Brodogradnja/Shipbuilding Volume 67 Number 2, 2016.

Nowrouz Mohammad Nouri

Saber Mohammadi

http://dx.doi.org/10.21278/brod67202 ISSN 0007-215X

eISSN 1845-5859

A MULTI-OBJECTIVE APPROACH FOR DETERMINING THE

NUMBER OF BLADES ON A NACA MARINE PROPELLER

UDC 629.5(05) 629.5.035:629.5.024.71

Original scientific paper

Summary

This paper numerically discusses how the performance of a NACA marine propeller is affected
by the number of blades, which is one of the most crucial geometrical parameters determining
the performance of a propeller. Results are presented in terms of the hydrodynamic and
structural parameters. The results show that changing the number of blades changes the
hydrodynamic efficiency, torque, thrust, cavitation behaviour and structural stiffness of the
propeller nonlinearly. Furthermore, it is shown that the propellers structural lifetime is
shortened by increasing the number of blades. Hence, the propeller’s number of blades is a
multi-objective function and will be discussed in this research. The applied tool which is used
to study the hydrodynamic performance of the propellers is a RANS-based CFD one and the
FEM is considered to study the structural behaviour of the propellers.

Keywords: marine propeller; number of blades; CFD; FEM; multi-objective function

1. Introduction
The geometry of a marine propeller is very complicated because of the numerous
geometrical parameters involved, such as number of blades, diameter, pitch ratio, and camber
and thickness distributions. The designer must adjust these parameters so that the propeller
produces the required thrust and has the minimum possibility of cavitation occurring,
acceptable efficiency, low vibration, and high structural lifetime. Each parameter has specific
effects on the designer’s objectives and must be carefully adjusted. For example, increasing the
pitch ratio of a propeller augments the thrust produced by the propeller and would nonlinearly
change its hydrodynamic efficiency. Furthermore, changing the pitch ratio of the blades of a
propeller alters the angles of the blades relative to the axis of rotation and affects their structural
behaviour. Therefore, its increment would reinforce the blades against the bending in the axial
direction.
The developments which have occurred in computational resources have made
researchers use numerical methods in order to accurately analyze the flow fields around
complicated devices. Therefore, the hydrodynamics researchers have been interested in
simulating the flow around the propellers by means of RANS-based Computational Fluid
Dynamics (CFD) methods for the last decade. Several research projects have been performed

15
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

on marine propellers in open water or wake conditions using numerical methods [1-8]. For
example, Ji et al. (2011) [6] performed an unsteady numerical simulation of cavitating turbulent
flow around a highly skewed marine propeller. They used the commercial code CFX based on
Reynolds-averaged Navier-Stokes (RANS) equation with SST-k-ω turbulence model. The
maximum difference of pressure fluctuation between the calculation and experiment reached
20% in this study. Peng et al. (2013) [7] investigated the effects of turbulence models on
predicting the performance of a specific marine propeller in the open water condition. They
showed that eddy-viscosity turbulence models predicted the hydrodynamic coefficients of the
propeller more accurately than Reynolds-stress models. Furthermore, they concluded that all of
the eddy-viscosity turbulence models predicted about the same performance for the propeller.
In spite of the extensive research done on predicting the performance of propellers, the effects
of geometrical parameters on the performance of propellers have rarely been discussed. For
example, skew and rake are among the typical investigated parameters by the researchers in
recent years [1, 3, 5].
One of the most important geometrical parameters of a propeller is the number of blades. This
parameter is related to the propeller’s geometrical configuration, the occurrence of cavitation,
the extent of deflection of the blades (vibrational behaviour), the propeller’s structural lifetime,
and its hydrodynamic efficiency. The importance of the subject is revealed in the design
procedure of a marine propeller using a specific mass and a specific expanded area ratio. Note
that this ratio is estimated for enabling the propeller overcomes the cavitation occurrence.
Hence, the effects of the division number of the mentioned properties which is indicative of
how the number of blades affects the performance of the propeller should be discussed before
starting the design process. The designer has to exactly know what happens if he increases or
decreases the number of blades even one. As mentioned before, the determination procedure of
the number of blades of a propeller is a multi-objective approach, which was the topic of this
research. The 4-digit NACA airfoils are selected as the cross sections of the propellers in this
research because they could be simply modelled and also provide a good comparability between
the results of the research. Furthermore the 4-digit NACA airfoils have acceptable similarities
to the applied cross sections in the geometries of common propellers. Moreover, the overall
shape of the blades is determined similarly to the basic shape of the un-skewed marine
propellers before modifying their edges, especially in the outer radius ratios. In this way, the
considered propellers in this research are similar to common marine ones and the obtained
results of this research are reliable in the case of such propellers. The obtained results would
help the designers to make accurate decisions on determining the number of blades on marine
propellers.
In this research, a finite volume RANS-based CFD tool is used to predict the
hydrodynamic performance of the propellers in the steady state way. Furthermore, a FEM tool
is used to predict their structural behaviour. The hydrodynamic analyses of the propellers are
performed and the obtained pressure distributions around the propellers are used in FEM
models to evaluate the structural stresses on the blades and the resulting deformations.

