CAVITATION TUNNEL TESTS ON DUCTED PROPELLERS

1 2 3 4 5
Daniele BERTETTA , Chiara BERTOGLIO , Francesco CONTI , Cesare M. RIZZO , Michele VIVIANI

1-2-4-5
DITEN – Naval Architecture and Marine Engineering Group - University of Genoa – Genova – Italy
3
Fincantieri C.N.I. – Naval Vessel Business Unit – Genova – Italy

daniele.bertetta@libero.it, chiara.bertoglio@hotmail.it, francesco.conti@fincantieri.it,

cesare.rizzo@unige.it, michele.viviani@unige.it

ABSTRACT
Propeller design requirements are nowadays more and more stringent, demanding not only to provide high
efficiency and to avoid cavitation, but including also requirements in terms of low induced vibrations and
radiated noise. Ducted propellers may provide the opportunity to reduce these negative effects or to
increase the efficiency of heavily loaded propellers.
This paper presents the development of a measuring device for the cavitation tunnel, in order to evaluate
the forces on the duct. The measurement tool was designed and realized in house and it is at the same
time the support of the duct to be placed in the cavitation tunnel and the device measuring the acting
forces.
By means of the developed measuring device, two ducted propellers with different characteristics (namely
with accelerating and decelerating duct) have been tested at cavitation tunnel. The results of these two
test series were compared with towing tank open water test data, showing a good agreement.
Moreover, cavitation tests were also carried out, focusing the attention on the phenomena on the duct and
on other interesting characteristics, such as thrust breakdown, providing useful data for future validations
of numerical propeller design and analysis tool.

1. INTRODUCTION

Propellers performance improvement holds a great interest in modern marine research. In

particular, ducted propellers, are regarded as a very interesting solution for marine propulsion
from various points of view, as reported in the pioneering works of the 60s and 70s (e.g.
Oosterveld 1970) and in more recent applications (e.g. Baltazar 2011). Depending on the duct
shape, it is possible to increase the propeller efficiency or to improve its cavitating behaviour
reducing also side effects such as radiated noise and pressure pulses.
The Kort nozzle is known as the first annular airfoil considered for a ducted propeller design and
has proven to be an effective tool to increase efficiency of heavily loaded propellers and to
improve propeller bollard pull behaviour. It belongs to the so called “accelerating nozzles”
because of its tendency to increase flow speed at the propeller disc; in this functioning
condition, the pressure distribution induced by the propeller on the duct surface results in a
positive thrust of the nozzle, whose effect is to increase the propeller efficiency at low values of
advance coefficient; for this reason, many vessels such as tugs, trawlers, push boat and tankers
are fitted with this kind of propellers.
Ducted propellers with decelerating type of nozzle present a wide range of advance coefficient
in which flow is decelerated. As a consequence, their efficiency is lower, and they are used for
different purposes; main advantage is the static pressure increase due to flow deceleration,
which results in a postponed cavitation inception. Cavitation avoidance has gained more and
more interest because of its consequences, such as blade erosion, radiated noise and pressure
pulses, which have always been very important subjects for warships, and nowadays represent
an important topic also for passenger ships in order to improve comfort onboard and in general
for all merchant ships for environmental reasons.
From the point of view of ducted propeller design, it is very difficult to model the interaction
between the nozzle and the propeller, even if many numerical codes have been developed over
the last decades. Lifting line and lifting surface methods are widely used for this purpose,
however, results obtained with such methods are not always sufficiently accurate in evaluating
mutual influence between propeller and duct, especially for higher design advance coefficients
at which the duct decelerates the flow. As a consequence, a calibration process on the basis of
a BEM or RANSE analysis is often necessary.
From this point of view, validation of numerical codes is still needed, since for this kind of
propellers available experimental data are rather limited compared to conventional propellers,
especially when considering cavitation behavior and related effects (thrust breakdown, radiated
noise, pressure pulses). Moreover, this kind of propellers are more subject to scale effects,
which have still to be further analysed.
In order to investigate ducted propeller functioning, and in particular their cavitating behavior
and the related side effects, a series of model tests have been planned. In the present paper,
the design of a device for the measurement of forces acting on the propeller duct during the
tests in the cavitation tunnel will be described, showing that good results were obtained in terms
of ducted propeller mechanical characteristics.
Two ducted propellers (with accelerating and decelerating ducts, later on referred to as DP1
and DP2), whose open water characteristics were measured at towing tank, have been tested in
order to check the measuring device functioning. Successively, an extensive experimental
campaign, in terms of cavitation observation, thrust breakdown of both propeller and duct has
been carried out at the cavitation tunnel of the University of Genova and results are reported.

