Ship Propulsion
description
Transcript of Ship Propulsion
VIShip Propulsion
405
Skewed Propeller Design for Minimum Induced Vibrations M. A. Mosaad, W.Yehia
Faculty of Engineering, Port Said University, Egypt, [email protected],[email protected]
ABSTRACT
Propeller skew is the single most effective design parameter which has significant influence on reducing
propeller induced vibration without sacrificing the efficiency. Applications of propeller skew for a certain
propeller almost without specified criteria. The goal of this paper is to present a proposed concept design
criteria for propeller skew. Computational results for the flow patterns of skewed propellers with different skew
angles, with presence and absence of cavitation inception are presented. The computational results are carried
out by FLUENT software using unstructured grids, based on Reynolds-Averaged Navier-Stokes computational
fluid dynamics method. Finally, comparative analysis of numerical results of the simulations is presented for
the selection of the best propeller skew angles. The overall results suggest that the proposed approach is
practicable for propeller designs for minimum induced vibrations.
Keywords: Propeller skew, CFD simulation, RANS
1. INTRODUCTION
During recent years Computational Fluid-Dynamics (CFD) models have demonstrated to rapidly become
effective tools to analyse marine propeller single-phase flows. In contrast to this, cavitation presents complex
two-and multi-phase flow phenomena that are still difficult to accurately simulate (Salvatore and Streckwall,
2009).
Cavitation occurs on nearly all ship propellers. It may lead to expensive problems if not acknowledged in an
early design stage. The two most frequently occurring problems are vibrations and noise in the afterbody due to
cavitation-induced pressure fluctuations on the hull, and cavitation erosion on propeller blades and appendages.
Early recognition of these adverse effects is important, not only to ensure compliance with contract
requirements, but also because often cavitation has to be controlled at the cost of propeller efficiency (Tom and
Terwisga, 2006). To ensure that the propeller meets the requirements that relate to comfort (vibration and
noise) and safe and economic operation (erosion), model scale experiments or computations that address
cavitation are to be conducted prior to construction (Tom and Terwisga, 2006).
Due to high operational costs of experimental investigations it is highly desirable to be able to study cavitation
with reliable CFD techniques (Lifante and Frank, 2008).
The paper presents a comparative study of a RANS prediction of flow pattern characteristics of two families of
skewed propellers. These propellers are characterized by the presence or absence of cavitation inception. The
first family is a skewed propeller family of DTMB-P4119 to be studied as non-cavitating propellers. The
second is a family of INSEAN-E779A as cavitating propellers.
406
2. NUMERICAL SETUP AND COMPUTATIONAL SCHEMES
General conservative form of the Navier – Stokes equation is presented as the continuity equation
Continuity equation,
(2.1) Where: ρ = density, [kg/m3]
iu = is the velocity component in the ith direction, m/s (i =1, 2, 3) and S m = source terms. In case of incompressible flows the density is considered to be constant. Since the propeller flow has been
considered as steady and incompressible, the continuity equation gets modified as,
0)( ii
ux (2.2)
The momentum equation will be,
iij
ij
i
jij
i
Fgxx
p
uux
ut
)()( (2.3)
Where:
ijl
l
i
j
j
iij x
uxu
xu
32)]([ , (2.4)
ij = is the Reynolds stress tensor
p = static pressure, [N/m2]
gi = gravitational acceleration in the i-th direction , [m/s2]
Fi = external body forces in the i-th direction and, N
ij is the Kronecker delta and is equal to unity when i=j; and zero when i j.
The Reynolds-Averaged form of the above momentum equation including the turbulent shear stresses is given by:
(2.5) Where:
'iu = is the instantaneous velocity component, m/s (i = 1,2, 3).
mi
i
Suxt
)(
407
In the present work, the SST (Shear Stress Transport) k- ω turbulence model is chosen for turbulence closure.
The SST k-ω model is currently one of the most widely used turbulence models for propeller flow simulation
(Krasilnikov and Jiaying, 2009).
For the cavitating propeller cases, the cavitation model was activated, using a multi-phase CFD setup with
water and water vapor under normal conditions as the working fluids (Lifante and Frank, 2008).
Regarding the Boundary Conditions for cavitation cases were set in the same way as for the non-cavitating
cases. The only difference was at the exit boundary, where a constant exit pressure was set to match the given
cavitation number (σ) (Shin and Kawamura, 2004).
The outlet boundary condition with a static outlet pressure based on the cavitation number can be calculated as
given in (Lifante and Frank, 2008):
(2.6)
Where:
Pout= outlet pressure,[pa], Pv= vapour pressure, [pa], σn = rotation cavitation number
(2.7)
D =Propeller diameter, [m], N: Revolutions of propeller, [rps], P: Static pressure at point of interest, [pa]
3. PROPOSED CONCEPT DESIGN FOR MINIMUM PROPELLER INDUCED VIBRATION
In this concept design three elements were identified as being influential in determining propeller vibratory
response. The three elements of importance are pressure fluctuation, propeller loading, and cavitation inception.
The objectives of the proposed concept design are:
• Minimize pressure fluctuation, within the neighbourhoods of the propeller flow field.
• Blades elements unloading throughout minimization pressure distribution of chordwise elements along
spanwise of the propeller diameter.
• Avoid cavitation inception which dramatically magnifying the propeller induced vibratory forces
The achievement of the three objectives in design will result in many successful propellers.
4. NON-CAVITATING PROPELLER
This study aims to analyze a family of skewed propeller of different skew angle to assess the influence of skew
in the objectives of the concept design. The selected propeller geometry is DTMB-P4119 which is a right
handed, three-bladed fixed-pitch propeller with pitch diameter ratio of 1.084 of typical diameter D=0.305 m,
the full details of geometry data for this propeller was given in (Brizzolara and Villa, 2008).
The original design of this propeller is without skew, i.e. skew angle=zero. Different propeller geometries of
the same propeller dimensions have been modeled with only difference in the skew angles. Skew angles ( S )
applied from 15:75 degrees with increment of 15 degree. The geometries of these propellers are shown in Fig.
4.1 for the simulation purposed design advance coefficient J=0,833 was selected.
408
4.1 Spanwise Pressure Distribution The pressure distribution on the blade surfaces is an important factor for blade designs, considering the
cavitation suppression and material strength issues (Daqing, 2002; Abdel-Maksoud et al.,1998)
CP=(P-Po)/ [0.5 (ND)2] (4.1)
Where:
P= Static pressure at point of interest, [Pa]
Po=Reference Pressure at infinity, [Pa]
Figures 4.2: 4.3 present comparison of spanwise pressure distribution as a term of pressure coefficient Cp
versus distance from the leading edge non-dimensionalized by the chord length (X/C) at 0.7 R, 0.9 R as
examples for the Skewed Propeller Family of DTMB-P4119.
4.2 Pressure Fluctuation
The propeller, as a main excitation source of ship vibration related problems, the predominant factor for ship
structures vibrations is pressure fluctuations.
In the present study of DTMB-P4119 family of skewed propellers the numerical results of pressure fluctuation
have been predicted. Figure 4.4 shows a direct comparison between the resultant pressure fluctuations at 0.7 R
of the studied geometries.
4.3 Influence of Skew on Tip Speed
The logic resultant consequence of skew application which play role in reducing the blade pressure loading and
fluctuation is the increase in tip speed. Figure 4.5 shows measurements of the circumferential speeds on the
propellers tips for different skew angles. The velocity measurements analysis was that the propeller's skew
angle has only an insignificant influence on the mean values of the tip flow velocity
4.4 Discussion of non cavitating propeller Results
Application of propeller skew has been shown to be effective in reducing in blade loading along the spanwise
of propeller diameter. This reduction can be easily investigated along the applied skew range of 0:60 degree
(Figures 4.2: 4.3) Skew of 75 degrees results in increase of the negative pressure i.e. the propeller back in the
tip region at 0.9 R.
Concerning the pressure fluctuation the increase of propeller skew almost improve the pressure fluctuation in
the propeller flow field neighborhoods (Figure 4.4).
The increased velocity as a direct consequence of pressure reduction has been also investigated. Fig. 4.5 shows
the slight increase in the propeller tip speed which might be negligible.
Finally, based on the aforementioned analysis, a moderate skew range of 45:60 degree is recommended from
hydrodynamic and vibration points of view.
409
.deg0.0S .deg15S
.deg30S .deg45S
Fig. 4.1 Skewed Propeller Family of DTMB-P4119
Fig.4.2 Chordwise Distribution of pressure coefficient for DTMB-P4119 Skewed Family at 0.7 R, J=0.833
.deg60S .deg75S
410
Fig.4.3 Chordwise Distribution of pressure coefficient for DTMB-P4119 Skewed Family at 0.9 R, J=0.833
Fig.4.4 Pressure Fluctuation of DTMB-P4119 Skewed Family at 0.7 R, J=0.833 (cont.)
411
Fig.4.4 Pressure Fluctuation of DTMB-P4119 Skewed Family at 0.7 R, J=0.833 (cont.)
Fig.4.5 Circumferential Tip Speeds of DTMB-P4119 Skewed Family, J=0.833
412
5. Cavitating Propeller The purpose of this study is to examine the proposed concept design and the criteria of skew application for
attest case of cavitating propeller with different number of blades. The propeller model selected for the present
study is INSEAN (Italian Ship Model Basin) E779A which is a four blade propeller, 4.5 degree skewed, with a
uniform pitch (pitch/diameter = 1.1), a forward rake angle of 4° 3” and a diameter of 227.2 mm.
Three other geometries have been also modelled by skew angles of 45, 60, and 75 degrees to apply and
examine the proposed concept design. This was to build a family of skewed E779A propeller. Figure 5.1 shows
these geometries.
.deg5.4S .deg45S .deg60S .deg75S
Fig. 5.1 Skewed Propeller Family of E779A
Fig.4.5 Circumferential Tip Speeds of DTMB-P4119 Skewed Family, J=0.833 (cont.)
413
For the simulation purposes, the following operating condition is considered: Uniform flow at speed V = 5.808
m/s and propeller rotational speed n = 36.0 rps, (advance coefficient J = 0.71); cavitating number of σn = 1.763
(Salvatore and Streckwall, 2009).
5.1 Cavitating Flow
Cavitating flow condition is simulated at the design advance coefficient and cavitation number is presented in
the following. Figure 5.2, 5.3 compares the predicted extensions of cavitating regions on the propeller face and
back.
.deg5.4S .deg45S .deg60S .deg75S
Fig.5.2 Back Cavitation on Skewed Propeller Family of E779A
.deg5.4S .deg45S .deg60S .deg75S
Fig.5.3 Face Cavitation on Skewed Propeller Family of E779A
5.2 Spanwise Pressure Distribution
Figures 5.4: 5.6 show a chordwise distribution of cavitating pressure at 0.6, 0.7, and 0.9 R for the purpose of
comparison of the application of the proposed concept design for the Skewed Propeller Family of E779A
propeller.
Fig. 5.4 Chordwise Distribution of pressure coefficient for E779A Skewed Family at 0.6 R, J=0.71
414
5.3. Pressure Fluctuation
Figure 5.7 shows the pressure fluctuation at 0.7 R for the family of E779A skewed propellers.
Fig. 5.5 Chordwise Distribution of pressure coefficient for E779A Skewed Family at 0.7 R, J=0.71
Fig. 5.6 Chordwise Distribution of pressure coefficient for E779A Skewed Family at 0.9 R, J=0.71
Fig. 5.7 Pressure Fluctuation of E779A Skewed Family at 0.7 R, J=0.71
415
5.4. Influence of skew on tip Speed For the cavitated propeller test case Figure 5.8 shows measurements of the circumferential speeds on the
propellers tips for different skew angles. As shown in the figures, the increase in the mean value of tip flow
velocity is small and can be also negligible.
5.5. Discussion of cavitating propeller Results
The analysis of non-cavitating propeller results came with a recommended beneficial skew range of 45:60
degree. The efficiency of this range examined for a cavitating propeller and has shown success in the objectives
of the concept design. The recommended skew range decreases the propeller blade elements loading along the
propeller diameter (Figures 5.4: 5.6). This blade unloading reduces the cavity volumes developed on the
propeller back (Figures 5.2, 5.3). The sheet cavitation developed on the original design of the propeller model
with 4.5 degree skew has been transferred to only slight tip cavitation by 60 degree skew. While 75 degree
skew which exceeds the recommended range results in excessive negative pressures on the propeller back, and
reproduced higher cavity volume on the propeller tip region. The pressure fluctuation also decreased by
implementing the proposed skew (Figure 5.7). Regarding the effect of skew on the tip speed slight increment
has been visualized (Figure 5.8).
Fig. 5.8 Circumferential Tip Speeds of E779A Skewed Family, J=0.71
9. Conclusions
Proposed propeller skew has a beneficial effect in reduction of the pressure fluctuation and blade hydrodynamic
unloading, moreover achieving higher margin against cavitation inception. Increase in skew will not always
416
reduce the propeller vibratory forces. Excessive skew can result in higher hydrodynamic loading of blades of
the negative pressure near tip regions. Regarding cavitation inception, it is concluded that a rise in the propeller
skew has been resulted in reproducing of cavitation volumes on the propeller reloaded blade elements. Form
the vibration point of view cavitation inception is dramatically magnifies the induced vibratory effects. The aim
should be a good balance and combination with propeller loading, and finding the optimum solution for
successful designs. The velocity measurements analysis was that the propeller's applied skew angles have only
an insignificant influence on the mean values of the tip flow velocity. To summarize, a moderate skew of 45:60
degree is proposed from a hydrodynamic and vibration points of view.
References: Francesco Salvatore, F., Streckwall, H., “Propeller Cavitation Modelling by CFD -
Results from the VIRTUE 2008 Rome Workshop,” First International Symposium on Marine Propulsors smp’09, Trondheim, Norway, June 2009
Tom, J.C., Terwisga, Van. “Cavitation Research on Ship Propellers- A Review of Achievements And Challenges,” Sixth International Symposium on Cavitation
CAV2006, Wageningen, The Netherlands, September 2006
Lifante,C., Frank, T., “Investigation of Pressure Fluctuations Caused by Turbulent And Cavitating Flow Around A P1356 Ship Propeller,” ANSYS Germany GmbH, Otterfing, Germany, NAFEMS Seminar: Wiesbaden, Germany, 2008
Krasilnikov, V., Jiaying, S., “CFD Investigation in Scale Effect on Propellers with Different Magnitude of Skew in Turbulent Flow,” First International Symposium on Marine Propulsors smp’09, Trondheim, Norway, June 2009
Shin, R, Kawamura, T. “Propeller Cavitation Study Using an Unstructured Grid Based Navier-Stokes Solver,” ASME Journal of Fluids Engineering, September 2004
Brizzolara, S., Villa, D., “A systematic comparison between RANS and Panel Methods for Propeller Analysis,” 8th International Conference on Hydrodynamics, Ecole Centrale, Nantes, 2008
Li Daqing (2002),“Validation of RANS Predictions of Open Water Performance of A Highly Skewed Propeller with Experiments,” Proceedings of the Conference of Global Chinese Scholars on Hydrodynamics, Vol 18, Issue 3, July 2006, Pages 520-528
Abdel-Maksoud, M., Menter, and F., Wuttke, H., “Viscous Flow Simulations for Conventional and High-Skew Marine Propellers”, Ship Technology Research, Vol. 45, No. 2, 1998.
417
Investigation of the Influence of Skew and Rake into the Propeller
Performance and Cavitation Serkan Ekinci, Fahri Çelik, Yasemin Arikan Dept. of Naval Architecture and Marine Engineering, Yıldız Technical University, Turkey, [email protected], [email protected], [email protected] Abstract
In this study a numerical method for the investigation of the influence of skew and rake values of the
blade sections of marine propellers over the cavitation and performance characteristics is presented.
The propeller blade surface is presented with source and vortex singularities. For this purpose a
computer code based on lifting surface theory is used.
This method is applied to a conventional propeller (Seiun-maru CP) in full scale and operating in non-
uniform wake by giving skew and rake for different angles.
The influence of this two geometric parameters are investigated over the propeller thrust, torque and
efficiency values and cavitation characteristics.
Keywords Propeller design, Propeller performance, Cavitation
1. Introduction In the design of a ship the most important element which influence the ship performance is the
selection of a wake adapted, non-cavitating and correctly positioned optimal propeller.
With an optimal propeller, a safer cruise is provided since effects like cavitation induced noise,
vibration, erosion and performance loss can be reduced to a minimum. For this reason it will not be
necessary to change neither the propeller nor the rudder during certain periods, also a significant
reduction in the first design and maintenance cost can be provided. On the other hand the fluid
medium in which the propeller is operating has a complex structure, a convenient representation of the
flow for the design and analysis of the optimal propeller and the realization of the calculations in this
direction is necessary. To date, both numerical and experimental studies are made on this topics and
still continue to be made.
Nowadays parallel to the developments in the computer technology, there exists several methods for
the propeller design and analysis based on the circulation theory (Lifting line, lifting surface, boundary
element methods, etc.) (Lee and Kinnas, 2005, Ghassemi, H., 2009, Bal et.al., 2009, Çelik et.al.,
2010). In the last decade three new improvements which have an impact in these topics has come out.
These are CFD (RANS Solvers) methods, high-speed camera techniques and PIV techniques (Kulczyk
418
et.al., 2007, Arazgaldi et. al.,2009, Kuiper, 2010, Szantyr, 1994, Bal, 2008, Bal, 2010). With the skew
given to propeller blade sections, especially the vibration and the noise occured due to cavitation can
be reduced. Because skewness, provides the entrance of the radial sections of the propeller blades in
the wake field in a smooth and gradual manner, and as a result of propeller-hull interaction the radial
variation of the blades in the water are fullfilled. On the other hand rake is applied to obtain the
effective propeller diameter and can be applied also against ventilation, efficiency lost in the aft,
formation of vibration.
In this study the influence of rake and skew over the cavitation form and propeller characteristics are
investigated for Seiun Maru CP with a computer code based on lifting surface method.
