ADAI

*.
V - - -

PROPELLERS

)R. ER. M. V@. C. OOSTERVELl

WAKE ADAPTED DUCTED
PROPELLERS

DR. IR. M. W. C. O O S T E R V E L D
(ASSISTANT MANAGING DIRECTOR, NETHERLANDS SHIP MODEL BASIN).

P U B L I C A T l O N No. 345
N E T H E R L A N D S SHIP M O D E L B A S I N
WAGENINGEN -NETHERLANDS
WAKE ADAPTED D U C T E D PROPELLERS
 CONTENTS

1. introduction
2. General relations for flow and force action
3. Calculations on ducted propellers with vortex theory
4. Potentialities of nozzles
5. The flow accelerating nozzle
6. The flow decelerating nozzle
7. General discussion of test lesults
8. Application of wake adapted nozzles
8.1. Introduction
8.2. Application of wake adapted nozzles behind tankers
8.3. Application of wake adapted nozzles behind twin-screw ships
8.4. Ducted propeller with circulation control
9. Ringpropellers
10. Conclusions
Appendix
Summary
Samenvatting
References
Nomenclature
Acknowledgement
 1. INTRODUCTION

A ducted propeller consists of a combination of an annular airfoil and an impeller,

acting as a propulsion unit. A schematic view of a ducted propeller is given in Fig. 1.
The axial force acting on the impeller of a ducted propeller usually differs from the
net thrust of the system. A positive or negative force may act on the nozzle depending
on the nozzle shape and the operating condition. Due to the nozzle action the velocity
at the impeller plane can be either less than or greater than the velocity at the pro-
peller plane of a conventional screw with the same diameter and speed of advance.
Insight into the shape of the nozzle profile of a ducted propeller can be gained by
Fig. 2. Here the flow through different types of ducted propellers is superimposed on
the flow through an open propeller. Both the open propeller and the ducted propellers
are designed for the same mass flow rate and velocity in the ultimate wake. Conse-
quently the thrust and ideal efficiency of these systems are equal.
The ducted propeller with the accelerating flow type of nozzle is now used extensive-
ly in cases where the ship screw is heavily loaded or where the screw is limited in
diameter. The accelerating nozzle offers a means of increasing the efficiency of heavily
loaded propellers. The nozzle itself produces a positive thrust.
In the case of the decelerating flow type of nozzle, the nozzle is used to increase the
static pressure at the impeller. This ducted propeller system is the so called pumpjet.
The duct will produce a negative thrust. This nozzle may be used if retardation of
propeller cavitation is desired. For naval ships a reduction in noise level can be ob-
tained which may be of importance for tactical reasons.

FIG.1. Ducted propeller.

FIG.2. General form of stream-
lines enforced by different nozzle
types.
-ating

Although the idea of surrounding a propeller by a nozzle is already very old, it has
lasted until the early 1930 before the ducted propeller came into practical use.
LUISASTIPA(1) and later on KORT(2) experimentally proved the advantages which
can be obtained by application of the accelerating nozzle. These investigations clearly
show that an increase in efficiency can be obtained by this nozzle at heavy screw loads.
Primarily due to the work done by KORT,the application of ducted propellers behind
certain ship types (tugs, pushboats, trawlers) has become common practice. This may
be the reason that the accelerating ducted propeller is frequently referred to as the
'KORT'nozzle.

Many studies on ducted propellers have been made during the last 40 years. An
extensive summary of this work was made by SACKSand BURNELL (3) in 1960, A
general review of the more recent theoretical studies on ducted propellers has been
given by WEISSINGER and MAASS(4).
Among the theoretical studies on ducted propellers the investigations of HORNand
AMTSBERG (5), K~~CHEMAN and WEBER(6) and DICKMANN and WEISSINGER (7) may
be mentioned in particular. Especially, the work of DICKMANN and WEISSINGER (7)
was the first step to the development of more refined theories on ducted propellers.
This paper was the base for the work which has been performed at Karlsruhe by
DICKMANN, WEISSINGER, WIEDEMER, BOLLHEIMER, BRAKHAGE, MAASS and RAUTMANN.
Some of the basic ideas used at Karlsruhe were also used by other investigators
working at Therm (ORDWAY, RITTER,GREENBERG, HOUGH,KASKEL, Lo, SLUYTER,
SONNERUP), at NSRDC (MORGAN, CASTER, CHAPLIN,VOLGT),at Viday (NIELSEN,
KRIEBEL, MENDENHALL, SACKS,SPANGLER) and by many others.
Most of the theoretical investigations on ducted propellers were concentrated to a
large extent on the linearized theory and on axisymmetrical nozzles in a uniform
flow.
A discussion of the adequacy of the different studies is beyond the scope of the
work described in the following. The design of a ducted propeller depends on a large
number of variables. For the design of an optimum ducted propeller system it is at-
tractive therefore to have a sound theoretical calculation method available supported
by carefully selected systematic experiments.

Recently, a comparison of theory and experiment on ducted propellers was given

by MORGAN and CASTER (8). Tests on ducted propellers are scarce, however, and most
of these tests are restricted to isolated applications.
By far the most extensive systematic experiments on ducted propellers for appli-
cation on ships have been performed at the NSMB by VANMANEN(9, 10, 12) and
VANMANEN and SUPERINA (11). These investigations concerned nozzles of the accele-
rating flow type. This work gives valuable advices for the design of optimum ducted
propeller systems from the viewpoint of efficiency for ships. A review of this work was
given by VANMANEN and O~~TERVELD (13) in 1966.
Recently, the results of systematic experiments on ducted propellers of the decele-
rating flow type were given by OOSTERVELD (14).

In the following a general discussion of the design of ducted propellers is given.

Besides, some new developments in the field of ship propulsion by special types of
ducted propellers will be discussed.
 2. G E N E R A L R E L A T I O N S FOR FLOW A N D FORCE ACTION

In the combined arrangement of duct and propeller, the axial force on the impeller
differs usually from the net thrust of the system. The force set up on the nozzle will be
positive or negative depending on the nozzle shape and the operating condition.
Insight into the working principle of a ducted propeller can be gained by the appli-
cation of the fundamental momentum relationship. Fig. 3 shows the simplified
system, by which the ducted propeller can be replaced.
The total thrust T acting on the fluid due to impeller and nozzle is:

applying the momentum- theorem over the control volume given in Fig. 3. The thrust
Tp developed by the impeller can be obtained by calculating the pressure difference
across the impeller disk using BERNOULLI'Sequation:

The kinetic energy E lost in the impeller slipstream is given by:

Hence the ideal efficiency q, of the propulsion device is defined by:

FIG.3. Control volume used for mo-

mentum considerations.
where,

The mean axial velocity at the impeller plane can be written as:

The mean axial velocity due to the impeller is:

Hence, the mean axial velocity U, induced by the nozzle at the impeller becomes:

The mean value of the static pressure at the impeller, P, can be calculated from

Hence the mean static pressure coefficient CPmancan be written as:

The ideal efficiency qr, the velocity ratio Vp/VAand the pressure coefficient CPmean
are
given as functions of the thrust coefficient CT and for different values of the thrust

FIG.4. Ideal efficiency of ducted propeller

system.
FIG.5. Mean axial velocity at impeller
plane of a ducted propeller.

ratio r in Figs. 4 through 6. The inflow velocity of the impeller of a ducted propeller
can be either less or greater than the inflow velocity of an open propeller under equal
conditions. The thrust ratio r of an open propeller is equal to 1. It can be seen from
Fig. 5 that when due to the nozzle action the flow rate through the impeller is in-
creased, a positive axial force or thrust is developed on the nozzle (thus r < 1). In

FIG.6. Mean static pres-

sure at impeller plane of
a ducted propeller.
the case of a nozzle which reduces the flow rate through the impeller, a negative
axial force is acting on the nozzle (thus o > 1).
It follows from Fig. 4 that, if the diameter of a ducted propeller system is kept
constant, the ideal eflciency rl, of the system increases with decreasing values of the
impeller thrust-total thrust ratio T. Thus the ducted propeller with the flow accele-
rating type of nozzle offers a means of improving the efficiency if compared with an
open propeller.
Fig. 6, shows that the static pressure at the impeller increases with increasing values
of the thrust ratio o. 'Ram pressures' at the impeller plane are built up when o exceeds
a value of l. A reduction of the flow rate by the nozzle results in an increment of the
static pressure at the impeller. The increment is attractive from a point of view of
retardation of screw cavitation. The duct itself, however, will produce a negative
thrust. In order to compensate for the increased nozzle drag, the impeller loading
must be enlarged. An improvement of the cavitation characteristics of the screw will,
therefore, only be obtained if the gain in static pressure exceeds the unfavourable
effect of the increased screw loading.
It may be concluded from these considerations that the application of a shroud
around a propeller can attribute to a higher efficiency or better cavitation charac-
teristics. The elementary theory used so far takes into account only the state of the
flow at the impeller and infinitely far upstream and downstream. It does not consider
what happens to the flow outside these regions.
To derive a relation between the shape of the nozzle profile and the characteristics
of the nozzle-propeller system, the assumption can be made that the wake area A,
and the exit area of the nozzle AEx (see Fig. 3) are equal. If the disc area of the im-
peller is denoted by A,, the following relation between the impeller disk area-nozzle
exit area ratio, the thrust coefficient CT and the thrust ratio z holds:

Calculations of the area ratio Ao/AExhave been made and the results are presented
in Fig. 7.

For a more refined evaluation of the characteristics of ducted propellers, use can
be made of theoretical and experimental methods. In the following, a simplified
method has been described based on a vortex theory, which enables the calculation of
the characteristics of systematic series of nozzle shapes as well as the more general
characteristics of ducted propellers. In addition, the results of open-water tests with
systematic series of ducted propellers both with the flow accelerating and with the
flow decelerating nozzles are given.
 3. CALCULATIONS O N D U C T E D PROPELLERS
WITH VORTEX T H E O R Y

A large number of theoretical investigations have been conducted on ducted

propellers during the past thirty years. An extensive summary of this work was made
by SACKS and BURNELL (3) in 1960, Recently a general review on theoretical studies of
ducted propellers was given by WEISSINGER and MAASS(4).
The calculations on ducted propellers may be classified into two types:
- the direct problem of ducted propellers in which the geometry of the ducted pro-
peller system is given and the pressure distribution and the force on the duct must
be determined.
- the inverse problem of ducted propellers in which a certain combination of impeller
forces and duct forces are given and the shape of the nozzle must be determined.
Both the direct and the inverse problem require the solution of a singular integral
equation. The present investigation is concerned with the inverse problem although a
slightly different approach has been chosen. In this approach the nozzle shape was
determined for a given uniform loading of the actuator disk representing the impeller
and for a given circulation of the nozzle. This method is very attractive, because the
computation is simple. However, it gives only the possibility of determining nozzle
shapes which will operate satisfactorily for a given flow condition. This method gives
also the possibility of determining the more general characteristics of ducted propellers
in an easy way.
The calculations on ducted propellers were based on the following assumptions.
The forward velocity of the ducted propeller system was assumed to be sufficiently
large, the nozzle loading and the blade loading of the impeller sufficiently low to
permit the application of the linearized theory. The effect of the hub shape on the
flow field was neglected. Further, the fluid was assumed to be inviscid and incompres-
sible. The ducted propeller system considered consists of an annular airfoil of finite
length and of an impeller having an infinite number of blades.
The mathematical model of the ducted propeller configuration can be composed by
means of vortex distributions. The impeller was regarded as a uniformly loaded
actuator disk set normal to the free stream. It was driven to' rotate with an angular
velocity m. Free trailing vortices started from the disk at hub and tip radius. The
flow around the nozzle was represented by a sinusoidal bound ring vortex distribution
with zero strength at the leading and trailing edges and by a ring vortex distribution
with zero strength at the leading edge and equal to the strength of the circumferential
component of the helical trailing vortices at the impeller disk. The resulting mathe-
matical model is summarized in Fig. 8.

Assume the vortex strength per unit impeller disk area at the radius of the impeller
to be equal to y (R,), the strength of the different vortex distributions then becomes:
- vortex strength per unit impeller disk area at radius r
- strength per unit area of the trailing vortex sheet starting from the impeller disk at
the tip radius

- strength per unit area of the trailing vortex sheet starting.from the impeller disk at
the hub radius

\/l + p212 .'l 1 y (R,)

- strength per unit area of the bound ring vortex distribution along the nozzle

sin 8
P7 (RI) (for - a L < x < 0 )
2 dz=T)
y (R,)sin 8 (for - aL < x < (1 - a) L)

FIG.8. Representation of ducted propeller by

vortex distributions.
where,

while the definitions of R,, R,, r, X, L, a and so on are given in Fig. 8.

The thickness effect of the nozzle profile on the flow field can be taken into account
in the linearized theory by representing the nozzle by a distribution of ring sources
and sinks along a cylinder with radius R, (see Fig. 8). In the case of very thin nozzle
profiles, the strength of the sink and source distributions representing the nozzle pro-
file can be calculated by considering the flow around the profile as two-dimensional.
Then the local source strength is equal to the derivative of the profile thickness:

where h(x) denotes the local thickness of the nozzle profile.

