern tube bearings ~2.0 75 psi
The number of line bearings, after removal of the weight supported by the stern
bearing(s) and the gear or motor bearings can be obtained simply by,
Ninteger > Wremaining/Wcapability where Wcapability = PD2(L/D)
9-5
Generally, the best design has the fewest bearings and the greatest flexibility giving a
robust design, a tolerance for errors in manufacturing and wear down in service.
With too few bearings, the system is flexible but the unit loads P may be exceeded.
With too many bearings, the system is inflexible leading to problems with bearing
and gear wear due to manufacturing errors, hull flexing in a seaway, hull curvature
due to deck heating by the sun, etc.
Preliminary bearing placement guidelines for a D diameter shafting system in
order to start a design are as follows:
shaft size minimum spacing maximum spacing
10 ≤ D ≤ 16 inch 14 D ≤ x ≤ 22 D
16 ≤ D ≤ 30 inch 12 D ≤ x ≤ 20 D
The use of the larger spacing needs to be confirmed by lateral vibration natural
frequency calculations.
6. determine dynamic characteristics and loads – vibrations and shock
7. determine shafting details - tapers, flanges, fillets, bolting, sleeves, coatings, etc.
Bearing Examples. Figure 9.3 shows a typical disk lubricated white metal line bearing.
Figure 9.3 Self Aligning Disk Lubricated Line Bearing [2]
9-6
Figure 9.4 shows a typical tilting pad thrust bearing. The tilting pads align to transmit thrust
through the development of a wedge of oil between the pad and the thrust collar. Figure 9.5
shows a roller type line bearing on the left and a roller type combined line and one-direction
thrust bearing on the right.
Figure 9.4 Tilting Pad Thrust Bearing [2]
Figure 9.5 Roller Line and Thrust Bearings [1]
9-7
Figure 9.6 shows a typical oil lubricated stern tube from a single screw ship. This particular
example has a long L/D ~ 2 stern bearing, which supports the heavy cantilevered propeller
weight, and a shorter L/D ~ 1 forward bearing that supports the shafting. The oil supply comes
from a head tank that maintains a passive positive pressure even in the event of a ship electrical
power failure. Seals are provided at both the propeller and the aft peak bulkhead to seal the stern
tube.
Figure 9.6 Oil Lubricated Stern Tube and Bearings [2]
Shafting Alignment. The shafting arrangement and preliminary design determines the number
of bearings and the longitudinal x position of the bearings. Shafting alignment determines the
vertical z (and possibly the lateral y) position of the bearings. The long stern tube or stern
bearing may also be slope bored to better align the bearing with the journal along the shaft
deflection curve.
The design goals of the alignment include the following:
1) the bearings must not become overloaded; they must not exceed P max;
2) the bearings cannot become unloaded since this changes the shafting span and the
whirling (lateral vibration) natural frequencies;
3) there should be fairly equal loading on the line bearings;
4) there should be nearly equal loading on the reduction gear (or electric motor) bearings
to ensure good tooth life (rotor alignment), e.g. within 6000 lbs or less;
5) the system must be able to tolerate cold startup and hot operating conditions;
9-8
6) the system should be able to tolerate bearing wear down, hull deflection, and shipyard
errors in construction.
The typical English units are bearing reaction/support forces in lbs and bearing placement and
movement in mils = 0.001 inch. SI units would be in tonnes or Newtons and mm. Typical
bearing movement or changes in service that must be considered are as follows:
slow speed gear casing heat up (thermal growth between base and bearings) 15-30 mils
marine diesel crankshaft heat up (obtain data from vendor) less, shorter
water lubricated strut bearing wear down 200 mils
water lubricated stern tube bearing wear down 185 mils
worst case water lubricated bearing wear down (5 yrs; Great Lakes) 450-500 mils
The structural analysis of the shafting can be done by hand or using MatLab, specialized
shafting codes, or general purpose Finite Element Method codes; e.g., NASTRAN. The practical
use of these analysis results is greatly aided by the definition of Reaction Influence Numbers
(RIN) or coefficients,
RINij (or Iij)  change in reaction of bearing i in lbs (or N or t)

due to a change in position of bearing j up 1 mil (or mm)
Early shafting alignments were along a straight line with concentric bearing centers. Beginning
in the 1950’s, there was a shift to “fair curve alignment” in which bearings are placed
approximately along the shafting static deflection curve with this placement anticipating the heat
up and wear down expected in service [5]. The reduction gear bearings are placed low in the
cold condition in anticipation of their heat up to the hot condition. As noted above, most long
large-ship stern tube bearings are slope bored to attempt to gain more bearing contact, minimize
the bearing peak loading, and improve bearing performance. The bearings forward of the stern
tube bearing are placed in anticipation of the wear down expected in the stern bearing.
For this development, we will use a coordinate system with the bearing reaction force [lbs]
positive upward to support the shaft and the bearing displacement positive upward from a
straight line reference line [mils]. Thus, heat up displacements will be positive and wear down
displacements will be negative. The RIN coefficients and the bearing forces or reactions must
satisfy the following conditions:
9-9
the sum of all reactions (static) is the total weight of the shafting system;
the sum of all the RIN’s due to movement of bearing j must be zero;
the alternating RIN’s must change sign; loading bearing i unloads its neighbors;
the matrix of RIN’s is symmetrical;
high RIN’s indicates that the system is stiff and sensitive to misalignment;
low RIN’s indicates that the system is flexible and robust.
The vector of bearing forces or reactions R0 are calculated for the straight cold condition and the
vector of reactions R as the result of any vector of displacement changes  are then given by,
R = R0 + [RIN]  Ro vector of bearing reactions in straight cold condition

 vector of bearing movements from the straight line
R vector of bearing reactions after alignment or movement 

and where [RIN] is the symmetric matrix of Reaction Influence Numbers. In general, we have
R’ = R0 + [RIN] (design alignment + heat up + wear down + hull deflection+ manufacturing error)
and R2 = R1 + [RIN] 1-2
where 1-2 is the vector change in position of the bearings between states 1 and 2. Consider the
RIN matrix for a simple three bearing system. It will have the following form when viewed in a
tabular form,
bearing numbers
1 2 3 properties
bearing 1 I11 I12 I13 all Iii > 0 all i
numbers 2 I21 I22 I23 Iij = Iji, all i, j
3 I31 I32 I33 Iij = 0 j = 1, …, n
i
Coefficient I22 is the change in the reaction at bearing 2 when it is raised 1 mil, which will be
positive. Coefficient I12 is the change in the reaction at bearing 1 when bearing 2 is raised 1 mil,
which will be negative since bearing 2 will assume some of bearing 1’s load.
Consider the example system shown in Fig. 9.7 that shows the bearing reactions,
exaggerated deflected shaft shapes, and RIN table for a diesel electric drive shafting system. It
shows (a) the original system with 5 bearings all aligned in a straight line, (b) the system with
bearing 4 raised 0.020” or 20 mils, (c) the system in a straight line condition with bearing 3
removed to increase its flexibility, and (d) the modified system with bearing 4 raised 40 mils.
The RIN matrix is shown for each number of bearings. Note that the system is much stiffer with
five bearings as reflected in the large RIN values.
9-10
Figure 9.7 Bearing Reactions and RIN Matrices for Diesel Electric Shafting System [6]
9-11
In case (b) in Fig. 9.7 we have,
R4 = 1.4k + 297(4 = 20) = 7340 lbs = 7.3 k where k is kips = 1000 lbs.
and R3 = 16k – 428(20) = 7440 lbs = 7.4 k
In case (a) if bearings 1 and 2 heat up and rise 20 mils in the hot operating condition, the
reactions will be,
R1 = 24k + 185(1 = 20) – 395(2 = 20) = 19,800 lbs

R2 = 37k – 395(20) + 941(20) = 47,920 lbs
R3 = 16k + 261(20) – 770 (20) = 5,820 lbs
R4 = 1.4k – 55(20) + 238(20) = 5,060 lbs
R5 = 67k + 3(20) –14(20) = 66,780 lbs
Note that this is not an acceptable condition because the two motor bearings 1 and 2 typically
need to be within about 6,000 lbs to maintain proper rotor alignment. If the stern bearing were
now to wear down 200 mils from the hot operating case, the reactions of the three aft bearings
will be,
R3 = 5,820 + 49(5 = –200) = – 3,980 lbs

R4 = 5,060 – 52(–200) = 15,400 lbs
R5 = 66,780 + 14(–200) = 63,980 lbs
Bearing 3 is now completely unloaded, or it may be on the top of the seat in this case, which
would not be considered acceptable.
Now suppose that we want to raise bearing 4 in case (c) so that bearings 1 and 2 will have
equal loading in a hot operating (+20 mils) condition. The growth from the cold straight line
alignment to the hot condition will give the reactions,
R1 = 19k + 108(20) – 168(20) = 17,800 lbs

R2 = 51k – 168(20) + 272(20) = 53,080 lbs
R4 = 9k + 71(20) – 133(20) = 7,760 lbs
R5 = 66k – 11(20) + 29(20) = 66,360 lbs
We can accomplish this design goal by raising bearing 4 by an amount 4 so that it will take the
load from bearing 2, i.e., we seek R1 = R2 where,
R1 = 17,800 + 714
R2 = 53,080 – 1334
9-12
and for the R1 = R2 equality this will require,
17,800 + 714 = 53,080 – 1334

2044 = 35,280 or 4 = 172.9 mils (0.173”)
In this realigned condition, the hot operating condition reactions will be,
R1 = 17,800 + 71(4 = 172.9) = 30,076 lbs

R2 = 53,080 – 133(172.9) = 30,084 lbs
R4 = 7,760 + 91(172.9) = 23,494 lbs
R5 = 66,360 – 29(172.9) = 61,346 lbs
confirming that the R1 = R2 goal is actually achieved. In this case, however, the line bearing 4
load may exceed the Pmax for this type of bearing requiring additional changes to the alignment.
Finally, consider the example from Marine Engineering shown in Figure 9.8. This
analysis shows the bearing reactions for the straight line (Line-in-line) condition, the cold as
aligned condition, the hot as aligned condition, the wear down condition, and the measured hot
condition obtained on sea trials for a multiple screw Naval vessel. Note that the reduction gear
bearings have nearly equal reactions in the hot condition to ensure good gear wear. The line
bearings have nearly equal loadings, except for the aft line bearing that is kept intentionally
lower in the design condition so that it can accept more load when strut and stern tube bearings
wear down over life.
References
1. Woud, H. K., and Sapersma, D., Design of Propulsion and Electrical Power Generation
2. Long, C. L., “Propellers, Shafting, and Shafting system Vibration Analyses,” Ch. X in
Harrington, R. L. (ed.), Marine Engineering, SNAME, New York, 1992.
3. Rules for Building and Classing Steel Vessels 2001, Part 5 Vessel Systems and Machinery,
American Bureau of Shipping Rules, Houston, TX, 2000.
4. Design Data Sheet DDS 243-1 “Propulsion Shafting,” NAVSEA.
5. Anderson, H. C. and Zrodowski, J. J., “Coordinated Alignment of Line Shaft, Propulsion
Gear, and Turbines,” Transactions SNAME, Vol. 67, 1959.
6. Ecker, W., “Propulsion Shafting,” Marine Engineer Professional Degree Thesis, University
of Michigan, 1972.
9-13
NA331 Course Notes
Figure 9.8 Example Bearing Reactions for Multiple Screw Vessel [2]
9-14
Chapter 10 Gearing, Clutches and Flexible Elements
Marine propulsion plants that do not use low-speed diesels, most forms of electric drive, or
direct connection to water jets or occasionally propellers require some form of mechanical
reduction gearing and associated clutches and flexible elements. Gearing resolves the
fundamental incompatibility between the optimum speed of the prime mover and the optimum
speed of the marine propeller or propulsor. Electric drive can be viewed as another way to
accomplish this speed transition. Clutches are used to attach and remove engines from the
system when multiple engines are used in a mechanical drive propulsion system. Flexible
elements are used to accommodate misalignment and the relative movement of components and
to isolate and damp vibratory excitation from reciprocating engines. Consult the short treatment
in Woud and Stapersma [1] and Mowers’ chapter in Marine Engineering [2].
Classification of Gearing. Mechanical gearing can be classified in a number of ways. First, it

can be classified by the number of stages of speed reduction. A stage consists of a small
diameter input gear (pinion) acting on a larger diameter output gear (gear) to produce a speed
change that is proportional to the ratio of their diameters. The gear ratio i = Dgear/Dpinion =
npinion/ngear where n is the rotation speed [rps]. The gear ratio can be expressed as a single
number i or as a ratio i:1. Marine gears can have single, double, or even occasionally triple
reduction. The typical range of their application is as follows:
reduction stages range of total gear ratio i typical application area

singe reduction 1 < i ≤ 8 - 10 medium-speed reciprocating engines
double reduction 8 - 10 < i ≤ 60 steam and gas turbines
triple reduction 45 < i ≤ 80 steam turbines and
smaller gas turbines
The reduction gears used with high-speed diesels can be double or triple reduction.
Gearing can also be classified as having a parallel, bevel, worm, or epicyclical
configuration. Parallel gearing is the most common in marine applications and consists of a
pinion driving a gear with the axes of the two components offset and parallel. A typical single
reduction, reversing parallel gear as might be used with a medium-speed diesel is shown in Fig.
10.1. This particular gear uses air inflated (pneumatic) clutches to engage the ahead or reverse
10-1
pinions. The engine input is to the spacer at the right. The ahead clutch engages the main pinion
shaft that passes through the first reversing path gear.
Figure 10.1 Single Reduction, Reversing, Parallel Gear Configuration
Bevel gears are used when the input and output axes are coplanar and intersect, usually at
90 degrees. These are seen in the drive trains for hydrofoils, SWATH vessels, and Z-drive
thrusters where the power must be transmitted at right angles downward through a strut to a
submerged hull or pod. Worm gears are used when the input and output axes are non-
intersecting and cross at 90 degrees. Applications are mostly in auxiliary equipment such as
winches.
Epicyclical gearing involves gears working inside and around other gears with concentric
input and output shafts. They consist of an inner pinion called the sun gear, an outer carrier ring
with internal teeth called the annulus, and three or more gears that operate between the sun gear
and annulus ring. Epicyclical gearing is becoming more common in high-speed craft applications
10-2
with restricted space, particularly with waterjets. The system can be implemented as either a
planetary, star, or solar gear characterized by the following:
epicyclical type input output fixed output rotation

planetary (3 ≤ i ≤ 12) sun gear planet gear carrier annulus same as input
star (2 ≤ i ≤ 11) sun gear annulus star gear carrier opposite to input
solar annulus star gear carrier sun gear same as input
Sketches of star and planetary reduction gear configurations are shown in Fig. 10.2. In the star
gear arrangement at the left, the star gear carrier is fixed and the star gears do not rotate around
the sun gear. The output is taken from the annulus or orbit gear. In the planetary gear
arrangement at the right, the annulus or orbit gear is fixed and the planet gears rotate about the
sun gear with the output taken from the planet gear carrier.
Figure 10.2 Star and Planetary Epicyclical Reduction Gear Configurations
Epicyclical gear sets offer significant weight and space saving potential and permit
concentric input and output shafts. For a long time, the limiting factor to their use was the lack
of machining capability to cut internal gear teeth at large diameters. The concentric shafts can
offer significant arrangement advantages over parallel gearing in restricted arrangement
situations. A comparison of alternative propulsion plants is shown in Table 10.1. In this
particular example, the medium-speed diesel system is able to use the optimum propeller rpm
giving a lower total required power. The low-speed diesel rpm is set by the engine design and
10-3
this results in the use of a suboptimal propeller for this application. The epicyclical gear allows
the use of optimum propeller rpm and with concentric input and output shafts that allows the
medium-speed engine to be placed further aft in the ship, thus, shortening the overall engine
room length.
Table 10.1 Example Comparison of Diesel Propulsion Plant Alternatives
engine type propeller engine hpBr engine room weight decks cut
rpm length above engine
low-speed diesel 122 12,600 11.25 m 404 t 4
medium-speed diesel
with parallel gearing 114 12,000 10.50 m 170/50 t 2
medium-speed diesel total/gear
with epicyclical gear 114 12,000 10.00 m 137/17 t 2
Figure 10.3 shows a recent steam plant combining gear that has double reduction on the LP
side and triple reduction on the HP side using a star and a planetary epicyclical gear for the first
two reductions. The LP side uses a planetary epicyclical gear for the first reduction and then the
two paths are combined with parallel gears.
Figure 10.3 Recent Double-Triple-Reduction Steam Plant Reduction Gear
Implementations with sun gear input and output from both the sun gear shaft and the
annulus or ring gear have been used to provide contra-rotating, concentric outputs for contra-
10-4
rotating propeller systems. Figure 10.4 shows a cutaway of a star-compound contra-rotating

propeller reduction gear using concentric outputs from both the sun and ring gear shafts.
Figure 10.4 Epicyclical Contra-rotating Propulsion Gear Set
Gearing can also be classified as spur, single helix, or double helix (herringbone). Spur
gears have their teeth parallel to the axis of rotation. These are simple gears that only have one
tooth on each gear in contact at any time. This can lead to rough, noisy operation. Spur gears
are usually never used in high-powered marine propulsion gearing. In single helix gears, the
teeth lie along helical paths along and around the axis of rotation. With helical teeth, many pairs
of teeth are in contact at any time giving a smooth, quiet transmission of power. This design has
the disadvantage, however, that axial forces are created; additional radial bearing forces are also
created. Figure 10.1 shows single helix marine gearing. Double helix gears consist of two
helical sections of teeth set at opposite angles with a manufacturing clearance or groove between
them. This design provides cancellation of the axial forces and locks the axial position of the
pinion with respect to the gear. Most high-powered marine applications utilize double helix
gearing. Figure 10.3 shows double helix gears used for both the epicyclical gear sets and the
combining pinions and large output gear.
Double reduction gears can also be classified by the number of power paths between the
first stage pinion and the second stage gear. With a single power path, the input pinion drives the
first stage gear, which drives the second stage pinion, which drives the large output gear – often
called the bull gear due to its large size. To make high-powered gears (primarily naval gears)
10-5
smaller and more compact, the first stage pinion can drive two parallel second stage gears that
drive two second stage pinions, both of which act on the bull gear. With two gears and pinions
sharing the load, the components can be made smaller for the same material loading. For the two
pinions to share the load equally, they must be manufactured and positioned precisely. These
gears are, thus, called locked train gears since the two parallel second stage pinions are in locked
positions relative to one another. Figure 10.5 shows a locked train gear for the gas turbine side
of a CODOG plant.
(rotated to the horizontal for clarity)

Figure 10.5 Schematic and Arrangement of Two Power Path CODOG Reduction Gear [1]
The medium-speed diesel input on the right side in Fig. 10.5 is single reduction. The gas
turbine input on the left is double reduction with two power paths between the first stage pinion
and the second stage gear. The schematic on the left has the two first stage gears rotated up to
the horizontal for visibility. The actual arrangement is shown in the end view on the right.
Principal Gearing Design Considerations. Tooth contact stress, tooth bending stress, and
torsional and bending deflection of the long slender pinions are the principal design
considerations for marine gearing.
Tooth contact stresses are the primary consideration for long term gear durability. If the
contact stress becomes too high, a local loss of material (pitting) occurs after a period of
operation. Contact stress is usually calculated based upon an adaptation of the Hertz formula for
the calculation of the maximum contact stress between two cylinders of radii r1 and r2,
10-6
respectively [3]. In this situation, the maximum compressive stress c [psi] is given by the Hertz
formula,
c = [0.175 (W/L) E (r1 + r2)/r1r2]1/2 E = Young’s modulus [psi]

ri = radii [in]
L = contact length [in]
W = loading [lbf]
Gear teeth are not cylinders, but involute profiles, which can be approximated as cylinders near
their line of contact. When this is done, we get the configuration sketched in Fig. 10.6.
Figure 10.6 Geometry of Gear Tooth Contact
Using these relationships in the Hertz formula gives, after a convenient rearrangement, a
gear loading parameter called the K factor defined as follows for steel gearing (E = 30 x 106 psi):
K factor = [c /4582]2 sin2= [Wt/(Ld)] • (1 + R1)/R1 d = pinion diameter 2r [in]

allowable actual Wt = tangential loading [lbf]
R1 = reduction ratio = R/r
Note that the tangential loading Wt will produce input torque Qin = Wt r at the input rpm and
output torque Qout = Wt R at the output rpm. The grouping of terms in the center square brackets
should be about the same for successful gears of the same material and material treatment. Thus,
the K factor has evolved as a measure of marine gear contact stress loading. The details of a
particular application determine the actual K factor as given in the right hand expression in this
equation. The allowable K factor is given by the central part of this equation. Notice that the K
factor is proportional to c2/E.
The allowable K factor depends upon the following:
(1) gear band hardness (since pinions are usually 20-30% harder than the gears)
(2) ship operating profile (since pitting is a time dependent phenomenon, the percent time
at each loading is significant)
10-7
(3) rpm of the gear set (since hydrodynamic oil separation of teeth improves with greater
slip speed between the gear teeth)
The K factor has evolved as an experience factor that can be used in marine gear specifications
to limit tooth contact stresses and ensure durable gears. It does have a rational basis in the Hertz
formula, but it is usually set by acceptable experience with similar gears. Typical limits are:
gear application area max. K factor explanation of value

commercial first reductions of double reduction 140 continuous full power
commercial second reductions/single reductions 110 continuous, but lower rpm
naval non-nuclear first reductions of double reduction 225 little high power operation
naval non-nuclear second reductions/single reductions 175 same, but lower rpm
naval nuclear first/second reductions closer to more high power operation
commercial since not fuel limited
epicyclical gearing to 300
Thoma proposed a rational correlation between the allowable K factor for commercial gear
applications and the gear band material properties expressed as a function of hardness on the
Brinell Hardness Scale (BHN) as shown in Fig. 10.7 [4].
Figure 10.7 Proposed Relationship between Allowable K Factor and Gear Band Hardness [4]
Given an allowable K factor in a particular case, the K factor equation above can be used to
perform valuable preliminary sizing of marine gears. Examples will be presented below. Also
see Marine Engineering [2] for a example sizing of a double reduction gear. The Michigan Gear
Sizing Program (GSP) [5] implements the sizing algorithms of Marine Engineering [2] and
Balukjian [6] and provides as useful way to obtain preliminary gear sizing information needed
for preliminary arrangements design.
10-8
Tooth Bending Stresses. The controlling aspect of detailed gear tooth design is the
bending stress developed in the teeth by the tangential loading Wt. The teeth are subjected to
periodic loading which can cause fatigue failure at the tooth root due to bending stresses. The
design problem is greatly complicated by the dynamic loading, vibration, etc. Semi-empirical
formulae including many correction or experience factors are utilized in the gear design. The
loading usually dictates the use of small, short teeth (high diametral pitch, a high number of teeth
per unit of diameter). Small teeth give lower bending moments and smooth, quiet operation.
Refer to Marine Engineering [2, pp. 330-331] for a simplified treatment and to Thoma’s paper
on marine gear design [7].
Bending and Torsional Deflection of Pinions. The second important design factor that
establishes the overall size of gear sets is the flexibility of long slender pinions. For a given K
factor (level of contact stress), more torque can be transmitted by increasing the gear contact
length L. Limiting L, however, is the problem that as the pinion becomes too long and slender
the pinion deflects due to bending and torsion to the extent that there is loss of tooth contact and
a concentration of the loading where the contact remains. This usually limits pinion length to 1.0
≤ L/d ≤ 2.5 in order to keep the tangential (circumferential) deflection of the pinion due to
torsion and bending to less than 0.001”. See Marine Engineering [2, pp. 334-335] for equations
for calculating these deflections.
Gear Sizing Examples. We will consider sizing examples for standard parallel gearing and
epicyclical gears based upon the K factor specifications.
Example 1: Suppose two identical medium-speed diesels producing PB = 12,000 hpBr each at N1
= 388 rpm are to drive a propeller at N2 = 122 rpm through a single reduction combining gear
with a K factor specification of K = 110. What are the minimum acceptable pinion and gear
diameters? What would be the separation distance between the two input shaft centerlines? The
configuration is shown in the sketch.
10-9
We will assume that the pinion length-diameter ratio L/d = 2.5 to limit the pinion
deflection. Equating tangential speeds at the pitch circle contact point,
N1r = N2R so that N1/N2 = R/r = 388/122 = 3.18 = R1
input torque per pinion Q1 = 33000 • 12 in/ft • PB/(2N1) = 63025 • 12000/388

