First International Symposium on Marine Propulsors

smp09, Trondheim, Norway, June 2009

Aspects of Propeller Developments for a Submarine

Poul Andersen1, Jens J. Kappel2, Eugen Spangenberg3
1

Dept. of Mechanical Engineering, Technical University of Denmark (DTU), Kongens Lyngby, Denmark
2
J.J. Kappel A/S, Hilleroed, Denmark
3
ThyssenKrupp Marine Systems (HDW), Kiel, Germany

ABSTRACT

Design and development of propellers for submarines are

in some ways different from propellers for surface
vessels. The most important demand is low acoustic
signature that has priority over propeller efficiency, and
the submarine propeller must be optimized with respect to
acoustics rather than efficiency. Moreover the operating
conditions of a submarine propeller are quite different.
These aspects are discussed as well as the weighing of the
various propeller parameters against the design
objectives.
The noise generated by the propeller can be characterized
as thrust noise due to the inhomogeneous wake field of
the submarine, trailing-edge noise and noise caused by
turbulence in the inflow.
The items discussed are demonstrated in a case study
where a propeller of the Kappel type was developed.
Three stages of the development are presented, including
a design of an 8-bladed propeller where the thrust
fluctuations as well as the thrust noise were significantly
reduced relative to a 7-bladed propeller. Results of
measurements are in good agreement with calculations.
Keywords

submarine propeller, propeller acoustics, Kappel propeller,

propeller noise.

The noise criteria to be considered include thrust noise at

blade rate, trailing-edge noise, turbulence noise and
structural noise.
The parameters that can be varied during the design
process are: diameter, blade number, skew and rake,
circulation distribution (pitch and camber) and trailing
edge geometry. For submarines there are generally fewer
restrictions on the diameter than for normal surface ships
and increasing diameter means decreasing specific
propeller thrust and higher average inflow velocity.
In the present case several designs were examined.
Parameters varied include diameter, number of blades,
blade shape (skew and rake and in particular tip
geometry) and propeller turning rate. However, the paper
focuses on the development and application of propellers
of the Kappel type (non-planar lifting surface). Results of
calculations, model test and full-scale tests are compared.
2 DESIGN OBJECTIVES

The naval submarine is a complicated vessel built to

fulfill a multitude of conflicting requirements. Leaving
aside the crew and weaponry contained and supported by
the submarine, some of the objectives to be targeted by
the overall vessel and by the propulsion system and its
propeller are:
x

low acoustic, optical, electromagnetic, thermal

and communication signatures,

large operational window with respect to

endurance, speed and depth,

high manoeuvrability in surfaced

particular in submerged condition,

best possible form and arrangement of hull and

control surfaces to achieve a low hydrodynamic
resistance and a smooth hull wake,

propeller with low excitation level with respect

to noise and vibrations,

propeller totally free of cavitation with adequate

margins,

propeller with high efficiency.

1 INTRODUCTION

Design and development of propellers for submarines are

in some ways different from propellers for surface
vessels. The most important requirement for a submarine
propeller is a low acoustic signature. In the design
process the steps in the design procedure include
estimates of speed and power and weight as the usual
standard procedure. The main feature for the design of an
acoustically optimised propeller is the frequent check of
its acoustic properties. There are generally four
conditions that must be considered:
x
x
x
x

deep diving, turning and diving,

periscope depth, diving and with acceleration,
surface operation and high acceleration mode,
wind milling, near the surface.

and

in

Figure 3. It illustrates the main feature in the design

process of an acoustical optimized propeller: the frequent
check of the acoustics. Optimization can be achieved by
defined parameters which can influence each other. All
design parameters should be considered simultaneously
as far as possible. The evaluation of a solution and its
possible consequences for the general design is done
under the design step designated "feedback analysis".

Figure 1: Typical U-boat wake. Low-velocity regions can

be seen close to the hub at 1.30 and 3 o'clock positions
and symmetrically at the port side. A remarkable region
of relatively high velocity is present around 12 o'clock for
the outer radii.
A typical submarine wake is shown in Figure 1. It can be
seen that it differs significantly from most single-screw
surface ship wakes being without the usual low-velocity
region in the upper part of the propeller disk, but with
remarkable influence of the control surfaces just upstream
of the propeller. This, in combination with the demand for
adequate cavitation-free margin of propeller operation
together with the strict requirements to noise, as outlined
in Figure 2, is a challenge for the propeller designer.

Figure 3: Design spiral for submarine propellers.

