You are on page 1of 12

Ocean Engineering 59 (2013) 152–163

Contents lists available at SciVerse ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

A study on the behind-hull performance of marine propellers astern


autonomous underwater vehicles at diverse angles of attack
Arash Nemati Hayati, Seyed Mohammad Hashemi, Mehrzad Shams n
Department of Mechanical Engineering, K.N. Toosi University of Technology, Tehran, Iran

a r t i c l e i n f o a b s t r a c t

Article history: In this study, we investigated the performance of a typical propeller behind an autonomous underwater
Received 7 March 2012 vehicle (AUV) at a fully turbulent flow regime at different angles of attack by the application of
Accepted 2 December 2012 computational fluid dynamics (CFD) method. For this purpose, preliminarily, numerical results for the
Available online 8 January 2013
open water performance of the propeller were validated against the empirical data. Then, the study of
Keywords: the behind-hull effects and self-propulsion point of the AUV were carried out for diverse propeller to
Marine propeller hull diameter ratios (Dprop/Dhull). After all, the performance of the propeller with minimum behind-hull
Behind-hull performance effects was investigated at different angles of attack and the results were discussed for high and
AUV moderate propeller loads. Our findings demonstrate that despite similar hydrodynamic efficiency for
CFD method
propeller at different angles of attack, the generated thrust and consequently the required propeller
Angle of attack
torque were significantly higher than those in straight motion. In addition, for the prediction of the
behind-hull performance of propellers, the propeller thrust and torque coefficients in open water
condition were at least 8% over-estimated for AUV straight motion and 7.5% under-estimated for
descent at high angle of attack.
& 2012 Elsevier Ltd. All rights reserved.

1. Introduction Second group of studies were done on the prediction of the


propeller performance in non-uniform flow and cavitating condi-
In recent decades, computational fluid dynamics (CFD) meth- tions. Some of these studies include Watanabe et al. (2003),
ods have been developed remarkably on the basis of Reynolds Kawamura et al. (2006), Ji et al. (2010), Zhang (2010), in which
averaged Navier–Stokes (RANS) equations in the analysis of the effect of the propeller–hull interaction on the propeller
marine propellers. Although the numerical modeling of flow performance was neglected due to the complexity of mesh
around various solids is possible, in case of propellers it is difficult generation, as well as the large amount of time-consuming
owing to several factors such as geometry complexity, turbulent computations. Instead, experimental wake behind the hull in
flow occurrence, flow separation, boundary layer effects, wake the absence of the propeller was considered as the inflow velocity
field and cavitation incidence (Hayati et al., 2012). to the propeller. Although this was a useful assumption, it lacked
Many works were carried out on the performance prediction of the accuracy because the propeller–hull interaction was
conventional and high-skewed propellers by CFD method. These neglected in those simulations.
works may be categorized into two major groups: First group Despite the studies mentioned above, in case of propellers
includes the studies on the capability of this method for the behind AUV hulls, numerical studies have been rarely done to find
performance prediction of marine propellers in open water out the propeller–hull interaction and performance of the pro-
condition. Some of these works include Abdel-Maksoud et al. peller in descent or climb at constant angles of attack. Thus, this
(1998), Martı́nez-Calle et al. (2002), Rhee and Joshi (2005), study intends to provide new information on the hydrodynamic
Morgut and Nobile (2012). In these literatures, the accuracy of performance of AUV propellers at different angles of attack. Due
CFD method for the performance prediction of marine propellers to the scarce effect of the rear control rudders with respect to the
working in uniform inflow was totally explained. These studies propulsion system in axis-symmetric underwater vehicles, it was
were a basis for further implementation of CFD method for the not presented in this study. Different angles of attack with regard
studies in this field. to the posture of the underwater vehicle were considered for a
comprehensive comparison and different flow characteristics
were examined around the propellers in each case. The results
n
Corresponding author. Tel.: þ98 21 8867 7272. of this paper would be useful for AUV propeller designers to
E-mail address: shams@kntu.ac.ir (M. Shams). obtain a more accurate estimation on the propeller performance

0029-8018/$ - see front matter & 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.oceaneng.2012.12.014
A. Nemati Hayati et al. / Ocean Engineering 59 (2013) 152–163 153

Nomenclature Rp Hull total drag with propeller (N)


Rh Hull total drag without propeller (N)
J Advance ratio (–) Z Number of blades (–)
va Advance velocity at propeller plane (m/s) P/D Pitch ratio (–)
a Angle of attack deg, (1) Cp Pressure coefficient (–)
y Angle of rotation deg, (1) D Propeller diameter (m)
AE/AO Blade area ratio (–) n Propeller hydrodynamic efficiency (–)
c Chord length (m) T Propeller thrust (N)
F External body forces (N) Dprop/Dhull Propeller to hull diameter ratio (–)
r Fluid density (kg/m3) Q Propeller torque (N m)
m Fluid viscosity (kg/(m/s)) n Rotational speed of the propeller shaft (rps)
PN Free stream pressure (kPa) Ps Static pressure (kPa)
g Gravitational acceleration (m/s2) t/c Thickness ratio (–)
dh/D Hub ratio (–) KT Thrust coefficient (–)
v Hull advance velocity (m/s) t Thrust deduction factor (–)
L/Dhull Hull length to diameter ratio (–) KQ Torque coefficient (–)
L Hull overall length (m) w Wake (factor) (–)

