Numerical analysis of the container vessel's self-propulsion at different

rudder deflection angles

Michał Wawrzusiszyna, Radosław Kołodzieja, b, Sebastian Bielickia
a Ship Design & Research Centre, CTO S.A, 80-392 Gdańsk, Szczecińska 65, Poland
b Gdansk University of Technology, 80-233 Gdańsk, Narutowicza 11/12, Poland

Michal.Wawrzusiszyn@cto.gda.pl, Radoslaw.Kolodziej@cto.gda.pl, Sebastian.Bielicki@cto.gda.pl

1. Introduction

Nowadays, CFD becomes one of the most commonly used research method in ship hydrodynamics,
limited to the analyses of hull resistance in calm water. With continuously improving computing power
and increasingly more accurate numerical methods it is possible to simulate more complex cases. State
of the art CFD tools also enable development of new ways of assessing ship maneuvering performance.
This paper presents an attempt on using CFD for evaluation of the coefficients used in the formulation
of rudder forces applied in the ship manoeuvring model. These coefficient are normally evaluated in
captive tests of the hull with working propeller and rudder deflected at different angles; the paper
presents the results of CFD simulation of this kind of experiment. The test case used in the analyses is
the well known the KRISO Container Ship (KCS). The computations were carried out at model scale
1:50, for which the reference model test results are available. Comparison of CFD and experimental
results is presented.

2. Mathematical model

There are many approaches to decomposition of forces acting on the ship during manoeuvring described
in literature. According to MMG standard method [1] they can be presented as sum of following
components:
X = XH + XP + XR (1) where:
Y = YH + YR (2) X, Y, N - Surge force, lateral force, yaw moment
N = NH + NR (3) XH, YH, NH - Surge force, lateral force, yaw moment acting on the hull
XR, YR, NR - Surge force, lateral force, yaw moment acting on rudder
XP - Surge force generated by the propeller

Effective rudder forces and moment are expressed as:

XR = –(1 − tR)FNsin 𝛿 (4)

YR = –(1 + aH)FNcos 𝛿 (5)
NR = –(xR + aHxH)FNcos 𝛿 (6)
Fig. 1: Coordinate
system
where:
FN - Rudder normal force
tR - Steering resistance deduction factor
aH - Rudder force increase factor
xH - Longitudinal coordinate of point of application
xR - Longitudinal coordinate of rudder position (~0.5LPP)

Mathematical model of maneuvering ship includes certain parameters that are unknown at initial design
stage (aH, xH and tR) thus they can be evaluated only by the means of model tests or numerical analyses.
The evaluation consists in analysis of forces acting on hull and rudder in vessel moving straight ahead
with rudder deflected at certain angles and constant speed, when forces YH and NH on right hand sides
of equations (2) and (3) are equal to zero. Forces XH+XP are assumed to be constant for considered
propeller rate of revolution and vessel speed (constant propeller advance ratio).
3. CFD Simulation

The computations were carried out at model scale 1:50 using the Reynolds Averaged Navier-Stokes
Equations (RANSE) method. The CFD method applied is based on previous publications for NuTTS
conferences [2][3]. Meshing and flow simulations were conducted with use of Star CCM+ 2019.1.
Analyses were done with the use of the Estimating Hull Performance (EHP) module.
As turbulence model the Realizable K-Epsilon (two-layer all-y+ wall treatment) was used. The mean
value of y+ on the hull was about 3.2 and below 1.0 in rudder/propeller region. Main particulars of the
hull and propeller are presented below.

Propeller Main hull particulars Unit Value

Value
CP572 Model scale [-] 1:50
Diameter [m] 0.160 Length b.p. m 4.600
Pitch ratio [-] 1.240 Length of waterline m 4.649
Breadth m 0.644
Hub ratio [-] 0.333
Draught m 0.216
Expanded Displacement volume m3 0.416
0.640
area ratio [-]
Fig. 3: Propeller Surface wetted area m2 3.781
Direction of
CP572 Left Block coefficient [-] 0.651
rotation
Midship section
[-] 0.985
coefficient

Configuration for propulsion analyses is presented in Fig. 4. The

flow was computed in the rectangular domain of the following
dimensions: [6L; 5L; 2.5L], where L is the hull length. Analyses
were divided into three parts:
- Mesh sensitivity study;
- Bare hull computations;
- Appended hull computations with working propeller.

Mesh sensitivity study with bare hull (half domain) was done. The
size of mesh was analysed against influence on resistance value.
Taking into account almost constant value of resistance for meshes
3, 4 and 5, mesh No. 3 was used for further computations as the
optimal compromise providing the mesh-independend solution
(see the table below). Fig. 4: Propulsion arrangement

Mesh size Size of base Relative

Resistance
No. [Num. of y+ element [m] resistance
[N]
cells] [%]
1 1 840 000 9.65 1.000 11.226 94.45
2 2 410 000 5.64 0.095 11.310 94.57
3 3 330 000 3.27 0.085 11.996 100.30
4 4 510 000 3.25 0.750 11.992 100.27
5 7 960 000 2.82 0.600 11.960 100.00

During analyses with propeller the hull was fixed to reduce computation time. The free surface was
modelled. Values of hull trim (0.078⁰) and sinkage (-0.0046m) for propulsion analyses were determined
from resistance computations.
Total mesh size for analyses with working propeller was about
8 000 000 cells (see Fig. 5). Seven rudder angels were analysed:
0⁰; ±10⁰; ±20⁰; ±35⁰.
For resistance and propulsion computations a constant inlet
velocity was set to 1.31m/s. Water density was set to
998.540kg/m3 and dynamic viscosity was set to 1.0122×10-3 Pa-s.

