POLISH MARITIME RESEARCH 3 (103) 2019 Vol. 26; pp.

65-77
10.2478/pomr-2019-0047

NUMERICAL AND EXPERIMENTAL STUDY ON SEAKEEPING

PERFORMANCE OF A HIGH-SPEED TRIMARAN WITH T-FOIL IN
HEAD WAVES

Ang Li a
Yunbo Li b
a
Harbin Engineering University, China
b
Shanghai Maritime University, China

ABSTRACT

The longitudinal motion characteristics of a slender trimaran equipped with and without a T-foil near the bow are
investigated by experimental and numerical methods. Computational fluid dynamics ( CFD) method is used in this
study. The seakeeping characteristics such as heave, pitch and vertical acceleration in head regular waves are analyzed
in various wave conditions. Numerical simulations have been validated by comparisons with experimental tests. The
influence of large wave amplitudes and size of T-foil on the longitudinal motion of trimaran are analyzed. The present
systematic study demonstrates that the numerical results are in a reasonable agreement with the experimental data.
The research implied that the longitudinal motion response values are greatly reduced with the use of T-foil.

Keywords: trimaran; T-foil; URANS; longitudinal motion; model test

INTRODUCTION outrigger layout on the resistance and seakeeping performance

were analyzed. In [5] the effects of the stagger of the side
Improving a ship’s seakeeping performance and enhancing hulls on the motions in heave and pitch of a trimaran hull
its navigational performance under stormy wave conditions were investigated. Despite the model experiments were still
has always been crucial research topics in the marine industry. valuable the use of CFD method was getting in common. CFD
More and more research works on high-performance vehicles approaches widely used by researchers can be classified into
and stabilization techniques have been made in recent years. two main categories, namely, based on potential theory and
Trimarans are currently of interest for many new high- RANS method. For the prediction of seakeeping performance
speed ship projects due to their high level of hydrodynamic some methods could be found in the literature : in [6] the
efficiency. high-speed slender body potential flow theory (also called
Compared with monohull and catamaran, the trimaran two and a half dimensional potential flow theory, 2D+t or
has excellent navigation and high-speed performance [1, 2]. 2.5D), was first developed. The method effectively solved
Large amount of efforts have been focused on developing the shortcomings of the STF method in the prediction of
experimental and numerical methods to predict the multihull ship seakeeping [7]. In [8] the vertical motions of
seakeeping performance of trimarans. In the work [3] two a trimaran with different Froude numbers were calculated by
different designs for target trimaran were studied. The first using 2.5D theory. The authors demonstrated that the results
one had a round bilge hull and asymmetric side hulls and of the 2.5D method for prediction of the trimaran wave-
the other one had a Deep-V hull and symmetric side hulls. induced motions are in agreement with the experimental
Comparison of the resistance and seakeeping qualities of the results. In [9] a series of model tests in regular head waves
two different trimaran designs tested in model scale were were conducted for one trimaran with different outrigger
presented in their paper. In [4] model tests were conducted on layouts at the Froude numbers of 0.471 and 0.628. The authors
multiple trimarans, and the influence of trim, hull form and pointed out that the numerical model of 2.5D is based on the
linear potential flow theory and hence cannot capture the

POLISH MARITIME RESEARCH, No 3/2019 65

Unauthentifiziert | Heruntergeladen 19.10.19 13:07 UTC

influence of the violent green water on the heave motions, are simulated. The obtained numerical results are validated
and that this could cause an error in the numerical results. with the use of experimental data. During the study, the
In the traditional potential theory viscosity and non- grid convergence tests are also carried out for the URANS
linearity effects are not taken into account, thus it may approach. The longitudinal motion of the ship is examined
cause errors. As computational facilities have become more in various wave conditions. Unsteady wave patterns and the
powerful and accessible, CFD tools are now increasingly used time history results of the longitudinal ship motion responses
to predict ship motions. In [10] a numerical method for the in waves are simulated by using the CFD tool. To investigate
prediction of motion performance of a trimaran with a long the influence of T-foil on hydrodynamic performance of
bulb and a transom stern was developed and its accuracy was high-speed trimaran in regular head waves, two cases are
assessed by the comparison with experiments. The unsteady considered: bare trimaran and trimaran with T-foil. The
Reynolds-averaged Navier-Stokes equations are solved with influence of large wave amplitudes and size of T-foil on the
the density function technique for the free surface treatment. longitudinal motion of trimaran are analyzed.
The research showed that the heave and pitch motions were
predicted with reasonable accuracy. In [11] ship motions and
added resistance of a high-speed trimaran in regular head DESCRIPTION OF THE TRIMARAN AND
waves were predicted by using CFD tools. In [12] CFD tools T-FOIL
were used to investigate the optimization of a trimaran yacht
equipped with axebow. The study was focused on the analysis A numerical simulation and small-scale model test of the
of resistance and seakeeping qualities which could provide trimaran were carried out. The body plan of the main hull
the best performance to the yacht. and outriggers are shown in Fig. 1. The main characteristics
When the trimaran navigates under high speed, the of the model are given in Tab. 1.
longitudinal motion responses may be large. The vertical The dimensions of the T-foil installed on the trimaran
acceleration will cause severe seasickness of passengers and model near the bow are indicated in Tab. 2. The T-foil is
crew [13]. In order to reduce longitudinal motion responses composed of a vertical foil and an airfoil with the attack angle
of fast ferries, different kinds of motion control devices have of ±5°. It is assumed that the counter-clockwise rotation angle
been developed to increase comfort of passengers. In recent of the T-foil is positive. The geometrical image of the trimaran
years, T-foils has been widely installed on fast ferries due equipped with the T-foil and stern flap is shown in Fig. 2.
to its small size and obvious effects. It is fixed under the
bow, providing the restoring force and moment, and it can
reduce the pitch and heave motion of the fast ferries [14]. The
INCAT corporation engaged in the study of anti-vertical
motion appendages, developed fixed and retractable T-foil and
cooperated with Maritime Dynamics Inc. (MDI) to design
the Ride Control System (RCS) which can adjust the restoring
force and moment generated by T-foil when the ship moves in
waves [15]. In [14-16] many numerical simulations and model Fig. 1 Lines drawing of the trimaran
tests of the high speed monohull craft with T-foil and flaps
were conducted. The authors pointed out that the vertical
Tab. 1 Main characteristics of the trimaran model
acceleration and MSI (Motion Sickness Index) of the fast ferry
were reduced by 65 and 35% , respectively , with the actuators
moving under control. In [17] an experimental study (model Main feature Value
tests) was conducted on a trimaran with controlled T-foil,
Middle Hull Side Hull
three different ship models including a bare hull, a hull with
Length overall (m) 3.343
non-controlled T-foil and a hull with an active controlled
T-foil. In the study, the efficacy of the non-controlled and Breadth (m) 0.641
controlled T-foil were reported and discussed. There are Depth (m) 0.249
limited studies based on URANS method on the reduction Length at waterline (m) 3.0 1.071
of vertical motion of the trimaran by means of T-foil. Beam at waterline (m) 0.24 0.051
The present study primarily deals with the seakeeping
Draft (m) 0.122 0.043
performance of a trimaran with T-foil near the bow in regular
head waves, investigated by using experimental and numerical Displacement (kg) 45.99
methods. Longitudinal radius of gyration (m) 0.807
In the present study, the numerical simulations and Ratio of main hull displacement and
14.333
experimental tests for the prediction of the ship motions outriggers displacement
of trimaran with and without T-foil in regular head waves
are carried out. The numerical method is based on URANS
method. Both free heave and pitch motions in head waves