2. Numerical simulations
2.1 Computational Fluid Dynamics
The governing equations for analysing the flow fields are consisted of momentum, mass,
and energy conservation equations. In this research, since isothermal flow is assumed, the
energy conservation equation is not needed. Through applying the Reynolds averaging method
on the governing equations of an incompressible Newtonian fluid without any external forces,
the governing equations can be stated as follows:

16
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

ui
0
xi
(1)
( ui ) (p) ui uj
( ui u j ) ui u j
t xj xi xj xj xi xj

Where u i are the fluctuating parts of the velocity components, u i are the mean values of the

velocity components, and ij u i u j is Reynolds stresses tensor. Reynolds stresses could not
be obtained from the mentioned governing equations, so some additional equations are needed.
Applying Boussinesq’s eddy-viscosity assumption, these terms are related to the mean velocity
gradients and the turbulent Eddy viscosity and could be stated for the incompressible flow as
follows:
 u u j 
 ij  t  i
  (2)
 x j x i 

Where t is the turbulent Eddy viscosity. The turbulent viscosity of the flow could be obtained
by means of turbulence models. The most famous turbulence models are two-equation models,
i.e., k-ε and k-ω models, because they are relatively simple, and they are accurate in numerous
types of flow fields. The RNG1-k-ε turbulence model is used in this research because it is more
responsive to rotating flows and streamlines’ curvatures [9]. The RNG-k-ε turbulence model’s
governing equations are as follows [10]:

k 1   t  k 
uj        P 
x j  x j   k  x j  k
 1   t     2
uj       C P  C 2
x i  x j     x j  1 k k k
(3)
u i u j u i
Pk   t (  )
x j x i x j
k2
t  C    t

Where k is the turbulent kinetic energy, ε is the turbulent kinetic energy dissipation rate and p k
is the turbulent kinetic energy production term. In equation (3), C  ,C 1 ,C 2 ,  k ,   are
constants.
The method which is used to model the rotational motion of the propeller is the moving
reference frame one. Thus, the rotational motion is applied to the computational domain with
the same rotational speed as the propeller, but in the opposite direction. Therefore, the velocity
vector is changed from the absolute form to the relative form according to v v r ω r ,
where r is the position vector relative to the origin and ω is the rotational velocity vector, and
the governing equations are solved in the relative form. This approach is the simplest technique
for modelling the problems with rotary boundary conditions, while the flow which is entering

1
Re-Normalization Group
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

to the rotary part is uniform relative to the axis of rotation, in a steady state way and has been
applied by many other researchers [4-8]. Hence, this technique is used in this research.
A finite-volume, RANS-based CFD code is used in this research. The pressure-velocity
coupling method is SIMPLEC2 [11], the pressure discretization method is PRESTO3, and the
discretization method used for convective terms is QUICK4 [12].
2.2 Structural analysis
The governing equations of the applied FEM code are as follows:
M u C u K u F (4)

Where M  , C  and K  are mass, damping, and stiffness matrices respectively, and u , u ,
and u are acceleration, velocity, and displacement vectors respectively. The applied external
forces used in the FEM model are represented as the F vector which includes:
- The pressure force over the propeller blades which is obtained from the CFD solution
and is computed as Fh   Pds .
A

- The centrifugal force which is computed as Fce    p re  2dv where re is the radial
v

position of the centre of the element,  p is the density of the propeller material, and 
is the rotational speed of the propeller [13].
2.3 Dimensionless hydrodynamic parameters of the propellers
In order to express the hydrodynamic parameters of a propeller in open water condition,
it is common to use the dimensionless parameters shown below:
T
Thrust Coefficient : KT = (5)
ρn D
2 4

Q
Torque Coefficient : KQ = (6)
ρn D
2 5

TVa
Efficiency : η= (7)
2πnQ

Va
Advance Coefficient : J= (8)
nD
where T is the propeller’s produced thrust in Newton (N), Q is the axial torque in Newton-meter
(N.m), Va is the advance velocity in meters per second (m/s), n is the revolutions per second
(rps), D is the diameter in meters (m), and  is the density in kilograms per cubic meter (
kg m 3 ) [14].

