Design of Passive Anti-Roll Tanks For Roll Stabilization in The Nonlinear Range
Design of Passive Anti-Roll Tanks For Roll Stabilization in The Nonlinear Range
Design of Passive Anti-Roll Tanks For Roll Stabilization in The Nonlinear Range
Technical Note
Design of passive anti-roll tanks for roll
stabilization in the nonlinear range
Khaled S. Youssef, Saad A. Ragab, Ali H. Nayfeh *, Dean T.
Mook
Department of Engineering Science and Mechanics (MC 0219), Virginia Polytechnic Institute and
State University, Blacksburg, VA 24061, USA
Abstract
The best way of reducing roll motion is by increasing roll damping. Bilge keels are the
most common devices for increasing roll damping. If more control is required, anti-roll tanks
and fins are used. Tanks have the advantage of being able to function when the ship is not
underway. Our objective is to develop design procedures for passive tanks for roll reduction
in rough seas. This paper focuses on the design of passive U-tube tanks. The tank-liquid
equation of motion is integrated simultaneously with the six-degree-of-freedom (6DOF) equa-
tions of the ship motion. The coupled set of equations is solved by using the Large Amplitude
Motion Program ‘LAMP’, which is a three-dimensional time-domain simulation of the motion
of ships in waves. The unstabilized and stabilized roll motions of a S60-70 ship with forward
speed and beam waves have been analyzed. For high-amplitude waves, the unstabilized roll
angle exhibits typical nonlinear phenomena: a shift in the resonance frequency, multi-valued
responses, and jumps. The performance of a S60-70 ship with a passive tank is investigated
in various sea states with different encounter wave directions. It is found that passive anti-
roll tanks tuned in the linear or nonlinear ranges are very effective in reducing the roll motion
in the nonlinear range. The effect of the tank damping, frequency, and mass on the tank
performance is studied. Also, it is found that passive anti-roll tanks are very effective in
reducing the roll motion for ships having a pitch frequency that is nearly twice the roll
frequency in sea states 5 and 6. 2001 Elsevier Science Ltd. All rights reserved.
0029-8018/02/$ - see front matter 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 2 9 - 8 0 1 8 ( 0 1 ) 0 0 0 2 1 - X
178 K.S. Youssef et al. / Ocean Engineering 29 (2002) 177–192
1. Introduction
During the past century, reducing roll motions of a ship in seaway has received
considerable attention by naval architects. This attention has been prompted by: (a)
consideration of the comfort of passengers on cruise ships; (b) military demands for
steadier gun platforms; (c) reduction in cargo damage; and (d) attempts to increase
the efficiency of crew. Anti-roll tanks are the most commonly used devices to stabil-
ize the roll motion of ships that are not underway. The shifting weight of the fluid
exerts a roll moment on the ship and, by suitable design, this can be made to damp
the wave-excited roll motion. Anti-roll tanks can be classified as either passive or
active, depending on whether the fluid is allowed to move entirely due to the motion
of the ship or is moved with a pump. Passive anti-roll tanks have the advantage of
having no moving parts and requiring little maintenance. Moreover, they operate
well at slow speeds where the use of fins is ineffective.
Froude (1874) was the first to use anti-roll tanks to reduce roll motion. He installed
water chambers in the upper part of the ship. The free-surface effect of the water
tank lengthened the period of the rolling motion and reduced the ship’s stability;
consequently, the system was abandoned. Frahm (1911) was the first to understand
the importance of placing the horizontal leg or cross duct of the U-tube below the
center of gravity of the ship and thereby to take advantage of the stabilizing compo-
nent developed by the horizontal acceleration of the water. The active counterpart
of Frahm’s passive tank was conceived by Minorsky (1935). A restoring moment
was developed by transferring the water directly with a proper phase from one leg
of the U-tube tank to the other at a high rate. A full-scale version of this system
was installed on the destroyer USS Hamilton. In still-water trials, it was found that,
by activating the stabilizer, the Hamilton could be made to roll as much as 18°. In
actual operation, the self-excited oscillations in the control system caused violent
fluctuations in the hydraulically operated pitch mechanism of the impeller. This had
the effect of destroying the phase relationship of the water transfer and thus pre-
venting adequate stabilization of the Hamilton at sea.