66 POLISH MARITIME RESEARCH, No 3/2019

Unauthentifiziert | Heruntergeladen 19.10.19 13:07 UTC

 Tab. 2 Dimensions of the T-foil` The forward speed U was set at 3.408 m/s, and the
corresponding Froude number was equal to 0.628. For
Principal particulars Value
Airfoil shape NACA0012 the entire seakeeping tests, the wave amplitude changed
Span (m) 0.144 with wave length. The wavelength λ varied between 1.5 m
Chord (m) 0.096
and 6.75 m, thus the ratio varied between 0.50 and 2.25.
Aspect ratio 1.5
Rotating angle of the flap () The numerical simulations were conducted in the same test
Width of the vertical foil (m) 0.01 conditions as for the model tests. The CFD simulations and
Length of the vertical foil (m) 0.060
model tests were performed for eleven different conditions
Height of the vertical foil (m) 0.024
Longitudinal position of installed foil (distance between listed in Tab. 3, ( identified by their case numbers). The wave
vertical centre line of foil and 10th theoretical station of 1.320
encounter frequency [ Hz] is calculated from the formula
the trimaran body) (m)
for head seas (where - the gravitational
acceleration, and - the ship forward speed [ m/s]).
Tab. 3 Simulation and model test conditions

Encounter frequency
Condition number

Wave amplitude

Wave frequency
Wave length
(λ/LPP)

fe (Hz)
f (Hz)
ζ (m)
Fig. 2 Geometrical image of the trimaran with T-foil

ζ*k
EXPERIMENTAL SETUP

Model tests were conducted at the Harbin Engineering C0 (calm

University towing tank whose dimensions are: 108 m, 7 m and water)
— — — — —
3.5 m in length, width and depth, respectively. There is a wave
C1 0.50 0.015 1.020 3.292 0.063
maker at one end of the towing tank, and a wave dissipation
bank at the other end of the towing tank. The maximum C2 0.75 0.0225 0.833 2.348 0.063
speed of the towing carriage over the tank is 6.5 m/s. The C3 0.88 0.025 0.769 2.060 0.059
waves generated by the wave generator are measured with C4 1.0 0.03 0.721 1.857 0.063
the use of a wave probe. The four - motion measurement C5 1.25 0.03 0.645 1.554 0.050
device (heave, pitch, roll and surge) is used to measure the
C6 1.38 0.03 0.614 1.437 0.046
ship model motions.
C7 1.60 0.04 0.570 1.280 0.052
The bare trimaran hull model and the model fitted with a
T-foil and stern flap are used for the model tests. The model C8 1.75 0.04 0.545 1.194 0.048
free to trim and sink was run first in calm water to obtain C9 2.0 0.05 0.510 1.078 0.052
the basic resistance. The models were then run in a seaway to C10 2.25 0.05 0.481 0.986 0.047
allow the ship wave-generated motions to be observed. The
heave, pitch and vertical acceleration at 19th hull station near
the bow in different wave conditions were measured. Fig. 3 NUMERICAL SIMULATION METHOD
shows the experimental setup for the model test.
This section provides details of the numerical simulation
approaches used in this study and discussion of the numerical
methods applied to the current CFD model.

