The 10th TSME International Conference on Mechanical Engineering

10th – 13rd December 2019

Pattaya, Thailand

AME002

Development on Design of Optimum Contra-Rotating

Propellers under Slipstream Contraction Condition

Abstract. From previous study, contra-rotating propellers (CRPs) are designed based
on lifting line theory where radius of free vortex lines are adjusted in order to
correspond to the slipstream contraction condition. The calculus of variations method
is then applied on equations that derived from the lifting line theory. The effect of
slipstream contraction is included in order to find the optimum CRPs where the
required thrust with minimum power can be achieved under the condition that CRPs
are operated in axial uniform inflow. Results from calculation are then converted into
a 3-dimensional CRPs configuration and is simulated by CFD method as preliminary
verification. Comparison study on results obtained from calculation and CFD shows
some differences in thrust and torque on each propeller. To find the cause of these
differences, further investigation is done and presented in this study. It is found that
the inflow velocity that appears on each propeller is differ from the assumption used in
calculation in which inflow velocity appearing on propeller equals to initial inflow
velocity that entering through CRPs. The modification is done by introducing
correctors for fore and aft propellers. These correctors are multiplied with the initial
inflow velocity to represent new values of inflow velocities appearing on propellers.
Initial inflow velocities on propellers are then replaced by these new values. Through
the same calculation process, new CRPs configurations are obtained and simulated by
CFD method. Correctors are adjusted in each calculation to find proper values that
conform both results from calculation and CFD. Through different trying out of
corrector values, it can be concluded that the inflow velocity appearing on propeller
not only changes differently in each condition but also between fore and aft propellers.
It’s found that the inflow velocity is directly proportional to advance ratio and
inversely proportional to thrust coefficient. This is because at low advance ratio and
high thrust coefficient, the propeller has high rotational speed and strong circulation
which tend to create stronger flow field turbulence. This causes the reduction of
inflow velocity on propeller. Moreover, the inflow velocity on aft propeller is always
less than that on fore propeller, as aft propeller has to encounter with wake flow
behind the fore propeller while fore propeller approaches to more uniform flow.

1. Introduction
From previous study [1], the contra-rotating propellers (CRPs) are studied based on Laskos [2] by
taking the slipstream contraction condition into consideration. The CRPs are designed by using the
lifting line theory where the radius of free vortex line is adjusted to correspond to results obtained
from Hoshino [3] in order to include the effect of slipstream contraction in the calculation process.
The CRPs are specified to operate in axial uniform inflow and the diameter of aft propeller is chosen
to be equal the diameter of the contracted slipstream to make use of the rotational loss behind the fore
propeller. The equations obtained from lifting line theory are used to state the auxiliary function for
finding the optimum CRPs that provide the required thrust with minimum power. This auxiliary
function of CRPs is then applied with the calculus of variations method and solved by using the
method stated in Coney [4] to find the optimum circulations distributing on propeller blade. The cross
section type of propeller blade is chosen as NACA2412 and the chord length distribution is defined by
following Oossanen’s study [5] to give the expanded area ratio of propeller approximately equal to
0.5. These optimum circulations with chosen blade cross section profile and chord length are used to
find the section lift coefficient and the orientation of blade cross section by following the experimental
data from Abbot and Doenhoff [6]. The elements of oriented blade cross section are combined into 3-
dimensional CRPs and are simulated by CFD program called ANSYS Fluent in order to check the
calculation results. The detail of simulation can be found in our previous work [1]. The comparisons
between calculation and CFD results for varied sets of optimum CRPs at varied advance ratios show
some differences in thrusts and torques on propellers. Considering thrusts obtained from different
advance ratio CRPs, results from CFD show that thrusts on both fore and aft propellers is considerably
lower for higher advance ratio CRPs but results from calculation show that only thrust on fore
propeller is lower moderately while thrust on aft propeller is higher for higher advance ratio CRPs to
maintain the total required thrust. Moreover, torques on propellers in low advance ratio CRPs are
much higher than torques obtained from calculation. Therefore there is a need to further investigate the
cause of these different results between calculation and CFD.

