The Choice of the Propeller'

By J. D. v a n M a n e n 2

in this paper the four main requirements for a propeller are dealt with. These require*
ments concern efficiency, cavitation, propeller-excited forces and stopping abilities. In a
propeller diagram the characteristic efficiency curves for different conditions are explained. A comparison of the optimum efficiencies for various types of propulsors is
given, and the applications on a 130,000-dwt tanker are considered. Cavitation-inception curves both for a specific propeller and for systematic propeller series are discussed. Predicted torque and thrust fluctuations, based on model*test data, and the
results of measurements on the full-size ship are compared. Finally a quasi-steady testing
technique, developed to analyze different types of stopping maneuvers, is described.

IN this paper an a t t e m p t is made to explain in an

instructive way the results of applied research in the
propulsion of ships. In particular it is hoped that it
will be instructive to those who are active in the field of
ship design.
The main requirements for a ship propeller are:
1 High efficiency.
2 %linimum danger of cavitation erosion.
3 2~/[inimum propeller-excited vibratory forces.
4 Good stopping abilities.
5 Favorable interaction with the rudder, to improve
maneuverability.
6 Dependability--minimum
vulnerability.
7 Low initial and maintenance costs.
In the following sections the author has restricted himself to a discussion of the requirements mentioned under
points 1 through 4.
There are two important diagrams, with the aid of
which insight into nearly every propulsion problem can
be obtained. These are:
(a) The diagram giving the relation between the
thrust coefficient Kr and torque coefficient KQ and the
advance ratio J of the propulsor is shown in Fig. i.
(b) The velocity and force diagram of a screw-blade
element is shown in Fig. 2.
Results such as those given in Fig. 1 were obtained
from an "open-water test" of a given screw model. In
such tests the screw model is driven from behind. The
propulsion motor and the measuring apparatus are
housed in a boat which is a considerable distance behind
the screw model and is connected to the carriage of the
towing tank. In this manner the thrust T and the
torque Q can be measured for constant values of rotative
speed n and varying advance speed Va without the
x Lecture held at a Seminar of The Royal Institute of Engineers,
Delft, The Netherlands, 1965.
2 Assistant Director, Netherlands Ship ~[odel Basin, Wageningen, The Netherlands.
158

KT=

~'----q Q"O--SrZ

KT

~P=~

Va

" o-5

0,6F
g

"

KI
~ 0,2

. . . . . . . . . .

0.2

0.4

0,5

0,8

l.O

12

--j

Fig. t

Relation between thrust coefficient K~, torque coefficient

KQ and advance coefficient J of propeller

influence of the ship t h a t ultimately is to be driven by

the propeller.
As a rule, thrust and torque are given in nondimensional form
T
K T

"

p(nD)2D

T
2

q
I![Q --

p(nD) 2D2D

pD4n 2

(thrust coefScient)

pD~n 2

(torque coefficient)

where
p = mass density of fluid
a measure of rotative speed of screw
D 2 = a measure of screw-disk area
nD=

These thrust and torque coefficients are plotted as a

function of the advance coefficient J, which is the ratio
between the speed of advance Va and the rotative speed
nD.

%/lost design problems can be solved for a particular

MARINE TECHNOLOGY

' f
id

[dT i

~,

/c. - ~ /
.~:,:-x/~i~ -/

rr/2

~
,

........

;va

P/{
dQi

dO

Fig. 2

Force and velocity diagram of screw-blade element~at

radius r
5

-Bp

10

15

20

30

40

50

60

t,0

Fig. 4 Bp -- ~ diagram

......

0,E

a,

o.a.

11.8 " " "

~\\\x \~\!/_~/\ \a \ .~/

1

O.t

1.1

"A',,. /
0,2

\.*/ \ ~ / -

1.3

0.9

%
&2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

J 0.7

Fig. 3 KT-KQ-Jdiagram
0,5
5

screw using a diagram such as Fig. 1. If, for instance, for

the screw concerned, the speed of advance Va and the
rotative speed n are known, then the thrust and torque
can be read off, or, with known Va and Q the rotative
speed n can be found. Thus, when two out of the four
quantities Va, n, T and Q are given, the other two can be
determined from the diagram.
In Fig. 2 the force and velocity diagram of a screwblade element at a certain radius r is given. The thrusting action of the screw induces velocities in the fluid.
The magnitude of these induced velocities depends on the
screw loading. If the speed of advance Va (or ship
speed Vs) is decreased, while the rotative speed is held
constant, the screw loading will increase and the induced
velocities will increase at a rate proportional to the increase in lift force dL and the effective angle of attack
OZl.

The induced velocities G, which are, to a good approximation, at right angles to the resultant incoming
velocity V, can be resolved into axial and tangential
components c~ and c~. At the screw disk the induced
velocities are one-half of their ultimate values far behind
the screw.
IlK the diagram, Fig. 2, the following symbols are indicated:
c~ror 7rnd = tangential speed of blade element at
radius r
dD~
= profile drag of blade element
APRIL 1966

-Bp

10

15

20

30

40

50

60

Fig. 5 Characteristic curves in a B~-3 diagram

dT~, dQ.; = thrust and torque force of blade element

without influence of profile drag

dT, d(2 = thrust and torque force including influence

fl
3~

of profile drag
= hydrodynamic pitch angle uncorrected for
induced velocities
= hydrodynamic pitch angle corrected for
induced velocities

This force and velocity diagram forms the basis for the
lifting-line theory for ship propellers. This theory will
be treated in greater detail in the section "Cavitation of
the Screw-Blade Sections."
This diagram is also helpful when analyzing propulsion
problems using quasi-steady considerations, see also subsequent sections.
Efficiency of the Propeller

An important source of data for screw design are the

results of open-water tests with systematical screw series.
A systematic screw series consists of a number of screw

models, in which only the pitch ratio P/D is varied. All

other characteristic dimensions, including diameter D,
number of blades z, blade-area ratio Ao/A, blade planform, form of blade sections, blade thicknesses and hub159