3.1 Design matrix

Whereas the design spiral describes the design as a

process the objectives and parameters must be specified.
The overall requirements to the submarine leads to
objectives specifically related to the design of the
propeller. These objectives are to be achieved in each of
the operating conditions mentioned in the introduction of
this paper.
To meet the objectives the submarine propeller must be
carefully optimized and balanced with respect to the
design parameters available. Some of the implications of
this process can be demonstrated if the objectives defined
and the parameters available are seen as a design matrix,
cf. Table 1.
4 DESIGN CONSIDERATIONS

The issues presented in the design matrix in Table 1 lead

to a number of detailed design considerations.
4.1 Cavitation margin

Figure 2: Typical submarine propeller noise/frequency

limit.

3 DESIGN PROCEDURE

The design procedure of the propeller should ideally

include feedback to the design of the submarine hull and
its control surfaces. The progress is iterative and can be
illustrated with the classical design spiral. A similar
design spiral can be outlined for the propeller design, cf.

Most surface ships experience propeller cavitation during

normal operation. The design effort for surface ship
propellers is therefore directed towards avoidance of
harmful cavitation in way of erosion and towards
moderate levels of propeller excited noise and vibration.
However, propeller cavitation must be completely
avoided for a submarine in submerged condition in order
to secure the lowest possible noise signature. For this
reason, for a submarine in submerged condition, the
cavitation margin that determines the non-cavitating
operating regime is very important. In general, the
cavitation margins are defined by inception of:

Table 1 Design matrix. Influence: X little or none, XX moderate, XXX strong.

OBJECTIVES
cavit. free range

low noise and vibration level

tion

1.&2.
order
nat.
freq.

TE
noise

antising.

high
efficiency

XXX

XX

XXX

XX

XX

XX

XXX

XXX

XX

XX

pitch distribution

XXX

XXX

XX

camber distribution

cavit.
margin,
general

cavit.
margin
tip

excita-

orthogonal blade length

XX

section profile

XX

circulation distribution

PARAMETERS

stress
limit

XXX

XXX

XX

XX

skew

XX

XX

XX

XX

area ratio

XX

XX

XX

XXX

XX

area distribution

XX

XX

XX

XX

XX

diameter

XX

XX

XX

XXX

XX

XX

RPM

XX

XX

XX

XX

XX

XX

blade number

XXX

XXX

XXX

XX

thickness ratio

XX

XX

XX

XXX

XXX

XX

XX

trailing edge thickness

anti-singing edge form

XX

XXX

material

XXX

XX

suction-side cavitation

suction-side tip-vortex cavitation

pressure-side cavitation

pressure-side tip-vortex cavitation

hub-vortex cavitation

Suction and pressure-side cavitation can normally be

avoided by allowing adequate room in the classical
cavitation bucket. This may lead to a thicker section
profile than used for a comparable cavitating propeller of
a surface ship.
Tip-vortex cavitation is apart from ambient pressure
less vapor pressure dependent on the vortex intensity
and the viscosity of the fluid. In turn the vortex intensity
depends on the gradient of circulation at the outer part of
the propeller blade. Submarine propellers are therefore
generally tip relieved in order to obtain sufficient margin
against tip-vortex cavitation. This leads to a less-thanideal circulation distribution with respect propeller
efficiency. As tip-vortex cavitation is also dependent of
viscosity a correction of the results observed in the model
tank will have to be applied to arrive at full-scale
expectations, for example according to McCormick. The
formation of tip-vortex cavitation observed in model tests
seems very dependent on local tip geometry. An
explanation could be that the amount of boundary-layer
material ending up in the vortex core depends on the

blade tip geometry. This would in turn have an effect on

the cavitation margin as more boundary-layer material
would increase the viscous and turbulent boundary core.
The hub vortex consists of the combined blade root
vortices. Similar to the tip vortex, the hub vortex is
dependent on the circulation gradient at the blade root.
The intensity of the hub vortex can be reduced by
relieving the blade loading towards the root. Similar to tip
relieving this leads to reduction in propeller efficiency.
Alternative means to suppress hub vortex cavitation may
be used by applying a boss cap with small lifting surfaces
which reduces the rotation of the hub vortex core.
Furthermore, hub vortex cavitation can be suppressed if
the boss cap is given a form that will generate enough
turbulence to increase the vortex core and thereby avoid
cavitation.
Cavitation performance for submarine propellers is
generally illustrated in a diagram presenting the cavitation
number N as function of the trust loading KT at which
the cavitation starts to occur. This presentation is not
directly suitable when comparing propellers where
parameters such as propeller diameter are different.
Instead, it is suggested to compare the relative margin
against cavitation 100KT/KT as a function of the
operational depth Tpc for a given speed, where KT is the
design value and KT is the difference between the
design value and the cavitation limit of KT, cf. Figure 4.

