International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 6
Number 1
March 2012
57
1 INTRODUCTION
For a long time ship manoeuvring characteristics
were considered to be of marginal importance. Nev-
ertheless, from the operational point of view of in-
land vessels, manoeuvring characteristics play a
crucial role. Operating in restricted waterways sub-
stantially increases the risk of accidents and disasters
as a result of manoeuvring errors. To increase level
of safety on the waterways, it is important to ensure
that vessels have appropriate manoeuvring charac-
teristics. For this purpose it is necessary to develop
research methods for determining the manoeuvring
properties of inland ships.
Currently, one of the more commonly used re-
search methods in ship hydrodynamics are numeri-
cal methods. With improvements in computing pow-
er and increasingly more accurate CFD methods it is
possible to simulate more complicated cases. This
paper presents a numerical method for prediction of
ship manoeuvrability. It is based on a mathematical
model of the ship equations of motion, described in
section 3. This model requires knowledge of the co-
efficients of the hydrodynamic forces acting on the
ship. In section 4 there is presented a method for de-
termining the hydrodynamic coefficients using CFD.
On inland waters the most restrictive rules are
described in Rhine Regulations. The severity of the
Rhine Regulations is the result of difficult conditions
on the Rhine: a strong current and high traffic inten-
sity. Vessels that fulfil these requirements are al-
lowed to operate on other waterways in Europe.
These regulations specify requirements for: a mini-
mum speed of the vessel, the ability to stop, the abil-
ity to move backwards and turning manoeuvrability.
In addition, there is specified an additional manoeu-
vre - "evasive action" , which is analogous to zig-
zag test and is a specific requirement for inland wa-
ters. The main characteristic of this evasive action
[1] is the moment of switching the rudder to the op-
posite side, which occurs when the ship reaches the
desired angular velocity.
2 MATHEMATICAL MODEL OF SHIP
MOTION
Motion simulation is based on two coordinate sys-
tems. The first one is the global, stationary coordi-
nate system. The ship position and orientation is de-
termined in the global coordinate system. The
second one is the local coordinate system. The cen-
tre of the local coordinate system is located at the
centre of gravity of the ship. Equations of motion of
the ship are written in the local coordinate system.
Figure 1 shows the two coordinate systems used in
the paper. It also shows the velocity components of
the vessel.
Application of CFD Methods for the
Assessment of Ship Manoeuvrability in Shallow
Water
T. Górnicz & J. Kulczyk
Wrocław University of Technology, Poland
ABSTRACT: Safety in water transport plays a significant role. One way to increase safety in the waterways is
to ensure that ships have proper manoeuvrability. Evaluation of manoeuvring properties performed at an early
stage of design can detect problems that later would be difficult to solve. To make such an analysis the nu-
merical methods can be used. In the paper the nu
merical method to evaluate the ship manoeuvrability on the
shallow water is presented. Additionally author shows the procedure of determining hydrodynamic coeffi-
cients on a basis of CFD calculation. The simulation results of inland ship was compared with experimental
data.
58
Figure 1. Coordinate systems and velocity components of the
vessel.
The movement of the ship is considered to be
planar motion (limited to three degrees of freedom).
This study only analysed small velocities at
which sinkage and trimming of the ship is minimal.
Therefore, these phenomena are not considered in
the research..
Interaction between the hull, rudder and propel-
lers are only simulated by appropriate coefficients.
In inland vessels the centre of gravity is far lower
than in marine ships. In addition, the centre of gravi-
ty is located near the centre of buoyancy. As a result
of this , the roll of the vessel is relatively small and
could also be neglected.
Wavelength on inland waterways is relatively
small in proportion to the length of the vessel. In the
study the influence of waves on the trajectory of the
ship was not taken into consideration.
The equations of motion (1) were written for the
centre of gravity of the ship. The left side of equa-
tions describes the ship as a rigid body. On the right
side of equations there are hydrodynamic external
forces (X, Y) and the hydrodynamic moment (N)
acting on the vessel.
zz
mu mvr X
mv mur Y
Ir N
−=
+=
=
(1)
where:
mvessel mass; ulongitudinal velocity; v
transversal velocity; ryaw rate; ˙- dot, the deriva-
tive of a variable over time; I
zz
moment of inertia
of the vessel; X,Y,N hydrodynamic force and mo-
ment acting on the vessel according to the axes of
local coordinate system. Detailed information about
model can be found in [2].