The total induced velocities due to the different vortex and source distributions
representing the impeller and the nozzle can be calculated according to the law of
Biot-Savart if the main dimensions of the system (LID; ulL; dlD; S/L and the thick-
ness distribution of the nozzle), the advance coefficient of the screw (J = V,/nD),
and the loading of the impeller and nozzle are given. The impeller thrust and torque,
the thrust on the nozzle and the ideal efficiency of the system can be calculated now
for the chosen values of the design parameters. In addition, the shape of the nozzle
and the pressure distribution along the nozzle can be computed. The method for the
calculation of the ducted propeller characteristics is, in a more detailed way, given in
the Appendix.
It is shown in the Appendix, that for the chosen load distributions of impeller and
nozzle, the shape of the nozzle is completely determined by the values of the thrust
coefficient CTo,the thrust ratio zo and the main dimensions of the ducted propeller
system. Here, the subscript '0' denotes that a ducted propeller system is considered
with an impeller rotating at infinite angular velocity or with an advance coefficient J
equal to zero. Then, the tangentially induced velocities and consequently the losses
due to the rotation of the fluid are zero. The case that the impeller rotates with an
infinite angular velocity coincides with the case that the impeller rotates with finite
angular velocity whi1e.a stator is used to eliminate the rotational losses. Thus starting
from given thrust coefficients C,,, thrust ratios z, (thus at J = 0) and the main di-
mensions the characteristics of ducted propellers can be calculated. It must be noted
that for the thickness distribution of the nozzle profiles of all nozzles given in the
following, the NACA four-digit wing section basic thickness form was used. First, the
shape of the nozzle profile and the pressure distribution along the nozzle can be
derived. Secondly, the thrust coefficient CT, thrust ratio z and the ideal efficiency q ,
of such a ducted propeller can be calculated over a range of advance coefficients J.
This method of calculation was followed for the determination of the characteristics
of systematic series of flow decelerating type nozzles as will be given in section 6. In
the following section the theory will be used for the determination of the more general
characteristics of ducted propellers.
 4. POTENTIALITIES O F NOZZLES

The choice of the shape of the nozzle profile and the nozzle length depends mainly
on requirements with respect to efficiency, danger of flow separation on the nozzle,
and critical cavitation indexes on nozzle and propeller.

Flow separation occurs if the nozzle is too heavily loaded. If flow separation occurs
on the interior or exterior surface of the nozzle, the drag of the nozzle will increase
sharply. Moreover, the propeller will be operating in a highly irregular flow if separa-
tion occurs on the interior surface of the nozzle. In either case, the efficiency of the
system would begin to decrease. Therefore, flow separation on the nozzle surface
should be avoided. In order to evaluate, the maximum loading of the nozzle without
separation, the flow around the nozzle profile can be assumed to be two-dimensional.
By analogy to the two-dimensional airfoil flow, the approximation that flow separa-
tion occurs if the lift coefficient, CL,exceeds a value of 1.&l .5 can be used. It must be
investigated, however, whether this rough approximation gives a reliable criterion for
the nozzle loading.
The method described in section 3 for the calculations of nozzle propeller charac-
teristics enables the analysis of the sectional lift coefficient CLof nozzles under design
condition :

together with

Analyses of the sectional lift coefficient C, have been made for nozzles with different
thrust ratios z and different thickness ratios of the nozzle profile S/L. The result is
given in Fig. 9, which shows that the loading of the accelerating type of nozzle is
strongly restricted by the risk of flow separation on the interior surface of the nozzle.
The risk of flow separation on the exterior surface of the decelerating nozzle, how-
ever, is small even in the case of large thrust ratios z.

In order to investigate the effect of nozzle length on overall efficiency, the drag of
the nozzle must be calculated. The efficiency of the ducted propeller system may be
written as q = qrqfn,where the efficiency factor due to nozzle drag qfN is:
FIG. 9. Sectional lift coefficient
of nozzle profile.

where CDNdenotes the sectional drag coefficient of the nozzle.

The flow around the nozzle is fully turbulent because the value of the Reynolds
number based on the profile length is beyond 106.The drag in this region of Reynolds
numbers may be expressed by the following empiric relation (15):

where C, is the turbulent skin-friction drag coefficient. Calculations of the efficiency

factor qfn have been made. The results are given in Fig. 10 which shows that the
efficiency losses due to nozzle drag become substantial if the ducted propeller system
is lightly loaded.
 Cl
FIG.10. Efficiency factor due to frictional nozzle drag of ducted propeller.

Investigations of the optimum nozzle shape from the viewpoint of efficiency lead
to the conclusion that a maximum contraction of the nozzle should be aimed at.
Consequently, the impeller thrust-total thrust ratio t must be as small as possible.
The contraction is limited by the risk of flow separation on the interior surface of the
nozzle. If by analogy to the two-dimensional airfoil flow the approximation is used
that flow separation occurs if the lift coefficient CLexceeds a value of l, then the op-
timum value of the thrust ratio t for a given length-diameter ratio LID of the ducted
propeller system can be obtained from Fig. 9. Hence, the optimum ideal efficiency
of the ducted propeller can be obtained as a function of the nozzle length-diameter
ratio LID and the thrust coefficient C, using the results presented in Fig. 10. Analyses
of the optimum ideal efficiency and the efficiency losses due to nozzle drag have been
made for ducted propellers with different nozzle length-diameter ratios LID and for a
thickness ratio SIL = 0.15 of the nozzle profile. The result is given in Fig. 11. From
this diagram it can be seen that the use of a nozzle leads to an increase of the efficiency
at higher screw loads (C,> 1 - 2). At these screw loadings a nozzle with a length-
diameter ratio LID = 0.5-1.0 is advisable. For a shorter nozzle the gain in ideal
efficiency will be lower, whereas for a longer nozzle the gain in ideal efficiency will be
counter-balanced by the frictional nozzle drag.

The application of the decelerating type of nozzle results in an increment of the

pressure at the impeller and a reduction of the pressure at the exterior surface of the
nozzle. Consequently, the risk of cavitation on the exterior surface of the decelerating
nozzle becomes substantial. The earlier described method for the calculation of the
ducted propeller characteristics enables the calculation of the static pressure along the
nozzle and at the impeller plane under design condition. The minimum static pressure
at the exterior surface of the nozzle and the mean static pressure at the impeller plane
 L h - a z s C~91.o c - 0 8 4
0 5 0 1. a78
loo a73
200 *- a69

I deal efficiency
E tficiency taking ~ n t oaccount
the eftrct of trictional nozzle
drag.

FIG.I l . Optimum ideal efficiency and efficiency losses due to nozzle drag of a ducted propeller
system.

are denoted by P,,, and Pm,,. The following non-dimensional pressure coefficients
are introduced :

where P, and * p VA2are the static pressure and the dynamic pressure of the un-
disturbed stream respectively. Calculations of Cpm,,and Cm,,, have been made for
nozzles with different length-diameter ratios LID and different maximum thickness
ratios SIL. Minimum pressures which may occur at the exterior surface of the decele-
rating nozzle are given in Fig. 12.

FIG.12. Minimum static pressure at the exterior surface

of decelerating nozzles.
 The reduction of the flow rate inside the decelerating type of nozzle, results in an
increment of the static pressure at the impeller. This increment is attractive from a
point of view of retardation of screw cavitation. The duct itself, however, will produce
a negative thrust (z > 1). In order to compensate for this thrust loss (induced nozzle
drag), the screw loading must be increased. An improvement of cavitation properties
of the screw will therefore only be obtained if the gain in static pressure at least com-
pensates the unfavourable effect of the increased screw loading.
An approximate method of calculation has been followed here in order to deter-
mine whether an improvement of the screw cavitation characteristics can be obtained
by application of a nozzle. A comprehensive set of data on the cavitation properties
of propellers according to the vortex theory (see Ref. (16)) permits the introduction of
the following empirical relation for the minimum pressure on the screw blades:

Pmin -P m - n f2.4 + 0.6 Z] CT

[chin]
screw
=
+ P VA2
- -- 8 [AE/A,+ 0.21

where the blade area ratio and the number of blades of the screw are denoted by
AE/A, and Z respectively. The ducted propeller may now be regarded as operating in
open-water with an equivalent uniform stream velocity (VAf U,), and an equivalent
static pressure of the undisturbed stream (P, + +
4 p VA2 - Q (V, UN)2). UN
denotes the mean axial velocity at the impeller due to the nozzle action. The relation
between UN, CT and z has been calculated with the aid of the momentum theory
given before.

FIG.13. Minimum static pressure at impeller blades

of a ducted propeller.
 Hence, the minimum static pressure coefficient on the impeller blades of the ducted
propeller may be written as:

-P m - C2.4 + 0.6 Z] zCT 2

[ ~ h i n ]ducted propeller -
Prnin

+ P V'?
_ -
8 [AE/Ao+ 0.21
+I-(I+$)

Analyses of the pressure coefficient Cprn,,corresponding to ducted propellers with

different thrust coefficients C, and thrust ratios z and for AE/Ao= l and Z = 5 were
made. The result is presented in Fig. 13. This diagram shows that, for the particular
screw considered (A J A , = 1, Z = 5 ) only for low values of C,, the flow decele-
rating type of nozzle affects favourably the cavitation properties of the screw. If
ducted propellers are considered with larger blade area ratios of the impeller or with
more rotor rows or both, the flow decelerating type of nozzle may even for larger
values of the thrust coefficient C, favourably affect the cavitation properties of the
screw.
If we compare the results presented in Figs. 12 and 13 then it can be seen that for
the particular screw considered the minimum pressures occurring at the impeller
blades are much lower than the minimum pressures at the exterior surface of the nozzle
even at large thrust ratios z.
 5. T H E FLOW A C C E L E R A T I N G NOZZLE

From the theoretical calculations given in the preceding sections it was concluded
that the use of an accelerating nozzle leads to an increase of efficiency at higher screw
loads (CT> 1). In addition, at higher screw loadings a long nozzle is advisable, while
at light screw loadings a short nozzle should be used. At high screw loads, however,
the nozzle must not have a larger LID ratio than about 1.O. Above this value the loss
in efficiency due to the frictional nozzle drag will still be larger than the gain in ideal
efficiency which can be obtained by application of a nozzle.
Extensive experimental investigations with screw series in accelerating nozzles
have been performed by VANMANEN(9,10, 12) and VANMANENand SUPERINA (1 l),
in order to supply data for the design of ducted propellers for optimum efficiency.
These tests were carried out with the usual tank apparatus for open-water tests
with ducted propellers. Fig. 14 shows the test equipment which is fitted under the
carriage running over the basin during the tests. The usual routine of open-water tests
was followed; the rpm of the screw was kept constant and by varying the speed of
advance the desired value of the advance coefficient J was obtained. The rpm was
chosen as high as possible to obtain a high Reynolds number. Speeds above 3 m/sec.
could not be investigated on account of the towing carriage.

FIG.14. Test equipment for open-water tests with ducted propellers.

The tests with the different types of accelerating nozzles were first carried out with
B-series screws. Later on, however, special screw series for use in nozzles were de-
signed.
The investigations on accelerating nozzles were focussed on the determination of
the optimum nozzle shape from the viewpoint of efficiency. Therefore, the influence
of variations in camber ratio of the nozzle profilefll and in the angle between the
nose-tail line of the nozzle profile and the propeller shaft a, on the characteristics,
was systematically determined. A review of the tested nozzle shapes is given in Table
1. The dimensions of these nozzles are given in Fig. 15.
FIG. 15. Profiles of nozzles
through 8, and 10 and 11. NOZZLE NO8 (L/b=0.5) I /

-/
NOZZLE N0.7

4'
NOZZLE N0.6

0
NOZZLE N0.5

NOZZLE N0.3 (Lbs0.50)

I

TABLE
1. Particulars of nozzles nos. 2 through 8 and 10 and 11
Nozzle no. LID SlL AL a1 Profile
(degr.1

12.7 NACA 4415

12.7 NACA 4415
12.7 NACA 4415
15.2 NACA 4415
10.2 NACA 4415
12.7 NACA 54 15
12.7 NACA 3415
12.7 NACA 5415
12.7 NACA 5415
 The test results with nozzle no. 7 in combination with the B 4-55 screw series are
given in Fig. 16. For particulars of the B 4-55 screw series see (17), (18). In Fig. 16,
the results are plotted in the conventional way with the coefficients:

as functions of the advance coefficient J = V,/nD with the pitch ratio P/D of the
screw as parameter.
From the results of open-water tests with the nozzle shapes summarized in Table 1,
it appears that the smallest optimum diameters and the highest efficiencies are ob-
tained with the lowest impeller thrust-total thrust ratios z. This is in accordance with
the conclusion that a maximum contraction of the nozzle should be aimed at.
 2. Particulars of nozzles nos. 18, 19 and 20
TABLE

Nozzle no. LID s/L flr, a~ Profile

(degr.1
18 0.50 0.15 0.09 10.2 NACA 9415
19 0.50 0.15 0.07 10.2 NACA 7415
20 0.50 0.15 0.05 10.2 NACA 5415

NOZZLE No.20

I NOZZLE Na19

FIG.17. Profiles of nozzles nos. 18, 19

B,- - . .
NOZZLE No.18

.-.-.

and 20. I
Lately, three new nozzle types were tested which did have some advantages over
the types tested before. These nozzle profiles meet the following practical require-
ments :
I. An axial cylindrical part at the inner side of the nozzle at the location of the screw
2. A diverging part of the nozzle behind the screw
3. A maximum thickness ratio of the nozzle profile of 0.15.
The dimensions of these nozzles are summarized in Table 2. The nozzle profiles
are given in Fig. 17. An angle between the nose-tail line of the nozzle profile and the
shaft axis cc, of 10.2 degrees has been chosen for these nozzles. The camber ratios of
these new nozzle profiles exceed the value of 0.05 so far investigated. The camber
ratios for the nozzles no. 18, 19 and 20 were respectively 0.05;0.07 and 0.09.Only
very slight differences in the test results with the B 4-55 screw series in combination
with these nozzles were found. The open-water test results of nozzle no. 19 in combi-
nation with the B 4-55 screw series are given in Fig. 18.
On the basis of these test results a, from the structural point of view, simple nozzle
shape was chosen. This nozzle (nozzle no. 19A) did have, as already mentioned, an
axial cylindrical part at the inner side of the nozzle at the location of the screw, the
outside of the nozzle profile is made straight and the trailing edge of the nozzle is
thicker. Besides, it must be noted that a nozzle with a thick trailing edge will have
better qualities when stopping than a nozzle with a sharp trailing edge. The profile
of nozzle no. 19A is shown in Fig. 19.
 IN NOZZLE m 1 9

0 U 0.3 W OS U J Q 7 W 09 10 11 I2

FIG.18. Open-water test results of nozzle no. 19 with B4-55 screw series.