reason large skew angles are often avoided in case of

submarine propellers.
4.4 Orthogonal
distribution

Figure 4: Relative margin against cavitation, i.e. cavitation

starts at the deviations in thrust coefficient shown. The
results are based on model tests without scale corrections
for propellers 4 and 6 discussed in Section 6.2.

blade

length

and

circulation

The circulation distribution has been discussed in 4.1 with

respect to load relieving of tip and root in order to obtain
sufficient margin against cavitation. However, the
application of "non-planar lifting surfaces" (Andersen et
al. (2005)) changes the circulation significantly. It is no
longer meaningful to present the circulation as a function
of the radius, instead circulation as a function over the
orthogonal blade length is used. This length is measured
orthogonally to a surface of helicoidal streamlines passing
the propeller blade. For a conventional propeller with a
radial generator line the orthogonal blade length is that of
the radial line independent of skew, whereas the
orthogonal line with or without skew might be about 10
per cent longer for the propeller with non-planar lifting
surfaces. A typical circulation distribution for a Kappel
propeller with non-planar lifting surfaces and a
conventional propeller is shown in Figure 5. Note that
gradient or derivative of the circulation distribution and
the maximum circulation is smaller in case of the Kappel
propeller. Both facilitate the adjustment of the load
distribution with respect to increase of cavitation margins
and blade-area distribution with respect to reduction of
unsteady forces.

4.2 Reduction of wake-induced propeller forces

The most important parameter with respect to propellerinduced forces is the wake field itself which for many
reasons should as fair as possible, be it surface ships or
submarines. However, with a given wake field the
number of propeller blades can probably be considered
the most important issue. Increasing the number of
propeller blades will reduce the unsteady force on each
propeller blade, but not necessarily the total unsteady
force. An example of unsteady force calculations with a
panel method for propellers with various geometries and
blade numbers are discussed in section 6.3.
Increase of the propeller diameter may increase or
decrease the unsteady forces depending on the radial
distribution of the unsteadiness in the wake field. An
example of variation of unsteady forces with propeller
diameter calculated by a panel method is shown in section
6.3.
4.3 Skew, pitch and camber distribution

The application of skew can reduce in particular the firstorder thrust and torque fluctuations and improve the
cavitation behaviour as commonly seen in surface ships.
However, a fairly large skew angle is needed to reduce
torque and thrust fluctuations significantly. Large skew
angles lead to fairly complicated higher-order flexural
modes of the propeller blades not to mention the blade
stresses when the propeller is going astern - for which

Figure 5: Spanwise circulation distribution of a

conventional and a Kappel propeller designed for the
same task. S0 is the length parameter along the orthogonal
blade.
5 ACOUSTICS

The effectiveness of a submarine is highly dependent

upon its noise signature, the special task of the boat
giving it mainly a passive role and control function. The
hydrodynamic flow noise dominates in particular the
performance of the sonar system. The own noise of the
boat correlates with the detection range, with respect to
detecting other ships and to being detected. Three types
of noise can be identified:

Thrust noise, dominating the noise level in the

low frequency range (<100 Hz),

Trailing edge noise (200-1000 Hz),

Broad band noise (100-1000 Hz), caused by

turbulence in the inflow.

5.1 Thrust fluctuations

The oscillating thrust of the propeller is probably the

dominating acoustic source. On the other hand thrust
fluctuations are not the only source as far as noise is
concerned. Measurements in full scale indicate
differences from theoretical calculations. In general, a
reduction of thrust fluctuation by 50 per cent yields a
reduced value for the noise level of 6dB which
theoretically halves the detection range of the submarine.
The acoustic level (dipole source) is proportional to the
harmonic order m, blade-number Z, turning rate n, and
the m-harmonic fluctuating force and consequently the
diameter, skew and number of blades of the propeller
must be selected with care.