In the past few decades, the performance of anti-roll tanks has received consider-
able attention. Chadwick and Klotter (1954) investigated the use of tanks for the US
Navy. Van Den and Vugts (1964) and Stigter (1966) studied the performance of
various types of anti-roll tanks for the Netherlands Ship Research Center. Field and
Martin (1976) evaluated the performance of passive U-tube and free-surface roll-
stabilization systems. Webster et al. (1988) studied the performance of free-flooding
anti-roll tanks during the major upgrade of the USS Midway. Later, Lee and Vassalos
(1996) investigated the use of flow obstructions inside the tank.
Recently Abdel Gawad et al. (2001) studied the performance of passive U-tube
anti-roll tanks. They modeled the ship motion by a single degree-of-freedom in roll.
They presented a detailed parametric study on the effect of tank damping, mass,
location relative to the ship CG, and tuning. They found that a well-tuned, well-
designed tank can be very effective in reducing the roll motion. In the present paper,
we consider the 6DOF equations of motion of the ship and design anti-roll tanks
for the stabilization of roll motion. A subroutine for integrating the tank equation is
K.S. Youssef et al. / Ocean Engineering 29 (2002) 177–192 179
added to the Large Amplitude Motion program ‘LAMP’ and all of the equations are
integrated simultaneously in the time domain. LAMP simulates in the time-domain
the motion of floating bodies in waves with or without forward speed. We have
studied the unstabilized and stabilized roll motions of a S60-70 ship with forward
speed and beam waves. The effects of tank damping, frequency, and mass on the roll
damping are studied for various sea states with different encounter wave-directions.
Nayfeh and Khdeir (1986a,b) studied the nonlinear roll motion of ships in various
sea conditions. Nayfeh et al. (1973), Nayfeh (1988), Nayfeh and Oh (1995), and Oh
et al. (2000) studied theoretically and experimentally the nonlinear coupling between
pitch and roll motions in the presence of a two-to-one internal resonance. In this
paper, we also investigate the effectiveness of passive anti-roll tanks for the stabiliz-
ation of roll motion at this resonance condition.
LAMP is currently regarded as the most nearly complete and fully functional ship-
motion code available for time-domain calculations. The body is assumed to be rigid,
three-dimensional, floating on the free wavy surface, and undergoing arbitrary six-
degree-of-freedom (6DOF) motion. LAMP is based on incompressible potential-flow
theory and predicts large-amplitude motions of a surface-piercing rigid body in a
prescribed or arbitrary seaway. Surface tension is not taken into account and the
water depth is considered infinite. The exact boundary conditions are applied on the
instantaneous submerged portion of the hull surface and the free-surface condition
is linearized around the incoming-wave surface. This approximation can be justified
in principle for small wave slopes and slender bodies.
A boundary-element approach is used. The instantaneous wetted surface of the
body is divided into a number of panels over which linearized transient free-surface
sources are distributed. The problem is formulated in a global or ground-fixed coordi-
nate system, which allows for arbitrary large-amplitude motions. A general waterline
integral is used to account for arbitrary translations and distortions of the body water
plane. Diffraction is taken into account by adding the incident wave contribution to
the boundary condition on the body. The hydrostatic and hydrodynamic forces are
calculated, then the rigid-body equations of motion are solved to update the position
and orientation of the body. More details about the LAMP theory and its implemen-
tation can be found in the manual LAMP for code version 2.9.2 available from
Science Applications International Corporation (1998).
LAMP is controlled by input files which can be generated by the user directly or
through a LAMP pre-processor. Depending on the user specification, it generates
several different output files. LAMP post-processors are designed to help the user
process these output files, perform additional computations related to the ship
motions and wave loads, and view the computed results graphically.
180 K.S. Youssef et al. / Ocean Engineering 29 (2002) 177–192
However, because the equations of motion of the ship themselves are coupled, there
are indirect interactions between the tank-liquid motion and the remaining degrees-
of-freedom.
Thus, the equation governing the tank angle t takes the form
attẗ⫹bttṫ⫹cttt⫹[at4f̈⫹ct4f]⫽0 (1)
where f is the roll angle of the ship and
at4⫽Qt(rd⫹hr),
ct4⫽Qtg,
att⫽Qtwr 冉 冊
w hr
2hd wr
,
btt⫽Qtqfwr 冉 冊
w hr
⫹
2h2d w2r
,
ctt⫽Qtg⫽ct4,
and
1
Qt⫽ rtwrw2xt
2
where xt is the length of the tank (in the fore/aft direction). More details of the
derivation of this equation are given by Lloyd (1989) and Abdel Gawad et al. (2001).