(550 ft lbf/hpBrs • 60 s/m) = 1,949,238 in lbf
Using the allowable K factor and noting that the tangential force Wt = Q1/r and L = 2.5d = 5r
K = 110 = [Wt/(2rL)] • (1 + R1)/R1 = [Q1/(r 2r 5r)] • (1 + 3.18)/3.18
Solving r = {[Q1/(10K)] • (1 + 3.18)/3.18}1/3 = 13.26” and R = 3.18r = 42.15”
so the resulting shaft separation is tentatively s = 2r + 2R = 110.8”
Since we assumed a rather long pinion with L/d = 2.5 (2.0 maximum is recommended), we
should also check the torsional and bending deflection of the pinion. Using the equations from
Marine Engineering [2], we will get a torsional and bending deflection of 0.0022” which means
that the gear probably should be redesigned with an L/d of 2.0. Continuing with example,
however, we can bring in one more design issue. Gear teeth must be cut by machines with
tooling that can only cut teeth in standard discrete sizes. When we use the gear bending stress
formula from Marine Engineering [2] for a unit loading U which has an acceptable range 6000 ≤
U ≤ 8000 where,
U = Wt NDP/L where NDP is the normal diametral pitch (teeth per inch of diameter)
for this example NDP = UL/Wt = 6000 L/(Q1/r) = 2.71
The available diametral pitches (gear cutters) are in 1/2 intervals between 2 and 4 so we can
round up to obtain the smaller of the nearest tooth sizes, NDP = 3 giving,
NDP • d = 3 • 2 •13.26” = 79.56 teeth
The gear must have an integer number of teeth around the circumference of its pitch circle so we
must select 80 teeth and modify the diameter to match; i.e., d = 80/3 = 26.67” and r = 13.33”
(instead of the initial estimate 13.26”). The teeth ratio of the gear and pinion must also be the
gear ratio; i.e., R1 = number of teeth on gear/number of teeth on pinion. So the number of teeth
on the gear will need to be 80 • 3.18 = 254.4, but this also must be an integer so the gear must
10-10
have either 254 or 255 teeth, which will give a possible an output rpm of 122.2 or 121.7,
respectively. We might choose 122.2 as acceptable giving 254 teeth on the gear. Adjusting the
diameter of the gear for this number of teeth gives,
D = N/NDP = 254/3 = 84.66” giving R = 42.33”
and a final shaft separation s = 2r + 2R = 111.33” (0.47% higher)
This distance must be checked with the width of the selected engines to ensure that there is
adequate (0.8 – 1.0 m) maintenance access distance between the two engines. The Michigan
Gear Sizing Program [5] performs the preliminary estimate calculations based on K factors only
for the most common parallel gear configurations, without the added consideration of ensuring
there are an integer number of teeth.
Example 2: Suppose a single medium-speed diesel producing PB = 12,000 hpBr at N1 = 520 rpm
is to drive a propeller at N = 122 rpm through a single reduction star epicyclical gear with a K
factor specification of K = 250. The overall gear ratio R’ and total input torque QT will be,
R’ = 520/122 = 4.26 and QT = 63025 • 12000/520 = 1,454,423 in lbf

The configuration is shown in the sketch,
Assume that L = 2r1 for all components, which would be typical. Now in a star configuration,
axis points A and B are fixed in space so by geometry and the relative motion of the tooth
contact points we have,
r1N1 = r2N2 = r3N3 so r3/r1 = N1/N3 = 4.26 giving r3 = 4.26 r1

r2 = (4.26 – 1)r1/2 = 1.63 r1
We next need to check how many star gears Z can fit within the annulus ring, using the sketch
10-11
and distance a must exceed 2r2 so conservatively we can assume a > 2r2 + 2”. Thus,
sin(2/2Z)min = (a/2)/( r1 + r2) = (r2 + 1)/(r1 + r2) = (1.63r1 + 1)/2.63r1
neglecting the 2” for the moment, sin(/Z)min ≈ 1.63/2.63 = 0.62
giving Zmax ≤ /sin-1(0.62) = 4.7 and as an integer Zmax = 4
With Z = 4 star gears, the power is split four ways (four power paths giving part of the reason the
epicyclical gear can be more compact), so each star gear receives Q = QT/4 = 363,605 in lbf =
Wtr1. We can now use the K factor to establish r1 noting that the local gear ratio between the sun
gear and the star gears is R1 = r2/r1 = 1.63,
K = 250 = [Wt/(2r1L)] • (1 + R1)/R1 = [Q/(r1 2r1 2r1)] • (1 + 1.63)/1.63

r1 = [363,605 • 2.63)/(1000 • 1.63)]1/3 = 8.37” and L = 2r1 = 16.74”
r2 = 1.63r1 = 13.64”
r3 = r1 + 2r2 = 35.66”
The pinion deflection now needs to be checked. It can be found to be a low 0.0008” in this case
using the equations from Marine Engineering [2]. Checking the tooth bending to get the NPD
and number of teeth yields the following for a selection of tooth size NDP = 2.5:
sun pinion number of teeth 42 final radius r1 = 8.40”
star gears number of teeth 69 final radius r2 = 13.80”
annular ring gear number of teeth 180 final radius r3 = 36.00”
The final gear ratio R’ = r3/r1 = 4.29 with a final output rate of N3 = 520/4.29 = 121.2.
Example 3. Repeat example 2 using a planetary epicyclical reduction gear. The only change is
in getting the ratios of the radii. The solution is presented in the following sketch and analysis:
10-12
Clutches and Coupling. Clutches are required in plants using multiple engines. In combined
plants (CODAG, CODOG, etc.), they must permit the various prime movers to come on line and
assume a share, or all, of the load. In multiple identical or similar prime mover plants, the
clutches are needed so that any single engine can be used at low load and the reliability
advantage of continued operation with an engine down can be realized. Various couplings are
used to accommodate system misalignment or movement or to modify the torsional vibration
natural frequencies of the system and damp out the transmission of vibratory torques from
reciprocating engines into their associated gearing.
Clutches can be classified as energy absorbing or overrunning clutches. Energy absorbing
(or friction) clutches utilize various friction elements. Some friction clutches are operated
pneumatically using air pressure to inflate elements that engage the clutch. The clutches used in
the reversing reduction gear shown in Fig. 10.1 are essentially annular tires that are inflated to
push the clutch surfaces radially inward (or axially in other designs) for engagement. When the
air pressure is released centrifugal force moves the clutch surface outward for disengagement.
The air is supplied down the center of the pinion shaft through a rotating air seal. Friction
clutches must absorb slip and heat, without failure, while synchronization is achieved between
the input and ouput elements. Upon full engagement there are no losses and  = 1.0.
Other friction clutches are operated by hydraulics in which oil pressure is used to move
clutch plates, perhaps axially, to engage a pack of annular plates. The alternating clutch plates
are splined onto the input and output shafts. An example of this type of clutch is shown in Fig.
10.8, which shows a Renk Tacke planetary reduction gear with an internal, oil-operated clutch
pack that holds the annulus ring when power is to be transmitted. This gear allows the input
prime mover to be idled with no output – the condition shown in the figure. The sun, planet, and
annulus gears all rotate, but the planet carrier, the output mechanism, does not. For engagement,
oil pressure is supplied to the annular piston at the right of the clutch pack and it moves to the
left, compressing the clutch pack to stop the annulus ring gear. With the annulus stopped, the
planet gear carrier rotates providing the gear’s output.
Overrunning clutches utilize mechanical friction to create the engagement. A leading
example of this type of clutch mechanism is the Synchronous Self Shifting (SSS) overrunning
clutch. These clutches are used primarily in combined plants where one prime mover must be
brought on line and then take over the powering of a system that is already in operation. They
10-13
transmit power only in one direction so they are used in CRP propeller applications. The input
must be synchronized with the operating system before engagement can be made. A simplified
schematic of this type of clutch is shown in Fig. 10.9. The input is to the outer clutch ring that
has a short internal ring of gear teeth used for engagement. The output is from the inner shaft.
Figure 10.8 Planetary Gear with Internal Oil Operated Clutch Pack [1]
The shifting mechanism moves axially along a helical spline to engage the gear teeth when
the input shaft reaches synchronization and then exceeds the output speed by a small amount.
The engagement is initiated by a series of pawls around the shifting mechanism that engage
when synchronization is reached. As long as the output shaft has a higher rpm, the pawls are not
engaged. When the speed of the clutch ring exceeds that of the output shaft, the pawls engage
shoulders on the outside of the SSS unit causing it to rotate which drives it forward along the
spline bringing the gear teeth into engagement. Note that there is no external power, signal, or
command required. Note also that upon engagement, and as long as power is being supplied,
there is a positive mechanical gear connection that will have zero losses. SSS clutches can be
seen in the CODOG plant combining gear shown in Fig. 10.5. Both the diesel and the gas
turbine have SSS clutches so that they can exchange supplying power to the CODOG system.
10-14
Figure 10.9 Schematic of Synchronous Self Shifting Clutch Mechanism
Various hydraulic couplings are used as clutches, to isolate reduction gears from the torque
variations produced by reciprocating engines, and to provide vibration damping. There are two
main types of hydraulic couplings: the Fottinger-type hydrodynamic coupling and the hydraulic
torque converter. These can be used as clutches by simply draining their oil.
The hydrodynamic coupling consists of a driving impeller (pump) unit on the input shaft
and a driven runner (turbine) unit on the output shaft. They operate in an oil filled cavity. A
drawing of a hydrodynamic coupling and its torque and efficiency characteristics for constant
input speed are shown in Fig. 10.10. The input and output portions of the coupling are connected
only by a circulating vortex created in the oil in the annulus between the two units. Power is
transmitted by the slip (difference in rotation speed) between the two units. As can be seen in
Fig. 10.10, the efficiency is linear with output speed and, thus, it falls off rapidly with increasing
slip (input speed - output speed). At the rated point, however, the efficiency can be 97-98%.
An example of the use this type of coupling is shown in the diesel input to the CODOG gear
shown in Fig. 10.5. The use of the fluid coupling allows the diesel to be idled with the coupling
oil drained and then it can be smoothly brought on line by filling the coupling.
10-15
Figure 10.10 Hydrodynamic Coupling [8, 9]
The hydraulic torque converter adds a stator guide wheel element to the arrangement as
shown in Fig. 10.11. The stator adjusts the oil circulation so that there are constant input
conditions to the radial pump unit on the input shaft. With a stator, the flow from this unit is
essentially constant for constant input speed. Thus, the radial turbine that drives the output shaft
exhibits approximately the same quadratic efficiency and linear, rising toward stall characteristic
seen with steam and two-shaft gas turbines. Both types of hydraulic couplings serve to isolate
gearing from the torsional vibratory inputs from reciprocating engines. “Gear ratio” changes up
to 5:1 for low speed maneuvering below the minimum diesel rpm can be achieved by changing
the oil level.
Figure 10.11 Hydrodynamic Torque Converter [9]
10-16
Reversing Reduction Gears. The operation of reversing reduction gears during crash back
maneuvers requires careful control to ensure efficient stopping and clutch survival [10]. The
clutches in these gears have severe service because they must absorb heat energy from the
slipping when the shaft is slowed, stopped, and then accelerated in reverse. To reverse, the
throttle is reduced to prevent over speeding, the ahead clutch is disengaged, and the astern clutch
engagement is initiated. Before the astern clutch can synchronize and lockup, it must slow down
the propeller (which the ship movement is trying to turn ahead), stop the propeller, and then
accelerate the propeller in the astern direction. All this time slip is occurring across the astern
friction elements causing them to heat, i.e.,
Qin = Qout = Qtransmitted

Pin = Pout + Pheat
QinNin = QinNout + Qin(Nin – Nout) = QinNout + QinNslip
The slip power Pheat = QinNslip goes into clutch heating. Electrical or friction shaft brakes can be
used to help stop the shaft and reduce the amount of heat that the astern clutch must absorb.
Careful control of timing is needed [10]. A schematic of the crash stop is shown in Fig. 10.12.
Slip does not begin in the ahead clutch until enough air has bled to make the torque capability of
the clutch less than the input torque. Lockup occurs at the end of slip when the torque capability
of the astern clutch exceeds the transmitted torque.
Figure 10.12 Reverse Reduction Gear Torques during Crash Back
Flexible Elements. Various flexible units are used in propulsion systems to accommodate
relative movement and misalignment, isolate gearing from the torsional vibratory inputs from
reciprocating engines, and provide torsional damping. The use of resilient mounts on diesels and
gas turbines for sound isolation make the relative movement compensation increasingly
important. The following types of flexible elements are common:
10-17
1) Dental couplings are shafts with gears at each end which operate inside of housings
with internal teeth. Thus, they transmit torque but allow axial movement such as
occurs between a turbine and its attached gearing during heat up. See Fig. 10.13.
Figure 10.13 Dental Coupling between Steam Turbine Rotor and Gear Pinion [2]
2) Quill shafts are long flexible shafts used in gearing (usually locked-train gears)
such as between first reduction gears and second reduction pinions. These shafts
are flexible in torsion and also allow axial movement. Their torsional flexibility
helps the load sharing between parallel power paths. These shafts often extend
through the gears and are attached to the outer ends of the gears to achieve
maximum possible length for added flexibility. See Fig. 10.14.
Figure 10.14 Example Quill Shaft between Reduction Stages [11]
10-18
3) Torsional flexible couplings use various metallic and elastomer elements in shear,
compression, or bending to alter the natural frequency of shafting systems in
torsion, isolate gears from the vibratory input from a reciprocating prime mover,
and to add vibration damping to the system. A typical torsional coupling using
radial metallic leaf springs in an oil filled cavity is shown in Fig. 10.15. The input
is to the innerstar (9) attached to the inner end of the spring packs (1); the ouput is
taken from the flange (5) that is attached to the outer end of the spring packs. The
cavity is filled with oil to increase the damping. Similar devices are also used at
the free end of diesel engine crankshafts to add damping and modify the system
natural frequencies by adding a significant inertia at the outer clamp ring.
Figure 10.15 Torsional Coupling [Geislinger]
3) Alignment flexible couplings of various types using diaphragms, membranes,

elastomer elements, etc. are used to accommodate radial, axial, and angular
misalignment and movement within shafting systems. These elements are fairly
common at small powers, but rare in large ships. Units using advanced glass and
carbon fiber composites are used on high-speed ferries and other high
performance applications. In long configurations, these become light-weight
composite torque tube shafts. Units are manufactured in combination with
torsional flexible couplings and clutches to provide compact, multipurpose
couplings. Figure 10.16 shows composite couplings between high-speed diesels
10-19
and epicyclical gears and then long composite torque tube couplings (shafts)
connecting to waterjet input shafts.
Figure 10.16 Waterjet System Composite Flexible Couplings and Shafts [Geislinger]
References
2. Mowers, G. P., “Reduction Gears,” Ch. IX in Harrington, R. L. (ed.), Marine Engineering,
3. Seeley, F. B. and Smith, J. O., “Contact Stresses,” Ch. 11 in Advanced Strength of Materials, 2nd
Ed., John Wiley, New York, 1959.
4. Thoma, F. A., “An Up-to-Date Look at Marine Gear Tooth Loading,” Marine Technology, Vol.
7, No. 2, April, 1970.
5. Parsons, M. G., Li, J., and Singer, D. J., "Michigan Conceptual Ship Design Software
Environment User's Manual," University of Michigan, Department of Naval Architecture and
Marine Engineering, Report No. 338, July 1998.
6. Balukjian, H., “A Computerized Method for the Preliminary Design of Main Reduction Gears,”
Naval Engineers Journal, Feb., 1972.
7. Thoma, F. A., “An Updated Approach to Marine Gear Tooth Bending Strength,” Transactions
SNAME, Vol. 80, 1972.
8. Taylor, D. A., Introduction to Marine Engineering, Butterworth-Heinemann, Oxford, UK, 1990.
9. Pickert, H., “Applicability of Hydrodynamic Transmission to Ship Propulsion with Gas Turbines
and CODAG Plants,” ASME Paper No. 70-GT-85, 1970.
10. Parsons, M. G., and Altemos, E. A., "The Effect of Shaft Brakes on Emergency Stopping Head
Reach with Reversible Reduction Gear Propulsion Systems," Proceedings of the 1975 Summer
Computer Simulation Conference, July 21-23, 1975.
11. Payne, C. N., Naval Turbine Propulsion Plants, U. S. Naval Institute, Annapolis, 1958.
10-20
Chapter 11 Engine-Propulsor Matching
We have introduced the various marine prime movers and the mechanical gearing,
electrical systems, shafting, and components that can connect the prime mover to the propulsor.
This chapter has three primary goals. First, it reviews the various power definitions and margins
that relate the prime mover sizing to the resistance of the vessel. Second, it covers the various
types of propulsors that are typically found on ships: conventional propellers, controllable-
reversible pitch (CRP) propellers, podded propulsors and thrusters, waterjets, etc. Finally, it will
address the overall system design problem of matching the prime mover to the propulsor – and
vice versa. In this treatment, the propulsor is considered to be part of the propulsion machinery
system and part of the marine engineer’s responsibility, even though the naval architects also
considers it part of their purview. Consult the chapter in Woud and Stapersma [1] and the
relevant chapters in Principles of Naval Architecture [2] and Marine Engineering [3].
In general, the propulsion plant involves the prime mover(s) that produce torque Q [kN m
or ft lbf] at an engine speed Ne [rpm] or ne [rps], a propulsor that produces a thrust T [kN or lbf]
at some propulsor speed Np and ship speed V, and a transmission system that may include
gearing or electric components to transmit torque and thrust and modify rotation speed as
needed. The marine engineer must properly match and combine these elements into a single
propulsion system producing thrust T = T(engine throttle , propeller pitch P, Ne, V). Of course,
the propeller pitch may be variable or it may be fixed in the initial design and manufacture. If
the propulsor is a waterjet, the equivalent pump characteristics are fixed in design. Recall that,
in general, the power is P = 2QN/60, where the torque, rpm, and power are a common point in
the system and the units are properly handled.
Power and Efficiency Definitions
The determination of the required propulsion power and engine sizing requires working
from a hull total tow rope resistance prediction RT to the required installed prime mover brake
power PB. It is important to briefly review the definitions used in this work [2, 4]. The approach
used today has evolved from the tradition of initially testing a hull or a series of hulls without a
propeller, testing an individual or series of propellers without a hull, and then linking the two
together through the definition of hull-propeller interaction factors. The various powers and
11- 1
efficiencies of interest are shown schematically in Fig. 11.1. It is obviously essential that one is
precise in stating which power is being discussed in any particular context. The hull without a
propeller behind it will have a total tow rope resistance RT at a speed V that can be expressed as
a power called the effective power PE,
PE = RT V/1000 [kW] (11.1)
where the resistance is in Newtons and the speed is in m/s. In English units with lbs and ft/s,
respectively, the power is in hpBr and the divisor is 550 ft lbf/hpBrs.
Figure 11.1 Location of Various Power Definitions
The open water test of a propeller without a hull in front of it will produce a thrust T at a
speed of advance VA with an open water propeller efficiency o. This can be expressed as a
thrust power PT,
PT = TVA /1000 [kW] (11.2)
These results for the hull with all appendages but without the propeller and for the propeller
without the hull can be linked together by the definition of the hull-propeller interaction factors
defined in the following:
VA = V(1 – w)
T = RT/(1 – t)
P = or (11.3)
where w is the Taylor wake fraction, t is the thrust deduction fraction, P is the behind the hull
condition propeller efficiency, and r is the relative rotative efficiency that adjusts the open water
11- 2
propeller efficiency o to its efficiency behind the hull. The wake fraction w reflects the slowing
of the flow within the ship’s boundary layer and wake. The thrust deduction t reflects the
increase in the resistance of the ship due to the acceleration of the wake when there is a propeller
behind the ship. Note that r is not a true thermodynamic efficiency and may, thus, assume
values greater than one.
Substituting eq. 11.3 into eq. 11.1 and eq. 11.2 yields the relationship between the thrust
power and the effective power,
PT = PE (1 – w)/(1 – t) (11.4)
from which we define the convenient grouping of terms called the hull efficiency H,
H = (1 – t)/(1 – w) = PE/PT (11.5)
The hull efficiency can be viewed as the ratio of the work done on the hull PE to the work done
by the propeller PT. Note that H is also not a true thermodynamic efficiency and may also
assume values greater than one.
The delivered power input to the propeller PD is related to the output thrust power from the
propeller PT by the behind the hull efficiency eq. 11.3 giving when we use eq. 11.5,
PD = PT/P = PT/(or) = PE/(Hor) (11.6)
The grouping of efficiencies dividing PE was previously called the Quasi-Propulsive Coefficient
or QPC; currently this is called the propulsive efficiency D = Hor.
The shaft power PS is defined at the output of the reduction gear or electrical transmission
process, if installed, and the brake power PB is defined at the output flange of the prime mover.
When steam machinery is purchased, the vendor typically provides the high-pressure and low-
pressure turbines and the reduction gear as a combined package so steam plant design typically
estimates and specifies the shaft power PS, since this is what the steam turbine vendor must
provide. When diesel, natural gas or gas turbine prime movers are used, the gear may be
provided separately so the design typically estimates and specifies the brake power PB, since this
is what the prime mover vendor must provide. The shaft power PS is related to the delivered
power PD transmitted to the propeller by the stern tube bearing and shaft seal efficiency s and
the line shaft bearing efficiency b by,
11- 3
PS = PD/(sb) (11.7)
The shaft power PS is related to the required brake power PB by the transmission efficiency of
the reduction gear or electrical transmission process t by,
PB = PS/t (11.8)
Combining eqs. 11.6, 11.7, and 11.8 now yields the needed relationship between the effective
power PE and the brake power at the prime mover PB,
PB = PE/(Horsbt) (11.9)
Power Margins
In propulsion system design, the design point for the equilibrium between the prime mover
supplied power and the propulsor consumed power is usually the initial sea trials condition with
a new vessel, clean hull, calm wind and waves, and deep water. The resistance is estimated for
this ideal trials condition. A power design margin MD is included within or applied to this
resistance prediction, and thus the effective power, in recognition that the estimate is being made
with approximate methods based upon an early, incomplete definition of the design. This is
recommended since most designs today must meet the specified trials speed under the force of a
contractual penalty clause. It is also necessary to include a power service margin MS to provide
the added power needed in service to overcome the added resistance from hull fouling, waves,
wind, shallow water effects, etc. not included in the trials condition resistance estimate. When
these two margins are incorporated, eq. 11.9 for the trials design point becomes,
PB reqd(1 – MS) = trials match point PB = PE (1 + MD)/(Horsbt) (11.10)
The propeller is designed to achieve this equilibrium point on the initial sea trials, as shown in
Fig. 11.2. The design match point provides equilibrium between the engine capability curve, the
prime mover at (1 – MS) throttle and full rpm (the left side of the equality in eq. 11.10), and the
propeller load with (1 + MD) included in the prediction (the right side of the equality). This
brake power PB reqd now represents the minimum brake power required from the prime mover.
11- 4
The engine(s) can, thus, be selected by choosing an engine with a total Maximum Continuous
Rating (or selected project engine rating) which exceeds this required value; i.e.,
Project Selected MCR ≥ PB reqd = PE (1 + MD)/{Horsbt(1 – MS)} (11.11)
sea trails design

match point eq. 11.10
service margin 15%
ideal diesel engine

load curves without effect of
turbocharging limitations
Figure 11.2 Propulsion Trials Design Match Point with Power Service Margin MS = 15%
Commercial ship designs typically have a power design margin MD of 3 to 5% depending

upon the risk involved in not achieving the specified power. With explicit estimation of the air
drag of the vessel, a power design margin MD of 3% might be justified for a fairly conventional
hull form using the best parametric resistance prediction methods available today [4]. The power
design margin for Navy vessels usually needs to be larger due to the relatively larger and harder
to estimate appendage drag on these vessels (appendage drag is up to 25% compared with 3-8%
for commercial designs). The U. S. Navy power design margin policy [5] includes a series of
categories through which the design margin decreases as the design becomes better defined and
better methods are used to estimate the required power; i.e.,
category description MD
1a early parametric prediction before the body plan and appendage configuration 10%
1b preliminary design prediction made prior to the model PE test 8%
2 preliminary/contract design after PS test with stock propeller and corrections 6%
3 contract design after PS test with model of actual propeller 2%
11- 5
Commercial designs typically have a power service margin MS of 15 to 25%. In principle,

this should depend upon the dry docking interval; the trade route, with its expected sea and wind
conditions, water temperatures, and hull fouling; and other factors. The idealized power output
of a diesel prime mover varies as N’ = N/No at constant throttle as shown in Fig. 11.2, where N is
the propeller rpm and No is the rated propeller rpm. Thus, diesel plants need a relatively larger
power service margin to ensure that adequate power is available in the worst service conditions.
The service margin might be somewhat smaller with steam or two-shaft gas turbine prime
movers since their power varies as (2 – N’)N’ and is, thus, much less sensitive to propeller rpm.
The power service margin might also be somewhat lower with a controllable pitch propeller
since the pitch can be adjusted to enable the prime mover to develop maximum power under any
service condition. Conventionally powered naval vessels typically have power service margins
of about 15% since the maximum power is being pushed hard to achieve the maximum speed
and it is used only a relatively small amount of the ship’s life. Nuclear powered naval vessels
typically have higher power service margins since they lack the typical fuel capacity constraint
and are, thus, operated more of their life at high powers.
It is important to note that in the margin approach outlined above, the power design margin
MD is defined as a fraction of the resistance or effective power estimate which is increased to
provide the needed margin. The power service margin MS, however, is defined as a fraction of
the MCR that is reduced for the design match point on trials. This difference in the definition of
the basis of the percentage for MD and MS is important. Note that if MS is 20% this will increase
PB in eq. 11.11 by 1/(1 – MS) or 1.25, but if MS were defined the same as MD it would only be
increased by (1 + MS) = 1.20. This potential 5% difference in the sizing of the main machinery
is significant. Practice has been observed in Japan and also occasionally in the UK where both
the power design margin and the power service margin are defined as increases of the smaller
estimates, so precision in contractual definition of the power service margin is particularly
needed when purchasing vessels abroad.
Propulsion Efficiency Estimation
Use of eq. 11.11 to size the prime mover requires the estimation of the six efficiencies in
the denominator. Hydrodynamic resistance and hull-propeller interaction estimation methods
can provide estimates of the hull efficiency Hand the relative rotative efficiency r [4, 6].
11- 6
Estimation of the open water propeller efficiency o in early design can utilize propeller
optimization methods [6]. Guidance for the stern tube and line bearing efficiencies are as
follows [3]:
sb = 0.98, for machinery aft