Most of the tests discussed so far were carried out with the Wageningen B-screw
series. However, frequently a screw type with wider blade tips (Kaplan type) than the
B-series screw type is preferred since the former is less susceptible to cavitation at the
blade tips. Extensive investigations performed at the NSMB have led to the design
of screw series, having uniform pitch and flat face sections. The main reasons for the
choice of this type of pitch distribution and these blade sections are the advantages
met when casting this type of screw. The results of the experiments mentioned show
that this type of screw did not have drawbacks with respect to efficiency and cavita-
tion. These screw series were designated the Ka-series.

NOZZLE Nal9A

FIG. 19. Profile of nozzle no. 19A.

30
 TABLE3A. Dimensions of the Ka-screw series
rlR 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Length of the blade from centre line 30.21 36.17 41.45 45.99 49.87 52.93 55.04 56.33 56.44 Length of blade section
sections in percentages to trailing edge at 0.6 R =
of the maximum length
of the blade section from centre line 36.94 40.42 43.74 47.02 50.13 52.93 55.04 56.33 56.44
at 0.6 R to leading edge
total length 67.15 76.59 85.19 93.01 100.00 105.86 110.08 112.66 122.88
.-

Max. blade thickness in percentages of the diam. 4.00 3.52 3.00 2.45 1 90 1.38 0.92 0.61 0.50 Maximum thickness at
centre of shaft = 0.049 D
Distance of maximum thickness from leading edge 34.98 39.76 46.02 49.13 49.98 - - - -
in percentages of the length of the sections

TABLE3 ~ Ordinates
. of the Ka-screw series
- -

Distance of the ordinates from the maximum thickness

From maximulll thickness to trailing edge From maximum thickness to leading edge
rl R 100% 80 % 60 % 40 % 20 % 20 % 40% 60 % 80 % 90% 95% 100%
Ordinates for the back
0.2 - - 38.23 63.65 82.40 95.00 97.92 90.83 77.19 55.00 38.75 27.40 -
0.3 - 39.05 66.63 84.14 95.86 97.63 90.06 75.62 53.02 37.87 27.57 -
0.4 - 40.56 66.94 85.69 96.25 97.22 88.89 73.61 50.00 34.72 25.83 -
0.5 - 41.77 68.59 86.42 96.60 96.77 87.10 70.46 45.84 30.22 22.24 -
0.6 - 43.58 68.26 85.89 96.47 96.47 85.89 68.26 43.58 28.59 20.44 v

0.7 - 45.31 69.24 86.33 96.58 96.58 86.33 69.24 45.31 30.79 22.88 -
0.8 - 48.16 70.84 87.04 96.76 96.76 87.04 70.84 48.16 34.39 26.90
-
0.9 51.75 72.94 88.09 97.17 97.17 88.09 72.94 5 1.75 38.87 3 1.87
1.o - 52.00 73.00 88.00 97.00 97.00 88.00 73.00 52.00 39.25 32.31 -

Ordinates for the face

7.29 1.77 0.1 - 0.21 1.46 4.37 10.52 16.04 20.62 33.33
4.62 1.07 - 0.12 0.83 2.72 6.15 8.28 10.30 21.18
2.36 0.56 - v 0.42 1.39 2.92 3.89 4.44 13.47
0.68 0.17 - - 0.17 0.51 1.02 1.36 1.53 7.8 1
J

Note: The percentages of the ordinates relate to the maximum thickness of the corresponding section.
FIG.20. Particulars of the Ka 3-65; Ka 4-55; Ka 4-70 and Ka 5-75 screw series.

Up till now open-water tests with the Ka 3-65; Ka 4-55; Ka 4-70 and Ka 5-75
screw series in combination with nozzle no. 19A, have been performed. The particulars
of these screw models are given in Table 3 and in Fig. 20. The screws were located in
the nozzles with a uniform tip clearance of 1 mm (about 0.4 percent of the screw
diameter D).
The fairing of the open-water test results was performed with the aid of a CDC
3300 computer by means of a regression analysis. The thrust and torque coefficients
KT, KTNand KQ were expressed as polynomials of advance coefficient J and pitch
ratio P/D:

+A2, ( ) +A ( )J - - - - - - - - - - + A,,, (i) +

2
J ~
With the aid of a regression analysis the significant terms of the polynomials and the
values of the corresponding coefficients were determined. From this analysis it was
found that most of the terms with high powers of J and PJD were insignificant. The
results for the Ka 3-65; Ka 4-55; Ka 4-70 and Ka 5-75 screw series in combination
with nozzle no. 19A are given in Table 4.
These results can be used directly for solving problems which arise when designing
and analyzing ducted screw propellers if a computer is available. Also the results
will be given in graphical form. With the aid of a tape-controlled drawing machine the

FIG.21. Open-water test results of Ka 3-65 screw series with nozzle no. 19A.
 4. Form of ~olynomialand coeflicients of Ka-screw series in combination with nozzle no. 1 9 ~
TABLE
coefficients KT, K,, KQ and v, were drawn in the conventional way as a function of
J. The diagrams of the Ka 3-65; Ka 4-55; Ka 4-70 and Ka 5-75 screw series in com-
bination with nozzle no. 19A are given in Figs. 21 through 24.
For design purposes various types of more practical diagrams can be derived from
the KT - KTN- KQ - J diagram. The most widely encountered design problem of
screw propellers or ducted propellers for ships is the one where the speed of advance
of the screw V,, the power to be absorbed by the screw P, and the number of revolu-
tions N are given. The diameter D is to be chosen such that the greatest efficiency can
be obtained. The problem of the optimum diameter of the system can be solved in an
easy way by plotting qo and J as a function of
FIG.23. Open-water test results of Ka 4-70 screw series with nozzle no. 19A.

The well-known coefficient B, see (17) and (18) is related to the above dimellsionless
variable K&I~/J"'" by the equation:

where
N = number of revolutions per minute
P = power in HP
VA = speed of advance in knots
In the usual diagram, the design coefficient Bp is the base and a new speed ratio is
used. This speed ratio is defined as :

d=-N
- -D 101.27
VA J
in which
D = screw diameter in feet.
 The Bp - 6 diagrams of the different Ka screw series in nozzle no. 1 9 are
~ given in
Figs. 25 through 28.
In cases where VA,T and N or V,, T and D are given, the problem of determining
the optimum diameter or the optimum number of revolutions can be solved by plot-
ting q0 and J as functions of

K , / J ~and K,/J'

For the different K. screw series the curves for optimum diameter (on base of
K;/' J-'I2 and K:" J-') and optimum rpm (on base of KT J-') are given in
Figs. 29, 30 and 31 respectively.

In order to obtain data for analyzing the manoeuvres of ships, submersibles, dril-
ling vessels (dynamic positioning) and so on fitted with ducted propellers, open-water
tests with the Ka 4-70 screw series in combination with nozzle no. 19A were conducted
FIG. 27. Br6 diagram of Ka 4-70 screw sereis with nozzle no. 19A.

FIG. 28. Bp-6 diagram of Ka 5-75 screw series with nozzle no. 19A.
FIG.30. Curves for optimum diameter of Ka screw series in nozzle no. 19A.
 0.3 04 0.5 0.6 07 1 2 3 L 5 6 7 10
K % ~
FIG.31. Curves for optimum rpm of Ka screw series in nozzle no. 19A.

over a wider range of advance coefficients. In addition, it was of importance to have

data available to determine the thrust which can be developed at bollard pull condi-
tion either with the propeller running ahead or astern for towing vessels (tugs, push-
boats).
The ducted propellers were tested at the following combinations of speed and rpm :
1st quadrant: speed ahead, rpm ahead. In this quadrant the results of normal open-
water tests are given. The measurements are, however, extended to
operating conditions where the thrust of the ducted propeller system
becomes negative. The advance angle /3 (tan P = V,/0.7 xnD) varies
between 0 and 90 degrees,
0" <B <90°
2nd quadrant: speed ahead, rpm astern
90" G /3 G 180"
3rd quadrant: speed astern, rpm astern
180" </3 <270"
4th quadrant: speed astern, rpm ahead
270" <360"
The measurements were performed with the usual apparatus for open-water tests
with ducted propellers. The immersion of the propeller shaft was equal to the screw
FIG.33. Open-water test results of Ka 4-70 screw series with nozzle no. 19A in four quadrants.

In Fig. 32, the open-water results with the reversely mounted duct and impellers
with speed and rpm astern (3rd quadrant) are given.
The test results over the whole range of operating conditions of the ducted propeller
system are given in Fig. 33. The usually accepted way of presenting the results of open-
water tests in the form of the thrust and torque coefficients KT, KTN,KQ as functions
of the advance coefficient J is unsuitable for representation of the ducted propeller
characteristics in all of the cases encountered during the four-quadrant measurements.
When n approaches zero, the advance coefficient J as well as the thrust and torque
coefficients KT and KO approach infinity and thus these coefficients cannot be used to
describe the propeller characteristics. Therefore the modified thrust and torque co-
efficients
FIG.34. Profile of nozzle no. 37. NOZZLE W 3 7

KaL-70 SCREW SERIES

IN NOZZLE N0.37
AHEAD

FIG.35. Open-water test results of Ka 4-70 screw series with nozzle no. 37.
are given as functions of the advance angle 1:

in the diagram. This is about the same representation as used by Mr~rovrcr-r(19),

BAKERand PATTERSON (20) and VANLAMMEREN et a1 (18).
For towing vessels (especially pushboats), the thrust which can be developed at
bollard pull condition either with the propeller running ahead or astern is of utmost
importance. In such cases it is attractive to use a nozzle with a relatively thick trailing
edge. Such a nozzle (nozzle no. 37) is shown in Fig. 34.

FIG.36. Open-water test results of Ka 4-70 screw series with nozzle no. 37; with reversely mounted
nozzle and impeller.
 0"

FIG.37. Open.-water test results of Ka 4-70 screw series with nozzle no. 37 in four quadrants.

FIG.38. Result of four-quadrant

measurements of Ka 4-70 screw
series with nozzle no. 19A ap-
proximated by Fourier series of
20 terms.
TABLE
5. Coefficients of Fourier series by which results are a~~roximated
of four-auadrant mea
ments with Ka 4-70 series screws in nozzle no. 19A

10
c*, 11
12
13
14
15
16
17
18
19
20
 FIG. 39. Result of four-quadrant
measurements of Ka 4-70 screw
series with nozzle no. 37 approxi-
mated by Fourier series of 20 terms.

The results of the four-quadrant measurements with nozzle no. 37 in combination

with the Ka 4-70 screw series are shown in the Figs. 35 through 37. The open-water
test results over the normal test range (speed and rpm ahead) and the results with the
reversely mounted nozzle and impeller with speed and rpm astern (3rd quadrant) are
givenintheformof KT - KT, - K, - J diagrams in Figs. 35 and 36 respectively.
The test results over the whole range of operating conditions are given in the form of
a C; - - C; - p diagram in Fig. 37.

In ship control studies, it is necessary to have a mathematical representation of the

data suitable for use on a computer. Therefore calculations were carried out to re-
present the results of the four-quadrant measurements by a Fourier series. The thrust
and torque coefficients C;, and C; were approximated by the following series:

C; = C [A(k) cos j k + B(k) sin pk]

k=O

ditto
ditto

In Figs. 38 and 39, the test results and their approximations by a Fourier series of
20 terms are given for the Ka 4-70 screw series in nozzle no. 19A and nozzle no. 37
successively. From the results given in these diagrams it can be seen that the charac-
teristics of ducted propellers in the four-quadrants can be adequately represented by
a Fourier series of 20 terms. The values of the Fourier coefficients of the ducted
propellers with nozzles no. 19A and no. 37 are given in the Tables 5 and 6.
 6. T H E FLOW DECELERATING NOZZLE

In section 3 a simple method for the calculation of the characteristics of ducted

propellers was discussed. This method has been followed for the determination of the
characteristics of systematic series of ducted propellers with decelerating nozzles.
The data used for the computations and the results are given in Tables I, 11and I11
of the Appendix. Some results are presented in Figs. 40 through 45. Figs. 40 and 41
show the effect of the thrust ratio z, on the shape of the nozzle and the character-
istics of the ducted propeller systems. The effects of the nozzle length-diameter ratio
LID and the thickness ratio of the nozzle profile S/L are shown in Figs. 42 through 45.
The pressure distribution along the interior and the exterior surface of the various
nozzles are given in Figs. 46 and 47.
Based on the theory a number of nozzle shapes with the following variations of the
design coefficients were manufactured and tested:
- variations in the thrust ratio zo at the design thrust coefficient C,, (nozzles nos.
30, 31 and 32).
- variations in the length-diameter ratio LID (nozzles nos. 33, 35 and 36)
- variations in the thickness ratio of the nozzle profiles SIL (nozzles nds. 33 and 34).
The design coefficients for these nozzles are summarized in Table 7. The nozzle
shapes are presented in Figs. 48, 49 and 50 and are tabulated in Table 8.

FIG.40. Meanlines and characteristics of systematic series of nozzles with CTo= 0.5 and 7, = 1.0;
1.2 and 1.4.
 g '10
L. =lL
Tl0-067L
---- YL=015 11,-0686

Go -10
L,, = t 2
-X-o 'llO-Q738
,,--
s/L-015 71,-0746

FIG.41. Mean lines and characteristics of systematic series of nozzles with CTo= 1.0 and 7, = 1 .O;
1.2 and 1.4.

FIG. 42. Mean lines and characteristics of systematic series of nozzles with = 0.5; = 1.2 and
different length diameter ratios.
 -s~=o
----. yL - 'leans
a15 'li0.716

All nozzles : CL -1.0

FIG.43. Meanlines and characteristics of systematic series of nozzles with CTo= 1.0; 7, = 1.2 and
different length diameter ratios.

AU nozzles: CL-0.5
T.0 -12

FIG.44. Meanlines and characteristics of systematic series of nozzles with CTo= 0.5; T,, = 1.2 and
with different thickness ratios of the nozzle profile.
 W

all nozzles

0.6 I \
S/L-OIS ?b-P716
c%O iI I
01.

02
1
l
All nozzles:
-
CTo-tD
to 1.2
Lk0.6

'0 4 a 8 j ? l.6 2D
d/o="
Meanlins and characteristics ot systematic
series of nozzles with CT~-UJ: X1
-,2. and Q =Gh-'tio
ditterent thickness ratios X.(SlL-0 ; 0.05 i 0.10
and a15 ).
0

FIG.45. Mean lines and characteristics of systematic series of nozzles with CTo= 1.0; T~ = 1.2 and
with different thickness ratios of the nozzle profiles.