2. DESCRIPTION OF EXPERIMENTAL SETUP

2.1 Cavitation Tunnel

The experiments were carried out at the Cavitation Tunnel facility of the University of Genoa
(DITEN), represented in Figure 1, in which conventional propeller cavitating behaviour (e.g.
Savio et al.2009) and SPP propellers characteristics (e.g. Ferrando et al.2007) are usually
tested.
The facility is a Kempf & Remmers closed water circuit tunnel with a squared testing section of
0.57 m x 0.57 m, having a total length of 2 m. Optical access to the testing section is possible
through large windows.

Figure 1. University of Genova Cavitation Tunnel

The nozzle contraction ratio is 4.6:1, and the maximum flow speed in the testing section is 8.5
m/s. Vertical distance between horizontal ducts is 4.54 m, while horizontal distance between
vertical ducts is 8.15 m. Flow speed in the testing section is measured by means of a differential
venturimeter with two pressure plugs immediately upstream and downstream of the converging
part. A depressurization system allows obtaining an atmospheric pressure in the circuit near to
vacuum, in order to simulate the correct cavitation index for propellers and profiles (2D and 3D).
The tunnel is equipped with a Kempf & Remmers H39 dynamometer, which measures propeller
thrust, torque and RPM. As usual, a mobile stroboscopic system allows to visualize cavitation
phenomena on the propeller blades. Moreover, cavitation phenomena visualization in the
testing section is also made with two Allied Vision Tech Marlin F145B2 Firewire Cameras, with a
resolution of 1392 x 1040 pixels and a frame rate up to 10 fps.

2.2 Measuring device

The decision of developing a special purpose self made measuring device for the measurement
of forces acting on the duct was aimed at obtaining an instrumentation specifically suited for the
cavitation tunnel features as well as to fully exploit the uncertainties of the measurements. In
fact, design of the instrumentation is fully known in all parts. Moreover, the measurement
system was developed also in view of the testing of conventional foils and it can be further
improved and adapted for different testing purposes, since the necessary design and
manufacturing skills were gained during this project.

Figure 2. Cavitation Tunnel test section set-up (left: top view, right: side view)

Instead of adopting a commercial set up configuration, a cavitation tunnel window was modified
to hold an aluminum alloy plate coupled by welding to an aluminum alloy hollow bar, as shown
in Figure 2 (Bertoglio 2011).
The hollow bar is at the same time the support of the duct as well as the measuring device.
Several bar section configurations with different internal and external shapes and sizes were
considered and preliminary calculations were carried out for various steel, composites and
aluminum alloy bars. The material and the bar section was finally conceived having the
following main characteristics:

1. Maximum acting stress values suitably lower than yielding of the material at the maximum
estimated duct thrust (500 N approx.),
2. Adequate fatigue strength, taking into account the fluctuating load behavior during service
induced by both the propeller and other hydrodynamic phenomena on the duct,
3. Limited deformations, i.e. elongation, deflections and rotations of the cross section, in order
to avoid interference between the duct and the rotating propeller during the tests,
4. Adequate device sensitivity, possibly lower than 0.2% of the nominal load capacity

By considering the structural model of the measuring bar, the strain field was studied to select
the most appropriate locations of strain gauges. The fundamentals of the model that was taken
into account for the preliminary analysis is shown in Figure 3.
As it can be noted, the duct force was considered applied in the center of the duct, even if this is
not exactly the case in reality. This assumption highly simplify the design of the measurement
device. Also, the duct itself was considered perfectly rigid, only transferring the forces to its
support through the flange.