Table

T r a n s f o r m a t i o n o f the

K~

KQ

Diagram

Into B~-6 D i a g r a m

KQ

,S

200

0.506

210

0.482

j2s

~D=0.6

0.8

1.0

0.182

0.0100

00234

00398

0.161

0.0108

0.0244

0.0410

Bp
1.2

1.4

0.6

08

1.0

1.2

1.4

0.0612 0.0840

18.2

27.8

36.3

45,0

52.3

0.0624

0.0854

21.3

32.1

41.5

51.3

60,0

36.7

47.3

58.1

679

220

0460

0.144

0.0115

0.0254

0.0420

0.0636

0.0868

24.7

230

0.440

0.128

0.0122

0.0262

00430

0.0647

00879

28.4

4 1 . 7 53.4

65.5

76.4

240

0.422

0116

0.0128

0.0269

0.0438

0.0656

0.0890

32.4

4 7 . 0 60.0

73.4

85.5

250

0.405

0.104

0.0134

0.0275

00446

00666

0.0900

36.7

52.3 67.0

81.8

95.2

260

0.390

0095

0.0138

0.0282

0.0453

0.0672

0.0909

41.0

5 8 . 4 74.1

90.3

1050

270

0.375

0.086

00143

0.0287

0.0460

0.0680

0.09t7

45.9

65.1

82.4

100.2 116.3

280

0.362

0.079

0.0148

0.0292

0.0466

0.0687

00924

51.0

71.7

90.6

110.0 127.6

79.3 I00.0

121.1 140,5

8 6 , 4 108.8 131.6 152,5

290

0.349

0.072

0.0152

0.0297

0.0472

0.0693

0.0932

56.7

300

0.339

0.066

0.0156

0.0301

0.0477

00698

0.0937

62.2

j = 101.277

8p=33.08\/ ~

diameter ratio dh/D are fixed for this series. The results
of open-water tests for such a screw series are given in the
Kr-KQ-J diagrams, Fig. 3.
The propeller efficiency ~ can be expressed in terms of
these nondimensionM coefficients as follows:
TVa
K~. J
~ - 2~rQn - KQ 2~r
By interpolation in the K~.-I~-J diagram of a screw
series most problems, which arise when designing or
analyzing screw propellers, can be solved.
Tile most widely encountered design problem is t h a t
where the speed of advance of the fluid into the screw
disk Va, the power to be absorbed by the screw P and
the number of revolutions n are given. The diameter D
is to be chosen so t h a t the greatest efficiency can be obtained. This is done as follows:
By choosing discrete values of the diameter D, the
corresponding values of the advance ratio J and the
torque eoettlcient K can be calculated. From the K~KQ-J diagram, Fig. 3, the corresponding pitch ratios P / D
and the efficiency ~; can be read off for each diameter
chosen. Plotting the values of ~ as a function of the
diameter will allow the diameter leading to the o p t i m u m
etficieney to be chosen.
I n order to simplify this frequent design problem, the
K~-KQ-J diagrams can be transformed into another diagram, from which the o p t i m u m diameter D can be read
off directly when the speed of advance Va, the power P
and the rotative speed n are given. For this purpose a
design coel~cient B,~ has been formed from the torque
coefficient K and the advance ratio J in such a way t h a t
the screw diameter is eliminated:

B~
160

NPV~ - 33.08 (ICQ'~V~

Va2V~

\ j; ]

I n the coefficient B~, N is the number of revolutions per

minute, the power P is in horsepower and the speed of
advance Fa is in knots (1 knot = 0.5144 m / s e e = 1.689

fps).
I n the usual diagram, the design eoelficient ]3~, is
the base and a new speed ratio 5 is used. This speed
ratio is defined as
ND
Va

101.27
J

in which D = screw diameter in feet.

The manner in which the Ks-K~-J diagrsbm is transformed into the B,j-8 diagram is shown in Table 1.
Fig. 4 gives an example of a B~-5 diagram for a particular screw series.
In Fig. 5 some characteristic curves in the Bp-5 diagram
are shown:
(a) O p t i m u m ~,, for P / D = eonst. This curve goes
through the points where the tangents to the curves of
equal efficiency (vp = const) are horizontal ( P / D =
const). The o p t i m u m ~,,-values correspond to the peaks
of the ~h~-curves in ;he K~,-Ke-J diagram, Fig. 3.
(b) O p t i m u m r,~, for J = const. This curve goes
through the points of contact between the curves of 3 =
const and those for s~ = const. These o p t i m u m s~values coincide with those on the envelope of the efficiency
curve in the K,r-KQ-J diagram, Fig. 3.
(c) O p t i m u m ~ for the most favorable diameter D.
This curve connecbs the points of contact between the
curves of ss = const and their vertical tangents (Bp =
const; P, N and Va are given).
(d) O1)timum v,, for the most f~vorable number of
revolutions N. This is the locus of the points of contact
between the curves of constant efficiency (,~ = eonst) and
MARINE

TECHNOLOGY

P
COASTERS
.
.
.
.

TWIN-SC.REW SINGLE SCREW

SHIPS
i CARGO SHIPS
.

TANKERS

0,80

TRAWLERS

PROPELLERS

'
060

SERIES 4 - 7 0
'

o,Ts

Vs

- -

SHP

I ~ ~)O.7,00m

.1s.8 k . - -

=27720hp

~-_.
3",~

,~ D, 6 7 4 m

<I~T~

6142m

~.-

f
W ....

PROPELLERS

- -

R...,oo

PROPELLERS [

~ -

T A N K E R : D I S P L =156.000 m 3
dw
=130,000ton

TUGS

--

- -

D=8.00m

'

0.50

o.es

0=7.60m

o.7.2om

~5.78m
1 ~

O. 7/,7,
I . (~

I 0,

--

~?(--

~p
i opt.
O . 6 0 L - - ~
Z,

0.40
FULLY CAVITATINB PROPELLERS
J
VERTICAL AXIS
o.

I0

'

15

20

"~
"

~"

OF B L A D E S

I
25

30

40

50

6(

70 80 90 fGO

I25

~50

200

Comparison of optimum efficiency values for different

types of propulsors

Table 2
NUMBER
OF
BLADES

curves on which P, D and Va have constant values.