FIG.46. Pressure distribution along interior and exterior surface of nozzles with Cro = 0.5 and
7, =1.0; 1.2 and 1.4.
FIG.47. Pressure distribution along interior and exterior surface of nozzles with CTo= 1.0 and
7, = 1.0; 1.2 and 1.4.

FIG.48. Particulars of nozzles nos. 30, 31 and 32.

57
 FIG.49. Particulars of nozzles nos.
33, 35 and 36.

Nozzle 34

FIG.50. Particulars of nozzles nos. 33 and

34.
 7. Design parameters of tested nozzle shapes.
TABLE
Nozzle number CTO 70

8. Particulars of nozzles nos. 30 through 36.

TABLE

xlL S (4lL
--
nozzle
number
30 31 32 33 34 35 36

- 0.500 + 0.0112 - 0.0073 - 0.0232 - 0.0083 + 0.0005 - 0.0067 - 0.0071

- 0.416 + 0.0130 + 0.0008 - 0.0096 + 0 + 0.0007 - 0.0008 - 0.0034
- 0.333 f0.0139 +0.0067 f0.0006 f0.0058 +0.0008 +0.0033 -0.0008
- 0.250 + 0.0128 + 0.0093 + 0.0067 + 0.0083 + 0.0011 + 0.0056 + 0.0004
- 0.167 + 0.0102 + 0.0091 + 0.0085 + 0.0083 + 0.0012 + 0.0055 + 0.0012
- 0.084 + 0.0059 + 0.0060 + 0.0063 + 0.0058 + 0.001 1 + 0.0037 + 0.0008
+ 0 0 + O + O f O f 0 0 0
+ 0.084 - 0.0081 - 0.0096 - 0.0112 - 0.0010 - 0.0009 - 0.0064 - 0.0054
+ 0.167 - 0.0186 - 0.0229 - 0.0276 - 0.0023 - 0.0021 - 0.0152 - 0.0158
+ 0.250 - 0.0316 - 0.0404 - 0.0496 - 0.0042 - 0.0038 - 0.0389 - 0.0279
+ 0.333 - 0.0476 - 0.0622 - 0.0773 - 0.0066 - 0.0058 - 0.0460 - 0.0417
+ 0.416 - 0.0661 - 0.0881 - 0.1106 - 0.0092 - 0.0083 - 0.0675 - 0.0592
+ 0.500 -0.0874 -0.1182 -0.1498 -0.0124 -0.0112 -0.0951 -0.0816

The experiments were all carried out with a series of five bladed Kaplan type
screws (Kd 5-100 series). The pitch distributions of the screws depend on the velocities
induced by the nozzle at the impeller plane and on the radial load distribution of the
screw. The design method is discussed at length in (11) and (13). Particulars of the
screw models are given in Table 9 and in Fig. 5 1. The screws were located in the nozzles

9. Particulars of screw models of the Kd 5-100 series

TABLE

Diameter D 240 mm
Number of blades z 5
Pitch ratio (at 0.7 D) PID 1.0-1.2-1.4-1.6-1.8
Blade area ratio AE/Ao 1.00
Blade outline Kaplan type
Blade section NASA 16-parabolic camberline
Propellers indicated by nos. - 3930-3931 -3932-3933 -3934
Screw nu3930

p pressure side SO.'

lp back side -5'12'

Particulars
All screw:
D=2LOmm
2.5
46- a167
ASB; 1.0
Screw m3930 Ph-1.0 (at Q7R)
.. no.3931 .. -U ..
.. m393 .. .l& ..
.. n0.3933 .. -1.6 ..
no.393L - -18 ..

FIG.51. Particulars of screw models of the

Kd 5-100 screw series.
with a uniform tip clearance of I mm. (about 0.4 percent of the screw diameter D).
It must be noted that in the presentation of the test results the diameter now denotes
the tip diameter of the screw.

The tests were carried out with the usual tank apparatus for open-water tests of
ducted propellers as described in the preceding section.
All the open-water test results were faired by computer and plotted in the conven-
tional way with the coefficients K,, KT,, KQ and v, as functions of the advance co-
efficient J. The diagrams are given as Figs. 52 through 58. The experimentally ob-
tained relations between the thrust coefficient C, and the thrust ratio z of the various
nozzles are given in the Figs. 59 and 60. In addition, the design thrust coefficient C,
and the design thrust ratio z are given in these diagrams.

FIG.52. Open-water test results of Kd 5-100 screw series with nozzle no. 30.
FIG.55. Open-water test results of Kd 5-100 screw series with nozzle no. 33.
FIG.58. Open-water test results of Kd 5-100 screw series with nozzle no. 36.
FIG. 59. Relationship
between C, and r of
nozzles nos. 30, 31 and
32.

FIG. 60. Relationship

between CT and T of
nozzles nos. 33, 34, 35
and 36.
 The relationship between the thrust coefficient CT and the thrust ratio z of nozzle
no. 33 in combination with the various screws of the Kd 5-100 screw series is shown in
Fig. 60. It can be seen from this diagram that, as might be expected, this relationship
is approximately independent of the pitch ratio of the screws. The small differences
which occur may be explained by the differences in rotational losses and radial load
distributions of the screws and by frictional effects.
The relationships between C, and z of nozzles nos. 30, 31 and 32 (Fig. 59) and
nozzles nos. 34, 35 and 36 (Fig. 60) were, therefore, only given for the screw with a
pitch ratio of 1.4. Although a number of simpliFying assumptions were made in the
development of the theory for the numerical calculations of the nozzles shapes (it was
a linearized theory, the impeller was represented by a uniformly loaded actuator disk,
tip clearance effects were not taken into account, the effect of friction was neglected)
a reasonable agreement between the calculated relationship of CT and z of the various
nozzles and those obtained by the open-water tests was found.

FIG. 61. Optimum relationship

between q,, 6 and Bp of nozzles
nos. 30, 31 and 32.

FIG. 62. Optimum relationship

between q,, 6 and Bp of nozzles
nos. 33, 34,35 and 36.
 The open-water efficiency q, and the diameter coefficient 6 of the various nozzles
were given for optimum condition from the viewpoint of efficiency in the Figs. 61
and 62 on a base of Bp. In combination with Figs. 59 and 60, these diagrams show
clearly that by comparing the properties of the different nozzles, the open-water
efficiency decreases with ihcreasing thrust ratio z. The diameter coefficients of the
various nozzles are nearly the same.

The prevention of strong vortices along the hub of the decelerating ducted propeller
is of importance because these vortices may result in hub vortex cavitation. This can
be achieved by the application of vanes or a stator located downstream of the impeller.
It may also increase the efficiency if the vanes eliminate the rotational losses without
incurring excessive frictional resistance.
Based on the theory as discussed in (1 1) and (13), a six-bladed stator was designed
for application downstream of the Kd 5-100 screw with P/D = 1.4 in nozzle no. 33.
Particulars of the stator are given in Fig. 63. Open-water tests with the Kd 5-100
screw series in combination with nozzle no. 33 and the six-bladed stator were per-
formed. The result is given in Fig. 64. The stator was fitted to the nozzle, consequently,
KTNdenotes the thrust coefficient of nozzle and stator.

FIG. 63. Particulars of six-bladed

stator.
FIG.64. Open-water test results of Kd 5-100 screw series and nozzle no. 33 fitted with stator.

In the case of application of a screw or a ducted propeller behind a ship, a part of

the rotational losses of the screw or impeiler is already eliminated due to the presence
of the rudder in the propeller slipstream. Consequently, the gain in efficiency due to
the stator will be lower as may be expected from comparing the results given in Figs.
55 and 64.
For comparison purposes, model tests have been carried out to determine the effect
of the presence of a rudder on the characteristics of a ducted propeller system. A more
or less common rudder configuration was therefore fitted to nozzle no. 33 and open-
water tests were performed with this system in combination with the Kd 5-100 screw
series. Particulars of the nozzle-rudder system are given in Fig. 65. The results of the
open-water tests are presented in Fig. 66.
 FIG.65. Particulars of rudder
fitted to nozzle no. 33.

The optimum curves for efficiency of the Kd 5-100 screw series in combination with
successively nozzle no. 33, nozzle 33 with six-bladed stator and nozzle no. 33 with
rudder, are given in Fig. 67. From this diagram it can be seen that both the stator
and the rudder give an increase of the efficiency. The ducted propeller equipped with
stator has the smallest optimum diameter.
FIG.67. Optimum relationship b-tween qv, 8 and B, of Kd 5-100 screw series and successively
nozzle no. 33 and nozzle no. 33 fitted with stator and with rudder.
 7. G E N E R A L DISCUSSION O F TEST RESULTS

In section 6, it was found that the relation between the thrust coefficient CT and tke
thrust ratio z of a ducted propeller system uas approximately independent of the pitch
ratio of the impeller. This result was for the Kd 5-100 screw series in combination
with nozzle no. 33 shown in Fig. 60.
The experimentally obtained relation between CT and z of the Ka 4-70 screw
series with nozzle no. 19A is shown in Fig. 68. The results presented in this diagram
confirm the above mentioned conclusion.
In section 2, the following relation between the impeller disk area-nozzle exit area
ratio A,/A,,, the thrust coefficient CT and the thrust ratio z was derived:

In the derivation of this expression it was assumed that the tangentially induced
velocities and consequently the losses due to the rotation of the fluid were zero.
For the range of impeller pitch ratios of ducted propellers normally considered, it
can be expected that the rotation of the fluid slightly influences the forces developed
on the nozzle or the nozzle thrust. A rough indication of the decrease in propeller
thrust due to the rotation of the fluid can be obtained from axial flow pump theory.
With the aid of this theory, the ratio between the propeller thrust influenced by rota-
tion and the propeller thrust for zero rotational losses CTpcan be written as:

Some results of the calculations performed with these relations are also presented in
Fig. 68.
It must be noted that in the derivation of these relations a large number of simpli-
fyJng assumptions were made. Therefore this theory is only useful to make a qualitative
analysis of the characteristics of ducted propellers. It can be seen from Fig. 68, that
the tendencies found from the theoretical calculations and the experimental results are
in good agreement with each other.

From all model tests performed with accelerating and decelerating nozzles as
described in the sections 5 and 6, it can be deduced that for the considered nozzles
(with length-diameter ratios LID between 0.4 and f.O), there exists a fixed relation
between C,, z and the ratio between the disk area of thenozzle. This result is shown in
Fig. 69. The numbers of the considered nozzles are indicated in the diagram. This
diagram gives a simple relation between the nozzle geometry and the characteristics
 FIG. 68. Relationship
between C* and t of
nozzle no. 19A.

of the ducted propeller system. The result will be used in the following sections for the
design of various special nozzle shapes.

Another general result which can be obtained from the model tests is presented in
Fig. 70. In this diagram the open-water efficiency of ducted propeller systems is given

FIG.69. Relationship between C=,T and nozzle geometry.

76
 I 30
DECELERATING NOZZLES
31 33 34 35 32
I

36
FIG.70. Relationship between q,, CTand nozzle geometry.

on a base of the nozzle area ratio Ao/AExand for different values of the thrust co-
efficient C=.By combining the results presented in Figs. 69 and 70 it can be seen that
with decreasing value of the area ratio A,/A,,, the thrust ratio z decreases and the
efficiency qo of the ducted propeller system increases. Thus with increasing positive
loading of the nozzle, the efficiency of the ducted propeller system increases. The load-
ing of the nozzle is limited by the risk of flow separation on the interior surface of the
nozzle. A criterion for the nozzle loading was given in section 4.
 As a comparison, the optimum curves for open-water efficiency v,,, diameter co-
efficient 6 and thrust ratio z of the Ka 4-70 screw series in nozzle no. 19A, the B 4-70
screw series and the Kd 5-100 screw series in nozzle no. 33 are given in Fig. 71 on a
base of Bp. Screws of the B 4-70 screw series are usually applied behind single screw
cargo ships (see for instance (l 7)). Typical B, and CTvalues for different ship types are
indicated in Table 10. The lightly loaded screws of fast ships are on the left side of
Fig. 71 while the heavily loaded propellers of towing vessels are on the right.
It can be seen from Fig. 71 that at low B,-values the open-water efficiencq of both
the accelerating and the decelerating nozzle decrease with respect to the efficiency of
the B 4-70 screw series. This fact can be explained by the relative increase of the fric-
tional and the induced drag of the nozzle (see also Fig. 11). The curves of the diameter
coefficient 6 of the accelerating and the decelerating nozzle almost coincide; the B 4-70
screw series has a larger optimum screw diameter. It is interesting to note that the
curves of the diameter coefficient based on the maximum diameter of the system 6*
of both accelerating and the decelerating nozzle and the B 4-70 screw series almost
coincide.
The accelerating nozzle (nozzle no. 19A), if compared with a conventional screw
propeller (B 4-70 screw series), gives rise to an improvement in open-water efficiency
v,, in the case of heavy screw loads (see Fig. 71). The decelerating nozzle (nozzle no.
33) has a relatively low open-water efficiency.

FIG.71. Optimum relationship between v,, 6 and Bp of the Ka 4-70 screw series in nozzle no. 19A,
the B 4-70 screw series and the Kd 5-100 screw series in nozzle no. 33.
TABLE
10. Typical Bp and CT values for different ship types

Torpedo's
Twin-screw ships
Fast warships (frigates, destroyers)
Single screw cargo ships
Coasters
Tankers
Trawlers
Towing vessels (tugs, pushboats)

In the case of heavy screw loads (a11 types of towing vessels) the attractiveness with
regard to propulsive efficiency of application of the accelerating nozzle has been
demonstrated in practice in the course of the past thirty years.
For middle and lower screw loads (tankers and single-screw cargo ships) the in-
crease in propulsive efficiency through application of an accelerating nozzle strongly
depends on the stern-nozzle configuration. In the case of lightly loaded screws the
application of a decelerating nozzle may become attractive when other requirements
than the efficiency influence the choice. For naval ships it is of utmost importance
that the inception speed, or the speed at which cavitation phenomena occur at the
screw will be as high as possible. For tactical reasons a minimum sound radiation of
propeller noise due to cavitation is necessary for these ships.

