Sixth International Symposium on Marine Propulsors

smp’19, Rome, Italy, May 2019

Generation and propagation of noise from cavitating marine propellers

Kostas Belibassakis, Gerasimos Politis

School of Naval Architecture & Marine Engineering, National Technical University of Athens, Greece

ABSTRACT frequencies (below 300 Hz), ambient noise levels have

Commercial ships contribute significantly to the been increased by 15-20 dB over the last century
background noise in the sea, and propeller is considered a (Wittekind et al. 2014, McKenna et al. 2012). On the side
main source concerning shipping noise. International of mitigation measures, IMO has recently issued
Organisations have recently issued guidelines for the guidelines for the reduction of underwater noise from
reduction of underwater noise from commercial shipping
commercial shipping to address adverse impacts on
in order to address adverse impacts on marine life,
including technical guidelines for the designers, marine life, including technical guidelines for the
shipbuilders and ship operators to minimize underwater designers, shipbuilders and ship operators to minimize
radiated noise from ships. In this work, the development underwater radiated noise from ships (IMO 2014).
of an unsteady propeller analysis BEM is presented and
discussed, providing us with the basic information and The broad bandwidth of the produced noise spectrum
data that are subsequently used to estimate noise indicates that shipping is a complex noise source
generation and propagation in the sea environment. The composed by the superimposition of several contributors.
acoustic prediction is based on the time-dependent The propeller noise is considered as main contributor to
pressure and cavitation volume data that are used for the commercial shipping noise, especially under cavitating
solution of FW-H equation by means of the Farassat conditions, whereas, below cavitation inception on-board
formulation. The model is further elaborated to include
machinery and hydrodynamic noise due to wave-structure
the reflective effects of nearby boundaries of free surface
(Lloyd mirror effect) and from the ship hull, which are interaction and wave breaking are also important.
expected to influence the sound directivity characteristics. Generally, propeller noise is dominant at high speeds. As
Focusing on the estimation of shipping noise in the low for the flow-induced noise due to ship-hull
frequency band of the spectrum, the applicability of the hydrodynamic interactions it is considered to be
above noise generation models, in conjunction negligible far from the vessel (AQUO 2015).
with suitable underwater acoustic propagation models,
is subsequently discussed, for the estimation of the In this work, underwater noise spectrum induced by
acoustic noise footprint from many sources representing marine propellers is estimated using a numerical model.
the travelling ships in an examined geographical region. The prediction of unsteady performance of marine
Key words: propeller noise, unsteady load, cavitation, propellers with and without cavitation is based on a
directivity effects, boundary effects numerical lifting surface theory, which provide us with
the basic information and data that are subsequently
1. Introduction exploited to study noise generation. In general, the
Commercial ships contribute significantly to the acoustic prediction is based on the time-dependent
background noise in the sea. This is due to the fact that pressure and sheet cavity volume data that are used as
the number, size and speed of ships continuously increase input for the Farassat formulation (Farassat 2001) of the
leading to increasing noise levels. Various studies, FW-H equation (Ffowcks Williams & Hawkings, 1969).
covering the period 1950-2007, indicate strong For the numerical formulation of noise prediction in the
relationship between commercial shipping activity and low-frequency band a simplified model based on
global economic growth (Frisk 2012). Furthermore, the appropriate contributions of acoustic monopoles and
importance of shipping noise and its impact on marine dipoles is derived. The unsteady sheet cavity volume
environment is also demonstrated by the fact that at low fluctuations are modeled by monopole sources, whereas,
the fluctuations of the unsteady pressure jump on
propeller blades are modeled by corresponding dipole
terms. Having calculated the noise spectrum at any point
near the sound source, the directivity patterns are
obtained, indicating anisotropic characteristics.
Subsequently, the present model is further elaborated to
study the reflective effects of nearby boundaries of free
surface (Lloyd mirror effect) and ship hull. It is shown
that free-surface effects, strongly influence the
underwater noise propagation in the marine environment
comparatively to the omni-directional source assumption.