2. Background Theory and Results of Previous Study

The CRPs are designed based on the lifting line theory. The lifting line is divided into elements where
the circulation of bound vortex inside elements represents elements of blade cross section and ability
in generating load. Inside the hub, there is also hub vortex that has the same magnitude of circulation
as bound vortex circulation but with opposite direction at imaginary position based on Kerwin and
Leopold’s study [7]. Moreover, there is also a free vortex in downstream which caused by the
variation of bound vortex. This free vortex line’s direction which is aligned with the direction of
resultant flow velocity that leave from lifting line along with its radius at any distance in downstream
is adjusted according to results obtained from Hoshino [3]. The radius of aft propeller, which placed
behind the fore propeller at distance d, is adjusted to be equal to the radius of fore free vortex line in
order to make use of the rotational loss behind the fore propeller. These circulations from the propeller
itself and from other propeller could cause the induced velocity at each blade element according to the
Biot-Savart law.

The lift force (L) acting on each blade element is the cross product of the bound circulation and
resultant flow velocity multiplied with fluid density at that blade cross section. The viscous force (D)
due to fluid viscosity also occurs which its direction aligns with the resultant flow velocity (V* ). The
force and velocity diagram are shown in Figure 1.

From Figure 1, the flow velocity is divided into tangential and axial direction with subscript t and a
respectively as well as all forces acting on the blade element. By summing all tangential and axial
forces on every blade elements (n) with thickness r and the total blade number (Z), thrust (T) and
torque (Q) acting on propeller could be found from equation (1) and (2). Furthermore, the free vortex
lines due to hub vortex are gathering in downstream behind the hub. This results in the low pressure at
that area and causes the hub drag (Dh ) which could be found from Wang [8]. The hub drag (Dh ) is
shown in equation (3) where rh is the hub radius and (1) is the circulation at the hub.
n
1
T=Z [r i + ut,s i + ut,m (i)]  (i)  c i CD V * i 2 [Va + ua,s i + ua,m (i)] r (1)
2
i=1
 n
1
Q=Z [Va + ua,s i + ua,m (i)]  (i) c i CD V * i 2 [r i + ut,s i + ut,m (i)] r(i)r (2)
2
i=1
1 rh
Dh =  ρ( log + 3)(Zf f (1)  Zr r (1))2 (3)
16π 0.1Rf

Lf

u'a,f
V∗ ua,f
D
Va
ωf rf
ut,f

u′t,f L

V∗
u′a,r
ua,r D
Va
ωr rr
ut,r
u′t,r

Figure1: Force and Velocity Diagram on Blade Cross Section

The thrust and torque equations are used to form the auxiliary function. This auxiliary function is then
applied along with the calculus of variation method to find the optimum circulation distributions that
resulted in the required thrust with minimum power for the CRPs. To simplify the optimum
calculation process, the drag force is initially neglected and later included to obtain the final result.
The circulations obtained is used to indicate the section lift coefficient and the orientation of blade
element under defined cross section profile and chord length according to Abbot and Doenhoff [6].

Blade elements are combined together for the complete CRPs and is simulated by using CFD program.
The flow domain and its boundary conditions are shown in Figure 2. Meshes inside flow domain are
generated by hybrid method and inserted with the inflation layers in area close to CRPs’ surface. The
CRPs are simulated by sliding mesh method with SST k- model.

Converged results obtained at each time step are averaged and compared with the calculation results.
The condition of CRPs are shown in Table 1 where 3 sets of optimum CRPs are found at 3 different
advance ratios while keeping other conditions the same. Results from 3 sets of optimum CRP from
both calculation and CFD are shown and compared in Figure 3 and Figure 4.
 Outer surface : No shear stress