These curves can be constructed easily starting from a
certain value of B~ and ~ (for instance points on the ~,optimum curve for optimum diameter) and reducing
the rotative speed by, say 10, 20, 30, 40 and 50 percent
The B , and ~-values will then also be reduced by 10, 20,
30, 40 and 50 percent.
The four typical curves coincide at one point in the left
upper corner of the diagram. This point corresponds
to the maximum of the envelope of the efficiency curves
in the open-water characteristics, Fig. 3.
If both the screw diameter D and the rotative speed
are free to be chosen for a given power P and speed of
advance Va, the optimum propeller will correspond to
this optimum point. As a rule, however, either the
resulting diameter is too large or the resulting number
of revolutions too small for practical purposes.
It is of great importance that the propeller designer
makes himself thoroughly familiar with one type of design
chart. I t is advised, therefore, that the propeller designer restricts himself to the open-water characteristics
(Kr-K~-J diagrams) and the B~-8 diagrams when making
use of ,systematic screw-ser~es data. If rJecessalT, the
open-water characteristics can be transformed in such a
manner that the required calculations for any given design requirements are reduced to a bare minimum. Examples are the B,-8 diagrams for given values of the
speed Va, rpm N and thrust T and the t*-~ diagrams for
the construction of the tow-force diagram of a tug [1].a
The advantage of the reduced amount of calculations
when using a special diagram does not outweigh the disadvantages arising from the use of an unfamiliar diagram
In Table 2 a survey of the systematic screw series
tested by the NSMB is given. Results are available in
the four fonns discussed previously [1-3].
a Numbers in brackets designate References at end of paper.
APRIL 1966

NUMBER

CR. O.P Twin Tripte

scr, scr.
3-3 3-5-3

Fig. 7 Propulsive coefficients for a 130,000-ton tanker

PROPELLER5

8p

Fig. 6

- -

3
4
5
6
7

Survey of

WageningenB-Series

BLADE AREA-RATIO
0.30

0.38
0.36

0,50
0.40

0.65
055

0.45

0,80
0.70

0,60
0.50

0.75
0.65

0.56

0.85 1.00
1,05
0.80

0.70

0.85

Open-water tests have been performed with systematic

series for types of propellers other then the conventional
screw. Some of those results are summarized in Fig. 6.
In this figure curves are given showing the highest
obtainable efficiency ~p for different types of propellers
as a function of Bj~. At the top of the figure the ranges
of B~-values typical for different ship types are indicated
The lightly loaded screws of fast ships are at the lefthand side, while the heavily loaded propellers of towing
vessels are at the right. Such a diagram can give a quick
indication which type of propeller will give the best
efficiency for a given type of ship.
In particular, it can be seen from the diagram that
propellers in nozzles are to be recommended for heavy
loadings, such as occur in tugs, trawlers and large tankers.
For heavy towing a long nozzle with a chord-diameter
ratio of 0.83 is preferable to a short nozzle with a chorddiameter ratio of 050 [4].
For fast ships, contrarotating propellers appear to give
a higher efficiency than conventional screws, such as
the B 4 . 7 0 or broad-bladed screws such as the Gawn 3110.
Information on the optimum efficiency of fully eavitating and vertical-axis propellers has been included in the
diagram [5], [6 ].
Fig. 7 gives the results of calculations of the propeller
for a tanker with a deadweight of 130,000 tons, a ship
161

Table 3

Required Power for a Tanker

Displacement volume = 156,000 m3; deadweight = 130,000 tons

RPM = 100; V, = 15.8 knots; 100 percent SHP = 27,720 hp.
D = 7.20 m
Conventional screw z
z
z
z

=
=
=
=

4, percent ....
5, percent ....
6, percent ....
7, percent ....

104
103
103
102

7.60 m
100
100
100
101

8.00 m
96
98
100
103

Counterrotating propellers
z = 4,5;

D = 6.57 -- 5.78 m

95%

Propeller with nozzle

D = 7.00 m
Resistance increase due to
ttogner stern and nozzle attachment
. . . . 6 percent

89%

Twin-screw ship
z = 3 - - 3 ; D = 7.47m
Resistance increase due to bossings
. . . . 5 percent

100%

Triple-screw ship
z=3--5--3;
D = 6.74 -- 6.42 m
Resistance increase due to bossings
. . . . 4 percent

Fig. 8 Arrangement of propeller in nozzle for a tanker model

94%

speed of 15.8 knots and an installed power of 27,720 shp.

The rotative speed of the propeller has been fixed at 100
rpm. The propulsive coefficient, which is a measure of
the efficiency of the propeller including the hydrodynamic
interaction of the ship's hull and the propeller, is plotted
vertically.
A t the left-hand side of the diagram results are given
for conventional ship propellers with 4, 5, 6 and 7 blades.
The calculations have been carried out for three different
diameters: D = 7.20, 7.60 and 8.00 m. I n order to give
a complete survey of the potentialities of a conventional
screw propeller it is necessary to carry out calculations
for a range of number of blades, diameters and rotative
speeds. I n this diagram the rotative speed has been
fixed at 100 rpm and it is obvious t h a t for the higher
number of blades (6 and 7) a screw diameter of 8.00 m
is too large; the highest efficiency is found for a diameter
of 7.60 m. For the lower number of blades the efficiency
continues to increase with increasing diameter.
I t might be the question if a diameter of 8.00 m gives
already the optimum for a 4-bladed propeller for this
ship. For lower rotative speeds, for instance 80 rpm f o r
turbine-driven ships, these tendencies are intensified.
To the right, results are given for a set of contrarotating
propellers (CR) and a propeller in a nozzle (DP = ducted
propeller). Results are also given for a twin-screw
arrangement of two 3-bladed propellers and a triplescrew ship with two 3-bladed and one 5-bladed propeller. The increase in resistance ( E H P ) due to the
nozzle a t t a c h m e n t or due to the brackets for the twinscrew or triple-screw configuration are specified to be 6, 5
and 4 percent, respectively. For each configuration the
optimum diameter, at 100 rpm, is indicated.
162

I n Table 3 the required power for each case is given,

using 27,720 shp as 100 percent.
The large saving in S H P which can be obtained using
a shrouded propeller should be noticed.
As tanker sizes continue to increase, the advantage
of shrouded propellers will become greater. I n Fig. 8,
the installation of a propeller in a clear-plastic shroud
on the afterbody of a tanker model is shown.
For the very big tanker the contrarotating propeller
leads to a power reduction of 5 percent compared to
conventional screws. The advantage of this type of
propulsion for large tankers lies primarily in the smaller
allowable screw diameters and thus a possible reduction
in the danger of vibration caused by the fluctuating
propeller-force field. As shown by Fig. 6, the contra~
rotating propeller m a y become a serious competitor of
the conventional screw propeller for fast cargo ships [7 ].
Cavitation of the Screw-Blade Sections

The lifting-line theory for ship screws gives an idea of

the induced velocities, pressures and forces along the
different radii of the screw blades. This information
forms the basis of considerations of cavitation and
strength.
Fig. 9 gives schematically the screw with a bound
vortex or circulation F at the radius r and trailing helicoidal vortices. The circulation F of a screw-blade section is defined as the line integral of the flow field along
a closed curve around this profile. This line integral
is the integration of the product of a line element ds and
the component of velocity tangent to this line element.
B y choosing the closed curve along which the line integral
is to be evaluated, in the manner indicated in Fig. 9, a
relation can be derived easily between the circulation F
of a screw-blade section and the tangential-induced
MARINE TECHNOLOGY