at straight motion, as well as descent of the AUV at constant investigated, in which the horizontal direction intersects the hull
angles of attack. symmetry axis at a point in the middle length of the hull
(Fig. 1(b)).
Based on the limitations of control surfaces in the provision
2. AUV postures of lift forces the general range for the angle of attack in the
motion of AUV is between 01 and 151. Thus, in this study three
Considering the motion of an underwater vehicle at a totally different conditions were investigated, including straight motion
submerged level, two principal planar postures can be men- and the AUV descent at two constant angles of attack equal to
tioned: First, the straight movement of the vehicle with the a ¼51 and 151.
assistance of the propulsion system, while the control rudders
remain their pre-adjusted state in order to guarantee the straight
motion: This condition is reputed as zero angle of attack, 3. Models
illustrated in Fig. 1(a). In this situation, the horizontal motion
direction coincides with the hull symmetry axis; Second: the 3.1. Hull model
descent or climb of the vehicle with the simultaneous assistance
of the propulsion system and control rudders: This posture is A typical middle-cylindrical hull form (Fig. 2), which is the
reputed as positive or negative angle of attack for descent or most common class of AUVs, was designed for this study. The
climb, respectively. In this study, positive angles of attack were overall dimensions were considered as a scale model with total
length and outer diameter of 2.857 m and 0.286 m, respectively.
In order to minimize wake field effects and maximize hydro-
dynamic performance, the hull length to diameter ratio (L/Dhull)
was specified to the ideal value of 10. For the fore part, an elliptic
form was determined with major to minor diameter ratio of 3 to
minimize the overall drag on the basis of the literature by Suman
et al. (2010), which indicated that this value is appropriate for
minimizing the hull total drag. For the rear part a conic form with
201 tail angle was determined to minimize the wake effects
(Burcher and Rydill, 1994).

3.2. Propeller model

Wageningen B-screw propeller introduced by Troost (1940) is


still a well-known conventional series in marine industry and is
an appropriate candidate for different studies according to the
extensive set of experiments executed on this series. Therefore, a
typical B3-50 propeller with a moderate pitch ratio (P/D ¼0.8)

Fig. 1. Principal planar motions of the AUV. (a): Straight motion (a ¼ 01)
(b) Descent (a 401). Fig. 2. Main dimensions of the designed hull (mm).
154 A. Nemati Hayati et al. / Ocean Engineering 59 (2013) 152–163

was opted for this study. In the initial stage of determining the 4. Numerical solution
self-propulsion point of the AUV, the general criteria for the
propeller to hull diameter ratio (Dprop/Dhull), which is widely used 4.1. Governing equations
among the propeller designers was considered between 0.4 and
0.7 (Burcher and Rydill, 1994). In this study, the aim of the Computations were executed by the application of a finite
propeller selection was not only to balance the propeller thrust volume CFD package (FLUENT 6.3), which is established on the
with the hull drag, but also to minimize the behind-hull effects. basis of CFD codes and is a prominent commercial software
The latter purpose was mainly considered to obtain more package in various problems in the study of marine propellers
accurate predictions for the propeller behind-hull performance and AUVs according to several previous literatures in this field of
at non-zero angles of attack. study, such as: Martı́nez-Calle et al. (2002), Benini (2004), Barros
Thus, four propellers with different diameters equal to et al. (2008), Jagadeesh et al. (2009) etc.
D ¼0.114 m, 0.143 m, 0.172 m and 0.200 m were examined. The The governing equation for mass conservation in each principal
principal geometrical parameters of the propellers are shown in direction can be written as:
Table 1. Propeller model 4 with 0.200 m diameter is illustrated in @
Fig. 3. In order to offset the influence of hub geometry, the hub ðrui Þ ¼ 0 ð1Þ
@xi
configuration was not presented in this study. For all cases the
hub ratio (dh/D) and pitch ratio (P/D) were fixed to 0.169 and 0.8, where r is the fluid density, ui is the velocity component in each of
respectively, where dh, D and P are hub diameter, propeller the principal directions (x, y and z). In addition, momentum
diameter and propeller geometrical pitch, respectively. Conse- conservation equation in principal directions can be written as:
quently each propeller has unique hub diameter and geometrical @ @ @p @tij
ðrui Þ þ ðrui uj Þ ¼  þ þ rg i þ F i ð2Þ
pitch according to these ratios and its diameter. @t @xj @xi @xj
where Fi and gi represent body forces and gravitational accelera-
Table 1
tion and tij the Reynolds stress tensor defined as:
Main geometrical specifications of the modelled propellers.
 