whereas for surface ships the immediate problem is

propeller singing, cf. Carlton (2007). For this reason
submarine propellers are rarely provided with antisinging edges, unless there are specific problems.
Theoretical work on trailing-edge noise has been done by
Lighthill (1952) and Blake (1986), and in addition
measurements exist, cf. Carmargo et al. (2005). These
measurements show clear evidence of vortex shedding
and correlation with the noise level, indicating that the
frequency increases somewhat with flow speed and that
the harmonic content also increases with angle of attack
of the blade sections.
The concentrated fluctuation intensity may be expressed
by (1st approach by Ross (1987)) I 5.6*10-10. For a
blade section the typical sound level is in the range of
27dB (f 350 Hz). The Strouhal number is estimated as
0.2 for all conditions.
In a case with a full-scale propeller with a blunt trailing
edge the noise level was nearly 50dB indicating most
probably induced trailing-edge noise in accordance with
the results of Figure 7.
The overall blade shaping, the trailing-edge geometry and
the boundary layer in the local region are decisive for the
trailing-edge noise, cf. Blake (1986). Reshaping the
trailing-edge region altogether shows some progress.
Wang et al. (2006) show results of a numerical optimization of trailing-edge noise where a substantial
reduction in the vortex shedding was obtained with a
corresponding improvement in the acoustic spectrum, in
particular in the low-frequency range.

Figure 6 Reduced noise (80 per cent reduction of thrust

fluctuation) for the 8-bladed propeller relative to the 7bladed propeller.
This is illustrated in Figure 6 where a large reduction in
the thrust-fluctuation noise has been obtained by
increasing the blade number from 7 to 8. The reduction in
thrust variation is calculated by the boundary-element
method, cf. Table 3, and the noise predicted by the
procedure by Ross (1986) on the basis of these results.
These results are in accordance with measurements, cf.
Figure 10.
Although thrust fluctuations can be reduced by a good
propeller design it is important to emphasize that the
wake optimization should begin early in cooperation with
the propeller designer. A poor flow field is not a good
baseline for the process of minimizing noise.
5.2 Trailing-edge noise

The trailing-edge noise should not be neglected. For

submarines the whole noise spectrum is of concern,

Figure 7 Trailing-edge noise generation and vortex

strength as function of trailing-edge geometry. The
Kappel propeller examined generally all have a form
factor of 0.22 as indicated.
In Figure 7 results of a simple analysis is presented of the
correlation between trailing-edge geometry, trailing-edge
vortex strength and noise. The analysis is made following
the procedure by Blake (1986) for two-dimensional

sections where the trailing edge is varied locally, from

very blunt shapes to very sharp forms. The shapes are
described by a form factor. From the figure it is seen that
there is a correlation with the trailing-edge vortex
strength, the more blunt trailing edge giving more noise.
In general, the noise starts at frequencies larger than
200Hz up to 1000Hz, and the strength varies with the
blade section radius, the wake harmonics and partly with
the local coherency of the vortices. For high Reynolds
numbers (larger that 106) the situation in unclear.
In the worst case singing will occur where the trailingedge noise gives strong tonal spectra and blade structural
regions synchronise their vortex shedding in the so-called
lock-in effect. Generally, the structural dynamic
behaviour should be taken into account in the concept of
propeller design. The blade geometry influences the
eigenmodes, but the highly damped material Sonoston
helps in some cases to reduce the peak magnitude
somewhat.
5.3 Inflow-induced noise

The inflow to the propeller is a stochastic flow field

superimposed the steady inflow field. The turbulence
intensity and the integral length scale are decisive values.
The inflow turbulence increases the noise-spectrum,
especially if the turbulence length is equal or close to the
blade distance. Although this phenomenon has not been
experienced with open propellers it occurs for pumpjets
with high blade numbers.
5.4 Perspective

For further developments of understanding of the flowinduced noise, as introduced in this paper, a viable and
promising alternative would be systematic acoustic measurements, for example using "piezo-technique" on the
surface of the blade (trailing-edge region), last but not
least to verify theoretical calculations. The approach by
Lighthill equations (Lighthill (1952)) can be designated
as a first estimation. If enhanced predictions are to be
made the efforts will increase strongly. More comprehensive software packages (hybrid procedure) exist on
the market and use more or less the principles of Lighthill
analogies as well. The applicability of this software is demonstrated for some examples, but the applications for
the propeller design process are most doubtful. What are
needed first are accurate experiments in model and full
scale which are essential to scaling the theoretical
calculations.
6 EXAMPLE

The development of a propeller for a particular submarine

design is described in the following and it demonstrates
some of the problems involved in the design process.
6.1 Initial design stage

For the initial stages of design four designers were invited

to present their best proposal for the submarine propeller.
The propellers were tested at Potsdam Model Basin,
(Schiffbau-Versuchsanstalt Potsdam, SVA), Germany, in
their cavitation tunnel in open water and behind a dummy