1.00

0.75 -

2trrc t .
..7

zr
0,50

CL
0.25
i

Fig. 9

Relation between circulation F and tangential induced

velocity ct
0.S

Fig. 11

CL -- t / l

1.0

&p
q

1.5

2.0

2.5

A p / q diagram for circular-arc profiles

- -

0.8

0.7

"

ADDITIONAL UNSTEADY

).80

~/~

0.6

0.5

/f
0,

Fig. 10

Oosterveld's h y d r o d y n a m i c m o d e l for screw blade as

vortex sheet

velocity c~ at the same radius r. The curve shown in

Fig. 9 consists of a coaxial cylinder, of radius r, cut open
along a line parallel to the axis. F a r forward of the
propeller c~ = 0, so t h a t t h a t part of the boundary has
zero contribution to the li~le integral. The two longitudinal boundaries make equal but opposite contributions
and, hence, cancel each other. Far aft, the contributim~
is 2retch. When the axial cylinder is flattened out, as at
the right-hand side of Fig. 9, the choseu boundary surrounds the bound vortices of the z-screw blades and thus,
according to Stokes' law, we find
zF = 2rrrc,

For a finite number of blades, say 4 or 5, the value c,

/'or the tangential induced velocity far aft the screw will
occur only at the helicoidal vortex sheets. Between
these vortex sheets the tangential induced velocities will
be less. On a circle with radius r the tangential induced
velocities will have an undulating magnitude with a maxin m m c, at the free vortices and a minimum in between.
T h e ratio between the mean and the m a x i m u m induced
tangential velocity is given by the Goldstein reduction
factor:

0,3
CL
I 0,2

o.,
o

//;"
0.02

APRIL

1966

get

0.06

t/t

0.0~

0.I0

0.12

0.14

OJ6

O.'le

Fig. 12 Relation between lift coefficient CL, ratio t / l and

cavitation number ~r, respectively, pressure coefficient Ap/q for
circular-arc profiles

I t follows for practical screws t h a t for the circulation

around the screw blade we have
Z r = 27rrCtme~n = 27rVKCt

The product of the lift coefficient and chord I of a

blade section is a basic p a r a m e t e r in the calculation of
the cavitation properties of the blade section. According
to the law of Kutta-Joukowski the lifting force of a
screw-blade element, see Fig. 2, is
dL = pVFdr

where V is the resultant velocity of the screw-blade section. B y definition

d L = CL}pV2ldr

and hence
Cd

{~trnean~

I
0.04

2 r _ 47rrKc, _ 47rD @x) -}c~_

V
zV
z
V
163

Lo

'

: T =

~.

//

Mcr

V2

v'

-1

\F'ACE OR

MT

~ NACA 6-SEI~IES
:

\ ' VISIBLE
\

oES,GN CL-O

".

"?"

//

kD

'

-04

-02

oi,~//-i
0

0~. 2 1 / .....L
02
04
--CL

L~
08

06

1Jo
I
L

R e l a t i o n b e t w e e n critical M a c h n u m b e r and p r e s s u r e
coefficient Ap/ q

......
-- I H R U S T C O E F F I C I E N T

where z = nondimensional radius = r//~.

With the aid of Fig. 2 a geometric relation can be
derived between lc,/V and the hydrodynamic pitch
angles 3 and 3~
~-C t

s i n 3, t a n

(3, -

3)

For a known velocity distribution the forces on the

screw-blade element can now be calculated easily.
This paper will be restricted to these elementary remarks based on the steady lifting-line theory for ship
propellers. Further details can be found in the available
literature, such as [3]. For a clear insight into the propeller action, the construction of more complicated hydrodynamic models m a y often be necessary. Theories
have been developed in which the screw blades are regarded as vortex sheets, both for steady mid unsteady
phenomena. Fig. 10 gives Oosterveld's hydrodynamic
model for the screw blade as a vortex sheet. The additional unsteady vortices are created when the screw
passes a region of low speed of advance, such as occurs
when the blade passes the stern post.
The creation of lift at the different radii of the screw
blade can be accomplished by camber and b y angle of
attack. The importance of the distribution of lift between these two for good cavitation properties can be
determined from:
(a) Characteristics for two-dimensional profiles giving the relation between the lift coefficient CL, the angle
of attack a and the cavitation number % or the pressure
coefficient Ap/q, see Figs. 11, 12 and 13.
(b) Cavitation-inception curves such as a, -KT diagram for a ship screw, Fig. 14.
(c) Cavitation-inception curves superimposed on a
B~-8 screw-series diagram, Fig. 15.
I n Fig. 11 the m a x i m u m pressure coefficient Ap/q =
(V 2. . . . V2)/V 2is given as a function of the lift coefficient
CL for four circular-are profiles with thickness-chord ratio
of 0.04, 0.07, 0.11 and 0.15.
Three areas can be discerned:
(a) An area with low CL-values, where a small change
164

/ BACK OR

. . . .

Fig. 13

2= , Vm

Ap

M~=v-~,/c-

Fig. 14

KT

Characteristic curves in a a - K T diagram

in CL (and, h e n c e iu angle of attack a) gives a large

change in Ap/q (pressure-side cavitation).
(b) An area where the lowest Ap/q-values are reached
and a change in C~ gives almost no change in Ap/q (the
area of shock-free entrance}.
(c) An area of high CL-values where a small chauge
in CL corresponds to a large change in Ap/q (suction-side
cavitation).
I n Fig. 12 the dsJ;a of Fig. 11 have been replotted in a
manner, more instructive to the screw designer. This
figure indicates, for a certain value of z (or Ap/q) the
thickness-chord ratio t/l where a m s x i m u m variation in
the lift coefficient C'L can be sustained without cavitation.
T h e point at the extreme right-hand side of' the C~-t/l
loop for a given Ap/q-value is tlhe point where a is about
0 deg and where shock-free entrance changes into pressure-side cavitation. For calculation of propeller cavitation according to steady lifting-line theory, a margin of
safety against cavitation is generally used in the calculations, for instance ~xp/q = ~ -- 20 percent.
Fig. 12 shows ths,t this reduction in cavitation nmnber
affects only the determination of the o p t i m u m thicknesschord ratio t/1 in order to obtain a m a x i m u m C~ variation at a given Ap/q.
Experimental ai~(t theoretical data, such as given in
Figs. 11 and 12, ~re scarce. An extra source for information, however, is given by the calculations for the
inception of supers(talc phenomena in aerodynamics [8 ].
There is a simple re?ation between the m a x i m u m pressure
coefficient Ap/q of :L profile and the critical M a t h number, as is shown in Fig. 13. The critical M a t h nmnber is
the ratio between t h a t advance velocity V, when at some
point on the profile the velocity of sound c has been
reached, and t h a t Yelocity of sound:
Mar

--

V
c

V
V .....