@ui @uj 2 @ui
Number of blades (Z) 3 tij ¼ m þ  m d ð3Þ
@xj @xi 3 @xi ij
Blade area ratio (AE/AO) 0.5 where dij is the Kronecker delta, which is unity when i and j are
Pitch ratio (P/D) 0.8
equivalent and zero in other cases. Considering Eqs. (1)–(3), the
Hub ratio (dh/D) 0.169
Thickness ratio (t/c) at (0.7R) 0.0473 Reynolds averaged momentum equation is derived as:
Skew angle (1) 5    
@  @  @ @ui @uj 2 @ui @p
Rake angle (1) 15 (Backward) rui þ rui uj ¼ m þ  m 
Handedness Right-handed @t @xj @xj @xj @xi 3 @xi @xi
Section type Inner sections: airfoil @  
Outer sections: segmental
þ rui uj ð4Þ
@xj
Diameter (D) (m) Model 1: 0.114
Model 2: 0.143 For turbulence description, standard k–o model with auto-
Model 3: 0.172 matic wall functions (mixed formulation), developed by Wilcox
Model 4: 0.200
Chord length (c) at (0.7R) (m) Model 1: 0.0412
(1998) was employed where k and o represent turbulence kinetic
Model 2: 0.0517 energy and turbulence specific dissipation rate, respectively and
Model 3: 0.0621 are obtained from the following equations:
Model 4: 0.0723  
@ @ @u @ m @k
ðrkui Þ ¼ tij i rb ko þ mþ t
n
ðrkÞ þ ð5Þ
@t @xi @xj @xi sk @xi
 
@ @ o @u @ mt @k
ðroÞ þ ðroui Þ ¼ a tij i rbo2 þ mþ ð6Þ
@t @xi k @xj @xj so @xj
where sk and so indicate turbulent Prandtl numbers for k and o,
respectively and b and bn were obtained from experimental
formulations by Wilcox (1998). Furthermore, turbulent viscosity
mt is obtained from the combination of k and o according to the
following equation:
rk
mt ¼ ð7Þ
o
This turbulence model has high accuracy in the prediction of
flows which involve both velocity decrease and flow separation
astern the AUV hull. In addition, it is proved that the mentioned
model can accurately simulates the flow parameters in the
proximity of the propeller plane (Hayati et al., 2012). The auto-
matic wall function switches between standard low-Re formula-
tion and wall functions on the basis of the grid spacing of the
near-wall cell. Consequently, usage of this kind of wall treatment
leads to improving the accuracy of the flow simulation near the
walls surfaces. The selected solver for the equations was 3D
segregated. The employed algorithm for pressure–velocity cou-
pling was SIMPLE with standard discretization for pressure and
Fig. 3. B3-50 propeller model. second order upwind discretization for momentum, k and o. The
A. Nemati Hayati et al. / Ocean Engineering 59 (2013) 152–163 155

turbulence intensity of the inflow was set to 4%. The absolute 4.2. Grid generation
conversion criteria for the residuals of the equations including
continuity, momentum, turbulence kinetic energy and turbulence An unstructured hybrid mesh was employed for grid genera-
specific dissipation rate were set to 10  5 throughout the study to tion (Fig. 4). For the propeller, blades surfaces were meshed with
obtain the results with acceptable accuracy. triangular cells. Smaller triangles with the sides of approximately
0.001D (where D is the propeller diameter) were used for regions
near to the root, blades edges and tips for better flow simulation
in these regions. The average value of y þ on the propeller surfaces
at all operating conditions was 40 with 13 elements in the
turbulent boundary layer which allows the simulation of the flow
in the proximity of the propeller surface with suitable accuracy.
Note that y þ is defined as:
ut r
yþ ¼ y ð8Þ
m
where y, r, m are the normal distance from the wall, fluid density
and viscosity, respectively, and ut is the friction velocity defined
as:
 0:5
tw
ut ¼ ð9Þ
r
And tw is wall shear stress. A moving mesh zone was
generated around the propeller for better prediction of the flow
between the propeller blades. For the hull, small triangles with
sides of approximately 0.007L (where L is the hull overall length)
were used for the whole body. The average value of y þ on the hull
surface at all operating conditions was 30 with 14 elements in the
turbulent boundary layer. Other surfaces were meshed with
larger triangles with sides of approximately 0.14L. Finally the
remaining regions in the domain were meshed with appropriate
growing rate of 1.2 with tetrahedral cells.

4.3. Boundary and operating conditions

In order to simulate the AUV motion at different angles of


attack in a fully submerged level and investigate the behind-hull
performance of the propeller, a cylindrical control volume was
assumed around the connected hull and propeller with inlet and
outflow boundary conditions for the inflow to the hull and the
flow exit astern the propeller, respectively (Fig. 5). A no-slip wall
condition was used on all solid surfaces, including (hull and
propeller). The system was surrounded by external walls located
far from the propeller. The domain dimensions (Table 2) were
considered large enough to avoid external effects such as reversed
flows and induced pressure on the performance of the hull and
propeller.
Fig. 4. Generated mesh for the numerical solution: (a) General view (b) Astern the The rotating speed of the propeller shaft was set to 25 rps
hull. (revolutions per second) throughout this study. The inflow

Fig. 5. Arranged control volume for the numerical solution.


156 A. Nemati Hayati et al. / Ocean Engineering 59 (2013) 152–163

Table 2 the former corresponds to the decrease of the propeller load.