2. Unsteady propeller analysis

The propeller hydrodynamic analysis in inhomogeneous

inflow conditions (ship’s wake due to hull’s boundary Figure 1. Representation of propeler blades and trailing
layer) is based on a vortex lattice method (Kerwin & Lee, vortex sheets using vortex elements on the mean camber
1978, Lee 1979). Given the induced unsteadiness due to surface.
ship wake and propeller’s rotating motion, a time-
stepping method is applied for the solution. The complete Subsequently, moving in the downstream direction, the
propeller discretization is obtained by a vortex lattice trailing vortex lines gradually deform to smoothly match
consisted of quadrilateral vortex element on the mean the ultimate wake trailing vortex lines. In the case of the
camber sufaces; see Fig.1. The method is modified in deformed wake model an additional contraction of the
order to include blade thickness and sheet cavitation blade wake is superimposed; see Fig.1.
effects by including source-sink distribution with
appropriate density.
2.1 Boundary Value Problem
To account for the effects of trailing vorticity, a free
The propeller blades are considered to be a set of
vortex sheet downstream of the blades has been
symmetrically arranged thin blades, rotating with a
incorporated to the model. Since the geometry of this
constant angular velocity about a common axis in an
additional boundary is unknown, it has to be determined
unbounded, incompressible fluid. The symmetric
as part of the solution. This is usually accomplished by
arrangement along with the hypothesis of weak
free wake analysis, based on a wake relaxation method to
interaction between the blades, are exploited in order to
incorporate the nonlinear effects. Taking into
reduce the computational cost. The presence of
consideration the unsteadiness of the problem, the grid
extraneous boundaries such as the rudder, the hub and the
representing the free wake surface is forced to
ship hull, are neglected, except that the last is recognized
dynamically evolve with time; see, e.g., Politis (2004).
as the body generating the non-uniform flow field (ship’s
According to this method some initial wake geometry is
wake). The blade boundary layer and shed vortex wake
assumed for which wake grid panels are defined.
thickness is assumed to be thin so that the fluid rotation
However, assuming that the shape of hydrodynamic
due to the propeller is confined in a thin layer.
trailing vortex sheets can be defined by means of wake
model, the problem can be approximately reduced to an For the incoming velocity field, it is also assumed the
integral equation over the solid surface coupled with the smallness of angle of attack, as well as the smallness of
Kutta condition, which substantially accelerates the spatial variation. The propeller approximately treated by
numerical solution. separating the velocity field in two terms: the first one
refers to the rotational wake, whereas the second to
In the case of wake models the radial distribution of the
perturbation velocity due to the existence of propeller.
hydrodynamic pitch angle, are selected to be given by a
The last can be considered irrotational outside the blade
simplified formula, in terms of the main geometrical and
boundary layer. Under these considerations, the Euler
hydrodynamic parameters; see, e.g., Lee (1979), Politis &
equations can be numerically integrated, giving us a
Belibasakis, (1990). Within the transition wake region,
decoupled formulation for the velocity and pressure fields
trailing vortex lines emanate from the trailing edge of
(Belibasakis & Politis 1998).
each blade aligned to the local trailing edge angle
bisector.
The principle of conservation of mass is the governing
equation throughout the fluid encompassing the blades,
the shed wake, and the cavity. The total fluid velocity w
in the propeller frame of reference is given by:
w = U + A + Ω × x + u, u = ∇Φ , (1)

where U and Ω are propeller translational and rotational

speed, respectively, A denotes the ship’s wake, and u is
the propeller’s disturbance velocity field. The latter is
obtained as solution of the Laplace equation for the
disturbance flow potential outside the propeller blades
and trailing vortex sheets, as well as the cavity surface:

∇2Φ = 0 . (2)
Figure 2. Representation of propeler blades and trailing
The unknown disturbance potential is numerically
vortex sheets using vortex and source-sink elements on
approximated by imposing the following boundary
the mean camber surface.
conditions in order to form a close set of equations:
More, specifically, a relation is obtained for the spanwise
2.1.1 Tangency condition on the blade surface
shed vorticity from the trailing edge of the blade in the
In a blade-fixed coordinate system rotating with the trailing vortex sheets, in terms of the temporal variation
propeller, the impermeability condition on the blade and of the circulation around the blade section at each
solid surfaces is described as: spanwise position, consistent with Kelvin's theorem for
nw = 0 , (3) the conservation of circulation.