Z Z

Rigid duct

Figure 3. Sketch of the structural model for the measuring device

Twenty-four strain gauges were located on the internal surface of the hollow section bar, hence
there is no need to protect them, being in a non-wetted area. Namely, the strain gauges were
bonded in way of the quadrants of the hollow section bar, i.e. at intersections with y-axis and z-
axis in order to measure the bending stresses generated by the force components along the
flow and by the ones perpendicular to it, accommodating them as far as possible in the rather
limited space inside the hollow bar. Their number and electrical connections were defined in
order to maximize the sensitivity of the instrumentation and to minimize the effect of
disturbances like temperature variations, axial forces along the bar, connecting wires
resistance, etc. In particular, the acting forces are measured by two half bridges configurations
(Hoffman 1989), one measuring the thrust on the duct along the flow direction (y-axis in Figure
3) and the other measuring the lateral force (z-axis in Figure 3).
An in house built signal conditioning system was used to supply the bridges, to acquire, filter
and amplify the signals: indeed, this conditioning unit is a rather conventional one, having a
stabilized power supply of 10V DC ±1mV, a gain of 100, and a 10kHz low pass filter.
As the measurement system behaves in essence as cantilever beam, the maximum stress and
strains occur in way of the plate where it is fixed. However, in order to limit the boundary effects
on the measurements, strain gauges were bonded suitably far from the fixity. Though,
accounting for the rather small size of the hollow section of the bar, impairing the bonding
procedure of gauges, and considering the need to maximize the instrument sensitivity, the
distance of the strain gauge from the plate was maintained of the same order of magnitude of
the diameter of the cross section of the bar. Basically, gauges were oriented along the bar axis
(x-axis in Figure 3) to measure bending strains, even if additional strain gauges were bonded as
well in the expected shear stress directions, in particular aiming at torque measurements.
As earlier mentioned, material and cross sections of the bar were chosen also to avoid any
interference of the duct with the rotating propeller when the bar bends and deforms; as a matter
of facts the clearance is extremely small (0.6 mm for DP1 and 1 mm for DP2), because of the
model scale. The maximum deflection and cross section rotation of the duct support was
quantified and the design of the device was found satisfactory. It is worth noting that the effect
of the window deflection when the internal pressure of the tunnel is decreased due to vacuum
necessary for the cavitation test conditions should be accounted for to avoid the contact
between the duct and the propeller due to vertical displacement (i.e. along x-axis in Figure 3).
Truth to tell, fatigue limit was slightly overcome, but this was accepted taking into account that
fatigue strength check was carried out considering the maximum nominal acting force (500 N),
for which a certain safety margin had already been taken.
A direct calibration of the measurement system was carried out by means of calibration weights
that were used to apply known forces on the device, both in the axial and lateral direction
separately as well as in oblique directions (  45°). The sensitivity was found within the required
specifications, i.e. about 0.2%. The calibration procedure was repeated several times with
satisfactory results and statistical analyses of measurements were carried out assessing the
accuracy of the instrumentation.
Thanks to this wide analysis, the cross talk between the two bridges (for lateral and axial forces
measurement) was quantified and corrections were considered during the tests, even if this
effect appears rather limited.
Moreover, hysteresis of the device was shown not to be a problem, even after severe loading
cycles (i.e. applying the maximum nominal load of about 500 N).
In order to avoid vortex shedding from the hollow bar in the tunnel flow, which can affect bar
integrity and decrease dramatically its fatigue life, a screening foil was adopted, as shown in
Figure 4. The foil shape was selected in order to postpone cavitation.
Even if the foil effect reduces considerably the device resistance, preliminary tests without the
ducted propeller at varying flow speed and tunnel internal pressure were carried out, aimed at
measuring the reference resistance to be subtracted from the total force.
The small clearance between the duct and the rotating propeller represented one of the most
challenging problems: deformations and vibrations of the measuring device induce gap
reductions, which were checked during the system device, but deflections of the tunnel window
due to inner vacuum, resulted in a duct vertical displacement larger than the expected one and
consequently in a further reduction of the clearance. Deflections larger than expected are
attributed to the effect of the gasket between the aluminium alloy plate and the tunnel window.
In addition, duct lateral vibration was noted during measurements, fully visible at high advance
coefficients of the tested propellers, especially for the accelerating duct.
The reason of these vibrations is probably due to vortex shedding experienced by the
accelerating duct, clearly visible during cavitation tests. High frequency acquisitions were
performed to better understand the phenomenon and data acquired were investigated in the
frequency domain by Fourier transform analysis. Different resonance peak frequencies were
detected and force intensity quantified for various advance coefficients and revolution rate of the
propeller, in order to have a full characterization of the phenomenon.