Global domain dimensions.
T
KT ¼ ð11Þ
Distance from the frontal tip of the hull to rn2 D4
Inlet boundary Outlet boundary Radial boundary Q
KQ ¼ ð12Þ
L 3L 0.5L rn2 D5

KT J
Z¼  ð13Þ
P/D=0.8
K Q 2p
1
KT (Hayati et al. (2012))
where T and Q are the propeller generated thrust and torque at
10K (Hayati et al. (2012))
the open-water performance, KT and KQ are thrust and torque
0.9 Q
coefficients, respectively, which represent the amounts of the
η (Hayati et al. (2012))
KT (Exp.) propeller generated thrust and torque. Z is the propeller hydro-
0.8
10KQ (Exp.) dynamic efficiency, which represents the ratio of the generated
η (Exp.)
thrust to the required torque for the propeller rotation.
0.7 It was obtained that the average amount of error for the
prediction of the propeller thrust and torque coefficients with
0.6
current numerical method was about 2.5% and 12% for thrust and
torque coefficients, respectively, which demonstrate the capabil-
K , 10K , η

ity of the executed numerical method (Hayati et al. (2012)).


Q

0.5
T

0.4
5.2. Behind-hull performance

First, the study of the hull resistance in the absence of the


0.3
propeller was carried out for four different mesh resolution levels
at Re¼4.07  106 (Table 3). In each case, the average amount of
0.2 y þ was maintained to 30 on all solid surfaces. Due to the axis
symmetric hull profile, the steady-state simulations were used.
0.1 The predicted drag for meshes with fine resolution was very close
to that for meshes with very fine resolution. Thus, the results
0 were obtained on the fine meshes and the hull total drag at
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 different inflow velocities is shown in Table 4.
J One of the primary influences of a propeller behind an under-
water vehicle is the augmentation of the hull total drag with
Fig. 6. Comparison between the numerical and empirical results for the propeller
respect to that in the absence of the propeller. This fact, which is
model 4 in open water condition.
considered as an important point in the process of the propeller
design for an underwater vehicle is represented as thrust deduc-
velocity varied between 1.43 m/s and 2.50 m/s, which corre-
tion factor defined as:
sponds to the length-based Reynolds numbers of 4.07  106 and
7.11  106, respectively. The flow regime was considered fully Rh
t ¼ 1 ð14Þ
turbulent same as a previous study by Philips and Turnock (2010) Rp
at Re 4 106 as it was proved that the assumption of transition
flow around prolate spheroids at Re4106, does not improve the Table 3
prediction of flow characteristics near the fore part of the hull Hull total drag without propeller with different mesh resolution levels at
(Kim et al., 2003; Clarke et al., 2008). Re¼ 4.07  106.

Mesh Mesh size Mesh size Number of Hull total drag


resolution at the hull far from cells without
5. Validation level surface (L) the hull propeller (Rh)
(L) (N)

5.1. Open water performance Coarse 0.010 0.14 1,650,336 24.34


Mid 0.008 0.14 2,884,284 32.85
In a previous literature by the same authors (Hayati et al., Fine 0.007 0.14 4,105,374 36.38
2012) the hydrodynamic performance of the propeller with Very fine 0.006 0.14 5,129,532 36.41
D ¼0.200 m and P/D ¼ 0.8 m (model 4 in Table 1) was compared
with experimental data for a range of advance ratios in open
water condition (Fig. 6). The principal parameters describing the
performance of the propeller are as follow: Table 4
Hull total drag without propeller at different inflow velocities.
va
J¼ ð10Þ Re v (m/s) Hull total drag without
nD
propeller (Rh) (N)
where va, n and D represent the advance velocity, rotational speed
of the propeller shaft (rps) and the propeller diameter. J is the 4.07  106 1.43 36.38
advance ratio, which is a kinematic parameter representing the 5.09  106 1.79 61.92
6.11  106 2.15 88.73
velocity seen by the propeller. The relation between the advance 7.11  106 2.50 119.54
ratio and the propeller load indicates that the augmentation of
A. Nemati Hayati et al. / Ocean Engineering 59 (2013) 152–163 157

where Rh is the hull total drag in the absence of the propeller at a propellers with different diameters were obtained by the numer-
determined advance velocity, while Rp is the hull total drag in the ical method and the results are compared to the theoretical
presence of the propeller. In order to obtain this factor, four results presented by Burcher and Rydill (1994) in Fig. 7. In that
propellers with different diameters were attached to the hull and literature, the investigated hull profile was the same as the
the predicted drag in each case was obtained at inflow velocities designed hull in this study, although the dimensions of the model
between 1.43 m/s and 2.50 m/s (Table 5). Note that the operating and the flow regime were not presented there and these factors
conditions of the propellers were set to n ¼25 rps and J ¼0.5 rps. may cause scarce discrepancies between the obtained results. It is
The increase in the hull total drag in the presence of the shown that by increasing the propeller diameter, the propeller
propellers is well shown in all cases with respect to the hull total influence on the hull total drag and consequently the thrust
drag without propeller (Table 4). The thrust deduction factors for deduction factor decreases significantly.
The pressure coefficient of the designed hull without/with the
Table 5
propellers with D ¼0.114 m and 0.200 m are shown in Fig. 8. Note
Hull total drag with propeller and the propeller generated thrust at different that the pressure coefficient is a ratio of the relative static
inflow velocities. pressure to the dynamic pressure defined as:

Re v J Propeller n Hull total drag Propeller P s P1


Cp ¼ ð15Þ
(m/s) diameter (rps) with propeller thrust 1=2rv2
(D) (m) (Rp) (N) (T) (N)
where Ps and PN represent the static and free stream pressure at the
4.07  106 1.43 0.5 0.114 25 38.63 16.65 hull proximity and far field, respectively, r is the fluid density and v
5.09  106 1.79 0.5 0.143 25 64.81 41.22
6.11  106 2.15 0.5 0.172 25 91.96 86.27
represents the inflow velocity to the hull. The fluctuations in the
7.11  106 2.50 0.5 0.200 25 121.85 157.72 pressure coefficient are due to the changes in the flow regime along
the hull length from the flow inlet at the hull fore (x/L¼0) to the hull
tail (x/L¼1) and the turbulent boundary layer effects. Significant
0.07 reduction in the pressure field was occurred at the joints of the fore
Num and rear part to the cylindrical part of the body that is mainly due to
Burcher and Rydill
the changes in the hull profile. In the presence of the propeller, the
0.06
flow pressure around the hull tail decreases, which leads to higher
pressure difference between the frontal and rear part of the hull, the
0.05
augmentation of the pressure drag and consequently the increase in
the hull total drag. Higher propeller diameter induces less reduction
on the pressure around the hull tail (x/L¼1) with respect to lower
0.04 propeller diameters (Fig. 8). Therefore, it is derived that propeller
t

with the largest diameter (D¼0.200 m) has the least effect on the
hull total drag and consequently has the minimum thrust deduction
0.03 factor as previously shown in Fig. 7.
Another important behind-hull effect of the propeller is the
variation of wake field behind the hull and the induced velocities
0.02
around this location. Wake factor is a practical parameter con-
sidered in the procedure of propeller design for underwater
0.01 vehicles and is defined as:
0.4 0.5 0.6 0.7  va 
D /D w ¼ 1 ð16Þ
prop hull v

Fig. 7. Comparison between the numerical and theoretical results for thrust where va is the actual inlet velocity to the propeller, v the advance
deduction factor. velocity and w the wake factor, which is a dimensionless

2
6
Re = 4.07×10 , Without propeller
6
Re = 4.07×10 , Dprop = 0.114 (m), J = 0.5
1.5 6
Re = 7.11×10 , Dprop = 0.200 (m), J = 0.5

1
Cp

0.5

−0.5

−1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x/L

Fig. 8. Longitudinal distribution of pressure coefficient on the the hull, without/with propeller.
158 A. Nemati Hayati et al. / Ocean Engineering 59 (2013) 152–163

Fig. 9. Wake distribution astern the hull: (a) Re¼ 4.07  106, without propeller (b) Re¼ 4.07  106, D¼ 0.114 m, J¼ 0.5 (c) Re¼ 7.11  106, without propeller
(d) Re ¼7.11  106, D¼ 0.200 m, J¼ 0.5.

parameter indicating the wake field strength at the behind of the 1


hull. Note that nominal wake is defined as the wake field behind the KT
hull in the absence of the propeller, while total wake is defined as 0.9 10KQ
the wake field behind the hull with the attached propeller. Based on
η
Eq. (16), the wake field behind the hull for D¼0.114 m and 0.200 m 0.8
are represented in Fig. 9. A remarkable distinction is observed
between the nominal and total wake fields due to the propeller 0.7
influence. The distribution of the nominal wake (Fig. 9(a) and (c))
are symmetrical; Furthermore, as the radial distance from the centre 0.6
KT, 10KQ, η

of the plane located on the hull symmetry axis increases, the wake
effect diminishes until it disappears at external radial distances. 0.5
Considering Fig. 9(b) and (d), wake distribution changes in a way
that in locations close to the propeller blades, a negative wake 0.4
region appears in both cases. This fact is due to the propeller
induced flow velocities, which enhance the flow velocity astern 0.3
the hull (va) to the values higher than the advance velocity (v) and
therefore, according to Eq. (16), negative wake is observed. Further- 0.2
more, the diagrams indicate that in case of larger propellers
(D¼0.200 m), wake effect decreases slightly with respect to small 0.1
propellers (D¼0.114).
Considering the behind-hull effects explained above, the 0
largest propeller with D¼ 0.200 m had the minimum behind- 0 0.2 0.4 0.6 0.8
hull effects. In the meantime, the study on the self-propulsion J
point of the AUV was done for the propellers with the operating
conditions mentioned in Table 5. The behind-hull hydrodynamic Fig. 10. Behind-hull performance of the propeller at straight motion (a ¼ 01).
performance of the propellers at straight motion (a ¼ 01) was
obtained and is shown in Fig. 10. Assuming the operating point of the study, not only for its minimum hull-interactions, but also for
the propellers at J¼0.5, the amounts of propeller thrust for the consideration of the self-propulsion point of the AUV.
different propeller diameters, were compared in Table 5 with
the hull total drag in the presence of the propellers. Apparently,
the results indicate that the propeller with D ¼0.200 m provides 6. Results and discussion
the self-propulsion point and the rest of the propellers produce
lower thrust than the hull total drag with the attached propellers. In Section 5.2, all simulations were executed at zero angle of
Thus, the propeller with D ¼0.200 m was selected for the rest of attack, which corresponds to the straight motion of the AUV at a
A. Nemati Hayati et al. / Ocean Engineering 59 (2013) 152–163 159