2. Lifting surface analysis method In this study a lifting surface method based on Szantyr’s work is developed to predict the performance
and cavitation regions on the blade of propellers operating in a non-uniform and uniform velocity field
(Szantyr, 1994). In this method, the hydrodynamic loading on the propeller blades is replaced by
appropriate distribution of vorticity while the thickness of the propeller blades is modeled by the
appropriate distribution of sources and sinks. These singularities are distributed on the surfaces built
up by the meanlines of the propeller blade sections. Similarly, cavitation on the blade surfaces is
modeled by the appropriate distribution of sources. This phenomenon is regarded as a continuous
process of deformation of the original geometry of the blade. The kinematics boundary condition is
the basis for the formulation of the lifting surface equation. This condition requires that the resultant
relative velocity of flow at the lifting surface should be tangent to this surface. Detailed descriptions of
this method can be found in (Szantyr, 1994).
0)..()1
()(
)1
(..)1
(..)1
(..4
1
nRVpS
dSprnpcqpq
pvS psSdS
prpondSprpvn
pSdS
prpn
where
n , unit length vector normal to the lifting surface
p , vorticity distribution on the propeller blades
pv , vorticity distribution in the variable zone of the propeller free vortex system
po
, vorticity distribution in the steady zone of the propeller free vortex system
pq , source distribution modelling the propeller blade thickness
pcq , source distribution modelling the quasi-steady sheet cavity thickness on the propeller blades
rp, distance between the area element dS and the point of calculation
419
Sp, area of the propellers blades Spv, area of the variable zone of the propeller free vortex system Sps, area of the variable zone of the propeller free vortex system
V
, inflow velocity
, angular velocity of propeller rotation
,R radius at which the point of calculation is located.
3. Numerical Application The Seiun-maru CP propeller is used in this study. The operating conditions of the propeller are given
in Table 1 while geometry of propeller is given in Table 2. The non-uniform wake distribution of this
propeller is given in Fig. 1.
Fig. 1. Nominal wake distribution at propeller disc for Seiun-maru CP
Table 1. The operating conditions of the Seiun-maru CP
For the analysis of different propeller geometries; φ = 0°, 15°, 30° and 45° skew angles and ε = 0°, 5°,
10° and 15° rake angles are given to the Seiun-maru propeller. The new skew quantities and rake
quantities are given according to the non-dimensional radius (r/R) for the Seiun-maru propeller in
Table 3 and Table 4, respectively.
Delivered power, PD (kW) 360
Design speed, VS (Knot) 9
Rate of rotation, RPM 90.7
Propeller diameter, D (m) 3.6
Mean wake fraction 0.24
Number of blades, Z 5
Direction of rotation Right handed
420
Table 2. Geometry of original Seiun-maru CP propeller
Non-dimensional radius (r/R)
Pitch ratio (P/D)
Chord distribution(m) Skew (m) Rake (m)
0.2 0.95 0.7060 -0.0560 0.0378
0.3 0.95 0.8240 -0.0570 0.0568
0.4 0.95 0.9240 -0.0510 0.0757
0.5 0.95 1.0010 -0.0385 0.0946
0.6 0.95 1.0510 -0.0175 0.1135
0.7 0.95 1.0570 0.0165 0.1324 0.8 0.95 0.9860 0.0630 0.1514 0.9 0.95 0.7940 0.1180 0.1703
0.95 0.95 0.6000 0.1540 0.1797 1.0 0.95 0.0000 0.1830 0.1892
Here:
D: Propeller diameter R: Propeller radius P: Pitch r: Coordinates of propeller sections along the radius C: Chord length in each section
Table 3. Amounts of new skew values of the Seiun-maru propeller
Table 4. Amounts of new rake values of the Seiun-maru propeller
Fig. 2. View of the Seiun-maru CP propeller for different skew angles
r/R 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.95 1 φ=0° 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 φ=15° 0.096 0.145 0.193 0.241 0.289 0.338 0.386 0.434 0.458 0.482 φ=30° 0.208 0.312 0.416 0.520 0.624 0.727 0.831 0.935 0.987 1.039 φ=45° 0.360 0.540 0.720 0.900 1.080 1.260 1.440 1.620 1.710 1.800
r/R 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.95 1 ε=0° 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 ε=5° 0.031 0.047 0.063 0.079 0.094 0.110 0.126 0.142 0.150 0.157 ε=10° 0.063 0.095 0.127 0.159 0.190 0.222 0.254 0.286 0.302 0.317 ε=15° 0.096 0.145 0.193 0.241 0.289 0.338 0.386 0.434 0.458 0.482
=0 =15
=30
=45
=0 =15 =30 =45
421
Fig. 3. Definition of rake for the Seiun-maru CP propeller
For the obtained new geometries all the calculations are made for four different advance coefficients
(J=0.225, 0.451, 0.677 ve 0.903). For four different rake and skew angles, the performance
characteristics (thrust coefficient (KT) ,torque coefficient (KQ) and efficiency (ηο)) of the Seiun maru
propeller are given in Fig. 4. The sheet cavitation regions formed in the discussed operating conditions
are given in Fig. 5.
(a)
(b)
Rake=0°
00.10.20.30.40.50.60.70.8
0 0.2 0.4 0.6 0.8 1
J
K T,10K
Q,ηο
Skew=0
Skew=15
Skew=30
Skew=45
10KQ
KT
ηο
Rake=5°
00.10.20.30.40.50.60.70.8
0 0.2 0.4 0.6 0.8 1
J
K T,10K
Q,ηο
Skew=0
Skew=15
Skew=30
Skew=45 10KQ
KT
ηο
Rake
422
(c)
(d)
Fig. 4. Propeller characteristics of the Seiun maru propeller for different rake and skew angles.
Fig. 5. Sheet cavitation regions of the Seiun maru propeller for different operating conditions
Rake=10°
00.10.20.30.40.50.60.70.8
0 0.2 0.4 0.6 0.8 1
J
K T,1
0KQ,η
ο
Skew=0
Skew=15
Skew=30
Skew=45 10KQ
KT
ηο
Rake=15°
00.10.20.30.40.50.60.70.8
0 0.2 0.4 0.6 0.8 1
J
K T,1
0KQ,η
ο
Skew=0
Skew=15
Skew=30
Skew=4510KQ
KT
ηο
-0.20
0.00
0.20
0.40
0.600.80
1.00
1.20
1.40
1.60
0.2 0.4 0.6 0.8 1
J
Shee
t cav
itatio
n ar
ea x
10(m
2 )
Rake=0
Rake=5
Rake=10
Rake=15
Skew =0°
Skew =15°
Skew =30°
Skew =45°
423
Fig. 6. Efficiency values for different rake angles in J=0.903 operating condition
Conclusions In this study an approach for the propeller analysis based on the lifting surface method is used. The Seiun-maru propeller is selected as a sample application. The results obtained can be summarized as follows: 1) It is observed that the efficiency of the propeller increases in all conditions about %2 while the rake angle is constant and the skew angle increases. The highest efficiency is obtained in the J=0.903 condition as ηο =0.672978 while rake=0° and skew=45° (Fig. 6). 2) It is observed that the efficiency shows a slightly decrease with growing rake angle by constant skew angle (approximately %0.195). 3) As expected while the rake angle is constant, with the increase of the skew angle the sheet cavitation regions of the propeller for all conditions are decreasing (Fig. 5). The lowest cavitation regions are obtained while rake=0° and skew=45°. As a result of 4×4×4=64 analysis the best convenient geometry is obtained in terms of performance and cavitation for rake=0° and skew=45°.
References
Arazgaldi R., Hajilouy, A. and Farhanieh, B., 2009, Experimental and Numerical Investigation of Marine Propeller Cavitation, Scientia Iranica Transaction B-Mechanical Engineering, Vol.16 No: 6 pp: 525-533. Bal, S., 2008, Numerical investigation of cavitating marine propellers, Naval Architecture and Marine Technology Technical Conference, Vol.1, pp.239-249 (In Turkish) Bal, S. and Güner, M., 2009, Performance of Podded Propulsors, Ocean Engineering Vol.36 No: 8, pp: 556-563. Bal, S., 2010 Hydrodynamics performance of marine propellers, Shipyard, Journal of Shipbuilding and Sub-Industry, Vol.17, pp.32-36 (In Turkish) Celik, F., Güner, M. and Ekinci, S., 2010, An Approach to the Design of Ducted Propeller, Scientia Iranica Transaction B-Mechanical Engineering, Vol.17 No: 5 pp: 406-417. Ghassemi, H., 2009, The Effect of Wake Flow and Skew Angle on the Ship Propeller Performance, Scientia Iranica Transaction B-Mechanical Engineering, Vol.16 No: 2 pp: 149-158. Kuiper, G., 2010, New developments and Propeller Design, 9th International Conference on Hydrodynamics, Shanghai. Kulczyk, J. S.,Skraburski, L. and Zawislak, M., 2007, Analysis of Screw Propeller DTMB 4119 using the FLUENT System, Archives of Civil and Mechanical Engineering, Vol:7, No:4, pp:129-136. Lee, H. and Kinnas, S.A., 2005, A BEM fort he Modeling of Unsteady Propeller Sheet Cavitation Inside a Cavitation Tunnel, Journal of Computational Mechanics, Vol.37, No: , pp:41-51. Szantyr, J.A., 1994, A method for analysis of cavitating marine propellers in non-uniform flow, International Shipbuilding Progress, Vol.41, No.427, pp.223-242.
J=0.903
0.6
0.62
0.64
0.66
0.68
0 5 10 15
Rake (degree)
ηο
425
*ĐTÜ, Faculty of Naval Arch. and Ocean Eng., Istanbul, Turkey, [email protected] **TUBITAK BILGEM Information Technologies Institute, Kocaeli, Turkey, [email protected]
426
considerably. In summary, an accurate prediction of propeller cavitation in a non uniform wake field can
provide a good basis upon which to predict overall propeller noise.
In another approach, cavitation analysis on propeller blades together with the prediction of the total
hydrodynamic performance may be obtained by using a verified lifting surface algorithm (Szantyr, 1994).
The approach used in this study is the model propeller concept. It can be conceptualized that the model
propeller is to be tested in a cavitation tunnel. Therefore, hydrodynamic performance as well as the
cavitation patterns may be obtained using a lifting surface algorithm. Using a semi empirical model, the
broadband noise spectrum of the model propeller may be calculated. Adjustments may then be applied to
the broadband noise spectrum to scale the results up to that of the full size propeller.
The broadband noise spectrum thus obtained represents the steady state noise spectrum generated by the
propeller. To impart an extra level of realism into the model, the broadband noise spectrum is then
modulated with a blade pattern in the time domain. Sometimes this is also referred to as the "DEMON"
component.
2. Formulation of the Problem
Direct calculation of the broadband noise is a complex procedure. The use of appropriate statistical
methods is regarded to be more efficient rather than the direct calculation. In the current work, a semi
empirical model similar to Brown’s equation is developed for the prediction of sound pressure level Ls
(dB) (Carlton, 2007),
[ ]4 3
2163 10 10 40 10TipCPs Tip Disti
D Tip
VAZD nL log log K log log Hf A V
= + + + +
(1)
Where
Z: Number of Blades D (m): Diameter of propeller np (RPM) : Propeller rate of revolution AC (m2): Mean sheet cavitation area on propeller blade AD (m2): Propeller disk area VTip (m/s): Propeller tip speed
iTipV (RPM): rotation rate (RPM) of the start of tip vortex
KTip = 60 but for deeply submerged propellers (e.g. submarine) 80 (Odabaı, 1987) HDist (m): Hydrophone placement distance Equation (1) is valid for fp<10 kHz where the center frequency fP lies at the peak of the broadband noise
spectrum. In the present study, fp is determined using the formulae 3.2 2 1 21 3
2.0 6 1 21 3
4400 1.722
1100 1.722
i in s n
pn n
i in s n
pn n
PfD
PfD
σ σσ σ
σ σσ σ
−
−
= => <
= => ≥
(2)
427
which are commonly utilized in connection with pump cavitation (Okamura and Asano, 1988). Here, Ps is
the static pressure (lbs/in2), and nσ is the cavitation number which is defined as
2 212
s Vn
P
P P
n Dσ
ρ
−=
(3)
where PV is the vapor pressure of water (1700–2400 Pa) and n
iσ is the incipient cavitation number derived from
21 2 1
1i i
C n n
D n n
AA
σ σσ σ
− = −
(4)
Equation (1) determines the sound pressure level at the peak frequency where f = fp. Otherwise, the sound
pressure level is obtained by
−
= => <
= => > (5)
Here, the constants A and B are determined from the continuity characteristics of the noise spectrum and
they are given by 0.007 0.2
;p ps s
p p
L LA B
f f
−
= =
(6)
The sound pressure level ps
L is obtained from the use of Equation (1) for the peak frequency fp. Hence,
Equation (5) together with Equation (1) yields a broadband noise spectrum over the frequency range of
interest.
The present work is carried out assuming scale model propellers compatible with the dimensions of
Emerson Cavitation Tunnel at University of Newcastle. The goal is to predict the noise spectrum of a 30
cm diameter propeller. Under these conditions the hydrophone distance HDist from model is taken to be
0.435 m.
Later, an approximation to the fullscale noise levels is carried out using the scaling laws recommended
by the Cavitation Committee in of (ITTC, 1987). The increase in the noise level in moving from model to
full scale is given by, 2 2
( ) 20 logz x y y y
P M P P P PP
M P M M M M
D r n DL dBD r n D
σ ρσ ρ
=
(7)
and the frequency shift is expressed as (see: Carlton, 2007)
P P
M M
f nf n
= (8)
In the above equations, the subscripts P and M refer to the fullscale and the model scale respectively, r is
the reference distance at which the noise level is predicted, prescribed as 1 m for both the fullscale and
the model scale calculations. σ is the cavitation number taken to be the same value for both the fullscale
and model, n is the propeller rate of rotation and ρ is the mass density of water assumed to be 1000 kg/m3
for the scale model in the cavitation tunnel conditions and 1025.9 kg/m3 for the full scale propeller in the
428
sea water conditions. Furthermore, setting y=2 and z=1, the expression for the increase in the noise level
reduces to 2
3( ) 20 log 1025.9 PP
M
nL dBn
λ =
(9)
The parameter λ is the scale ratio between the model (30 cm diameter) and the fullscale propellers.
3. Propeller DEMON Component
Once a broadband noise model has been established for the propeller under study, a modulation model
can be developed to impart a realistic time domain signature to the broadband noise model. This is
sometimes referred to as the "DEMON" component (Kummert, 1993), (Nielsen, 1999). As the propeller
blade rotates about the shaft axis it passes through different regions of wake flow and turbulence, which
results in cavitation and associated peaks in generated noise. The cavitation noise envelope exhibits both
cyclical and random components which may be attributed to the wake flow pattern and to the turbulence
in the wake flow. This gives rise to a cavitation noise amplitude, or envelope, which continually varies as
the propeller rotates about the shaft axis (Ross, 1987). The cyclical variation of the propeller cavitation
noise envelope is modeled as a form of amplitude modulation by Lourens (Lourens and Prcez, 1998) by
(Kummert, 1993) and by (Nielsen, 1999). The authors of the cited papers utilize the amplitude
modulation model for the purpose of analyzing observed real propeller signals.
In this work, the aim is to synthesize an artificial propeller signal. In order to generate the artificial signal,
the propeller noise spectrum obtained in the previous section is used as input. The propeller noise
spectrum may be represented by the Npoint discrete frequency domain function X(k). The frequency
index k is related to actual frequency by the relation
<<⋅−≤≤⋅
=NkNNfkN
NkNfkf
s
s
2/,/)(2/0,/
(10)
X(0) corresponds to the DC (zero frequency) component of the spectrum and X(N/2) corresponds to the
Nyquist frequency fs/2. The remaining components are complex conjugates of one another such that X(N
k) = X*(k).
To simulate a natural time domain signal, a random phase is associated with the spectral value X(k)
which results in the randomized spectral value Xr(k). The random phase is generated using a uniform
random variable U on the interval (0,1).
<<−
<<
==
=
NkkNX
kekX
kkkX
kXN
r
NUi
N
r
2*
22
2
)(
0)(
,0)(
)( π
(11)
The time domain signal of the broadband noise is attained by taking the discrete Fourier transform of the
randomized frequency spectrum. In other words,
429
{ })()( kXDFTnx r= Pa
(12)
The modulator function m(t) is a sum of sinusoids and harmonics in the form of a Fourier series,
60/
)2cos()(0
Pshaft
nshaftn
nf
tfnAtm
=
=∑∞
=
π
(13)
The coefficients of the sinusoidal modulating function and its harmonics are represented by the coefficients A0, A1, ..., A∞. Coefficient A0 is the average (DC) value of the modulator function. Coefficient A1 is the magnitude of the sinusoid of the shaft turn rate frequency fshaft. The coefficients A2, A3, ..., are the magnitudes of the harmonics of the shaft frequency. The time domain broadband noise x(n) (equation 12) is modulated with the modulator function m(t) (equation 13) to produce the final output signal y(n). Given that t = n/fs the function m(t) may also be written as m(n/fs). Thus the resulting modulated propeller noise y(n) is given by
)()/()( nxfnmny s ⋅= Pa (14)
Various constraints need to be introduced to the above mentioned parameters. In order to ensure a well formed modulation, the constant coefficient A0 should be greater than the sum of the other coefficients. Furthermore, in order to ensure that the modulation does not create a change in average power level of the signal, the sum of the coefficients squared must equal unity. Thus,
∑
∑∞
=
∞
=
+=
≤<=
≥
1
2212
02
1
10
1
10,
0
nn
nn
n
AA
AA
A
αα
(15)
The coefficient α controls the modulation level, which adjusts the variation in amplitude about the mean
level.