Fig. 13 gives ait example of the relation between the

critical M a t h number M~, and the lift coefficient C~ of a
MARINE

TECHNOLOGY

'

N,

~-SUCTION

1,6

__

SIDE CAVITATION

........

._

1.4

1.2

1.e

0.8

0.6

1.8

2.2

2.6

-v%

Fig. 15

0.8

3.0

3.4

3.8

5.4

5.8

6.2

Vibration Generated by Propeller

W
0,2

~/~

~/2

~/4

--0

Fig. 16

Axial wake distribution b e h i n d single-screw ship

model

profile without camber but with six thickness-chord

ratios [8 ].
For the investigation of screw models in the cavitation
tmmel it is important to show the different results in a
characteristic diagram. Fig. 14 gives, in an instructive
manner, the onset of the different types of cavitation.
Such a diagram can be made for every screw model by
systematically varying the cavitation number ~ at certain values of the thrust eoefflcient K~. From such a
diagram it can be ascertained if the screw model is too
near the inception of pressure-side cavitation at the
design condition. B y reducing the camber somewhat
and compensating for this camber reduction by a pitch
increase, the curves for the onset of pressure-side and
suction-side cavitation can be shifted to the left, making
pressure-side cavitation less likely.
Tests in the cavitation tunnel with systematically
varied screw series can give data about the onset of pressure-side or suction-side cavitation. Burrill [9] has
systematically tested some screw series and has indicated
the lines for cavitation inception in a B~-6 diagram, see
Fig. 15. From this diagram it is evident that the curve
APRIL

5.0

for optimum diameter for a given power P, rotative speed

Y and speed V a , lies in the region of suction-side cavitation. By a slight change in camber of the blade sections
the area more favorable from the viewpoint of cavitation
can be moved toward the position of the Dopceurve.

...........

0.4

4,6

Cavitation-inception curves in a B~-6 diagram

0.6

4.2

1966

Usually the variations of the flow field at the screw

can be split up into two components, i.e. :
(a) The radial variation, especially of the axial veloeities. This variation does not lead to unsteady phenomena at the screw. A propeller working in such a
velocity field has a steady flow and force pattern. Moreover, the propeller can be adjusted to this radially nonuniform flow by an appropriate distribution of pitch and
camber, and optimum efficiency and cavitation properties may be expected in such cases.
(b) The circumferential (at a given radius) variation
of both axial and tangential velocities. This nonuniformity is the origin of the periodically fluctuating force
pattern and the unsteady pressure distributions along the
blade chords, and determines the dynamic pressure
fluctuations induced by the propeller on the stern.
In Fig. 16 art example is given of the wake distribution
behind a single-screw ship model. In this figure only
the axial velocity component is given in the form of the
local wake fraction w = (Vo - - V ~ ) / V , , where V, is the
ship speed.
Although often attempted [10], the experimentM
determination of the tangentiM-veloeity components
usually meets many difficulties. The 5-holed spherical
pitot tube of Van der Hegge-Zijnen still gives no consistent
results for velocities below 1 m/see. These tangentialvelocity components are much smaller than the axial
components but for an accurate theoretical analysis of
the unsteady phenomena at the screw propeller the circumferential inequality caused by the tangential velocities may not be neglected in the future.
A simple quasi-steady analysis with the aid of a K ~ - K o 165

)(

10

I
[

">~

tl\

[NNRR

0 t')

60\
SECT'ON.~
40 B0
/~ ~/~n

80~/V-SHAPEO\

/30~0

~ o,~,,y I

8~ I U-SI'4APEDSECTION
Fig. 18

FIGURES INDICATE TFtE

BLADE ANGLES IN DEGREES

]Fig. 17

Thrust eccentricity calculated by Stuntz, Pien, Hinterthan, and Ficken

J diagram as proposed by Schuster [11 ] may give a qualitative picture of the forces generated by a screw in a circumferentially nonuniform flow field.
For every 5 or 10 deg of the circumference an instantaneous examination of the blade is made. The axial
wake velocities are regarded as constant at each blade
position. With the aid of the open-water characteristics
(K~-KQ-J diagram) of the particular screw, the time
history of the thrust and torque can be found. The path
of the center of the thrust will be symmetrical with regard to the longitudinal centerplane of the ship when the
tangential wake velocities are neglected. This path will
be swept z-times every revolution for a z-bladed propeller.
Usually the region of maximum wake velocity above the
propeller axis will be broader (thicker) than that below
the propeller shaft. The closed path on which the center
of thrust is moving will lie mostly above the propeller
axis.
When the tangential wake velocities are included, the
rotative speed of the screw blades will be smaller when
entering the peak of the wake and larger when leaving
this peak. This will cause a shift of the path of the center
of thrust to starboard for a screw that rotates clockwise
and to port for a screw that rotates counterclockwise.
The shape of the sections in the ships afterbody has a
pronounced influence on the position and form of this
path, Fig. 17.
Because of this eccentric position of the thrust, horizontal and vertical bending moments are created in the
propeller shaft.
166

Coordinates, elastic deformations and forces

Looking more closely at the variation of torque, it

is obvious t h a t a dynamic-force pattern is created in the
propeller shaft in way of the propeller because of the
circumfereni;ial inequality of the wake and, hence, the
torque-generating Force. These horizontal and vertical
transverse forces and bending moments have to be
absorbed mainly by the sterntube and the sternpost.
The forces and moments acting on the propeller working in the flow field behind the ship can be divided into
six components:
Axial. Thrust and torque.
Transverse. A transverse force, because of the
circumferential inequality of the torque force (unbalance
of torque) ; a vertical bending moment due to the thrust
eccentricity.
Vertical. A vertical force and a horizontal bending
moment for the same reasons as stated in the foregoing,
see Fig. 18.
The experimentM determin'~tion of the thrust and
torque fluctuations of a screw model behind a model of a
single-screw ship was carried out successfully for the first
time by Krohn and Wereldsma [12]. T h e y carried out
their measurements of the hydrodynamic forces created
by the propeller using a measuring shaft of very great
stiffness.
M a n y systematic and individual experiments have
been done using the apparatus of Krohn and Wereldsma
[13,14]. The systematic experiments give information
about the influence of number of blades, the shape of the
afterbody and the position of the propeller shaft.
The frequencies of the periodic force fluctuations due
to the propeller running in the flow field behind the ship,
will be equal to the number of revolutions of the propeller times the number of blades (the blade frequency)
or a multiple thereof.
Regarding the influence of the nmnber of blades, the
characteristic difference between propellers with even
and odd numbers of' blades must be mentioned.
MARINE TECHNOLOGY

TORQUE VARIATIONS

F\
,,,/~,,,

i4A

VERTICAL BENDING MOMENT(propetter

,i\"

!'