= 0.97, for machinery amidships (11.12)
SNAME Technical and Research Bulletin No. 3-27 can provide guidance for the transmission
efficiency with mechanical reduction gears [7],
t = g = ∏ (1 – i) where i = 0.010 for each gear reduction (11.13)

i
i = 0.005 for the thrust bearing
i = 0.010 for a reversing gear path
Thus, a single reduction, reversing reduction gear with an internal thrust bearing used in a
medium-speed diesel plant would have a gearing efficiency of about t = 0.975. Note that since
test bed data for low-speed diesels usually does not include a thrust load, t = 0.995 should be
included in direct connected low-speed diesel plants to account for the thrust bearing losses in
service. With electric drive, the transmission efficiency includes the efficiency of the electrical
generation, switchboard, transformer, frequency conversion, electric motor, and gearing (if
installed) [7]; i.e.
t = genSwTrFcmg where gen = electric generator efficiency (11.14)

Sw = switchboard efficiency
Tr = transformer efficiency (if installed)
    Fc = frequency conversion efficiency
    m = electric motor efficiency
g = reduction gear efficiency (eq. 11.13)
The SNAME bulletin [7] includes data for this total transmission efficiency t depending upon
the type of electrical plant utilized. In general, in conventional AC generation/AC motor
electrical systems t varies from about 88 to 93%, in AC/DC systems t varies from about 85 to
90%, and in DC/DC systems t varies from about 80 to 86% each increasing with the rated
power level of the installation. Further, all the bearing and transmission losses tend to increase
as a fraction of the transmitted power as the power drops below the rated condition. Thus, the
percentage loss will be higher at part load than at the design condition.
11- 7
Propulsor Characteristics.
We need to develop and review the basic characteristics of the principal types of marine
propulsors and how these are displayed for design and analysis. We will begin by looking at the
conventional fixed-pitch propeller. The families of these characteristics versus pitch-diameter
ratio P/Dp are also the range of characteristics available with a controllable-reversible pitch
(CRP) propeller. The primary difference is a 1-2% lower open water efficiency since CRP
propellers have a larger hub to accommodate the pitching mechanism and provide adequate
ligament strength between the blades. Waterjets are well established and are becoming even
more important with the interest in high speed vessels. Additional options such as
supercavitating propellers, vertical axis (Voith-Schneider) propellers, nozzle propellers, and
surface piercing propellers have special areas of application.
Fixed-Pitch Propellers. Propellers develop their thrust by developing lift due to their
operation at an angle of attack relative to the incoming flow. It is useful first to review the
incoming flow diagram for a propeller blade section as shown in Figure 11.3 (unfortunately
showing a now uncommon blade shape). The pitch P of the propeller is the distance the
propeller would advance in one revolution if there were no slip between the water and the blade.
Thus, the pitch or pitch angle  = tan–1(P/Dp) can be thought of physically in the same way as
the blade angle of attack with which it correlates. The pitch P is usually non-dimensionalized by
the propeller diameter Dp and used as the pitch-diameter ratio P/Dp. The relative flow into the
propeller blade is a vector combination of the inflow due to the propeller rotation npDp and the
propeller’s speed of advance Va leading to a relative inflow VR at a hydrodynamic pitch angle 
= tan–1(Va/npDp). This gives the blade section a physical angle of attack relative to the inflow
of –  that will produce lift and drag on the “airfoil” blade. These forces can be resolved
into a thrust T and a tangential force Q/Rp that leads to the torque on the propeller. In general,
the pitch is a function of radius since  is varying with r. To provide a single value to describe a
propeller, the value at the 0.7r, where r = Rp = Dp/2, in the region where the propeller develops
its greatest lift is typically used to describe a propeller. From a physical viewpoint, the pitch-
diameter ratio can be thought of as a measure of the angle of attack of the blade section.
11- 8
Increasing pitch increases the loading on the propeller resulting in a greater thrust, more torque,
slower rotation, and closer operation to cavitation limits.
We are interested in the torque-rpm characteristics of propellers and the thrust they will
develop at each speed of advance so that we can establish the load on the propulsion system.
The most useful way to present the “open water” (without the hull) performance of a propeller is
in terms of non-dimensional coefficients KT and KQ defined as follows:
thrust coefficient KT = KT(P/Dp, J) = T/(np2Dp4)

(function of)
torque coefficient KQ = KQ(P/Dp,J) = Q/(np2Dp5)
advance coefficient J = Va/npDp
In English units, these are diameter Dp in ft, Va in ft/s, np = Np/60 in rps, thrust T in lbs, torque Q
in ft lbf, and water density  of 1.989 slugs or lbf s2/ft4 for salt water. In SI units the
corresponding units are m, m/s, rps, N, Nm, and kg/m3, respectively.
Figure 11.3 Flow Diagram Relative to a Blade Section [1]
The open water propeller efficiency is related to these propeller coefficients by,
o = o(P/Dp, J) = thrust power/input power = (TVa/1000)/(2Qnp/1000) = JKT/(2KQ)
Since KT and KQ are functions of J and P/Dp, the open water efficiency o is as well. Typical KT,
KQ, and J characteristics appear as shown in Fig. 11.4 for the Wageningen B-Screw Series that is
11- 9
commonly used for preliminary design of moderately cavitating large propellers. The family of
propellers for various P/Dp shown in Fig. 11.4 is for the B4-70 family meaning that these B-
Screw Series propellers all have Z = 4 blades and an expanded area ratio Ae/Ao = 0.70. The
expanded area Ae is the area of the propeller blades when they are flattened out (see PNA for the
precise definition [3]); Ao is the area of the propeller disk Dp2/4. Figure 11.4 has the blade
section drag scaled to a full-scale Reynolds number of Re0.75R = 3.0 x 106.
Figure 11.4 Characteristics for Wageningen B-Screw Series B4-70 Propellers [9]
11- 10
Some properties of the characteristics shown in Fig. 11.4 should be noted. The KQ is
plotted as 10KQ so it can be presented on the same scale as KT and o. Both KT and KQ are
nearly linear in J; decreasing with increasing J. The thrust KT and open water efficiency o go to
zero together since zero thrust (output) yields zero efficiency. This occurs at a value of J just
above the numerical value of P/Dp (at about J ~ 1.05 P/Dp). Characteristic 10KQ is above KT.
The thrust, peak efficiency, and J at peak efficiency all increase with increasing P/Dp. Propeller
design usually seeks to place the design propeller operating point slightly to the left of (below)
the peak efficiency possible from the propeller. This is done so that if the operating J is missed
slightly on the higher J side in operation, the propeller does not operate on the right side of the
peak where the efficiency declines rapidly toward zero.
The Propeller Law. The Propeller Law is an idealized propeller load that is very useful in
preliminary work when nothing better is known. It is also useful in reasoning about engine-
propeller matching situations prior to obtaining detailed numerical results. It is based upon the
assumptions that,
(1) ship resistance is proportional to speed squared, RT  V2,
(2) the hull-propeller interaction factors (w, t, r) are independent of speed, so that RT  T
and V  Va.
The first assumption implies that the resistance is dominated by frictional resistance so it applies
to displacement hulls when the speed-length ratio Vk/√Lf is less than 1 or the Froude number Fn
is less than 0.3. Using these assumptions and the definitions of the non-dimensional propeller
coefficients we can obtain,
TV2  KTn2 from the KT definition

V2  J2n2 from the J definition
Therefore,
KT/J2 = constant
Looking at the propeller characteristics, KT decreases with increasing J so this result can only be
true if both KT and J are constant along the Propeller Law load curve. So under the assumptions
of the Propeller Law, J, KT, KQ, and o must all be constant for fixed P/DP regardless of V, N, or
n anywhere along any constant RT  V2 curve. This is a powerful result that we can use in many
11- 11
applications. This result yields the following for any constant resistance condition that follows
assumption (1) at constant P/Dp:
NnV propeller rpm is proportional to ship speed
TV n
2 2
Q  V2  n2
PT  V3  n3 and  PS  PB, if the other  in eqn. 11.9 are also independent of speed.
This model breaks down when assumptions (1) and (2) are not valid such as for displacement
hulls above Vk/√Lf ~ 1 where wave resistance becomes large and the proportionality can be as
high as V9; for planing hulls when in the semi- and fully-planing modes; for other non-
displacement hulls such as Surface Effect Ships and hydrofoils; and when the resistance
condition changes due to fouling, towing, Sea State, etc. In the latter case, the condition will still
have T  V2 etc., but the constants of proportionality will be different.
The Propeller Law is a very good model for most displacement hulls of commercial design
since Vk/√Lf < 1. Even when it is approximate, it is still useful for reasoning about engine-
propeller matching situations. Note that even when the Propeller Law is not valid, we can use
the predicted RT, w, t, and r versus V results from power predictions or model tests with the
propeller characteristics to develop comparable Q and P curves versus n or V. The Propeller
Law is, however, very useful for preliminary work, estimates, and design reasoning.
Controllable-Reversible Pitch (CRP) Propellers. CRP propellers offer major

advantages at added cost and weight. Some propellers have variable pitch – even two discrete
pitches in some cases – but they are not reversible, so we will specifically use reversible (R) in
the general designation. The CRP design makes available the full array of characteristics as
shown in Fig. 11.4 plus negative pitch performance since the P/Dp is an additional control
variable that can be actively changed to yield any set available of KT, KQ, o, and J
characteristics. When the Propeller Law applies, there will be different torque Q  V2  n2 and
power P  V3  n3 characteristics for each pitch-diameter ratio.
The CRP propeller offers significant advantages at the added cost and weight of the
mechanism, controls, larger hub that leads to a 1-2% lower open water efficiency, etc. A CRP
propeller can cost the order of a $1-2M more than a conventional propeller for a large ship. It
can, however, allow more optimal operation of the vessel under all service conditions including
11- 12
towing, one engine down, etc.; provide improved stopping and maneuvering; permit the use of
simpler (non-reversing) gearing; enable the use of highly turbocharged engines, and permit
operation at constant rpm to enable the generation of constant frequency power from shaft
generators.
Controllable-reversible pitch propellers use various mechanical designs to accomplish the
variable pitch. In general, commercial CRP propellers have tended to have only Z = 4 blades
which makes the mechanism design and the structural design of the hub easier. Military CRP
propellers are also available in Z = 5 designs and certainly higher blade numbers are feasible, but
more difficult to design and build. Some designs use a hydraulic piston in the hub with oil
provided down the center of the shaft through an oil distribution (OD) box that extends around
the shaft or connects to the forward end of the shaft. Other designs utilize a push rod that
extends down the shaft to move the blade pitch, etc. In all cases, the propellers utilize a hollow
shaft. Figure 11.5 shows an example CRP propeller mechanism. This design has a piston in the
hub that moves the blades through lever arms that are connected to the blade trunions.
Figure 11.5 Typical Controllable-Reversible Pitch Propeller Mechanical Design
Four-Quadrant Characteristics. The propeller characteristics such as shown in Fig. 11.4

apply only to the normal operation of the propeller with the rotation forward and the ship moving
11- 13
forward. This is termed the first quadrant of propeller operation. Design studies of crash stop
and crash ahead maneuvers with a fixed-pitch propeller require that its characteristics be defined
for all four quadrants of operation; i.e.,
hydrodynamic pitch angle
First quadrant – normal ahead operation n > 0, V > 0 0 <  < 90
Second quadrant – maneuvering or crash stop n < 0, V > 0 90 <  < 180
Third quadrant – normal astern operation n < 0, V < 0 -180 <  < -90
Fourth quadrant – maneuvering or crash ahead n > 0, V < 0 -90 <  < 0
The normal presentation using KT, KQ, and J is not useful in this full range of operations since all
three are infinite when the shaft rpm passes through zero. To provide characteristics more useful
for the numerical simulation of maneuvers, a modified set of non-dimensional coefficients is
commonly used, e.g.,
CT* = T/[0.5 (Va2 + (0.7npDp)2)Dp2/4]
CQ* = Q/[0.5 (Va2 + (0.7npDp)2)Dp3/4]
 = tan–1[Va/(0.7npDp)]
When either Va or np pass through zero individually, these coefficients remain finite. Note that
the hydrodynamic pitch angle  is usually taken at the reference 0.7 radius. Four quadrant
characteristics for the Wageningen B-Screw Series B4-70 Propellers are shown in Fig. 11.6.
Propulsion and control studies with CRP propellers require the use of “four quadrant” CRP
propeller characteristics that keep the shaft rotation positive, but vary the ship speed and
propeller pitch from full positive to full negative. The best available systematic model for this
situation was published by Strom-Tejsen and Porter [10] based upon propeller tests conducted by
Gutsche and Schroeder [11]. This model is available at the University of Michigan as a
FORTRAN subroutine that can be used in simulation studies with CRP propellers.
Note that all of the first-quadrant and four-quadrant data typically used for ship
maneuvering simulations, such as shown in Figs. 11.4 and 11.6, are quasi-equilibrium data. This
assumption is usually acceptable for system design studies. In extreme conditions of transient
propeller operation, however, propeller cavitation can lead to a thrust breakdown leading to
characteristics that differ from the quasi-equilibrium results. Detailed design studies may require
explicit correction for transient conditions.
11- 14
Figure 11.6 Four-Quadrant Characteristics for Wageningen B4.70 Propellers [9]
Nozzle Thrusters. Nozzle propulsors involve an airfoil nozzle ring around the propeller that
either accelerates (pumpjets) or decelerates (Kort nozzles) the streamline flow that would result with
11- 15
no nozzle. The pumpjet was evaluated on the U. S. destroyer Witek following World War II with
little success. It is used today in submarines and torpedoes primarily for noise reasons. The Kort
nozzle, named after it inventor, was introduced in 1933 for use in applications requiring high thrust
for the available diameter, such as tugs and fishing vessels and thrusters These are of the greatest
practical use today. The nominal (dashed) contracting flow line and accelerating and decelerating
nozzle flow lines are shown in Fig. 11.7 [adapted from 12]. The design properties of series of Kort
nozzles were studied by van Manen and Oosterveld [12].
Figure 11.7 Flow Lines through Nozzle Propellers [adapted from 12]
Characteristics for nozzle propellers can be displayed in the typical KT, KQ, 0 and J format.
Data for the Ka series Kort nozzle propeller with a 4-70 propeller is shown in Fig. 11.8. This design
is typical with a nozzle length to internal diameter ratio /D = 0.5 and angle  = 9˚. The duct airfoil
ring is oriented so that it develops a lift with a forward component that can cancel the added drag of
the nozzle and provide added thrust. The KT displayed Fig. 11.8 is the net for propeller and the
11- 16
nozzle so this can be used as usual for the entire composite propulsion device. The thrust coefficient
for the nozzle alone is shown as KTN.
Figure 11.8 Characteristics for Ka 4.70 Propeller in Nozzle 19a [12]
Kort nozzle propellers can be used to increase the zero speed bollard pull by as much as 30-
40% with the same diameter propeller. It can also be used in other diameter limited applications.
For the same efficiency, the optimum Ka propeller can be about 30% less in diameter than its B-
Series counterpart. The usefulness of a Kort nozzle has been related to the Taylor propeller
coefficient BP = NPD0.5/VA2.5 where N is propeller rpm, PD is developed power in hpBR, and VA is
speed of advance in knots. It should be considered in single screw vessels if BP > 30 and used in
tugs with BP > 100 [2].
11- 17
Podded Propulsors. The use of podded propulsor has grown in recent years. Often these
use electric motors (often permanent magnet motors to reduce diameter) located in pods below
the hull that drive direct-connected fixed- or variable-pitch propellers. In most cases the pods are
360˚ azimuthing so that the rudder and its appendage drag can be eliminated. The packaged
nature of these units offers major arrangements advantages and installation efficiencies. The
pods can have the propeller in front (pulling) or behind (pushing) or both. A leading 4-27 MW
design is shown in Fig. 11. 9.
Figure 11.9 Cutaway of Variable-Pitch Mermaid Podded Drive [Rolls Royce website]
Typical installations on large cruise ships utilize multiple pulling pods to achieve the
required total power. Figure 11.10 shows the triple ABB Azipod installation on a recent cruise
ship. The two outer pods are steerable and the centerline pod is fixed in this design.
Figure 11.10 Triple Pod Installation on a Cruise Ship [ABB website]
11- 18
Recent designs have introduced the use of a conventional mechanical forward propeller
with a centerline, azimuthing counter-rotating pod as a way to achieve the hydrodynamic
benefits of counter-rotating propellers (to recover the rotational energy normally left in a
propeller wake) without the complications of concentric shafts and seals. A recent design is
shown in Fig. 11.11.
Figure 11.11 Counter Rotating Conventional-Podded Drive [ABB website]
Waterjets. The primary propulsor alternative for high-speed craft is the waterjet, which
becomes competitive with conventional and surface piercing propellers at vessel speeds above
30-40 knots. Waterjets are axial pumps that take water from an inlet in the vessel bottom and
produce a thrust from a jet of water that is emitted from the stern. The jet direction can be
controlled laterally if desired to provide steering; a reverse deflector bucket can be used to
provide reverse thrust. Fig. 11.12 shows a typical cross-section of a marine waterjet.
Waterjet characteristics are presented as the thrust developed at various vessel speeds and
this can be matched to the thrust required by the vessel at each speed. In waterjet design, a
typical thrust deduction for a displacement hull is about t = 0.05 and this can be used to move
from the ship resistance RT to the thrust required from the waterjet. Typical characteristics from
a waterjet are shown schematically in Fig. 11.13 as a function of percent shaft speed (n); also
shown is the resistance curve for a typical high speed planning vessel. In this illustration, the
thrust deduction appears to be taken as 0 since the thrust developed by the waterjet is being
directly compared to the hull resistance rather than Treqd = RT/(1 – t).
11- 19
Figure 11.12 Typical Marine Waterjet [Lips, 1]
Figure 11.13 Typical Thrust Characteristics of a Waterjet [1]
Engine-Propeller Matching.
The matching problem involves picking the drive and load design variables so the trials
design condition equilibrium point is where the designer wants it and then insuring that the off-
design equilibrium conditions will also be acceptable. In general, we have the following
problem:
11- 20
In general the problem is more sensitive with the reciprocating engine since its allowable
operating region in the PB versus Ne plane is much more restrictive, particularly with heavy
turbocharging. The operating region is more forgiving with steam turbines, two-shaft gas
turbines or electric motors. Recall that the reciprocating engine will only develop the full PB
you have paid for when both the rated torque Q and rated rpm Ne are achieved. The marine
engineer, therefore, needs to be sure that the load and engine are properly matched so the engine
will be able to develop its rated power when needed in service.
The Trials Design Point. The usual design point problem is to pitch the propeller so that
the ship makes its design speed on initial trials when using rated Ne and PB(1 – MS) power,
where PB here is the engine MCR or the “project MCR” selected for the design. Recall that this
implies a new hull, propeller, and engine condition; negligible wind and waves; deep water, etc.
The Service Margin MS has been included to ensure that the vessel can tolerate service
conditions with significant waves, wind, fouling, shallow water, etc. without exceeding full
throttle and rated torque on the engine. The trials condition matching problem is shown
schematically in Fig. 11.2. Consider how this design can be approached if the propeller diameter
is limited by the shape of the hull aft and the required propeller-hull clearances set Dp = Dp max
that will usually give the highest propeller efficiency, i.e.,
KT = RT/[(1 – t) • np2Dp4] all known except the gear ratio and np

J = V(1 – w)/npDp all known for the desired design speed V, except np
11- 21
Eliminating the unknown np between these two equations gives,
KT = RT /[(1 – t) • Dp2V2(1 – w)2] • J2 = constant • J2
This required KT curve can be plotted on the family of propeller characteristics for the propellers
of the desired blade number and enough expanded area to satisfy cavitation limits. Each point of
intersection with the KT curves gives the operating point for that pitch-diameter ratio and a
corresponding efficiency at the intersection J. Searching over all possible P/Dp will reveal the
maximum efficiency o and the desired optimum design. The propeller np, can then be obtained
by solving the resulting J for np. This result can set the required gear ratio to utilize rated engine
rpm at this trials equilibrium condition. Such a solution is shown in Fig. 11.14 for the optimum
Wageningen B4-70 propeller when the constant is found numerically to be 0.533 for RT =
1,200,000 N, w = 0.2, t = 0.1,  = 1025 kg/m3, Dp = 6.0 m, and V = 10.29 m/s (Vk = 20 knots).
Figure 11.14 Solution for Optimum B4-70 Propeller Design (without cavitation verification)
11- 22
We would have to verify separately that this expanded area ratio (Ae/Ao = 0.70) would be
acceptable from a cavitation viewpoint. For this example, the KT = 0.533 J2 curve equals the
P/Dp = 1.0 KT curve at J = 0.63. At this J value, the efficiency for this propeller is about 0.615.
Repeating this process for all P/Dp ratios gives a locus of o for all propellers that will produce
the required thrust. The maximum efficiency propeller is found at about P/Dp = 1.10 giving
about o = 0.620. In practical design, this process can be performed using numerical
optimization methods such as the Michigan Propeller Optimization Program (POP) which solves
this problem including a suitable cavitation constraint based on the 5% and 10% Burrill lines [6].
Off-Design Conditions with a Fixed-Pitch Propeller. We will consider a series of

examples to illustrate how these problems can be reasoned and solved.
Example 1. A small single screw service tanker is powered by two ideal diesels that in
the normal load condition will develop a rated power of 1000 kW each at the full throttle
10 bar BMEP and 180 propeller rpm (N). Suppose that in the light load (empty)
condition, the propeller would have to operate at 200 prpm to absorb 2000 kW from the
engines. Assume that the Propeller Law applies. Adding the power produced by the two
parallel engines at each rpm yields,
a. Determine the BMEP and kW on each engine if the engine throttles are reduced
identically and the engines are restricted to the maximum continuous propeller rpm of N
11- 23
= 180 rpm in the light load condition, point a. Note “o” designates a reference value.
Assume the normal service load condition consumes the full engine MCR.
Propeller Law light load condition PB/PBo = (N/No)3 so PB = PBo(N/No)3

= 2000(180/200)3 = 1458 kW
and PB = 729  BMEP • N for each diesel so
BMEP = BMEPo(PBNo/PBoN) = 10 bar (729 • 180)/(1000 • 180) = 7.29 bar
or 72.9% throttle
b. Determine the BMEP, propeller rpm, and kW from a single engine if the other is
declutched in the light load condition. Assume the throttle is increased from part a. to
achieve the highest permissible power.
BMEP = 10 bar PB/PBo’ = N/No’ engine PB/PBo = (N/No)3 propeller load

for the engine PB = 1000(N/180); for the propeller PB = 2000(N/200)3
These must be in equilibrium at point b, so equate and solve for N
1000N = 360,000 • N3/8,000,000 giving N2 = 22,222 N = 149.1 rpm

and PB = 1000(149.1/180) = 828.2 kW
c. Determine the vessel speed on one engine in the normal load condition.
engine PB/PBo’ = N/No’ PB/PBo = (N/No)3 propeller load