3. EXPERIMENTAL CAMPAIGN SETUP

After developing the measurement device described in previous section, a first experimental
campaign on two different ducted propellers (Figure 4), with an accelerating (DP1) and a
decelerating (DP2) duct, was carried out.
Main characteristics of the propellers are reported in table 1, where D is propeller diameter, AE
is expanded area, AO is disc area, P and C are pitch and chord values, Z is blade number. Duct
chord length was taken equal to half the propeller diameter in both cases.

Figure 4. Experimental setup (DP1 on the left, DP2 on the right)

 Table 1. Main characteristics DP1 ducted propeller

DP1 DP2
D [mm] 230 230
AE/AO 0.689 0.725
P/D (0.7R) 1.566 1.354
C/D (0.7R) 0.491 0.498
Z 4 4

For the two configurations, a rather comprehensive test campaign were carried out, including
usual open water characteristics, cavitation observations and thrust breakdown evaluation.
Cavitation tests were also performed employing stock brackets in order to reproduce the typical
wake of a twin screw ship, as shown in Figure 4.

4. MEASUREMENT RESULTS

4.1 Open Water Tests

As a first step, open water tests have been performed to evaluate the agreement between
cavitation tunnel results and SVA (Heinke 2011) towing tank measurements in order to check
the accuracy of the new developed experimental device, keeping in mind the different working
conditions (open water vs. closed tunnel). As an example, results for DP2 ducted propeller are
reported in Figure 5, in which values are given as a percentage of the design ones.

Figures 5 and 6 - Comparison between DP2 propeller characteristics at towing tank (SVA) and
cavitation tunnel (left: without shift, right: with shift)

As it can be noted, the usual shift in advance coefficient between towing tank and cavitation
tunnel curves presents in this case slightly larger values (about 5% with respect to usual 1-3%).
Reason for this may be the strong blockage effect due to the invasiveness of the measuring
device and of the duct itself into the tunnel test section; nevertheless further investigations are
planned in order to analyze this effect.
As usual practice, a shift factor in the advance coefficient has been evaluated for the total thrust
curve. By shifting measured values, a rather satisfactory agreement is obtained (see Figure 6),
thus indicating that the shifting procedure, which is a usual practice for conventional propellers,
is also applicable for ducted propellers. Thereafter, cavitation tests were performed keeping
constant KT values, thus implicitly overcoming the differences in advance coefficient.

4.2 Cavitating Tests

Cavitating behaviour of both propellers was deeply investigated, considering various cavitation
phenomena in terms of inception and extension at different functioning conditions, ranging from
rather low loadings to rather high loadings (±60% Ktot with respect to design point). For both
propellers, usual cavitation phenomena have been observed (sheet and bubble cavitation), and
they are not discussed in detail here for sake of shortness; on the contrary, attention is focused
on the typical phenomena of ducted propellers (tip leakage vortex and duct cavitation),
analyzing their dependence on propeller loading and differences between cavitating behavior of
accelerating and decelerating duct.
As it can be noted in Figure 7, instead of the classical tip vortex typical of a conventional
propeller, in this case a different vortex pattern was observed, connecting the tip of the blade
with the internal duct surface; this kind of vortex is well known in literature, and it is referred to
as “tip leakage vortex”.
Tip leakage vortex is due to the interference of the duct surface with the vortex structure;
recirculation of flow on the blade tip is partially prevented by the small gap and this leads the tip
vortex to assume a different configuration and a different behaviour. In general, for conventional
propellers, tip vortex inception value is strongly dependent on the advance coefficient (i.e. on
propeller loading) and it is replaced by vortex from sheet face for high advance values; in this
case (and for both propellers), it was observed that tip leakage vortex is present in the whole
range of advance coefficient analysed, with an almost constant inception cavitation number.