The terms inside the square brackets in Eq. (1) couple the motion of the liquid
to that of the ship. The roll-stabilizing moment exerted on the ship by the liquid
inside the tank is given by ⫺[at4ẗ+ct4t]. The tank natural frequency wt is given by
冪a ⫽冪w w+2h h
ctt 2ghd
wt⫽ (2)
tt r r d
which depends on the tank dimensions and the gravitational acceleration g. This fact
will be used for tank tuning as discussed later. The geometric parameters rd, hr, wr,
w, and hd are defined in Fig. 1, and qf is the coefficient of linear damping in the tank.
The tank equation is coupled with the 6DOF equations of motion of the ship:
three translational equations (surge, sway, and heave) and three rotational equations
(roll, pitch, and yaw). These equations are solved by the LAMP code. For more
details about the derivation of these equations, see Fossen (1994). A subroutine for
integrating the tank equation is added to the LAMP code so that all of the equations
are integrated simultaneously in the time domain.
The coordinate system (x, y, z) is fixed to the ship. The x-axis lies in the plane of
symmetry of the hull and points toward the bow with the origin (x=0) at midship.
The y-axis is perpendicular to the plane of symmetry and points toward the starboard,
and the z-axis also lies in the plane of symmetry and points downward. The roll
motion of a S60-70 ship with Froude number=0.2 (nondimensional forward speed)
in beam waves has been studied. The length (LBP=400 ft) between perpendiculars
of the S60-70 ship is taken as the reference length. The nondimensional ship mass
is calculated to be 0.005715. The nondimensional mass moments of inertia Ixx, Iyy,
and Izz of the ship in the ship-fixed coordinate system are taken to be 0.00001429,
0.00035719, and 0.00035719, respectively. The products of inertia are taken to be
zero. The linear natural frequency ws of the roll is approximately 1.45. The wave
amplitude is nondimensionalized according to the ship length.
First, the unstabilized (without the tank) roll motion of this ship in beam waves
with different amplitudes and encounter frequencies is studied in the linear and non-
linear ranges. In Fig. 2 we show variation of the roll angle with the nondimensional
encounter wave frequency. For a small wave amplitude (e.g. 0.0015), the response
is single-valued and linear. Resonance is clearly predicted at the natural roll fre-
182 K.S. Youssef et al. / Ocean Engineering 29 (2002) 177–192
Fig. 2. The roll response of the unstabilized ship as a function of the nondimensional encounter-wave fre-
quency.
quency of 1.45. As the wave amplitude increases, the inherent nonlinearity of the
system bends the response curves to the right and shifts the frequency at maximum
amplitude to values higher than the natural frequency. The peak amplitude increases
with increasing wave amplitude. The curve connecting the peak amplitudes is called
the backbone curve. It is bent to the right, indicating that the nonlinearity is of the
hardening type. The multi-valuedness of the response also produces jumps.
In the design of passive anti-roll tanks, it is important to tune the natural frequency
of the tank to the natural frequency of the ship in roll. Therefore, the performance
of a passive tank/ship system that is tuned in this way may deteriorate at high wave
amplitudes because of the shift in the frequency where peak amplitudes of the
response occur due to nonlinear effects.
First, we consider the linear range. A case is studied with a wave amplitude of
0.0015, see Fig. 3. In this case, the tank is tuned to the linear natural frequency 1.45
of the ship. The roll angle is shown for three values of the nondimensional damping
coefficient qf, namely 0, 0.1, and 0.17. In this case, the tank with different damping
coefficients reduces the unstabilized roll motion by about 95% at the resonance fre-
quency. For the tank with smooth walls (qf=0), we note that there are peaks in both
the low and high encounter-wave frequency regions. The tank adds a degree-of-
K.S. Youssef et al. / Ocean Engineering 29 (2002) 177–192 183
Fig. 3. Roll angle as a function of the encounter-wave frequency for different damping coefficients qf
in the linear range in a beam wave with an amplitude=0.0015.
freedom to the system. By increasing the tank damping coefficient, these two peaks
are also reduced significantly.