2
Semi-Implicit method for pressure linked equations-consistent
3
Pressure staggering option
4
Quadratic upstream interpolation for convective kinematics

18
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

2.4 Computational Domain

In order to solve the governing equations of the flow field around the propeller, a
computational domain must be considered around the propeller in such a way that the
dimensions of the domain do not affect the simulation results. Figure 1 shows that the
considered computational domain has had a cylindrical shape.

Figure 1: The computational domain

The applied boundary conditions are:

- Velocity inlet at the left end of the domain in Figure 1.
- Pressure outlet at the right end of the domain in Figure 1.
- Velocity inlet on the lateral face of the computational domain in Figure 1.
- Wall boundary condition with zero velocity relative to the domain on the propeller’s
blades.
Figure 2 shows the dimensions of the considered domain.

Figure 2: Dimensions of the computational domain (D is the propeller’s diameter.)

The dimensions of the computational domain are considered in such a way that the effects
of the propeller’s revolution on the flow field do not reach the boundaries. In the other words,
the flow becomes as uniform as possible near the domain faces. So, the domain must be large
enough especially for high blade loadings (high rates of revolution). We performed a study over
the different dimensions of the computational domain and tracked the dependency of the
produced thrust on the domain’s dimensions. The results are reported in Figures 3&4. It can be
concluded that the dependency of the produced thrust on the domain’s dimensions would be
minimized by using the domain’s dimensions reported in Figure 2.
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

Figure 3: Independency of the CFD results from the cylindrical domain’s diameter

Figure 4: Independency of the CFD results from the cylindrical domain’s height downstream the
propeller
The mesh is generated in the computational domain in such a way that the hexahedral
elements covered the entire domain, except for a small area near the propeller’s blades because
of some complexities. So, the computational domain is divided into several zones in order to
create a hexahedral dominant mesh in all zones, except for a small cylinder around the propeller
in which the tetrahedral elements are generated. Figure 5 shows the generated mesh in the
computational domain.

Figure 5: The generated mesh in the computational domain: left) the entire domain; right) around the

propeller

2.5 Validation
In order to evaluate the solution’s procedure, a standard 6-bladed B-wageningen propeller
with a pitch ratio of 1.2, an expanded area ratio of 1 and a diameter equal to 25 cm was
considered. The dependency of the hydrodynamic efficiency of the selected propeller on the
number of elements in the computational domain is studied and the results are presented in
Table 1. The magnitude of error is computed by using equation 9.
Efficiencyi -Efficiencyi-1
Efficiency errori = ×100 (9)
Efficiencyi-1

20
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

Table 1: The effects of the mesh sizing on the propeller’s performance

Element size
Element size in Blade
in the zones Number Efficiency
a small cylinder surface
No. far from the of Error
around the mesh
propeller Elements (%)
propeller (%D) (%D)
(%D) (106)
1 1.5 to 7.5 1.5 7.5 to 50 0.55 -
2 1.5 to 5.0 1.5 5.0 to 50 0.65 3
3 1.5 to 3.5 1.5 3.5 to 50 1.20 2.6
4 1.5 to 2.5 1.5 2.5 to 50 1.57 2.2
5 0.8 0.8 0.8 to 50 12.5 2.15
As it is shown in Table 1, by using a mesh sizing equal to 1.5 percent of the propeller’s
diameter on the propeller blades and in a small cylinder around the propeller a grid consisting
of about 1.6 million elements is produced. The dependency of the hydrodynamic efficiency on
the number of elements would be minimized by using such grid and the results would be
independent from the number of elements.
In order to demonstrate the validity of the obtained simulation procedure, the
hydrodynamic coefficients of the considered propeller were extracted. The operational
conditions of the propeller were adjusted so that the maximum allowable thrust over the
propeller was produced based on Keller’s cavitation criterion as follows [14]:
Ae (1.3  0.3z )T
 K (10)
A0 D 2 (P0  Pv )

Ae
where is the minimum expanded area ratio which makes the propeller able to overcome the
A0
cavitation occurrence, z is the propeller’s number of blades, D is the propeller’s diameter in
meters (m), P0 is the summation of the atmospheric and hydrostatic pressure on the propeller’s
plane in Pascal (pa), Pv is the vapor pressure in the ambient temperature in Pascal (pa), T is the
propeller’s thrust in Newton (N), and K is the constant which depends to the number of
propellers and the type of the vessel and is assumed to be 0.2 is this research, regarding the
reference [14]. Based on equation 10, the maximum allowable thrust over the considered
propeller was estimated about 1750 N in an advance coefficient equal to 0.8 in this research.
The adjustments of the operational parameters have been performed so that the thrusts which
were lower than 1750N were produced by the propeller. The maximum error happened in the
prediction of the produced thrust and was about 8% in the considered operational range [15].
The validation diagram is plotted in Figure 6.