Contact surface Outlet surface

ωfront ωrear

Solid surface : No slip condition

Inlet surface

Figure 2: Flow Domain and Its Boundary Conditions for Simulating in CFD program

Table 1: Conditions Used for Obtaining Optimum CRPs

Propeller Model A B C
Advance ratio : J= 2πVa ⁄Dp,f 0.8 1.0 1.2

Rotational velocity :  ωf ωr [rad/s] 8.0809 6.4647 5.3873

Fore propeller’s diameter : Dp,f [m] 2

Hub radius ratio : xh,f rh ⁄Rf 0.2

Blade number : Zf Zr [blades] 4

Distance between fore and aft propellers ratio : d/Rf 0.45

Thrust coefficient : CT = Treq ⁄(0.5πR2f V2a 0.512

Torque ratio : q = Qr ⁄Qf 0.8

Fluid density : ρ [kg/m ] 998.2

Inflow velocity : Va [m/s] 2.05778

 2000 A
1900 B
C
1800
Thrust [N]

1700
1600 fore, Calculation
aft, Calculation
1500 fore, CFD
1400 aft, CFD

1300
0.8 1 1.2
Advance Ratio, J

Figure 3: Thrust Comparison between Calculation and CFD

900 C

B
800
A
torque [Nm]

700

600 fore, Calculation

aft, Calculation
500 fore, CFD
aft, CFD
400
0.8 1 1.2
Advance Ratio, J
Figure 4: Torque Comparison between Calculation and CFD

Comparing results obtained from Calculation and CFD shows that the differences exist as shown in
Figure 3 and Figure 4. On the fore propeller, results from CFD and Calculation indicate the same trend
for thrust which decreases at higher advance ratio. However thrust obtained from CFD is higher than
that obtained from Calculation for low advanced ratio. This is in contrast with higher advanced ratio
where thrusts obtained from CFD is lower than those obtained from Calculation. On the aft propeller,
the trend is difference as thrusts obtained from CFD decrease at higher advanced ratio while thrusts
obtained from Calculation increase at higher advanced ratio.
For torque on propeller, CFD provides higher torques on both fore and aft propellers than those
obtained from Calculation especially at low advance ratio.

3. Analysis Procedure
Considering assumptions used in Calculation and conditions for CFD simulation, there are various
possible reasons why results obtained from CFD and Calculation are difference such as assumptions
about constant pitch of free vortex lines in calculation, the difference between inlet boundary in CFD
simulation and lifting line position in Calculation, the wake flow that happened behind the fore
propeller, the error in converting Calculation results into 3-dimensional CRPs and etc. However, after
thoroughly investigated, one of possible reasons that should be first investigated is the difference in
inflow velocity that appearing on each propeller. In Calculation, the inflow velocity is defined at the
lifting line where the middle of propeller blade is placed while in CFD, it is defined at the inlet
boundary which differs from the position of lifting line or the middle of propeller blade. This is
because the domain of flow field has to cover all parts of CRPs and there are some constraints in
generating meshes. Therefore the inlet boundary used in CFD has to be shifted from the propeller
blade to upstream of the flow as shown in Figure 5. There is a distance between the inlet boundary and
lifting line in which the inflow velocity could be induced and changed by the rotating propeller before
it reaches the propeller blade. Moreover some turbulence in flow field that caused by rotating propeller
also affects the inflow velocity appearing on the propeller as well. This could happen especially for the
aft propeller that has to encounter the turbulent flow that leave from the fore propeller.

Inlet boundary
Inflow velocity, V

Lifting line

Calculation CFD

Figure 5: Difference in Defining the Position of Inflow Velocity between Calculation and Fluent

Two correctors are used for investigating the inflow velocity as the cause of difference between
Calculation and CFD results. One is for fore propeller (w ) and the other is for aft propeller (w ).
Each of them is multiplied with inflow velocity and placed as modified inflow velocity appearing on
propeller as shown in Figure 6. This modified inflow velocities are used in Calculation process.
Equations obtained are solved to find the new optimum circulation distribution and hence the new
CRPs are formed and simulated by CFD.
 u'a,f
u ,

V , =w V
ωr
u,
u′ ,

u′ ,
u,
V , = w V
ωr
u,
u′ ,

Figure 6: Velocity Diagram of Propeller Blade Cross Section with Modified Inflow Velocities

4. Analysis Results
Results of new CRPs from Calculation and CFD are compared. Tests are done under the same
conditions as described in Table 1 with two additional different thrust coefficient 0.412 and 0.612 in
each advance ratio. The proper values of correctors are found through trial and error process and
shown in Table 2. Figure 7 and Figure 8 show results obtained from both Calculation and CFD that
include these correctors in the analysis.