0.35 fully turbulent flow regime. The behind-hull performance of the


0 deg selected propeller at different angles of attack is represented in
5 deg
Fig. 11. It should be mentioned that in the simulations at non-zero
15 deg
0.3 angles of attack, the unsteady computations were executed with a
time-step size of 1.11  10  4 s, which corresponds to the rotation
angle of 1.01, and the computations were continued for several
0.25 complete cycles until no significant variations were observed in
the amounts of the propeller thrust and torque coefficients.
Note that in case of descent in xy plane (Fig. 1(b)), the
KT

0.2 propeller thrust and torque are obtained from the following
equations:
T ¼ T x cos a þ T y sin a ð17Þ
0.15

Q ¼ Q x cos a þ Q y sin a ð18Þ


0.1 where Tx, Ty, Qx and Qy are the propeller thrust and torque
components in principal directions in the motion plane
(Fig. 1(b)) and a is the angle of attack. Comparing the amounts
0.05 of the generated thrust at different angles of attack, it is derived
0 0.2 0.4 0.6 0.8 that in descent of the AUV at high angles of attack, more thrust is
J generated by the propeller, with respect to low angles of attack
0.4 and straigth motion (Fig. 11(a)). Generation of more thrust, leads
0 deg to the requirement for more power for the propeller rotation as
5 deg Fig. 11(b) indicates that more torque is needed for the propeller
15 deg
rotation at high angles of attack. However, in case of propeller
0.35
hydrodynamic efficiency no remarkable increase is observed in
the behind-hull efficiency of the propeller at different angles of
attack (Fig. 11(c)). For a better comparison, the amount of
0.3 increase in the propeller thrust and torque coefficients from
a ¼01 to 151 are shown for a range of advance ratios in Fig. 12.
10KQ

Note that the increase percentage is defined as:

0.25 K T a ¼ 15 K T a ¼ 0
DK T % ¼  100 ð19Þ
K Ta ¼ 0

K Q a ¼ 15 K Q a ¼ 0
DK Q % ¼  100 ð20Þ
0.2 KQa ¼ 0

It is observed that for a wide range of advance ratios, including


high and normal propeller loads the amount of increase in
0.15 propeller thrust and torque coefficients from straight motion to
0 0.2 0.4 0.6 0.8 descent at a ¼ 151is approximately 17%. This fact shows that the
J required power for the propeller rotation at high angles of attack
is significantly higher than that at zero angle of attack.
0.6 In order to compare the flow characteristics around the
propeller at a ¼01 and a ¼151, first the static pressure distribution
on the propeller was investigated for high and normal propeller
0.5 loads (J¼0.245 and 0.5, respectively) and is shown in Fig. 13. It is
derived that in both advance ratios the overall range of the flow
pressure at the pressure side of the propeller is higher at a ¼151
0.4 with respect to straight motion (a ¼01). This fact, which is well
monitored on the region between the mid-chord line and leading
η

edge leads to higher thrust and torque generation as discussed in


0.3 Fig. 11(a) and (b) previously. Another important point in this
comparison is that in case of descent, the pressure distribution on
the propeller blades are not identical unlike the case of straight
0.2 movement. Fig. 13(b) and (d) show that the pressure distribution
on each of the blades depends on the blades position with respect
0 deg
5 deg to the flow entering the propeller. For a better comparison,
0.1 15 deg considering the chordwise pressure distribution on the propeller
blades at r/R ¼0.7 at high and normal loads (J¼0.245 and 0.5,
respectively) it can be seen that the range of the pressure
0 0.2 0.4 0.6 0.8 coefficient on the propeller blades at a ¼151 is different for each
J of the blades (Figs. 14 and 15). Note that y is the angle of rotation,
Fig. 11. Comparison between the behind-hull performance of the propeller
which determines the blades location in a way that y ¼01
at different angles of attack: (a) Thrust coefficient, (b) Torque coefficient and corresponds to the upper blade and the direction of rotation is
(c) Hydrodynamic efficiency. clockwise from the pressure side view of the propeller (Fig. 13). It
160 A. Nemati Hayati et al. / Ocean Engineering 59 (2013) 152–163

Fig. 12. Increase in the propeller thrust and torque coefficients from a ¼01 to 151.