GOVERNING EQUATIONS

An URANS approach was applied to calculate the added

resistance and ship motions in regular waves by using the
STAR-CCM+ commercial CFD software [18]. The governing
equations are the RANS equations and continuity equation
Fig. 3 Experimental setup for seakeeping tests for mean velocity of the unsteady, three-dimensional

POLISH MARITIME RESEARCH, No 3/2019 67

Unauthentifiziert | Heruntergeladen 19.10.19 13:07 UTC

incompressible flow. The averaged continuity and momentum fractions. The grid is simply refined to enable the variations in
equations are given by Eq. (1) and Eq. (2) in tensor form in volume fraction to be more accurately captured. In this work,
the Cartesian coordinate system. a second-order convection scheme was used throughout all
the simulations to accurately capture interface between the
(1) phases. The flow equations were solved in a segregated manner
by using a predictor-corrector approach. Convection and
diffusion terms in the RANS equations were discretized by a
(2) second-order upwind scheme and a central difference scheme.
The semi-implicit method for pressure-lined equations
(SIMPLE) algorithm [19] was used to resolve the pressure-
where u is the relative averaged velocity vector of flow velocity coupling. The turbulent model selected in this study
was a SST k-w turbulence model which has the ability to
between the fluid and the control volume, ui 'u j ' is the simulate the complex flow problems with flow separation and
Reynolds stresses and p is the mean pressure. For Newtonian strong adverse pressure gradients. In order to simulate the
fluid under three-dimensional incompressible flow, the mean seakeeping behaviour of the trimaran, the Dynamic Fluid
Body Interaction (DFBI) model was used to allow the model
shear stress tensor τ ij can be expressed as follows : to move free in heave and pitch modes. The DFBI model
enabled the RANS solver to calculate the exciting force and
(3) moments acting on the hull due to wave forces, and to solve
the governing equations of rigid body motion to reposition
the grid body [20]. The overset mesh allowed all motions to
where µ is dynamic viscosity. be simulated without the risk of badly deforming cells, thus
The equation for the translation of the centre of mass of enabling the simulation of large amplitude motions.
the ship body is given as:
C OM PU TAT IONA L D OM A I N , B OU N DA RY
(4) CONDITIONS AND MESHES

where m represents the mass of the body, f is the resultant A general view of the computational domain with the
force acting on the body and v is the velocity of the centre trimaran is shown in Fig. 4. The boundary conditions are
of mass. An angular momentum equation of the body is given in Tab. 4.
formulated in the body local coordinate system with the
origin in the centre of the body:

(5)

where M is the tensor of moments of inertia, is the angular

velocity of rigid body, and n is the resultant moment acting
on the body. The resulting force and moment acting on the
ship are obtained from fluid pressure and shear force acting
on each boundary face of the body. The translations of the
ship are estimated according to the computed velocity and
pressure fields in the flow domain.

PHYSICAL PHENOMENA MODELLING

Fig.4 General view of computational domain and boundaries
The “Finite Volume Method” (FVM) was applied to the
Tab. 4 Boundary conditions
spatial discretization. The “Volume of Fluid “ (VOF) method
was used to model the free surface, either with a flat wave
Symmetry
Boundary

Trimaran

in the simulation of calm-water resistance or first-order

Bottom
Outlet
Inlet

Side

Top

regular wave in the calculation of ship motion in waves.

The VOF model is based on the assumption that the same
basic governing equations as those used for a single-phase
Boundary conditions

problem can be solved for all the fluid phases present within
Symmetry Plane

the domain, as it is assumed that they will have the same

Pressure Outlet
Velocity Inlet

Velocity Inlet

Velocity Inlet

Velocity Inlet

No-Slip Wall

velocity, pressure and temperature. This means that the

equations are solved for an equivalent fluid whose properties
represent the different phases and their respective volume

68 POLISH MARITIME RESEARCH, No 3/2019

Unauthentifiziert | Heruntergeladen 19.10.19 13:07 UTC

 Two different coordinate systems were applied to predict boundaries were selected as velocity-inlet. The symmetry
ship performance. One is the earth-fixed coordinate system plane forms a symmetry condition. The trimaran hull was
and another one is the moving coordinate system located at regarded as a rigid body and set as no-slope wall. In order
the centre of gravity (CG) of the trimaran. The origin of the to prevent the undesirable effect of the reflected waves from
inertial frames was located on the base line of the ship at its aft the outlet boundary, artificial wave damping capability of the
perpendicular. The origin of the ship-fixed reference frames software package [22] was applied to the outlet boundary in
was situated at their centre of gravity (CG). First , the flow background region with a damping length equal to1.5 ship
field was solved, and the exciting force and moments acting length.
on the hull were calculated in the earth-fixed coordinate There are three methods , namely : that of rigid motion,
system. Then, the forces and moments were converted to deforming mesh motion and overset mesh, which can be
the moving coordinate system. The motions of the trimaran chosen. The rigid motion and deforming mesh motion
were described by translations and rotations with respect to methods have distinct disadvantages compared to the overset
the centre of gravity in the earth-fixed coordinate system. mesh method when simulating bodies with large amplitude
The equations of motions were solved to calculate the ship motions. The rigid motion approach causes difficulties for
velocities and then they were converted back to the earth- free surface refinement, especially in pitch, and deforming
fixed coordinate system. These information sets were then meshes may lead to cell quality problems. The authors of [23]
used to find the new location of the ship and grid system. showed that the overset grid method was more accurate in
The overset grid system was re-positioned after each time the prediction of the motion response of an SSB catamaran
step. The authors of [21] recommend that for simulations in under regular wave conditions when heading in the wave. The
the presence of incident waves, the inlet boundary should overset region moves with the hull over a static background
be located (1÷2) LBP away from the hull, whereas the outlet region. In fact, using the overset mesh saves computational
should be positioned( 3÷5) LBP downstream to avoid any wave costs and allows to generate a sufficiently refined mesh
reflection from the boundary walls. The trimaran symmetry configuration around the free surface and the body. Since
about its centre plane of the middle hull allowed to discretize the overset mesh was used, two different regions , named
only one-half of the computational domain. Therefore only background region and overset region , were established to
a half of the trimaran and control volume were considered simulate ship responses in calm water and waves. An interface
in all the calculations; thus, the symmetry plane formed the was produced between the background region and overset
centreline domain face to reduce computational time. The region to transfer fluid information. Mesh generation was
mid-plane of the ship is located at y=0 and ship draught (T) performed by using the automatic meshing facility in STAR-
is z=0. The dimensions of the boundaries are shown in Fig. 5. CCM+ software. A trimmed cell mesher was employed to
which gives front and side views of the computation domain. produce a high-quality grid for complex mesh generating
problems. The ensuing mesh was formed primarily of
unstructured hexahedral cells with trimmed cells adjacent
to the surface. Prism layers were applied around the trimaran
to resolve the turbulent boundary layer at the no-slip surface
and mesh control volumes were produced to refine the meshes
near the free surface and the hull surface together with the
surfaces of anti-pitching appendages. Fig. 6 presents a general
view of the computational mesh showing the background and
overset domains. As shown in Fig. 6, near the free surface
and around the hull, local mesh refinements were applied to
resolve the wave pattern and the incident wave field. Fig. 7
provides a closer look at the mesh around the ship hull and
Fig. 5. The dimensions of the computational domain anti-pitching appendages.