Finally it must be noted that the flow velocity at the impeller disk of a ducted pro-
peller is far less sensitive to variations in ship speed than the flow velocity at an or-
dinary propeller. Consequently the power absorption of a ducted propeller is relative-
ly less sensitive to variations in the ship speed. This feature has also been an important
factor in the case of application of ducted propellers behind tugs, pushboats and
trawlers. All these ships must operate satisfactorily at different loadings (towing and
free-running) of the screw. For towing vessels the thrust which can be developed at
bollard pull condition, either with the propeller running ahead or astern, is of im-
portance. In such cases, as stated in section 5, itheuse of a nozzle with a relatively thick
trailing edge is attractive. Such a nozzle (nozzle no. 37) was shown in Fig. 34.
The optimum curves for open-water efficiency v, and diameter coefficient 6 of the
Ka 4-70 screw series in combination with successively nozzle no. 19A and nozzle no.
37 are given in Fig. 72 on a base of Bp. From this diagram it c m be seen that, as may be
expected due to the increase of the nozzle drag by the relatively thick trailing edge of
the nozzle profile, the ducted propeller system with nozzle no. 37 has a lower open-
water efficiency than the system with nozzle no. 19A.
In Figs. 73 through 76, the thrust force at bollard pull condition and with the im-
peller running ahead and astern is given on a base of the power with the diameter and
FIG.72. Optimum relationship between q, 6 and Bp of the Ka 4-70 screw series in respectively
nozzle no. 19A and nozzle no. 37.

the rpm as parameters. The result with nozzle no. 19A is given in Figs. 73 and74
whereas this result for nozzle no. 37 is given in the Figs. 75 and 76.
The thrust-power ratios for the different nozzle configurations and flow conditions
are for instance for a power absorption of 5000 HP and an impeller diameter of 4 m:

nozzle no. 1 9 ~ahead:

, TIP = 13.6 kg/HP
astern : = 8.3 kg/HP
nozzle no. 37 ahead : = 14.8 kg/HP
astern : = 12.2 kg/HP

From these results it can be seen that nozzle no. 37 at bollard pull condition has the
best characteristics, even with the impeller running ahead.
FIG.74. Bollard pull thrust of nozzle no. 19A with Ka 4-70 series screw FIG. 73. Bollard pull thrust of nozzle no. 19A with Ka 4-70 series screw
with P/D = 1.0; Impeller running astern. with P/D = 1 .O. Impeller running ahead.
 8. APPLICATION O F WAKE A D A P T E D NOZZLES

Before going into details of the design of wake adapted nozzles the wake behind a
ship will be discussed. Usually the variations of the flow field at the screw can be split
up into two components (see also (21)):

1. The radial variation, especially of the axial velocities. 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 non-uniform flow by an
appropriate distribution of pitch and camber, and optimum efficiency and cavi-
tation properties may be expected in such cases.

2. The circumferential (at a given radius) variation of both the axial and tangential
velocities. This non-uniformity is the origin of the periodically fluctuating pressure
distributions along the blade chords and leads to the unsteady force-pattern at the
screw and at the stern of the ship. These periodically fluctuating pressure distri-
butions lead to unsteady cavitation phenomena which may be serious from a
viewpoint of erosion and noise radiation. In addition, these unsteady pressure
distributions lead to propeller induced vibrations.
The inflow velocity of the screw can be made more constant over the screw disk by
surrounding the propeller by a non-axisymmetrical nozzle which is adapted to the
wake distribution and flow direction behind the ship.

In the case of a single-screw ship the intake velocity will be lower in the upper part
of the screw disk than in the lower part. Consequently, the propeller is more heavily
loaded in the upper part of the screw disk. The inflow velocity of the propeller can be
made more constant over the screw disk by a non-axisymmetrical nozzle which
accelerates the flow in the upper part of the screw disk and decelerates the flow in the
lower part of the disk.

In the case of twin-screw ships the propellers operate in a varying inflow, due to the
shaft inclination. The inclination is a consequence of the fact that the propeller shaft
has a sizeable inclination to both the horizon and the buttock lines in way of the pro-
peller. The non-axisymmetrical nozzle must be designed here, from the viewpoint of
retardation of screw cavitation, in such a way that the actual effectiveincidence changes
of the blade sections of the impeller will be as low as possible during a revolution.

Investigations with respect to special nozzle configurations will be described in the

following sections.
 From Fig. 71 it could be seen that propellers in nozzles are to be recommended for
heavy screw loadings such as occur on towing vessels, trawlers and very large tankers.
The possibilities of using nozzles on towing vessels (tugs, push-boats) have been
adequately demonstrated in practice.
The trend in the design of tankers has been towards large dimensions at neatly
constant speed and therefore high-powered ships. Although the rpm has been de-
creased and consequently the diameter of the screw was chosen larger, the loading
coefficient of the screw of these ships has still become larger. Therefore, according to
Fig. 71, a larger reduction in shaft horsepower due to application of a ducted pro-
peller in comparison with a conventional screw may be expected for these ships when
their size is increased.
Based on results, as presented in Fig. 71, propulsion tests have been performed at
the Netherlands Ship Model Basin with a large number of tanker models equipped
with ducted propellers. Some results of these tests were given in (22, 23, 24, 25). The
investigations performed by LINDGREN et a1 (26) may be mentioned also in this respect.
The results of these tests generally confirm the conclusion that an increase in pro-
pulsive efficiency can be obtained by application of a ducted propeller for this ship
type if compared with a conventional screw. Some of the more extensive tests will be
discussed here.

The total propulsive efficiency of a self-propelled ship depends on the open-water

efficiency of the screw and on the mutual interference of screw and hull. A screw
fitted behind a ship normally increases her drag, because of the decreasing pressure
(suction field), which accompanies the acceleration of the flow in front of the screw.
This additional drag is called thrust deduction. The open-water efficiency of a screw
with an accelerating nozzle is, for the range of screw loadings of tankers, better than
that of an open-screw, but the higher thrust deduction of the first system may reduce
this gain.
Therefore, tests have been carried out with a model of a 95.000 TDW tanker with
conventional stern and fitted with a conventional screw or one of a series of ducted
propeller systems. The nozzles were designed for different acceleration ratios of the
flow. The design of these nozzles was based on the results presented in Fig. 69. The
profiles of the tested nozzle shapes are given in Fig. 77. The nozzle area ratio Ao/AEx
of the different nozzles are also indicated in the diagram, since this ratio is one of the
important characteristics of a ducted propeller.
Analysis of the reduction in DHP due to the use of the diierent nozzles in com-
parison with the conventional screw propelIer has been made for a ship speed of
16.5 knots and for the loaded and ballast condition of the vessel. The result is given in
Fig. 77. This diagram clearly shows that for the loaded as well as for the ballast
condition the most accelerating nozzle (nozzle 11) still gives the largest reduction in
DHP. Further, it can be seen that an increase in DHP of about l percent is found by
making the trailing edge of the nozzle profile round instead of sharp.
 130
ballast c m d ~ t ~ o n
Shtp sped l65 knols
,120

-110
OHP nozzle
DHPopen screw

- 100

ago

- 0 80
0.8 W U) A~/A~;' L2 1.3

FIG.77. Reduction in DHP due to the application of different axisymmetrical ducted propeller
systems behind tankers.

It is a well known fact that the propulsive efficiency of a tanker can be further in-
creased by application of a ducted propeller, in combination with a cigar-shaped stern.
The cigar-shaped stern tends to make the flow more uniform and to bend the flow in
a horizontal direction. In this respect a non-axisymmetrical nozzle behind a conven-
tional stern may also be attractive.
Measurements performed by WERELDSMA (27) show that for large tankers a pro-
peller thruat eccentricity of 0. l D is common practice due to the fact that the propeller
operates in the non-uniform wake. The inflow velocity of the propeller can be made
more constant over the screw disk by a non-axisymmetrical nozzle which is adapted to
the wake distribution behind the ship. This nozzle accelerates the flow in the upper
part of the screw disk (by increasing the exit area of the nozzle) and decelerates the
flow in the lower part (by decreasing the exit area of the nozzle). A view of a tanker
with a conventional stern and equipped with a non-axisymmetrical ducted propeller is
given in Fig. 78. The non-axisymmetrical nozzle is still cylindrical at the inside from the
leading edge of the nozzle to the impeller; only the aft part is non-axisymmetrical but
still cylindrical.
FIG.78. View of stern of tanker fitted with non-axisymmetrical ducted propeller.

Tests have been carried out with the model of a 95.000 TDW tanker fitted with suc-
cessively different non-axisymmetrical nozzles. The profiles of the tested nozzle shapes
are given in Fig. 79. It can be seen from this diagram that the maximum diffusor angles
(in the upper part of the screw disk) of the nozzles V, Va and V11 are equal to the
diffusor angle of nozzle 11. Nozzle V1 has a larger diffusor angle. Nozzles V, Va and
V1 are made non-axisymmetrical in the same degree. Nozzle V11 is less non-axi-
symmetrical.
An analysis of the reduction in DHP due to the use of the different nozzles in com-
parison with the conventional screw propeller has again been made for a ship speed of
16.5 knots and for the loaded and the ballast condition of the vessel. The result is
given in Fig. 79. The curves obtained with the axisymmetrical ducts are also indicated
in the diagram. From this diagram it can be seen that nozzles V and Va give the largest
reduction in DHP. Nozzle V1 which is the most accelerating nozzle type of the series
 I30
ballast condition
Ship speed 165knob.
/
/
l20

-1.10
OHP nozzle
DHPopcn screw
iao

P90

0.80
R8 0.9 l0 AdAEl.l 12 13

FIG.79. Reduction in DHP due to the application of different non-axisymmetrical ducted propeller
systems behind tankers.

probably suffers from flow separation in the upper part of the screw disk at the inside
of the nozzle, as the diffusor angle of the nozzle is too large. Nozzle Va is slightly
better than nozzle VII. It can be seen once more that a small increase in DHP is
found by making the trailing edge of the nozzle profile round instead of sharp. By
comparing the results presented in Figs. 77 and 79 it can be seen that the best non-
axisymmetrical nozzle (nozzle Va or V) in comparison with the best axi-symmetrical
nozzle gives a reduction in DHP of about 2 percent.
In Table 11 the results are compared of a large number of model self-propulsion
tests, performed with tankers with conventional and cigar-shaped stern arrangements
and fitted with conventional screws, ducted propellers (nozzles I and Ia) and non-
axisymmetrical ducted propellers (nozzle Va).
From this table it can be seen that the conventional stern with non-axisyrnmetrical
ducted propeller gives a reduction in DHP which is still larger than can be obtained by
 11. Reduction in DHP if compared with tanker with conventional stern and screw
TABLE
Configuration Loaded Ballast
condition condition

Ship with conventional stern and axisymmetrical ducted propeller 2-6%

Ship with conventional stern and non-axisymmetrical ducted reductions
propeller 6-9 % about 2-3 %
larger
Ship with cigar-shaped stern and axisymmetrical ducted propeller 5-8 %

application of a cigar-shaped stern with an axisymmetrical nozzle. In addition, it can

be expected from the homogenizing effect of the non-axisymmetrical nozzle on the
inflow velocity of the impeller that the non-axisymmetrical ducted propeller offers a
means of minimizing propeller induced vibration and cavitation problems.
The suitable non-axisymmetrical nozzle shapes for large tankers discussed so far
were determined experimentally. For the development of a more refined design
method the wake-field behind the ship must be known. Normally wake surveys are
made without running propeller. In the case of tankers which have a full form a
region of flow separation occurs above the propeller. This causes quite different
wake fields with and without the propeller running. This fact is illustrated in Fig. 80.
In this diagram the results of wake surveys conducted on the model of a 240.000 TDW
tanker are given. The pitot-tube wake surveys were made in a plane about 0.15 D
upstream of the screw propeller. The tests were performed for the full load condition
and at the service speed of the ship.
It can be seen from the wake survey without running propeller that the mean flow

WITHOUT PROPELLER
I I I V .AXIAL WAKE VELOCITY

FIG. 80. Wake of large tanker with and without running propeller.

88
velocity in the upper part of the screw disk is larger than in the lower part. Based on
this result it should be expected that the screw is relatively more heavily loaded in the
lower part of the screw disk. This is in contradiction with the results of measurements
of the thrust excentricity and the results of observations in a cavitation tunnel.
However, from analyzing the wake survey with running propeller it can be seen
that the mean velocity in the upper part of the screw disk is lower than in the lower
part. Due to the presence of the ship's hull and probably due to flow separation, it is
more difficult to suck water in the upper part than in the lower part of the screw disk.
It is clear that this result is in accordance with the usually found thrust eccentricity
and the results of cavitation observation tests. A more refined design of the non-
axisymmetrical nozzle must therefore in future be based on the results of a wake sur-
vey with running propeller.
In addition, the determination of a wake survey with running propeller in the case
of the application of a non-axisyrnmetrical ducted propeller will give the final answer
how far this system has made the flow more uniform at the impeller disk.
1050

1025
DHP
DHP nordru

FIG.81. Reduction in DHP

due to application of ducted
propeller systems with differ-
ent lengthdiameter ratios
LID behind a tanker.
 Finally a systematic series of tests with a tanker model (210.000 TDW) fitted
with ducted propellers with different nozzle length-diameter ratios were performed.
The considered nozzle profiles are given in Fig. 8 1. In this diagram, the results of the
model self-propulsion tests are also given. These nozzles all have the same nozzle
area ratio A,/AEx and yet the shortest nozzle (LID = 0.40) is the best. However, the
difference with the nozzle with LID = 0.55 is small.