Figure 7. Kt tot=130%Kt tot_des σn=150% σn_des - DP1 propeller

Moreover, tip leakage vortex and vortex from sheet face have been jointly observed in certain
functioning conditions (see Figures 8 and 9), showing that recirculation due to back-face
differential pressure is not the sole cause of tip leakage vortex development.

Figures 8-9. Kt tot=70% Kt tot_des σn=150% σn_des - DP1 propeller

Considering figure 9, the different angle between propeller pitch and tip leakage vortex pitch is
clearly visible. Such difference, which may be due to the lower advance speed at duct due to
the boundary layer, is noticeable and it occurs with a similar trend for all the investigated
advance coefficients. Further investigations of this phenomenon are planned, aiming at its
reduction, since it appears to be the most significant cavitating phenomenon as well as a rather
important radiated noise source.
An important feature which was observed for both propellers is that, even if cavitation
phenomena inceptions were searched for in a wide range of thrust coefficient values (±60% with
respect to design point), no strong back sheet nor sheet face cavitation have been clearly
observed, with very limited extensions, sometimes at the tip, sometimes ad mid-span of the
blade. This is probably due to the fact that, for both propellers, at low advance coefficient the
duct has an accelerating effect on flow, while at higher values flow speed is reduced; then, a
change in the advance coefficient of the propulsor leads to a smaller change at the propeller
disc, leading to a “flattening” of the inception points of all phenomena and improving the
propeller off design behavior, especially regarding cavitating phenomena related to propeller
functioning with a certain angle of attack. The mentioned effect may also have an influence on
the tip leakage vortex behaviour previously analysed.
A strong difference in cavitating behaviour between the two propellers has been obviously
detected regarding duct cavitation patterns. The accelerating duct is employed for highly loaded
propellers, and so it has been designed to be shock free in a flow strongly convergent to the
propeller disc, with a significant radial component; this is evident if we consider the geometry of
the classical Kort nozzle adopted also for DP1 ducted propeller (see Figure 10).

Figure 10. Kort nozzle geometry

When considering lower loading functioning conditions for the DP1 ducted propeller, a negative
angle of attack results, due to the low contraction of the flow to the propeller disc; leading to
vortex shedding and “external duct bubbles” just aft the duct leading edge; this phenomenon,
visible in Figure 11, is present only in correspondence to low loadings of the propeller (30%
below design point), and increases its strength moving towards less loaded conditions .
In addition to this, internal bubble cavitation has been detected for the accelerating duct at the
axial coordinate in which the inner duct surface becomes almost cylindrical, immediately before
the propeller disc, as shown in Figure 12. This is probably the region of the inner surface of the
duct where pressure is lower due to the suction effect of the propeller blade, which leads to
early cavitation; this phenomenon is amplified by the rather high loading of present propeller.

Figure 11-12. Kt tot=70% Kt_tot_des σn=100% σn_des (left), Kt tot=130%Kt_tot_des

σn=100% σn_des (right) – DP1 propeller
In the case of the decelerating nozzle, no internal cavitation has been detected during tests,
consistently with the decelerating nature of the duct; on the contrary, for low σn and very low Kt
values, bubble cavitation has been observed on the external surface as it can be seen in Figure
13. It has to be remarked that this is a very off design condition; the phenomenon appears to
have a violent inception and a rapid diffusion along the duct profile chord, starting just before
half duct; a strong increase in propulsor radiated noise due to implosion of bubbles has been
noted during tests.

Figure 13. Kt tot=60%Kt_tot_des σn=80% σn_des – DP2 propeller

A phenomenon of great interest related to cavitation is the thrust breakdown that has been
characterized for the two ducted propellers at every working point of the cavitation bucket in
order to relate the phenomenon with cavitation inception. This analysis is not very common,
since tests at cavitation tunnel are in most cases carried out without measuring duct forces, with
the sole aim of visualising cavitation phenomena. The presently adopted testing configuration
allows to measure forces on the duct and thrust breakdown effects with cavitation, providing
useful information for propeller designer and for numerical codes validation purposes. In
particular, a clear definition of duct and propeller influence on thrust breakdown is obtained,
allowing to focus attention on most important phenomena. In Figure 14, an example is given
referring to the DP2 ducted propeller in a particularly off-design working point, with a rather low
propeller loading, for which the duct cavitation is well developed. A strong correspondence
between cavitation phenomena and thrust breakdown is present; in particular, it is interesting to
note the duct thrust falling at the inception of external duct bubbles between σn=90% σn_des and
σn=75% σn_des, evident sign of the violence of the previously mentioned cavitation
phenomenon.