where, n is the vector normal to the blade surface. Using

2.1.3 Conditions on the trailing vortex sheets
the following representation for the disturbance velocity
field: The velocity jump across each trailing vortex sheet is
tangential to these surfaces (kinematic condition), and the
1 σr 1 r
u( x 0 ) = ∫ dS ( x ) + ∫ γ× dS ( x ), (4) pressure is continuous across this blade wake (dynamic
4π S B ∪ SC
r 3
4π SB ∪ SW
r3 condition). The latter are approximately satisfied by
propagating the vorticity emitted at the trailing edge to
where, r = x 0 − x , σ is surface source-sink distribution
each blade wake model.
defined by means of Eq.(3) on the blade and cavity
surface, and γ the surface vorticity on the blade and 2.2 Numerical solution scheme
wake surface; see Fig.2. In the Vortex Lattice method the
continuous distributions of source-sink and vorticity are The propeller hydrodynamic problem is solved at each
modelled by discrete source-sinks and vortices. Using time step, and the singularity distributions are calculated,
Eq.(4) in the boundary condition Eq.(3), it results in a from which velocity is computed using Eqs.(1) and (4).
linear algebraic system from which the discrete Once the velocities are found, the pressure distribution is
singularities on the propeller blades are calculated. The obtained from Bernoulli’s equation, as follows:
above representation ensures the satisfaction of condition
p − p(
Ι)
∂Φ 1
at infinity at each time step, in the form of demanding the
ρ
=− −
∂t 2
(
2 2
w − q + A = 0, ) (5)
vanishing of the perturbation velocity at large distances
where, p (
from the propeller. Ι)
denotes the incident field pressure in the
2.1.2 Kutta condition at the trailing edge moving coordinate system fixed to the propeller and ρ is
In the case of ideal flow, Kutta condition prevents the the density of the water; see also Belibassakis & Politis
existence of infinite velocities at the trailing edge. In the (1998). For operation at specific value of the advance
present study, necessitating equal pressure (dynamic coefficient:
Kutta condition) as the trailing edge is approached from
U + Am
both sides of the blade surface and provides us J= , (6)
nD
information concerning the vorticity field on the blade
wake in terms of the solution of the integral equation in where, U + Am is the mean value of the propeller
previous time steps. inflow at its disc, n the propeller revolutions per minute
and D = 2R is the propeller diameter, the unsteady forces
and moments, including thrust and torque, are
subsequently calculated by pressure integration on
propeller blades. The results concerning the thrust and
toque coefficients are obtained as follows:
T Q
KT = and KQ = . (7)
ρ n2 D4 ρ n2 D5

The corresponding coefficients corresponding to each

symmetrical part (each blade) of the propeller are dented
as KT ,1 and KQ ,1 , respectively.
Figure 3. Calculation of cavity surface on the suction side
3. Propeller blade cavitation at each section of the propeller blade.
In the case of cavitating flow, the prescribed vapor
Boundary Condition (DBC) requires that the pressure,
pressure provides an additional criterion that controls the
which is calculated using Bernoulli's equation (5),
cavity formation and decay. Some important phenomena
everywhere inside the cavity volume to be constant and
such as general cavity inception and surface tension are
equal to the given cavity pressure:
not considered. It is assumed that the cavity starts at the
leading edge of the blade, and that only the suction side of p = pc . (10)
the blade is cavitating (Lee, 1979, Kinnas & Young
2003). The non-dimensional parameters which 3.2 Numerical solution of the cavitating problem
characterize the cavitating flow field are the cavitation
The cavity boundary is determined by an iterative process
and Froude numbers, defined on the basis of the propeller
in which the dynamic boundary condition is satisfied on
rotational speed (n) as follows:
an approximate cavity surface and the kinematic
p − pv n2D boundary condition is used to update the surface. In the
σn = , and Fn = , (8)
0.5ρ n 2 D 2 g first iteration, the panels representing the cavity are
placed on the foil surface. In subsequent iterations, the
where, p is the atmospheric and hydrostatic pressure at
cavity surface is updated using the kinematic condition,
the propeller shaft center, pv denotes the vapour pressure and the boundary elements are moved to the updated
and g the gravity acceleration. surface. The local cavity length ℓ(u 2 , t ) , defined as the
arc length of the projection of the cavity (on the nose-tail
3.1 Conditions on cavitating propeller axis along u 1 ) at each radial position u 2 = const is
The problem concerning the unknown disturbance determined from the following requirement (closure
velocity and potential in the case of cavitating propeller is condition) at each spanwise strip:
solved by imposing additional boundary conditions due to
h(u 1 = ℓ(u 1 , t ), u 2 , t ) = 0 . (11)
the presence of the cavity. Introducing a curvilinear
coordinate system (u 1 , u 2 , u 3 ) as shown in Fig.3, and For a given cavitation number σ n , the above requirement
is used as the basis of an iterative solution scheme to
denoting by h (u 1 , u 2 , t ) the thickness of the cavity
calculate the cavity’s boundary and its temporal variation.
normal to the blade surface at the point (u 1 , u 2 ) at time As an example of application we consider two 5-bladed
t, the Kinematic Boundary Condition (KBC) is used to model propellers, one without skew (KP068) and one
determine the position of the cavity surface, given by the with 70deg skew (KP070) for which detailed
following equation: experimental data are available in (Kim et al. 1988). The
D 3 basic dimensions are: D=0.25m, P/D=1.2 (at 70% of tip
u − h (u 1 , u 2 , t )  = 0 , (9)
Dt 
radius), expanded area ratio 72.5%. The propellers
operate in an axial wake, as shown in Fig.4. In this case,
where, the material derivative appearing in the above the operating condition concerning the advance
equation is formulated in terms of the total flow velocity coefficient is J=0.889. Moreover, in the experiments the
on the surface of the cavity. Moreover, the Dynamic Froude number is Fn = 10.45 and the cavitation number
is σ n = 3.
Figure 4. Distribution of axial wake (Kim et al. 1988) on
the propeller disc.