8.3. APPLICATION OF WAKE ADAPTED NOZZLES BEHIND TWIN-SCREW SHIPS

For naval ships it is of importance that the inception speed, or the lowest speed at
which cavitation phenomena on the screw will occur, should be as high as possible.
For tactical reasons a minimum sound radiation of propeller noise due to cavitation
is necessary for these ships. Usually, fast naval ships (frigates, destroyers) are twin-
screw vessels and typical B, or CT values for the screws of these ships were given in
section 7. These values were B, 10-20; CTcz 0.5-1.0. From Fig. 13 it could be
seen that the application of a decelerating ducted propeller may be attractive for fast
naval ships, as this nozzle can attribute to a retardation of propeller cavitation.
In the case of twin-screw ships, the propellers operate in a varying inflow, due to
the shaft inclination. This inclination is a consequence of the fact that the propeller
has a sizeable inclination to both the horizon and the buttock lines in way of the
propeller. Wake data indicate that the flow is probably following the buttock lines
closely and should have a fairly uniform inclination over the disk. A view of the pro-
peller and the hull in way of the propeller is given in Fig. 82. In addition the velocity

hull buttock in way of

propeller centerline

propeller disc
Looking inboard looking forward

FIG.82. View of propeller and hull of twin screw ship.

diagrams of a screw blade-element at a radius 0.7 R are given in this Figure for dif-
ferent positions of the blade during a revolution.
From the viewpoint of retardation of screw cavitation, the application of a non-
axisymmetrical nozzle may be attractive here. This nozzle must be designed is such
a way that the actual effective incidence changes of the blade sections of the impeller
will be as low as possible during a revolution. A part of the stern of a twin-screw ship
fitted with non-axisymmetrical nozzles is shown in Fig. 83. Particulars of the nozzle
are given in Fig. 84. This nozzle (the starboard nozzle) accelerates the flow velocity
with respect to the mean flow velocity at the starboard side (0 = 90")by increasing the
exit area of the nozzle and decelerates the flow at the port side (0 = 270") by decreas-
ing the exit area of the nozzle. The nozzle is still cylindrical at the trailing edge. The
leading edge of the nozzle is adapted to the flow direction.

lookinq inboard Lookinp forward

FIG.83. Hull of twin-screw fitted with non-axisymmetrical nozzles.

The velocity diagrams of a screw blade-element at a radius 0.7 R are given in Fig. 83
for the different angular positions of the blade. These diagrams show that the angle of
incidence of the flow with respect to the blade element will be constant during a
revolution. However, the incidence velocity of the blade element will vary.
Preliminary results of tests performed with a non-axisymmetrical nozzle at different
shaft inclinations in a cavitation tunnel of the NSMB have shown that an improve-
ment in the cavitation characteristics of the impeller can be obtained. Further in-
vestigations will be performed to prove the reliability of this concept.
 FIG.84. Particulars of non-axisymme-
trical nozzle.

8.4. DUCTED
PROPELLER WITH CIRCULATION CONTROL

From tests performed with different ducted propeller systems it was found that the
characteristics of the system strongly depend on the ratio between propeller disk area
and the exit area of the nozzle. With an increase of the exit area of the nozzle, the
circulation around the nozzle profile increases, the nozzle acts more accelerating and
gives a larger contribution to the thrust of the ducted propeller system. In addition,
the efficiency of the system increases. The loading of the nozzle, however, is restricted
by the danger of flow-separation on the nozzle.
The circulation around the nozzle profile and the velocity at the impeller are strong-
ly influenced by the exit area of the nozzle. If it is possible to make the exit area of
the nozzle adjustable, for instance by means of a large number of flaps, then such a
ducted propeller may operate favourably at different operating conditions, as with a
controllable pitch propeller. In addition, it must be possible to deflect the propeller
jet and generate transverse forces by an eccentric adjustment of the exit area of the
nozzle.
From a practical point of view such a flap mechanism at the trailing edge of the
nozzle can hardly be realized. The circulation around the nozzle profile can, however,
also be effected by (a combined suction and) injection of water at the trailing edge of
the nozzle. Such a concept is shown in the Figs. 85 and 86.
FIG.85. Nozzle with combined suction and injection of water at the trailing edge of the nozzle.

By a combined suction and injection of water at the trailing edge of the nozzle the
following objects can be pursued:
- by injection of water at the exterior surface of the nozzle (eventually combined
with water suction at the interior surface) as indicated in Fig. 85, the circulation
aropnd the nozzle will be increased. The nozzle actsmore accelerating, the thrust
of the nozzle increases and the efficiency of the system becomes probably higher.
- the suction and injection of water along the trailing edge of the nozzle at the in-
terior or exterior surface offers a means to accelerate or decelerate the inflow
velocity to the impeller. By a proper choice of the flow rate through the slit at the
trailing edge the inflow velocity of the impeller can be kept constant at different
speeds of the ship.
Thus the screw can absorb the full power at constant rpm at different ship speeds.
This is attractive for tugs and trawlers (towing and free running condition) but also
for cargo ships (loaded and ballast condition, added resistance in waves).
- non-axisymmetrical suction and injection of water along the trailing edge of the
nozzle, as shown in Fig. 86, offers a means of deflecting the propeller jet. Conse-
quently a transverse force will act on the nozzle.
In this way the nozzIe can be used for the steering of the ship.
In the near future the qualities of this concept will be determined by performing
tests with a ship model fitted with a ducted propeller with circulation control.
FIG.86. Nozzle with non-axisymmetrical suction and injection of water at the trailing edge for
steering.
 9. RINGPROPELLERS

A ringpropeller is a propeller with a ring fitted to the blade tips. This ring rotates
with the propeller. The ring may be countersinked in a conventional nozzle or may
have a profile similar to the profile of the nozzle of a ducted propeller. These two
configurations are shown in Fig. 87.

FIG.87. Particulars of ringpropellers. RlNGPROPELLER IN NOZZLE

It was concluded in section 4 that the use of a nozzle only leads to an increase of
the efficiency at higher screw loads, where the gain in ideal efficiency exceeds the loss
in efficiency due to the frictional nozzle drag. It is clear that in the case of the ring-
propeller a substantial additional drag acts in circumferential direction on the ro-
tating nozzle. This extra torque will decrease the efficiency of the ringpropeller in
comparison with a conventional ducted propeller system.
In order to investigate the effect of the additional drag of the rotating ring on the
overall efficiency, calculations have been carried out for a series of ringpropellers in
nozzle no. 19A. If we neglect the tangential induced velocities due to the impeller then
the following relations hold for the axial and tangential flow velocities with respect to
the nozzle ring:

and
and the decrease in thrust and increase in torque due to the rotating ring may be
written as:

and

AC, = 2 C,--
b v,
D F'"
(-) 2

where C, is the turbulent skin-friction drag coefficient.

For a Ka 4-70ring propeller screw in nozzle no. 19A, for instance, we can derive
the relationship between Bp and qo from the test results with the corresponding Ka
4-70series screw in nozzle no. 19A by taking into account the additional drag of the
rotating ring with :

b p ] ring propeller
= p+ ACoBp
CQ

where CT, CQ,J, BP and qo correspond with the considered Ka 4-70series screw.
Calculations of the relationship between Bp and q, of the ring propeller series have
been made and the result is given in Fig. 88. From this diagram it can be seen that
over the whole range of operating conditions the efficiency losses due to the additional
drag of the rotating ring are substantial. Consequently, the rotating ring must be
made as short as possible.

Model tests have been performed with a systematic series of ring propellers in
combination with nozzle no. 19A. Propeller models of the existing Ka 4-70series
(pitch ratios P/D = 0.6;0.8;1.0; 1.2 and 1.4) were used for these tests. Each model
was provided with a ring attached to the blade tips. Particulars of the Ka 4-70 ring
propeller series in nozzle no. 19A are given in Fig. 89.
The tests were carried out with the usual tank apparatus for open-water tests with
ducted propellers. The usual routine of open-water tests was followed; the rpm of the
screw was kept constant and by varying the speed of advance the desired value of the
advance coefficient J was obtained. The open-water test results were faired and plotted
in the conventional way with the coefficients KT KTN,KQ and q, as functions of the
advance coefficient J. The diagram is given in Fig. 90.

Tn addition, open-water tests have been conducted with a series of ringpropellers

where the ring did have a profile similar to the shape of the nozzle of a ducted pro-
 0.0

0.7

nozzle no.19A (theoretically).

0.6 I I

0.5

0.L

?l0 I
a3 Model tests:

Reb = b _a35.106
0.2 v
Cf = 0.006
0.1

0
3
- 4 5 7 10 20 30 GO 50 70 100 150 200 300
BP
FIG.88. Efficiency of ringpropeller series in nozzle taking into account the additional drag of the
rotating ring.

-c----

FIG.89. Particulars of Ka 4-70 ringpropeller series in nozzle no. 19A

97
FIG.90. Open-water test results of Ka 4 7 0 ringpropeller series in nozzle no. 19A.

peller. The particulars of the ring and of the four bladed propellers with a blade-area
ratio of 0.55 are given in Fig. 91. This ringpropeller series was called the R 4-55 series.
The results of these investigations were already given by VAN GUNSTEREN (28) and
KELLER (29). The open-water test results of this series were faired and plotted in the
conventional way with the coefficientsKT,KQ and v. as functions of J. The diagram is
given in Fig. 92.

Finally, open-water tests have been carried out with a multiple nozzle arrangement
with ringpropeller and ringstator proposed by VAN GUNSTEREN. The forward part
of the nozzle ring is fixed and is fitted to a stator. This stator has 7 adjustable vanes.
The aft part of the nozzle ring is fitted to the blade tips of a screw propeller. This ring
is rotating with the propeller. This ducted propeller system is shown in Fig. 93.
FIG.91. Particulars of R 4-55 ringpropeller series.

FIG.92. Open-water test results of R 4-55 ringpropeller series.

 MULTIPLE NOZZLE-
RING PROPELLER ARRANGEMENT

FIG.93. Particulars of multiple nozzle-ringpropeller system.

Due to the fact that the nozzle profile has a slot, the nozzle may be very heavily
loaded. Therefore, a strongly flow accelerating type of nozzle profile was chosen
with a relatively large nozzle exit area. By giving different angles to the vanes of the
stator, the system acts more or less like a controllable pitch propeller. So the system
may operate satisfactorily at different loading conditions.
Tests were conducted with this system with screws of the Ka 4-55 series with pitch
ratios of P/D = 1.2; 1.4 and 1.6 fitted to the aft rotating part of the nozzle. The vanes
were designed such that at zero angle setting all the rotation was taken out of the flow
by the screw with P/D = 1.4 and for the advance coefficient J = 0.35. In addition,
tests were carried out with the screw with P/D = 1.4 and with the vanes at different
angle settings. The test results are given in the Figs. 94 and 95.
As a comparison the optimum curves for open-water efficiency v, and diameter
coefficient S of the B 4-70 screw series, the Ka 4-70 screw series in nozzle no. 19A, the
Ka 4-70 ring propeller series in nozzle no. 19A, the R 4-55 ringpropeller series and
the multiple nozzle-ring propeller system are presented on a base of the power
coefficient Bp in Fig. 96.

It can be seen from the diagram that all ducted propeller systems with a part or the
total nozzle ring attached to the propeller tips have a lower open-water efficiency than
the conventional ducted propeller and the conventional screw series. The decrease in
efficiencycan be explained by the increase of the frictional drag of the system due to
the drag of the ring attached to the propeller tips. The efficiency losses due to the in-
crease in frictional drag pointed out to be more important than the gain in efficiency
due to the reduced tip clearance losses over the range of screw loads considered. It
can also be seen from Fig. 96 that the ducted propeller systems with a ring attached to
the blade tips have a lower optimum diameter than the conventional ducted propeller
and conventional screw series.
FIG.95. Open-water test results of multiple nozzle-ringpropeller system. FIG.94. Open-water test results of multiple nozzle-ringpropeller system.
 L R 4-55 series
5 multiple nozzle-rin

01.. 1 0 0 .

0- 0
3 1 5 7 10 20 30 L0 50 70 ' 100 200 300
BP
FIG. 96. Optimum relationship between q., 6 and Bp of B 4-70 screw series, Ka 4-70 screw series in
nozzle no. 19 A, the Ka 4-70 ringpropeller series in nozzle no. 19A, the R 4-55 series and the multiple
nozzle-ringpropeller system.

Finally, the thrust which can be developed at bollard pull condition with the im-
peller running ahead was calculated for all these ducted propeller configurations. The
calculations were carried out for a power absorption of 5000 HP and a maximum
diameter of the system of 4 meters. For each system the pitch ratio of the impeller
was so chosen that the thrust-power ratio at bollard pull condition was the greatest.
The results are given in Table 12.
From this table it can be seen that for the bollard pull condition, the Ka 4-70 series
screw in nozzle no. 19Aand the multiple nozzle ring propeller system have the largest
thrust-power ratios. However, also the Ka 4-70 ring propeller series in nozzle no. 19A

TABLE
12. Thrust at bollard pull condition for different ducted propeller configurations

B 4-70 screw series 0.70 52400

Ka 4-70 series in nozzle no. 1 9 ~ 1.10 68000
Ka 4-70 ringpropeller series in nozzle no. 1 9 ~ 1.20 57400
R 4-55 ringpropeller series .1.25 56600
Multiple nozzle with ring stator and screw 1.70 65500
and the R 4-55 series have larger thrust-power ratios than the conventional B 4-70
screw series.
In a recent publication by VANGUNSTEREN (28) the advantages and disadvantages
of ringpropellers were discussed. Advantages of the ringpropeller in comparison with
the conventional screw propeller are:
- *superiorin off-design condition
- good efficiency if the conventional propeller operates at restricted diameter.
- protection of the propeller by the ring (manoeuvring in ice).
However, a disadvantage is, a relatively low efficiency. In comparison with the
conventional ducted propeller it can be stated that the ringpropeller has a lower
price, there are no problems of fitting a nozzle to the hull and there are no tip clearance
problems (cavitation, centering and so on). For coasters and trawlers, for instance,
the application of ringpropellers may be attractive.