The determination of the harmonic coefficients Ak is relatively more complex. The ship propeller noise
typically exhibits a strong harmonic at the blade frequency fbr defined as
,...3,2,160
=
=
k
HzZnkf Pbr (16)
Hence the value for the coefficient AZ is typically greater than the other coefficients. 4. Numerical Example
The developed model is applied to a four bladed propeller whose principal dimensions are provided in
Table 1. Table 1. Main Particulars of the Propeller
Number of Blades 4 Boss Ratio 0.276 Propeller Diameter (m) 2.100 Rake 0 Pitch Ratio at 0.7R 0.8464 Skewback (degrees) 40 Expanded Blade Area Ratio 0.55 Direction of Rotation Right Handed
430
The offsets of the blade sections and the hub as well as the details of the trailing and leading edges of the
blades are obtained from (Takinaci et al., 2000).
The 3D representation of the propeller is provided in Figure 1. The nonuniform wake field in which the
propeller operates is shown in the form of a velocity ratio contour plot in Figure 2.
Figure 1. 3D representation of propeller. Figure 2. Non uniform wake field in terms of velocity ratios.
The sample test condition based upon the cavitation tunnel test obtained from (Takinaci et al., 2000) are
outlined in Table 2. Table 2. Fullscale and corresponding test conditions for noise modeling.
Ship’s Speed (knots) σn nM Tunnel Speed
v (m/s) 10.0 5.0175 978 3.35
The predicted and measured sound pressure spectrum levels for the test condition in Table 2 are presented
in Figures 3 and 5 for the model and in Figures 4 and 6 for the fullscale propeller. In these figures, the
logarithmically scaled xaxis represents the center frequencies (f) in Hz while the linearly scaled yaxis
represents the sound pressure levels (Ls) in dB re 1 Pa, 1 Hz, 1 m. A common practice in the analysis
and presentation of the noise levels is to reduce the values of Sound Pressure Levels – Ls (SPL) in each
1/3 Octave band to an equivalent 1 Hz bandwidth.
Figures 3 through 6 compare the experimentally measured and the predicted noise levels. The curve
consisting of two relatively smooth sections is the predicted spectrum. The other curve shows the
experimentally measured spectrum. The predictions are in fairly good agreement with the experimental data.
Having established a broadband noise spectrum prediction, the next step is obtaining an audio model by
modulating the broadband noise spectrum. The generated propeller broadband noise for Condition in
Table 2 is modulated by the procedure described previously using the parameters listed in Table 3.
Table 3. Propeller Noise Modulator Harmonic Input Values
nP 197 A0 0.986295 A3 0.051330 A6 0.051330 Z 4 A1 0.108666 A4 0.179160 α 0.5 A2 0.051330 A5 0.051330
431
The determination of the harmonic coefficients Ak of the modulator function is not straightforward, and is
there is a lack of empirical study of the parameters associated with amplitude modulated propeller noise
in the literature (Nielsen, 1999). Nevertheless as a general principle, for a Z blade propeller, the Zth
harmonic component is expected to be dominant. Typical to commercial propellers, one blade exhibits
significantly higher cavitation than the other blades resulting in a rhythmic pattern which may be detected
by ear or on a DEMON graph. Thus the first harmonic is also expected to be dominant relative to the
others. Based on these heuristics, and in consultation with sonar personnel, a set of parameters was
chosen which reasonably simulated both the auditory sound effects of, as well as the expected output of
DEMON analysis, of a civilian cargo vessel propeller.
The time domain version x(n) of the spectrum from Figure 4 is attained using Equation 12, and the result
is shown in Figure 7. The graph of x(n) is limited to a one second period for the purpose of clarity.
The modulator function m(t) defined in Equation 14 is used to modulate the broadband noise signal in order to
simulate the propeller noise. The parameters in Table 3 are used in Equation 13. The resulting function m(t) is
shown in Figure 8.
By modulating the broadband noise x(n) with the modulator function m(t) as shown in Equation 14 the output
propeller signal y(n) is obtained. As can be seen in Figure 9, the amplitude of the original broadband noise x(n)
varies with the peaks introduced by the modulator function m(t).
Figure 9: Propeller Modulated Noise
Figure 7: Broadband Noise (Time Domain Version of Spectrum) Figure 8: Modulator Function
432
It is concluded that the empirical prediction of broadband noise followed by a modulation technique as
outlined in this work may be used to generate a realistic time series audio signal. The full audio signal in WAV
or MP3 format may be downloaded from the web site (Web 1).
5. Conclusions
An empirical prediction model of broadband noise for marine propellers is developed. The model is composed
of two components: firstly, the empirical prediction of the frequency domain broadband noise, and secondly,
modulation of the noise in the time domain.
The results of the empirical prediction model are seen to be in general agreement with the available
experimental data. In order to impart a realistic audio character to the spectrum thus obtained, a modulation
model is also employed. As a result the data obtained in the frequency domain is converted into an audible
output.
For the future work, it is desired to enrich the empirical prediction algorithm and the modulation parameters
further with more empirically gathered data.
6. References
Atlar, M., Takinaci, A.C., Korkut, E., Sasaski, N. and Aono, T., Cavitation Tunnel Tests for Propeller Noise of a FRV and Comparisons with FullScale Measurements, 4th Int. Symp.on Cavitation, Pasadena, USA, 2023.07.2001.
Carlton, J. S., Marine Propellers and Propulsion, 2nd ed., Butterworth Heinemann, 2007.
Hazelwood, R. A. and Conelly, J., “Estimation of Underwater Noise A Simplified Method,” Int. J. Soc. For Underwater Tech., vol. 26, no. 3, pp. 97103, 2005.
ITTC, Cavitation Committee Report, 18th International Towing Tank Conference, Kobe, Japan, 1987.
Kehr, Y. Z. and Kao, J. H., “Numerical Prediction of the Blade Rate Noise Induced by Marine Propellers,” Journal of Ship Research, vol. 48, no. 1, pp. 114, March 2004.
Kummert, A., Fuzzy Technology Implemented in Sonar Systems, IEEE J.Oceanic Eng., vol.18, no.4, pp.483490, 1993.
Lourens, J. G. and du Prcez, J. A., “Passive Sonar ML Estimator for Ship Propeller Speed,” IEEE J. Oceanic Eng., vol. 23, no. 4, pp. 448453, Oct 1998.
Nielsen, R. O., “CramerRao Lower Bounds for Sonar BroadBand Modulation Parameters,” IEEE J. Oceanic Eng., vol. 24, no. 3, pp. 285290. July 1999.
Odabaı, A. Y., “Cavitation Inception and Prediction of BroadBand Noise Levels,” British Maritime Technology, Tech. Rep. W1607, March, 1987.
Okamura, N. and Asano, T., “Prediction of Propeller Cavitation Noise and Its Comparison with FullScale Measurements,” J.S.N.A. Japan, vol. 164, 1988.
Ross, D., Mechanics of Underwater Noise, Peninsula Publishing, 1987, pp. 253285.
Szantyr, J. A., “A Method for Analysis of Cavitating Marine Propellers in Nonuniform Flow,” International Shipbuilding Progress, vol. 41, no. 427, pp. 223–242, 1994.
Takinaci,A.C., Korkut, E.,Atlar, M.,Glover, E.J.,Paterson, I.,“Cavitation Observation and Noise Measurements with Model Propeller of a Fisheries Research Vessel,” Dept. of Marine Tech., Univ.of Newcastle, Rep. MT200056, 2000.
Web 1, Takinaci, A. C.. Propeller Noise Simulation. [Online]. Available : http://www.gidb.itu.edu.tr/staff/ takinaci/ noisesimulation/propnoisesimulation.html
Yoshimura, Y. and Koyanagi, Y., “Design of a Small Fisheries Research Vessel with Low Level of UnderwaterRadiated Noise,” Journal of the Marine Acoustics Society of Japan, vol. 31, no. 3, pp. 137145. 2004.
433
* *** ** **The flow around a marine propeller is one of the most challenging hydrodynamics problems.
Computational fluid dynamics (CFD) has emerged as a potential tool in recent years and has
promising applications. The goal of this paper is to provide complete guidelines for geometry
creation, boundary conditions setup, and solution parameters of the flow around rotating propeller.
These guidelines are addressed to handle propeller simulation problems in order to achieve quick,
more accurate solution with less computational cost. In this paper CFD results for flow around a
marine propeller are presented. Computations were performed for various advance ratios following
experimental conditions. ReynoldsAveraged NavierStokes (RANS) method combined with an
extensive validation of two different turbulence models k–ε and k–ω was applied for the flow
simulation. The computations enable direct comparison of the reliable CFD results with the
experimental data.
Propeller flow, CFD simulation, RANS
The flow around the propeller is complex due to its geometry and the combined rotation and
advancement into water. The availability of numerical techniques and low cost highspeed
computational capability has made a major impact on the analysis of complex flows. Computational
fluid dynamics with regard to these specific applications is still in a process of evolution, as is
evidenced from specific validation studies in current research literature.
RANS computations offer that possibility, and such viscousflow computations start to be used in the
practical ship design; but how accurate these predictions are, and to what extent they depend on the
turbulence modelling used, is not really known (Tomasz and Paweł, 2010). Despite of the great
advancement in the CFD technologies and feasibilities of the approaches for marine propeller flows,
some issues need to be addressed for more practicable procedures. The complexity in geometry, mesh
generation and turbulence modelling are the main obstacles. In fact a marine propeller is a very
complex geometry, with variable section profiles, chord lengths and pitch angles, and in operational
conditions it induces rotating flow and entails tip vortex (Sileo et al., 2006).
In order to verify the reliability of the CFD simulations, the flow about a propeller model DTMB
P4119 propeller is investigated, and the results are compared with the existing available experimental
results.
********The flowaroundamarinepropeller is oneof themost challenginghydrodynamicsproblems.Computational
fluiddynamics(CFD)hasemergedasapotentialtoolinrecentyearsandhaspromisingapplications.Thegoal
ofthispaperistoprovidecompleteguidelinesforgeometrycreation,boundaryconditionssetup,andsolution
parametersoftheflowaroundrotatingpropeller.Theseguidelinesareaddressedtohandlepropellersimulation
problems in order to achieve quick,more accurate solutionwith less computational cost. In this paperCFD
results for flow around amarine propeller are presented.Computationswere performed for various advance
ratiosfollowingexperimentalconditions.ReynoldsAveragedNavierStokes(RANS)methodcombinedwithan
extensivevalidationoftwodifferentturbulencemodelsk–εandk–ωwasappliedfortheflowsimulation.The
computationsenabledirectcomparisonofthereliableCFDresultswiththeexperimentaldata.
The flow around the propeller is complex due to its geometry and the combined rotation and
advancement into water. The availability of numerical techniques and low cost highspeed
computationalcapabilityhasmadeamajorimpactontheanalysisofcomplexflows.Computational
fluid dynamics with regard to these specific applications is still in a process of evolution, as is
evidencedfromspecificvalidationstudiesincurrentresearchliterature.
RANScomputationsofferthatpossibility,andsuchviscousflowcomputationsstarttobeusedinthe
practicalshipdesign;buthowaccuratethesepredictionsare,andtowhatextenttheydependonthe
turbulence modelling used, is not really known (Tomasz and Paweł, 2010). Despite of the great
advancementintheCFDtechnologiesandfeasibilitiesoftheapproachesformarinepropellerflows,
someissuesneedtobeaddressedformorepracticableprocedures.Thecomplexityingeometry,mesh
generation and turbulence modelling are the main obstacles. In fact a marine propeller is a very
complexgeometry,withvariablesectionprofiles,chordlengthsandpitchangles,andinoperational
conditionsitinducesrotatingflowandentailstipvortex(Sileoetal.,2006).
Inorder toverify the reliabilityof theCFDsimulations, the flowaboutapropellermodelDTMB
P4119propellerisinvestigated,andtheresultsarecomparedwiththeexistingavailableexperimental
results.
Numerical studieswere carried outusing computational fluiddynamics (FluentR12) to obtain the
open water characteristics of propeller model as well as the distribution of pressure on the blade
434
surface.Thiseffortinvolvesstandardizationofthecomputationalgriddomainwithproperchoiceof
thegridsizeandcontrolvolumearoundthepropellerunderinvestigation.
ThegeneralconservativeformoftheNavierStokesequationispresentedasthecontinuityequation
Continuityequation,
∂ρ/∂t+∂(ρui)/∂xi=Sm (2.1)
Whereρthedensityinkg/m3,uithevelocitycomponentintheithdirectioninm/s(i=1,2,3)and
Sm represemts the source terms. In case of incompressible flows the density is considered to be
constant.Since thepropellerflowhasbeenconsideredassteadyand incompressible, thecontinuity
equationgetsmodifiedas
∂(ρ)/∂=0 (2.2)
Themomentumequationwillbe,
+++−=+ ρ
∂τ∂
∂∂ρ
∂∂ρ
∂∂ )()( (2.3)
Where
δ
∂∂
∂∂
∂∂τ
32)]([ −+= (2.4)
istheReynoldsstresstensor,pstaticpressure,N/m2,gravitationalaccelerationinthethdirection,
m/s2,externalbodyforcesinthethdirectioninN,δistheKroneckerdeltaandisequaltounity
when;andzerowheni≠j.
TheReynoldsAveragedformoftheabovemomentumequationincludingtheturbulentshearstresses
isgivenby:
( ) ( ) ( )
′′−
∂∂
+∂∂
−
∂∂
−
∂
∂+
∂∂
∂∂
=∂∂
+∂∂ ρρρ
32
(2.5)
Where:istheinstantaneousvelocitycomponent,m/s(=1,2,3).
Inorder tocharacterize turbulence,additionalconservationequations (orclosureequations) forkε
andkωhavetobesolved.
Thepropellermodelconsideredinthepresentstudy isDTMBP4119designedat theDavidTaylor
Model Basin. P4119 is a threebladed fixedpitch propeller of typical diameter D=0.305 m, the
geometrydataforthispropellerwasgivenin(Sileo,2007)and(Villaetal.,2008).
Acommercial computerprogramHydroComp.PropCadused forgenerationofbladeprofiles,was
used to generate the blade profiles for the geometry of P4119 screw propeller. The program
transformsinputdataintothecoordinatesofcloudpointsinspace.Thepointsdescribetheshapeofa
propellerbladesurfaceandthentheycanbeconnectedintocurves,surfacesandavolume.Propeller
dataareimported inNURBS(NonUniformRationalBSpline)ModellerRhinoceros4.0asa third
435
partyprogramtobuildacompletesolidpropellermodel.Thebladeshavebeensimplymountedonan
infinitelylongcylinder,whichservesasthehubandshaft,toavoidthestagnationpointonthehub
closetothepropeller(Sileoetal.,2006)and(Sileo,2007).
Figures3.1,3.2showthesurfacemodelandsolidmodelrespectively.
Fig.3.1DTMBP4119PolySurfaceModelFig.3.2DTMBP4119SolidModel
A cylindrical volume is used to simulate the physical domain.
The inlet is at 1.5D upstream, the outlet at 3.5D downstream;
solidsurfaceson thepropellerbladesandhubarecentredat the
coordinate system origin and aligned with uniform inflow. The
outerboundaryisat1.5Dfromthehub.Thedomainwaschosen
depending on research performed by (Kulczyk et al. 2007) and
domain dependence studies in (Amminikutty et al., 2006).
Illustration of the schematic diagram of the propeller
computationaldomainisshowninFig.4.1.
Unstructured tetrahedral cells were generated using the FLUENT, to define the control volume
(Amminikutty et al., 2006). Figure 5.1 shows surface meshes on the propeller blade and boss
surfaces.Themesheswereusedtogeneratea3Dmeshinsidethedomainvolume;Fig.5.2showsthe
gridinthepropellerneighbourhood.
Anotherimportantparameteristhequalityofthemesh:theelementscannotbetoomuchdistorted;
otherwisetheobtainedresultswillnotbecorrect.It isbesttoassumethemaximumcellequiangle
skewbelow0.9(Kulczyketal.2007).
Fig.5.1SurfaceMeshonBladesandHubFig.5.2Gridinpropellerneighbourhood
Fig.4.1ComputationalDomain
436
Boundary conditions were set to simulate the flow around a rotating propeller in open water: A
movingreferenceframeisassignedtofluidwitharotationalspeedequaltothepropellerrpm.Wall
formingthepropellerbladeandhubwereassignedarelativerotationalspeedofzerowithrespectto
adjacentcellzone.Auniformspeedaccordingtodifferentadvancecoefficientwasprescribedatinlet.
At outlet pressure outlet boundary condition was set. The outer boundary Nonviscous wall was
assignedfortheouterdomainboundarywithrelativerotationalspeedofzerowithrespecttoadjacent
cellzone(Dunnaetal.,2010).
ThecomputationconditionswerebasedontheresearchdonebyJessup(1989)sincetheexperimental
resultshadbeentakenfromit.Therotationalspeedwassetat600rpm.Theadvancecoefficientwas
changedbychanging thevelocityof inflow.Thecomputationswereperformedusing twodifferent
turbulencemodelsk–εandk–ωTherotationalmotionofthepropellerwasmodelledbyimmobilizing
the latter and rotating the calculation domain in theopposite direction this gives exactly the same
resultsasifthepropellerwererotating(Kulczyketal.2007).Computationsforoneoperationpoint
took about 2448 hr.During this time the program computed about 2,0003,000 iterations. Solver
parametersettingsforthepropelleropenwatersimulationsincludingphysicalconstantsareshownin
Table7.1.
Table7.1:SolverParameters
Theopenwatercalculationwascarriedoutatthesamerunningconditionsasusedintheexperimental
setupat(Jessup,1989)thecalculationwascarriedoutforadvancecoefficientsintherangefrom0.5
to1.1,similartotheexperiment.Thepressurefieldonthebladesshowslowpressureonthesuction
side; the back of the propeller and high pressure on the pressure side; the face of the propeller.
Figures8.1,8.2showthepressuredistributiononthebothsuctionandpressuresidesrespectivelyfor
thetestedpropelleratadvancecoefficientof0.5.