Figure 8: Model of Propeller 8, the 8-bladed Kappel

propeller as discussed in Section 6.3.
model of the submarine. The initial design stage was
followed by a second design stage in which the design
restrictions with respect to the first order natural blade
frequency were relaxed. Two propellers were tested in the
model tank at SVA in the second design stage. Two
further designs were tested in the model tank in a third
design stage. All propellers are fixed pitch and righthanded. Propellers 4, 6 and 8 were of the Kappel type.
The detailed geometries of the propellers were quite
different. The main parameters of the eight propeller
designs are listed in Table 2 with Propeller 4 as reference.
In this paper the particular design work with three Kappel
designs will be described even though several other
designs were also developed and tested.
Propeller 4 (Kappel) of the first design stage was tested in
full scale with the original trailing-edge geometry and
with a trailing-edge geometry modified corresponding to
an anti-singing edge for a surface ship, sloped from the
suction side. This modification reduced the noise level
significantly in the frequency range 150 to 350 Hz, as
indicated in Figure 9. Propeller 4 showed moderate thrust
fluctuations after the modification.
6.2 Second design stage

Based on the model-test and full-scale results of the first

design stage and on relaxed conditions with respect to
first-order natural blade frequency, two further designs
were made in the second design stage. Relaxation of the
design requirements made possible a substantial
improvement of the cavitation margin and small reduction
in the thrust fluctuation. Propellers 5 and 6 (Kappel) were
designed in this stage.

Table 2. Propeller designs

1st design stage

2nd design stage

3rd design stage

Propeller
1

Propeller
2

Propeller
3

Propeller
4*

Propeller
5

Propeller
6*

Propeller
7

Propeller
8*

Diameter (relative)

1.06

0.96

0.96

1.0

0.94

1.0

0.94

1.0125

Pitch ratio

0.83

1.02

1.05

1.03

1.04

1.09

1.21

1.09

Skew, degree

13.75

-15.03

22.99

30.8

31.6

29.7

31.4

29.6

Blade area ratio

0.52

0.50

0.76

0.59

0.61

0.59

0.59

0.59

No. of blades

*Propellers 4, 6 and 8 are of the Kappel type.

results confirm the significant reduction found by the
calculations when changing from the seven to eight
blades. The model tests for Propeller 6 (seven blades) and
Propeller 8 (eight blades) are shown in Figure 10. The
difference between the seven-bladed and the eight-bladed
propellers is significant and the difference is more than
10dB which correlates well with the theoretical Lighthill
calculation based on the values of the panel calculation.

Figure 9: Envelopes of measured full-scale thrust

fluctuations for Propeller 4 (cf. Table 2) with original and
modified trailing edges.

6.3 Third design stage

The harmonic analysis of the wake of the submarine

indicated that reduced propeller excitations might be
expected in case a six-bladed and particularly an eightbladed propeller were selected instead of a seven-bladed
propeller. For this reason the propeller forces and
moments were calculated for propellers in the behind
condition by use of the boundary-element method in the
third design stage. The propellers were assumed to be
non-cavitating. The calculations were carried out for the
seven-bladed versions, Propeller 4, Propeller 6 and
furthermore for eight-bladed propellers with diameters
(relative) of 1.0, 1.025, 1.05 and 1.1 (versions of
Propeller 8). Table 3 gives an overview of forces and
moments calculated. According to these calculations the
design changes from Propeller 4 to Propeller 6 resulted in
a minor reduction of thrust fluctuation. However, the
change from seven to eight blades reduces the thrust and
torque fluctuations to about 1/5. The fluctuation of the
vertical force component is of the same magnitude for the
blade numbers seven and eight, whereas the horizontal
fluctuation increases 2 times when changing to eight
blades.
The thrust fluctuations were furthermore investigated in
the circulating model tank of SVA. The model tank

The optimum diameter with respect to efficiency is found

to be in the range 1.0 to 1.025 for the eight-bladed
propeller version. The thrust and torque fluctuations
calculated by the panel method increases with diameter,
whereas the horizontal and vertical force fluctuations
decrease marginally. The intensity of the tip vortex will in
principle decrease as the diameter is increased, thereby
increasing the margin against tip vortex cavitation.
However, the cavitation margin in general will decrease
as the diameter increases with constant RPM. An eightbladed propeller with a diameter of 1.0125 (Propeller 8)
was designed and tested in the model test facilities of
SVA.