Figure 6: Comparison of the simulation results with the experimental data [15] [15]
So, it was observed that the applied CFD tool which was used with the specified settings
has been sufficiently accurate in analysing the flow around the propellers.
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

Furthermore, the independency of the FEM results from the number of elements were
studied and it was concluded that by using a grid consisting of about 800,000 elements the FEM
results would not depend on the number of elements any more. Note that a tetrahedral structural
element which was defined by 10 nodes with three degrees of freedom at each one was used in
the FEM analyses. This element has had quadratic displacement behaviour. This structural
element is well suited for modelling the propeller’s structure. Because of the lack of
experimental data on the structural behaviour of the considered B-series propeller, the
mentioned FEM code was validated by some simple cases such as cantilever beams. It was
assumed that the results of the structural analyses of the propellers will be accurate enough if
the FEM simulations are precisely performed.
Note that in the present study the simulations are performed in order to compare the
results between some different models. So, it is important to have the computational tool which
is sophisticated enough from the ranking of the designs perspective and the absolute accuracy
of the predictions is not the main concern and it is sufficient to be fairly enough.

3. Case study
The main objective of this research is to present a multi-objective determination approach
for the number of blades on a single marine propeller. This multi-objective approach includes
several important considerations, i.e., preventing cavitation (related to the pressure distribution
around the propeller), hydrodynamic efficiency, deflection of the blades (related to the
vibrational motion of the blades), maximum Von-Mises stresses at the existing hydrodynamic
pressure (related to the structural lifetime), and the axial length of the propeller (related to the
physical space of the propeller). In order to perform this research, some propellers with similar
geometrical parameters but with different numbers of blades are considered. So, the following
assumptions are made:
1- The diameter is equal to 40 cm.
2- The expanded area ratio is equal to 1.23 and similar shapes are considered for the
leading and trailing edges between all models.
3- The pitch ratio is equal to 1.5.
4- The four-digit NACA cross sections with the location of the maximum thickness equal
to 30% of the length of the chord from the leading edge and with the location of the maximum
camber equal to 50% of the length of the chord from the leading edge are used in all radius
ratios.
5- The camber ratio is equal to 2% over the blades.
6- A similar linearly-decreasing distribution of the magnitudes of the maximum thickness
is used in the radial direction.
7- The numbers of the blades are 3, 4, 5, 6, 7, 8, 9, 11, 13 and 15 in models 1 to 10.
8- The weight of the propellers is about 4 kg.
In this way, we considered ten different models of the propellers with similar geometrical
parameters but with different numbers of blades in order to find the effects of the number of
blades on the propeller’s performance out. Table 2 provides the radial geometrical properties
of ten different models.

22
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

Table 2: Radial properties of the models

t(mm) f/c (%)
x (r/R) Cross Section Number of blades
(1 to 10) (1 to 10)
0.2 9.84 2 4-digit NACA
0.3 8.76 2 4-digit NACA
0.4 7.68 2 4-digit NACA
0.5 6.6 2 4-digit NACA
0.6 5.52 2 4-digit NACA
0.7 4.44 2 4-digit NACA
0.75 3.9 2 4-digit NACA 3, 4,
0.8 3.36 2 4-digit NACA 5, 6,
0.85 2.82 2 4-digit NACA 7, 8,
0.9 2.28 2 4-digit NACA 9, 11,
0.92 2.06 2 4-digit NACA 13, 15
0.94 1.85 2 4-digit NACA
0.96 1.63 2 4-digit NACA
0.97 1.52 2 4-digit NACA
0.98 1.42 2 4-digit NACA
0.99 1.3 2 4-digit NACA
1 1.2 2 4-digit NACA
where t is the maximum thickness of the cross section, f is the maximum camber of the section,
c is the chord length of the section, r is the radial position of the section, and R is the overall
radius of the propeller. The material which is used for the structure of the propellers is stainless
steel and Table 3 presents its physical properties.
Table 3: Physical properties of stainless steel
Property Magnitude
Modulus of elasticity (Pa) 2  1011
Poisson ratio 0.3
Density ( kg m ) 3
7850

Tensile yield strength (Pa) 2.5 108

Tensile ultimate strength (Pa) 4.6  108
Based on the mentioned assumptions, the weights of all propellers are the same, i.e., about
4 kg without the hub. One of the created geometries is shown in Figure 7.