Next, we analyze the nonlinear range. Two cases, with wave amplitudes equal to
0.01 and 0.005, are studied (see Figs. 4 and 5). The frequency of the tank/liquid
system is taken to be equal to 1.59 to match the frequency at which the maximum
roll amplitude occurs for the unstable ship. In Fig. 4, the maximum amplitude of
the roll motion is reduced to about 10% of the unstabilized case by using tanks with
either qf=0.1 or with qf=0. The tank with damping reduces the roll motion a little
more for low encounter-wave frequencies and a little less for high frequencies, com-
pared to the tank with smooth walls, see Figs. 4 and 5. In both cases, a typical
response can be noticed. For low encounter frequencies, the tank amplifies the roll
motion; and for high frequencies, the tank reduces the roll motion. In Fig. 5, the
maximum amplitude of roll motion is about 7% of the unstabilized case.
In this section, we compare the performance of two tank/ship systems, in one the
natural frequency of the tank/liquid system is tuned at the linear frequency of 1.45
and in the other it is tuned at the nonlinear frequency of 1.59. A wave amplitude
of 0.01 is used. Fig. 6 shows the roll angles for the two tanks without damping
184 K.S. Youssef et al. / Ocean Engineering 29 (2002) 177–192
Fig. 4. Roll angle as a function of the encounter-wave frequency for different damping coefficients qf
in the nonlinear range in a beam wave with an amplitude=0.01.
Fig. 5. Roll angle as a function of the encounter-wave frequency for different damping coefficients qf
in the nonlinear range in a beam wave with an amplitude=0.005.
K.S. Youssef et al. / Ocean Engineering 29 (2002) 177–192 185
Fig. 6. Roll angle as a function of the encounter-wave frequency for different natural frequencies wt of
the tank, without damping (qf=0) in the nonlinear range in a beam wave with an amplitude=0.01.
(qf=0) as functions of the encounter-wave frequency, and Fig. 7 shows the responses
with damping (qf=0.1). From these two figures, we note that the system tuned in the
nonlinear range has a slightly better performance than the one tuned in the linear
range.
The effect of the mass of the liquid in the tank on the performance of the tank/ship
system is studied in the nonlinear range with a wave amplitude of 0.01, see Fig. 8.
The natural frequency of the tank is taken to be 1.59 and the damping coefficient
is taken to be 0.1. Three values of the tank/ship mass ratio, namely 3, 4.5, and 6%,
are investigated. The results show that the larger the mass of the liquid is, the better
the tank/ship performance is. In the rest of this paper, the tank mass ratio is taken
to be 4.5%.
In this section, we study the use of passive tanks to stabilize the roll motion
induced by a two-to-one internal resonance. The internal resonance occurs when a
ship has a natural frequency in pitch that is nearly twice its natural frequency in
roll. The seaway is chosen to be a long-crested wave. The unstabilized-ship responses
186 K.S. Youssef et al. / Ocean Engineering 29 (2002) 177–192
Fig. 7. Roll angle as a function of the encounter-wave frequency for different natural frequencies wt of
the tank, with damping (qf=0.1) in the nonlinear range in a beam wave with an amplitude=0.01.
Fig. 8. Effect of tank mass on the roll angle as a function of the encounter-wave frequency with damping
(qf=0.1) in the nonlinear range in a beam wave with an amplitude=0.01.
K.S. Youssef et al. / Ocean Engineering 29 (2002) 177–192 187
for various sea states and different encounter-wave directions are given in Fig. 9. It
follows from Fig. 9 that the roll increases significantly for head waves in sea states
5 and 6. This is the result of the two-to-one internal resonance between the pitch
and roll modes and the pitch is being excited with a primary resonance. The time
histories of the roll and pitch angles for sea state 5 are plotted in Figs. 10 and 11,
respectively. The roll instability that results from the influence of the two-to-one
internal resonance is clearly identified.