The inlet boundary was located one model length ahead

of the model’s forward perpendicular; the outlet boundary
- three model lengths behind the model’s aft perpendicular;
the bottom - two model lengths below the calm water surface;
the top - one model length above the calm water surface
and one model length between symmetry boundary and
side boundary. The velocity-inlet boundary condition was
set in the positive x - direction, where first - order incident
regular waves were generated. The velocities of the incident
wave were directly set to the inlet boundary to generate
the required regular wave. The negative x- direction was
modelled as a pressure-out. The top, bottom and side Fig.6 A general view of the meshes for the whole domain

POLISH MARITIME RESEARCH, No 3/2019 69

Unauthentifiziert | Heruntergeladen 19.10.19 13:07 UTC

 Where Z a, Φ , A and ζa represent the first Fourier Series
(FS) harmonic amplitudes of heave, pitch, acceleration
and incident wave amplitude time histories, respectively.
Corresponding Fourier transform equations may be referred
to [25]. In this paper, the longitudinal motions of heave and
pitch were evaluated at the ship’s centre of gravity (CG) and
the vertical acceleration was evaluated at the 19th hull station
near the bow.

VERTIFICATION
Fig.7 (a) Meshes around the bow (b) Meshes around the T-foil
A verification study was carried out to assess the numerical
uncertainty USN. Grid convergence test was performed at the
VERIFICATION AND VALIDATION FOR wave-ship length ratio λ/L = 1.38, which corresponds to the
CFD SIMULATION near-resonant physical conditions for heave motion [9], since
large motions and accelerations tend to cause the highest
In this section, following the method presented in [24], numerical errors. The verification parameters included the
verification and validation (V&V) was performed to quantify wave amplitude ζa, the heave and pitch transfer functions
errors and uncertainties in CFD simulations. TF3, TF5 and vertical acceleration at 19th hull station (A19)
It is necessary to mention that the transfer functions of near the bow, respectively.
heave, pitch and acceleration motions were calculated as The convergence studies were performed for three cases
follows: by using systematically refined grids and time steps. Often in
implicit unsteady simulations, the time step is determined by
(6) the flow properties. In [26] there is recommended the use of at
least 100 time steps per period for motion responses. In this
paper, a small time step (1/256 of the encounter period) was
(7) used as a base time step. The coarse and fine mesh systems
are derived by reducing and increasing cell numbers per
wavelength and cell height on free surface , respectively, by
(8) using a factor of √2 [27] as well as cell numbers on and around
the ship hull, which is affected by the mesh refinement on free
Tab. 6 Results of grid convergence study.

Tab. 7 Results of time step convergence study.

70 POLISH MARITIME RESEARCH, No 3/2019

Unauthentifiziert | Heruntergeladen 19.10.19 13:07 UTC

surface. The simulation time step is set to be proportional to solutions corresponding to fine SK1, base SK2, and coarse SK3
the grid size [28, 29, 30]. values for the kth input parameter. To assess the convergence
The grid convergence study was carried out with the condition, Rk acc. Eq. (9) is used as the convergence ratio:
small time step in three calculations in which
the grid size was systematically coarsened by a uniform (9)
refinement ratio of in each direction, while the
time step convergence study was carried out with a time step where is the difference between the
refinement ratio of on the fine grid. The base and results of base mesh and fine mesh, and is
coarse mesh were generated by changing the base size of grid the difference between the results of coarse mesh and base
only in the overset region, which means that the background mesh. Subscript k refers to the kth input parameter. SC is the
mesh configuration was not altered [20]. This enabled the corrected simulation value calculated as where S
incident head waves to be modelled efficiently through the is the simulation results, is the corrected value. is the
computational domain. Otherwise, the incident head waves corrected uncertainty.
would not have been captured well with a coarser grid, thus Tab. 8 Results of grid convergence study.
leading to wrong results. Although the mesh changed, the
size of prism layers around the hull surface remained the
Parameter (%D) (%D) E(%D)
same. The value of yplus on the hull surface were between
8.069 2.5 8.447 0.844
30÷100 as shown in Fig. 8 and Fig. 9. The mesh cell numbers
for each mesh configuration are listed in Tab. 5. 11.014 2.5 11.294 -4.589
13.137 2.5 13.373 5.463
12.154 2.5 12.409 1.319

As shown in Tab. 6 and 7, monotonic convergence was

achieved for , and . The uncertainty estimated
for the wave amplitude, motion transfer functions and vertical
acceleration at 19th hull station were reasonable small. The
grid uncertainty (9.108%) and the time step uncertainty
Fig.8 Yplus distribution on the hull surface
(10.496%) are relatively large for the pitch transfer
function these values are reduced to 2.592% and 3.024%
, respectively, when the corrected uncertainties and
are estimated. The time step uncertainties are generally
larger than those of the grid uncertainties , which indicates
that the numerical results are relatively more sensitive to the
time step resolution.