External hydrodynamic forces and the hydrody-
namic moment acting on the ship were written like
in the MMG model [3], as a sum of the compo-
nents.
HPR
HR
HPR
XX X X
YY Y
NN N N
= ++
= +
= ++
(2)
where:
X
H
, Y
H
, N
H
hydrodynamic forces and moment act-
ing on the bare hull; X
P
, N
P
- force and moment in-
duced by the operation of propellers, X
R
, Y
R
, N
R
forces and moment induced by flow around the rud-
der.
2.1 Hull
To determine the hydrodynamic forces induced by
flow around the hull the following mathematical
model was used [3]:
2
24
32
23
32
23
() ( )
()
H x T vv vr y
rr vvvv
H y v r x vvv vvr
vrr rrr
H zz v r vvv vvr
vrr rrr
X m u R u X v X m vr
Xr X v
Y mv Yv Y mu r Y v Y v r
Y vr Y r
N jrNvNrNv Nvr
N vr N r
−− + + + +
++
= ++− + + +
++
=+++ + +
++
=
(3)
where:
m
x
, m
y
added mass coefficients, in x and y direc-
tion; j
zz
added inertia coefficient; R
T
(u) hull re-
sistance; X
vv
, X
vr
,..., Y
v
, Y
r
,... - coefficient of hydro-
dynamic forces acting on the hull; N
v
, N
r
, ...
coefficients of hydrodynamic moment acting on the
hull.
2.2 Rudders
Hydrodynamic forces induced by rudder laying can
be calculated on the basis of equations (4). The
model was taken from[4].
(1 ) sin
(1 ) cos
( ) cos
R RN
R HN
R R HH N
X tF
Y aF
N x ax F
δ
δ
δ
=−−
=−+
=−+
(4)
where:
t
R
- coefficient of additional drag; a
H
- ratio of addi-
tional lateral force; x
R
x-coordinate of application
point of F
N
; x
H
- x-coordinate of application point of
additional lateral force; δ rudder angle; F
N
normal
hydrodynamic force acting on the rudder.
The value of normal force F
N
was determined
on the basis of model (5).
59
( )
( )
( )
2
2
2
2
0
0
0,
1 (1 · ( ))
2 (2
()
1
1
0.6
1
cos
1 (1 )
·
·
N RN R
RR
R
P
R
P
P
RR
P
RR
RR
F ACU
U w Cg s
K ss
gs K
s
D
h
w
K
w
s wU
nP
w
ww
w
xr
ρ
η
η
β
α δ γβ
ββ
=
=−+
−−
=
=
=
=−−
=
=
′′
=
(5)
where:
ρ water density; A
R
rudder area; C
N
normal
force coefficient; U
R
- effective rudder inflow ve-
locity; C coefficient, dependent on ruder angle
sense (C≈1.0); Dpropeller diameter; h
R
height of
rudder; w
P
- wake fraction at propeller location; w
R
-
wake fraction at rudder location; w
R0
- effective
wake fraction at rudder location, in straight ahead
ship motion; U total velocity of vessel; β drift
angle; n – rotational speed of propeller; P – propeller
pitch; x’
R
non-dimensional x-coordinate of appli-
cation point of F
N
; r’ – non-dimensional yaw rate.
2.3 Propellers
This paper uses a mathematical model of hydrody-
namic forces generated by two propellers.
12
12
(1 )( )
(1 )( )
P
P
X tT T
N tT Td
=−+
=−−
(6)
where:
t - thrust deduction factor; T
1
,T
2
thrust generated
by propellers; d- distance from the axis of propeller
to symmetry plane of the vessel, in y direction.
Propeller thrust was determined on the basis of
the relation (6).