Recently, a special type of ring propeller system has been proposed by SPARENBERG
(39). This system has a ring around the blades of which the profile varies circurn-
ferentially. The basic idea is that in this way the strong trailing vortices of the blades
can be evenly spread, resulting in a gain in efficiency. In other words, due to the finite
number of screw blades there will be a variation in the velocities in circumferential
direction in the wake of a screw. This variation will lead to additional kinetic energy
losses. By designing the nozzle in such a way that the flow between the blades will be
more accelerated than at the blades the flow in the slipstream will be more uniform.
Probably a small increase in efficiency can be obtained by application of such a ring
nozzle. However, the production of this nozzle ring seems quite difficult and expensive.
 As a result of the investigations the following conclusions can be made:
1. The accelerating nozzle offers a means of improving the efficiency of a propulsion
system at heavier screw loads (CT > 2-3). The nozzle which accelerates the flow
as much as possible is the best from the viewpoint of efficiency. This nozzle has
the lowest impeller thrust-total thrust ratio 7. The loading of the nozzle and con-
sequently the acceleration of the flow is limited, however, by the risk of flow separa-
tion on the nozzle.
2. Application of a decelerating nozzle improves the cavitation properties of a pro-
pulsion system if the gain in static pressure compensates the unfavourable effect
of the increased screw loading. This is the case for lightly loaded screws with normal
blade area ratios and number of blades.
3. Ducted propellers have a fixed relation between the thrust coefficient C,, the
thrust ratio z and the impeller area-nozzle exit area ratio, which is approximately
independent of the particulars (for instance pitch ratio) of the impeller considered.
This relation determines the range of application of the nozzle.
4. The ducted propeller system with a nozzle ring which is partly or completely
fitted to the impeller blades has a relatively low efficiency due to the additional
drag of the rotating nozzle ring. The system has, however, a reasonable thrust-
horsepower ratio at bollard pull condition. For some types of ships (coasters,
trawlers) the application of a ring propeller has attractive features.
5. The application of a wake adapted or non-axisymmetrical nozzle behind a full
single screw ship (tanker, bulkcarrier) offers a means of making the inflow velocity
of the propeller more constant over the screw disk. Consequently, this nozzle
should minimize problems concerning propeller induced vibrations and cavitation.
In addition, the applications of such a nozzle leads to a reduction in DHP.
6. In the case of twin-screw ships, a non-axisymmetrical nozzle offers a means of
minimizing the actual effective incidence changes of a blade section during a
revolution. Consequently, the non-axisymmetrical nozzle improves the cavitation
properties of the screw. This is of particular importance for fast naval ships.
7. For the design of ducted propeller systems it is attractive to have a calculation
method available supported by systematic experiments. Tests on ducted propellers
were scarce. The results of the systematic tests presented here may help to fill up
the lack of experimental data on ducted propellers.
Appendix
THEORETICAL ANALYSIS O F DUCTED PROPELLERS

The calculations of the ducted propellers were based on the following assumptions.
The forward velocity was assumed to be sufficiently large, the nozzle loading suffi-
ciently low to permit the application of the linearized theory. The effect of the hub
shape on the flow field was neglected. The ducted propeller system considered consists
of an annular airfoil of finite length and of an impeller having an infinite number of
blades.
The mathematical model of the ducted propeller configuration can be composed
by means of vortex distributions and sink and source distributions. The impeller was
regarded as a uniformly loaded actuator disk set normal to the free stream. It was
driven to rotate with an angular velocity a. Free trailing vortices started from the
propeller disk at hub and tip radius. The flow around the nozzle was represented by a
sinusoidal bound ring vortex distribution with zero strength at the leading and trailing
edges and by a ring vortex distribution with zero strength at the leading edge and
equal to the strength of the circumferential component of the helical trailing vortices
at the irnpeller disk.
The resulting mathematical model is summarized in Fig. 8.

Assume the vortex strength per unit impeller disk area at the impeller tip to be
equal to y (R,), the strength of the different vortex distributions then becomes:
- vortex strength per unit impeller disk area:

- strength per unit area of the trailing vortex sheet starting from the impeller disk at
the tip radius :

- strength per unit area of the trailing vortex sheet starting from the impeller disk at
the hub radius :

- strength per unit area of the bound ring vortex distribution along the nozzle :

sin 8
PY (R,) -- (for - a L < X <0)
2J a (l - a)
and

Y1 (R,) sin 0 (for - a L < X < ( l - a)L)

where,

p = @R1
-; R0
jZ = -; X + (a - 0.5) L = 0.5 Lcos 8
VA R1

and the definitions of R,, R,, r, X, L, a and so on are given in Fig. 8.

The thickness effect of the nozzle profile on the flow field can be taken into account
in the linearized theory by representing the nozzle by a distribution of ring sources
and sinks along a cylinder with radius R, (see Fig. 8).
In the case of very thin nozzle profiles, as considered here, the strength of the sink
and source distributions representing the nozzle profile can be calculated by consider-
ing the flow around the profile as two-dimensional.
Then the local source strength is equal to the derivative of the profile thickness:

where h (c) denotes the local thickness of the nozzle profile.

2. CALCULATIONS
OF THE FLOW FIELD

The total induced velocities can be calculated according to the law of Biot-Savart
if the main dimensions of the ducted propeller system (Ro, R,, L and a), the thickness
distribution of the nozzle (h (c)), the loading of duct and impeller (y (R,)) and
(yl (R1)), the rotational velocity of the impeller (U) and the undisturbed stream
velocity (V,) are given. In the following, we will give the induced velocities in the
point ( X , r) of the flow field due to the various vortex distributions and the source and
sink distributions.
The relations are made non-dimensional in the following way:

y = - ?(R,)
- - ~l(R1) g (C') h (43
Y1 =- S <F) = - h (C) =-
v, VA VA R1
where a, t and r denote the axial, tangential and radial induced velocities respectively.
(1) Induced velocities due to the vortices representing the impeller disk

W:" (X,?) = 0
W,'"(K,?) = T A , , , (X,?)

A,,, ( X , i ) = A / d i , /
n/ 2

d$
2 cos 2 + --
X
L 0
[x2 + ( F , + F)' - 4Fl? sin2 $1 3/2

(2) Induced velocities due to the trailing vortex sheet starting from the impeller disk
at the tip radius

W i 2 )(X,?) = p7A2 ,l (X,F)

W:" ((a,?)= 'yA2,, ( X , i )
C2'(g?)= p?A2,3 (P,?)
a, nI2

A,,, (i.3 = ! l d r j d $ +
p2 (1 ? COS 2 $)
7G
0 0
[{pX - 2)' + p2 {(l +
- 4F sin2 $)l3"

AZs2(X,?) = - ! / d t J d$ [- -
p (F + cos 2 $)
7G ----------- -1312
0 0

p [p3 - t ]cos 2 $
(g?)= - - ! I d r J d$
ft 1- - - - - - - - - - - - - - 1312
0 0

(3) Induced velocities due to the trailing vortex sheet starting from the impeller disk
at the hub radius :

F3'(P,?) = ~ 9 ~(P,?) 3 , ~
Wi3' ( 2 , ~=) jiA3,2 (X,?)
W j 3 )(Z,?) = pjiA3,3 (X,?)
W n12

A ~( 2, , ~~=
) -I j d r J d$
+
~p' (A F cos 2 $)
X
0 0
[{PP - r}' + +
p' {(l ?)' - 4Ai sin2
 CO r/2

A3,2 (X,?) = - l J d r / d$
(F
/l2 + 1 COS 2 $)
n C- - - - - - - --- - -- -1312
0 0

m n/2

A ( X ) = 1JdrJ
7r
d$ c - - - - - - -ap- -[pX- - -- -z]- COS 2 l,b 1312

0 0

(4) Induced velocities due to the discontinuous ring vortex distributions along the
nozzle :

(X,?) = 0
-
w,'~'(X,?) = p7A4,3 (X,?)

A4,1 (XY?)= -
(1+ i cos 2 $) sln 8
[(P - + (1 + - 4 i sin2 $l311

(5) Induced velocities due to the sinusoidal ring vortex distribution along the nozzle:
-( 5 ) - - .>, .
W,, (x,r) = Y I A S , l(Z,?) .:;+,.

A5,l (P,?) = ' J d t J d l ( ,

+ i cos 2 $) sill 0
-
(1
71 [ ( P - o2+ ( 1 + - 4 i sin2 ~ l ] ~ ~ ~
-09 0

- r) cos 2 + sin 8
-
(P
(X,?) = -- C- - - - - - - - - - -1312
n
-09 0
(6) Induced velocities due to the ring sources and sinks along the nozzle representing
the thickness of the nozzle:

Wi6' (X,?) = 0
+ cos 2 $)
V
'J J
(X,?) = -
n
-av
d;S
0
d$ g (t) (F -

The total induced velocities may be written as:

w a (E,?) = pYA1 + ylAz ( f , r ) + A,

(V) (X,?)
A, (X,?) = A2,, (E,?) + A,,, (X,?) + A4,, (X,?)
A2 (X,?) A s , , (X,?)
;)=

As (g,?) A6,, (X,?)

;)=

Vt (X,?) = ?A3 (X,?)

A3 (X,?) = A,,, (P,?) + A,,, (X,?) + A3,2 (X,?)
Wr (273 = ~ 7 A (X,?)
4 + +,A5 (X,?) + A7 (X,?)
A4 (E,?) = A2,3 (X,?) + A3,3 (X,?) + A4,3 (X,?)
A, (X,?) = A s , , (X,?)
A7 (X,?) = A6 $ 3 (X,?)

where the coefficients A are functions of X, F, the main dimensions of the ducted
propeller system (thus R,, R,, L and a) and the thickness distributions (h?(c)) of the
nozzle only. These coefficients are independent of the vortex densities y (R,) and
y , (R,) and the advance coefficient p.

3. CALCULATION
OF THE CAMBERLINE OF THE NOZZLE PROFILE

The shape of the camberline of the nozzle profile can be obtained from the relation:

where the definition of s (X) is given in Fig. 8.

4. PRESSURE
DISTRIBUTION ALONG THE NOZZLE

With the aid of Bernoulli's theorem the pressure distribution along the nozzle can
be calculated. The static pressure coefficients Cp+(x) and C,-@) are defined by:

P , and V , are the static pressure and the velocity of the undisturbed flow, P t ( x )
and P-(X)are the static pressure at respectively the inner and outer side of the nozzle
as a function of the location.
Calculation of the pressure distribution along the nozzle gives:

R , sin 0
c; (4= ~ j i + R, - 2-4, (%l)] +
ji, [sin 0 - 2A2 ( K , l)] - 2As (*,l)
R , sin 0
C , (3) = p? - R2 - 2A1 (*,l)] +
-
yl [- sin 0 - 2A2 (%l)]- 2A6 (%,l)
where,

R , = 1 R2 = 0 for .-art < E < 0

R1=OR2=1 for 0 < E < (1 - a ) ~

5. CALCULATION
OF THRUST, TORQUE AND EFFICIENCY

The non-dimensional thrust and torque coefficients are defined by

where T, Tp,TNand Q denote the total thrust, the impeller thrust, the nozzle thrust
and the torque respectively.
The thrust and torque coefficients become:

( 1 -014
CTN= 7; [- 4 / A (X , 1) sin 8 d g ]
-04
+
o . ( 1 -014
2

-art
(l-a)rt (l-a)#

+ {- 4 1 A, ( X , 1) sin 0 dF +8
J A2 (X,l)-
- dX

The efficiency of the ducted propeller system is defined by

 The shape of the camberline of the nozzle profile s(x), the pressure distribution
along the nozzle Cp+(x) and C,-(X), the thrust and torque coefficients CTP, CTN,
C, and CQand the ideal efficiency q, of the ducted propeller system were given in the
reIations (1) through (8). From these relations it can be seen that s(x), Cp+(x),C,-(X),
CTNand pco are completely determined by:

- The main dimensions of the ducted propeller system given by R,, R,, L and a,
and thickness distribution h(4) of the nozzle profile.
,
the vortex density y ,
the product of the vortex density y and the advance ratio p.
The thrust coefficient CTpand consequently the total thrust coefficient C, and the
ideal efficiency q, depend not only on the product pji but also on 7.
The impeller is represented by an actuator disk rotating with an injinite angular
velocity if the undisturbed stream velocity V,, and the product p? are kept constant
and the advance coefficient p becomes infinite. Then the vortex density ji goes to
zero. Besides, the tangentially induced velocities and consequently the losses due to the
rotation of the fluid become zero. The case that the actuator disk rotates with an in-
finite angular velocity coincides with the case that the actuator disk or the impeller
rotates with finite angular velocity while a stator is used to eliminate the rotational
losses. The total thrust coefficient C,, the thrust ratio r and the ideal efficiency qi
in that case are denoted by C,,, to and qio.
The shape of the camberline of the nozzle profile is also completely determined by
the values of C,,, zO the main dimensions of the ducted propeller system and the
thickness distribution of the nozzle profile. Calculations of the thrust coefficient, CT
the thrust ratio z and the ideal efficiency q, at various advance coefficients p were
made for a number of nozzle shapes given by CTO,z0 the main dimensions of the duct-
ed propeller system, A, q, a and the thickness distribution of the nozzle profile. In
addition, the shapes of the camberlines of the nozzle profiles and the pressure distri-
butions along the nozzle surfaces were calculated. The data used for the computations
and the results are given in the tables I, I1 and 111 respectively.
Table I presents the thrust coefficients C,,, the thrust ratio zo the main dimen-
sions of the ducted propeller system and the thickness ratio of the nozzle profile for
which computations were carried out. All the considered nozzle shapes had a NACA
four-digit basic thickness form.
Table 11 gives the effect of the rotational velocity of the screw on the ideal effi-
ciencyq,, the thrust coefficient C, and the thrust ratio z of the ducted propeller systems.
Finally, Table 111 presents the mean lines and the pressure distributions along the
nozzles of the various ducted propellers. In addition, the ideal efficiency v, for
J = 0 (vto)is given in this table.
 I. Thrust coefficient CTO,thrust ratio
TABLE and main dimensions of nozzles for which compu-
tations were carried out.
nozzle CTO 70 LID alL SlL dlD
number
 11. Effect of the advance coefficibnt J on the thrust coefficient CT and the thrust ratio 7 of the
TABLE
considered nozzle shapes
Nozzles Nr. A Nozzles Nr. B Nozzles Nr. C
J CT/CTO 7/70 CT/~TO 7/70 CT/~TO 7/70

Nozzles Nr. D Nozzles Nr. E Nozzles Nr. F

CTICTO 7/70 CT/CTO 7/70 CTICTO 7/70
 111. Mean lines, pressure distributions along the nozzle profile and ideal efficiency of the
TABLE
considered ducted propeller systems
Nozzle Tio x/R s(x)/R CP' ( X ) CP- ( X )
number
TABLE
111 (continued)

Nozzle rlio x/R sW/R CP (X)

+ CP-(X)
number
 I l l (continued
TABLE
Nozzle 90 x/R s(x)/R CP' ( 4 CP- ( X )
number
TABLE
111 (continued)
- -

Nozzle Tie X/& s(x)/R CP ( X )

+ CP- ( X )
number
TABLE
111 (continued)

Nozzle ?ro x/R s(x)/R '

CP (X) CP-(X)
number
TABLE
111 (continued)
Nozzle Y)IO x/R &)/R CP' ( X ) CP-( X )
number
 111 (continued)
TABLE
Nozzle ?ro x/R s(x)/R
number
TABLE
111 (continued)
- -

Nozzle 1)lo x/R s(x)lR CP'(X) CP- (-4

number
 111(continued)
TABLE
Nozzle 710 x/R s(x)lR CP'(X) CP-(X)
number
 SUMMARY

For the application behind ships two different types of ducted propellers are consider-
ed; the ducted propeller where the nozzle accelerates the flow at the propeller and the
ducted propeller where the nozzle decelerates the flow at the propeller. The first type
of ducted propeller is now extensively used in cases where the ship screw is heavily
loaded or where the screw is restricted in diameter. The accelerating nozzle offers a
means of increasing the efficiency of heavily loaded propellers. The second type of
ducted propeller is used to increase the static pressure at the impeller. This nozzle may
be used if retardation of propeller cavitation is desired.