Table 2: Proper Values of Correctors for Each Set of CRPs

Thrust coefficient Advance Propeller
w w
(C ) ratio Model
0.8 A1 1.003 0.99
0.412 1.0 B1 1.012 0.995
1.2 C1 1.024 1.01
0.8 A2 0.994 0.97
0.512 1.0 B2 1.004 0.98
1.2 C2 1.019 0.997
0.8 A3 0.981 0.951
0.612 1.0 B3 0.995 0.961
1.2 C3 1.015 0.983
 1500 A1 B1
C1
1450
1400
Thrust [N]

1350
1300
1250
1200
1150
1100
0.8 1
Advance Ratio, J 1.2
(a)

1900 A2 B2
C2
1800
fore, Calculation
1700 aft, Calculation
Thrust [N]

fore, CFD
1600 aft, CFD
1500
1400
1300
0.8 1 1.2
Advance Ratio, J
(b)

2350
A3
2250 B3 C3
2150
Thrust [N]

2050
1950
1850
1750
1650
0.8 1
Advance Ratio, J 1.2
(c)

Figure 7: Thrust Comparison between Calculation and CFD with Correctors included for CRPs
at (a) = 0.412, (b) = 0.512 and (c) = 0.612.
 850
C1
750
B1
Torque [Nm]

650 A1
550

450

350
0.8 1 1.2
Advance Ratio, J
(a)

950 C2

850 B2 fore, Calculation

Torque [Nm]

aft, Calculation
750 A2 fore, CFD
aft, CFD
650

550

450
0.8 1
Advance Ratio, J 1.2
(b)

1100 C3

1000 B3
Torque [Nm]

900
A3
800

700

600

500
0.8 1
Advance Ratio, J 1.2
(c)

Figure 8: Torque Comparison between Calculation and CFD with Correctors included for CRPs
at (a) = 0.412, (b) = 0.512 and (c) = 0.612.

From Table 2, modified inflow velocities that appearing on propellers are found to be different from
the initial inflow velocity about 0.3 - 2.4% for fore propeller and 0.3 - 4.9% for aft propeller under
given conditions. By investigating at different conditions, it could be concluded that the value of
inflow velocity on propeller is directly proportional to advance ratio and reversely proportional to
thrust coefficient. Values of correctors for aft propeller are always less than those for fore propeller.
This is because the initial inflow velocity at the inlet boundary is highly induced by the propeller
before it reaches the fore propeller. However at some conditions such as at low advance ratio and high
thrust coefficient, the strong turbulence occurred by high rotational speed and strong circulation of
propeller results in reduction of inflow velocity. By applying correctors for inflow velocity, results
from both Calculation and CFD are closer with each other. Still some discrepancies in values
especially for torques at low advance ratio CRPs could be seen in which the maximum difference is
12.6% on fore propeller of CRPs at thrust coefficient equal to 0.412 and advanced ratio equal to 0.8.
The reason is the drag coefficient used for calculation is approximately lower than that used in CFD
and hence the calculated torque is underestimated. Torque difference between Calculation and CFD
could be seen especially at low thrust coefficient and low advance ratio CRPs since the drag force is
aligned close to tangential direction and also with high rotational speed CRPs results in higher
magnitude of torque on propeller.