The time step value was changed during computations:

ts=0.030s – for development of a free surface and resistance
stabilisation
ts=0.001s – when propeller was rotated by 2.94⁰ per one time step.
Fig. 5: Mesh presentation

Designed pitch ratio set on propeller geometry was P/D0.7=1.24 while in experiment P/D0.7=0.80. The
constant propeller revolutions n=8.165 [RPS] were set according to propeller thrust value TP=13.0[N]
from model test results, where rudder was not deflected. Simulation of propeller rotation in the domain
was solved by using sliding mesh.
Global forces in i and j direction on rudder and hull were monitored. Moment acting on the entire ship
model (rudder, propeller and hull) was measured relative to z-axis located in hull LCB (x=2.23m).

4. Results of CFD analyses

Detailed results of computations are presented in below table and in Figs. 6-7.
URmean
Rudder Propeller Hull Force
Force Force Force Force Moment In front
angle thrust resistance XFN
YFN [N] FN [N] X [N] Y [N] N [Nm] of rudder
δ [deg] TP [N] RH [N] [N]
[m/s]
-35.0 14.54 24.60 -9.62 13.24 -16.36 -10.40 15.85 -34.35 1.289
-20.0 13.37 19.22 -3.95 13.45 -13.99 -6.16 16.15 -35.15 1.336
-10.0 13.05 16.81 -1.47 7.27 -7.41 -4.08 9.20 -19.64 1.351
0.0 13.17 16.12 -0.59 -0.19 0.19 -3.29 0.09 -0.26 1.362
10.0 13.13 16.75 -1.17 -7.59 7.68 -3.96 -9.17 19.48 1.357
20.0 13.55 19.08 -3.50 -13.26 13.66 -5.85 -15.75 33.70 1.344
35.0 14.48 24.06 -8.77 -18.37 20.08 -9.88 -21.68 46.58 1.306

25

20

15
Tp, RH [N]

10

5 Fig. 6: Thrust and hull resistance

Tp RH
comparison for different rudder
0 angles
-40 -30 -20 -10 0 10 20 30 40
Rudder deflection [deg]
Fig. 7: Comparison of flow and pressure distribution for different rudder angles
(left column port side, middle – starboard, right – top view)
During the resistance computations of bare hull the water velocity at propeller disc and rudder position
was measured. The mean values were then used to calculate the wake fraction coefficients of propeller
wP0 and rudder wR0 as presented below.

XR: URmean = (1-wR)U

XP: UPmean = (1-wP)U

Fig. 8: Wake fraction coefficients

UPmean URmean wP0 wR0

U
in propeller disc in front of propeller wake rudder wake fraction
[m/s]
[m/s] rudder [m/s] fraction coeff. [m/s] coeff. [m/s]
0.893 0.954 1.309 0.271 0.319

In order to determine the hydrodynamic coefficients of rudder (tR, aH and xH), forces acting on a hull
XH+XP, YH and moment NH can be expressed as a function of FNsinδ and FNcosδ [4]. It turns out that
their relationship is almost linear for given propeller load, therefore derivatives can be approximated as
a constant value (Fig. 8).
The results for -35º rudder deflection were substantially different from other results. It seems it is the
consequence of the flow separation. Therefore the data for this particular rudder angle, were not taken
into consideration during rudder coefficients calculation.
tR aH xH xR
0.426 0.262 -0.346 -0.500

Fig. 8: Rudder coefficients approximation (red cross – points excluded from analyses)
5. Conclusions

 Considerable difference between water flow in rudder section for portside and starboard can be
noticed.
 The coefficients resulting from CFD vary substantially from the experimental values. The difference
may arise from neglecting in the experiment the force component generated on the rudder horn.
Despite the simulations of turning test based on CFD and experimental coefficients show that the
sensitivity of the model to the values of these coefficients is rather small the influence of rudder
horn forces will be analysed to enhance the approach.

1 5
CFD EXP
4
EXP CFD
3

x/Lpp 2

0 0
tr ah xh 0 1 2 3 4 5
y/Lpp

Fig. 9: Comparison of CFD and experimental results (left) and the results of turning simulation based
on CFD and experimental input.

6. References
1. Yasukawa Y.Y.H., "Introduction of MMG standard method for ship maneuvering", Technol, 2015
2. Wawrzusiszyn M., Kraskowski M., Król P., Bugalski T., (2018). “Experimental and numerical
hydrodynamic analysis of propulsion factors on R/V Nawigator XXI with pre-swirl stator device”,
21st Numerical Towing Tank Symposium
3. Wawrzusiszyn M., Bugalski T., Hoffmann P., (2015). “Numerical simulations of ship hull-propeller
interaction phenomena”, 18th Numerical Towing Tank Symposium
4. Bielicki S.: Opracowanie matematycznego manewrowania statku na fali. Część1: identyfikacja
charakterystyk kadłuba i opracowanie modelu matematycznego ruchu statku na wodzie spokojnej,
Technical Report No. RH-2018/T-104, Gdańsk, December 2018