::/0 v
"...i

~.../

~/

.*20

180
Propeller

270

position

HORIZONTAL

LO ,

P~

BENDING

ii

oo

oo0

18o.

,70.

,,'o.

z.,
Z-5

"7~ t /~f,~\.(-'~,'/ I / f ' t f ! ~'

.-5\1
.10

MOMENT

o.=t25

Ii//~ !

,o

v--v-

360

THRUST VARIATIONS

,..15~

",._.1

90

wlwgh~ududed)

!,,/!,

',!/
\ /

Z=6

1 / I', /i ~,~
'.,_,"
.J

-~.,.
9I

I//'

180
270
PropetEer position 8

360

Fig. 19 Effect of number of blades on dynamic propeller forces, excited in "behind"

condition
Table 4

Formulas A p p r o x i m a t i n g

Transverse Forces a n d

Bending Moments,

Excited by Propeller ( 1 5 - K n o t Tanker)

FORMULAE APPROXIMATING THE TRANSVERSE FORCES AND BENDING MOMENTS EXCITED BY
APROPELLER (15KNOTS TANKER)
1) HORIZONTAL TRANSVERSE

:/FyO'7D/ = 0 . 1 2 , 0 . 0 7 3

FORCE

SIN(4~3+80 )

Z,4)

LTz g e m j

2) VERTICAL TRANSVERSE

FORCE

=0.12+ 0.150 SIN (513 101 )

ZI 5)

:r]/Fx'O7D/=006.0.076 SJN (413.126 )

Z= 4)

[Tz gem j
=006,0130

3) HORIZONTAL BENDING MOMENT

4) VERTICAL BENDING

MOMENT

:[
Tx
]=0.007, 0008
LFz gem 0.7DJ

:[

-Ty

APRIL 1966

( Z = '5)

SIN (413+147 )

(Z=4)

=0.007-0019

SIN (513+159 )

(Z= 5)

]oo.o32-o.o12

S~N(4~,.130~

CZ- 4>

=0032+0053

SIN (513 +155 )

( Z = "5)

IFz gem 0.7D]

For a screw propeller with an even number of blades,

the fluctuating forces of two opposite blades will give rise
to a larger total thrust and torque amplitude because two
blades pass simultaneously the stern and its associated
peaks in wake velocities. The transverse force and
bending moment of one blade will be compensated more
or less by those of the opposite one.
For propellers with an odd number of blades, the blades
will pass, alternatingly, the upper and lower wake peak.
The total thrust and torque fluctuations will thus be
smaller than for an even-bladed propeller. For an odd
number of blades the transverse forces and bending
moments, the favorable mutual compensation experienced by the even-bladed propeller will not occur. Fig. 19

SIN ( 5 p , 2 0 1 o)

gives an illustration of results of measurements on 47 5and 6-bladed screw models in the wake of a ship model.
A statistical investigation of the experimental data
on thrust and torque fluctuations of some 40 different
ship models, tested at the N S M B , leads to the following
conclusions:
1 No systematic relation can be found to exist between the amplitudes of the force fluctuations and the
principal ship-shape parameters such as block coefficient,
prismatic coefficient and screw diameter-ship length ratio.
2 For prismatic coefficients of the afterbody between 0.73 and 0.79 it was ascertained for 4-bladed propellers t h a t with a probability of about 80 percent the
following results will be obtained: The amplitude of the
167

Table 5

~z lz*~z

td:z

*~z

*~z

symbol

Coefficient

Hydrodynamic
moment of inertia

+ ; z C z =rz

1i

measured

17 6 10_~ kgmsec21
i
i

Hydrodynamic
torsional damping

0.86 10 .2 kgm
I

Hydrodynamic

mass

32"10-2

Hydrodynamic
axial damping

2,7

kg sec
m

F
Acceleration

Velocity

coupling

coupling

~]

14 10 -3 kg sec 2

=~

023

kg sec

L~J LL~J
Fig. 20

Scheme of coupled differential equations of screw shaftthrustblock system

first harmonic 4 of the torque fluctuation will be 6 ~ percent of the mean total torque, the amplitude of the first
harmonic of the thrust fluctuations will be 10 percent
of the mean total thrust, and the amplitudes of the higher
harmonics will be substantiMly lower.
And likewise for a 5-bladed propeller: The amplitudes
of the first and second harmonics of the torque fluctuations will be, respectively, 1 ~ and 1 percent of the total
torque, and the amplitudes of the first and second harmonics of the thrust fluctuations will be, respectively, 2
and lJ/~ percent of the total thrust.
Deviations from these indications larger than 2 percent absolute do not occur.
3 Fine-ended vessels, which includes most fast ships,
can have substantially greater force fluctuations.
In Table 4 a review of the formulas which approximate the transverse forces and moments generated by a
4-bladed and a 5-bladed propeller behind a 15-knot
tanker is given. For the loading of the shaft in the vertical direction besides the hydrodynamic forces the weight
of the propeller hag to be taken into account.
The mean value of the propeller-generated transverse
forces can be neglected compared to the propeller weight
from a viewpoint of static shaft loading. The static
bending moment, lifts the propeller up and reduces the
deflection of the sterntube. Reckoning has to be held,
however, with a large bending moment in the shaft in
way of the screw plane.
Comparing the dynamic behavior of a 4-bladed and
5-bladed propeller, it is noted that:
(a) The fluctuations in the transverse force of the 54 The first harmonic has the blade frequency.
168