PB = 1000(N/180) PB = 2000 (N/180)3
These must be in equilibrium at point c, so equate and solve for N
1000(N/180) = 2000(N/180)3 (N/180)2 = 0.5 N = 127.3 rpm
and since V  N, the speed will be 127.3/180 = 0.707 times the two engine full throttle
speed in the normal load condition.
Example 2. Consider two diesel engines driving a combining gear through torque
converter hydraulic couplings. Show that with only one engine operating, the torque
converter effectively introduces a “variable gear ratio” that enables the operating engine
to develop its full power. Assume that the Propeller Law applies. With two engines
operating, the combination of the reduction gear and the slip in the torque converter give
an effective “gear ratio” r = Ne/Np. Now to find the new Np’/Ne, we know that the
11- 24
propeller load has not changed. Using the Propeller Law then, we know that we are still
operating along the same PE  V3 load curve resulting in constant J, KQ, KT, o, etc.
At the engine, we also know we are giving the torque converter the same input Ne and QB
or KQ. The output of the torque converter will be the pair of Qout and Np that will produce
the needed constant value of KQ. The gear ratio effect on Qout and Np’ must be
considered as well as the losses in Qout to arrive at the Qp and Np. In general, an iterative
solution is required. If, however, we assume ideal characteristics for the single torque
converter; i.e., Q = (Qo/2) • (2 – N/No) if it has maximum efficiency at the design point A
(see Fig. 10-11). In this expression Qo and No are the propeller torque and rpm at point A
with both engines running, we have
KQ = Q/n2D5 = constant so Qo/No2 = Q/N2

where o is point A and (Q, N) are for one engine point C
thus, Qo/No2 = Qo(2 – N/No)/(2N2)

or 2(N/No)2 + N/No – 2 = 0
solving for the one positive solution gives N/No = 0.781
We have effectively introduced an additional “gear ratio” r’ = 1/0.781 = 1.28. The slip in
the torque converter will increase until the needed output torque is achieved increasing
the effective gear ratio by 28%.
11- 25
Solutions using Contours: Suppose a diesel engine is matched at the “corner” or

MCR point of the engine operating envelope for some service condition and we want to
know the operating condition with a +30% increase in resistance at all speeds due to a tow or
a –30% resistance decrease due to a ballast condition. Thus, we would have the following in
terms of the hull PE versus V and the prime mover PB versus Ne planes:
We know Vo, No, Jo, KTo, KQo, To, Qo, PEo, and PBo at point A and we know the required PE
curves for the two new conditions as shown. With ideal engine constraints, we also know the
+30% resistance case will be torque limited at point B and the –30% resistance case will be rpm
limited at point C. With our ideal Propeller Law model, we also know we will have cubic PB 
N3 load curves through B and C. What we do not know is the appropriate constant of
proportionality or constant J for these two new load curves. Therefore, we cannot numerically
obtain the equilibrium in the PB-Ne plane.
We must work for an equilibrium in the PE-V plane to find the appropriate J for these new
conditions, because that is where we know the new loads (±30%). For the +30% case, we do
know that it will occur at some state having the same torque, Qo, as A. What we need to do is
generate a contour of the PE produced by all states having Q = Qo, but N < No. Where this
contour crosses the PE needed curve, yields point B in the PE-V plane. We then have the correct
constant J and o for the new +30% load case equilibrium and can find PB and Ne at point B.
For the –30% case, we know our equilibrium will occur at some state having the same rpm,
No, as A. We, therefore, need to generate a contour of the PE produced by all states having N =
No, but V > Vo. Where this crosses the PE needed curve, we have point C in the PE-V plane. We
then have the correct J and o for the new –30% load case and can find PB and Ne at point C.
11- 26
To generate these contours given Vo, No, Jo, KTo, KQo, To, Qo, PEo, and PBo, we start with
the ratios obtained from the definitions as shown in Table 11.1. This table develops the
relationships needed to develop constant rpm contours, constant torque contours (for
reciprocating engines), and contours of the torque developed by steam turbines and two-shaft gas
turbines. For each, there is a recommended independent variable and a preferred order in which
to calculate the values. We will illustrate the use of this approach in a comprehensive example
below.
Table 11.1 Basic Propeller Characteristics Relations for Constant Pitch P/Dp
Basic Relationship Specialized Form of Relations – Constant P

constant rpm No constant torque Qo Q/Qo = (2 – N/No)
J/Jo = VNo/VoN N/No = 1, J/Jo = V/Vo V/Vo = JN/JoNo V/Vo = JN/JoNo
KQ = KQ(J, P/Dp) not needed KQ = KQ(J) KQ = KQ(J)
KT = KT(J, P/Dp) KT = KT(J) KT = KT(J) KT = KT(J)
KQ/KQo = QNo2/QoN2 [2(N/No)2 –
= PBNo3/PBoN3 not needed N/No = (KQ/KQo)1/2 N/No – KQ/KQo] = 0
KT/KTo = TNo2/ToN2 T/To = KT/KTo T/To = N2KT/No2KTo T/To = N2KT/No2KTo
PE/PEo = TV/ToVo PE/PEo = TV/ToVo PE/PEo = TV/ToVo PE/PEo = TV/ToVo
independent variable V J J
computation order V/Vo, J/Jo, J, KT, KT, KQ, KQ/KQo, KT, KQ, KQ/KQo,
to reach PE vs. V KT/KTo, T/To, KT/KTo, N/No, V/Vo, KT/KTo, N/No, V/Vo,
PE/PEo, PE T/To, PE/PEo, PE T/To, PE/PEo, PE
From the basic relationships shown in Table 11.1, we have six equations in seven unknowns
(for constant pitch P and leaving out PB). Thus, when we assume some relationship for any one
variable, such as constant N or constant Q, we can develop a consistent set of values for the other
six variables (the contour) using any convenient variable as the independent variable.
Example 3. We want to select an engine and reduction gear for a small single-screw
displacement hull in salt water. To reduce the number of options we will use a specific
propeller design which we can assume to be readily available.
11- 27
Hull data: PE @ 13 knots from towing tank data 924 hpBr; wake fraction w = 0.20; thrust
deduction t = 0.15; relative rotative efficiency r = 1.00
Propeller data:
Dp = 10.8 ft.; KT = 0.400 – 0.30J and KQ = 0.0500 – 0.030J for P/Dp = 1.0
Transmission data:
available gear ratios (r) 4.5:1 5.0:1 5.5:1
gear efficiency g = 0.96
cost of gear same for all ratios
stern tube and line bearing efficiency sb = 0.99
Engine data: (similar to GM EMD Series of turbocharged engines)

number of cylinders 8 12 16
rated rpm (Ne) 800 800 800 rpm
rated PB 1530 2260 3060 hpBr
cost of engine increasing 
PB rated point
rated torque
limit engine operating
engine limits
operating assuming
limits
ideal diesel characteristics
rated rpm limit
0 Ne
800
A. The initial objective is to select the engine and reduction gear appropriate for this
application. We are satisfied with the above rated conditions as our Project MCR, but
want a 15% Service Margin when establishing the design power for each engine.
1. First we need to determine the speed range in which we are interested for this ship.
For this example, we will assume that a speed of 15 knots is desired with a speed between
13 and 17 considered satisfactory. For simplicity, we will also assume the resistance of
the ship follows the Propeller Law; i.e., PE = PEo(V/Vo)3. Using this we can develop the
11- 28
PE versus V curve for the ship using PEo = 924 hpBr at Vo = 13 knots as the reference
condition.
V V/Vo (V/Vo)3 PE
13 1.00 1.00 924
14 1.08 1.25 1154
15 1.15 1.54 1419
16 1.23 1.86 1723
17 1.30 2.24 2066
2. Knowing we want an engine/propeller matching at the rated 800 rpm of the engine,
we must pick a reduction gear so that the propeller can provide the needed PE at some
speed. We will assume the desired speed is 15 knots and analyze the three available gear
ratios to determine which will produce a propeller rpm that will produce the needed PE =
1419 hpBr near 15 knots. From the following analysis, it can be seen that the 5:1 gear
ratio gives the PE we need at 15 knots. This particular analysis is for English Engineering
System units. In SI units, the 1.689 ft/s per knot is replaced by 0.5144 m/s per knot, the
550 ft lbf per hpBrs is replaced by 1000 Nm per kWs, and remaining units are consistent.
0.050-0.030J
11- 29
3. We can now plot a constant rpm contour for this propeller being driven by an 800
rpm engine through a 5:1 ratio gear as shown in the following analysis. Plotting this
engine produced PE–V curve with the propeller load PE–V curve, we find the intersection
at 15 knots which is the design matching point B.
4. We can now determine the appropriate engine for this ship from the list available.
Using a 15% Service Margin in establishing the design point power, the required PB will
11- 30
be as calculated below. The 12-cylinder engine should be selected since it produces the
required PB with about a 15% Service Margin.
5. Now assume that the hull resistance increases 40% at all speeds due to fouling while
the ship is in service and that the hull-propeller interaction factors remain constant. We
seek the vessel speed and rpm capability in this condition if the rated torque and rpm of
the 12-cylinder engine are not to be exceeded. First, we can generate a new vessel
resistance curve and plot this on the PE–V plane using PE = PEo(V/Vo)3.
6. This gives the new propeller load curve. The engine will be either torque limited or
rpm limited with this new load. To establish which occurs, we can locate the full rated
point of the engine (2260 hpBr at 800 rpm) in the PE–V plane to establish if the engine
will be torque or rpm limited with the increased resistance. Plotting this point A on the
PE–V plane, we can notice that it falls on the Ne = 800 constant rpm contour as it should.
The rated point is on the left of the new +40% resistance PE curve so the engine will be
rpm limited in this resistance condition. Note that even a 40% resistance increase does
not require the use of the full 15% power service margin. If the rated point were to fall to
the right of the increase resistance PE curve, the engine would be torque limited.
11- 31
7. Knowing now that the engine will be rpm limited, the new operating point will be the
intersection of the new propeller load curve and the already developed constant Ne = 800
contour between the engine design and rated points, point C. This equilibrium is at 13.41
knots. We could check the PB required to produce the required PE at this intersection.
The resulting PB required is less than the rated 2260 hpBr as expected.
KQ = 0.050 – 0.030J = .0311
(0.680)(1.062)(1)(0.99)(0.96)
11- 32
B. Now suppose an accident occurs a few years after this vessel enters service and the
engine must be replaced. The engine designs have been improved slightly increasing the
rated power for the 12-cylinder engine to 2500 hpBr at 800 rpm. The original gear is to be
reused. In order to fully utilize the greater power of the new engine, it is possible to
repitch the existing propeller to P/Dp = 1.1 which will give the new characteristics KT =
0.45 – 0.30 J and KQ = 0.062 – 0.035J.
1. We could begin this section by directly calculating a constant rpm PE contour for the
new engine and pitch ratio as we did in part A.3. To illustrate the use of ratioing various
points along a constant rpm, torque, or power contour relative to a reference point as a
way of saving calculations, we will utilize that approach here. We first locate the new
rated point A’ in the PE–V plane to establish what type of contour will be needed to
obtain the equilibrium points in the normal and +40% resistance conditions and to
provide the needed reference point for the ratioing.
2. This new rated point A’ falls between the +40% resistance propeller load curve and
the normal resistance propeller load curve. The engine will, thus, be rpm limited in the
normal resistance condition and the new equilibrium point B’ can be found by developing
a constant rpm PE contour using the rated point as the reference and assuming speeds
faster that at the rated point. Plotting this contour on the PE–V plane shows an
equilibrium point at 16.25 knots in the normal resistance condition. We can find the
resulting PB to be 2353 hpBr using an approach similar to part A.7.
11- 33
3. In the +40% resistance condition, the engine will be torque limited and the new
equilibrium point C’ can be found by developing a constant torque PE contour using the
rated point as the reference and assuming J values lower than at the rated point. Plotting
this contour on the PE–V plane shows a equilibrium point C’ at 14.3 knots in the +40%
resistance condition. Interpolating J versus V indicates that this equilibrium corresponds
to J = 0.678 allowing the engine rpm to be calculated. The engine can be seen to be just
slightly torque limited in this case.
11- 34
Control of Controllable Reversible Pitch (CRP) Propellers
The use of a CRP propeller offers many advantages such as improved stopping and slow
speed maneuverability, elimination of gear and engine reversing capabilities, and a reduced
number of engine maneuvering cycles. It also offers the possibility of better matching the
propeller load and the prime mover under all operating conditions. This is possible because two
controls are available: engine throttle and propeller pitch. It is usually not wise, however, to
leave their independent selection to the discretion of the operator. The losses possible with poor
operation exceed the gains possible by optimal, coordinated operation. The designer/vendor
usually establishes a fuel-pitch control program so that the operator has single lever (0-10
forward and reverse) control on the bridge. The program defines a correlation between throttle
and pitch as a function of the bridge handle position designed to achieve some particular design
objective.
There are three primary approaches to establishing the pitch for part load operation of a
CRP propeller [13]. First, one could set pitch to give the rpm that will maximize propeller
efficiency at each speed:
Np* = N that yields max{o(N)}|V constant , a propeller only viewpoint
Second, one could set pitch and throttle to give minimum engine specific fuel rate at each power:
Ne* = N that yields max{E(N) = 1/sfr}|PB constant , an engine only viewpoint
Neither of these is really the correct overall system approach; Ne* is, however, closer to the
correct result. The proper overall system goal is the minimum fuel per mile for each speed:
N* = N that yields min{mf}|V constant = max{oE • Hrsbg/PE(V)}|V constant
The expression following the dot (•) is essentially constant for modest changes in pitch and rpm,
so the system view is to maximize the product of o and E = 1/sfr. Usually Ne* < N* < Np*
and, as noted, N* is closer to Ne* than Np*. Correctly setting pitch to yield minimum mf can
save 3-5% over using a constant pitch program. As will be explained below, many designs just
use Ne* and that will give most of this potential improvement.
11- 35
Some designs just use constant pitch control. This will work effectively, but will provide
none of the potential fuel advantage of a CRP propeller over a fixed-pitch propeller in normal
operations.
Some simple designs can use separate setting of pitch and throttle with the operators given
tabular or graphical guidance on how to correlate P and N to yield acceptable fuel rates.
Indication and alarms must be provided, however, to ensure the operator will not overload the
engine (excessive pitch and low rpm can overload the prime mover; excessive Q for a given rpm
setting). The engine room can have an override capability to prevent an overload request by the
bridge command. The prime mover normally has a governor to control rpm with an overspeed
limit feature.
Other applications use constant rpm control when auxiliaries, such as attached generators,
require constant rpm operation. Recall that at low propulsion powers, in particular, the constant
rpm engine may be operated in a poor specific fuel rate region so this approach might be limited
only to higher propulsion powers. The usual approach is to hold throttle constant and actively
control pitch to maintain the constant rpm. Features must be included, however, to ensure that
this controller does not try to overload the prime mover.
As mentioned above, most large applications today utilize single lever bridge control. This
approach is simpler for the operator and ensures that the designer’s objectives (such as minimum
fuel per mile) are met. There are two primary approaches for controlling fuel economy with
CRP propellers [13]. These are shown schematically in Fig. 11.15 and described in the
following:
• Correlation of Speed and Pitch can yield N* (minimum fuel per mile). In this
approach, the control handle sends two signals. One signal sets the propeller pitch. The
other is a correlated signal to the prime mover governor that controls the fuel rate (Q) to
give the correct N. The pitch signal must be overridden, however, if the prime mover
torque Q limit is reached. The problem with this approach is that the N versus P
correlation varies with ship operating condition and the number of engines in operation.
The operator or computer could select from among a finite set of predetermined
programs, but this is an added complication. This is especially true when the program is
implemented mechanically in bridge control pneumatic system cams. This approach is,
thus, best suited for simpler plants with only limited changes in ship operating conditions.
11- 36
• Correlation of Speed and Torque can yield Ne* (minimum sfr, which achieves most of
the desired minimum fuel per mile goal). The control handle sends two signals. One
signal sets prime mover fuel rate (Q), so the prime mover is always protected against over
torque. The other is a correlated signal to the pitch controller that controls pitch to give
the correct Ne. The advantage here is that this program is now independent of ship
operating condition and the number of engines in operation; only one program is needed.
By setting Q and N, this approach guarantees constant KQ operation and this result is
fairly close to the minimum fuel per mile goal (N*) solution. This approach is best suited
for more complicated plants and/or changing ship operating conditions.
Correlation of Speed and Pitch Correlation of Speed and Torque
Figure 11.15 Schematics of Single Lever Bridge Control Strategies [13]
Matching Engines and Waterjets
The matching problem becomes easier with waterjets due to their unique characteristics.
First, the waterjet is a mixed or axial flow pump in which the power consumed is related only to
its impeller design and the shaft rpm cubed. The power consumed is independent of the speed
of the vessel, which is not the case with a propeller. This yields,
waterjet PD = R(N/1000)3 P  N3
propeller PD  N3  V3 if the Propeller Law applies
11- 37
Here R is the impeller rating, the kW absorbed by the impeller at 1000 rpm. If the Propeller
Law does not apply, the power absorbed by the propeller in the steady state is still directly
dependent upon the vessel speed to some power, at least locally. This means that with a waterjet
the engine rpm directly sets the engine’s load, independent of vessel speed, and the jet actually
operates as a dynamometer on the engine. This essentially eliminates any concern about the
engine becoming overloaded by the load at low speeds. With a propeller, the steady engine load
is determined by the propeller resistance curve since this is the power that the propeller must
produce at steady state. With a waterjet, the steady engine load is determined only by the throttle
setting that determines the rpm regardless of vessel speed.
The design matching is established in terms of the thrust required by the vessel; i.e., Treqd =
Rt/(1 – t). The thrust produced by a waterjet is established by the Momentum Equation which
gives T = m(Vout – Vin). This thrust depends upon the vessel speed and the power delivered to
the waterjet regardless of vessel loading, seaway, wind, etc. The power, through the rpm,
determines the mass flow rate m and Vout. The ship speed and inlet wake fraction determine the
Vin = Vs(1 – w). Figure 11.13 shows typical waterjet thrust curves as a function of vessel speed
Vs and impeller percent rpm (n is % shaft speed in this illustration). Notice in this case that the
cavitation problem is present for waterjets at low vessel speed where there is insufficient vessel
speed to imposed the required pressure at the inlet of the impeller to prevent cavitation. This is
in contrast with propellers that have their primary cavitation problems at high vessel speed and
high propeller loading.
Figure 11.16 shows typical thrust capabilities of a waterjet and a fixed-pitch propeller with
the thrust required by a planing boat versus speed. The waterjet has a thrust characteristic
similar to that shown in Fig. 11.13; it rises with decreasing boat speed until it is limited by the
continuous cavitation limitation of the impeller. Two resistance curves are shown and each
exhibits the typical “hump characteristic” as the boat begins to plane. Two curves are shown
because the propeller driven boat will have significant appendage drag due to the shaft, struts,
and rudder that can reach 10-15% or even higher. The waterjet installation will have none of this
appendage drag. The thrust capability of the fixed-pitch propeller is also shown. It is limited by
the high-speed diesel torque limits at low vessel speeds and by engine rpm limits at speeds above
the design point. This means that while an engine driving a propeller cannot develop full power
in the light vessel load condition due to the engine overspeed (rpm) limit, the waterjet can
11- 38
continue to develop full power and, thus, achieve higher boat speeds. This is important on high-
speed ferries that operate with varying loads. This was also very important in the Blue Ribband
speed record breaking trans-Atlantic crossing of the Destriero that was able to greatly increase
its speed as it burned out its large fuel load. The large difference between the thrust required and
the thrust available in the hump region means that the waterjet can accelerate the vessel much
faster below 20 knots.
Figure 11.16 Comparison of Thrust with Waterjet and Fixed Pitch Propeller [14]
The comparison of the efficiency of waterjet and propeller propulsion needs to recognize
the significant difference in the appendage drag between the two options. Alexander [14]
defines a propulsion efficiency PC as follows:
PC = bare hull resistance x boat speed/engine power

He argues that this is a more appropriate measure, than the usual use of D = RTV/1000PD, when
considering waterjets since the total resistance RT includes a significant contribution from the
appendage drag that is due directly to the use of a propeller (from the shafts, struts, and rudders
that are eliminated with the waterjet). Figure 11.17 shows typical values for this PC for selected
propellers and waterjets and for 25% larger waterjets and 25% larger diameter propellers for a
particular small two-shaft planing vessel. The selected waterjets can be seen to be superior to the
propellers above a speed of about 33 knots and this advantage grows with speed. The larger
waterjets and propellers improve the PC and the crossover speed at which the waterjets are
superior then drops to about 24 knots. It should also be noted that since many waterjets are
11- 39
directly connected to their prime mover, the gearing losses not included in PC can also be
eliminated.
Figure 11.17 Propulsive Coefficient without Appendage Drag Comparison [14]
Blount gives the following ranges for the wake fraction and thrust deduction with flush
inlet waterjets [15]:
displacement hulls FnV/(g≤ 1 0.00 ≤ w≤ 0.02 0.00 ≤ t≤ 0.08

semi-planing hulls 1 ≤ Fn≤ 2.5 0.02 ≤ w≤ 0.04 t = 0.05
planing hulls 2.5 < Fn w = 0.05 – 0.07 ≤ t≤ – 0.02
The thrust deduction can be used to obtain the required thrust from the total resistance for use in
matching a waterjet with a hull. The wake fraction relates the duct inlet velocity to the hull
velocity. For a more rigorous treatment of the waterjet thrust deduction, considering also the
effects of the jet on sinkage and trim, refer to the recent work in the Journal of Ship Research
[16].
Blount also gives the comparison shown in Fig. 11.18. This shows his view of the range of
achievable overall propulsive coefficients with flush inlet waterjets, surface piercing propellers
and conventional submerged propellers as a function of vessel speed. The waterjet can be seen
to be superior above about 30 knots so waterjets are of great interest in high-speed ferries and
high-speed naval applications. A second, somewhat conflicting, view of achievable propulsor
only efficiencies is shown in Fig. 11.19 [17].
11- 40
Alison gives a comprehensive and detailed analysis of waterjet propulsion for fast vessels
[18]. This analysis goes into much greater detail relative the waterjet design, cavitation, and
efficiency than is feasible here. We will consider some more details of the waterjet pump design
in the next chapter.
Figure 11.18 Comparison of Achievable Overall Propulsive Coefficients [15]
Figure 11.19 Comparison of Achievable Propulsor Only Efficiencies [17]
11- 41
References
2. van Manen, J. D. and van Oossanen, P., “Propulsion,” in Principles of Naval Architects,
Vol. II, SNAME, Jersey City, NJ, 1988.
4. Parsons, M. G., “Parametric Design,” Ch. 11 in Ship Design and Construction, Vol. I,
SNAME, Jersey City, NJ, 2003.
5. NAVSEA Design Data Sheet DDS 051-1, 1984.
6. Parsons, M. G., Li, J., and Singer, D. J., "Michigan Conceptual Ship Design Software
Environment User's Manual," University of Michigan, Department of Naval Architecture and
Marine Engineering, Report No. 338, July 1998.
7. Parsons, M. G., Singer, D. J. and Denomy, S., “Integrated Electric Plants in Future Great
Lakes Self-Unloaders,” Journal of Ship Production and Design, 27-4, November, 2011,
pp. 169-185.
8. “Marine Diesel Power Plant Performance Practices,” SNAME T&R Bulletin No. 3-27,
1976.
9. van Lammeren, W. P. A., van Manen, J. D., and Oosterveld, M. W. C., “The Wageningen
B-Screw Series,” Transactions SNAME, Vol. 77, 1969.
10. Strom-Tejsen, J. and Porter, R. R, “Prediction of Controllable-Pitch Propeller Performance
in Off-Design Conditions,” Paper VII B-1, Third Ship Control Systems Symposium, Bath,
UK, 1973.
11. Gutsche, F. and Schroeder, G., “Freifahrversuche an Propellern mit Festen und
Verstellbaren Flugeln ‘Voraus’ und ‘Zuruch’,” Schiffbauforschung, Vol. 2, No. 4, 1963.
12. van Manen, J. D. and Oosterveld, M. W. C., “Analysis of Ducted Propeller Design,”
Transactions SNAME, Vol. 74, 1966.
13. Schanz, F., “The Controllable Pitch Propeller as an Integral Part of the Ship’s Propulsion
System,” Transactions of SNAME, Vol. 75, 1967.
14. Alexander, K., “Waterjet versus Propeller Engine Matching Characteristics,” Naval
Engineers Journal, May, 1995.
15. Blount, D. L., “Achievements with Advanced Craft,” Naval Engineers Journal, September,
1994.
16. Eslamdoost, A., Larsson, L. and Bensow, R., “Waterjet Propulsion and Thrust Deduction,”
Journal of Ship Research, Vol. 58, No. 4, 2014.
17. Eames, M. C., “Advances in Naval Architecture of Future Surface Warships,” Transactions
RINA, Vol. 123, 1981
18. Allison, J., “Marine Waterjet Propulsion,” Transactions SNAME, Vol. 101, 1993.
11- 42
Chapter 12 Machinery Vibration, Excitation and Mitigation
Ship vibrations can be hull and structural or machinery vibrations. Vibration is the
oscillatory motion that can occur in all bodies with mass and flexibility. Ship hull vibration is
beam bending vibration of the entire hull: vertical, lateral or torsion. General structural vibrations
involve the local vibration of panels, deck houses, masts, etc. Machinery vibration can be the
propulsion shafting vibration or the vibration of and within individual pieces of equipment.
Propulsion shafting vibration is usually considered in three modes: torsional, longitudinal, or
lateral:
Individual machinery components also experience vibration due to the general excitation of the
space or its own lack of balance or deterioration. Vibration monitoring is used today to assess the
condition of most rotating equipment. Long’s chapter in Marine Engineering gives a dated but
reasonably good overview of shafting vibration [1].
Excitation is the force or moment that causes a transient or steady vibration. Shafting
vibration is excited by the following:
 Propeller: As the propeller moves through the circumferentially varying wake field, it
experiences an oscillating angle of attack and this produces an oscillatory lift on each
blade. This produces an oscillatory thrust and torque and oscillatory lateral forces and
moments. Mass unbalance and gyroscopic excitation of the propeller can also excite
lateral vibrations.
 Reciprocating engines: The periodic cylinder pressure and reciprocating motion of the
piston, piston rods, and connecting rods yield oscillatory torque and to a lesser extent an
oscillatory thrust on the crank shaft of a reciprocating engine.
 Seaway: Motions in a seaway produce thrust and torque excitation on the propeller due to
its influence on the inflow to the propeller.
 Ice: Propeller milling of ice blocks can produce large short term or transient excitation.
12-1
Marine propulsion plant vibrations lead to the following concerns:

 Motion: Motion can produce wear in couplings, etc. and fatigue in operators. Human
fatigue is a factor in vibration limit specifications.
 Alternating stresses: Alternating stresses can lead to a fatigue failure in any structure.
They are one of the principal factors in torsion shafting vibrations.
 Unloading (backlash) of gear teeth or thrust bearings: If the alternating component of
torque (or thrust) exceeds the steady component of torque (or thrust), gear teeth (or the
thrust bearing) will be unloaded and will chatter back and forth between the two faces.
 Excitation of adjacent machinery and structure: Even though a particular level of vibratory
motion may not harm the shafting, it can pump large amounts of vibratory energy into the
ship and this can lead to excessive motion or failure of other machinery or structure. This
is usually the limiting consideration in longitudinal vibration.
We will begin with a brief review of the core of a first course in mechanical vibrations.
12.1 Mechanical Vibration Review
The theory of linear vibration [2, 3, 4] is well developed and it is usually acceptable for the
small displacement vibration occurring in marine machinery. This approach assumes that there
are linear springs or stiffness (Hooke’s law – force proportional to displacement) and linear
damping (viscous or equivalent viscous damping - force proportional to velocity). In linear
vibration, superposition holds. If excitation Q1 yields vibration 1 and excitation Q2 yields
vibration 2, then excitation (aQ1 + bQ2) produces vibration (a1 + b2).
Vibratory motion can be divided into two types, free vibration and forced vibration:
 Free vibration. Free vibration is the transient vibration that occurs due to initial conditions.
This will occur at one or more natural frequency of the system. These natural
frequencies are fundamental physical properties of the system. Transient vibration is
usually not of great interest in marine machinery – the exception being operations in ice.
They usually die out quickly due to energy dissipation or damping in the system.
 Forced vibration. Forced vibration is the steady-state vibration that usually occurs as the
result of external excitation. This will occur only at the excitation frequency in linear
vibration. If the excitation frequency is near any natural frequency of the system,
resonance occurs and dangerously large amplitudes can occur. Damping, which
dissipates energy in friction, propeller motion, flexure of structure, etc., can limit the
resonance amplitude to acceptable levels.
12-2
Engineering systems usually have small damping. We, therefore, can often ignore damping at first
to evaluate (estimate) the natural frequencies and mode shapes of the system. If forced vibration
results are then needed near any resonance, it is then necessary to estimate the damping.
Periodic motion is motion that repeats in equal time intervals called the period . Otherwise
the motion is aperiodic. In periodic motion, x(t + ) = x (t) for all t, with period  [sec/cycle] and
frequency f = 1/ [cycles/sec = Hz, Hertz]. Harmonic motion is motion that can be represented
by a sine or cosine function. Thus, x(t) = X1cost + X2sint = Xcos(t – ), with single amplitude
X = (X12 + X22)1/2, phase angle = tan-1(X2/X1), and circular frequency  = 2f [rad/sec]. With
displacement x(t) = Xcost, velocity is dx/dt = - Xsint = Xcos(wt + /2), and acceleration is
d2x/dt2 = - X2cost = X2cos(t + ) and as expected physically, the acceleration leads and
produces the velocity that leads to the displacement. A harmonic function can also be represented
using a complex or phasor notation:
Fourier or harmonic analysis can be used to represent any periodic motion or excitation by
a series of harmonic components. If the system is linear, superposition holds and we can
decompose a periodic excitation into its harmonic components, establish the response to each of
these components separately, and then superpose (add) the responses to obtain the total response.
A Fourier series representation of a periodic function uses,
12-3
where
We will leave the derivation for the text books and use software when needed to decompose
signals. We can also use an equivalent complex Fourier series; i.e.,
An infinite Fourier series is needed to “exactly” represent a general periodic function. In practice,
a finite series is adequate. When a function is close to harmonic only a few terms may be needed.
The following sketches illustrate two marine machinery examples:
Note that if a function is even (symmetrical about 0, f(t) = f(-t)), it will contain only cosine terms
(all bn = 0); if it is odd (anti-symmetrical about 0, f’(t) = -f’(-t) after the a0/2 mean value term is
removed), it will contain only sine terms (all an = 0).
12.1.1 One-Degree of Freedom Systems
Recall that a degree-of-freedom is a coordinate needed to describe the motion of an object.

Thus, a particle has three (x,y,z) degrees-of-freedom in translation. Any real, flexible body has an
12-4
infinite number of degrees of freedom. In most practical problems, we can consider parts of the
system rigid with respect to the rest allowing the consideration of a small, finite number of degrees
of freedom. Similarly, it is often acceptable to consider parts of the system massless with respect
to the rest allowing the use of concentrated masses and massless springs.
Free vibrations. Many practical problems can be modeled as single-degree-of-freedom
systems. Further, the fundamental character of all vibration is shown by the study of the one
degree-of-freedom system. Consider a concentrated mass and a massless spring and apply
Newton’s law,
This result is generally true. If we use the static equilibrium (or steady motion) condition to define
our perturbation vibration coordinate x, we can neglect all forces and motions that makeup the
static equilibrium (or steady state motion) condition. Here we use x defined from the static
equilibrium deflection and the weight W and static deflection  do not appear when we apply
Newton’s law in the perturbation coordinate. We just have,
12-5
Now we can add damping that is a model for any energy dissipating influence on the system. The
most common model is linear or viscous damping modeled as a dashpot with the damping force
proportional to the velocity. Consider,
The free vibration is the homogeneous solution to this differential equation (F(t) = 0). We assume,
In general B ≠ 0 and this must be true for all t so we must have the characteristic equation,
The critical damping cc = 2(mk)1/2 = 2mn is the value that makes the second ± term zero. The
solution is the superposition of the solutions for the two roots, i.e.
12-6
The solution character depends on the value of the damping factor . There are three cases:
Or using Euler’s formula to express things in terms of sine and cosine and recognizing that the
coefficients Ai must be real,
which gives a damped sinusoidal motion at the damped natural frequency,
Forced vibrations. We can now add harmonic excitation in a general form with a phase
angle  relative to t = 0, which could just be one component of a harmonic analysis of a general
periodic excitation. The effect of the excitation phase shift is just a shift in the zero point on the
time scale so the linear response will have the same phase shift.
12-7
The differential equation is now,
The solution has the homogeneous (transient) or complementary solution xc(t), already found
above, plus a particular or steady-state solution at excitation frequency  of the form:
Differentiating the particular solution and substituting this into the differential equation yields the
solution where the amplitude A = X can be non-dimensionalized by the static solution F/k to yield
a magnification factor and forced response phase angle:
The peak or maximum (resonance) response will occur at the freqeuency that maximizes X,
Note that for small damping p < d < n converge, but they are different. The complete solution
is the sum of the complementary xc(t) and particular solution xp(t), i.e.,
12-8
Note that when the initial conditions are used to obtain B1 and B2, the entire expression must be
used. Analyze the forced (steady-state) part first to obtain X, , and ; then use the entire
expression and the initial conditions to obtain B1 and B2.
Figure 12.1 summarizes the general results for the particular solution xp(t) of the force
excitation case and Fig. 12.2 summarizes the general results for the base motion excitation and
rotating unbalance excitation cases. These can be used with the homogeneous solution xc(t). With
these general results available, there is no need to solve the equations again. Just use these results
with n = (k/m)1/2 and  = c/cc and cc = 2mn.
Figure 12.1 General One-dimensional Linear Vibration Results for Force Excitation
12-9
Figure 12.2 General One-dimensional Linear Vibration Results for Base Motion and
Rotational Unbalance Excitation
12-10
12.1.2 Multi-degree-of-freedom systems.
Most marine engineering vibration problems must be modeled as multi-degree-of-freedom

systems. We will review the solution for a two-mass system to illustrate the fundamental
characteristics of all of these situations:
 an n degree-of-freedom system will have n natural frequencies ni;
 free vibration at each ni yields a characteristic relationship among the coordinate
motions called an associated principal mode or mode shape i;
 free vibration is superposition of motion in the n-principal modes each at its own ni ;
 forced vibration will be at the excitation frequency and resonance occurs at each ni.
Consider a general series two-mass system:
We can neglect damping and determine the undamped natural frequencies by assuming free
vibrations (F1 = F2 = 0) and harmonic motion xi = Aisint, differentiating and substituting into
the equations of motion. The nontrivial solution (Ai ≠ 0) yields the characteristic equation,
12-11
To continue, we will consider a specific numerical case with m2 = 2m1 = m and all ki = k to
solve for the natural frequencies and mode shapes. The characteristic equation will be quadratic
in 2 yielding two natural frequencies i or for this example,
To obtain the amplitude ratios or mode shapes i, we can use either equation of motion to give,
Now use 12 and 22 to obtain 1 and 2, respectively,
Thus, free vibration will occur at 1 with x1 always 0.732x2 and at 2 with x1 always -2.732x2.
The motion in mode shape 2 at 2 will be at about twice the frequency of the motion in 1 at
1for this numerical example. The general solution for free vibrations will be superposition of
the two modes with the amplitudes and phase shifts determined by the initial conditions, i.e.,
12-12
Note that the each higher frequency mode shape has one more node (location of zero motion in
that mode) than the next lower frequency.
If this system were excited by a force F1sint on body 1 only (F2 = 0), the forced
vibrations particular solution for this numerical example would be,
The basic characteristics of this motion apply to all multi-degree-of-freedom vibrations:

 forced oscillation is at the excitation frequency ;
 resonance occurs in each i response when  = i;
 a superposition of motion in the i exists (provided damping is proportional to the mass
and/or the stiffness)
If we take a general matrix approach to forced vibrations of a multi-degree-of-freedom

system, we have:
12-13
The complex matrix in the brackets is called the dynamic matrix D and the general damped forced
vibration solution will be as follows:
To solve this general problem, all we need is computer software that will invert the complex
dynamic matrix D() or solve the system of complex linear equations as represented by (1) above.
MATLAB can solve a complex system of equations Ax = b using left division x = A\b or inversion
of matrix A using inv(A) and solving for x = inv(A)*b. In general, the left division operation is
the faster and more stable option. Recall that the characteristic equation for an n discrete
mass/coordinate system will be a polynomial that contains the eigenvalue  = 2 to the nth power.
For an axial or lateral vibration problem that has the system restrained by a spring to ground
representing the thrust bearing and its foundation (axial) or shaft bearings and their foundations
(lateral), this polynomial will be of the form (a2n + b2(n-1) + …+c) = 0 and there will be n non-
zero eigenvalues 2. For a torsional problem, which can have free rotation, this polynomial will
be of the form 2(a2(n-1) + b2(n-2) + … +c) = 0 with one root 2 = 0 (period = ∞ with no node
in the associated mode shape) corresponding to the free rotation. In this case, the dynamic matrix
D() will be singular and its inverse will not exist; it will, therefore, be necessary to use the left
division operator within MATLAB for torsional problems to solve Ax = b.
The eigenvector matrix (mode shapes) and eigenvalue matrix E (natural frequencies) of the
system Ax = 0 can be obtained in MATLAB using the command [V, E] = eig(A). The mode shape
vectors appear as the columns of V = [ 1, 2,…] and E is a diagonal matrix with the eigenvalues
i = ni2 down the diagonal in ascending order. The first ni is associated with 1; the second ni
is associated with 2; etc.
12.2 Shafting Vibration Modeling
Practical modeling of a propulsion shafting system for torsional vibration can utilize the
assumption that there are discrete masses and massless springs. This is true because the mass
moment of inertia (the “mass” in rotation) depends upon R4 where R is the body radius. Thus,
turbine rotors, propellers, and reduction gear pinion and gear sets appear as rigid masses relative
12-14
to the shafting that can be considered as massless springs. The impact of the gearing also has to
be modified since the gearing changes torques by the gear ratio and the effective stiffnesses and
masses are modified by the gear ratio squared at each gear set. Figure 12.3 shows the modeling
used to analyze the torsional vibration of a containership propulsion shafting system consisting of
a single screw, a double reduction gear and an HP and LP turbine. This might also be one of the
two gas turbine per shaft military system on a twin-screw DDG51 class destroyer. The system can
be effectively modeled using six degrees-of-freedom. There is torsional damping at the propeller
and the two turbine rotors and structural damping from the twisting of the shafting elements. The
pinion/gear sets are modeled by single equivalent mass moments of inertia. Thus, the complex
dynamic matrix analysis D() can have dimension six.
Figure 12.3 Example Turbine System Torsional Modeling [5]
In axial vibration, the mass depends upon R2 and the mass of the shafting becomes significant
relative to the other masses. Because the analysis requires only the accurate prediction of the
lowest few natural frequencies, it is acceptable to model the distributed shafting by a modest
number of discrete masses that have the same total mass and center of gravity as the actual shafting
segments. Figure 12.4 shows the axial vibration discrete mass model for the same containership
12-15
as considered in Fig. 12.3. Because the axial forces are transmitted by the thrust collar into the
thrust bearing and then down into its foundation and the ship structure it is necessary to model the
thrust bearing, its foundation and a portion of the ship structure below the bearing. Since a dental
coupling is used in introduce axial freedom between the first and second reductions of the gear, it
is only necessary to model the second reduction pinion/gear set. Effective engineering predictions
of the lowest few natural frequencies can be obtain using seven discrete masses to model the
propulsion shafting leading to a ten degree-of-freedom model.
Figure 12.4 Example Turbine System Axial Modeling [5]
In reciprocating engine propulsion plants torsional vibration is of primary concern and the
system can be modeled with a discrete mass moment of inertia for the flywheel and each cylinder
set of the engine. Figure 12.5 shows the discrete mass modeling used to analyze a single-screw,
low-speed diesel propelled handy-sized bulk carrier. Again in torsion, the mass of the shafting
12-16
and the crankshaft journals can be considered as massless relative to the remaining parts of the
system. There is structural damping and damping at the propeller and cylinder sets. The shaft
couplings are modeled as discrete mass moments of inertia and a total of eleven masses are needed
to effectively predict the two lowest natural frequencies that are relevant in design.
Figure 12.5 Example Diesel System Torsional Modeling [6]
12.3 Machinery Vibration Excitation
We will introduce the two principal sources of machinery vibratory excitation: the propeller
and a reciprocating engine, if installed. In order to do the analyses, we also need the mass/inertia
and hydrodynamic characteristics of the propeller and the mass/inertia characteristics of the
engine.
12-17
Propeller characteristics. A general coordinate system for the dynamics of a propeller is

shown in Fig. 12.6.
Figure 12.6 Propeller Coordinate System [7]
There are six degrees-of-freedom in general and the propeller can be considered to be a rigid
body relative to the flexible shafting system. The equations-of-motion for the vibrating propeller
can be written in general as,
with m the physical mass, J the rotational mass moment of inertia, and Jd the diametral mass
moment of inertia about a diameter in y or z completing the mass matrix M. Spreadsheets are
available of obtaining these properties. The first two coordinates are axial x and torsional x and
the last four are the translation (y,z) and rotation (y,z) in the lateral plane mode. Force fh is the
hydrodynamic force on the propeller due to its perturbation velocity and acceleration, which for a
fully-immersed propeller can be expressed as,
where Ma is the added mass matrix and Cp is the propeller damping matrix. This expression gives,
12-18
where fs are the structural forces and moments that attach the propeller to the rest of the system
and fe are the excitation forces and moments acting on the propeller. It can be shown theoretically
that the propeller added mass matrix Ma has the form [7, 8],
where the zeros indicate that there is no inertia coupling between the axial/torsional modes and the
lateral mode. Coefficient m21 is the propeller inertial coupling from torsion to axial motion. There
is symmetry (m33 = m55) and reciprocity (m21 = m12), but matrix Ma is not symmetrical
(m35 = - m53) due to the handedness of the propeller. Theory also shows that the propeller damping
matrix Cp has the same form. Software and regression equations for the terms in the propeller
added mass and damping matrices as a function of propeller characteristics can be obtained from
our earlier work [7, 8]. The preliminary design regression equations were included in Carlton’s
excellent book, unfortunately with a few typographical errors [9]. Data given in Marine
Engineering [1] contains potentially serious errors.
Without either inertial or damping coupling between the axial/torsional motion and the lateral
motion, the treatment of lateral vibrations as an uncoupled mode is justified. Traditionally, the
axial and torsional modes have also been treated as decoupled. It has been recognized for some
time, however, although it is not recognized by Marine Engineering or even the American Bureau
of Shipping Rules until recently, that the axial and torsional modes are coupled by the propeller
characteristics. If the propeller vibrates axially, it experience a variable inflow which creates a
variable blade angle of attack and a variable lift that can be resolved into both an axial force fx and
a torsional moment qx. The coupling effects can become significant if a natural frequency of axial
vibration is close to a torsional natural frequency. With the coupling characteristics m21 and c21
available, the coupled problem can be treated if needed [5].
Propeller excitation. The treatment included in Marine Engineering [1] is worth review.
When the propeller operates in a circumferentially varying wake field Va = V(1 – w(r,)) due to
12-19
the hull forward of the propeller, the propeller experiences a periodic change in the blade angle of
attack as illustrated in Fig. 12.7 for a single screw propeller. The inflow is harmonically analyzed
to obtain the Fourier coefficients for the x, y, and z components of the flow. There is also a radial
variation of the average wake that has important influence and implications for excitation
mitigation. There is a fundamental difference between the axial wake field behind a single screw
vessel and a twin screw vessel. A single screw wake typically has a high wake region at the top
and, to a lesser extent, at the bottom of the disk due to the hull and skeg forward of the propeller
resulting in larger even harmonics of the wake. The twin screw wake has a lower high wake region
when the blade passes through/near the hull boundary layer and then it is low around the rest of
the disk resulting in larger odd harmonics of the wake.
Figure 12.7 Typical Variations of Advance Angle of a Blade Section in One Revolution on a
Single-Screw Vessel [1]
A theoretical analysis of the propeller excitation [7, 10] shows the following:
 hydrodynamic vibratory forces and moments occur only at the blade rate frequency and its
harmonics, i.eZ where Z is the blade number,  is the rotation frequency, and the
harmonic number is  = 1, 2, 3, ….;
 vibratory thrust and torque are produced by only the Z harmonics of the wake;
 lateral vibratory forces and moments are produced by only the Z ± 1 wake harmonics;
 excitation due to propeller imbalance or blade damage occurs only at the rotation frequency;
 propeller skew can be utilized to significantly reduce any selected component(s) of
vibratory forces and moments [11].
12-20
The general form and harmonic characteristics of the single-screw and twin-screw wakes
and the second and third bullets above lead to the following general design observations:
 thrust and torque excitation is greatest behind single-screw hulls and lowest behind twin-
screw hulls with even number bladed propellers;
 lateral excitation is greatest behind twin-screw hulls and lowest behind single-screw hulls
with odd number bladed propellers.
Typical results for the torque, thrust, vertical bending moment and horizontal bending moment
on a single-screw vessel versus blade number are shown in Fig. 12.8. A typical torque harmonic
analysis from a Z = 4 propeller operating behind a single-screw ship produced,
Q(t) = 86.48 + 5.41cos(4t – 14.4˚) + 1.06cos(8t – 83.3˚) + 0.02cos(12t – 0.2˚) … ft LT
Note that the blade rate excitation is r = 5.41/86.48 = 0.0626 = 6.26% of the steady torque and
this dominates the excitation with the twice blade rate excitation down to 1.23% and the third
0.02%.
Figure 12.8 Typical Single-Screw Propeller Alternating Torque, Thrust and Moments [1]
12-21
Reciprocating Engine Excitation. Torsional vibration analyses for plants with

reciprocating engines are typically provided by the engine vendor. The marine engineer needs to
understand enough about the character of this vibration to interpret these reports. Excitation is
created by the reciprocating engine parts that impart an unsteady torque on the crankshaft at its
rotation frequency and the first few harmonics. The variable torque produced by gas pressure
within the cylinders occurs at the rotation frequency and its harmonics in a two-stroke cycle
engine, but at twice the rotation frequency and its harmonics in a four-stroke cycle engine. Its
character also depends upon the firing order of the cylinders. One of the best teaching
illustrations of this situation comes from the classic MIT mechanical engineering text by Den
Hartog [12].
The cylinder set in a reciprocating engine has complex motion involving pure rotational
(crank throw), pure translation (crosshead, piston rod, and piston), and combined translation and
rotation (connecting rod). To model this situation in an idealized torsional model, each cylinder
set is represented by an equivalent mass moment of inertia Je that has the same average kinetic
energy per cycle as the complex cylinder set. The time varying error in this average model is
then represented by a Fourier series for the time varying inertial self-excitation torque that the set
applies to the crankshaft [12]. This takes the form,
QI = a sint – b sin2t – c sin3t + …
This is an inertial vibratory excitation that will vibrate the engine even if the fuel is turned off.
The engine is also subject to a periodic gas pressure torque due to the combustion of the
fuel. The first three terms of the gas pressure torque is illustrated for a four-stroke cycle engine
in Fig. 12.9 [12]. Note that these first 3 harmonics have a magnitude more than twice the
average full load torque per cylinder. This occurs because the signal is very unlike a sine wave
and many terms are needed in the Fourier series to effectively represent this history. A typical
gas pressure torque at reduced power and no load (just air pumping) are also shown. An
example harmonic analysis for a four-stroke cycle engine normalized by the mean torque Q0 has
coefficients as follows:
Harmonic n 1 2 3 4 5 6 … 16 17 18
Order number k ½ 1 1½ 2 2½ 3 8 8½ 9
Qk/Q0 2.16 2.32 2.23 1.91 1.57 1.28 0.08 0.05 0.04
12-22
The order number k is the number of excitation cycles per engine rotation that will be k = ½, 1,
1½, 2, 2 ½, … for a four stroke engine and k = 1, 2, 3, … for a two-stroke engine. Note that it is
not until the 16 harmonic that the magnitude get down to that typically experienced in propeller
excitation. Many harmonics are needed to describe the gas pressure torque rather than a just few.
Figure 12.9 Four-Stroke Engine Gas Pressure Torque and First Three Harmonics [12]
The gas pressure cycle will require one revolution in a two-stroke cycle engine and two
revolutions in a four-stroke cycle engine. The gas torque excitation frequencies ee will be at,
ee = 2kNe/60 where k = 1, 2, 3, … two-stroke

k = ½, 1, 1½, 2, … four-stroke
and these harmonics can be in resonance with any natural frequency  at the engine rpm Ne,
Ne = 60/2k critical rpm
The resonance conditions in reciprocating engines (and other marine vibrations situations)
are often called critical frequencies or criticals. A common notation for reciprocating engine
torsional criticals is the following:
12-23
e.g. 3-II or II – 3 or II/3 where II is the number of nodes in the mode shape
of natural frequency  , and
3 is the engine gas torque order k in resonance with 
The critical engine rpm Ne must be within the operating range of the engine to be of practical
interest.
A meaningful vibratory response requires that there be resonance and then also that there is
significant net excitation in that critical. This depends upon the number of cylinders and the
firing order of the cylinders since the input from the n cylinders can be out of phase and cancel
or they can be in-phase and add. The firing order is set by the mechanical configuration of the
crankshaft since we know that a cylinder must fire when it is near top dead center (TDC). Figure
12.10 shows a crankshaft for an in-line 8-cylinder four-stroke engine with firing order 1-3-5-7-8-
6-4-2. Each cylinder fires as it reaches TDC during the correct rotation of the shaft.
Figure 12.10 Crankshaft for a Four-Stroke Engine with Firing Order 1-3-5-7-8-6-4-2
The effects of the cylinder phasing can be treated using phasors or illustrated graphically
using phase diagrams. We can first consider the phase diagram by defining the following phase
angle from the input from cylinder 1 to the input of cylinder i in harmonic order k,
ki = ki where i = crank angle from reference cylinder 1 to cylinder i
k = order number under consideration
This depends upon the firing-order as determined by the crankshaft design.
12-24
Consider the example for the engine shown in Fig. 12.10 that shows the crank diagram (end
view of the crankshaft) and the phase diagrams for various k. Note that the patterns repeat.
Some orders can be seen to tend to be additive, called major orders, and others tend to cancel
(minor orders and half major orders), but this is not the whole question.
If we analyze the vibratory energy input to the engine [12], it will have the form:
Eeek = Qk’|i exp(jki)| 1 where the |…| is a complex vector magnitude