VALIDATION

To determine modelling errors, the numerical results were

compared to the experimental data. The experimental data
uncertainty was assumed to be of a rather low value UD =2.5%.
The validation uncertainty Uv and the comparison error E
were calculated and summarized in Tab. 8. As indicated in
Tab. 9, the wave amplitude, the heave, pitch transfer function
Fig.9 Yplus distribution on the hull surface as well as the vertical acceleration can be considered validated
since the comparison error E is smaller than the validation
Tab. 5 The final cell numbers for each mesh configuration in the grid uncertainty Uv. The uncertainty levels are estimated to be
convergence study. 8.447%, 11.294%, 13.373% and 12.409%, respectively.

Mesh con-
Grid name Grid number (Million)
figuration RESULTS AND DISCUSSION
Background Overset Total
G1 Fine 1.092 5.040 6.132 After performing the validation, the main findings of this
paper are summarized and discussed as follows.
G2 Base 1.092 2.229 3.321
G3 Coarse 1.092 1.055 2.147

The results of grid and time step convergence studies are

listed in Tab. 6 and 7. Let’s consider the situation of three

POLISH MARITIME RESEARCH, No 3/2019 71

Unauthentifiziert | Heruntergeladen 19.10.19 13:07 UTC

WAVE MEASUREMENT angles different resistance characteristics were revealed.
Tab. 10 shows that the trimaran with T-foil attack angle of -5o
For the purpose of monitoring the waves generated at the demonstrates preferable calm water resistance characteristics.
inlet, a wave probe located 0.5 LWL [25] in front of the hull was Therefore, the T-foil attack angle of -5o was used as a basic
set to record the wave elevation. Fig. 10 shows the recorded condition in subsequent research on the influence of ζa*k
time history of the wave elevation at . The mean and size of T-foil on the trimaran motions.
first- order harmonic wave amplitudes of the last ten encounter Tab. 9 Calm water resistance of the trimaran without anti-pitching appendages
periods was 0.0284 m. The errors between the simulated wave
amplitude and the intended wave amplitude were equal to
5.33%. The slight difference was considered acceptable for the EFD (N) CFD (N)
current cell size and time step. The measured wave amplitudes
were used to calculate the motion transfer functions in the Resistance (N) 36.84 36.50
CFD simulation.
Error (%) --- 0.92

Tab. 10 Calm water resistance of the trimaran with different T-foil attack
angles

T-foil T-foil T-foil T-foil

(-5): (-5): (+5): (+5):
EFD (N) CFD (N) EFD (N) CFD (N)

Resistance (N) 39.804 39.116 40.242 39.186

Error
--- 2.74 --- 2.62
(%)

ΔR (%) 8.05 7.17 9.23 6.85

Fig.10 Time history of wave elevation at the wave probe, λ/L=1.38

SHIP MOTION RESPONSE OF BARE TRIMARAN

Following the model tests and numerical simulations of the

CALM WATER RESISTANCE OF TRIMARAN WITH trimaran with and without anti-pitching appendages in calm
AND WITHOUT T-FOIL water, the motion response of the trimaran was investigated.
The Response Amplitude Operators (RAOs) of heave, pitch
To investigate the calm water resistance of the trimaran motions and vertical acceleration at 19th hull station were
with and without T-foil, numerical simulations and compared with the model test results as shown in Figs. 11÷13.
experimental tests were carried out. Tab. 9 presents the In Fig. 11 it can be observed that the CFD method over-
experimental and numerical results of the calm water estimated the heave response values around the resonant
resistance of the bare trimaran. The total resistance from period, while under-estimated the heave response values
model tests exceeded the total resistance from numerical for short and long wave conditions. Fig. 12 indicated that
simulations by about 0.92%, which indicates that the numerical the CFD method under-estimated the pitch response values
method was able to simulate the calm water resistance of around the resonant period. Fig. 13 showed that the CFD
the trimaran with a reasonable small error. Tab. 10 shows method was in a good agreement with the experimental
the experimental and numerical values of the calm water results around the resonant period, while under-estimated
resistance of the trimaran at different values of T-foil attack the vertical acceleration at 19th hull station for short and long
angle (±5o). The total resistance from experimental tests also wave conditions. In general, the CFD method used in this
exceeded the total resistance from numerical simulations. paper was capable of calculating the longitudinal motions of
The errors between the experimental (EFD) and CFD the bare trimaran with a relative high speed.
results turned out to be 2.74 % for -5o T-foil attack angle and
2.62% for +5o T-foil attack angle, respectively. Compared
to the resistance of the bare trimaran in calm water, the
resistance of the trimaran with different T-foil attack
angles increased by 8.05% for T-foil (-5o): EFD, 7.17% for
T-foil (-5o): CFD, 9.23% for T-foil (+5o): EFD, 6.85% for
T-foil (+5o): CFD. As a result of the application of the anti-
pitching appendage ( i.e.T-foil ) the calm water resistance
of the trimaran increased, and for different T-foil attack