Parameter SettingSolver 3DSegregated,Steady,
ImplicitVelocityformulation Relativetoadjacentcellzone
Viscousmodel Standardkε,kωTurbulentmodel
Waterdensity 998.2kg/m3Gradientdiscretization GreenGaussCellBasedPressurediscretization BodyForceWeightedMomentumdiscretization FirstOrderUpwindTurbulentkineticenergydiscretization FirstOrderUpwind
Parameter SettingTurbulencedissipationrate FirstOrderUpwind
Pressurevelocitycoupling SIMPLEBladesurfaceboundarycondition Wall(noslip)
Outersurfaceboundarycondition Wall(allowsslip)
Waterinletboundarycondition VelocityInlet:Inflowatadvancespeed
Outflowboundarycondition Pressureoutlet
437
Fig.8.1PressureDistributiononsuctionsideFig.8.2PressureDistributiononpressure
The study of the flow field shows that the propeller accelerates the flow over the blades and
introduces swirl in the flow downstream of the propeller, as expected. Figures 8.3: 8.7 show the
pathlinesaroundthepropellerblades.Thesefiguresgiveconfigurationsabouthowthewhirlsorwake
formedbehindthepropelleratdifferentadvancecoefficients.
Fig.8.3PathlinesaroundpropellerbladesatJ=0.5 Fig.8.4PathlinesaroundpropellerbladesatJ=0.7
Fig.8.5PathlinesaroundpropellerbladesJ=0.833 Fig.8.6PathlinesaroundpropellerbladesatJ=0.9
Fig.8.7PathlinesaroundpropellerbladesatJ=1.1
Thethrustandtorquecoefficientsareadequatelyestimatedincomparisonwithmeasureddataforthe
range of studied advance coefficients. The computed thrust and torque on the propeller were
converted into the dimensionless thrust coefficient, torque coefficient and the efficiency was
calculated.Moreover,thestudywasperformedviatwoturbulencemodelskε,kω,andtheoutcomes
of both models were compared with the experimental results conducted by (Jessup, 1989) The
computationresultsarepresentedinFig.7.8,whichpresentsthecomputedthrust,torquecoefficients
andefficiency,withthecorrespondingexperimentaldata.
438
The curves trends with varying advance ratios are well predicted. However, CFD solutions over
predict, and the discrepancy increases with increasing propeller load, i.e., decreasing the advance
coefficient,J.This tendencyseems tobeprevalent inall theRANSCFDsimulationsandmightbe
unavoidabledue to the experimental conditionshardly conformable inCFD, such as the effectsof
tunnelwall,inflowspeednonuniformity,andhubandbossconfigurations(RheeandJoshi,2003).
KTExp10KQExpEtaExpKTk10KQkKTk10KQkEtakEtak
ωω
ε
εε
ω
Fig.7.8OpenwatercharacteristicsofDTMBP4119propeller
Theanalysisofmarinepropellerishighlycomplexwhichhasmanyconsequences,andanalyzermust
beawarethatitisdifficultandtimeconsuming,unlessonehasareadymethodologyofcarryingout
suchinvestigations.Byapplyingtheprovidedguidelinesinthispaperquickandcorrectresultscanbe
obtained.Particular focus shouldbeplaced in themeshgeneration,boundaryconditionssetup,and
turbulencemodelling, which reveal to be crucial for the quantitative comparison of the computed
resultsandfortheefficiencyofthenumericalcalculations.
CFD results were compared with open water characteristics and found in good agreement. ;
Differences between computed and experimental results are less than 5% and 7% for Thrust
Coefficient()andTorqueCoefficient()whilethestudyconsiderskωasaturbulencemodel.On
the other side, kε turbulence model these differences became 7%, 10% for the same parameters
respectively.Thekεmodelisnotquiteappropriateforsimulatingpropellerflowbecausetheresults
inthemodelhavebeenseentobeoverpredicted.Theuseofakωmodelisdeemedsufficientfor
propellerapplications.
Tomasz B., Paweł H., “Numerical Simulation of the Flow around Ship and Rotating Propeller,” 18th InternationalConferenceonHydrodynamicsinShipDesign,Safety,andOperationHYDRONAV,Gdańsk,2010
Sileo,L.,Bonfiglioli,A.,Magi,V.,2006,“RANSEsSimulationoftheFlowpastaMarinePropellerunderDesignandOffdesign Conditions,” 14th Annual Conference Computational Fluid Dynamics Society of Canada (CFD 2006), Queen'sUniversityatKingston,Ontario,Canada,July1618.
439
Sileo, L., 2007, “Low Reynolds Number Turbulent Flow Past Thrusters Of Unmanned Underwater Vehicles,” 2ndInternationalConferenceonMarineResearchandTransportation,ICMRT,Ischia,Naples,Italy
JessupS.D.,1989,“AnExperimentalInvestigationofViscousAspectsofPropellerBladeFlow,”theCatholicUniversityofAmerica.
Villa D, Gaggero S, Brizzolara S. , 2008, “A systematic Comparison between RANS and PanelMethods for PropellerAnalysis.InternationalConferenceonHydrodynamics,Nantes,France.
Kulczyk,J.,Skraburski,Ł.,Zawiślak,M.,2007,“Analysisofscrewpropeller4119usingtheFluentsystem,”ArchivesofCivilAndMechanicalEngineeringASME,Vol.VII,No4.
Amminikutty V., Anantha V., Dhinesh G. , 2006, “Dynamic Characteristics ofMarine Hubless Screw Propellers”, 5thInternationalConference,onHighPerformanceMarineVehicles,810November,Australia
DunnaSridhar,BhanuprakashT.,DasH. ,2010,“FrictionalResistanceCalculationsonaShipusingCFD,”InternationalJournalofComputerApplicationsVolumeII–No.5
RheeS.H.,Joshi,S.,2003,“CFDValidation foraMarinePropellerUsinganUnstructuredMeshBasedRANSMethod,”ProceedingsofFEDSM'03,the4thASMEJSMEJointFluidsEngineering,ASME,SummerConference,Honolulu,Hawaii,July611,pp.17.
441
Practical Approaches for Design of 4 Bladed Wageningen B Series
Propellers Serkan Ekinci, Fahri Çelik , Yasemin Arıkan Abstract
Although there has been important developments in marine propellers since the first use in 1850, the
main concept has been conserved. Parallel to the developments in the computer technologies in the last
50 years, methods based on the circulation theory are often used for the design and analysis of
propellers. Due to the few number of design parameters and the ability to predict the propulsion
performance, the use of systematic propeller series based on open water model experiments is
widespread. The design with standard propeller series is usually made with diagrams developed by
model experiments. Reading errors during the use of these diagrams are inevitable.
In this study practical approaches are presented for preventing reading errors and time loss during the
design with standard propeller series. For the conditions that the initial design variables of propeller
(delivered power (PD), propeller rate of rotation (n) and propeller advance speed (VA)) or (propeller
thrust (T), propeller rate of rotation (n) and propeller advance speed (VA)) are given, practical and
useful methods based on diagrams and empirical formulas for the design and performance prediction
of 4 bladed Wageningen B propeller series are offered. In addition design applications for both design
conditions mentioned above are realized. The results obtained from the presented method are
compared with those of open water diagrams and a good agreement between the results is observed.
Keywords: Propeller design, propeller series, propeller performance, open water diagram
1. Introduction In ship hydrodynamics, fixed pitch propellers named also screw propellers have an important place
among the propulsion systems to propel a ship. The screw propellers which were put into practice
firstly in the middle of 19th century have kept its position as the best suitable propulsion system for
170 years. Although there has been important developments in the propeller design and the propulsion
systems in this long period of time, a change in the main concept of screw propellers was not
observed. It is seen that these propellers will be used for longer periods of time through their high
efficiency and suitable use.
The aim of the propeller design is to obtain the optimum propeller which operates with minimum
power requirements against maximum efficiency and an adequate revolution number. Usually two
methods are used in the propeller design. The first is to use diagrams obtained from open water
propeller experiments for systematic propeller series, the second is to use mathematical methods
442
(lifting line, lifting surface, vortexlattice, BEM (boundary element method)) based on circulation
theory. After 1950s, depending on the developments in the computer technology, great improvements
have been realized in the second method for the design and analysis of propellers (Lerbs, 1952,
Eckhardt and Morgan, 1955, Glover, 1974, Szantyr, 1987, Güner, 1994, Greeley and Kerwin, 1982,
Pien, 1961, Kerwin et. al., 1997, Lee and Kinnas, 2005, Ghassemi, 2009, Bal and Güner, 2009). In the
last decade, several new developments have appeared in the design and analysis of propellers; namely
CFD methods (RANS solvers), high speed camera techniques and PIV techniques (SanchezCaja, et.
al., 2001, Kulczyk et. al., 2007, Arazgaldi et. al., 2009, Çelik et. al.,2010, Tukker and Kuiper, 2004,
Kuiper, 2010).
In the first stage of propeller design generally open water diagrams of systematic propeller series are
used. These series consist of propellers of which blade number (Z), propeller blade area ratio (BAR),
pitch ratio (P/D), blade section shape and blade section thickness vary systematically. The most well
known and applied propeller series are the Wageningen (Troost) screw series. Except these, there are
also the Gawn (Froude), Japanese AU, KCA, Lindgren (Ma), NewtonRader, KCD, Gutsche and
Schroeder pitch controllable propeller, Wageningen nozzle propeller, JDCPP propeller series which
took part in the literature through various studies (Carlton, 2006).
Some studies about the propeller design and analysis methods developed recently are as following:
Tanaka and Yoshida in (Tanaka and Yoshida, 2003) have developed a computer program for propeller
designers which transforms the dimensionless tables obtained from propeller series experiments to
numerical graphics in great accuracy. In a similar study a computer program is developed by
Koronowicz et.al. in (Koronowicz et. al., 2001) for the design calculations for a propeller which is
analyzed in the real velocity environment. In the study calculations for the scale effects in the velocity
field where the propeller is operating, corrections in the velocity field due to the rudder, maximization
of the propeller efficiency, optimization of accurate blade geometry in terms of cavitation and
strength, blade geometry depending on induced pressure forces and optimization of the blade number
are made. A multipurpose optimization method is developed by Benini in (Benini, 2003) for
Wageningen B propeller series using an algorithm for maximizing both the thrust and torque
coefficients under a constraint determined according to cavitation. Unlike classical lifting line
methods, Çelik and Güner in (Çelik and Güner, 2006) suggested a improved lifting line method by
modeling the flow deformation behind the propeller with free vortex systems. In his study Olsen in
(Olsen, 2004) has developed a method to calculate the propeller efficiency with the help of energy
coefficients including the propeller loss. He has compared his findings with the results obtained from
vortexlattice method. Hsin et.al. in (Hsin, 2010) have applied in their study a method derived from
the adjoint equation of the finite element method to two propeller design problems. Matulja and
Dejhalla in (Matulja and Dejhalla, 2008) have realized the selection of the optimum skew propeller
geometry with artificial neural networks. Roddy et.al. in (Roddy et.al., 2006) have presented a method
443
based on the artificial neural network for the prediction of thrust and torque values of the Wageningen
B series propeller method. Chen and Shih in (Chen and Shih, 2007) have realized the propeller design
by the use of Wageningen B series propellers by considering the vibration and efficiency
characteristics using genetic algorithm. A similar study has been done by Suen and Kouh in (Suen and
Kouh, 2000).
In this study practical design approaches for Wageningen B series propellers are presented for the
propeller design conditions in which the delivered power (PD), the advance speed (VA) and the
revolution number (n) or propeller thrust (T), propeller rate of rotation (n) and propeller advance speed
(VA))are known. A set of propellers for a wide loading range is generated by the use of polynomials
presenting the open water diagrams of Wageningen B propeller series. The effects of the Reynolds
number are not included in the calculations. The set of propellers is four bladed and is obtained in such
a manner that all the blade area ratios (BAR) and pitch ratios (P/D) of the Wageningen B series are
included. The propeller design and performance characteristics are presented by means of empirical
formulas and diagrams to propeller designers as a practical tool for the use during the preliminary
design stage. Besides the propeller design, the diagrams and empirical formulas can also be used for
the propeller performance prediction.
2. Propeller Design with Standart Propeller Series In the initial design stage of a ship it is necessary to predict the performance of its propeller. For this
purpose, the Wageningen B screw series, since also having low cavitation, are used commonly for the
design of the propeller.
The experimental data of these series are firstly reported by Troost in (Troost, 1938). Later some
corrections are made on the series by taking the scale effects into account and the results are given by
van Lammeren et.al. in (Lammaren, 1969). A detailed regression analysis is made for the performance
characteristics of the Wageningen B series propellers by Oosterveld and van Oossanen in (Oosterveld
and Ossannen, 1975). They have presented the propeller open water characteristics of the Wageningen
B series for the Reynolds number equal to 2x106 as polynomial functions in (1) and (2). Later this
study was expanded by including viscous corrections for different Reynolds numbers (Carlton, 2006).
The variable parameters relating these series are the number of propeller blade (Z), the blade area ratio
of propeller (BAR) and the pitch ratio (P/D).
∑=
=47
1)()()/()(
(1)
∑=
=39
1)()()/()(
(2)
The propeller characteristics are expressed as below:
444
Thrust coefficient: 42
ρ= (3)
Torque coefficient:
52 ρ
= (4)
Advance coefficient:
= (5)
Open water propeller efficiency: π
η2
= (6)
Propeller thrust loading coefficient: 2
8
π
= (7)
The Wageningen B propeller series are general purpose series. These series are expressed in open
water diagrams where the KTKQJA curves are showed for propellers with constant blade number (Z)
and blade area ratio (BAR) but variable pitch ratios (P/D). Because the open water experiments are
made in fresh water, this must be considered in the design calculations.
3. Practical Design Approaches of Marine Propellers for 4 Bladed Wageningen B Series
In this section empirical expressions and diagrams are presented for the practical design and
performance analysis for four bladed propellers of the Wageningen B series. For two different design
case (PD, n, VA or T, n,VA are given), a set of propeller for each condition is generated by changing
systematically the advance speed (VA) and blade area ratio (BAR) of the main propeller including all
the four bladed Wageningen B propellers (BAR: 0.41.0; P/D: 0.51.4). Then the P/D, KT, KQ, JA and
ηo curves and expressions for the design of new propellers are given for different blade area ratios in
the case that (PD, n, VA) or (T, n, VA) are known. A computer code based on polynomials of the
Wageningen B series is used in the applications of the design method. The main propeller data used
for generating the set of propellers for each BAR (0.4, 0.55, 0.70, 0.85, 1.0) is given in Table 1.
Table 1. Main propeller design input data
Delivered power, PD (kW) 648
Speed of advance, VA (m/s) 4.372
Propeller rate of rotation per minute, RPM 380
Propeller diameter, D (m) 2.12
Number of blades, Z 4
Blade area ratio, BAR 0.70
According to the initial design variables, propeller design approaches for two conditions are as
following:
445
Propeller Design Case 1:
Initial design variable requirements of the propeller are given below:
Delivered power, PD in kW Propeller rate of rotation, n in rps Ship speed, VS in m/s Taylor wake fraction, w Number of blades, Z The necessary blade surface area ratio (BAR) required to minimize the risk of cavitation can be
determined using an appropriate cavitation criteria, such as that due to Burrill in (Burrill, 194344).
The torque requirement of a propeller can be expressed as a function of the advance coefficient
= by taking the diameter from this formula, replacing it in the expression 52 ρ
= and
by rearranging it, the equations below can be obtained:
555
2
.2
==
πρ (8)
k =5
2
2
πρ= (9)
For the design condition where the initial design variables, delivered power (PD), propeller rate of
rotation (n) and propeller advance speed (VA) of the “k” expression (9) are known, the propeller
design is carried out as follows: The torque requirement curve KQJA according to Equation (8) is
drawn over the Wageningen B open water diagram. The intersection points of this curve which
express the torque requirement of the propeller and the KQ curves of different P/D values of the open
water diagram describes the possible design solutions. The optimum efficiency curve is obtained by
drawing vertical lines from the intersection points to efficiency curves. And the maximum point of this
curve, which represents the highest efficient propeller among different solutions satisfying the
requirements, is read. Later the JA, P/D, KT, KQ and ηo values of the optimum propeller are read.
The set of propellers is generated by considering the PD and n values of the main propeller (Table 1)
constant and by changing the advance speed VA (2.522.5 knots, in 0.5 steps, 39 values) by containing
all the P/D (0.51.4) of the Wageningen B series. For this condition the propeller design is based on
only the (k) value, so the main propeller data is used to generate different values of (k) including all
propellers for 4 bladed Wageningen B series. And, P/D, JA, KT, KQ and ηo curves are presented in
dependence of k0.2 in Figure 15.
446
0.4
0.6
0.8
1
1.2
1.4
1.6
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Pitc
h ra
tio (P
/D)
k0.2
BAR=0.4
BAR=0.55
BAR=0.7
BAR=0.85
BAR=1.0
Figure 1. Variation of non dimensional pitch ratio (P/D) due to k0.2
0
0.2
0.4
0.6
0.8
1
1.2
0.5 1 1.5 2 2.5 3 3.5 4 4.5
Adv
ance
coe
ffici
ent
( JA
)
k0.2
BAR=0.4
BAR=0.55
BAR=0.7
BAR=0.85
BAR=1.0
Figure 2. Variation of advance coefficient (JA) due to k0.2
0.13
0.14
0.15
0.16
0.17
0.18
0.19
0.2
0.21
0.22
0.5 1 1.5 2 2.5 3 3.5 4 4.5
Thru
st c
oeffi
cien
t (K
T)
k0.2
BAR=0.4 BAR=0.55
BAR=0.7 BAR=0.85
BAR=1.0
Figure 3. Variation of thrust coefficient (KT) due to k0.2
447
0.1
0.2
0.3
0.4
0.5
0.5 1 1.5 2 2.5 3 3.5 4 4.5
Torq
ue c
oeffi
cien
t (10
K Q)
k0.2
BAR=0.4
BAR=0.55
BAR=0.7
BAR=0.85
BAR=1.0
Figure 4. Variation of torque coefficient (KQ) due to k0.2
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.5 1 1.5 2 2.5 3 3.5 4 4.5
Ope
n w
ater
effi
cien
cy (
η 0)
k0.2
BAR=0.4
BAR=0.55
BAR=0.7
BAR=0.85
BAR=1.0
Figure 5. Variation of open water efficiency (ηo) due to k0.2
As the pitch/diameter ratio (P/D), advance coefficient (JA), thrust coefficient (KT), torque coefficient
(KQ) and propeller efficiency (ηo) are given in dependence of (k) in the graphics above, these can also
be described in a general polynomial form (10). While the design of a propeller with the given PD, n,
VA in (k) can be carried out with the graphics, it can also be executed practically with the use of (10).