1 = vibratory magnitude of cylinder 1
Qk’= magnitude of input torque in order k
The magnitude of the excitation energy input, therefore, depends upon the magnitude of complex
vector magnitude that depends upon both the mode shape i from the critical mode constructed
with the magnitude of cylinder 1 as 1 and the cylinder phasing relative to cylinder 1 ki due to the
firing order and resultant crankshaft mechanical design; depends upon the excitation per cylinder
Qk’ in the critical under consideration; and depends upon the resulting magnitude of the vibratory
oscillation of cylinder 1 1. The vibratory magnitude, therefore, depends upon the complex vector
magnitude |…| that can be shown graphically in the star diagram. Its elements are the magnitude
of cylinder i in the mode shape  (which is normalized for amplitude 1 at cylinder 1) at the phase
angle given by ki. The star diagram shown here is for the above example for the ½ - II critical.
Each vector in the sum has the magnitude from the II (two-noded) critical mode shape shown at
the angle indicated by the k = ½ phase diagram. The net vector shows the vector magnitude |…|
and its phasing, the net effect of the vector combination of the vibratory input from all 8 cylinders.
12-25
These contributions can cancel or they can be additive. If there is no net energy input there will
be no vibration. For illustration, the following mode shape II is assumed within the engine:
The star diagram illustrates that the firing order as reflected in the crankshaft mechanical design,
the engine order involved and the resonant mode shape  all matter. Note that in this example
vibratory mode, cylinders 3 and 7 are at nodes so they have no vibratory motion and cannot input
energy to the vibration.
In V-format engines there are two cylinders (one for each bank) connected to the same crank
throw. Thus, the two banks are coordinated with the same firing order but at a delay that will have
a different phase shift in each gas torque harmonic. The net input can be analyzed for the reference
bank and then this can be multiplied by the V-factor = 2|cos(kV/2)| that depends upon the V angle
V between the two banks of the V-engine [12]. It can seen that if V is correct for a given engine
order k, the two banks of a V-engine can completely cancel regardless the net energy input found
from the star diagram for the reference bank or Qk’.
Reciprocating engine torsional analyses are typically done under the assumption that all
cylinders are producing the same output, i.e. all Qki’ are equal. Most Classification Society rules
also require that they also be done with one cylinder down. This is to cover the case where the
fuel is secured to a cylinder because it is having a problem. This is occasionally done on low-
speed diesels, in particular. The star diagram can be used to find which cylinder will be the worst
case if it is not operating. Figure 12.11 shows the results for the limiting torsional stress in a low-
speed diesel installation compared with the Lloyd’s Rules. It shows the case with all cylinders
operating and with the worst-case cylinder shut down. It can be seen that the I-4, I-5, I-6, I-8 and
I-12 criticals are in resonance with significant results in the operating range of the engine. This is
12-26
probably a 6-cylinder engine with a 4-bladed propeller contributing to the vibration in the I-4 and
I-8 criticals. The I-6 critical exceeds the limit for continuous operation, but is acceptable for
“passing through” so the engine has a barred speed band forbidding continuous operation in the
rpm range 71 to 83. With one cylinder down, the stress exceeds the continuous operation limit
above about 72 rpm.
Key:
solid line – all cylinders operating
solid/dotted line – worst case
cylinder not operating
Figure 12.11 Limiting Torsional Vibration Stress for a Low-Speed Diesel Installation [13]
Studies have been done under the assumption that the Qki’ will vary probabilistically due to
cylinder-to-cylinder variations in the fuel supply, air supply, etc. This has shown that the complex
vector combination in the star diagram can be very nonlinear, i.e. small changes in Qki’ can make
much larger changes in the net excitation [6, 14]. Figure 12.12 shows results for a low-speed diesel
installation compared to the American Bureau of Shipping limit for continuous operation.
12-27
Figure 12.12 Stress Results with a Standard Deviation = 15% in Cylinder Output [6]
Note in Figure 12.12 that the stress in critical I-6 would require a barred speed range and it
has a normal distribution indicating that the influence is linear since this is a major order for this
6-cylinder engine and there is no node within the crankshaft of the engine. The stress in criticals
I-5 and I-4 are nonlinear with the response varying much more than the standard deviation of the
input Qki’. Even though this engine passes the ABS continuous operation limit in critical I-4 by
2/3 under the assumption that all Qki’ are equal, it could exceed this limit due to variation in the
Qki’.
12.4 Machinery Vibration Mitigation
A vibration problem usually occurs because resonance (e = ni) exists, there is significant
excitation at this frequency e, and there is inadequate damping at the natural frequency ni to
limit the response to an acceptable level. This provides the approaches to vibration mitigation that
can be the following:
 Lower the natural frequency. To remove or shift a resonance, the natural frequency
ni (k/m)1/2 can be lowered by increasing the mass or reducing stiffness in the system.
The diameter of the shafting can be reduced as permitted by regulatory bodies. Tuning
weights are added to the front end of reciprocating engines to lower natural frequencies.
Lowering ni can be effective with propeller excitation, in particular, since that excitation
generally follows the Propeller Law with torque proportional to Np2.
12-28
 Raise the natural frequency. The natural frequency ni can be raised by increasing
stiffness by strengthening foundations, increasing diameter of shafting, etc. This is
particularly effective if the critical can be moved out of the operating range. A change in
the engine manufacturer or the number of cylinders could shift the natural frequency in
resonance up or down.
 Increase damping . System damping is usually small and it has little impact on the
vibrations except near resonance. Structural damping is difficult to increase. Resilient
mounts can be used on reciprocating engines to add damping and isolate the vibratory
excitations of the engine from the ship in general. Flexible shaft couplings, such as
shown in Fig. 10.15, are the most effective way to add torsional damping to shafting
systems.
 Move the excitation frequency. The excitation frequency e can be shifted away from
resonance by changing the engine operating rpm, if possible, or by changing the number
of propeller blades Z. Since the major-order excitation is often the largest problem in
reciprocating engines, a shift from two-stroke to four-stroke would shift the major-order
excitation frequency downward.
 Reduce the magnitude of the excitation. The propeller excitation magnitude can be
reduced by hull form modifications that will smooth out the wake variation around the
propeller disk, but this is difficult after initial design. Various tunnels, fins, and ducts can
be added to the hull to smooth the inflow to the propeller. The wild card in reducing
propeller excitation, however, is the use of propeller skew s(x) as shown in Fig. 12.13.
Figure 12.13 Definition of Propeller Blade Skew [9]
The wake field behind a hull has some variation with radius, but this is modest. Regardless
this variation is known. When a propeller blade passes through a high wake region such as the top
12-29
of the disk on a single-screw ship, the entire blade tends to experience the same wake effect at the
same time. Thus, when the variable lift is integrated along the blade from hub to tip, the excitation
at each radius will tend to be in phase and additive. If there is a radial wake variation, it is included
in the propeller excitation analysis. By introducing skew s(x), a time shift is introduced in the
time at which each radius of the blade experiences the high wake region. This produces a phase
shift ns(x) that is be different in each harmonic n of the wake. As a result of this phase shift, it is
possible for the integration of a component of the variable lift, e.g. variable thrust, from the hub to
the tip to produce a greatly reduced or even zero net result if the optimum skew distribution is
utilized. Skew does not have an appreciable effect on propeller efficiency and only introduces a
modest weight and cost increase due to the more demanding propeller blade bending that results.
There is also a loss in propeller backing performance. The skew does have the additional benefit
that cavitation inception is delayed further reducing propeller radiated pressure that produces hull
vibration excitation over the propeller.
The Greenblatt’s propeller skew optimization program SKEWOPT [11] can be used to
optimize the quadratic or cubic skew distribution s(x) for a propeller of a given design to minimize
a weighted linear combination of the six vibratory forces and moments acting on a propeller in an
arbitrary wake field w(r, ). Thus if a vessel has a problem with a resonance with an axial natural
frequency and blade rate propeller excitation, it is possible to essentially eliminate the axial blade-
rate vibratory excitation by the use of an optimum skew distribution. Maximum reduction tends
to be about 100% tip skew, which is defined as the skew that will have s(R) = 2/Z rad., the blade
spacing.
References
1. Long, C. L., “Propellers, Shafting, and Shafting System Vibration Analysis,” Ch. X in
2. Thomson, W. T. and Dahleh, M. D., Theory of Vibration with Applications, 5th Edition,
Prentice Hall, Upper Saddle River, NJ, 1998.
3. Tse, F. S., Morse, I. E. and Hinkle, R. T., Mechanical Vibrations Theory and Applications,
2nd Edition, Allyn and Bacon, Inc. Boston, MA, 1978.
4. Meirovich, L., Elements of Vibration Analysis, 2nd Edition, McGraw-Hill, Inc., New York,
1986.
5. Parsons, M. G., “Mode Coupling in Torsional and Longitudinal Shafting Vibrations,” Marine
Technology, Vol. 20, No. 3, July 1983, pp. 257-271.
12-30
6. Nikolaidis, E., Perakis, A. N., and Parsons, M. G., "Torsional Vibration Analysis of Diesel
Engine Shafting System: A Structural Reliability Approach," in Haddad, S.D., (ed.),
Advanced Diesel Engineering, Ellis Horwood, Ltd., Chichester, UK, 1988.
7. Parsons, M. G., Vorus, W. S. and Richard, E. M., “Added Mass and Damping of Vibrating
Propellers,” University of Michigan Department of Naval Architecture and Marine
Engineering, Report No. 229, Ann Arbor, MI, Oct. 1980.
8. Parsons, M. G. and Vorus, W. S., “Added Mass and Damping Estimates for Vibrating
Propellers,” Proceedings of the Propellers ’81 Symposium, Virginia Beach, VA, May 26-
27, 1981, pp. 273-302.
9. Carlton, J. S., Marine Propellers and Propulsion, Butterworth-Heinemann, Oxford, UK,
1994.
10. Tsakonas, S., Breslin, J. and Miller, M., “Correlation and Application of an Unsteady Flow
Theory for Propeller Forces,” Transactions SNAME, Vol. 75, 1967, pp. 158-193.
11. Greenblatt, J. E., “SKEWOPT: Propeller Skew Optimization Program User’s and
Programmers Manual, University of Michigan Department of Naval Architecture and
Marine Engineering, Report No. 204, Ann Arbor, MI, Aug. 1978.
12. Den Hartog, J. P., “Multicylinder Engines,” Chapter 5 in Mechanical Vibrations, McGraw-
Hill, New York, 1956.
13. Woodward, J. B., Low Speed Marine Diesel, Wiley-Interscience Publication, John Wiley and
14. Nikolaidis, E., Perakis, A. N., and Parsons, M. G., "Structural Reliability of Marine Diesel
Engine Shafting Systems," Transactions of Society of Naval Architects and Marine
Engineers, Vol. 93, 1985.
12-31
12-32
Chapter 13 Design and Analysis of Fluid Systems
To this point we have focused on the prime movers, propulsion systems components, prime
mover-propulsor matching, and control of the overall propulsion system. We will now shift
gears a bit and focus on the fundamentals of fluid systems design. Fluid systems are used
throughout ships and the marine engineer has the responsibility to design, specify, detail, and test
these systems. Systems are provided for fuel, lube oil, cooling water, steam, potable water,
waste water, ballast, bilge water, liquid cargos, feed water, condensate, etc. [1]. The various
Heating, Ventilation, and Air Conditioning (HVAC) systems also fall into this general category
as well [2]. You should consult Marine Engineering for a descriptive treatment of these various
systems. We will concentrate here on the design fundamentals that are used in the development
and detailed design of these systems. This treatment will focus on the fluid systems.
System Diagrams
The basic requirements for fluid systems and HVAC systems are usually set by the system
diagrams. These diagrams show a schematic representation of the system and establish all
system operating characteristics, applicable specifications, and ordering data for all system
components. They are developed very early in the design process to define the design
functionality, to permit ordering of the purchased components, and to define what should be
developed on the arrangement drawings and CAD models which are developed much later. The
arrangement drawings today are usually developed within the overall three-dimensional
computer product model for the machinery spaces.
The following outlines what the system diagrams should, or at least may, contain:
 Schematic arrangement of the system showing all components including system
operational features, what components are served, isolation capabilities for failure or
maintenance, crossovers to other systems, etc.;
 Design pressure and flow for each part of the system;
 Specified head and flow and other specifications for all pumps;
 Sizing and specification for all piping – the diameter is set by the flow and maximum
acceptable velocities to avoid erosion; increased if necessary for acceptable pressure drop;
13- 1
 Type, location, and specification for each valve and component –

flow stopping or isolation for maintenance gate, ball, or butterfly valves
coarse flow control globe valves
fine flow control throttle globe valves
pressure relief relief valves
vacuum relief check valves
reverse flow prevention check valves
pressure regulation pressure regulating valves
pressure reduction reducing valves or orifices
vents and drains usually added as part of arrangements
 Type, location, and specification for all instrumentation for monitoring, control, or periodic
testing - pressure, differential pressure, vacuum, temperature, flow, level, salinity,
position, etc.
 Type, location, and specification of all special components - filters, strainers, etc.
 Special guidance notes for the designer’s use in developing the actual arrangements –
required/permissible slopes, orientation of components, required elevations, etc.
 Fabrication specifications – minimum bend radii, welding, brazing, inspection
requirements, etc.
 Testing requirements – hydrostatic tests, special operational tests, etc.
A typical system diagram showing a system schematic for a gas turbine fuel handling
system is shown in Fig. 13.1. Note that the diagram does not show the physical location of the
parts of the system, but shows the operational and functional characteristics of the system. Some
system diagrams do show bulkheads so that it is clear in which space each part of the system is
located. This particular example does not show the more detailed specifications for the piping
and components of the system. This additional information is usually or often included on the
system diagrams since it clearly shows the full intent of the designer and allows the purchasing
department to prepare purchasing documents on the basis of the diagram information. The
principal valve types used for flow control and isolation are shown in Fig. 13.2. The principal
valve types used for back flow prevention, pressure regulation, and pressure protection are
shown in Fig. 13.3.
13- 2
Figure 13.1 Schematic Diagram of Gas Turbine Fuel System
13- 3
Figure 13.2 Principal Types of Valves for Flow Control and Isolation [1]
Figure 13.3 Principal Types of Valves for Back Flow Protection and Pressure Control [1]
13- 4
Arrangements Drawings
The arrangement of the fluid systems is developed based upon the system diagrams
developed by the design engineer. These definitions show the actual arrangement and
dimensions of the system and provide the basis for the fabrication and construction. Today, they
are usually developed by designers (CAD drivers) within the three-dimensional ship product
model and then converted to two-dimensional paper drawings and sketches as needed for
production information. Prior to the evolution of three-dimensional computer modeling that
allows virtual “walk throughs” and the checking of all clearances, accesses, slopes, etc., high
priority (nuclear, etc.) systems were often schemed and developed in full-scale wood and
cardboard mockups or scale models before they were converted to drawings. Prior to on-block
outfitting, the systems were run on the ship after structural fabrication and the first system to
arrive at a location usually got the first choice of space. With the more concurrent development
of all systems in a region within the computer product model, a more system optimal
arrangement can be achieved.
In the development of the arrangements, the designer needs to consider minimum cost;
access to valves and equipment for operation, maintenance, and repair; smooth flow into
components to avoid erosion; visibility of local instrumentation; provision of adequate drainage;
access to shipboard welds for fabrication and inspection, such as radiography; access to make
needed freeze seals where needed; provision of adequate hangers for support under normal
operation and perhaps shock; and the provision of adequate flexibility for thermal growth. The
latter two issues require designing to proper guidelines or specific analyses must be performed
by the marine or structural engineer.
Design codes require specific thermal expansion analyses if the system operating
temperature is high. To illustrate this design issue, consider a length of straight 6” diameter
steam piping extending between two bulkheads and carrying steam at T = 800 F. This pipe
might have about a 1/2” wall thickness. When heated to 800 F, this pipe would expand  = T,
where  is the coefficient of thermal expansion of the steel in in/in F. For steel, this expansion
would be about 0.005 in/in and this would produce a compressive stress in the straight pipe if it
were constrained by the bulkheads. The resulting compressive stress if rigidly contained would
be  = E = 0.005 in/in • 30x106 psi = 150,000 psi. Steel has an allowable  = 25,000 psi so
13- 5
compressive failure and/or buckling would probably occur. Thus, high temperature piping must
be provided adequate flexibility through bends, curves, and expansion joints and loops to be
acceptable in service. This becomes particularly critical when multiple branches of a piping run
are constrained close to their junction points.
Materials
The marine engineer has important responsibilities in the selection of materials within fluid
systems. Further, he or she must monitor component specifications and design to ensure that
they will be compatible and acceptable from a materials standpoint. You should consult the
materials chapter in Marine Engineering [3]. In general, material selection needs to consider
first costs versus replacement cost; strength at design temperature; corrosion resistance; erosion
resistance; wear properties; and the fabrication methods necessary, such as welding, brazing, and
adhesives. Composite and plastic piping is in increasing use in cold water service. Stainless
steel piping in high-pressure nuclear applications receives the most extensive material quality
control, welding certification, and nondestructive testing.
One of the more subtle corrosion issues in fluid systems is galvanic corrosion. This can
occur when dissimilar metals are placed in contact or close to one another in an electrolyte, such
as seawater. In the presence of an electrolyte, they form a wet cell battery and the less noble
metal experiences accelerated corrosion while the more noble metal experiences cathodic
protection. Figure 13.4 shows the corrosion potentials or Galvanic Series of metals in flowing
sea water at ambient temperature [4]. The unshaded symbols show ranges exhibited by stainless
steels in acidic water such as may exist in crevices or in stagnant or low velocity or poorly
aerated water where Stainless Steel become active, while the shaded areas show the potentials of
Stainless Steel when is in passive state. A more complete presentation is available in Marine
Engineering [3].
Galvanic corrosion is the reason zinc sacrificial anodes are placed on the hull and rudder so
that they will corrode away and protect the more noble unpainted parts of the hull mild steel.
Some vessels apply an impressed electrical current field to the hull to provide the same
protection. Galvanic corrosion can also occur in fluid systems, particularly seawater systems,
13- 6
when dissimilar metals are used in the piping, weld metal, component bodies, valve and
component internal trim, etc.
Figure 13.4 Galvanic Series in Flowing Seawater [4]
Fabrication and Testing
The fabrication of fluid systems and the testing of these systems is a very important
consideration for the marine engineers responsible for these systems. The fabrication may
involve welding using various types of joints: butt welds, socket welds, etc. They might be
inspected using magnetic particle inspection, liquid penetrant inspection, radiography, etc.
13- 7
Welding procedures must be developed and qualified for certain special welds. The fabrication
may also involve brazed joints using various joint configurations and these may be inspected by
ultrasonic inspection. These issues are particular important in high-pressure nuclear and
submarine systems.
The marine engineer will also be involved in specifying, writing procedures for,
supervising, and evaluating the testing that must be performed on fluid systems prior to their
certification for use. This includes flushing, hydrostatic testing, operational testing, etc. of
modules and portions of systems leading to the operational testing of the entire systems during
dockside testing and sea trials. The increased use of equipment modules allows more testing in
advance of module installation on block.
Piping Pressure Drop Calculations
One of the important parts of sizing fluid system piping and the selection of pumps for
these systems is the calculation of the pressure drop in the systems due to friction. The sizing
usually begins by consideration of the required flow rate. The mass flow rate, flow volume, and
flow velocity in a pipe are related by continuity,
on a mass basis m = AV = constant

or on a volume basis Q = AV = constant
where Q is a volume flow rate m3/h or U.S. gallons per minute (gpm); A = Di2/4 is the flow
area; and V is the flow velocity averaged over the flow area in m/s or ft/s. Of course, proper
units have to be used in these equations. There are 7.4766 U.S. gallons in 1 ft3, 264.2 U.S.
gallons in 1m3, and a flow rate of 4.403 U. S. gpm equals 1 m3/h. The initial sizing of piping is
usually made by considering the flow that must pass through each line and the design and
maximum permissible velocities for the particular type of system. In general, maximum water
flow rates are about 3 m/s or 10 ft/s. A comprehensive table of recommended design velocities
and maximum velocities is shown in Table 13.1.
Fluid Properties. Recall from your fluids course, that a fluid is a substance that deforms
continuously under a shear stress. The fluid property viscosity is defined by Newton’s Law of
13- 8
Table 13.1 Recommended Design and Maximum Velocities in Marine Fluid Systems [1]
Viscosity that relates the shear stress  to the velocity gradient in the fluid as follows:
 = du/dy where  is the absolute or dynamic viscosity [units FT/L2]
For Newtonian fluids (essentially all marine engineering fluids, except LNG),  can be
considered a constant that is a strong function of temperature. In general, absolute viscosity
increases with temperature for gases and decreases with temperature for fluids as we saw for
marine fuels in Fig. 3.1. Saturated steam and slightly superheated steam are exceptions to this.
For most practical marine engineering calculations,  can be considered a constant at some value
based upon the average temperature. As noted in the fuels chapter, the units of viscosity are
many. The most common units of absolute viscosity are,
13- 9
 in SI units = 1 N s/m2
of 1 poise = 1 dyne s/cm2; the viscosity of water at 4˚C (39.2 F)
1 centipoise = 0.01 poise
1 lbf s/ft2 = 47.88 N s/m2 = 478.8 poise = 47880 centipoise
There is also the kinematic viscosity  defined as  = gc/ [units L2/T], where the usual
units resolution constant gc = 32.174 lbmft/lbfs2 appears in English units; in SI units gc = 1
kgm/Ns2. The most convenient English units are [lbm/ft3], weight density  = g/gc [lbf/ft3].
If the fluid density  is in [slugs/ft3], then  = / and  = g. The units of kinematic viscosity
are,
 in SI units = 1 m2/s
 of 1 stoke = 1 cm2/s with centistokes [cSt] more useful = 0.01 stoke
1 ft2/s = 0.092903 m2/s = (30.48)2 stokes = 92,903 cSt
Recall that specific gravity S is defined as S = (4˚C)/ water(4˚C). There are many other
units for fluid density and viscosity, but we will work primarily with the above units.
Energy and Head. If we apply the First Law of Thermodynamics to steady flow in a pipe
extending from point 1 to point 2,
gz1/gc + V12/2gc + u1 + 1q2 + P1/1 = gz2/gc + V22/2gc + u2 + 1w2 + P2/2
where recall (Table 2.2) that the units resolution constant gc = 1 kgm/Ns2 = 32.174 lbmft/lbfs2.
The head is defined as the mechanical energy per unit mass in this equation or,
H = P/ + gz/gc + V2/2gc [lbf ft/lbm or N m/kg]
The more commonly used definition of head in engineering practice is in length units [ft or m] so
this is multiplied by gc/g to give,
h = P/ + z + V2/2g [ft or m]
The three components that comprise this total head are the pressure head, the elevation head,
and the velocity or dynamic head, respectively. The sum of the pressure head and the elevation
head is called the piezometric, hydraulic, or static head.
13- 10
The usual treatment is to use a “one-dimensional” analysis for pipe flow with x taken along
the length of the pipe and V taken as the average velocity in the pipe or m/A. The flow, of
course, has a velocity distribution across the pipe diameter with a zero flow no slip condition at
the wall and maximum flow at the centerline. For turbulent flow, which is usually the case,
using this average velocity in the kinetic energy term is acceptably close when compared to how
well we know the effects of friction. In laminar flow (Re = DV/ ≤ 2000, pipe diameter D), the
kinetic energy found by squaring the average velocity can be as low as about 1/2 of the actual
kinetic energy found by integrating the velocity squared distribution across the pipe. Fluids texts
[5] include corrections in the case of laminar flow.
Incompressible Frictional Flow. For most pipe flow problems of interest to the marine
engineer, the fluid can be treated at incompressible,  and v constant. The classic Crane
Technical Paper 410M [6] offers the following guidance for when this assumption is valid for
oil, water, steam, air, etc.:
1. if the pressure drop is less than about 10% of the inlet pressure to the line, assume
that the flow is incompressible using vinlet or voutlet,
2. if the pressure drop is between about 10% and 40% of the inlet pressure, assume
that the flow is incompressible using vaverage = (vinlet + voutlet)/2,
3. if the pressure drop exceeds about 40% of the inlet pressure, treat the fluid as
compressible if gaseous or a vapor.
Compressible frictional flow is treated in fluids and some thermodynamics texts.
The Darcy-Weisbach Formula. If we apply the Momentum Equation to one-dimensional,

incompressible, horizontal flow in a differential element dx of constant diameter D pipe, we
obtain,
F = net change of momentum (mV) and from continuity m = AV
define a friction factor f = /V2/2gc or  = f V2/2gc
From momentum,
PA – (P + dP)A – Ddx = mVout – mVin = 0
so dP = – Ddx/A = – f(V2/2gc)Ddx/(D2/4)
13- 11
Giving for the frictional pressure drop (– dP) in a length dx = L,
dPL = (4f)(L/D)(V2/2)/(144gc) in English units; P[psi], V[ft/s]

dPL = (4f)(L/D)(V2/2) in SI units; P[Pa], V[m/s]
Expressing this frictional loss in terms of head gives,
hL = f (L/D)(V2/2g) where f = 4f
This result is called the Darcy-Weisbach formula. The friction factor f is presented in the Moody
Diagram shown in Fig. 13.5. In laminar flow, an exact solution for the friction factor is possible
giving f = 64/Re, where the Reynolds number in laminar pipe flow is Re = VD/< 2000. Similar
to the ITTC friction line for flat plates, the frictional drag in turbulent flow is dependent on both
the Reynolds number and the average roughness of the pipe internal surface expressed as  [ft or
m] or a non-dimensional relative roughness /D.
Figure 13.5 Moody Diagram [7, 8]
13- 12
Practical values for the relative roughness /D are given as a function of diameter for
various types of piping in Fig. 13.6. The friction factor f for fully turbulent flow, where f is no
longer dependent upon the Reynolds number, is given on right in Fig. 13.6.
Frank Vibrams’ drawn tubing d < 10 in.
[mm]
Figure 13.6 Relative Roughness of Pipe Materials and Fully Turbulent f [6]
A useful fit to the Moody diagram data for the turbulent region is the following:
f = 1.325/[ln(0.270/D + 5.74/Re0.9)]2 where Re = VD/