0
2
0
24
1 22
1
cos
·exp( 4.0 )
()
(
·
)
P
P
PP
P
T
T
P
P
w
U
n
T nDK J
K J a aJ
J
r
J
w
a
D
w
x
β
β
ββ
′′
=
=++
=
=
=
(7)
where:
T propeller thrust; J - advance coefficient; a
0
, a
1
,
a
2
– K
T
polynomial coefficients; w
P0
- effective
wake fraction in straight ahead ship motion; x’
P
non-dimensional x-coordinate of propeller..
Values of thrust coefficients used in the calcula-
tions were derived from experimental research.
3 DETERMINING OF HYDRODYNAMIC
COEFFICIENTS
3.1 The coefficients of the hydrodynamic forces
acting on the hull
The model of the hydrodynamic forces described in
section 2.1 requires knowledge of the values of hy-
drodynamic coefficients: X
vv
, X
vr
,..., Y
v
, Y
r
,..., N
v
,
N
r
, .... The literature, including [5], describes the
empirical formulas derived for marine ships. Due to
the complicated nature of these forces, the results
can be insufficiently accurate. For this reason, pref-
erence is given to other, more accurate method to
determine the coefficients of hydrodynamic forces.
One of these methods is the CFD calculation. In ad-
dition to the providing accurate results, numerical
calculations enable the model to take into account
specific operational conditions of the inland ship, for
example, that the impact of shallow water on the hy-
drodynamic forces acting on the hull.
o determine the coefficients of hydrodynamic in-
teractions it is necessary to have a database con-
taining the values of the hydrodynamic forces and
the corresponding to them values of velocity (u,v,r)
and the acceleration (
,,uvr

) of the ship. One way to
obtain such a database is to simulate with help of
CFD software the series of tests (manoeuvres): yaw ,
yaw with drift, sway test. Figure (2) shows the tra-
jectory of a ship during these manoeuvres.
The values of the added mass (m
x
,m
y
,) coeffi-
cients , added inertia (j
zz
) coefficient and the ship
resistance (R
T
(u)) can be obtained on the basis of
empirical methods or CFD calculations.
During the yaw manoeuvre the ship transverse
velocity v and acceleration
v
are zero. The equa-
tions of the hydrodynamic forces acting on the hull
can be simplified to the following form:
2
3
3
()
H x T rr
H x r rrr
H zz r rrr
X mu R X r
Y murYrYr
N jr N N r
u
r
=−− +
= ++
=−+ +
(7)
In the results of CFD simulation of the yaw ma-
noeuvre the relation between hydrodynamic forces
and moment (X
H
, Y
H
, N
H
) and velocity u, r, and ac-
celeration
u
is obtained. The least squares method
can be used on the results from yaw simulation to
approximate the following coefficients: X
rr
, Y
r
, Y
rrr
,
N
r
, N
rrr
.
60
During the sway manoeuvre the speed r and the
acceleration
,ru

are zero. The equations of hydro-
dynamic forces acting on the hull can be simplified
to the following form:
24
3
3
()
H T vv vvvv
H y v vvv
H v vvv
X R Xv X v
Y mvYvYv
N Nv
u
Nv
=−++
= ++
= +
(8)
When simulating the sway manoeuvre using CFD
the data showing the relation between hydrodynam-
ic forces acting on the hull (X
H
, Y
H
, N
H
) and veloci-
ties u and v is obtained. The least squares method is
used on the results from sway simulation to approx-
imate the following coefficients: X
vv
, X
vvvv
, Y
v
, Y
vvv
,
N
v
, N
vvv
.
The CFD simulation of the yaw with drift ma-
noeuvre provides the calculations for the rest of hy-
drodynamic coefficients: X
vr
, Y
vvr
, Y
vrr
, N
vvr
, N
vrr
.
More information about determining hydrody-
namic coefficients can be found[6]
Figure 2. The manoeuvres for determining the hydrodynamic
coefficients: A) yaw manoeuvre B) yaw with drift, C) sway
manoeuvre.