Figure 6. (a) Calculated cavitation pattern on the blades of

propeller KP070 in the wake of Fig.4, for various angular
positions. (b) Key blade cavity volume and (c) center of
volume, respectively, as predicted by the present model.
In general, the present method predictions are found in
good agreement with measured data (see Kim et al. 1988).
Figure 5. Circumferential variations of Kt,1 and radial
coordinate of center of blade thrust for the normal KP068 4. Propeller noise prediction model
and skewed KP070 propellers. The solid and dashed lines
represent the cavitating and non-cavitating conditions, As propeller rotates in the non-uniform wake, it is
respectively. subjected to unsteady pressure loads and cavitation which
lead to discrete tonal and increased broadband noise (Seol
The calculated results of unsteady hydrodynamic analysis et al. 2005). Low frequency noise is caused by the
for the non-cavitating and cavitating conditions are fluctuations of blade pressure and variation of unsteady
presented in Figs.5 and 6. In particular, Fig.5 shows the sheet cavitation volume. The former has dipole character-
circumferential variation of blade thrust coefficient for the istics and the latter is modelled by a bubble that acts as an
key blade (Kt,1). Dashed lines indicate the non-cavitating acoustic monopole. High-frequency noise is caused by
conditions and solid lines the effect of cavitation. In the sheet cavity collapse and/or by shock wave generation.
same plots, it also presented the radial position of the
thrust center in terms of the propeller radius (multiplied The formulation of Lighthill’s acoustic analogy proposed
by 0.1). In Fig.6 the calculated cavitation pattern on the by Ffowcs, Williams and Hawkings (1969) is:
blades of skewed propeller KP070 in the wake of Fig.4 is 1 ∂ 2 p′
presented, for various angular positions. Also, the blade − ∇2 p′ = g m + g d + g q , (12)
c 2 ∂t 2
cavity volume and the coordinates of its center are shown.
For comparison in Fig. 6(b) the calculated result where p′ denotes the acoustic pressure, c is the speed of
concerning the cavity volume of the unskewed propeller sound in the medium (c=1500 to 1550m/s for water). The
KP068, as calculated by the present method, is also various terms in the right-hand side correspond to forcing
presented by using red lines. from the acoustic monopole, dipole and quadrupole
source terms (Farassat & Myers,1988), defined as for the acoustic wave to travel between the observation
follows: point x 0 and the disturbance generating point xT ,k .
∂ ∂
g m = [ ρ un | ∇f | δ ( f )] , g d = − [li | ∇f | δ ( f )] , (13) Similarly, for the thickness effect we have:
∂t ∂xi
∂  un 
∂t ∫ f =0  r (1 − M r )  ret
4π pT′ ( x 0 , t ) = ρ   dS . (19)
∂2
gq = − Tij H ( f )  , (14)
∂xi ∂x j  Considering the same approximations, the above equation
where, f=0 indicates the moving surfaces un the is simplified in the following form:
corresponding normal velocity, li the loads, Tij is the ρ ∂
[un ]ret dS .
r ∂t ∫ f =0
4π pT′ ( x 0 , t ) ≈ (20)
Lighthill’s stress tensor and H denotes the Heaviside
function. The quadrupole term is important for strongly Moreover, we consider that the blades are very thin in
transonic flow phenomena manifested at higher
comparison to the cavity volume, and using:
frequencies. Given that the speed of sound in water is
dQ (t )
much greater than the flow velocities, and focusing on the ∫ f =0 un dS ≈ dtc , (21)
low-frequency part of the generated noise spectrum the