5%

0%
0% 50% 100% 150% 200% 250% 300% 350% 400%

-5%

-10%

-15%

-20%

-25%

-30%
σn

Kt Kt_prop Kt_duct

Figure 14. Thrust breakdown – DP2 propeller – Kt tot = 60%Kt_tot_des

5. CONCLUSIONS

In the present paper, the description of a newly developed measuring device for ducted
propellers is reported. Comparison between measurements at cavitation tunnel and towing tank
is presented, showing satisfactory agreement at the different working conditions. Following this
preliminary analysis, a complete characterization of the cavitating behaviour of two ducted
propellers with accelerating and decelerating type of duct has been presented, in order to have
an overview of the various type of cavitating phenomena that affect this kind of propellers.
The device described in this paper represents a very useful instrument to conduct full tests on a
ducted propeller similarly to a conventional propeller, monitoring all the involved parameters,
including duct thrust and cavitation phenomena. This allows acquiring interesting data about
thrust breakdown and relating these results with cavitation observations. Thanks to data
obtained by the above mentioned tests, it was and will be possible to obtain important
information for the design of ducted propellers and for the calibration of existing design codes,
verifying that the propeller delivers the expected thrust at the design advance coefficient without
occurrence of cavitating phenomena that are typical of this kind of propeller, such as the tip
leakage vortex, and whose nature can be better investigated by these tests.
Future activities will be devoted to the analysis of other important side effects of propeller
cavitation, i.e. propeller radiated noise and pressure pulses, with the aim of investigating
effectiveness of decelerating ducts for improvement of propeller performances.
In parallel, the measuring system design will be improved: the final goal is to obtain a six degree
of freedom system measuring the three force components as well as two bending moments and
torque, maintaining the desired sensitivity of the system.

5. ACKNOWLEDGEMENTS

The measurement tool was designed and realized in house in the framework of a research project partially
funded by the University of Genova.
Present work was carried out in the framework of the BESST project (Breakthrough in European
Ship and ShipbuildingTechnologies); the research has received funding from the European
Community's Seventh Framework Programme (FP7/2007-2013) under grant agreement n°
233980.

REFERENCES

Baltazar J., Falcao de Campos J.A.C., Bosschers J. “Open-Water Thrust and Torque Predictions of a
Ducted Propeller System With a Panel Method”, Second International Symposium on Marine
Propulsors SMP’11, Hamburg, Germany, June 2011
Bertoglio C. (2011) “Sviluppo di un Sistema di Misura per Eliche Intubate”, Tesi di Laurea Magistrale in
Ingegneria Nautica, Polo Marconi, Università degli Studi di Genova
Carrera G. (2010) “Dispense del corso di Automazione Navale per Ingegneria Nautica”, a.a. 2010-2011,
Università di Genova
Ferrando M., Viviani M., Crotti S. (2007) “Performance of a Family of Surface Piercing Propellers”,
nd
Proceedings of 2 International Conference on Marine Research and Transportation (ICMRT 2007),
p.C63-C70, Ischia, 28-30 June 2007
Heinke H.J., Heinke C. (2011) “Open Water Tests with the Free Running and Ducted Controllable Pitch
Propeller VP40140”, SVA report 157/3840
Hoffmann K. (1989) “An Introduction to measurements using strain gauges”, Hottinger Baldwin
Messtechnik GmbH, Darmstadt, Germany
Oosterveld M.W.C. (1970) “Wake Adapted Ducted Propellers”, publication n. 345, NSMB – Wageningen -
Netherlands
Savio L., Viviani M., Conti F., Ferrando M., (2009) “Application of computer vision techniques to measure
th
cavitation bubble volume and cavitating tip vortex diameter”, Proceedings of the 7 International
Symposium on Cavitation (CAV2009), p. 737-748, August 17-22, 2009, Ann Arbor, Michigan, USA