This result is within the typical accuracy of reading the Moody diagram graph Fig. 13.5.
13- 13
The Modified Bernoulli Equation. Recall that the Bernoulli Equation applies to ideal
(frictionless), incompressible, steady-flow without heat transfer or work. In one-dimension, the
First Law of Thermodynamics yields just,
H = P/ + gz/gc + V2/2gc = constant from point to point
In piping systems, we also have frictional losses in the piping and components (valves, bends,
etc.) and also have pumps to create a pressure increase in the fluid so we can write a modified or
“marine engineer’s” version of the Bernoulli equation to be between points 1 and 2,
enegy in H1 + Hp = H2 + HL energy out

energy added energy lost
by pump to friction
Consider the following simple piping system,
The modified Bernoulli equation will then be,
P1/ + gz1/gc + V12/2gc + Pp/ = P2/ + gz2/gc + V22/2gc PL/


   P1/ + z1 + V12/2g + hp = P2/ + z2 + V22/2g + hL
In this example, the two velocities will be essentially zero at the free surfaces of the fluid. The
flow rate will then be the flow rate that satisfies this equation since both Pp/ and PL/depend
upon the flow rate. The Darcy-Weisbach formula gives the frictional losses in the straight pipe
only. These are termed the “major” losses although they may not actually be the dominant losses
in practical marine engineering systems.
13- 14
Minor Losses. The pressure drop due to expansion and contraction of a pipe diameter,
entering and leaving a pipe, valves, fittings, components, bends, etc. are called “minor” losses.
The quotes are used here because they are often the dominant losses in practical marine
engineering problems. There are three ways to represent these losses:
resistance coefficient K = number of velocity heads lost
K = PL/(V2/2gc)
equivalent length L/D = length of straight pipe that will give the same pressure drop
L/D = K/f = PL/(f V2/2gc)
flow coefficient CV = flow of 60 F water through a component at 1 psi P, primarily

used for valves; defined only in English Engineering System units at this time
CV = 29.2 D2/√K
Components, bends, etc. are tested and the pressure drop results are presented in one of these
forms. A very small number of the resistance coefficients can be derived analytically. Examples
are included in fluids texts.
The resistance coefficients for pipe entrances and exits are shown in Fig. 13.7. At all
shapes of exit, the flow slows to zero losing the dynamic head so K = 1. The resistance
coefficients for a sudden expansion or sudden contraction are shown in Fig. 13.8; in both cases
the reference velocity head is in the smaller pipe of diameter d. The resistance coefficients for a
gradual expansion or gradual contraction are shown in Figs. 13.9 and 13.10, respectively. Again
in both of these, the reference velocity head is in the smaller diameter pipe. Resistance
coefficients for the added resistance for 90˚ pipe bends, in addition to the resistance due to the
pipe length around the bend, are shown in Fig. 13.11. Five pipe diameter radius bends are
typically the minimum radius bends permitted in marine engineering practice. When pipe is bent
to a tighter radius, it tends to lose circularity. A collection of resistance coefficients for typical
valves is shown in Fig. 13.12. Note that the constants in these data are just the L/D values for
the component; i.e., K = (L/D)f. In this presentation,  is the ratio of the diameters,  = d1/d2,
which is usually 1 in most applications. Note that for check valves, the minimum pipe flow
velocity needed to ensure full disc lift is also given. The data for valves is design detail
dependent and specific vendor data should be used, if available.
13- 15
Figure 13.7 Pipe Entrance and Exit Loss Resistance Coefficients [7]
Figure 13.8 Sudden Expansion and Sudden Contraction Resistance Coefficients [7]
13- 16
Figure 13.9 Gradual Expansion Resistance Coefficients [7]
Figure 13.10 Gradual Contraction Resistance Coefficients [7]
Figure 13.11 Pipe Bend Added Resistance Coefficients [7]
13- 17
Note:  = d1/d2 and

V with the overbar = specific volume [ft3/lbm]
Figure 13.12 Resistance Coefficients for Selected Valves and Fittings [older English units version of 6]
13- 18
Example 1. Suppose we want to find the flow of water at 50˚C between the following two tanks.
 = 1012.1 kg/m3 Crane A-6 in Appendix A
3 in. Schedule 40 pipe commercial steel internal D = 0.0779 m

Crane B-16 in Appendix A
25 m straight pipe
We can use the modified Bernoulli equation with no pump term,
We can set up a losses table for convenience.
Note: L/D = K/f

/D = 0.05 mm/77.9 mm = 0.00064
Assuming fully turbulent flow and taking f = 0.0175 from Fig. 13.6 to start,
13- 19
Note that this Reynolds number is not quite fully turbulent, so we might want to iterate using
Fig. 13.5 or the equation at the bottom of page 13-13 to give f(Re = 3.22 x 105) = 0.019. We
should also make sure that this velocity is an acceptable flow velocity in Table 13.1, which it is
for hot water.
Systems Design. The problem of sizing piping systems and pumps can be outlined as
follows:
1. For the required flows, pick pipe diameters to give acceptable velocities;
2. Determine the system pressure drops; possibly increasing diameters if the pressure
drops are excessive;
3. Check components to ensure P’s are acceptable: e.g., check valves have minimum
P to prevent chatter; strainers, etc. may have maximum P requirements; etc.;
4. Develop the head-flow characteristics for the system;
5. Select a pump by matching head-flow characteristics for the system and pump;
6. Verify that the pump will not cavitate at the operating condition(s).
Returning to the modified Bernoulli equation between two points, we can write,
H1 + Hp = H2 + HL
where all the terms, in general, depend upon the flow rate Q in the system. This can be
rearranged to isolate the pump head on the left hand side,
Hp = (H2 – H1) + HL
pump change in friction and
head piezometric head minor losses
(pressure and elevation terms)
13- 20
This formulation assumes that the kinetic energy terms are negligible or their difference is
negligible. If not, the velocity heads must also be included in H1 and H2. In this rearrangement
of the modified Bernoulli equation, the expression on the left is the pump characteristic Hp(Q)
and the terms on the right constitute the system characteristic [H2 – H1 + HL], also a function of
the flow rate Q. Pump characteristics are usually available from the vendor graphically so this is
a common way to solve these matching problems. The system characteristics will appear as
follows for circulating systems that form a closed loop and for transfer systems that move fluid
from one point to another:
These systems characteristics can then be matched with the pump characteristics to give a
general matching problem (equilibrium between the driver and the load) that might appear as
follows for a transfer system:
13- 21
Combining System Characteristics. For branched systems that are common in marine
fluid systems, you can combine the characteristics for series and parallel portions of the system
to obtain the combined characteristic of the overall system. The system portions might appear as
follows:
series paths have common flow and additive net head changes
parallel paths have common net head changes and additive flows
Consider combining the characteristics of the parallel branches BC 1 and BC2,
and then the series portions AB and BC can be combined to give the total system characteristic,
The combined characteristic for AC can now be matched with a pump to obtain the operating
point and then the flows in the parallel branches BC1 and BC2 can be established for this
13- 22
condition. In the following sketches, the upper diagram obtains the operating point where the
pump characteristic is in equilibrium with the total system characteristic for AC. At this total
flow QAC, we can obtain the total flow point on the combined BC operating curve. At this head,
we can then obtain the individual flows QBC1 and QBC2 that add to produce the total flow QAC.
pump
The marine engineer also encounters systems that have branches that do not rejoin as
above, but serve as terminal points in the system. Consider the following sketch,
We may want to size the piping to give Q1 and Q2 at some specified value using a given pump
with characteristic Hp, which is a function of (Q1 + Q2). One approach is to write an expression
for the pressure at the junction point PJ using the modified Bernoulli equation for each of the
three branches,
Branch 1 PJ = P1 + KJ1Q12
Branch 2 PJ = P2 + KJ2Q22 + gzJ2/gc
Pump branch PJ = P3 – K3J(Q1 + Q2)2 – gz3J/gc + Pp
13- 23
where KIJ is the constant from the Darcy-Weisbach formula using Vij = Qij/Aij and f(/Dij, Re)
and Dij for the system from i to j. This yields three equations and four unknowns (P J, DJ1, DJ2,
D3J), so there is no unique solution to this problem. The usual approach is to pick the diameters
that will minimize systems cost, subject to ensuring that the system has sufficient hydraulic
stiffness. System hydraulic stiffness is the ability to change Q1 without significantly affecting Q2
(the classic “scalding the person in the shower when the adjacent toilet is flushed” problem).
The requirement for stiffness leads to a more costly solution with large D3J compared to DJ2 and
DJ1. This makes PJ essentially independent of Q1 or Q2 since K3J is then small.
Characteristics of Pumps
Pumps are an extremely important part of marine propulsion and auxiliary systems. Some
of the main points related to their selection and application will be introduced here.
Classification of Pumps. There are two main classes of pumps: turbomachinery pumps
and positive displacement pumps. Turbomachinery pumps utilize hydrodynamics to develop the
pressure within the pump; an example would be a axial flow waterjet pump. Positive
displacement pumps use some type of volume change to create the pressure; an example would
be a reciprocating piston type bilge pump. The following outlines their main characteristics.
Turbomachinery pumps: further classified by the dominant direction of the flow in the
impeller; general service pumps
• centrifugal or radial flow pumps – the most common type of general shipboard pump
e.g. a main feed pump - 450 U.S. gpm at 1100 psi rated point
- low to moderate heads, low to high flows
- can use multiple stages to produce higher heads
- can use multiple suctions to produce higher flows
- will not operate with a dry suction so they must be primed manually or supplied
by a priming pump
- subject to internal cavitation if not provided sufficient inlet pressure
• propeller or axial flow pumps – high flow applications
e.g. a main seawater circulating pump - 45000 U.S. gpm at 10 ft. head
(0.4333 psi/ft H2O; 1.422 psi/m H2O)
- low heads, high flows
- must also be primed
- subject to internal cavitation if not provided sufficient inlet pressure
• mixed flow pumps – characteristic intermediate between axial and radial flow pumps
e.g. some smaller water jets
13- 24
Positive displacement pumps: further classified by the type of volume change;

used to develop high heads, create suction lift, or meter flow.
• reciprocating or piston pumps
e.g. bilge pump – 700 U.S. gpm, 125 psi, 15 ft suction lift
- low flow, any head
- low to moderate viscosity fluids
- will pull a suction so the pump can be self priming
• rotary pumps
e.g. fuel oil service pump – 32 psi, 350 U.S. gpm, 15 “Hg suction lift
- low flow, any head
- any reasonable viscosity fluid (fuels, oils, etc.)
- various mechanical concepts using screws, lobes, vanes, etc.
Figure 13.13 shows the general ranges of rated head and flow characteristics for the various
types of pumps. In marine service as noted above, positive displacement pumps are also used for
suction lift and flow metering and not just for the high head applications as implied in Fig. 13.13.
Figure 13.14 shows a sampling of the different types of pumps found on board ships. An axial
flow waterjet pump is shown in Fig. 11.10.
Figure 13.13 Pump Type Selection Chart based upon Head and Flow
13- 25
Figure 13.14 Sampling of Types of Pumps used On Board Ships [9]
13- 26
Definitions of Head, Power, and Pump Efficiency. Head has been defined above, but for
completeness it is repeated here from the pump viewpoint. Head is the mechanical energy per
unit mass of the fluid. Recall that it includes the piezometric (pressure and elevation) head and
the velocity head. In English Engineering System units:
Hp = E/m = Pp/ lbf ft/lbm m [lbm/s], Pp [lbf/ft2], Pp’ [lbf/in2]
hp = Egc/(mg) = Pp/ ft  [lbm/ft3],  [lbf/ft3], Q [U.S. gpm]
Ep = m Pp/ lbf ft/s energy rate added to fluid
The pump will deliver to the fluid a water or hydraulic horsepower WHPBr given by,
WHPBr = m (Pp/)/550 hpBr

= [Q • 0.1336 ft3/gal ] (Pp’/•144 in2/ft2)/(33,000 ft lbf/hpBr min)
= Q Pp’/1715.3 = Q hp S/3958.4 hpBr
where S is the specific gravity of the fluid being pumped. The term in square brackets is m
[lbm/min]. Recall that the English units resolution constant g c = 32.174 lbmft/lbfs2. In SI units
the water power becomes simply,
PW = m (Pp/) kW m [kg/s], Pp [kPa],  [kg/m3]
= Q hp S/367.47 kW Q [m3/h], hp [m]
The pump will be inefficient due to friction losses from the fluid, friction losses in packing and
bearings, leakage from the high to low pressure parts, turbine vector diagram type losses, etc. so
the driver must provide more power than the pump water power output PW (analogous to
delivered power and effective power for a propeller). The brake power required from the pump
driver is then,
BHPBr = WHPBr/p = Q Pp’/1715.3p = Q hp S/3958.4p hpBr
PB = PW/p = Q hp S/367.47p kW
where p is the pump efficiency defined by these equations. Typical head-flow, efficiency, and
power characteristics will be shown below.
13- 27
Turbomachinery Pump Similarity Laws. Scaling laws can be derived for “similar”
turbomachinery pumps that have geometric similitude, geometrically similar flow velocity
diagrams, and small Reynolds number (viscosity) effects [10]. These results are as follows:
Q/ND3 = constant 1 N [rpm]

gH/N2D2 = constant 2 D, impeller diameter
WHP or PW/N3D5 = constant 3 L, impeller length along axis of rotation
Within the limits of Reynolds number effects, we also have BHP  WHP. Using these results,
we can also establish the effect of changing the rpm N of a known pump (constant D) as follows:
QN
H  N2  Q2
BHP  WHP  N3  Q3 provided p is a constant.
This is the “Propeller Law” for turbomachinery pumps. Consider how these results can be used
to solve the problem of obtaining the operating point and efficiency for a pump in a particular
application if the head and efficiency data is given (as is often the case) only for the rated rpm of
the pump. The following example with H’ = h, the head in ft or m, illustrates:
The similarity results can also be used to scale data from a known pump to obtain approximate
characteristics for a geometrically similar pump of a new size operating at the same rpm. If the
diameter and length are changed by a scale ratio = D/Do = L/Lo, etc. we can derive from the
above,
Q  3; H  2; WHP  BHP  5
13- 28
Specific Speed. The specific speed is a constant for a “similar” series of turbomachinery
pumps that is very useful in the selection of pump type and preliminary design.
Specific speed Ns = the speed of the member of a series of “similar” pumps that will
produce unit discharge at unit head; usually at maximum p
using Q/ND3 = constant 1 and
h/N2D2 = constant 2’ and eliminating D we get,
NQ1/2/h3/4 = constant 4
So for two pumps in a similar pump series,
No = N(Q/Qo)1/2/(h/ho)3/4 and with Qo = 1 and ho = 1 this yields
Ns = N Q1/2/h3/4 where Q [gpm] is the flow per suction

N [rpm] is the pump speed
and h [ft] is the head per stage
Note that specific speed is dimensional with the above English units commonly used. Generally,
centrifugal pumps have low Ns ≤ 4000, axial flow pumps have high Ns ≥ 10 000, and mixed flow
pumps have intermediate Ns values. Figure 13.15 shows the range of typical efficiency of
various marine turbomachinery pumps and the effect of viscosity on the pump efficiency. Thus
if we know the specific speed for a pump, we can understand what type of flow it will have and
estimate the pump efficiency. In SI units, the specific speed is usually based on Q [m3/s], N
[rpm], and h [m] and, thus, has a different numerical value related by Ns(English) = 69.7 Ns(SI).
For water, the effect of viscosity and Reynolds number is negligible, but a change in
viscosity when using other liquids, such as oil, makes it a predominant factor and it is no longer
possible to preserve flow similarity under these circumstances. Figure 13.16 shows centrifugal
pump efficiency and a head correction factor h(Re)/h(Re ≥104) versus a pump Reynolds number
Re = Q/D, where density  [lbm/ft3], viscosity  [lbm/ft s], flow Q [ft3/s], and impeller tip
diameter D [ft].
The overall pump characteristics can be reliably predicted in early design on the basis of
the pump specific speed. These results are shown in Fig. 13.17 that shows generalized head,
efficiency, and power characteristics versus specific speed and flow for turbomachinery pumps.
These results can be useful in preliminary design before specific pump designs are selected.
13- 29
Figure 13.15 Typical Pump Efficiency for Marine Pumps versus Specific Speed
Figure 13.16 Effect of Viscosity on Centrifugal Pump Performance [11]
13- 30
Figure 13.17 Generalized Turbomachinery Pump Characteristics versus Ns
13- 31
Cavitation and Net Positive Suction Head. Cavitation can occur in pumps when the fluid
vaporizes in the lowest pressure region (highest velocity region) near the pump inlet. These
bubbles then collapse after the pressure increases sufficiently within the pump impeller. The
cavitation can lead to noise, vibration, loss of head and efficiency, and worst of all destructive
pitting erosion of the pump internal metal can occur. The cavitation occurs when the local
absolute static pressure falls to or below the vapor pressure of the pumped fluid at its
temperature. The required Net Positive Suction Head (NPSHr) is the pressure required at the
inlet of the pump to prevent cavitation inside the pump. The NPSHr is measured by the pump
manufacturer and usually provided as a function of flow rate with the pump head-flow
characteristics.
where hV(T) is the absolute vapor head of the fluid,
NPSH = hI – hV [ft or m] ≥ 0 vapor pressure Psat (Crane A-6) at T expressed as a
head, and hI is the absolute total head at the pump inlet
The vendor provides NPSHr(Q) or general values must be assumed as described below. The
system designer must ensure that the available NPSHa exceeds that required to be sure that
cavitation will not occur within the pump (NPSHa > NPSHr). Consider the following example:
If specific pump data is not available, the required NPSHr can be estimated using Thoma’s
Cavitation Parameter, the same for all pumps of equal specific speed Ns. Defined as follows:
13- 32
 = NPSHr/h where h is the head of the first stage in the same units as NPSHr
For estimates you can use formulas from the Hydraulic Institute using English specific speed Ns
units of Q [gpm], h [ft], and N [rpm],
single suction pumps  = 6.3 Ns4/3/106
double suction pumps  = 4.0 Ns4/3/106
Actual vendor data is obviously preferred, if available.
Pump and System Matching. The matching of pumps and piping has been discussed
above. Here we will just add the problem of combining the characteristics of two parallel,
unequal capacity pumps. Consider the situation shown in the sketch.
up
We want to combine the pumps and associated system characteristics at point J so that they can
be matched against the characteristics for the remainder of the system beyond J. In this case, we
do the matching point not with just the pump versus all of the rest of the modified Bernoulli
equation, but with the system before J versus the system beyond J. The head at J must, of
course, be the same from either side; it is a single pressure. We will have,
PJA = P1 + PA – PfA + PeA where Pfi is the branch i friction losses
PJB = P2 + PB – PfB – PeB and Pei is the branch i elevation term
relative to J
We can plot these two results for PJ versus the flow in each part of the system,
(graphs are symbolic – not to scale)
13- 33
These are parallel elements so their flows can be added for each value of P J,
We could then match this with a system characteristic PJ(Q = QA + QB) for the system beyond
point J to obtain the system operating point and flow rate of each pump,
References
1. Cassee, H. J., “Piping Systems,” Ch. XX in Harrington, R. L. (ed.), Marine Engineering,

2. Krinsky, J. L., “Heating, Ventilation, Air conditioning, and Refrigeration,” Ch. XXI in
3. Materials Staff, David Taylor Research Center, “Construction Materials” Ch. XXII in
4. Atlas Steel Technical Note #7 “Galvanic Corrosion”
5. Munson, B. R., Young, T. H., and Okiishi, D. F., Fundamentals of Fluid Mechanics, 4th
Edition, John Wiley & Sons, Inc., New York, NY, 2002.
6. “Flow of Fluids through Valves, Fittings, and Pipe,” Crane Company Technical Paper No.
410M, Metric Edition, 1999.
7. Evett, J. B. and Liu, C., Fundamentals of Fluid Mechanics, McGraw-Hill Book Company,
New York, 1987.
8. Moody, L. F., “Friction Factors for Pipe Flow,” ASME Transactions, Vol. 66, 1944.
9. Sembler, W. J., “Pumps, Compressors, Blowers, and Ejectors,” Ch. XIV in Harrington, R. L.
(ed.), Marine Engineering, SNAME, New York, 1992.
10. Woodward, J. B., “Head and Flow Calculations for Piping Design,” University of Michigan,
Department of Naval Architecture and Marine Engineering Report No. 103, Nov. 1979.
11. Stepanoff, A. J., Centrifugal and Axial Flow Pumps: Theory, Design, and Application,
Krieger Publishing Co., Melbourne, FL, 1993.
13- 34
Chapter 14 Heat Transfer and Heat Exchangers
There are many important applications of heat transfer in marine engineering. We will
have time to introduce the subject, treat some simple heat conduction problems typical of marine
applications, and consider heat exchanger selection, specification, and operating problems.
Mechanical Engineering courses in Heat Transfer would be a valuable elective. Texts by
Incropera and DeWitt [1] and Kreith and Bohn [2] are useful references. Recall from physics
that heat transfer is the process of energy transport that takes place due to a temperature
difference. There are two modes of heat transfer: (1) conduction which is energy transport by
direct molecular communication without appreciable displacement of the molecules, and (2)
radiation which is energy transport between separated bodies by electromagnetic wave
phenomenon. There is also the associated process called convection which is energy transport by
combined conduction and then mass transport or movement of a fluid. The convection can be
free convection when the fluid movement is due to density gradients only or forced convection
when the fluid movement is created by an external pump, fan, etc.
Conduction
Consider a differential element cube with sides dx, dy, and dz which has a heat flow q
[Btu/hr or J/hr] across each face and a heat source S per unit volume [Btu/hr ft3 or J/hr m3]
within the cube due to electrical resistance, fission of Uranium, nuclear radiation heating, etc.
During some period of time dt we can perform an energy balance on this cube and obtain,
heat inflow + heat generated by S = heat outflow + change in stored internal energy
(qx + qy + qz)dt + Sdxdydzdt = (qx + ∂qx/∂x dx + qy + ∂qy/∂y dy + qz + ∂qz/∂z dz)dt