Initial speed

kn

10,1

11.7

13.3

14.6

Speed at which tugs assist

effectively in braking and keep
the ship on course

kn

10

Reach at 20 R,RM. ahead

before tug assis:ance becomes
km
effective

3.1

2.9

2.6

2,3

Reach at 50 R.R~". astern

and 40 tons extra braking force
km
exerted by tugs

0.9

1.1

1.4

1.7

km

4.0

4.0

4.0

40

Head

Stopping Maneuvers for a lO0,O00-dwt Tanker for

Headreach of 4 km (2.5 miles)

reach

bladed propeller (although unimportant) are twice as

large as those for the 4-bladed propeller.
(b) The fluctuations in bending moment are much
higher for a 5-bladed than for a 4-blMed propeller.
Realizing that the ship designer generally has at hand
effective and relatively cheap means of avoiding axial
shaft vibrations (torque and thrust) and thai; he has to
reduce the excitation of horizontal hull vibrations to a
minimum, the 5-bladed propeller is to be regarded as an
unfavorable propeller compared to a 4-bladed one.
From results of recent systematic tests with, among
others, the Wageningen B-Series it could be deduced that
a 6-bladed propeller has about 3 percent less efficiency
than the comparable 4-bladed propeller [4]. The smaller
screw diameter, the larger screw clearances and the very
favorable pattern of the fluctuating forces (see Fig. 19)
are, however, distin,% advantages, justifying tlhe application of 6-bladed propellers for single-screw ships.
The propeller shift and the ship's stern are not infinitely stiff. Thu~, because of the described force
patterns, elastic deformations will occur.
The torsion and the axial displacement of the screw
owing to the elastic shaft give rise to hydrodynamic
coupling between tl'~e axial dynamic screw forces (thrust)
and the dynamic torsion forces (torque).
The deflections ,of the propeller shaft due to the
bending moments create gyroscopic phenomena at the
propeller.
A certain volume of water follows the unsteady movements of the screw blades, manifesting itself as an added
lnaSS.

The unsteady character of the screw loading will'induce in the screw race helicoidal trailing vortex patterns,
varying periodically [n strength, see Fig. 10. The energy,
carried away by this vortex system causes hydrodynamic
damping. All the hydrodynamic quantities of the screw
as a source of' vibration are summed up now. If it were
possible to calculate or determine experimentMly these
hydrodynamic quantities, then it would be possible to
predict the expected stresses in the stern construction and
in the propeller shaft resulting from the unsteady forces
of the ship propeller.
Fig. 20 gives the :~cheme of coupled differential equations of the screw-shaft-thrustblock system.
MARINE TECHNOLOGY

FI,~ * ~PIEEE

~!~i~i~i~i~i~i~i~i~i~i~i~i~i~i~i~i~i~i~i~i!i~i~i~i~i~i~i~!~i~i~i~iII~!i~i~i~i !~i~i~i~i~!~ii~!i!i~iii,'i!~ii;!i~!~i i~!i i i i i i li~i i i !ili!i !ii ~i i i~ii!i!~i iii!i~i~i ~i ~!ii i i ~!i~i !~!~!i!~ii !i !!~ii !i ~i i i!i ~i~i!~ii~i~i i!!~ii!!ii i i !i!i!~ii!i~i~i~i i~i i !!i!~ii i ~i!i!!!!~i~!ii ~i i!~i~i~i~i!~ii!i ~!!~ii i i i i ~i!i~i~i i~i i~i i i i~i i ~i i~i~!ii ~
Fig. 21

P r o p e l l e r exciter for d e t e r m i n a t i o n o f coefficients for h y d r o d y n a m i c mass a n d d a m p i n g a n d h y d r o d y n a m i c

coupling between thrust and torque vibrations

,5
i

TORQUE VARIATION

THRUST VARIATION

Fig. 20 the values for a model of a specific single-screw

cargo ship, as measured by Wereldsma, are given [15].
The results of Wereldsma's prediction of the torque
and thrust fluctuations based on model-test data and the
results of measurements on the full-size ship are compared
in Fig. 22. The good correlation between prediction and
measurement indicates that a new area of ship-model
testing has been opened as a service for the shipbuilding
and ship operating industry.

'1 . 5

g
o

I
0

"Jl
45

MEASURED - FULL SIZE

....

Stopping of Ships

90
135
PropH[er position e

180

SHIP

PREDICTION DERIVED FROM MOOELTEST-RESULTS

Fig. 2 2 C o r r e l a t i o n of m e a s u r e m e n t s o n full size a n d p r e d i c t i o n

of t o r q u e a n d t h r u s t variations, b a s e d o n model-test results

The hydrodynamic mass and damping and the hydrodynamic coupling between thrust and torque vibrations
have to be determined for the prediction of the stresses
in a given shaft configuration. Wereldsma has developed
a propeller exciter, Fig. 21, to evaluate the coefficients
appearing in the left-hand side of the equations. With
this exciter a given axial or torsional vibration can be imposed on a model screw at a certain load KT or K~ and at
a certain advance ratio J. In the column at the right in
APRIL 1966

With the aid of the quasi-steady velocities and forces

diagram acting on a screw-blade section an insight can be
gained into the force pattern around the screw during
stopping, Fig. 23.
When the rotative speed of the screw is reduced the
angle of attack, and, hence, the thrust, will decrease from
that at full-ahead power (phase 1). At about 70 percent of the normal ahead number of revolutions the thrust
will become zero and the screw turns freely (phase 2).
As the RPM is further reduced, a negative angle of
attack will result in negative lift and thrust. A further
reduction in RPM leads to such large negative angle of
attack that flow separation on the screw blades will
occur with an accompanied loss in lift. This separation
starts at about 30 percent of the RPM ahead (phase 3).
The decreasing lift causes a decrease in braking force
169

(T)

fuU power

. ~lOO

tugs

w turning stack

v.

I
$Lo~vty~|stern

make flmgt

'i
--*50

--m0

50

RPk~ OF THE PROPELLER

: !.50 1

zOlO
+100

50

.S,O

R "P 141 IN */=

-tO0
I

e:

12

L
i
SPEED REOUCTION

14- i9

.9:

knots I

i
3

----HEAO

Fig.