For the intermediate values of the blade area ratio (BAR) the linear interpolation method can be used.
The other propeller parameters (D, T) can be obtained from (JA) and (KT) expressions given in (3), (5).
In addition to propeller design, these equations and graphics can also be interpreted as useful and
timesaving tools for the prediction of the performance characteristics of an existing 4 bladed
propellers. The values of the coefficients in (9) due to the blade area ratio (BAR) are given in Table 2.
++++++= )()( 5/15/25/35/42.0 (10)
Here; k = 5
2
2
πρ; F(k), any of pitch ratio (P/D), advance coefficient (JA) etc.
448
Table 2. Coefficients of the F(k) equation due to blade area ratio (BAR)
BAR=0.4 F(k) a b c d e f g P/D 0.0469 0.658 3.6709 10.402 15.846 12.53 4.7135 JA 0.0182 0.2659 1.5607 4.7324 7.9103 7.1721 3.166 KT 0 0 0.0053 0.0535 0.1875 0.25914 0.2633
10KQ 0.0236 0.3287 1.818 5.0795 7.556 5.7157 1.9072 η0 0.003 0.0056 0.0331 0.0794 0.0085 0.4465 0.9475
BAR=0.55 F(k) a b c d e f g P/D 0.0437 0.6128 3.4116 9.636 14.614 11.494 4.3615 JA 0.0339 0.4778 2.6798 7.6654 11.878 9.7354 3.7278 KT 0 0 0.0036 0.0345 0.1156 0.1477 0.2038
10KQ 0.0205 0.284 1.5572 4.3023 6.3058 4.6823 1.5636 η0 0.0009 0.013 0.0747 0.1959 0.157 0.3042 0.9246
BAR=0.70 F(k) a b c d e f g P/D 0.0428 0.603 3.3695 9.53597 14.462 11.327 4.312 JA 0.0323 0.4573 2.5775 7.4106 11.541 9.5054 3.6688 KT 0 0 0.002 0.0192 0.0608 0.0664 0.1719
10KQ 0.0197 0.2735 1.5028 4.1506 6.0542 4.4478 1.4905 η0 0.0011 0.0154 0.0837 0.2046 0.1394 0.3806 0.9324
BAR=0.85 F(k) a b c d e f g P/D 0.029 0.4161 2.3818 6.9566 10.997 9.1227 3.8503 JA 0.0263 0.3758 2.1437 6.2669 9.9928 8.5175 3.4603 KT 0 0 0.0005 0.0016 0.0022 0.001 0.161
10KQ 0.0089 0.1301 0.7555 2.2266 3.5196 2.866 1.1633 η0 0.002 0.0283 0.1537 0.3935 0.3993 0.2119 0.8802
BAR=1.0 F(k) a b c d e f g P/D 0.0112 0.1747 1.1005 3.5816 6.3924 6.0879 3.175 JA 0.0182 0.2659 1.5607 4.7324 7.9103 7.1721 3.166 KT 0 0 0.0029 0.0296 0.1054 0.1517 0.1183
10KQ 0.006 0.0723 0.3214 0.6202 0.3805 0.2892 0.5878 η0 0.0042 0.0574 0.3057 0.7859 0.9192 0.1208 0.7864
Propeller Design Case 2:
Initial design variable requirements of the propeller are given below:
Propeller thrust, T in kN Propeller rate of rotation, n in rps Ship speed, VS in m/s Taylor wake fraction, w Number of blades, Z Blade area ratio, BAR In the case that total resistance of ship (RT) in a constant ship speed (Vs) is known, the propeller thrust
(T) is determined from thrust reduction factor (t) which can be obtained from empirical formulas or
model tests: )1/( −=
449
The thrust of a propeller can be expressed as a function of the advance coefficient
= by taking
the diameter from this formula, replacing it in the expression 42
ρ= and by rearranging it, the
equation below can be obtained:
41
44
2.
==
ρ (11)
k1 = 4
2.
ρ= constant (12)
For the design condition where the initial design variables, propeller thrust (T), propeller rate of
rotation (n) and propeller advance speed (VA) of the “k1” expression (12) are known, the propeller
design is carried out similarly with the propeller design method 1 explained above: The thrust curve
KTJA according to Equation (11) is drawn over the Wageningen B open water diagram. The
intersection points of this curve which express the thrust generated by the propeller and the KT curves
of different P/D values of the open water diagram are marked. Then the optimum efficiency curve is
obtained by drawing vertical lines from the intersection points to efficiency curves. The maximum
point of this curve represents the efficiency of the optimum propeller among different solutions
satisfying the requirements. Later the P/D, JA, KT, KQ and ηo values of the optimum propeller are read.
In this case, the set of propellers is generated by considering the T and n values of the main propeller
(Table 1) constant and by changing the advance speed VA (2.023.5 knots, in 0.5 steps, 42 values) by
containing all the P/D (0.51.4) of the Wageningen B series.
For this condition the propeller design is based on only the (k1) value, so the main propeller data is
used to generate different values of (k1) including all 4 bladed Wageningen B series propellers. And,
P/D, JA, KT, KQ and ηo curves are presented in relation to (k10.2) in Figure 610.
0.4
0.6
0.8
1
1.2
1.4
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
k30.2
Pitc
h ra
tio (P
/D)
BAR=0.4
BAR=0.55
BAR=0.7
BAR=0.85
BAR=1.0
Figure 6. Variation of non dimensional pitch ratio (P/D) due to k1
0.2
k10.2
450
0
0.2
0.4
0.6
0.8
1
1.2
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5k30.2
Adv
ance
coe
ffici
ent (
JA)
BAR=0.4
BAR=0.55
BAR=0.7
BAR=0.85
BAR=1.0
Figure 7. Variation of advance coefficient (JA) due to k1
0.2
0.13
0.14
0.15
0.16
0.17
0.18
0.19
0.2
0.21
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
k30.2
Thru
st C
oeffi
cien
t (K T
)
BAR=0.4BAR=0.55
BAR=0.7BAR=0.85
BAR=1.0
Figure 8. Variation of thrust coefficient (KT) due to k1
0.2
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
k30.2
Torq
ue c
oeffi
cien
t (10
K Q)
BAR=0.4
BAR=0.55
BAR=0.7
BAR=0.85
BAR=1.0
Figure 9. Variation of torque coefficient (KQ) due to k1
0.2
k10.2
k10.2
k10.2
451
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
k30.2
Ope
n w
ater
effi
cien
cy (η
o)
BAR=0.4
BAR=0.55
BAR=0.7
BAR=0.85
BAR=1.0
Figure 10. Variation of open water efficiency (ηo) due to k1
0.2
As the pitch/diameter ratio (P/D), advance coefficient (JA), thrust coefficient (KT), torque coefficient
(KQ) and propeller efficiency (ηo) are given as the function of (k10.2) in the graphics above, these also
can be described in a general polynomial form (13). While the design of a propeller with the given T,
n, VA in (k1) can be carried out with the graphics, it can also be executed practically with the use of
(13). For the intermediate values of the blade area ratio (BAR) the linear interpolation method can be
used. The other propeller parameters (D, Q) can be obtained from (JA) and (KQ) expressions given in
(4), (5).
The values of the coefficients in (13) due to the blade area ratio (BAR) are given in Table 3.
++++++= )()( 5/11
5/21
5/31
5/411
2.011 (13)
Here; k1 = 4
2.
ρ;
F(k1), any of pitch ratio (P/D), advance coefficient (JA) etc. Table 3. Coefficients of the F(k1) equation due to blade area ratio (BAR)
BAR=0.4 F(k1) a b c d e f g P/D 0.0347 0.5458 3.4182 10.885 18.65 16.579 6.7573 JA 0.0174 0.2802 1.8093 6.0106 10.932 10.622 4.77 KT 0 0 0.005 0.0537 0.2078 0.328 0.3223
10KQ 0.0166 0.2612 1.6337 5.1913 8.8426 7.7176 2.9026 η0 0.0002 0.0026 0.0071 0.0036 0.012 0.3375 0.9564
BAR=0.55 F(k1) a b c d e f g P/D 0 0.0351 0.4811 2.5104 6.2522 7.5568 4.2016 JA 0.0191 0.3046 1.9421 6.35 11.325 10.757 4.7295 KT 0 0 0.0047 0.0483 0.1723 0.2425 0.257
10KQ 0 0 0.0211 0.2313 0.8964 1.4523 0.9891 η0 0.0003 0.0055 0.0448 0.173 0.2764 0.1041 0.8864
BAR=0.70 F(k1) a b c d e f g P/D 0 0 0.05 0.5671 2.2977 3.9714 3.0688 JA 0.0266 0.4061 2.4661 7.6433 12.908 11.65 4.9102 KT 0 0 0.0044 0.043 0.1472 0.1967 0.2359
10KQ 0 0 0.0196 0.2157 0.8404 1.3689 0.9583 η0 0.0003 0.0023 0.0044 0.0697 0.1368 0.2061 0.9097
BAR=0.85
k10.2
452
F(k1) a b c d e f g P/D 0.0178 0.2879 1.8672 6.2242 11.296 10.788 5.0213 JA 0.0136 0.2202 1.4332 4.8136 8.8973 8.8772 4.1948 KT 0 0.0024 0.0331 0.168 0.3829 0.3772 0.2913
10KQ 0.0085 0.1352 0.8644 2.8218 4.9527 4.4646 1.8532 η0 0.0011 0.0176 0.117 0.3831 0.5806 0.0911 0.8226
BAR=1.0 F(k1) a b c d e f g P/D 0.006 0.1061 0.765 2.869 5.9513 6.6298 3.8647 JA 0.0084 0.1407 0.9514 3.3529 6.5894 7.1104 3.7094 KT 0 0 0.0002 0.0034 0.0192 0.0447 0.1563
10KQ 0.0004 0.0003 0.0425 0.3185 0.9619 1.3605 0.9979 η0 0.0017 0.0282 0.1812 0.5757 0.8779 0.3128 0.745
The practical design approaches presented in this work allow the design of a four bladed Wageningen
B series propeller or the performance prediction of an existing propeller (in dependence of only the (k)
value given as in (9) for Design Case 1 and the (k1) value as in (12) for Design Case 2) and the blade
area ratio. The present approaches are considered as a practical tool for propeller designers for the use
during the preliminary design stage as diagrams and empirical formulas.
4. Design Applications For verification and to show the usability of the present approaches, for each propeller design case
mentioned above, an application has been carried out for medium thrust loaded propeller of which
design variables are given in Table 4. While the variables of Propeller 1 are used for Design Case 1,
those of Propeller 2 are used for Design Case 2
From the design variables of these propellers with the use of the open water curves of the Wageningen
B series and the above presented empirical formulas the propeller designs are realized. The obtained
propeller design and performance results (P/D, KT, KQ, JA and ηo) are presented in Table 5. The design
results of the propellers are compared, and it is seen that the order of the error values is acceptable. Table 4.Design variables of the propellers
Propeller 1 Propeller 2 Propeller rate of rotation n, RPM 200 120 Delivered power PD, kW 6090 Propeller thrust T, kN 448 Ship speed Vs, knots 17.5 14 Taylor wake fraction, w 0.15 0.15 Number of propeller blade Z 4 4 Blade area ratio BAR 0.7 0.7
Table 5. Comparison of the design results of propellers
Design Case 1 Design Case 2 k=0.4010 CTH=1.210 k1=1.2443 CTH=1.0956
Open Water Present %Error Open Water Present %Error P/D 0.8100 0.7946 1.9049 0.84 0.8405558 0.0661705 JA 0.5580 0.5440 2.5099 0.588 0.5791836 1.499388 KT 0.1480 0.1486 0.4338 0.149 0.1472746 1.1579666 10KQ 0.2170 0.2078 4.2392 0.226 0.2228337 1.4010194 η0 0.6050 0.6024 0.4281 0.618 0.6168464 0.1866705
453
5. Results Recently although the propeller designs can be made with computer programs based on circulation
theory, the traditional design methods using propeller series based on model experiments remains as
the most applied method. Especially the practice of the design with a few design variables and the
availability of the performance prediction is the primary advantage of this method.
In this study practical design approaches are presented for four bladed Wageningen B series propellers
for the cases where (delivered power (PD), the advance speed (VA) and the propeller revolution
number (n)) or (propeller thrust (T), the advance speed (VA) and the propeller revolution number (n))
are known. The presented propeller design methods can be used in a rapid and accurate manner by the
given diagrams and empirical formulas. Additionally the performance characteristics of an existing
propeller can also be predicted.
For each design case the design applications of the present approaches for the propellers are realized.
The results obtained from these methods are compared with those of the open water diagrams, and a
good agreement is shown.
References
Arazgaldi R., Hajilouy, A. and Farhanieh, B., 2009, “ Experimental and Numerical Investigation of Marine Propeller Cavitation”, Scientia Scientia Iranica Transaction BMechanical Engineering, Vol.16 No: 6 pp: 525533
Bal, S. and Güner, M., 2009, “Performance of Podded Propulsors”, Ocean Engineering Vol.36 No: 8, pp: 556563.
Benini, E., 2003,“Multiobjective Design Optimization of BScrew Series Propellers Using Evolutionary Algorithms”, Marine Technology and Sname News, Vol.40, No:4, pp:229238.
Burrill, L.C., 194344, “Calculation of Marine Propeller Performance Characteristics.” 60.
Carlton, J.S., 2006 “Marine Propellers”, Second Edition 2006. CARLTON, J.S. “Marine Propellers”, Second Edition.
Çelik, F., Güner, M. and Ekinci, S., 2010, “An Approach to the Design of Ducted Propeller”, Scientia Iranica Transaction BMechanical Engineering, Vol.17 No: 5 pp: 406417.
Çelik, F., Güner, M., 2006, “An Improved Lifting Line Model for The Design of Marine Propellers” Marine Technology and Sname News, Vol.43, No:2, pp:100113.
Chen, J.H. and Shih, Y.S., 2007, “Basic Design of Series Propeller with Vibration Consideration by Genetic Algorithm”, Journal of Marine Science and Technology, Vol.12, No:3, pp:119129.
Eckhardt M.K. and Morgan, W.B., 1955, “A propeller design method”, 63.
Ghassemi, H., 2009, “The Effect of Wake Flow and Skew Angle on the Ship Propeller Performance”, Scientia Iranica Transaction BMechanical Engineering, Vol.16 No: 2 pp: 149158.
Glover, E.J., 1974, “A Design Method for Heavily Loaded Marine Propellers.” 116.
Güner, M., 1994, Ph.D. thesis, University of Newcastle upon Tyne, England, 1994.
Greeley, D.S. and Kerwin, J.E., 1982, “Numerical methods for propeller design and analysis in steady flow” 90.
454
Hsin, C.Y., 2010, “Application of the Adjoint Method to the Propeller Designs”, Journal of Hydrodynamics Ser.B, Vol.22 No:5, Supp.1, pp:501506.
Kerwin, J.E., Taylor, T.E., Black, S.D. and Mchugh, G.P., 1997, “A Coupled Lifting Surface Analysis Technique for Marine Propulsors in Steady Flow”, Proceeding Propeller/Shafting 1997 Symposium, Virginia.
Kulczyk, J. S.,Skraburski, L. and Zawislak, M., 2007, “Analysis of Screw Propeller DTMB 4119 using the FLUENT System”, Archives of Civil and Mechanical Engineering, Vol:7, No:4, pp:129136.
Kuiper, G., 2010, “New developments and Propeller Design”, 9th International Conference on Hydrodynamics, Shanghai, China.
Koronowicz, T., Chaja, P. and Szantyr, J., 2008, “A Computer System for the Complete Design of Ship Propellers”, Archives of Civil and Mechanical Engineering, Vol.8 No:1, pp:4959.
Lammaren, W.P.A. van, Manen, J.D. van, Oosterveld, M.W.C., 1969, “The Wageningen BScrew Series”, Trans. SNAME.
Lerbs, H.W., 1952, “Moderately Loaded Propellers with a Finite Numbers of Blades and an Arbitrary Distribution of Circulation.” 60.
Lee, H. and Kinnas, S.A., 2005, ”A BEM fort he Modeling of Unsteady Propeller Sheet Cavitation Inside a Cavitation Tunnel”, Journal of Computational Mechanics, Vol.37, No:”, pp:4151.
Matulja, D. and Dejhalla, R., 2008, “Neural Network Prediction of an Optimum Ship Screw Propeller”, Annals of DAAAM and Proceedings, pp:829830.
Olsen, A.S., 2004, “Energy Coefficients for a Propeller Series”, Ocean Engineering, Vol.31, No:34, pp:401416.
Oosterveld, M.W.C. and Ossannen, P. van, 1975, “Further Computeranalysed data of the Wageningen BScrew Series”, ISP,22.
Pien, P.C., 1961, “The Calculation of Marine Propellers Based on Lifting Surface Theory”, Journal of Ship Research, No:2.
Roddy, R.F., Hess, D.E. and Faller, W.E., 2006, “Neural Network Predictions of the 4Quadrant Wageningen BScrew Series”, pp:315335, COMPIT’06.