For the design of a ducted propeller it is attractive to have a theoretical calculation

method available supported by experiments. Tests on ducted propellers are scarce and
most of the tests are restricted to isolated applications. The work described here deals
with systematic experiments on ducted propellers and may contribute to fill up the
lack of experimental data.

The potentialities of ducted propellers both with nozzles of the accelerating and the
decelerating flow type are discussed based on simplified theories.
The open-water test results of systematic series of flow accelerating and flow dece-
lerating nozzles are given. These data are of importance for the design of optimum
ducted propeller systems as well from the viewpoint of efficiency as from the view-
point of cavitation.
In addition, the results of open-water tests with ringpropellers and ringpropellers in
nozzles are given.

Finally, the design and the application of non-axisymmetrical nozzles which are
adapted to the wake behind the ship, are discussed. In the case of single-screw ships
(tankers and bulkcarriers) the nozzle is designed in such a way that the flow at the
propeller plane becomes more uniform. Such a nozzle offers, next to a reduction in
DHP, a means of minimizing propeller induced vibration and cavitation problems.
In the case of twin-screw ships (fast naval ships; frigates, destroyers), the propellers
operate in a varying inflow, due to the shaft inclination. The non-axisymmetrical
nozzle is designed here, from the viewpoint of retardation of screw cavitation, in such
a way that the actual effective incidence changes of the blade sections of the impeller
will be as low as possible during a revolution.
 SAMENVATTING

Voor toepassing bij schepen komen twee typen schroef-straalbuissystemen in aan-

merking; het type waarbij de straalbuis de stroming ter plaatse van de schroef versnelt
en het type waarbij de stroomsnelheid ter plaatse van de schroef vertraagd wordt.
Schroef-straalbuissystemen van het eerste type worden veelvuldig gebruikt in ge-
vallen waarbij de schroef zwaar belast enlof beperkt in diameter is. De versnellende
straalbuis verhoogt het rendement van zwaar belaste schroeven.
Schroef-straalbuissystemen van het tweede type worden gebruikt wanneer het ge-
wenst is de druk ter plaatse van de schroef te verhogen. Door de verhoging van de
druk kan schroefcavitatie verminderd of voorkomen worden.

Het biedt voordelen, gezien het grote aantal variabelen, om het ontwerp van een
schroef-straalbuissysteem te baseren op een combinatie van theorie en experiment.
Weinig resultaten van proefnemingen met schroef-straalbuissystemen zijn beschik-
baar en bovendien betreffen deze proefnemingen vaak slechts een incidentele toepas-
sing. De resultaten van de proefnemingen met schroef-straalbuissystemen gegeven in
dit proefschrift dragen bij om het gebrek aan experimentele gegevens aan te vullen.

De eigenschappen van schroef-straalbuissystemen, zowel met versnellende als ver-

tragende straalbuizen, worden, gebaseerd op eenvoudige theorieen, afgeleid.
De resultaten van proefnemingen met systematische series versnellende en vertra-
gende straalbuizen worden gegeven. Deze gegevens zijn van belang voor het ontwer-
pen van optimale schroef-straalbuizen uit oogpunt van rendement of cavitatie. Daar-
naast worden de resultaten van vrijvarende proeven met ringpropellers en met ring-
propellers in straalbuizen gegeven.

Tenslotte wordt het ontwerp en de toepassing van asymmetrische straalbuizen die

aangepast zijn aan de volgstroom achter schepen besproken. Voor volle enkelschroef-
schepen (tankers, bulkcarriers) wordt de straalbuis zodanig ontworpen dat het snel-
heidsveld terplaatsevande schroef gelijkmatiger wordt. Een dergelijke straalbuis biedt,
naast een vermogensbesparing, ook de mogelijkheid om het gevaar voor trillingshinder
(opgewekt door de schroef) en voor erosie van de schroefbladen door cavitatie te ver-
minderen.
Bij dubbelschroefschepen (en in het bijzonder snelle Marine-schepen, zoals jagers
en fregatten) werken de schroeven in een niet gelijkmatige aanstroming door de
schuine asliggingen. De asymmetrische straalbuizen worden hier zodanig ontworpen
dat de invalshoekvariaties van een blad gedurende een omwenteling zo klein mogelijk
zijn. Het gevaar voor het optreden van schroefcavitatie kan hierdoor verminderd
worden.
 REFERENCES

l. STIPA,L.; 'Experiments with Intubed Propellers'. L'Aerotechnica pp. 923-953, August, 1931.
Translated by Dwight M. Miner, NACA. NACA TM 655, January, 1932.
2. KORT,L.; 'Der neue Diisenschrauben-Antrieb', Werft-Reederei-Hafen, Jahrgang 15, Weft 4,
February 15,1934.
3. SACKS, A. H. and J. A. BURNELL; 'Ducted propellers - A critical review of the State of the Art'.
Progress in Aeronautical Sciences, Vol. 2, New York 1962.
4. WEISSINGER, J. and D. MAASS;'Theory of the ducted propeller. A review'. 7th Symp. on Naval
Hydrodynamics, Rome, August 1968.
5. HORN,F. and H. AMTSBERG' 'Entwurf von Schiffdiisen systemen' (Kortdlisen), Jahrbuch der
Schifiautechnischen Gesellschaft, vol. 44, 1950.
6. K~~CHEMANN, D. and J. WEBER; 'Aerodynamics of Propulsion', MC. Graw-Hill Book Co., Inc.,
New York, 1953.
7. DICKMANN, H. E. and J. WEISSINGER; 'Beitrag zur Theorie optimaler Dusenschrauben', (Kort-
diisen), Jahrbuch der Schiffbautechnischen Gesellschaft, vol. 49, 1955.
8. MORGAN, W. B. and E. B. CASTER; 'Comparison of theory and experiment on ducted propellers',
7th Symp. on Naval Hydrodynamics, Rome, August, 1968.
9. VANMANEN,J. D.; &Open-watertest series with propellers in nozzles'. Int. Shipb. Progress,
vol. 1, no. 3, 1954.
10. VANMANBN,J. D.; 'Recent research on propellers in nozzles*, Int. Shipb. Progress, vol. 4.
no. 36,1957.
11. VANMANEN,J. D. and .A. SUPERINA; 'The design of screw propellers in nozzles', Int. Shipb,
Progress, vol. 6, no. 55, 1959.
12. VANMANEN, J. D.; 'Effect of radial load distribution on the performance of shrouded propellers',
Trans. RINA, 1962.
13. VANMANEN,J. D. and M. W. C. OOSTERVELD; 'Analysis of ducted propeller design', Trans.
SNAME, 1966.
14. WTERVELD, M. W. C.; 'Model tests with decelerating nozzles*, ASME, Symp. on Pumping
Machinery for Marine Propulsion, Philadelphia, May 1968.
15. HOEXNER, S. F.; 'Fluid Dynamic Drag', published by the Author 1968.
16. KELLER, 3. AUP'M;'Enige aspecten bij het ontwerpen van scheepsschroeven', Schip en Werf,
(December 1966).
17. 'Principles of Naval Architecture' Published by the SNAME, New York (1967). .
18. VANLAMMEREN, W. P. A.; J. D. VANMANEN and M. W. C. WTERVELD; 'The Wageningen
B-screw series'. Trans. SNAME, 1969.
19. M m o m , I. YA.; 'Investigation of hydrodynamic characteristics of screw propellers under
conditions of reversing and calculation methods for backing of ship'. Buships Translation 697,
1960.
20. BAKER, D. W. and C. L. P A ~ R S O'Representation
N; of propeller thrust and torque character-
istics for simulations' NSRDC,Technical Note, MEL 202169, April 1968.
21. VANMANEN, J. D.; 'The choice of the propeller'. Marine Technology, April 1966.
22. VANMANEN, J. D. and J. KAMPs; 'The effect of the shape of afterbody on propulsion'. Trans.
SNAME, 1959.
23. VANMANEN, 3. D.; M. W. C. OOSTERVELDand J. H. W r r r ~ ;'Research on themanoeuvrability
and propulsion of very large tankers'. 6th Symp. on Naval Hydrodynamics, Washington, 1966.
24. OOSTERVELD, M. W. C.; 'The application of non-cylindrical nozzles for large tankers and bulk
carriers'. Shipbuilding and Shipping Record, November 1968.
25. WTERVELD, M. W. C.; 'Ducted Propellers'. Seminar of the Royal Institute of Engineers, Delft,
The Netherlands, May 1969.
26. LINDGREN, H.; C. A. JOHNSON and G. DYNE;'Studies of the application of ducted and contra-
rotating propellers on merchant ships'. 7th Symp. on Naval Hydrodynamics, Rome, August 1968.
27. WERELDSMA, R; 'Some aspects of the research into propeller induced vibrations'. Int. Shipb.
Progress, vol. 14, no. 154, June 1967.
28. VANGUNSTEREN, L. A.; 'Ringpropellers', Paper read before SNAME and Institute of Marine
Engineers, Canada, September 1969.
29. ,KELLER,W. H. AUF'M; 'Comparative tests with B-series screws and ringpropellers', NSMB-
report no. 66-047, DWT,1966.
30. SPARENBERG, J. A.; 'On optimum propellers with a duct of finite length'. Mathematisch Instituut
Universiteit Groningen, report TW-54, 1968.
= axial distance between leading edge of nozzle and impeller plane.
= expanded blade area of screw
= exit area of nozzle
= disk area of screw
= loading coefficieut, Bp = 33.07 K ~ . J ~-''~ ' ~

= drag coefficient of nozzle, CD, = DN

t Pv; ~ D L
= skin-friction drag coefficient
= lift coefficient

= pressure coefticient, C, =
P - P,
t P v:
1
= thrust coefficient, CT =
f p V: nj4

= torque coefficient, CQ = Q
$ p V: 7114 0 2 . D

1
= thrust coefficient, C,* =
Q p [V: + (0.7 n n ~ ) ' ]n/4 D'
= torque coefficient, C; = Q
+ p [V: + (0.7 n n ~ )n/4
~ ] D
= propeller diameter
= drag of nozzle
= kinetic energy loss in propeller slipstream
= thickness distribution of nozzle
VA
= advance coefficient, J = -
nD

T
= thrust coefficient, K T = ---
pn2~4

Q
= torque coefficient, K Q = ---
pn2~'
= nozzle length
= number of revolutions per second and per minute
P = pressure
p, = static pressure of undisturbed stream
P = power
Q = torque
r = radius
RI = screw radius
Ro = hub radius of screw
T = thrust
U1, U, = increment of axial velocity at impeller and infinite far downstream of
impeller
Up,U, = mean axial velocity at impeller plane induced by impeller and nozzle
VA = undisturbed stream velocity
VP = mean axial velocity at impeller plane
Z = number of screw blades
A,/A, = blade area ratio of screw
A,/AEx = ratio between impeller disk area and exit area of nozzle
f/L = camber ratio of nozzle
LID = nozzle length-diameter ratio
P / D = pitch ratio of screw
S/L = maximum thickness ratio of nozzle
Mi = angle between nose-tail line of nozzle profile and propeller shaft

= advance angle at 0.7 R, /l= arctan VA

0.7 nnD
= vortex density
= vortex strength
101.27
= diameter coefficient or speed ratio, b = --
J
'li = ideal efficiency
q f ~ = efficiency factor due to nozzle drag
= open-water efficiency, v, = -
J KT
-
2n KQ
IZ = hub-diameter ratio of propeller
P = specific mass of water

= TP
impeller thrust-total thrust ratio, z = -
T
= kinematic viscosity of water
n wR1
= advancecoefficient, p = - = -
J VA
= angular velodity
To the board of Directors of the Netherlands Ship Model Basin I am grateful for the
permission to publish this research in the form of a thesis.
A special word of thanks must be given to the following institutes and companies
for their faith and financial support in different parts of this work and for their kind
permission to publish the results:
Office of Naval Research, U.S. Navy
Contract No. N 62558-3999
Contract No. N 62558-4555
Royal Netherlands Navy
Mobil Oil Corporation, New York
Lips N.V., Drunen
Esso International, New York.
Appreciation is also expressed to many members of the Special Projects Depart-
ment of the N.S.M.B. whose valuable assistance has contributed a great deal to the
completion of this study. Especially the readily available help of Mr. H. Nijding and
Mr. A. J. P. Huybens is thankfully acknowledged.