3.2 Rudder characteristic.
A mathematical model of hydrodynamic forces act-
ing on the rudders, described in section 2.2, requires
knowledge of the normal force coefficient (C
N
) . The
characteristics of isolated rudders have been de-
scribed in many sources, for example [7]. Working
conditions of rudders installed under the hull of the
inland vessels may be significantly different than
these from a single rudder. This is due to the pres-
ence of wake and propeller streams as well as the
impact of the limited depth of the waterway. In order
to determine the correct characteristics of the rudder,
more accurate methods should be used.
In studies to determine the characteristics of the
rudders CFD methods were used. The calculations
were carried out in two ways. Firstly, the approach
shown in Figure 3, was based on calculations of a
rudder located in the propeller stream. The geometry
of propeller was replaced by the disk with a pres-
sure jump. The value of the pressure jump was
equivalent to propeller thrust. Restriction of flow
around a rudder at the top and bottom edge was
simulated by two flat plates. The symbols c and d on
the scheme denotes distance between rudder and
plates. The main advantage of this approach is the
low complexity of the discrete model, which signifi-
cantly accelerates the calculations.
The second method used in the study was to build
a full geometric model of the entire hull with propel-
lers and rudders. This solution required the usage of
discrete models with a much larger number of ele-
ments. CFD calculations of the entire hull give more
accurate results but require large computing power.
Figure 4 shows an example of a discrete model of
the stern of the ship with the rudders and simplified
models of propellers.
Additional information on the rudder force nu-
merical calculations can be found in [8].
Figure 3. Scheme of model for rudder force calculations.
Figure 4. Example of discrete model of the stern of the ship
with propellers and rudders.
4 CFD METHODS
In the studies the Ansys Fluent commercial CFD
software was used. It is based on a finite volume
method. To solve the three-dimensional turbulent
flow the RANS method was used. The turbulence
model k-ε-Realizable was used. The boundary layer
was calculated using the Enhanced Wall Treatment
model. The result of research presented in [9]
shows that this model works best for flows with
large pressure gradients and a separation phenome-
non.
61
The calculations mainly used a structured mesh
with hexagonal elements. Unstructured (tetragonal)
grids were only used in the calculation of the hull
with rudders and propellers, due to the complicated
geometries involved.
To simulate the yaw, yaw with drift, sway ma-
noeuvre the moving mesh technology was tech-
nique was utilised. Parameters of mesh motion were
defined in an additional batch program (UDF) to the
system Ansys Fluent (UDF). The program is written
in C language.
5 RESULTS
Numerical calculations were performed for a model
of inland transport vessel (ship A). The scale model
was λ = 21.81.
For the same scale a physical model was created
and tests in the towing tank were performed. All
tests were performed in the shallow water. In this
paper the results for h/T=1.89 are presented. Table 1
contains the main parameters of the ship A.
Table 1. The main parameters of the ship A
______________________________________________
Parameter unit ship value
______________________________________________
LPP m 85.50 3.920
B m 11.45 0.525
T m 2.65 0.122
CB - 0.0853
LCB m 43.71 2.004
Number of propellers - 2
Number of rudders - 2
______________________________________________
Parameters of the propeller (model)
_____________
Parameter unit value
______________________________________________
D m 0.08
P/D - 1.1102
A
E
/A
0
- 0.7474
Z - 4
______________________________________________
Parameters of the rudder (model)
_____________
Parameter unit value
h
R
m 0.092
A
R
m
2
0.0103
Profile - N
ACA0012
______________________________________________
5.1 Direct Comparison
CFD calculations are characterized by a number of
restrictions and simplifying assumptions. In the nu-
merical calculations of yaw, yaw with drift and
sway test, which are necessary for determining the
value of hydrodynamic coefficients, a problem with
calculation stability occurred . After disabling two-
phase flow (air-water) the problem disappeared.
Therefore, the calculation of coefficients of hydro-
dynamic forces acting on the hull X
H
, Y
H
, and N
H
does not include the impact of the free surface. This
problem didn't appear in the calculation of the re-
sistance of the ship. In the subsequent research only
situations with low ship speed (Fr = 0.12) were ana-
lysed. When the ship speed is relatively low the in-
fluence of free surface on the hydrodynamic force is
negligible.