SanchezCaja, A., Rautaheimo, P. and Siikonen, T., 2001, “Simulation of Incompressible Viscous Flow Around a Ducted Propeller Using a RANS Equation Solver”, TwentyThird Symposium on Naval Hydrodynamics.
Suen, J.B and Kouh, J.S., 2000, “Genetic Algorithms for Optimal Series Propeller Design”, Marine Technology III Book Series: Marine and Maritime Vol:1 pp:237246.
Szantyr, J., 1987, “An interactive program for design of ship propellers”, The Naval Architect.
Tanaka H. and Yoshida, Y., 2003, “An Approach to Transformation of nonDimensional Diagrams into Nomograms and its Application to New Propeller Design System”, Marine Technology V, Book Series:Marine and Maritime, Vol.3 pp:91102.
Tukker, J. and Kuiper, G., 2004, “High Speed Video Observations and Erosive Cavitation”, PRADS Symposium.
Troost, L., 1938, “Open Water Tests Series with Modern Propeller Forms”, Trans.NECIES, Vol.54.
455
Effects of Recent Upgrades on Background Noise Levels of Emerson
Cavitation Tunnel
Emin Korkut*, Mehmet Atlar ** *I.T.U., Faculty of Naval Arch. and Ocean Eng., Istanbul, Turkey, [email protected] ** School of Marine Science and Tech., University of Newcastle, Newcastle, UK, [email protected]
Abstract Emerson Cavitation Tunnel (ECT) of Newcastle University has been a major facility for cavitation, noise
and propeller design studies, such as development of KCA, KCB, KCC and KCD series. This tunnel has
recently been modified to upgrade the existing measuring section with updated technical specifications.
The modifications included new honeycombs in the lower horizontal limb and high speed corner. The
measuring section is the same length of the previous measuring section, laser friendly, more accessible,
compliant with all of the existing dynamometers and capable to take dummy model hulls and other models
with ease. In addition to the new measuring section, a retro-fit style high speed insert was added to
increase the speed range over 10 m/s and hence providing the tunnel with a testing capability at higher
speeds. Low background noise level in cavitation tunnels is important to interpret correctly propeller noise
levels measured. In order to investigate the effect of the upgrades (new measuring section and with the
insert) on the background noise levels of the ECT an experimental study was carried out to measure the
tunnel background noise levels. This will provide information about the effect of new measuring section
and the insert on the background noise levels. The tunnel background noise level was measured previously
in 2000 and 2006. This paper includes the details of the above tests in terms of description of the testing
facility and the measurements associated with noise as well as the analysis and discussion of the these
characteristics.
Keywords: Emerson Cavitation Tunnel, Background Noise Level, New Test Section, Insert.
1. Introduction Emerson Cavitation Tunnel (ECT) of Newcastle University has been a major facility for cavitation, noise
and propeller design studies, such as development of KCA, KCB, KCC and KCD series. The tunnel has
been and still is an integral part of the teaching and research in the School. The Emerson Cavitation tunnel
is a worldwide respected facility, which actively competes in collaborative global projects and contributes
456
to the state-of-the-art in propulsion, water turbine, coating technology, ice research and numerous other
areas of research. Under the Hydro Testing Alliance (HTA) project (FP6), Joint Research Programme (JRP) 10 is called
“Noise Measurements” and in this programme Task 1 is entitled “noise measurements at model scale”
whilst its sub-Task 1.4 involves “Background noise measurements”. Within this sub-task Newcastle
University (UNEW) is committed to measure the background noise of their Emerson Cavitation Tunnel”
following the recent modifications made to their tunnel. The Emerson Cavitation Tunnel (ECT) of Newcastle University (UNEW) was recently upgraded by a
substantial grant from the Higher Education Funding Council for England (HEFCE)’s Strategic
Infrastructure Funds (SRIF - Phase3). The details of the upgrading are given in (Atlar, 2011) and involved
mainly followings (Atlar, 2011):
Replacement of the measuring section with larger window access Development and purchase of a high speed insert allowing speeds up to 9m/s Replacement of the old honeycomb section with an effective smaller celled unit Replacement of the guide vanes at the diffuser section end Addition of a new degassing system Replacement of the impeller bearing and dynamic balancing of the hub Upgrade of the tunnel control system to become automated Upgrade of the tunnel water fill/drain and pressure systems to become automated Renovation of the tunnel working area / building
It is anticipated that the above improvements, particularly on the measuring section and its vicinity -
through the replacement of the honeycomb at the contraction and guide vanes in the high speed corner-
will alter the flow quality and hence background noise characteristics of the tunnel positively. As well
known, low background noise level in cavitation tunnels is important to interpret correctly propeller noise
levels measured. It is therefore necessary to measure the background noise level of the upgraded tunnel
and to make a comparison with any available background noise data taken in the past. In order to
investigate the effect of the upgrades (new measuring section and with the insert) on the background noise
levels of the ECT an experimental study was carried out to measure the tunnel background noise levels.
This will provide information about the effect of new measuring section and the insert on the background
noise levels. The tunnel background noise level was measured previously in 2000 and 2006. This will
provide information about effect of new measuring section and the insert on the background noise levels.
457
This paper includes the details of the above tests in terms of description of the testing facility and the
measurements associated with noise as well as the analysis and discussion of the these characteristics.
Conclusions withdrawn from the study are also given. 2. Experimental Set-Up and Test Conditions Experiments were carried out in two groups at the Emerson Cavitation Tunnel (ECT) of the University of
Newcastle (UNEW), which is denoted as “Section 1” in this paper, has a cross section of 1.22 m by
0.81m, while smaller section, which is obtained by placing an insert in the larger section and denoted by
“Section 2” has a cross section of 0.8 m by 0.81 m. Atlar (2011) gives a complete description of the
facility and new upgrades. The first group was the background noise measurements in the larger section. The second group of tests
involved the same tests in the case of insert (smaller section). Table 1 summaries the test conditions.
Table 1. A summary of test conditions.
Test performed Test section
Tunnel water velocity V (m/s) Dynamometer, N (RPM) Vacuum condition
Nothing running
Section 1 (without insert)
N/A N/A
Atmospheric, Hst=0.25mHg &
Hst=0.50mHg
Impeller only 0, 2, 3, 4, 5 & 6
Dynamometer only 0 500, 750, 1000, 1250, 1500, 1750, 2000, 2500, 2750 & 2950
Impeller & Dynamometer on 2, 3, 4, 5 & 6 500, 750, 1000, 1250, 1500, 1750,
2000, 2500, 2750 & 2950 Nothing running
Section 2 (with insert)
N/A N/A
Impeller only 0, 3, 4, 5, 6, 7 & 8 Dynamometer only 0 500, 1000, 1500, 2000, 2500 & 2950 Impeller & Dynamometer on 3, 4, 5, 6, 7 & 8 500, 1000, 1500, 2000, 2500 & 2950
2.1 Noise Measurements
Noise measurements were carried out at three different conditions; atmospheric condition, Hst= 0.25 mHg
and Hst=0.50 mHg vacuum conditions, given in Table 1. The measurements have been recorded using a
Bruel and Kjaer type 8103 miniature hydrophone mounted in a water filled, thick walled, steel cylinder
placed on a 30mm thick plexiglass window above the propeller at a vertical distance of 0.459 m
coinciding with the tunnel centreline and shaft centreline of the dynamometer in the case of larger test
section, as shown in Figure 1. However the position of the hydrophone was above propeller at a vertical
distance of 0.459 m and at 0.2 m off the tunnel centreline in the case of insert (smaller test section), that is
shown in Figure 2. The signals from the hydrophone were collected by further Bruel and Kjaer equipment,
in this case a PC based Pulse digital acquisition and analysis system up to a frequency of 25 kHz.
458
Figure 1. Hydrophone position and the dynamometer in larger test section.
Figure 2. Hydrophone and the dynamometer position in the case of insert.
During the measurements the dissolved gas content of the water was measured about 11%, using a YSI55
oxygen meter. Previously various background noise measurements campaigns with the ECT were carried
out in 2000 and 2006 for investigating the noise characteristics of different model propellers. The 2000
campaign for the model propeller of a fisheries research vessel can be found in Takinaci, et al., (2000) and
Atlar et al., (2001). The 2006 campaign for the model of a coated tanker propeller were reported in Korkut
(2006) and Korkut and Atlar (2009). Full details of the recent 2008 campaign can be found in Korkut and
Atlar (2011). Although there are differences in the hardware and software arrangements of these
measurement campaigns, as well as the differences due to the tunnel modifications, it was thought that it
would be useful to include the results of these past measurements and to compare with the results of the
current measurement campaign. 3. Analysis and Presentation of Noise Measurements Although the measurements were carried out with the test sections 1 and 2 for different water velocities,
dynamometer speeds in different vacuum conditions given in Table 1, Figures 3 to 7 only show
comparisons of some background noise measurements of the tunnel for different test sections and
0.2 m
459
previously measured levels in 2000 and 2006. A common practice in the analysis and presentation of the
noise levels is to reduce the measured values of Sound Pressure Levels (SPL) in each 1/3 Octave band to
an equivalent 1 Hz bandwidth by means of the correction formula recommended by ITTC (1978) as
follows.
SPL = SPLm - 10log( f ) (1)
where SPL1 is the reduced sound pressure level to 1 Hz bandwidth in dB; re 1 Pa, SPLm is the measured
sound pressure level at each centre frequency in dB; re 1 Pa and f is the bandwidth for each one-third
octave band filter in Hz. The ITTC also required that the sound pressure levels be corrected to a standard
measuring distance of 1 m using the following relationship.
SPL = SPL1 - 20log (r) (2)
where SPL is the equivalent 1 Hz at 1 m distance sound pressure level (in dB; re 1 Pa) and r is the
vertical reference distance for which the noise level is measured.
0
20
40
60
80
100
120
140
10 100 1000 10000 100000
Centre Frequency (Hz)
SPL
(dB
; re
1 �Pa
, 1 H
z, 1
m)
Nothing running without insertat atmospheric
Nothing running with insert atatmospheric
Figure 3. Comparison of noise levels of tunnel in nothing running condition with and without insert at
atmospheric condition. The Emerson Cavitation Tunnel is a large steel structure, which is mounted on a special foundation
without any attachment to the surrounding building. In a survey study on the background noise of the
tunnel carried out by Clarke (1987), the facility was specified as “noisy” in nature requiring careful
measurement of the background noise contributed from the major sources, which are the surrounding
environment, impeller and dynamometer. Therefore, the tunnel has been recently upgraded in various
areas including the replacement of its measuring section to have a more accessible testing area as well as a
higher speed capability with the introduction of an insert. This upgrading was also combined with the
replacement of the old honeycomb section at the contraction with a new modern honeycomb to improve
460
the inflow to the new test section. Within the above framework it became an important objective to see the
effect of these improvements on the background noise characteristics of the tunnel and hence a
comprehensive noise survey was conducted.
0
20
40
60
80
100
120
140
10 100 1000 10000 100000
Centre Frequency (Hz)
SPL
(dB
; re
1 �Pa
, 1 H
z, 1
m)
Nothing running without insert atHst=0.5 mHg
Nothing running with insert atHst=0.5 mHg
Figure 4. Comparison of noise levels of tunnel in nothing running condition with and without insert at
Hst=0.5mHg condition.
Figure 5. Comparison of noise levels of in impeller running only at V=6 m/s with and without insert at
atmospheric condition.
461
Figure 6. Comparison of noise levels of tunnel in dynamometer running only at N=2000 RPM with and
without insert at atmospheric condition.
0
20
40
60
80
100
120
140
10 100 1000 10000 100000
Centre Frequency (Hz)
SPL
(dB
; re
1 �Pa
, 1 H
z, 1
m)
Without insert, V=5 m/s & N=1500 RPM atatmosphericWith insert, V=5 m/s & N=1500 RPM at atmospheric
Figure 7. Comparison of noise levels of tunnel in impeller and dynamometer running together at V= 5 m/s
and N= 1500 RPM with and without insert at atmospheric condition.
462
Figure 8. Comparison of noise levels of tunnel in impeller and dynamometer running together at V= 4 m/s
and N=1000 RPM measured recently with those measured in 2000 and 2006.
Figure 9. Comparison of noise levels of tunnel in impeller and dynamometer running together at V= 4 m/s
and N=1500 RPM measured recently with those measured in 2000 and 2006.
Background noise levels measured with the test section 1 in the nothing running condition are almost
similar to those measured with the test section 2 under the atmospheric condition (Figure 3). However
under the vacuum conditions noise levels with the test section 2 are slightly less than those with test
section 1, as shown in Figure 4. For the impeller running only condition at V= 3 and 4 m/s noise levels of
the test section 2 are less than those of test section 1. At V= 5 and 6 m/s a similar trend is also observed in
463
the low frequency range. However in the high frequency range the noise levels of the test section 2 are
higher than those of the test section1 (Figure 5). Background noise levels in the dynamometer in operation only with both test sections are almost similar,
as shown in Figure 6. In the case of impeller and dynamometer in operation together at V= 3 m/s a similar
trend is observed. At V=4 and 5 m/s the noise levels are similar for both test sections in the low frequency
range, however in the high frequency range a slight increase in the noise levels of the test section 2
compared to those of the test section 1. This can be seen in Figure 7. For impeller and dynamometer running together conditions, comparisons are made between the recent
results and the results measured in 2000 and 2006. In Figures 8 and 9 the background noise levels with
test section 1 of 2000 and 2006 are greater than those measured recently in 2010. This trend indicates that
the recent modifications made in the tunnel test section improved the background noise levels of the
tunnel. 4. Conclusions An experimental study was carried out to investigate the background noise levels of the Emerson
Cavitation Tunnel with recently installed two test sections. Some conclusions withdrawn from the study
are as follows:
Background noise levels measured with the larger (Test Section 1) and smaller (Test Section 2),
which is formed by using an insert in Test Section 1, indicate that the dynamometer is still the main
background noise source compared to the contributions from the tunnel’s main impeller and the
tunnel in silent (nothing running) condition. This may suggest that the tunnel background noise can
be further reduced by replacing the present electric motor of the dynamometer with a silent one.
Level of the background noise measurements increases with increasing level of vacuum applied to
the tunnel for all conditions.
Test Section 2 (with insert) has slightly less background noise levels than Test Section 1 (without
insert) for all conditions tested.
Based on the comparisons with some background noise measurements taken with the old measuring
section in the past, the recent upgrading made to the Emerson Cavitation Tunnel appears to have
improved the background noise levels at its measuring section.
5. Acknowledgements This paper is based on part of an EU- FP6 Project: Hydro testing Alliance, Joint Research Programme No.
10 (HTA-JRP10); Noise Measurements. The principal author was supported by the Scholl of Marine
464
Sciences and Technology during his stay in Newcastle for the tests. The authors thank Mr Ian Paterson,
the supervisor of the Emerson Cavitation Tunnel, for his help during the experiments.
6. References
Atlar, M., Takinaci, A.C., Korkut, E., Sasaki, N. and Aono, T., (2001), “Cavitation Tunnel Tests for Propeller Noise of a FRV and Comparisons with Full-Scale Measurements”, 4th International Symposium on Cavitation CAV2001, Pasadena, California, USA.
Atlar, M. (2011), “Recent Upgrading of Marine Testing Facilities at Newcastle University”, AMT’11 Conference Proceedings, Newcastle University, April 2011.
Clarke, M.A., (1987), “Noise Project, Newcastle University”, Report of Stone Vickers Ltd Technical Department, Report No: H93, UK.
ITTC, (1978) “Cavitation Committee Report”, 15th International Towing Tank Conference. The Hague, the Netherlands.
Korkut, E. (2006), “An Investigation into the Effects of Foul Release Coating on Propeller Performance, Cavitation and Noise Characteristics”, Report of School of Marine Science and Technology, Report No: MT-2007-004, University of Newcastle, UK.
Korkut, E. and Atlar, M., (2009), “An Experimental Study into the Effect of Foul Release Coating on the Efficiency, Noise and Cavitation Characteristics of a Propeller”, First International Symposium on Marine Propulsors, smp’09, Trondheim, Norway.
Korkut, E. and Atlar, M., (2011), “Background Noise Measurements of the Emerson Cavitation Tunnel following the Upgrading in 2008”, Report of School of Marine Science and Technology, University of Newcastle, UK.
PULSE Product Data, (2011), “Software for PULSE™ 15 incl. Types 7700, 7705, 7707, 7709, 7764, 7770, 7771, 7773, 7789 and 7797”, Brüel & Kjaer.
Takinacı, A.C., Korkut, E., Atlar, M., Glover, E.J. and Paterson, I., (2000). Cavitation observations and noise measurements with model propeller of a fisheries research vessel. Department Report No:MT-2000-56, Department of Marine Technology, University of Newcastle upon Tyne, Newcastle, UK.
465
Wake Adapted Propeller Design Application to Navy Ships
Ahmet Gültekin Avcı, Emin Korkut I.T.U., Faculty of Naval Arch. and Ocean Eng., Istanbul, Turkey, [email protected], [email protected]
Abstract Wake adapted propeller design is a one of the key issues to improve performances of ships by improving
efficiency of their propellers, hence reducing fuel consumptions of the ships as well improving hull
pressure, cavitation and noise characteristics. Within the above context, an MSc study has been carried out
numerically to investigate the effect of a wake adapted propeller design for a navy vessel. This paper
reports preliminary results of the study. The paper includes the details of the study, applications to the
navy vessel and discusses further improvement of the methodology.
Keywords: Lifting-line Method, Wake-adapted Propeller Design, Navy Ship.
1. Introduction Importance of propeller design has been increased due to requirements for environmental friendly, better
performance and efficiency in ship propulsion. To design a propeller is an optimization problem,
including efficiency, cavitation, vibration and noise characteristics that should be in acceptable limits. In
this respect wake adapted propeller design is a one of the key issues to obtain better performances of ships
by improving efficiency of their propellers, hence reducing fuel consumptions of the ships as well
improving hull pressure, cavitation and noise characteristics.