5. Conclusion
The lifting line theory with the effect of slipstream contraction is used for designing the CRPs.
Equations obtained that represent the relationship of distributing circulations, velocities and forces on
propeller are then solved by using the calculus of variations method to find the optimum CRPs that
provide the required thrust with minimum power under axial uniform flow. The optimum CRPs
obtained from calculation are simulated by using CFD program for preliminary verification. Result
comparison from both Calculation and CFD for different CRPs which suited for different advance
ratios shows some differences. Results from CFD gives higher thrust on aft propeller at low advance
ratio CRPs but gives lower thrust on both fore and aft propeller at high advance ratio CRPs than those
obtained from Calculation. After investigating on assumptions, it is found that the inflow velocity on
propeller is defined at different position between Calculation and CFD. In Calculation, the inflow
velocity is defined at the lifting line or the middle of propeller blade while in CFD it is defined at the
inlet boundary which is next to lifting line in the upstream direction. Since the flow domain has to
cover all parts of CRPs and some constraints in generating meshes which make it shifted from the
lifting line position. Moreover, the wake flow could occur behind the fore propeller due to propeller
rotation. This could make inflow velocities on propellers in CFD analysis differ from those that
assumed in Calculation. To verify that the difference of inflow velocity appearing on propeller causes
the difference of results between Calculation and CFD, two correctors are assumed and multiplied
with inflow velocities appearing on fore and aft propellers respectively. The proper correctors found
from trial and error process show that the inflow velocity on propeller has different value from the
initial value and it is directly proportional to advance ratio and reversely proportional to thrust
coefficient. The reason is that at lower advance ratio and higher thrust coefficient, the propeller has
higher rotational speed and higher circulation strength. This causes the flow more turbulent and
decrease the magnitude of inflow velocity. Moreover, the inflow velocities on fore and aft propellers
also have different values by which the inflow velocity on aft propeller is lower than that on fore
propeller because the aft propeller has to face the wake flow behind the propeller while fore propeller
encounters the uniform flow. Using the corrector for inflow velocity on propeller could improve
results from Calculation and CFD. However there are some differences in value of torque especially
on CRPs with low thrust coefficient and low advance ratio. This has to be further investigated to find
causes in which one might be the error in assuming drag coefficient. Furthermore, the inflow velocity
appearing on propeller could be changed not only the magnitude as assumed in this study but also the
direction. This could affect the thrust and torque on CRPs as well.

Nomenclature
c Chord length of blade cross section
CD Drag coefficient of blade cross section
CT Thrust coefficient
d Distance between fore and aft propeller
D Drag force
DP Propeller diameter
J Advance ratio
L Lift force
n Number of lifting line’s element
q torque ratio of aft propeller’s torque to fore propeller torque
Q Torque
R Propeller radius
rc Radius at control point
rh   Hub radius
rv Radius at vortex point
r Length of lifting line’s element
T Thrust
Treq Total required thrust
u Self-induced velocity
u' Mutually-induced velocity
Va Axial inflow velocity
V*  Resultant flow velocity on blade cross section
Z Blade number
 Flow density
 Bound vortex circulation
 Rotational velocity

Subscript
f Fore propellers
r Aft propellers
t Tangential direction
a Axial direction

References
[1]    Nattanun S and Chirdpun V 2017 Design of optimum contra-rotating propellers under slipstream
contraction condition The 8th TSME-ICoME (Bangkok)
[2] Laskos D 2010 Design and cavitating performance of contra-rotating propellers M.S. thesis
Dept. Mech. Eng. MIT (Cambridge MA)
[3] Hoshino T 1989 Hydrodynamic analysis of propellers in steady flow using a surface panel
method J.S.N.A. (Japan) 165-166
[4] Coney W B 1992 Optimum circulation distributions for a class of marine propulsors Journal of
ship research No.3 36 pp 210-222
[5] Oossanen P 1974 Calculation of performance and cavitation characteristics of propellers
including effects of non-uniform flow and viscosity NSMB Publication 457
[6] Abbott I H and Doenhoff A E 1959 Theory of wing section including a summary of airfoil data
General Publishing Company Ltd.(Toronto Ontario)
[7] Kerwin J E and Leopold R 1964 A design theory for subcavitating propellers The society of naval
architects and marine engineers (New York)
[8] Wang M H 1985 Hub effects in propeller design and analysis Ph.D. thesis Dept. Mech. Eng.
MIT (Cambridge MA)