+ CpdTdxdydz
where Cp is the specific heat [Btu/lbmF or J/kg˚C] and T is temperature [F or ˚C]. The second
development we need is the experimentally observed Fourier’s Law, i.e.,
14 - 1
qx = – k Ax ∂T/∂x
[Btu/hr] [ft2] [F/ft]
[J/hr] [m2] [˚C/m]
where the minus sign indicates that the heat flow is in the direction of decreasing T. Fourier’s
Law states that the rate of heat flow (qx) is proportional to the area normal to the flow (Ax) and
the temperature gradient in the direction of the flow (∂T/∂x). The constant of proportionality k is
the material property thermal conductivity which is experimentally determined and depends upon
the material, its state, and its temperature. The necessary English units are [Btu/hrftF or
J/hrm˚C] from the above, but engineering handbooks also use [Btu/hrft2F/in] so one has to be
careful. Values for typical materials using English units and temperature are shown in Fig. 14.1.
Note that the independent variable is temperature on an absolute scale [R].
Btu
Figure 14.1 Thermal Conductivity of Selected Gases, Liquids, and Solids [3]
14 - 2
Fourier’s Law can be substituted into the energy balance equation. For example, the
change term in the x-direction gives with Ax = dydz,
– ∂qx/∂x dx = [∂(k∂T/∂x)/∂x]dx dydz
Repeating this for y and z and dividing by dxdydzdt gives,
∂/∂x (k∂T/∂x) + ∂/∂y (k∂T/∂y) + ∂/∂z (k∂T/∂z) + S = Cp∂T/∂t
or when k can be assumed uniform in space this becomes the general heat conduction equation,
(k/Cp)[∂2T/∂x2 + ∂2T/∂y2 + ∂2T/∂z2] + S/Cp = ∂T/∂t
The leading term in parenthesis is called the thermal diffusivity  = k/Cp [ft2/hr or m2/hr]. In
steady-state problems without sources this reduces to the Laplace equation,
∂2T/∂x2 + ∂2T/∂y2 + ∂2T/∂z2 = ∂2T/∂r2 + (1/r)(∂T/∂r) + (1/r2)(∂2T/∂2) + ∂2T/∂z2 = 0
The first form is in Cartesian coordinates; the second form is in polar coordinates. We will
consider one-dimensional applications of this result that have many important applications.
Plane Wall. We can usually use k(Tave) in most engineering problems. We can use
Fourier’s Law directly in this case,
qx = – kAx ∂T/∂x = kAx(Tin – Tout)/x
For the temperature distribution through the wall, we can integrate the Laplace equation twice to
give,
d2T/dx2 = 0 so dT/dx = C1 and T = C1x + C2
Thus, the temperature distribution is linear through a uniform plane wall. The constants C1 and
C2 are selected to satisfy the boundary conditions of the problem such as,
1) a specified T, T(xi) = Ti
2) perfect insulation, qx(xi) = – kAx ∂T/∂x|xi = 0 so ∂T/∂x|xi = 0
3) line of symmetry, also ∂T/∂x|xi = 0
14 - 3
Thus, for a wall of thickness x with Tin on one side and a lesser Tout on the other,
T = C1x + C2
at x = 0 T = Tin so C2 = Tin
at x = x T = Tout = C1x + Tin so C1 = (Tout – Tin)/x
giving T = Tin – (Tin – Tout)x/x
Composite Plane Wall. A wall or bulkhead could consist of layers of materials of

different conductivity and thickness: steel, insulation, sheathing, etc. It might appear as
sketched,
We can use the plane wall results across each slab noting that in the steady-state the heat flow
through all the layers must be the same,
q = k12A(T1 – T2)/L12 or (T1 – T2) = qL12/(Ak12)
q = k23A(T2 – T3)/L23 or (T2 – T3) = qL23/(Ak23)
q = k34A(T3 – T4)/L34 or (T3 – T4) = qL34/(Ak34)

adding ____________________________________
(T1 – T4) = (q/A)(L12/k12 + L23/k23 + L34/k34)
giving
q = A(T1 – T4)/(L12/k12 + L23/k23 + L34/k34)
14 - 4
With this q, the temperature distribution can be obtained. The slope will change in each layer as
shown in the sketch due to the changing k.
This type of problem can also be approached using an electrical analogy by defining a
thermal resistance R. For the plane wall case this will be,
Ri = Li,i+1/(Aki,i+1) for slab i
giving for I slabs,

I
q = (T1 – TI+1)/ Ri
i=1
analogous to Ohm’s Law I = E/Re for series resistances where Re = Ri.
Axisymmetric Cylinder Wall. For a cylinder such as a pipe with no  and negligible z
variation, the Laplace equation in cylindrical coordinates yields,
∂2T/∂r2 + (1/r)(∂T/∂r) = (1/r)∂(r∂T/∂r)/∂r = 0 giving r∂T/∂r = C1 and T = C1lnr + C2
with C1 and C2 again selected to satisfy the boundary conditions. Fourier’s Law also gives,
qr = – kA ∂T/∂r = – k2r ∂T/∂r
Using separation of variables for this integration,

ro To
qr/(2k)  dr/r =  dT
ri Ti
giving
qr = 2k(Ti – To)/ln(ro/ri)
14 - 5
Composite Axisymmetric Cylinder. This development can be extended to treat a number

of concentric cylinders such as a pipe surrounded by insulation. The approach is similar to the
analysis of the multislab wall,
q = 2k12(T1 – T2)/ln(r2/r1) or (T1 – T2) = [q/(2)] ln(r2/r1)/k12
q = 2k23(T2 – T3)/ln(r3/r2) or (T2 – T3) = [q/(2)] ln(r3/r2)/k23

adding _________________________________________
(T1 – T3) = [q/(2)][ln(r2/r1)/k12 + ln(r3/r2)/k23]
giving
q = 2(T1 – T3)/[ln(r2/r1)/k12 + ln(r3/r2)/k23]
and by inspection the definition of the thermal resistance is given by 2Ri = ln(ri+1/ri)/ki,i+1 for
this case.
Incorporating a Surface Film and Convection. Consider next a colder plate which is
conducting heat from a warmer fluid as might be found in a plate heat exchanger. The fluid can
move so convection will be present either by forced flow or buoyancy. The plate’s surface also
has a thin surface film(s) due to corrosion, fouling, etc.
These two complications are modeled by introducing a heat transfer coefficient h [Btu/hrft2F or
J/hr m2˚C] and using,
film: q = hfA(T1 – T2)

convection: q = hcA(Tb – T1) for q at the surface bounding the fluid
14 - 6
For the film, values of hf are obtained experimentally or perhaps calculated if k and thickness are
known for something like an oxide layer. For the convection case, hc can be found analytically
in only a very few cases (such as laminar flow over a flat plate), so it is usually found
experimentally. The value of hc depends on the following:
1) the character of the fluid flow (laminar, turbulent, separated) reflected in the
Reynolds number Re
2) the properties of the fluid: , Cp, kf, 
3) the geometry of the body and the flow: body shape, arrangement of multiple bodies,
orientation of the flow, direction of gravity
In heat exchangers, the flow direction and tube spacing and arrangement affect hc. In terms of
the thermal resistance, the film has Ri = 1/(hfA) for the planar case and for the cylindrical case Ri
= 1/(2rhf) where r is the local radius at the film. In the convection case, hc just replaces hf.
Example: Consider the heat transfer through a plate type keel cooler on the outside of the hull.
Hot engine cooling fresh water is cooled by the sea water. To minimize corrosion the outside of
the plate surface is covered with a thin layer of copper cladding.
Using the results for the composite plane heat transfer,
q/A = hci(Tbi – T1) = hfi(T1 – T2) = k1/L1(T2 – T3) = k2/L2 (T3 – T4) = hfo(T4 – T5) = hco(T5 – Tbo)
14 - 7
Finally with q/A, we can find the intermediate temperatures T1 to T5 by working through each Ri
from one of the known boundary temperatures,
Radiation
The second mode of heat transfer is radiation which involves energy transfer by
electromagnetic wave phenomenon. It is important only where higher temperatures are involved.
The primary marine engineering example is heat transfer within a boiler; particularly between
the hot flue gases and flame in the furnace region and the surrounding water wall and water
screen tubes. In general, we are considering energy transfer by the visible light and infrared
wavelengths of electromagnetic radiation.
When radiant energy falls on a body, it absorbs some fraction  (absorptivity), reflects
some fraction  (reflectivity), and may transmit some fraction  (transmissivity) if it is not
opaque. Thus,
 ++  = 1
For most engineering materials,  ~ 0 except in very thin layers. A black body is an idealization
of a material that absorbs all energy  = 1. All real materials are gray bodies with  < 0.99. In
equilibrium, a body must emit as much energy as it absorbs for T to be constant, so
emissivity  = 
For examples, slate and water at right angles  = 0.95 – 0.99

polished steel  = ~ 0.07 at 100 F; ~ 0.14 at 1000 F
oxidized steel  = ~ 0.80
The physical law for radiation, comparable to Fourier’s Law for conduction, is Stefan-
Boltzmann’s Law of Total Radiation which experimentally gives for a black body using an
absolute temperature scale,
14 - 8
q = AT4
where  is the Stefan-Boltzmann constant in English units  = 0.1714 x 10–8 Btu/hrft2R4 or in SI

units 5.6704 x 10-8 W/m2K4. For a gray body this becomes,
q = 1730A[T/1000]4 with T on absolute scale
where  is material dependent and comparable to k in conduction.

When this is applied to heat transfer between two bodies in equilibrium, we obtain,
q12 = A1F12(T14 – T24) = – A2F21(T14 – T24)
where A1 is the area of body 1 visible from body 2 and F12 is a factor incorporating the geometry
(separation distance, orientation, size, etc.) of the two bodies and their grayness 1 and 2.
Heat Exchangers
The marine engineer needs to be able to perform preliminary sizing calculations for marine
heat exchangers. This work is needed in preliminary design and as an initial step in selecting and
specifying the requirements for heat exchangers to be included on a ship. There is also a need to
predict how an existing heat exchanger will perform in off-design conditions when its basic
properties (heat transfer surface area, materials, etc.) are known.
Types of Heat Exchangers. Heat exchanges can be classified in a number of ways. They
can be classified by the relative direction of the flow (parallel, counter, or cross) of the hot and
cold fluids with respect to each other. They can also be classified by the number of passes, the
number of times one fluid moves passed the full range of temperature of the other fluid.
single-pass, parallel-flow
14 - 9
single-pass, counter-flow
single-pass tube side, single-pass shell side counter-flow heat exchanger [1]
two-pass on tube side, single-pass on shell side heat exchanger [2]
If the two fluids cross each other at right angles, the heat exchanger is classified as having
cross-flow. Cross-flow heat exchangers are classified as either mixed or unmixed flow
depending upon whether or not the fluid is constrained to specific paths or channels. In mixed
flow, the fluid is able to completely mix across the width of the flow area. The design may have
one fluid in tubes and the second fluid free to mix around the outside of the tubes. The tubes
might have fins to improve the heat transfer, particularly with air heat transfer that is much less
effective. Other designs have channels in both directions created the mechanical construction.
14 - 10
cross-flow with one fluid mixed and the other fluid unmixed
cross-flow heat exchanger with both fluids unmixed
Log Mean Temperature Difference. We will use the nomenclature defined for the
single-pass, parallel-flow heat exchanger sketched above. The heat exchanger is essentially
adiabatic so it must satisfy:
heat from hot fluid = heat to cold fluid
q [Btu/hr] = mhCph(Thin – Thout) = mcCpc(Tcout – Tcin)
where Cp is the specific heat [Btu/lbmF or J/kg˚C]. The heat exchanger will also obey the
following for the total heat transfer across the heat transfer surface,
q = UoAo“average T” = UoAoTln
14 - 11
where Ao is the total outside surface area of the tubes (o) and Uo is the overall heat transfer
coefficient [Btu/hrft2F or W/m2˚C], which must correspond to this same outside surface area. In
preliminary design, one must establish Ao or at least UoAo so that a heat exchanger can be sized
properly. The temperature difference across the heat transfer surface “T” is complicated
because it is continuously changing along the heat transfer surface. The appropriate “average
T” is the log mean temperature difference (LMDT or Tln) that we will derive for a single-pass,
parallel-flow heat exchanger and then extend to other types of heat exchangers.
At a local element of heat transfer surface area, dA, there will be a heat transfer dq given
by the following if we assume that Uo is constant throughout the heat exchanger,
dq = UodAT, incremental heat transferred across dA
dq = – mhChdTh, incremental heat lost by the hot fluid
dq = mcCcdTc, incremental heat gained by the cold fluid
Here we have dropped the p subscript in Cp for convenience. These yield,
dTh – dTc = d(Th – Tc) = – dq[1/(mhCh) + 1/(mcCc)]
and using dq = UodAT = UodA(Th – Tc) we get,
d(Th – Tc)/(Th – Tc) = – Uo[1/(mhCh) + 1/(mcCc)]dA
This can be integrated to give,
ln(Th – Tc) = – Uo[1/(mhCh) + 1/(mcCc)]A + C
where C is the constant of integration determined by the boundary conditions. We have two
boundary conditions,
1) (Th – Tc) = Thin – Tcin at A = 0 so C = ln(Thin – Tcin)
2) (Th – Tc) = Thout – Tcout at A = Ao, the total heat transfer surface area
so ln[(Thout – Tcout)/(Thin – Tcin)] = – UoAo[1/(mhCh) + 1/(mcCc)]
Finally we can substitute,
1/(mhCh) = (Thin – Thout)/q and 1/(mcCc) = (Tcout – Tcin)/q
14 - 12
and with the definition of the log mean temperature difference Tln = q/UoAo, we must then have
for the single-pass, parallel-flow heat exchanger,
Tmax Tmin
Tln =[(Thin – Tcin) – (Thout – Tcout)]/ln[(Thin – Tcin)/(Thout – Tcout)]
or more commonly seen written as,
LMTD = Tln =(Tmax – Tmin)/ln(Tmax/Tmin)
This final result is the LMTD for a parallel-flow heat exchanger. It also applies to single-pass,
counter-flow heat exchangers if the Tmax is the maximum T and Tmin is the minimum T
across either end of the heat exchange surface or for example,
For more complex heat exchangers involving multiple passes or mixed and/or unmixed
cross flow, a comparable result can be derived. For general engineering practice, however, the
approach is to just use the easy to remember result for the single-pass, counter-flow heat
exchanger and then correct this result using a correction factor F available graphically as a
function of two parameters R and P defined as follows:
R = (mtCt)/(msCs) = (Tsin – Tsout)/(Ttout – Ttin) and P = (Ttout – Ttin)/(Tsin – Ttin)
Thus, we have Tln = FTln single-pass, counter-flow where F = F(R, P)
These correction factors F are shown for four common types of heat exchangers in Fig. 14.2 [1].
Note that in Fig. 14.2, the tube (t) side temperatures are in lower case, the shell (s) side
temperatures are in capitals, i indicates in, and o indicates out. In early design, we usually know
what we want the heat exchanger to do; i.e., mt, ms, Tsin, Tsout, Ttin, and Ttout. This yields Tln.
We also know q [Btu/hr or W]. The sizing problem is then to determine UoAo or Ao given Uo.
14 - 13
Figure 14.2 Correction Factors F to Counter-Flow Log Mean Temperature Difference (LMTD) [1]
14 - 14
Example. Establish the heat transfer area Ao required in an economizer that will heat
100,000 lbm/hr of feed water (Cp = 1 Btu/lbmF) from 285 F to 385 F in the tubes. The
shell side is heated by the flue gas in mixed cross flow. The flue gas enters at 630 F and
leaves at 315 F. The overall heat transfer coefficient for the air-to-water heat transfer is
estimated to be Uo = 15 Btu/hrft2F. We begin by drawing a temperature diagram as if this
were a counterflow heat exchanger to establish Tmax and Tmin.
Overall Heat Transfer Coefficient Uo. We will briefly survey the methods of obtaining
the overall heat transfer coefficient Uo [Btu/hrft2F or W/m2˚C]. We can note three approaches
here as follows:
1) Experience with Typical Heat Exchangers. The data shown in Table 14.1 is useful for
making preliminary estimates [2]. Notice the high values for heat exchange involving steam to
liquids and the low values for heat exchange involving gases.
2) First Principles. With sufficient information, we can apply our results for conduction
through concentric tubes to obtain Uo for a tube heat exchanger that transfers heat between two
fluids. With a metallic tube of inside radius ri and outside radius ro, we can see by inspection of
our earlier results that Uo referred to the tube outside surface area Ao = 2ro will be as follows
when we use the resistance Ri approach,
14 - 15
7
Uo = 1/( RiAo) = 1/[ro/(hciri) + Roxi(ro/ri) + Rfi(ro/ri) + roln(ro/ri)/k + Roxo + Rfo + 1/hco]
i=1
convection oxide film fouling film tube wall same terms on outside
on inside inside inside
You should convince yourself that this is a correct result by referring back to the earlier
development. The (ro/ri) in the first three terms is because the resistances (1/hci, Roxi, and Rfi) are
defined for the inside surface 2rik, but Uo is required for the outside surface 2rok. Here the
oxide films are reasonably predictable based upon tube material, fluid type, operating
temperature, etc. The fouling terms are much less predictable and often the fouling terms are
omitted from the above equation and the resulting Uo clean is reduced using a cleanliness factor,
F3, based upon experience (usually 0.85 to 0.90). This would yield,
5
Uo = F3/( RiAo) = F3/[ro/(hciri) + Roxi(ro/ri) + roln(ro/ri)/k + Roxo + 1/hco]
i=1
Table 14.1 Approximate Overall Heat Transfer Coefficients for Preliminary Estimates [2]
14 - 16
3) Empirical Formulae. There are a number of empirical formulae which give acceptable
values for Uo for design purposes. As an example, the formula from Marine Engineering [4] for
marine condensers is shown in Figure 14.3. The F3 (~ 0.85) in the condenser formula is the
cleanliness factor described above.
Figure 14.3 Heat Transfer Coefficient Model for Marine Condensers [4]
Heat Exchanger Effectiveness. Once a heat exchanger is sized for its rated condition, Ao
is fixed. Also Uo is essentially constant. If not, you could recalculate Uo for conditions other
than the rated condition. If we operate the heat exchanger in off-rated conditions, we usually
know UoAo, inlet temperatures, and fluid flows. The resulting outlet temperatures are the
unknowns and we can no longer calculate Tln directly. From the above derivation, we had the
following for parallel flow,
14 - 17
ln[(Thout – Tcout)/(Thin – Tcin)] = – UoAo/(mhCh)[1 + (mhCh)/(mcCc)]
or (Thout – Tcout)/(Thin – Tcin) = exp[– UoAo/(mhCh)[1 + (mhCh)/(mcCc)]]
Thus, if we know any three of the inlet and outlet temperatures, the flows, and UoAo, we can
obtain the fourth temperature.
Usually, however, both outlet temperatures are unknown. The heat exchanger effectiveness
has been derived as an aid in solving such problems. We will derive this result for the parallel-
flow heat exchanger considered above. We had,
q = mhCh(Thin – Thout) = mcCc(Tcout – Tcin)
giving Tcout = Tcin + mhCh/(mcCc)(Thin – Thout)
which can be substituted into the above equation to give,
(Thout – Tcin)/(Thin – Tcin) – mhCh(Thin – Thout)/[(mcCc)(Thin – Tcin)]
= exp[– UoAo/(mhCh)[1 + (mhCh)/(mcCc)]]
Now if we add 1 – (Thin – Tcin)/(Thin – Tcin) = 0 to the left side and manipulate, we have
1 – (Thin – Thout)/(Thin – Tcin) – mhCh(Thin – Thout)/[(mcCc)(Thin – Tcin)]
= exp[– UoAo/(mhCh)[1 + (mhCh)/(mcCc)]]
which gives the heat exchanger effectiveness h,
h = (Thin – Thout)/(Thin – Tcin)
= [1 – exp[– UoAo/(mhCh)[1 + (mhCh)/(mcCc)]]/[1 + (mhCh)/(mcCc)]
This is the heat exchanger effectiveness we used earlier in Chapter 5 for gas turbine regenerators
and intercoolers. The effectiveness is a function of the two parameters UoAo/(mhCh) and
(mhCh)/(mcCc). Thus if we know UoAo, mhCh, and mcCc and the inlet temperatures, we can find
Thout.
Heat exchanger effectiveness ratios have been derived in a more general form for various
types of heat exchangers and are available graphically as shown in Fig. 14.4 [1]. If we now
14 - 18
define, Cc’ = mcCc [Btu/hrF or W/˚C] and Ch’ = mhCh and designate the maximum of these as
Cmax and the minimum as Cmin, we can plot  as a function of two parameters, the number of heat
transfer units (NTU) = UoAo/Cmin and the ratio Cmin/Cmax. The general results are then,
 = (heat exchanger type, NTU, Cmin/Cmax)
q = Cmin(Thin – Tcin) , to give the new q [Btu/hr] given inlet T’s, flows, and UoAo
 = Ch’(Thin – Thout)/[Cmin(Thin – Tcin)] , for the new Thout
 = Cc’(Tcout – Tcin)/[Cmin(Thin – Tcin)] , for the new Tcout
where the  is available in Figure 14.4.
Example. Suppose we design a counter flow lube oil cooler to cool 80 U.S. gpm of oil
(Cp = 0.48 Btu/lbmF,  = 54.3 lbm/ft3) from 150 F to 120 F using seawater (Cp = 0.94
Btu/lbm,  = 64.0 lbm/ft3). The seawater inlet is 75 F. To prevent marine growth the
seawater outlet is limited to 95 F.
a. Assuming single-pass and Uo = 50 Btu/hrft2F, determine the heat transfer area and
the required seawater flow rate.
14 - 19
b. If the oil flow is reduced to 50 U.S. gpm and the seawater flow is left unchanged,
determine the resulting oil outlet temperature.
References
1. Incropera, F. P. and DeWitt, D. P., Introduction to Heat Transfer, 3rd Edition, John Wiley &
Sons, New York, 1996.
2. Kreith, F. and Bohn, M. S., Principles of Heat Transfer, 6th Edition, Brooks/Cole, 1993.
3. Jakob and Hawkins, Elements of Heat Transfer, 3rd Edition, John Wiley & Sons, New York,
1957.
4. Foster, R. J., “Main and Auxiliary Condensers,” Ch. XV in Harrington, R. L. (ed.), Marine
Engineering, SNAME, Jersey City, NJ, 1992.
14 - 20
Figure 14.4 Heat Exchanger Effectiveness for Various Types of Heat Exchangers [1]
14 - 21
14 - 22
Appendix A Excerpts from Crane Technical Paper No. 410M – [Chapter 13, reference 6]
Appendix B SI Units Thermodynamic Tables for Water [Chapter 2, reference 1, pp. 664-681]

Bai Giang MayTau Michigan

Uploaded by

Copyright:

Available Formats

Bai Giang MayTau Michigan

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bai Giang MayTau Michigan

Uploaded by

Copyright:

Available Formats

Informal Course Notes for

Professor Emeritus Michael G. Parsons, FSNAME

Version: September 2015

Department of Naval Architecture and Marine Engineering

energy prime transmission/

I.C. = internal combustion

Figure 1.1 Most Common Marine Propulsion Alternatives

Figure 1.2 Schematic of CODED Propulsion Plant for High-Speed Ferry

• Specific Fuel Rate. A specific fuel rate is an expression of thermodynamic efficiency.

• Rating Limitations. Marine machinery is usually available only in discrete sizes. It is

• Management Preference. Some ship owners have specific desires relative to

Chapter 2 Review of Thermodynamics

2.1. Pressure, Temperature, and Other Properties

pressure above atmospheric pressure

The standard atmospheric (barometric) pressure is 760 mm of mercury (Hg) in SI units

1 atm. = 760 mm Hg = 0.101 325 MPa absolute = 101.325 kPa absolute

Table 2.1 Most Common SI Unit Prefixes [2]

definition prefix symbol

1 N = 1 kg m/s2 (about ¼ lbf)

1 bar = 0.1 MPa = 105 Pa = 14.50379 psi

1 kgf/cm2 = 0.980 665 bar = 98 066.5 Pa = 14.223 36 psi

1 lbf = 4.448 222 N = 0.453 59 kgf

Temperature. Temperature can be expressed either on an absolute or common

The two common scales are related by

P vaporization line (Tsat, Psat)

Figure 2.2 Pressure-Temperature Vaporization Line for Water

saturated constant P line

Figure 2.3 Schematic T-v Diagram for Water

2.2 Heat, Work, Energy, and Power

as the movement or transfer of energy by virtue of a temperature difference. Heat is defined to

1 Btu = 778.17 ft lbf = 1.0550 kJ.

Table 2.2 Unit Conversion Factors Used to Resolve Units

778.17 ft lbf/Btu = 1 32.174 lbm ft/(lbf s2) = gc = 1

1 Nm/J = 1 1 kg m/(N s2) = 1

1 hpBr = 550 ft lbf/s

1 hpM = 4500 kgf m/min

The two horsepowers relate to the kilowatt as follows:

The three units of power relate to the Btu as follows:

1 hpBr = 550 ft lbf/s • 3600 s/h /(778.17 ft lbf/Btu) = 2544.43 Btu/h

2.3 The First Law of Thermodynamics

The First Law of Thermodynamics is a statement of the conservation of energy. It states

This is an experimentally observed fundamental law of nature.

Q + Q = W + W

(Q – W) = Q – W)

Using paths A and C, we similarly have,

(Q – W) = Q – W)

Subtracting these last two equations and rearranging we obtain,

(Q – W) = Q – W) (2.2)

1Q2 = U2 – U1 + m (V22 – V12)/2 + mg (z2 – z1) + 1W2 (2.4)

This is another statement of the First Law of Thermodynamics. This is a statement of

1 q2 = 1Q2/m = u2 – u1 + (V22 – V12)/2 + g (z2 – z1) + 1w2 (2.5)

Figure 2.7 Demonstration of the Existence of Thermodynamic Property Energy

By definition, enthalpy is a combination of thermodynamic properties so it too is a property, it

m= VdA = V = V/v (2.6)

control volume (c.v.)

constant mass m cv area Ae

exiting flow Ve, me

Figure 2.8 Steady-State, Steady-Flow Model

Conservation of mass (continuity) applied to the control volume requires that,