contributions by the latter term are neglected in the where Qc (t ) denotes the calculated cavity volume, as e.g.
present work. In the sequel Farassat formulation 1A shown in the first subplots of Fig.5, we finally obtain:
(Farassat 2001, Seol et al. 2005) is employed offering an
ρ Z
d 2Qc ,k (tr ) 1
integral representation of the solution of Eq.(12) forced pT′ ( x 0 , t ) ≈
4π
∑
k =1 dt 2 rc ,k
, rc ,k = x 0 − xQ ,k (tr ) , (22)
only by the monopole and dipole terms. The acoustic
pressure field is accordingly given by thickness and where xQ ,k ( t ) denotes the center of cavity volume of the
loading components, as follows: k-blade.
p '( x 0 , t ) = pT′ ( x 0 , t ) + pL′ ( x 0 , t ) . (15)
5. Results and discussion
The loading term is given by:
1d  p nrˆ  The predictive capability of the above simplified model is
4π pL′ ( x 0 , t ) = −
c dt ∫ f = 0   dS +
 r (1 − M r )  ret
indicated by rescaling the KP060 and KP070 propellers
from model to full scale (D=7m, n=100rpm, and
 p nrˆ  U=20.2kn), simulating a real case scenario of a container
+∫  2  dS , (16)
 r (1 − M r )  ret
f =0 ship. What is more, we assume a moving observer with
exactly the same forward speed as propeller’s one. The
where p denotes the pressure jump on the blade
selected observation point is set equal to 10 times the
surface, M r denotes Mach number in the r-direction and propeller radius (R) in order to satisfy the model’s
the integrand is calculated at retarded time. For relatively simplifications to be valid. Initially, it is taken on the
large distances (of the order of several propeller propeller shaft axis and the calculation is exploited in a
diameters) of the observation point from the propeller, we spectral analysis so as to derive and compare the noise
use the following leading-order approximation: frequency spectra.
r = x 0 − x ≈ x0 − xT ,k ( t ) , and rˆ ≈ ( x0 − xT ,k ( t ) ) / r , (17) Calculations based on the above approximate formulas,
Eqs. (18) and (22), significantly reduce the computational
where, xT ,k ( t ) denotes the center of thrust on the k-blade,
cost of the surface integrals (16) and (19), respectively.
which is the most important load on the lifting surface. Numerical results are presented in Fig.7 for the normal
Using the fact that the Mach number is very small, (KP060) and skewed (KP070) propellers. More specifi-
Eq.(17) eventually leads to the following simplification: cally, in these figures the frequency spectrum of loading
1 Z
dTɶk (tr ) xP − xT ,k ( tr ) term (dipole) for the non-cavitating (dashed lines) and
pL′ ( x 0 , t ) ≈ −
4π c
∑ k =1 dt rT ,k 2
+ cavitating (solid lines) conditions is shown, with refere-
nce to 1µPa/Hz2 at 35m. We observe the appearance of a
1 Z xP − xT ,k ( tr ) peak at the blade frequency, which is sharper in the case
4π
∑ Tɶ (t )
k =1
k r
rT ,k 3
, (18)
of the skewed propeller. It is seen that without cavitation
(dashed lines) the noise is reduced in the case of skewed
where Tɶk (tr ) denotes the fluctuating part of the k-blade
propeller. However, under cavitating conditions, an
thrust, as e.g. shown in the first subplots of Fig.5 for increase of the peak of the noise spectrum for the skewed
propellers KP069 and KP070, and tr = r / c denotes the propeller is observed, which is due to the increased
retarded time which represents the time which is needed volume as calculated by the present model (see Fig.6b).
 Figure 9. Modification of the directivity characteristics of
the acoustic field due to presence of: (a) hard hull
boundary (Neumman b.c.) and (b) soft hull (Dirichlet b.c.)