Propellers are usually designed first using propeller standard series data, such as Wageningen-B series, Au
series, Gawn series, KCA series, Ma series, Newton-Rader series, KCD series, Meridian series, etc.
(Carlton, 2007). In order to carry out propeller design, first of all, ship service speed, resistance of the ship
for the service speed according to ship type and the route and required engine power should be predicted.
After prediction of ship resistance by model tests or numerical methods, aft form of the ship should be
analyzed and a propeller with most suitable dimensions should be estimated initially in the first phase.
The other design parameters are diameter and number of blades for a given propeller type. Theoretically,
while number of blades increases, vibration decreases but it also decreases the efficiency. Other important
factors for the propeller design are propeller rate of rotation and cavitation, noise and hull pressure
466
characteristics (Carlton, 2007). Furthermore wake adaptation of the propellers which is designed, using
standard series data is necessary to obtain minimum vibration, noise and cavitation characteristics for the
maximum efficiency. This leads to further optimization of the propellers.
In the above context a study has been carried out numerically to investigate the effect of a wake adapted
propeller design for a navy vessel by using a Matlab code. The main objective of the program is to provide
accurate and powerful propeller design tool for use by users. The program developed for the wake
adaptation is applied to a hypothetical navy ship. Methodology used for the design program is given in
Section 2. Results of the program are included in Section 3. Finally some conclusions withdrawn from the
study are also given.
2. Design Methodology Propellers are designed to absorb minimum power and to give maximum efficiency, minimum cavitation
and minimum hull vibration characteristics. These objectives can be achieved in the following stages: a)
basic design b) wake adaptation and c) design analysis.
In the basic design, by using semi-empirical methods (i.e. standard series model propeller data and
cavitation diagrams) and simple beam theory, stress prediction, the diameter, pitch, blade surface area and
weight of the propeller are obtained. For the wake adaptation, by using analytical procedures (e.g. vortex
flow based lifting line methods) and simple blade section design methods the basic design is further
optimised (i.e. pitch distribution and sectional blade shape) with respect to the non-uniform axial wake
flow in which the propeller works.
In the design analysis stage, using advances analytical procedures (i.e. Lifting surface based methods), the
optimised design is analysed in 3-D wake. If the analysis demonstrates unsatisfactory performance of
cavitation hull pressures, shaft forces and moments, blade section geometry are modified by trial and error
until the problem is solved. In this paper the first two stages are used to obtain wake-adapted propeller
design. Design analysis stage is still in progress.
2.1 Lifting Line Method
Lifting line method was developed by Prandlt, which is based on potential flow, and the details of the
model can be found in Parndtl et al. (1934). In this method, which is based on the theory of circulation, a
blade is considered as a straight line. The circulation related to blade is determined by vortex filaments.
467
According to Helmholtz, these filaments cannot be restricted in the fluid. Vortex which formed
surrounding the trailing edge generates the airstream downwards. Abbott et al. states amplitude of the
airstream occurs anywhere of the blade section equals to the sum of the amplitudes of vortexes around
trailing edge (1959). A general scheme of lifting line theory is demonstrated in Figure 1 below.
Figure 1. Lifting Line Theory Scheme (Epps, 2010).
Propeller blade velocity and force diagram shows the velocities and forces on a 2D blade section through
the directions of ea and et in Figure 2.
Figure 2.Propeller blade velocity/force diagram (Epps, 2010).
In Figure 2, angular velocity is ae , inflow velocities are a a aV V e and t t tV V e ,induces
axial and induced tangential velocities are * *a a au u e ve * *
t t tu u e , the total resultant inflow
velocity is
* * 2 * 2( ) ( )a a t tV V u r V u (1)
Hydrodynamic pitch angle is *
*arctan a ai
t t
V ur V u
(2)
468
Also, angle of attack , blade pitch angle i , circulation re , inviscid Kutta-Joukowski force
* ( )rİF V e and viscid drag force VF . For a Z bladed propeller total thrust and total torque can be
written as:
cos sinh
R
i i v ir
T Z F F dr (^e ) (3)
sin cosh
R
i i v ir
Q Z F F dr (^e ) (4)
Power (P) consumed by the propeller is:
P Q (5)
where Q is the torque produced by the propeller and is rotational speed component. Then efficiency of
the propeller () can be expressed as:
sTVQ
(6)
where Vs is the ship speed.
2.2 Design Algorithm
Propeller design algorithm is developed based on Matlab codes. A typical design algorithm is shown in
Figure 3. Full details of the method can be found in Avci (2011).
3. Results and Discussions 3.1 Design Application to a Navy Ship
Propeller design studies have been carried out for a hypothetical navy ship. A scaled model of the
hypothetical navy ship was manufactured at Ata Nutku Ship Model Testing Laboratory of Istanbul
Technical University. Resistance, wake, propulsion and flow visualization tests were performed in the
Large Towing Tank of the Ata Nutku Ship Model Testing Laboratory, which is 160 m long, 6 m wide and
has a water depth of 3.4 m. The main particulars of the navy ship and propellers are given in Table 1 for
the design loading condition.
469
Table 1. Main particulars of hypothetical navy ship and propellers.
Length on waterline LWL (m) 85 Breadth Bmax (m) 13.55 Draught (midship) T (m) 3.38 Displacement (ton) 1913 Wetted surface area AWS (m2) 1122.75 Surface area of rudders AR (m2) 14.4 Total surface area of appendages AA (m2) 200 Bulbous sectional area AB (m2) 3.00
Block coefficient CB 0.48 Service speed VS 25 kn. Number of propeller - 2 Number of propeller blades Z 5 Diameter of propeller D (m) 3.4 Expanded blade area ratio AE/A0 0.85
Figure 3.Propeller design algorithm.
3.2 Resistance Prediction
Resistance prediction estimated using appropriate methodical series data or statistical analysis data, or
model test results in the code. For the ship in question comparison of resistance, hence effective power,
predictions for the navy ship is given in Figure 4 based on different methods and experiments.
470
Figure 4.Resistance prediction curves.
3.3 Propeller Design Using Standard Series Data
In this section, standard series data were used for the preliminary design. First, the propulsion factors,
wake fraction (wt), thrust deduction factor, (t) and relative-rotative efficiency values for the ship were
estimated using the commercial available Desp codes, which is based on Wagenningen-B series data and
then, these parameters were entered in the Matlab code. The propulsion coefficients obtained from Desp
calculations are shown in Table 2.
Table 2. Propulsion coefficients estimated.
VS RT T wt t H R 0 D [knots] [kN] [kN] [-] [-] [-] [-] [-] [-]
15 165.1 196.8 0.193 0.161 1.039 0.958 0.536 0.534 16 194.2 231.5 0.192 0.161 1.038 0.958 0.538 0.536 17 227.2 270.8 0.191 0.161 1.037 0.958 0.541 0.537 18 263.1 313.7 0.19 0.161 1.036 0.958 0.541 0.537 19 301.8 359.8 0.19 0.161 1.035 0.958 0.54 0.536 20 346.5 413.1 0.189 0.161 1.034 0.958 0.539 0.534 21 403.3 480.8 0.188 0.161 1.033 0.958 0.537 0.532 22 473.4 564.4 0.187 0.161 1.032 0.958 0.534 0.528 23 550.4 656.2 0.186 0.161 1.031 0.958 0.531 0.524 24 627.6 748.2 0.186 0.161 1.03 0.958 0.528 0.521 25 705.8 841.4 0.185 0.161 1.029 0.958 0.526 0.518 26 784.0 934.7 0.185 0.161 1.029 0.958 0.524 0.516 27 861.3 1026.9 0.184 0.161 1.028 0.958 0.522 0.515 28 937.0 1117 0.184 0.161 1.028 0.958 0.522 0.514 29 1010.4 1204.6 0.184 0.161 1.027 0.958 0.521 0.513 30 1074.5 1280.9 0.183 0.161 1.027 0.958 0.522 0.513
where RT is the total resistance, T is the thrust, H is the hull efficiency, R is the relative-rotative efficiency, 0 is
propeller open water efficiency and D is propulsive efficiency.
471
3.3.1 Matlab Calculations For the basic design stage, Matlab codes use Wageningen-B series data. If resistance values of the ship are
not taken from model tests, the codes for resistance prediction should be launched first. Four design
options exist, which are;
1. T, D, VA are known and Nopt is required,
2. PD, D, VA are known and Nopt is required,
3. T, N, VA are known and Dopt is required,
4. PD, N, VA are known and Dopt is is required.
The propeller diameter is known, so that the first option was chosen. KT/J2 curve and open water diagrams
obtained from the calculations for different P/D ratios are shown in Figure 5. A 3-D view of the designed
propeller is illustrated in Figure 6.
Figure 5.KT/J2 curves obtained for standard series data. Figure 6. 3-D View of standard series
design propeller
Results of the basic design using the standard series data are given in Table 3 for two options: the
propulsion factors estimated and taken from the model test results.
Table 3. Standard series design results. Type=1 : KT/J^2 condition wt and t estimated wt and t taken from model test results Number of propellers – INPUT 2.000 2.000 Number of blades – INPUT 5.000 5.000 Wake Fraction - INPUT OR ESTIMATED 0.234 0.080 Thrust Deduction Fraction - INPUT OR ESTIMATED 0.139 0.126 Propeller Diameter in meters(initial) - INPUT OR ESTIMATED 3.400 3.400 AE/A0 ratio (BAR) - INPUT OR ESTIMATED USING KELLER 0.75 0.75 Service Speed (knots) 25 25 Total Ship Resistance (kN) 760 760 Thrust for each propeller (kN) - CALCULATED 441 435 Advance Velocity (m/s) - CALCULATED 9.85 11.83 Hull efficiency – CALCULATED 1.12 0.950 Relative-rotative Efficiency – INPUT 1.000 1.000 Maximum Open Water Efficiency - OUTPUT 0.63 0.67 Pitch to Diameter Ratio (P/D)- OUTPUT 1.00 1.1 Rotation Rate (rpm), N – OUTPUT 273 275 BURRILL CAVITATION CONTROL 0.53<0.75 0.55 < 0.75
472
3.4 Wake-adapted Propeller Design
In order to design a wake-adapted propeller one should use an analytical procedure (e.g. vortex flow based
lifting line method) and simple blade section design methods. Then, the basic design is further optimised
(i.e. pitch distribution and sectional blade shape) with respect to the non-uniform axial wake flow in which
the propeller works.
After the basic design, the following parameters were obtained as: propeller diameter, D=3.40 m,
propeller rate of rotation N=275 RPM, pitch to diameter ratio P/D=1.1, expanded blade area ratio
AE/A0=0.75 and number of blades, Z=5. Nominal axial and tangential velocities obtained from the wake
survey are shown in Table 4.
Table 4. Inflow velocities obtained from wake survey.
r/R Axial (Vx/Vs) Tangential (Vt/Vs) 0.3 0.6116 -0.0252 0.5 0.9565 -0.0102 0.6 0.9617 -0.0089 0.7 0.9568 -0.0160 0.8 0.9591 -0.0108 0.9 0.9552 -0.0148 1.0 0.9411 -0.0137
3.4.1 Wake-adapted Design Using OpenProp Matlab Codes
OpenProp is an open source coded program which was developed by Massachusetts Institute of
Technology to design and analyse of propellers and turbines. The mathematical model of the code
depends on propeller vortex lattice lifting line method (Kimball and Epps, 2010). Epps also states that a Z
bladed propeller is considered as a straight radial lifting line including M numbers of panels by this
method. Figure 7 demonstrates the validation of the codes agreeing with the US Navy code PBD solutions
and experimental data on US Navy 4119 propeller (Kimball and Epps, 2010; Black, 1997).
Figure 7. Comparison of results of OpenProp, PBD and experiment on U.S. Navy 4119 circulation distribution and open water diagram (Kimball and Epps, 2010).
473
Propeller parameters to be used in lifting line method by MPVL program which is the early version of
OpenProp are shown in Table 5. According to Chung, results of MPVL match Pvl completely for the
heavily loaded propellers (2007).
Table 5. MPVL design parameters.
Number of blades 5 Rate of rotation 275 Diameter 3.4 Required thrust (kN) 435 Ship velocity (m/s) 11.83 Water density kg/m3 1025 Number of panels over the radius 50 Number of points over the chord 50 Max. iterations 200
A typical input and interface of MPVL used for single design is shown in Figure 8. c/D values are
calculated from the reference of Hoerner (1965). Performance results for each rate of rotations by single
design are given in Table 6.
Figure 8. MPVL interface.
Table 6. Performance results for different rate of rotation.
N (RPM) 235 250 275 300 Ct 0.3116 0.3116 0.3116 0.3116 Cp 0.3954 0.4035 0.4205 0.442 KT 0.123 0.1086 0.0898 0.0754 KQ 0.0249 0.0211 0.0165 0.0134
Va/Vs 0.909 0.909 0.909 0.909 0 0.7163 0.7019 0.6735 0.6408
474
Blade section thickness for each radius is calculated by the program and checked according to the strength
rules of Bureau Veritas. Figure 9 shows a view of the wake-adapted designed propeller. Geometric
properties of designed propeller are given in Table 7.
Figure 9. MPVL single wake-adapted designed
propeller.
4. Conclusions In this study, a wake-adapted propeller design program was developed using Matlab and an applied to a
hypothetical navy ship. Some conclusions withdrawn from the study are as follows:
The program at this stage is easy to use with a user friendly interface and very efficient for
propeller designs.
The parameters optimized by using MPVL parametrical design are propeller blade number,
propeller diameter and propeller rate of rotation.
Wake adaption is definitely required for appropriate propeller designs to obtain maximum
propulsion efficiency with minimum cavitation, noise and hull pressure characteristics.
Further work is to include a design analysis tool to obtain satisfactory cavitation hull pressures,
shaft forces and moments. This work is still in progress.
5. Acknowledgements
This study is supported by Postgraduate Thesis Support Program of the Graduate School of Science, Engineering and Technology. The authors would like to thank Assoc. Prof. Ali Can Takinaci for his helpful comments on the analyses.
r/R P/D Skew Xs/D c/D f0/c t0/c 0.317 0.66 -6.6 0 0.289 0.0051 0.1298 0.342 0.74 -7.2 0 0.304 0.0045 0.1183 0.398 0.93 -8.6 0 0.339 0.0037 0.0899 0.436 1.05 -9.4 0 0.361 0.0033 0.0755 0.527 1.19 -10.2 0 0.420 0.0024 0.0504 0.631 1.23 -8.1 0 0.471 0.0019 0.0351 0.738 1.22 -1.1 0 0.478 0.0014 0.025 0.837 1.18 9.9 0 0.456 0.0010 0.019 0.918 1.11 22.3 0 0.376 0.0007 0.0173 0.95 1.07 28 0 0.308 0.0007 0.019
0.999 1.01 37.6 0 0.119 0.000 0.051
Table 7. Geometric properties of the wake adapted propeller
475
6. References Avci, A.G., (2011). “Wake Adapted Propeller Design Application to Navy Ships”. MSc Thesis, Istanbul Technical University.
Abbott, I.H. and von Doenhoff, A. E. (1959). “Theory of Wing Sections, Including a Summary of Airfoil Data”. Dover.
Black, S.D. (1997). “Integrated Lifting Surface/Navier-Stokes Design and Analysis Methods for Marine Propulsors”. Ph.D. Thesis. MIT.
Carlton, JS. (2007). “Marine Propellers and Propulsion”. Butterworth-Heinemann, Elsevier, Second Edition.
Chung, SM (2007) “An Enhanced Propeller Design Program Based on Propeller Vortex Lattice Lifting Line Theory”. MSc. Thesis, MIT.
Epps, B.P. (2010) OpenProp v2.4 Theory Document, MIT Department of Mechanical Engineering Technical Report.
Hoerner, J, (1965) Fluid dynamic drag. New York.
Kimball R.W. and Epps, B.P. (2010). "OpenProp v2.4 propeller/turbine design code," http://openprop.mit.edu.
Prandtl, L, Tietjens, O.G., and Hartjog, J. (1934) “Applied Hydro and Aeromechanics”. London, England: McGraw-Hill Book Company, Inc.
477
Energy Saving in Trawlers: Practical and Theoretical Approaches Gaetano Messina Former research leader of the Institute of Marine Sciences, Marine Fishery Department, Italy, [email protected]
Abstract Due to a critical overfishing situation all over the Mediterranean waters, maintaining the productivity of a
trawler at acceptable levels calls for technological interventions, mainly aimed at reducing the fuel costs.
In order to discuss on energy savings in fishing, a trawler is a very suitable example, since its management
costs are strongly affected by the fuel consumed.
This paper tries to identify key areas to achieve fuel saving in fishing activities.
Many trawlers hulls request quite different powers to reach the same speed due to the fact that even small
modifications to the hull shape could provide significant variations of its resistance by sea waters.
Some analyses on cruising speed, hull shape and propulsion systems will be worked out in the paper, on the
base of some research results and experience based considerations, addressed to the hull and the propulsive
apparatus as well. 1. Introduction The productivity of a trawler could be expressed as a ratio between the fish catch value and the overall costs
to achieve this catch. The present fuel cost, which concerns most fishing fleets, is claiming technical
solutions for cheaper fishing vessel designs.
This paper is aiming at giving a contribution in this sense, offering some considerations mainly addressed to
the aspects of management costs of a fishing vessel as sea-going vehicle.
Due to the clear impossibility to not fish more, maintaining this productivity at acceptable levels calls for
technological interventions, mainly aimed at reducing the fuel costs. To discuss on energy savings in fishing,
a trawler is a very suitable example, since the management costs of this type of vessel are strongly affected
by fuel consumptions.
A fishing trip of a trawler consists of two fishing stages:
a) steaming from/to any fishing areas and
b) towing the fishing gear
The following worthwhile areas could be identified for investigation:
- steaming speed
- propulsion systems
While steaming to/from fishing grounds, the ship’s hull is the main user of the engine power and fishing
boats’ features could be improved by applying to their hulls some rules of naval architecture, till now almost
all neglected.