72 POLISH MARITIME RESEARCH, No 3/2019

Unauthentifiziert | Heruntergeladen 19.10.19 13:07 UTC

 considered identical to those applied to the simulations of
the bare trimaran. In considering the calm water resistance
characteristics of the trimaran with T-foil, only the T-foil
attack angle of -5o was taken into account in this section.
Response values of heave and pitch motions and vertical
acceleration at 19th hull station calculated by the CFD method
were compared with the results of the experiment as shown in
Figs. 14÷16. In Fig. 14 it can be observed that the CFD method
under-estimated the heave response values, especially for
the short waves (λ/L <0.88), which can be explained by that
for short waves higher requirements as to grid setting are
necessary. The amplitude of the heave response is reduced by
19.6% in case of the CFD method and 13.6% for experimental
results. Fig. 15 indicates that the CFD method was in a good
Fig.11 Heave RAO
accordance with the experimental results. The amplitude of
the pitch response is reduced by 39.1% for CFD method and
42.4% for experimental results. Fig. 16 shows that the CFD
method a little under-estimated the vertical acceleration at
19th hull station. The amplitude of the acceleration is reduced
by 41.3% for CFD method and 40.1% for experimental results.
Figs. 14÷16 obviously show that the T-foil can significantly
reduce the longitudinal motions response values of the
trimaran with a high speed.

Fig.12 Pitch RAO

Fig.14 Heave RAO

Fig.13 RAO of acceleration at 19th hull station ( A19)

SHIP MOTIONS OF TRIMARAN FITTED

WITH T-FOIL
The longitudinal motions of the trimaran at high speed
are severe. For the purpose of improving the seakeeping Fig.15 Pitch RAO
performance of the trimaran, T-foil was used as an anti-
pitching appendage to reduce the motion response of
the trimaran in waves. The selected wave conditions are

POLISH MARITIME RESEARCH, No 3/2019 73

Unauthentifiziert | Heruntergeladen 19.10.19 13:07 UTC

 amplitudes of the heave, pitch and acceleration were reduced
by 22.72%, 20.93% and 22.16%, respectively. With the wave
amplitude increasing, the T-foil had different influence on
values of heave and pitch motion and acceleration. The effect
of reducing heave motion appeared a little greater, but in
contrast, the effect of reducing pitch motion and acceleration
became lower, the reduction was from 39.1% to 20.93% for
pitch motion and 42.4% to 22.16% for acceleration motion.

Fig.16 RAO of acceleration at 19th hull station ( A19)

SHIP MOTIONS OF TRIMARAN

IN THE CONDITION OF ζa*K =0.1
Longitudinal motion characteristics of the trimaran with
and without T-foil were investigated in the condition of ζa*k
within the range of 0.046 (longest wave) and 0.063 (shortest
wave). In the considered situations, a high amplitude may Fig.17 Heave RAO
occur making a different influence on the ship motion in
waves. The longitudinal motions of the trimaran with and
without T-foil in the condition of ζa*k=0.1 were investigated.
The wave conditions for the CFD calculations are shown in
Tab. 11. As can be seen in Tab. 11, the wave amplitude ζa times
wave number is equal to about 0.1.
Tab. 11 Calm water resistance of the trimaran without anti-pitching
appendages

Wave length Wave amplitude

ζa*k
(λ/Lpp) ζa (m)

0.50 0.024 0.101

0.75 0.036 0.101

0.88 0.042 0.100

Fig.18 Pitch RAO
1.0 0.048 0.101

1.25 0.060 0.101

1.38 0.066 0.100

1.60 0.077 0.101

1.75 0.084 0.101

2.0 0.096 0.101

2.25 0.108 0.101

Response values of heave and pitch motions and vertical

acceleration at 19th hull station calculated by the CFD method
are shown in Figs. 17÷19. In the figures it can be observed that
Fig.19 RAO of acceleration at 19th hull station ( A19)
in the wave condition ζa*k=0.1, the T-foil played some role in
reducing the amplitudes of heave, pitch and acceleration. The

74 POLISH MARITIME RESEARCH, No 3/2019

Unauthentifiziert | Heruntergeladen 19.10.19 13:07 UTC

 INFLUENCE OF T-FOIL SIZE ON
TRIMARAN MOTIONS
To investigate the influence of size of T-foil on trimaran
motions, three sizes of T-foil were considered. During the
CFD simulations, the applied T-foil was of the dimensions
decreased and enlarged by 20% in each direction for the
horizontal wings ,with other parameters unchanged. The
dimensions of the so differentiated T-foils are shown in Tab.
12. The T-foil origin size was marked T2.
Tab.12 Dimensions of the different -size T-foils

Principal particulars T1 T2 T3
Airfoil shape NACA0012 Fig.21 Pitch RAO

Span (m) 0.1152 0.144 0.1728

Chord (m) 0.0768 0.096 0.1152
Aspect ratio 1.5 1.5 1.5
Rotating angle of the flap () -5 -5 -5
Width of the vertical foil (m) 0.01 0.01 0.01
Length of the vertical foil (m) 0.060 0.060 0.060
Height of the vertical foil (m) 0.024 0.024 0.024
Longitudinal position of foil
installation place (distance between
1.320 1.320 1.320
centre line of vertical foil and 10th hull
station of the trimaran) (m)