The figures 5,6,7 shows a comparison of results
obtained from CFD calculations and the experi-
mental test performed in the towing tank. The charts
show the course of the hydrodynamic forces X
H
, Y
H
and hydrodynamic moment N
H
during the yaw and
sway test. During the tests in towing tank the PMM
mechanism was used. The worst comparison be-
tween CFD and experiment is for X
H
. This force is
relatively small and it is very susceptible for differ-
ent factors and disturbances. The comparison for
hydrodynamic moment N
H
is much better. The hy-
drodynamic moment depends mainly on huge differ-
ence of pressure on both sides of a hull. And the
pressure field was accurately predicted by CFD cal-
culations.
Figure 5. Comparison of the results of the experiment and CFD
calculations for the characteristic of X
H
forces obtained during
yaw test.
Figure 6. Comparison of the results of the experiment and CFD
calculations for calculations for
Y
H
force obtained during yaw
test.
Figure 7. Comparison of the results of the experiment and CFD
calculations for the N
H
moment obtained during the sway test.
5.2 Indirect Comparison
In the towing tank the model tests of ship A were
performed. The following tests were performed:
yaw, yaw with drift and sway. On the basis of the
62
experimental results the coefficients of hydrody-
namic forces were calculated. Similarly, on the basis
of the CFD calculations the results of a second set
of hydrodynamic forces coefficients was deter-
mined. For both sets of coefficients the simulation of
standard manoeuvres was performed.
The manoeuvre results were obtained from the
author's program to simulate the motion of the ship.
The following figures show the comparison of the
results obtained from experimental tests and com-
puter simulation. The figures 8,9,10 illustrate the
characteristic charts for the turning circle manoeu-
vre, evasive action and spiral test.
Figure 8. Comparison of trajectory for turning circle manoeu-
vre, rudder δ = 35 °.
Figure 9. The comparison of characteristic chart of the eva-
sive action.
Figure 10. The comparison of characteristic chart of the spi-
ral test.
6 SUMMARY
The paper presents a numerical method for evaluat-
ing manoeuvring characteristics. Research on a
model of the inland vessel showed good agreement
between numerical calculations and experimental re-
sults.
Due to problems with the stability of the CFD
code only the mono-fluid calculations were per-
formed. Further studies will be carried out to calcu-
late the free surface effects, and taking into account
phenomena such as sinkage and trim.
BIBLIOGRAPHY
[1] COMMISSION DIRECTIVE 2008/126/EC European Par-
liament and of the Council laying down technical require-
ments for inland waterway vessels, Official Journal of the
European Union L 32/1.
[2] Właściwości manewrowe statku śródlądowego na wodzie
ograniczonej, Tabaczek T., Górnicz T., Kulczyk J., Zawi-
ślak M, Raporty Instytutu Konstrukcji i Eksploatacji Ma-
szyn, SPR, Politechnika Wrocławska, 2010.
[3] The Specialist Committee on Esso Osaka Final Report and
Recommendations, Conference Proceedings of the 23rd
ITTC, Volume II, Venice, 2002.
[4] Kijima, K., Tanaka, S., Furukawa, Y., Hori, T.On predic-
tion Method of Ship Maneuvering Characteristics, MAR-
SIM'93, St. John's, Newfoundland, Canada, 1993.
[5] The Manoeuvring Committee Final Report and Recom-
mendations, Conference Proceedings of the 23rd ITTC,
Volume II, Venice, 2002
[6] Marine Rudders and Control Surfaces, Anthony F. Molland
and Stephen R. Turnock, Elsevier, 2007.
[7] Metoda oceny właściwości manewrowych statku śródlądo-
wego. Górnicz T, Raporty Instytutu Konstrukcji i Eksploat-
acji Maszyn, PRE, Politechnika Wrocławska, 2010.
[8] Badanie wpływu głębokości wody na siły hydrodynamicz-
ne działające na ster statku śródlądowego. Górnicz Tomasz,
Kulczyk Jan, Tabaczek Tomasz, XXV. Plavebné dni, 2008,
[9] Computional Fluid Dynamics A practical Approach, Jiyuan
Tu, Guan Heng Yeoch and Chaoqun Liu, Elsevier, 2008