Figure 7: Geometry of the eight-bladed propeller: left) radial cross sections of one blade in the pre-

modelling step; right) the entire propeller

N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

As shown in Figure 7 and indicated in Table 2, seventeen different radial sections are
used in the geometry of each propeller, all of which have four-digit NACA cross sections with
similar properties as mentioned earlier. Also note that, as shown in Table 2, in order to keep the
weights of the propellers the same, we had to avoid changing the magnitudes of the maximum
thickness of the airfoils in the same radius ratios between different models. As the result, we
had to increase the number of blades on some propellers that had thicker radial sections relative
to their corresponding chord lengths. In the other words, the magnitudes of t/c(r)i are increased
by increasing the number of blades. As it is shown in Figure 7, the adopted geometries for the
propellers are different from common ones in the outer radius ratios. The reason of selecting
such configuration for the propellers was to retain the hydrodynamic load of the blades in their
outer parts and make the resultant produced force of the blades further from the roots. So, the
bending effect of the thrust would be reinforced and would make a loading condition for the
blades which is harder than usual. Therefore, we could have a strict judgment about the
propellers from the structural perspective which could increase the reliability of the results in
the case of common propellers. In the other hand, since the basic shape of each un-skewed
marine propeller would be similar to the presented propeller in Figure 7 (before modifying their
leading and trailing edges), the results deemed not to have conflict with the common cases.

4. Results
In this research, a velocity of 10 m/sec is assumed as the velocity inlet boundary condition
of the domain. The rotational speeds of the propellers are changed such that a similar limited
range of the advance coefficient is covered between different models.
The extracted hydrodynamic parameters of the propellers with different numbers of
blades are compared with each other in Figures 8 to11.

Figure 8: Hydrodynamic efficiency of the propellers versus the number of blades

Figure 9: Maximum efficiency of the propellers versus the number of blades

24
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

Figure 10: Magnitude of the thrust produced by propellers in the advance coefficient corresponding to
the maximum efficiency versus the number of blades

Figure 11: Magnitude of the axial torque of the propellers in the advance coefficient corresponding to the
maximum efficiency versus the number of blades
As shown in Figure 8 the advance coefficients which are belong to the optimum points
(maximum efficiency) are to some extent unchanged by changing the number of blades on the
propellers. Figure 9 shows that the maximum efficiency of the propellers occurs in the 6-bladed
propeller and has an efficiency that is about 1% higher than the 3-bladed propeller and about
4% higher than the 15-bladed propeller. The maximum efficiency changes versus the number
of blades nonlinearly. But, the maximum efficiencies of the 4-, 5-, 6-, 7-, 8- and 9-bladed
propellers differs each other by less than 1%. The effect of number of blades on torque and
thrust magnitudes at the maximum efficiency point is shown in Figures 10 and 11 respectively.
Figure 10 shows that the thrust produced by the different propellers in the optimum advance
coefficient augments as the number of blades increases up to nine blades (20% increase relative
to the 3-bladed propeller) and after that it incurs a small decrement towards the 15-bladed
propeller (8% decrease relative to the 9-bladed propeller). Furthermore, it is observed that the
slope of the thrust increment decreases by moving towards the 9-bladed propeller. The same
behaviour is observed for the axial torque in Figure 11.
In order to investigate the effect of the number of blades on the occurrence of cavitation,
the non-dimensional pressure coefficient is considered as follows:
P P
CP (11)
12 V 2

where P is the static pressure, P is the ambient pressure far from the propeller, and V is the
velocity of the flow far from the propeller. In this way, the minimum pressure coefficients at
different radial sections of the propeller are compared (Figure 12). Figure 12 also shows the
average pressure coefficient over the surfaces of the blades. To compute the pressure
coefficient, it is assumed that the propellers are immersed in the water to a depth of 5 m.
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

Figure 12: Comparing the minimum pressure coefficient at different radial sections between 3- to 15-
bladed propellers
Figure 12 shows that increasing the number of blades of the propellers while holding the
other geometrical parameters constant decreases the magnitude of the pressure coefficient at
the blades’ roots significantly. This effect is observed in the radius ratios between 0.2 and 0.6.
The contours of the pressure distribution over the 6- and 15-bladed propellers are presented in
Figure 13 in order to make the subject clearer.