The response values of heave and pitch motions and vertical

acceleration at 19th hull station calculated by the CFD method
are shown in Figs. 20÷22. In the figures it can be observed Fig. 22 RAO of acceleration at 19th hull station ( A19)
that each of the three T-foils of different size produced
different effect on the longitudinal motion response values
of the trimaran. The trends of the RAOs were the same for In addition, the influence of the change of T-foil size on the
the considered different-size T-foils. With the increase of calm water resistance of the trimaran , were studied. Tab. 13
the T-foil dimensions, their effect on pitching stabilization shows the results of the numerical calculations of the calm
became greater. The amplitudes of the heave, pitch and water resistance of the trimaran with T-foils of different size.
acceleration were reduced by 17.68%, 15.17% and 13.15% for Compared to the calm water resistance of the bare trimaran,
T1 foil and 29.72%, 25.73% and 31.44% for T3 foil, respectively. the resistance of the trimaran with T-foils of different size
increased by 5.28% for T1(-5o), 7.17% for T2 (-5o), 11.53% for
T3(-5o). With increasing T-foil dimensions, the calm water
resistance of the trimaran increases.
Tab.13 Calm water resistance of the trimaran with T-foils of different size

T1(-5): T2 (-5): T3 (-5):

CFD(N) CFD (N) CFD (N)
Resistance (N) 38.426 39.116 40.707
ΔR (%) 5.28 7.17 11.53

CONCLUDING REMARKS AND

DISCUSSION
Fig.20 Heave RAO In the present work, to investigate the seakeeping and
resistance characteristics of the trimaran at a relative high
speed (Fn =0.628), the unsteady RANS method was used

POLISH MARITIME RESEARCH, No 3/2019 75

Unauthentifiziert | Heruntergeladen 19.10.19 13:07 UTC

to simulate the calm water resistance of the trimaran, the of Mechanical Engineering, 4(2), (2007). pp. 183-195.
longitudinal ship motions (heave and pitch ) and vertical
acceleration at 19th hull station for a broad range of wave 2. Mynard, T. : Numerical and experimental study of wave
conditions. The wave conditions covered a series of wavelength resistance for trimaran hull forms. Vol.1, (2008), pp.117-132.
λ varying between 1.5 m and 6.75 m, thus the λ/L ratio varied
between 0.50 and 2.25. 3. Brizzolara, S., Capasso, M., Ferrando, M., et al. : Trimaran
The longitudinal motion responses of the trimaran at a hull design for fast ferry applications. Int. Conference on
relative high speed were rather intense. In order to reduce Ships Design and Shipping, Nav. 2003.
the longitudinal motions of the trimaran a T-foil was used
as an anti-pitching appendage which was located at the 19th 4. Bertorello, C., Bruzzone, D., Cassella, P., et al. : Trimaran
hull station near the bow of the middle hull of the trimaran. model test results and comparison with different high
The T-foil can significantly reduce the amplitude of the speed craft. Practical Design of Ships & Other Floating
longitudinal motions of the trimaran, but the increase of Structures, 1, (2001), pp. 143-149.
calm water resistance was a detrimental effect.
All experimental measurements obtained from the model 5. Hebblewhite, K., Sahoo, P.K. and Doctors, L.J.: A case
tests were used for comparison with numerical results. study: theoretical and experimental analysis of motion
Therefore, to avert scaling errors, all numerical investigations characteristics of a trimaran hull form, Ships and Offshore
were conducted in corresponding model scale. Structures, 2:2, (2007) , pp.149 – 156.
The STAR-CCM+ unsteady RANS solver was used for
the simulation of the trimaran behaviour. The method was 6. Fa lt i n s e n , O. M . : Nu me r ic a l pre d ic t ion
extended by an explicit six-degrees-of-freedom solver to of sh ip mot ions at h ig h for wa rd speed .
compute nonlinear ship motions. CFD simulations have been  Phil. Trans. R. Soc. , London Ser. A, 334, (1991), pp. 241-252.
validated by comparison with results of the experimental tests.
The verification study at the resonant frequency demonstrated 7. Salvesen, N., Tuck, E.O., Faltinsen, O.M. : Ship Motion and
that the numerical results were more sensitive to the time step Sea Loads. SNAME, 1970.
than to the mesh size. The validation was positive for all the
considered variables. 8. Wei, Y.F., Duan, W.Y., Ma, S. : Trimaran motions and
Comparisons between the experimental results and hydrodynamic interaction of side hulls, 9th International
numerical calculations showed that the unsteady RANS Conference on Fast Sea Transportation, FAST 2007, pp.
method predicts ship longitudinal motions with a high 413-421.
accuracy.
Compared with the results for the bare trimaran , the 9. Wang, S. M., Ma, S., & Duan, W. Y. : Seakeeping optimization
use of anti-pitching appendage in the form of T-foil greatly of trimaran outrigger layout based on NSGA-II. Applied
reduced the longitudinal motions. In the wave condition of Ocean Research, 78, (2018), pp. 110-122.
ζa*k=0.1, the amplitudes of the heave, pitch and acceleration
were reduced with the help of T-foil by 22.72%, 20.93% and 10. Sato, Y., Uzawa, K., Miyata, H. : Validation of motion
22.16%, respectively. In the wave condition of ζa*k of a value prediction method for trimaran vessels. International
between 0.046 and 0.063, the T-foil produced a little greater Conference on Numerical Ship Hydrodynamics 2007,
effect on heave motion, but its effect on pitch and acceleration (Vol.5).
was reduced by about 20%. With the increase of T-foil size,
the effect of reducing longitudinal motions becomes greater, 11. Wu, C.S, Zhou, D.C, Gao, L., et al. :). CFD computation
however, the calm water resistance of the trimaran increases of ship motions and added resistance for a high speed
by 5.28% to 11.53% compared with the bare trimaran. trimaran in regular head waves, International Journal of
The nonlinear effects on the longitudinal motions of Naval Architecture and Ocean Engineering, 3(1), (2011),
the trimaran in wave of a relatively high amplitude will pp. 105-110. doi: https://doi.org/10.2478/ijnaoe-2013-0051.
be further investigated experimentally in the future.
The use of T-foil can significantly reduce longitudinal motions 12. Poundra, G.A.P., Utama, I.K.A.P., Hardianto, D., et
but has a negative effect on resistance, therefore the combined al. : Optimizing trimaran yacht hull configuration
use of T-foil and stern flap will be further studied with the based on resistance and seakeeping criteria. Procedia
aim to realize the comprehensive optimization of resistance Engineering, 194, (2017), pp. 112-119.
and seakeeping performance of the trimaran.
13. Fang, C.C., Chan, H.S. : An investigation on the vertical
motion sickness characteristics of a high-speed catamaran
BIBLIOGRAPHY ferry. Ocean Engineering, 34(14), (2007), pp. 1909-1917.