Figure 7. Power density spectrum for the cavitating and

non-cavitating cases of propeller KP068 and KP070 with
reference to 1µPa at 35m.
Figure 10. Modification of the directivity characteristics
of the acoustic field on the vertical plane due to the free-
surface effect.
spreading law concerning the transmission losses:
TL[dB re 1µPa @ 1m] = 20 log r . (23)
The present model can be further elaborated to include the
effects of additional boundaries. As an example, the effect
Figure 8. Noise directivity 3D contours of cavitating
of hull boundary on the directivity characteristics of the
KP068 due to: (a) loading and (b) total noise.
noise generated by the propeller is shown in Fig.9. In
Also, it is evident that cavitation significantly modifies particular, we consider a simple (monopole) harmonic
(of the order of 20dB/Hz) the levels of the acoustic source representing the propeller located at the shaft axis
spectral density. It is also seen that the loading term of (near the frequency corresponding to the center of the
KP070 propeller presents a second peak around the second octave band), and an obstacle representing the hull
65Hz, which is due to the phase difference between the (L=100m and B=15m). Numerical results are presented
two terms of the loading noise. Furthermore, the modeling the hull surface as an ideal hard surface
directivity characteristics of the noise are presented in (Neumann b.c.) in Fig.9(a) and as soft surface (Dirichlet
Fig.8 in the case of the normal propeller. As expected the b.c.) in Fig.9(b), respectively. In reality, the effect of the
loading term has clear dipole characteristics, while the hull is defined by the reflective characteristics of the
total noise directivity pattern is more regular and mostly surface (dependent on the material, stiffness, coating,
affected by the manifestation of blade cavitation. Similar surface roughness etc) and is expected to be more close to
patterns are also obtained concerning the skewed the hard boundary condition (shown in Fig.9a), indicating
propeller. Given the fact that most studies report noise that the modification of the directivity pattern of the
results at the distance of 1m from the propeller, and propeller due to the hull is rather small. On the other
taking into consideration that the total noise directivity hand, the effect of free surface due to Lloyd mirror effect
pattern is close to omni-directional, and assuming by a pressure release boundary (zero acoustic pressure on
homogeneous acoustic parameters in the vicinity of the the free surface) is important, as shown in Fig.10, for the
propeller, the present results can been rescaled to a same as before frequency. The latter effect can be easily
distance of 1m by using the standard geometrical incorporated in the present simplified model by mirroring
the acoustic monopoles and dipoles of each propeller Farassat, F. , Myers, M.K., (1988). ‘Extensions of
blade above the horizontal plane, leading to the following Kirchhoff’s formula to radiation from moving surfaces’.
expression: Journal of Sound and Vibration 123 , 451–460.
1 Z dTɶk (tr ) xP − xT ,k ( tr ) Farassat, F., (2001). ‘Acoustic radiation deom rotating
pL′ ( x 0 , t ) ≈ − ∑
4π c k =1 dt rT ,k 2 − rɶT ,k 2
+ blades – the Kirchhoff method in aeroacousics’. Journal
of Sound and Vibration 239 785-800.
1 Z xP − xT ,k ( tr ) Ffowcs Williams J.E., Hawkings, D.L., (1969). ‘Sound
4π
∑ Tɶ (t )
k =1
k r
rT ,k 3 − rɶT ,k 3
, (24)
generated by turbulence and surfaces in arbitrary
motion’. Philosophical Transactions of the Royal
ρ Z d 2Qc ,k (tr )  1 1  Society A 264.
pT′ ( x 0 , t ) ≈
4π
∑ 2  −  , (25)
k =1 dt  rc ,k rɶc ,k  Frisk, G. (2012). ‘Noiseconomics: The relationship
between ambient noise levels in the sea and global
where rɶT ,k , rɶc ,k denote the distances from the image economic trends’. Scient. Rep. 2, Art.: 437.
points of the centers of thrust and cavity volume of each IMO (2014). Guidelines for the Reduction of