478
2. The Steaming Speed Let’s firstly discuss on steaming speed. Fuel consumption is closely linked to the delivered engine power
which, on turn, depends on ship’s resistance and speed.
A typical feature of the vessel resistance curve is of moderate increase at low speed with increasing steepness
in the higher speed regions. At the top of the speed range, the resistance increases with speed in the 6th to 8th
power.
Very high speed-length ratios for displacement hulls (about 1.3), corresponding to high ship resistances, are
peculiar to steaming. In order to reduce the resistance it would be enough to make the ship to operate at a
lower speed/length ratio.
Two main factors determine the shape of the resistance curve for a vessel:
- vessel displacement
- vessel length
Ship resistance is roughly proportional to its displacement. Some investigations show a 35 ÷ 45% resistance
increase for displacement increases by 50%. The vessel length determines the steepness of the resistance
curve at different speeds and, in practice, the maximum attainable speed of the vessel.
For a displacement type hull, there will be a practical upper speed limit which cannot be exceed, irrespective
of the increase in power applied. Therefore, a reduction in speed when the ship is steaming from one fishing
area to another and from there to the home port and vice versa, could allow a large fuel saving.
This could be accomplished by two different ways, i.e.:
- by lengthening the ship to realize as much length as possible, according to its requirements in terms of
stability, seaworthiness and working efficiency;
- by reducing the speed.
The energy saving rising from a steaming speed reduction will be consider here.
Many steaming tests have been carried out, over a research fishing trip at different engine revolutions taking,
as a starting and reference point, the fuel consumption to travel a given distance at a maximum speed of.
10.25 knots
Reducing the speed from 10.25 knots to 9.75 knots (i.e. by only half a knot) gives a fuel consumption
decrease, of about 18%. Generally speaking, lowering by 10% the free running speed reduces by 30-40% the
fuel consumed (per mile steamed).
Most of the fuel is consumed by applying the last rpm of the engine. When the rpm are increased from 80%
to 100% fuel consumption is doubled.
A flow meter should be installed on board the trawlers so to make the fisherman to carefully monitor the fuel
consumption and to practice more economic trawling trips.
3. Improved Hull Forms
Even though the speed of a trawler could not be increased, much could be done in order to highly reduce the
hull resistance.
479
Fishing vessels are not equally power consuming and require highly spreading effective
powers/displacement at the same relative speed. This is due to the fact that even small modifications to the
hull shape could provide significant variations in its resistance and means that there is room to improve their
performances from a powering point of view.
In order to give some quantitative indications on the relationship between the geometry and the resistance of
the hull forms and to get to a merit rank, calculations have been made on a set of 8 commercial trawlers.
Tables 1a and 1b show their characteristics. For each ship, the total resistance has been calculated by the following methods: Van Oortmerssen, Darvin,
Takagi, Inoi, Nakamura, Lap, Henschke, Taggart and Ridgely Nevitt.
The RT (kg) values have been averaged and referred to the full load dispacement Δ (t) of the ship.
Table 1a. General characteristics of the fishing vessels
LWL
LBP B T D L/B B/T SWS
VES
SELS
A 27.12 23.80 7.40 2.96 3.50 3.216 2.500 237 2387 231.0 B 29.95 27.25 7.30 3.10 3.90 3.733 2.355 327 3287 305.3 C 27.14 23.80 7.00 3.06 3.60 3.400 2.288 276 2772 224.7 D 28.25 26.35 7.50 3.00 3.56 3.513 2.500 320 3220 256.0 E 29.30 27.30 6.80 2.98 3.50 4.015 2.286 288 2891 252.0 F 33.70 31.40 8.00 3.10 4.10 3.925 2.581 408 4098 255.1 G 27.71 24.50 8.00 2.77 4.00 3.063 2.890 304 3058 359.8 H 25.12 22.00 7.20 2.90 3.40 3.056 2.483 269 2701 217.2
Dimensions: lengths (m), area (m2), volume m3, weight (kN).
Table 1b. General characteristics of the fishing vessels (continued)
CB CP CWP CM AWP AM xCF
[*]
xCB [*] L/1/3
VES
SELS
A 0.459 0.568 0.701 0.807 123.4 17.9 -2.777 -0.920 3.844 B 0.529 0.612 0.765 0.862 193.3 21.3 1.627 0.274 3.956 C 0.541 0.787 0.927 0.687 154.4 14.7 -0.303 -0.503 3.657 D 0.540 0.646 0.815 0.836 161.0 18.8 -1.673 -0.530 3.851 E 0.523 0.693 0.829 0.755 152.8 15.3 -1.538 -0.023 4.136 F 0.519 0.612 0.776 0.849 154.1 16.7 -1.418 -0.181 4.235 G 0.560 0.670 0.787 0.836 151.8 17.8 -1.475 -0.975 3.643 H 0.583 0.696 0.863 0.837 136.5 17.8 -1.280 -0.490 3.412
Note : The (-) indicates that CF and CB lie astern the amidship
Table 2. Values of RT/Δ as a function of the relative speed V/L
Fig. 1 shows that, for speed V/L < 1, the hull forms C e E exhibit the highest specific resistance while the
other hull forms exhibit the same specific resistance. For speeds V/L > 1, the specific resistance is notably
V/L A B C D E F G H
0.2 0.19 0.20 0.18 0.19 0.22 0.20 0.18 0.18 0.4 0.76 0.82 0.70 0.73 0.86 0.78 0.68 0.68 1.0 5.61 5.77 7.95 5.58 7.27 6.00 6.14 6.53 1.2 11.70 11.12 18.30 10.66 13.47 11.04 15.25 16.37
1.4 23.02 20.70 43.31 20.07 25.32 20.90 29.87 32.91
480
different for the hulls, with gaps more than about 100%. In the field of V/L = 1.25, which is peculiar for
fishing vessels, the hull D, is able to reach the best speed with lesser power. The results obtained from
systematic model tests at naval towing tanks allow outlining some general rules, which could help a designer
to draw a hull shape of higher efficiency. Among the parameters which influence the performance of a hull,
the prismatic coefficient, the longitudinal position of the maximum sectional area, the centre of buoyancy,
the half angle of entrance, the shape of bow and stern, are the most important ones.
Fig. 1. Average values of specific resistance as a function of the relative speed
The following further suggestions could be given for better fishing vessel designs:
- shifting afterwards the center of buoyancy gives good results. It should be placed at about 0,3LWL astern the midship;
- the value of the block coefficient CB should be around 0.52; - the prismatic coefficient CP is mostly affecting the resistance. Some results from studies on this
coefficient allow to state that higher CP give higher resistances. Its optimal value, for fishing vessels, seems to be around 0.58-0.60;
- an entrance angle iE = 20° could be assumed for good performances; - a transom stern seems better than a rounded stern. Taking into account such considerations, a model of fishing vessel has been designed and tested in a naval
tank. The fore body of this basic hull has been replaced by a bulbous bow. This modified model was tank tested as well. Both the models represent a fishing vessel of the following
features:
Length between perpendiculars LBP = 26.40 m
RT/Δ
481
Load waterline length LWL = 8.00m Beam B = 6.75 m Draft D = 2.87 m Prismatic coefficient CP = 0.590 Block coefficient CB = 0.447 Full load displacement = 249 t Both towing and self propulsion results for the two models are reported in Table 3.
Fig. 2 shows that up to about 7.5 knots, the bulbous bow shows worse effective power characteristics than
the basic hull but, in the same speed range, the bulbous bow is better as to the delivered power.
This confirms that:
- the bulb positively acts on the propulsive efficiency, in particular on the hull efficiency and therefore its
performances are more efficient for any operating speed at least in this case.
- Both the basic and bulbous bow form showed lower power requests than a commercial vessel of same
displacement. Table 3. Effective (PE) and delivered (PD) powers for both basic (1) and bulbous bow form (2) at speeds (V)
V [knots]
PE [HP] PD [HP] 1 2 % 1 2 %
5 10 12 + 20.00 23 19 - 21.00 6 18 21 + 16.70 37 32 - 15.62 7 30 33 + 10.00 59 53 - 11.32 8 50 46 - 8.70 93 85 - 9.41 9 77 67 - 14.92 137 125 - 9.60 10 112 97 - 15.46 197 170 - 15.88 11 179 169 - 5.91 299 260 - 15.00 12 343 321 - 6.85 543 492 - 10.36 13 674 582 - 15.80 1109 967 - 14.68 14 1203 1112 - 8.18 2153 1931 - 11.50
Fig. 2. Effective (PE) and delivered (PD) power curves for a trawler with and without a bulbous bow
482
4. IMPROVED PROPULSION SYSTEMS The power plant of a trawler typically consists of a diesel engine driving a fixed blade propeller which
exibits its best efficiency only at its designed point. Therefore, the efficiency of a fixed blade propeller,
designed for steaming optimal performance, will drop when trawling.
The vice versa is as well true.
In order to improve the propulsive efficiency some effective devices could be suggested. 4.1. Ducted Stern
Such device (fig. 3) consists of a duct structure put ahead of the propeller. It will modify the ship’s wake.
Model tests with and without such device revealed energy savings (5-10%) due to lesser hull resistance.
Fig. 3. Ducted stern (source: Alain Le Duff, modified)
Propeller B.3.50 - P/D=0,8
050
100150200250300350400450500550600650700750800850900
1,0 1,1 1,2 1,3 1,4 1,5 1,6 1,7 1,8 1,9 2,0 2,1 2,2 2,3
propeller diameter [m]
Del
iver
ed P
ower
and
Pro
pelle
r rpm
m
Pd
N
Fig. 5. Relation between powers (PD), revolutions (N) and propeller diameters (D) for the same thrust
483
4.2. Stator This structure, consists of putting some lifting flaps on the stern strut in order to reduce the loss of cinetic
energy due to the rotation of the propeller wake. It could be applied together with a ducted stern, The
efficiency could be improved by 2 5%.
Fig. 4. Stator (source: Alain Le Duff, modified)
4.3. Slowly running propellers Such propellers will gives an improved propulsive efficiency by increasing the amount of water through the
propeller disc. The same thrust could be produced with lesser engine power by reducing rpm and increasing
the propeller diameter.
The diagram of fig. 5 shows, for a particular propeller, how much power is requested at different rpm and
propeller diameters, to produce a thrust of 6000 kg.
As a rule of thumb, when the propeller revolutions are halved and the diameter is increased by 1/3, the
required power (and then the fuel consumed) will be reduced by ¼. Such indications are usually applied to
new vessels.
but quite often some owners replace both the engine and the propeller even on their already working
trawlers.
Further, a reduction of the blades number is effective to the fuel consumption. 4.4. Ducted propellers
A ducted propeller, i.e. a propeller fitted around with a ring-shaped profile, will produce the same bollard
pull with lesser engine power. For a trawler, the use of a ducted propeller will be power-saving.
Due to its smaller diameter, if compared with a conventional propeller, it could be installed also on already
existing trawlers.
It could be said that, rpm being constant, a ducted propeller having a smaller diameter (-10%) than the
conventional one, will produce a greater thrust (+25%).
Figure 6 shows a comparison between the powers required by a ducted (PK) and a conventional propeller
(PC) to develop the same thrust.
484
The above statements are also supported by some bollard pull tests carried out on a trawler firstly equipped
with a free propeller and then with a ducted one. Their performances are listed in Table 4.
The main engine was developing a maximum continuous power of 550 hp at 500 rpm.
For each engine rpm, both the corresponding pulls and the exhaust temperatures were taken.
The data reported in Tables 4 and 5, allow to say that a ducted propeller:
- compared to a free one, even of lesser diameter, running at the same rpm, gives a mean thrust increase of
about 26%;
- the thrust being equal, the ducted propeller gives a mean power saving of about 32%.
Fig. 6. Power required by a ducted (PK) and a conventional propeller (PC) to develop the same thrust
Table 4. Performances of the ducted and unducted propellers
PROPELLER unducted ducted
Z Number of blades 4 3
D Propeller diameter 1600 mm 1500 mm
P Propeller pitch 1040 mm 1350 mm P/D Pitch ratio 0.65 0.9
Table 5. Comparison between the bollard pulls (T), delivered powers (PD) and exhaust temperatures (S)
at the same rpm (N) of an unducted (1) and a ducted (2) propeller
N [rpm]
T [kg] S [°C] PD [HP] T/PD 1 2 % 1 2 1 2 % 1 2 %
385 3380 4240 25.44 360 338 180 170 - 5.55 9.66 12.11 25.36 400 3640 4600 26.37 375 360 200 192 - 4.00 10.40 13.14 26.35 415 3920 4950 26.27 420 383 225 215 - 4.44 11.20 14.14 26.25
485
Table 6. Powers (PD) and rpm (N) at the same bollard pull (T), for an unducted (1) and a ducted (2) propeller
T [kg] N [rpm] PD [HP]
1 2 1 2 % 3500 392 350 189 128 - 32.27
4000 419 374 232 156 - 32.76
4.5. Grim Wheel A Grim wheel is working as a waterturbine powered by the propeller wake. It is placed then in the
slipstream of the propeller and can freely running around its own axis. Its diameter is about 20% larger than
the propeller.
The exceed disc area works as a propulsor.
For an existing propeller, the revolutions number is fixed and the Grim wheel is an attractive way to
virtually increase its diameter.
/
Fig. 7. Efficiency improvement by a Grim wheel The energy savings range from 5 to 12%. A Grim wheel could be applied either to new or to already existing
propellers (fixed or c.p. type) when a proper room is available. Higher fuel savings could be obtained when a
Grim wheel is used in association with heavily loaded propeller.
The improvement of efficiency [fig. 7] depends on the Dg/Dp ratio and on thrust loading CT, given by
22 DVakTCT
where: density of the water D propeller diameter k numerical factor (k = 0.3925) Va propeller advance speed T propeller thrust The overall efficiency of a (Grim wheel/propeller) combination is comparable to a slow running propeller,
whose diameter is equal to the vane-wheel.
486
The difference between both is the number of revolutions. The rpm of the Grim wheel/propeller
combination is larger than the slow running propeller, resulting in a lower cost for machinary and shaftings.
5. Conclusions Some results coming either from direct calculations or model tests, have been discussed in this paper. They
allow to briefly conclude that:
- It seems convenient to reduce the steaming speed in order to achieve some fuel saving rate.
- It is possible to state a set of hull parameters, particularly suitable for a lesser fuel consumer fishing
vessel;
- Trawlers should not have to be overpowered, hoping to realize higher steaming speeds. A displacement
ship, like a trawler, could reach only a maximum speed imposed by its length; overpowers mean then
wasted energy.
- For an useful evaluation of the fuel consumption a suitable fuel-meter should be placed on board the
trawlers.
- To obtain substantial fuel savings, tank tests should be done because they are the most efficient mean to
ascertain the hull performances.
- The practical results ratify the usefulness of nozzle propellers for trawlers.
- Reducing the number of blades will reduce fuel consumptions.
- High propeller diameters running at low rpm will better the efficiency.
References
Bledsoe, M., Stefun G., “A study of the effect of a bulb on the series 60, 0.60 block coefficient model“, DTMB Hydromechanics Laboratory, 1955
Messina, G., “A fuel consumption based method to measure the fishing effort”. EU Research 97/0073, 2001
Gueroult, E.R., “An approach to the design of the new types of fishing vessels“. Fishing Boats of the world:3, Fishing News Books Ltd., 1967
Messina, G.,“Validità del bulbo prodiero nelle carene dei motopescherecci“, Quad. Lab. Tecnol. Pesca, vol. .II, N. 2 suppl., 1975
Fiorentini, L., Messina G. & Pietrucci A., “Analisi delle prestazioni di un motore diesel nelle navi da pesca al traino“, Quad. Lab. Tecnol. Pesca, vol. .III, N. 1, 1981
Fiorentini, L., Messina G. and Pietrucci A., “Prestazioni e consumi del motopeschereccio a strascico. Gazzettino della Pesca“, N. 7, 1982
Le Duff A., “L’art d’améliorer les rendements “, Initiatives Techniques, France Pêche.
Messina, G., “Correlazione tra forme di carena e risparmio di combustibile nelle navi da pesca al traino“ in Risparmio di energia nella pesca al traino, 1984
Messina, G., Panarello A., “Considerazioni sulla scelta del propulsore più economico nei motopescherecci“ in Risparmio di energia nella pesca al traino, 1984
Messina, G., “Prora tradizionale e con bulbo: due carene a confronto“. Gazzettino della Pesca, N. 10, 1984
Digernes, T., “Practical Guide for Fuel saving on fishing vessels“, FAO, 1984
Messina, G., “Osservazioni ed analisi su una carena con bulbo per pescherecci“, Atti NAV '86, 1986
487
Messina, G., “Energy saving on trawlers: some factors related to design and propulsion“, FAO Studies and reviews, N.62, 1987
Messina, G., Repetti G. and Wittjen E., “Sull'applicazione delle eliche intubate nei motopescherecci. Risultati e prospettive“ in Technics and Technology in fishing vessels, 1987
De Cock, J., “Fuel conservation by means of passive and active control“, CETENA 25th Anniversary Symposium, S. Margherita Ligure, 1987
Messina, G., “Energy saving on trawlers: some factors related to design and propulsion“, FAO Studies and reviews, N.62, 1987
Messina, G., “Appraisal and experimental checking of the performances of conventional and ducted propellers“, ICES - International Council for the Exploitation of the Sea. Oostend., 1988
Messina, G., Susat L., “ Some notes on the design of trawlers” in "Technics and technology in fishing vessels, 1989
Alkan, A.D., Kaya S., Messina G., Nabergoj R., “Design aspects of some Medi-terranean fishing vessels“ in Technics and Technology in fishing vessels, pp.285-292, 1997
Messina, G.,“Technological contributions to fuel savings in trawlers”, Conference on Energy Efficiency in Fisheries, Brussels, 2006
Messina, G., “Some advices to design safe and effective fishing vessels”, Nor-Fishing Technolo-gy, Conference, 6-10 August, Trondheim, 2006