Figure 13: Contours of the pressure distribution over the: left) 6-bladed propeller; right) 15-bladed
propeller
The effect of decreasing the pressure is weakened by moving from the roots of the blades to
their tips. If the cavitation zone is placed in radius ratios between 0.6 and 0.7, increasing the
number of blades up to five would reinforce the blades against the cavitation occurrence
slightly. While more increase in the number of blades would raise the risk of cavitation in these
zones. If the cavitation zone is placed in radius ratios between 0.7 and 0.9, increasing the
number of blades up to the six do not affect the cavitation occurrence significantly. While more
increase in the number of blades would raise the risk of cavitation in these zones. If the
cavitation zone is placed in radius ratios between 0.9 and 1, increasing the number of blades up
to eight would reinforce the blades against the cavitation occurrence slightly. While more
increase in the number of blades would raise the risk of cavitation in these zones. Figure 12
shows that the average pressure coefficient over the propeller’s blades decreases significantly
as the number of blades is increased. It means that the level of the pressure decrement in the
roots of the propellers’ blades has been severe enough which has specified the direction of the
change of the averaged pressure over the blades. For the 15-bladed propeller, the magnitude of
the minimum pressure coefficient is decreased by about 300% more at the propeller’s root and

26
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

by about 25% more at the propeller’s tip than is the case for the 3-bladed propeller. The average
magnitude of the pressure coefficient is decreased by about 360% between these two models.
Since the magnitude of the pressure distribution around the airfoil (and here around the
propeller) strongly depends on the thickness ratio, it is worth to pay attention to the thickness
ratio (%) distribution over the propellers’ blades which are shown in Figure 14.

Figure 14: Comparison of the maximum thickness ratio at different radial sections between 3- to 15-
bladed propellers
Figure 14 shows that, the magnitude of the maximum thickness ratio of the blades was
increased in every radial section by increasing the number of blades. Also, the increasing rate
of the thickness ratio was constant between different radial sections of two different models.
However, the magnitudes of the thickness ratio were increased significantly near the roots
which have made the propellers more prone to the cavitation risk in these locations (see Figure
14).
In this research, a specified thrust of about 600 N is considered, and the corresponding
hydrodynamic pressure is imported from the CFD solver for the corresponding FEM model.
Thus, we are able to compare the structural behaviour of the propellers under the same loading
condition. Since, the deflection of the propellers was very small compared to their diameters,
the interaction between the fluid and solid was neglected and there was a one-time importation
of the distribution of pressure from the CFD solver to the FEM solver. Figure 15 shows the
results.

Figure 15: The maximum deflection of the propellers over their leading edges
Figure 15, shows that the magnitude of the maximum deflection which occurs over the
leading edge of the propellers is reduced by increasing the number of blades up to six, but the
magnitude of this deflection is increased for seven or more blades. Since, the leading edges of
the propellers are rather thick, it is concluded that the number of blades has a nonlinear
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

relationship with the structural deflection or the propellers’ stiffness values. The maximum
magnitude of the deflection is reduced by about 65% by increasing the number of blades from
three to six and is reduced by about 60% by decreasing the number of blades from fifteen to
six. On the other hand, Table 4 indicates that increasing the number of blades augments the
maximum magnitude of the Von-Mises stresses on the roots of the propeller’s blades. Figure
16 shows the distribution of the Von-Mises stresses along the mid-section of the 7-bladed and
15-bladed propellers. The reason of the difference between the overall shapes of the blades in
Figure 16 is the difference between the blade numbers which has leaded to the different blade
shapes. Furthermore, the difference between the legends is because of the difference between
the stiffness of the blades. As it can be seen in this figure, the maximum Von-Mises stresses are
happened on the roots of the blades which are found in Table 4.

Figure 16: Distribution of the Von-Mises stresses along the mid-section of two types of propellers: left)
7-bladed propeller; right) 15-bladed propeller
Table 4: The relationship between the maximum magnitude of the Von-Mises stresses and the number of blades
Number of Maximum Von-Mises
blades stress (MPa)
3 14.8
4 20.2
5 23.6
6 24
7 25.2
8 26.2
9 29
11 29.8
13 31.3
15 34.3
Table 4 indicates that the maximum magnitude of the Von-Mises stresses over the blades
is increased by increasing the number of blades. Note that increasing the number of blades from
3 to 15 is increased the maximum magnitude of the Von-Mises stresses by about 130%. This
parameter is directly related to the structural lifetime of the propellers [16].
The data in Table 5 clearly shows that increasing the number of blades from 3 to 15 is
resulted in 80% reduction in the axial length of the propellers. Figure 17 shows the axial length
(the occupied axial space with the propeller) of a sample propeller.