1. Davis, M.R., Holloway, D.S.: A comparison of the motions of 14. Esteban, S., Giron-Sierra, J.M., Andres-Toro, B.D., et al. :
trimarans, catamarans and monohulls. Australian Journal Fast ships models for seakeeping improvement studies using

76 POLISH MARITIME RESEARCH, No 3/2019

Unauthentifiziert | Heruntergeladen 19.10.19 13:07 UTC

 flaps and t-foil. Mathematical & Computer Modelling,41(1), 27. Bøckmann, A., Pakozdi, C., Kristiansen, T., et al. : An
(2005), pp. 1-24. Experimental and Computational Development of a
Benchmark Solution for the Validation of Numerical
15. Haywood A.J., Duncan A.J., Klaka K.P. et al. : The Wave Tanks. ASME 2014, International Conference on
development of a ride control system for fast ferries. Control Ocean, Offshore and Arctic Engineering (2014), Vol.2, pp.
Eng. Pract. 3(5), (1995), pp. 695-702. V002T08A092-V002T08A092.

16. Giron-Sierra J.M., Esteban S, Andres D., et al. : Experimental 28. Kim, M., Hizir, O., Turan, O., et al.: Estimation of
study of controlled flaps and T-foil for comfort improvement added resistance and ship speed loss in a seaway. Ocean
of a fast ferry. In: IFAC Proceedings control applications in Engineering 141, (2017), pp. 465-476.
marine systems, vol. 34, (2001), pp. 261-266.
29. Kim, M., Turan, O., Day, S., et al. : Numerical studies on
17. Zong, Z., Sun, Y., Jiang, Y. : Experimental study of added resistance and ship motions of KVLCC2 in waves.
controlled T-foil for vertical acceleration reduction of a Ocean Engineering 140, (2018), pp. 466-476.
trimaran. Journal of Marine Science & Technology, (2018),
pp. 1-12. 30. Hizir, O., Kim, M., Turan, O., et al. : Numerical studies
on non-linearity of added resistance and ship motions of
18. CD-Adapco. User guide of STAR-CCM+ , Version 11.0.2, KVLCC2 in short and long waves. International Journal of
2016. Naval Architecture and Ocean Engineering ,(2018). https://
doi.org/10.1016/j.ijnaoe.2018.02.015.
19. Patankar, S.V., Spalding, D.B. : A calculation procedure for
heat, mass and momentum transfer in three-dimensional
parabolic flows. Int. J. Heat. Mass Transf. 15 (2), (2005), CONTACT WITH THE AUTHORS
pp. 1787-1806.
Ang Li
20. Tezdogan, T., Demirel, Y.K., Kellett, O., et al. : Full-scale e-mail: [email protected]
unsteady RANS CFD simulations of ship behavior and
performance in head seas due to slow steaming. Ocean College of Shipbuilding Engineering
Eng. 97, (2015), pp. 186-206. Harbin Engineering University
Harbin
21. International Towing Tank Conference (ITTC) : Practical china
guidelines for ship CFD applications. In: Proceedings of
the 26th ITTC, Rio de Janeiro, Brazil, 2011b. Yunbo Li
e-mail: [email protected]
22. Choi, J., Yoon, S.B. : Numerical simulations using
momentum source wave-maker applied to RANS equation College of Ocean Science and Engineering
model. Coast. Eng. 56 (10), (2009), pp. 1043-1060. Shanghai Maritime University
Shanghai
23. Sun, H.B., Jing, F.M., Jiang, Y., et al. : Motion prediction china
of catamaran with a semi-submersible bow in wave.
Polish Maritime Research 1(89) , Vol. 23, 2016 ,
pp. 37-44.

24. Stern, F., Wilson, R., Shao, J. : Quantitative V&V of CFD

simulations and certification of CFD codes. International
Journal for Numerical Methods in Fluids, 50(11), (2006),
pp. 1335-1355.

25. Sun, X.S., Yao, C.B., Xiong, Y., et al.: Numerical and
experimental study on seakeeping performance of a swath
vehicle in head waves. Applied Ocean Research, 68, (2017),
pp. 262-275.

26. International Towing Tank Conference (ITTC), Ocean

Engineering Committee : Final report and recommendation
to the 27th ITTC. In: Proceedings of the 27th ITTC,
Copenhagen, 2014.

POLISH MARITIME RESEARCH, No 3/2019 77

Unauthentifiziert | Heruntergeladen 19.10.19 13:07 UTC