SPEED OF SHIP

2 5

REACH

~ 6-4 knols

__

i
5

~'% j i~
8

IN k m

Headreach :for a 1 0 0 , 0 0 0 - d w t tanker; initial speed

14 knots

v,~0
Fig. 23 Relation between thrust and R P M at a constant ship
speed. Force and velocity diagrams for blade element of screw

dv
dv d s
dv
K : m . a = m ~ - = m ~ - ~ - = m ~d~ V

d s = rn KV--dv
vi

in w h i c h

S : Head

reach

A= D i s p l a c e m e n t

g = Gravitational
v i = Initial
vt= Terminal

K = Braking

Fig. 24

ton
acceleration

speed
speed

force

msec-

m.sec- 1
m see -1
ton

Integral for calculation of headreach of ships

until such time as the separation or profile drag become

large enough to predominate and the braking force again
increases (phase 4).
The continued increase of profile drag with the astern
operation of the propeller will further increase the braking fores (phase 5). At a high number of astern revolutions the probability of cavitation and of drawing air into
the propeller increases. The occurrence of one of these
m a y cause a decrease in braking force.
The typical S-characteristic, describing the thrust
between 100 percent R P M ahead and 100 percent R P M
astern operation, was described for the first time by
T h a u [16]. For the propeller alone, this curve can be
determined for quasi-steady operation, when the openwater screw characteristics are available for the ahead and
astern running condition [17]. These quasi-steady considerations are at the same time the basis for a method
for calculating the headreaeh. I n Fig. 24 an integral
is derived calculating the headreach from the basic law of
dynamics, t h a t
Force = mass X acceleration
170

The hydrodynamic added mass has been taken into account by tile use of a faetor 1.05.
The values of tee ratio V/K can be determined from a
model test for each speed V at different rotative speeds.
For a given ship displacement the braking force K call be
calculated for any combination of speed V and the rotalive speed, and the integral :for the headreach can be
determined.
The ratio between displacement and power A/SHP, or
as written in the integral of Fig. 24, the r,~tio A/K, is
very important for the length of the headreaeh. A large
displacement propelled b y a relatively small power will
give a long headreach (large value of the ratio A/SHP,
tankers). A low value for the ratio A/SHP, as for in-stance for destroyers and tugboats, will give a very short
headreach.
Analysis for a m a x i m u m allowable headreaeh of 4 k m
(2.5 miles) were made for a 100,000-dwt tanker at different initial speeds, with the assumption thai, tugs would
assist in the stopping maneuver. Table 5 is a review
of this analysis. The only possible maneuw~r is the one
starting from an i>itial speed of 10.1 knots. When braking from 10.1 to 7 knots with a rotative speed of 20 rpm,
the distance covered is 3.1 k m (2 miles). At. 7 knots the
tugs take hold and exert an extra braking force of 40 tons.
The rotative speed becomes 50 r p m astern and the tanker
stops after another 0.9 k m (0.6 mile). Operations requiring tugs to ntske fast at speeds greater than 7 knots
nmst be considered a very risky undertaking. For a
m a x i m u m allowable headreach of zi k m the initi~d speed
of the 100,000-to> tanker may, hence, not exceed 10.1
knots.
These conclusions are based on model tests. I t m a y
be possible, due to a conservative interpretation of the
scale effect, t h a t these results are somewhat pessimistic.
Owing to the lack of sufi%ient data from full-scale tests,
a correction of the data in Table 5 for scale effects is not
possible.
I n Fig. 25 the c~dculation of stopping of a 100,000-ton
tanker is given for an initial speed of 14 knots. The
different phases into which the whole maneuver can be
divided are indicated. An essential part of th.e maneuver
MARINE TI'CHNOLOGY

is the fact that the screw is to be stopped when the speed

reaches 6 knots and the tugs make fast. Otherwise the
vessel will loose steerageway.
Finally, it may be noted that data and testing methods
as treated in this review are important resources needed
in the choice of the type of propeller and the determination of its dimensions for a given application.
References

1 L. Troost, "Open Water Test Series with Modern

Propeller Forms," Trans. NECI, 1950-51.
2 W. P. A. van Lammeren, L. Troost, and J. G.
Koning, Resistance, Propulsion and Steering of Ships, The
Technical Publishing Company, H. Stare, Haarlem, 1948.
3 J . D . van Manen, "Fundamentals of Ship Resistance and Propulsion, Part B, Propulsion," International
Shipbuilding Progress, 1957.
4 J . D . van Manen, "A Review of Research Activities at the Netherlands Ship Model Basin," International
Shipbuilding Progress, 1963.
5 A.J. Taehmindji and W. B. Morgan, "The Design
and Estimated Performance of a Series of SupercavRating
Propellers," Second Symposium on Naval Hydrodynamics, Washington, 1958 (1960).
6 J. D. van Manen, "Ergebnisse systematischer
Versuche mit Propellern mit armb~hernd senkrecht
stehender Achse," Jahrbuch STG, 1963; Schip en We'll,
1964.
7 J . B . Hadler, W. B. Morgan, and K. A. Meyers,
"Advanced Propeller Propulsion for High-Powered
Single-Screw Ships," Trans. SNAME, vol. 72, 1964, pp.
231-293.
8 I. It. Abbott, A. E. yon Doenhoff, and L. S.
Stivers, "Summary of Airfoil Data," NACA Report 824,
1945.
9 L.C. Burrill and A. Emerson, ':Propeller Cavita-

APRIL 1966

tion: Further Tests on 16 in. Propeller Models in the

King's College Cavitation Tunnel," Trans. NECI, 196263 ; International Shipbuilding Progress, 1963.
10 J. D. van Manen, "Dutch die Schraube erregte
Sehiffssehwingungen," Schiffstechnilc, 1965; Schip en
We~f, 1965.
11 S. Schuster, "Propeller in Non-Uniform WakeCollection of Existing Work," Tenth ITTC, London,
England, 1963, Report of Propulsion Committee, Appendix 7.
12 J. Krohn and P~. Wereldsma, "Comparative Model
Tests on Dynamic Propeller Forces," International Shipbuilding Progress, 1960.
13 (a) J. Krohn, "Ueber den Einflusz der Propellerbelastung bei verschiedener Hintersehiffsform auf die
Sehub- und Drehmomentsehwankungen am Modell,"
Sch~i~ und Hafen, 1958.
(b) J. Krohn, "Ueber den Einflusz des Propellerdurchmessers auf die Schub- und Drehmomentschwankungen am Modell," Sch~stechnih, 1959.
14 J. D. van Manen and R. Wereldsma, "Propeller
Excited Vibratory Forces in the Shaft of a Single Screw
Tanker," International Shipbuilding Progress, 1960.
15 (a) R. Wereldsma, "Dynamic Behaviour of Ship
Propellers," Doctor's Thesis, Technological University,
Delft, 1965; Publication No. 255 of the NSMB.
(b) R. Wereldsma, "Experiments on Vibrating
Propeller Models," International Shipbuilding Progress,
1965.
16 W. E. Thau, "Propellers and Propelling Machinery. Maneuvering Characteristics During Stopping
and Reversing," Trans. SNAME, 1937.
17 It. F. Nordstr6m, "Screw Propeller Characteristics," Meddelanden Statens Skeppsprovningsanstalt
No. 9, 1948.

171