28
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

Figure 17: The side view of the 8-bladed propeller

Table 5: The relationship between the axial lengths of the propellers and the number of blades
Number of Axial length of the
blades propellers (mm)
3 200
4 150
5 120
6 100
7 85.7
8 75
9 66.6
11 54.3
13 45.9
15 39.5
The obtained results could be interpreted as follows:
- Regarding the results obtained from the simulations, it is obvious that the procedure
used to determine the number of blades on a propeller is a multi-objective function of
efficiency, cavitation, vibrational behaviour, structural lifetime, and physical space.
- In considering cavitation, it is apparent that increasing the number of blades on a
propeller changes its resistance against cavitation nonlinearly. Increasing the number
of blades of the propellers is weakened them in radius ratios between 0.2 and 0.6
against the cavitation occurrence. Changing the number of blades would affect the
cavitation occurrence in radius ratios between 0.6 and 1 slightly. The changes in these
zones are nonlinear. Generally, it can be concluded that the propensity for a propeller
to experience cavitation can be reduced by decreasing the number of blades if the
cavitation zones were located closer to the roots than the tips, while the propeller’s
weight and its expanded area ratio are kept constant. Because increasing the thickness
ratio of the blades nonlinearly changes the cavitation behaviour, it cannot be
concluded that every thicker blade is more prone to the cavitation occurrence relative
to a thinner one.
- Considering efficiency, the difference between propellers with three to nine blades is
less than 1%. However, the 15-bladed propeller has an efficiency value that is about
4% less than that of the 6-bladed propeller. The reason for this lower efficiency is
increasing the magnitude of the lift-to-drag coefficient of the radial sections by
increasing their thickness ratio up to a specified value. After that, the magnitude of
the lift-to-drag coefficient decreases with further increase in the thickness ratio [17].
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

Thus, the maximum efficiency of the propellers is increased by increasing the number
of blades from three to six, but it is decreased by any further increases in the number
of blades. Also, it is observed that the advance coefficient, which is related to the
maximum efficiency of the propellers, remains unchanged as the number of blades is
changed. The main reason is that the radial distributions of the effective angles of
attack are the same for all of the propellers. Generally, it is seen that increasing the
thickness ratio of the blades changes the efficiency of the propellers nonlinearly. So,
it cannot be concluded that every thicker blade has less efficiency relative to a thinner
one.
- Figure 10 shows that the magnitudes of the thrust that are produced by the propellers
at the optimum point are increased until the propellers have nine blades, but these
values are decreased when the number of blades exceeded nine. This phenomenon
occurs because of the relationship between the lift and drag coefficients and the
thickness ratios. As mentioned in [17], the lift coefficient increases up to a certain
value as the thickness ratio increases. If the thickness ratio continues to increase, the
lift coefficient will decrease after reaching the mentioned certain value of the lift
coefficient. On the other hand, the drag coefficient increases continuously as the
number of blades increases. Because the thrust produced by the propeller is directly
related to the lift coefficient of the radial sections and is inversely related to the drag
coefficient of these sections [18], a nonlinear nature is obviously seen in its behavior.
- Figure 11 shows that the magnitude of the axial torque applied on the propeller’s shaft
is augmented as the number of blades is increased up to nine, after which there is a
slight reduction in the torque. This event happens because the magnitude of the torque
applied on the propeller’s shaft is directly related to both lift and drag coefficients
[18]. The lift coefficient is decreased and the drag coefficient is increased by
increasing the number of blades from nine to fifteen. So, the torque is decreased, but
to a far less extent than the produced thrust.
- Considering the vibration (deflection), it is observed that the structural stiffness of the
propellers is nonlinearly changed by changing the number of blades. The greatest
stiffness occurs in the 6-bladed propeller which is about 60% greater than that of the
3- and 15-bladed propellers subjected to hydrodynamic pressure distributions that are
corresponded to the same magnitudes of thrust.
- Considering the lifetime of the structure, it is apparent that increasing the number of
blades increases the maximum Von-Mises stresses on the blades. It means that the
propeller would have a shorter lifetime.
- From the geometrical point of view, it is observed that the axial lengths of the
propellers are reduced by increasing the number of blades, which could be a very
important issue in the case of designing the ducted propellers for underwater
axisymmetric bodies.
So, it is observed that the procedure of determining the number of blades on a propeller
is a multi-objective one which must be performed according to the design criteria. According
to the obtained results in this research, the following flowchart (Figure 18) could be proposed
in order to determine the propeller’s number of blades:

30
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

Figure 18: The determination procedure of the propeller’s number of blades

At the end, it is worth to note that the Reynolds numbers which are based on the chord
lengths of the propellers’ blades in radius ratios equal to 0.7 have been over 1.25 million
between different models in this research. It shows that the simulations have been performed in
the fully turbulent flow regimes and the scale effects have been minimized as much as possible.

5. Conclusions
A multi-objective approach for determining the number of blades on a propeller is
presented in this article. Ten different models of propellers with 4-digit NACA cross sections
and the only variable being the number of blades are developed and analysed hydrodynamically
and structurally. It is observed that the procedure of determining the number of blades on a
propeller is a multi-objective function of hydrodynamic efficiency, cavitation risk, vibrational
behaviour, structural lifetime, and physical space, which must be considered with respect to the
design criteria.
N. M. Nouri, S. Mohammadi A Multi-objective approach for determining the
number of blades on a NACA marine propeller

Acknowledgment
We sincerely appreciate the assistance and guidance concerning various aspects of the research that were
provided by the staff of the Applied Hydrodynamics Laboratory at Iran University of Science and Technology.