Fig. 13. Pressure distribution on the pressure side of the propeller: (a) J¼ 0.245, a ¼01, (b) J¼ 0.245, a ¼151, (c) J¼ 0.5, a ¼ 01 and (d) J¼ 0.5, a ¼ 151.

can be seen that for descent at a ¼151 at both advance ratios, the the unsteadiness of the flow around the propeller at high angles
difference between the pressure coefficient on the pressure and of attack. Furthermore, the overall range of the velocity is
suction side of the upper blade is more than the other blades decreased for high angle of attack comparing to straight motion,
(Fig. 14(b) and Fig. 15(b)). This event may cause undesired which correlates to the analysis of static pressure around the
vibrations on the propeller shaft during its rotation, which lead propeller and consequently the difference in propeller perfor-
to induced vibrations on the hull at descent and consequently mance at different conditions as discussed previously.
cause difficulties in the stability control of the vehicle. Besides, in order to compare the total wake distribution astern
In order to investigate blade to blade inflow velocity to the the hull at high angle of attack, the total wake distribution at
propeller at different angles of attack, the distribution of the a ¼151 is shown in Fig. 17 for a moderate advance ratio. It is
actual inflow velocity to the propeller is shown for high and observed that in case of descent, the general range for the wake
normal propeller loads in Fig. 16. It is observed that in descent, factor increases with respect to straight motion (Fig. 9(d)), while
the region of maximum velocity was relocated from the tip of the the wake field is not distributed steadily unlike the straight
blades to internal sections. This fact is more remarkabale in the motion. The regions with negative wake factors were induced
upper blade (y ¼01) at a ¼151 (Fig. 16(b) and (d)), which is due to proportional to the blades positions, which corresponds with the
A. Nemati Hayati et al. / Ocean Engineering 59 (2013) 152–163 161

α=0 (deg), J=0.245 α=15 (deg), θ=0 (deg), J=0.245


40 30
Pressure side Pressure side
30 Suction side 20 Suction side

20
10
10
0
p

p
0
C

C
−10
−10
−20
−20

−30 −30

−40 −40
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x/L x/L

α=15 (deg), θ=120 (deg), J=0.245 α=15 (deg), θ=240 (deg), J=0.245
30 30
Pressure side Pressure side
Suction side 20 Suction side
20

10 10

0 0
p

Cp
C

−10 −10

−20 −20

−30 −30

−40 −40
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x/L x/L

Fig. 14. Chordwise distribution of pressure coefficient at r/R ¼0.7, J¼ 0.245: (a) a ¼01, (b) a ¼151, y ¼01, (c) a ¼ 151, y ¼ 1201, (d) a ¼151, y ¼ 2401.

α=0 (deg), J=0.5 α=15 (deg), θ=0 (deg), J=0.5


6 8
Pressure side Pressure side
Suction side 6 Suction side
4
4
2
2
p

0 0
C

−2
−2
−4
−4
−6

−6 −8
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x/L x/L

α=15 (deg), θ=120 (deg), J=0.5 α=15 (deg), θ=240 (deg), J=0.5
8 6
Pressure side Pressure side
6 Suction side Suction side
4

4
2
2
p

0
C

0
−2
−2

−4 −4

−6 −6
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x/L x/L

Fig. 15. Chordwise distribution of pressure coefficient at r/R¼ 0.7, J ¼0.5: (a) a ¼ 01, (b) a ¼151, y ¼01, (c) a ¼ 151, y ¼ 1201 and (d) a ¼ 151, y ¼2401.
162 A. Nemati Hayati et al. / Ocean Engineering 59 (2013) 152–163

Fig. 16. Distribution of the inflow velocity to the propeller blades. (a) J ¼0.245, a ¼ 01, (b) J¼ 0.245, a ¼151, (c) J¼ 0.5, a ¼01 and (d) J ¼0.5, a ¼151.

to the open water condition. The changes were averaged for


the investigated advance ratios, namely J¼ 0.1, 0.245, 0.5 and
0.676.
It can be seen that there is a reduction of 7.6% and 8.3% in the
values of the propeller thrust and torque coefficients, respectively
at zero angle of attack with respect to the open water perfor-
mance, while for descent, an increase of 6% and 5.3% can be seen
in the mentioned values, respectively at a ¼51, and for high angle
of attack (a ¼151) an increase of 7.7% and 7.4% is shown for the
propeller thrust and torque coefficients, respectively (Fig. 18).
This comparison demonstrates that the propeller thrust and
torque coefficients in open water condition are at least 8% over-
estimated for straight motion of the AUV, while in case of descent
at high angle of attack, they were at least 7.5% under-estimated
and the designer should consider these changes in the process of
the AUV propeller design for increasing the accuracy of propeller
selection and behind-hull performance prediction, on the basis of
the propellers open water performance diagrams.
Fig. 17. Total wake distribution astern the hull at J ¼0.5, a ¼151.

7. Conclusion
actual inflow velocity to the propeller shown previously in
Fig. 16(b) and (d). The behind-hull performance of four conventional propellers
Finally, based on the obtained results it would be fruitful to with different diameters were examined in detail. Different
compare the open water and behind-hull performance of the parameters were investigated astern the hull and it was shown
propeller to provide an estimation for the prediction of the AUV that high propeller to hull diameter ratio (Dprop/Dhull ¼0.7) mini-
propeller performance at different angles of attack based on the mzed the effect of behind-hull interaction. The hydrodynamic
available open water diagrams. The comparison was executed performance of the selected propeller was investigated at differ-
based on the following equations: ent angles of attack. The results indicated that for high angles of
attack more thrust was generated but at the same time more
K T a K T Open water
DK T % ¼  100 ð21Þ torque was required for the propeller rotation, which led to more
K T Open water
required power at non-zero angles of attack, while the propeller
K Q a K Q Open water behind-hull efficiency was not varied significantly. The examina-
DK Q % ¼  100 ð22Þ tion of the flow characteristics around the propellers at different
K Q Open water
angles of attack showed that induced vibrations on the hull was
DKT and DKQ are the amount of changes in the propeller thrust more possible at high angles of attack on the basis of unsteady
and torque coefficeint at different angles of attack with respect flow characteristics on the propeller blades. Besides, the study
A. Nemati Hayati et al. / Ocean Engineering 59 (2013) 152–163 163