Figure 10. Spectral analysis up to 120 Hz of thrust

fluctuations, Propellers 6 and 8 (respectively 7 and 8
blades). The propeller model is turning at 13 rps. The
reduction is more than 10 dB at the 1st order blade rate,
91 Hz respectively 104 Hz, when comparing Propeller 6
and 8.
By courtesy of Schiffbau Versuchsanstalt Potsdam (SVA).

Table 3: Total forces and moments acting on propellers in inhomogeneous inflow

Propeller

Propeller
4

Propeller
6

Propeller
8.1

Propeller
8.2

Propeller
8.3

Propeller
8.4

D, (relative)

1.0

1.0

1.0

1.025

1.05

1.1

No. of blades

Kx

0.25942

0.25081

0.25296

0.23448

0.21288

0.16736

kx

0.0517

0.0481

0.0089

0.01006

0.00918

0.00918

ky

0.0093

0.00767

0.00963

00.0086

0.00756

0.00557

kz

0.0071

0.00685

0.01724

0.01541

0.01364

0.01074

Mx

0.04326

0.04221

0.04296

0.03880

0.03440

0.02608

kmx

0.0076

0.00771

0.00171

0.00183

0.00170

0.00149

kmy

0.0032

0.00324

0.00519

0.00470

0.00416

0.00313

kmz

0.0051

0.0052

0.00964

0.00895

0.00817

0.00682

kx/Kx

0.1993

0.1918

0.0352

0.0429

0.0458

0.0549

ky/Kx

0.0358

0.0306

0.0381

0.0366

0.0355

0.0333

kz/Kx

0.0274

0.0273

0.0682

0.0657

0.0641

0.0642

kmx/Mx

0.1757

0.1827

0.0398

0.0472

0.0494

0.0571

kmy/Mx

0.0740

0.0768

0.1208

0.1211

0.1209

0.120

kmz/Mx

0.1179

0.1232

0.2244

0.2307

0.2375

0.2615

Kx*=kx/Kx+kmx/Mx

Kx*

0.3750

0.3745

0.0750

0.0901

0.0952

0.1120

Kyz*=ky/Kx+kz/Kx

Kyz*

0.2551

0.2579

0.4515

0.4541

0.4580

0.4790

+kmy/Mx+kmz/Mx

7 CONCLUSION

The development work for submarine propellers has been

described and the differences relative to surface
propellers outlined. The design process has been
illustrated for a propeller of the Kappel type applied to a
submarine design. The main concern has been a low
acoustic signature and one of the means of obtaining this
has been using eight blades, relative to seven blades used
initially. Theoretical calculations as well as model tests
demonstrate that lower thrust fluctuations were obtained
accompanied by lower thrust noise.
ACKNOWLEDGEMENTS

The authors thank ThyssenKrupp Marine Systems for

permission to publish the results of this propeller
development. Also, the authors thank the Potsdam Model
Basin (SVA) and in particular Dr. Reinhard Schulze for
making the result of the measurements available for the
paper.

REFERENCES

Andersen, P., Friesch, J., Kappel, J.J., Lundegaard, L. &

Patience, G. (2005), 'Development of a Marine
Propeller with Non-Planar Lifting Surfaces', Marine
Technology 42(3), pp. 144 - 158.

kx = kxmax - kxmin

Blake, W. K. (1986). Mechanics of flow-Induced Sound

and Vibration, Vol. II, Complex Flow-Structure
Interactions. Applied Mathematics and Mechanics,
17-II, Academic Press.
Carlton, J. (2007). Marine Propellers and Propulsion,
Second Edition, Butterworth-Heinemann, Elsevier.
Carmargo, H.E., Smith, B.S., Devenport, W. J. &
Burdisso, R.A. (2005). Evaluation and Calibration of
a Prototype Acoustic Test Section for the Virginia
Tech Stability Wind Tunnel, Report VPI-AOE-294,
Virginia Polytechnic Institute and State University,
Blacksburg, Virginia, USA.
Lighthill, M. J. (1952). 'On Sound Generated Aerodynamically, I. General Theory'. Proceedings of the
Royal Society of London, Series A, Mathematical and
Physical Sciences 211(1107) (Mar. 20, 1952), pp.
564-587.
Ross, D. (1987). Mechanics of Underwater Noise,
Peninsula Publishing.
Wang, M., Marsden, A.L. & Moin, P. (2006).
'Computation and Control of Trailing-Edge Noise',
Proceedings of the 26th Symposium on Naval
Hydrodynamics, Rome, Italy.