k-blade above the free surface, respectively. Underwater Noise from Commercial Shipping to
Address Adverse Impacts on Marine Life – non
6. Conclusions mandatory technical advices, International Maritime
Organization MEPC.1/Circ.833.
A numerical model is developed for the prediction of
Kerwin, J., Lee, C.S., (1978). ‚Prediction of steady and
noise generated from cavitating or non-cavitating marine unsteady performance of marine propellers by
propellers operating in unsteady inflow conditions in the numerical lifting surface theory’. SNAME Trans. 86.
wake of the ship. The hydrodynamic part is analyzed by a Kim, K.H. , Nguyen, P.H., (1988). ‘Propeller cavitationa
velocity-based vortex lattice method, providing the and cavitation induced pressure fluctuation: Comparison
unsteady pressure on the blades and cavitation data. The between theory and experiments’. Proc. SNAME
latter are subsequently used, in conjunction with Farassat propellers’88 Symposium (Virginia Beach).
formulation, to calculate acoustic radiation from moving Kinnas, S., Young, Y.L., (2003). ‘Modeling of
surfaces and predict the acoustic spectrum at a distance of Unsteady Sheet Cavitation on Marine Propeller Blades’.
several diameters from the propeller, representing the Intern. J. of Rotating Machinery 9, 263–277.
source of marine propeller noise. An approximate model Lee, C.S., (1979). ‘Prediction of Steady Unsteady
is derived, exploiting information and integrated data Performa-nce of Marine Propellers With or Without
concerning the time history of blade sheet cavity volume Cavitation by Numerical Lifting Surface Theory’. PhD
and the unsteady blade thrust. The latter are used to Thesis, MIT, Department of Ocean Engineering
calculate the monopole and dipole forcing terms of the Lighthill, M.J. (1952). ‘On sound generated
acoustic equation and derive the propeller acoustic aerodynamically, 1. General Theory’. Proceedings of the
Royal Society A211, 564–587.
spectrum in the low and moderate frequency band. Also,
the directivity characteristics of the propeller noise are McKenna, M.F., Ross, D., Wiggins, S.M., Hildebrand,
J.A., (2012). ‘Underwater radiated noise from modern
calculated, and the effect of nearby boundaries on
commercial ships’. J. Acoust. Soc. Am. 131 92-103.
underwater noise propagation are presented compara-
tively to the omnidirectional source assumption. In Politis, G.K., (2004). ‘Simulation of unsteady motion of
a propeller in a fluid including free wake modeling’.
particular, the effect of the free surface as a pressure
Engin. Analysis with Boundary Elements 28, 633–53.
release boundary (Lloyd mirror effect), and of the ship
Politis, G.K., Belibasakis, K.A., (1990). ‘Application of
hull, treated as hard and soft boundary, are illustrated.
Panel Methods to Linearized Lifting Surface Propeller
Future work will include the validation of present method Performance Problem’, Proc. 5th Intern Congress on
concerning the noise prediction by comparison against Marine Technology, IMAEM'90, Athens, Greece.
available data (e.g., Sharma et al. 1990 and similar). Seol, H., Suh, J.C., Lee, S., (2005). ‘Development of
hybrid method for the prediction of underwater propeller
References noise’. Journal of Sound and Vibration 288, 345–360.
AQUO (2015). Report R2.9, Ship Underwater Radiated Sharma S., Mani K., Arakeri V. (1990). ‘Cavitation
Noise Patterns, Achieve Quieter Oceans by shipping noise studies onmarine propellers’. Journal of Sound and
noise footprint reduction (FP7 Project No314227). Vibration 138, 255–283.
Belibasakis, K.A., Politis, G.K. (1998). ‘A non-linear Wittekind, D., Schuster, M., (2014). ‘Propeller
velocity based Boundary Element Method for the Cavitation Noise and Background Noise in the Sea’.
analysis of marine propellers in unsteady flow’. Proc. 1st Intern. Meeting – Prop. Noise & Vibr. 6-7
International Shipbuilding Progress 45 93-133. Nov. 2014, Istanbul.