Fig. 18. Average changes in the propeller thrust and torque coefficients at different angles of attack with respect to open water condition.

proved that the propeller thrust and torque coefficients in open marine propeller using a RANS CFD code. In: Sixth International Symposium
water condition are at least 8% over-estimated for straight motion on Cavitation (CAV2006), Wageningen, The Netherlands.
Kim, S., Lebanon, N.H., Rhee, S., Cokljat, D.., (2003). Application of modern
of the AUV, while in case of descent at high angle of attack, they turbulence models to vortical flow around a prolate spheroid. In: Proceedings
were at least 7.5% under-estimated and the designer should of 41st Aerospace Sciences Meeting and Exhibit, 6–9 January, Reno, NV, USA.
consider these changes in the process of the AUV propeller design. Martı́nez-Calle, J., Balbona-Calvo, L., Gonzalez-Perez, J., Blanco-Marigorta, E., 2002.
An open water numerical model for a marine propeller: a comparison with
experimental data. In: Proceedings of the ASME FEDSM’02, 2002 Joint US-
References European Fluids Engineering Summer Conference, 14–18 July, Montreal,
Canada.
Morgut, M., Nobile, E., 2012. Influence of grid type and turbulence model on the
Abdel-Maksoud, M., Menter, F., Wuttke, H., 1998. Viscous flow simulation for
numerical prediction of the flow around marine propellers working in uniform
conventional and high-skew marine propellers. J. Ship Technol. Res. 45, 64–71.
inflow. Ocean Eng. 42, 26–34.
Barros, E.A., Dantas, J.L.D., Pascoal, A.M., Sa, E., 2008. Investigation of normal force
Philips, B.A., Turnock, R.S., 2010. Influence of turbulence closure models on vertical
and moment coefficients for an AUV at nonlinear angle of attack and sideslip
flow field around a submarine body undergoing steady drift. J. Mar. Sci.
range. IEEE J. Oceanic Eng. 33 (No. 4).
Benini, E., 2004. Significance of blade element theory in performance prediction of Technol. 15, 201–217.
marine propellers. Ocean Eng. 31, 957–974. Rhee, S.H., Joshi, S., 2005. Computational validation for flow around marine
Burcher, R., Rydill, L., 1994. Concepts in Submarine Design. Cambridge Ocean propeller using unstructured mesh based Navier–Stokes solver. JSME Int. J.
Technology Series 2. Cambridge University Press. B-Fluid T. 48 (3), 562–570.
Clarke, D.B., Brandner, P.A., Walker, G.J., 2008. Experimental and computational Suman, K., Rao, D.N., Das, H.N., Kiran, G.B., 2010. Hydrodynamic performance
investigation of the flow around a 3-1 prolate spheroid. WSEAS Trans. Fluid evaluation of an ellipsoidal nose for a high speed under water vehicle. JJMIE 4
Mech. 3, 207–217. (5), 641–652.
Hayati, A.N., Hashemi, S.M., Shams, M., 2012. A study on the effect of the rake Troost, L., 1940. Open water test series with modern propeller forms. II. Three
angle on the performance of marine propellers. Proc. Inst. Mech. Eng. Part C J. bladed propellers. Trans. NECIES, 56.
Mech. Eng. Sci. 226, 940–955. Watanabe, T., Kawamura, T., Takekoshi, Y., Maeda, M., Rhee, S.H., 2003. Simulation
Jagadeesh, P., Murali, K., Idichandy, V.G., 2009. Experimental investigation of of steady and unsteady cavitation on a marine propeller using a RANS CFD
hydrodynamic force coefficients over AUV hull form. Ocean Eng. 36, 113–118. code. In: Fifth International Symposium on Cavitation (CAV2003). 1–4
Ji, B., Luo, X.W., Wu, Y.L., Liu, S.H., Xu, H.Y., Oshima, A., 2010. Numerical November, Japan, Osaka.
investigation of unsteady cavitating turbulent flow around a full scale marine Wilcox, D.C., 1998. Turbulence Modeling for CFD, second ed. DCW Industries, Inc.,
propeller. J. Hydrodyn. 22 (5), 747–752. La Canada, California.
Kawamura, T., Takekoshi, Y., Yamaguchi, H., Minowa, T., Maeda, M., Fujii, A., Zhang, Z.R., 2010. Verification and validation for RANS simulation of KCS container
Kimura, K., Taketani, T., 2006. Simulation of unsteady cavitating flow around ship without/with propeller. J. Hydrodyn. 22 (5), 932–939.

You might also like