References
[1] Abdel-Maksoud, M., Menter, F., & Wuttke, H. (1998). Viscous Flow Simulations for Conventional and
High-Skew Marine Propellers. Ship Technology Research 45, Schiffstechnik, 45(2), 64-71.
[2] Watanabe, T., Kawamura, T., Takekoshi, Y., Maeda, M., & Rhee, S. H. (2003, November). Simulation of
steady and unsteady cavitation on a marine propeller using a RANS CFD code. In Proceedings of the Fifth
International Symposium on Cavitation, Osaka, Japan.
[3] Ghasemi, H. (2009). The effect of wake flow and skew angle on the ship propeller performance. In Scientia
Iranica, 16(2), 149-158.
[4] Da-Qing, L. I. (2006). Validation of RANS predictions of open water performance of a highly skewed
propeller with experiments. Journal of Hydrodynamics, Ser. B, 18(3), 520-528.
http://dx.doi.org/10.1016/S1001-6058(06)60106-6.
[5] Hayati, A. N., Hashemi, S. M., & Shams, M. (2011). A study on the effect of the rake angle on the
performance of marine propellers. Proceedings of the Institution of Mechanical Engineers, Part C: Journal
of Mechanical Engineering Science, 0954406211418588.
[6] Ji, B., Luo, X., Wang, X., Peng, X., Wu, Y., & Xu, H. (2011). Unsteady numerical simulation of cavitating
turbulent flow around a highly skewed model marine propeller. ASME J. Fluids Eng, 133(1), 011102.
http://dx.doi.org/10.1115/1.4003355.
[7] Peng, H. H., Qiu, W., & Ni, S. (2013). Effect of turbulence models on RANS computation of propeller
vortex flow. Ocean Engineering, 72, 304-317. http://dx.doi.org/10.1016/j.oceaneng.2013.07.009.
[8] Shamsi, R., Ghassemi, H., Molyneux, D., & Liu, P. (2014). Numerical hydrodynamic evaluation of
propeller (with hub taper) and podded drive in azimuthing conditions. Ocean Engineering, 76, 121-135.
http://dx.doi.org/10.1016/j.oceaneng.2013.10.009.
[9] Orszag, S. A., Yakhot, V., Flannery, W. S., Boysan, F., Choudhury, D., Maruzewski, J., & Patel, B. (1993).
Renormalization group modeling and turbulence simulations. Near-wall turbulent flows, 1031-1046.
[10] Dong, Y., Duan, X., Feng, S., & Shao, Z. (2012). Numerical simulation of the overall flow field for
underwater vehicle with pump jet thruster. Procedia Engineering, 31, 769-774.
http://dx.doi.org/10.1016/j.proeng.2012.01.1100
[11] Van Doormaal, J. P., & Raithby, G. D. (1984). Enhancements of the SIMPLE method for predicting
incompressible fluid flows. Numerical heat transfer, 7(2), 147-163.
http://dx.doi.org/10.1080/01495728408961817.
[12] Kaya, F., & Karagoz, I. (2008). Performance analysis of numerical schemes in highly swirling turbulent
flows in cyclones. Current science, 94(10), 1273-1278.
[13] Ghassabzadeh, M., & Ghassemi, H. (2013). Determination of Hydrodynamics Characteristics of Marine
Propeller Using Hydro-elastic Analysis. Brodogradnja, 64(1), 40-45.
[14] Carlton, J. (2011). Marine propellers and propulsion. Butterworth-Heinemann.
[15] Barnitsas, M. M., Ray, D., & Kinley, P. (1981). KT, KQ and efficiency curves for the Wageningen B-series
propellers.
[16] United States. Department of Defense. (1987). Military Standardization Handbook: Metallic Materials and
Elements for Aerospace Vehicle Structures. Department of Defense.
[17] Shyy, W., Lian, Y., Tang, J., Viieru, D., & Liu, H. (2007). Aerodynamics of low Reynolds number flyers
(Vol. 22). Cambridge University Press.
[18] Benini, E. (2004). Significance of blade element theory in performance prediction of marine propellers.
Ocean Engineering, 31(8), 957-974. http://dx.doi.org/10.1016/j.oceaneng.2003.12.001.

Submitted: 16.05.2015. Nowrouz Mohammad Nouri, [email protected]

Saber Mohammadi
Accepted: 23.03.2016. School of Mechanical Engineering, Iran University of Science and
Technology, Tehran, Iran

32