The performance of the passive tank/ship system for the 6DOF ship motion is
investigated in these various sea states. The natural frequency of the tank/liquid
system is taken to be 1.59. In this analysis, the performance is analyzed for the two
nondimensional damping coefficients qf of 0 and 0.1. The roll amplitude is plotted
as a function of encounter wave directions in Figs. 12–15. Since the tank is tuned
in the nonlinear range, it significantly reduces the roll motion in sea states 3 and 4,
as evident in Figs. 12 and 13. In sea states 5 and 6, the unstabilized S60-70 ship
has high roll amplitudes due to the two-to-one internal resonance, which can also
be reduced by using the tuned passive tank, see Figs. 14 and 15.
Fig. 9. The amplitude for the unstabilized ship as a function of its heading in a regular sea for various
sea states.
188 K.S. Youssef et al. / Ocean Engineering 29 (2002) 177–192
Fig. 10. Time history of the unstabilized-ship roll motion in sea state 5.
Fig. 11. Time history of the unstabilized-ship pitch motion in sea state 5.
K.S. Youssef et al. / Ocean Engineering 29 (2002) 177–192 189
added to the ‘LAMP’, and all of the equations are integrated simultaneously and
interactively in the time domain. The unstabilized and stabilized roll motions of a
S60-70 ship with forward speed and beam waves are analyzed.
The unstabilized frequency–response curves of the roll exhibit typical nonlinear
characteristics. As the amplitude of the beam wave increases, the frequency at which
the maximum roll amplitude occurs is shifted to the right of the ship’s natural fre-
quency in roll and the frequency range for which multi-valued responses occur
increases.
A well-tuned tank can be very effective in reducing roll motions. In the nonlinear
range, a 90% reduction in the roll motion was achieved, while a 95% reduction was
obtained in the linear range. A passive tank tuned in the linear range can also be
effective in reducing the roll motion in the nonlinear range.
For head waves in sea state 5, a S60-70 ship whose natural frequencies in pitch
and roll are in the ratio of two-to-one has high roll amplitudes. This response can
be significantly damped by a passive tank.
The development of an improved model of the passive tank-liquid motion is in
progress. This tank consists of a series of U-shaped pipes placed side bye side along
the length of the ship. The 6DOF equations governing the tank-liquid motion will
be coupled with the 6DOF equations of motion of the ship. All these equations will
be integrated simultaneously in the time domain. Also an active control strategy to
tune the tank will be added to the system.
190 K.S. Youssef et al. / Ocean Engineering 29 (2002) 177–192
Acknowledgements
This work was supported by the Office of Naval Research under Grant No.
N00014-96-1-1123.
References
Abdel Gawad, A.F., Ragab, S.A., Nayfeh, A.H., Mook, D.T., 2001. Roll stabilization by anti-roll passive
tanks. Journal of Ocean Engineering (in press).
Chadwick, J.H., Klotter, K., 1954. On the dynamics of anti-roll tanks. Schiffstechnik 2, 23–45.
Field, S.B., Martin, J.P., 1976. Comparative effects of U-tube and free surface type passive roll stabiliz-
ation systems. Royal Institute of Naval Architecture 2, 73–92.
Fossen, T.I., 1994. Guidance and Control of Ocean Vehicles. Wiley, New York, pp. 5–30.
Frahm, H., 1911. Results of trials of the anti-rolling tanks at sea. INA Transactions 53, 183–197.
Froude, W., 1874. Considerations respecting the rolling of ships at sea. INA Transactions 14, 96–116.
Lee, B.S., Vassalos, D., 1996. An investigation into the stability effects of anti-roll tanks with flow
obstructions. International Shipbuilding Progress 43, 70–88.
Lloyd, A.R.J.M., 1989. Seakeeping-Ship Behaviour in Rough Weather. Ellis Horwood, Chichester, pp.
380–397.
Minorsky, N., 1935. Problems of anti-rolling stabilization of ships by the activated tank method. ASNE
47, 87–119.
Nayfeh, A.H., 1988. On the undesirable roll characteristics of ships in regular seas. Journal of Ship
Research 32, 92–100.
Nayfeh, A.H., Khdeir, A.A., 1986a. Nonlinear rolling of biased ships in regular beam waves. International
Shipbuilding Progress 33, 84–93.
Nayfeh, A.H., Khdeir, A.A., 1986b. Nonlinear rolling of ships in regular beam seas. International
Shipbuilding Progress 33, 40–49.
192 K.S. Youssef et al. / Ocean Engineering 29 (2002) 177–192
Nayfeh, A.H., Oh, I.G., 1995. Nonlinearly coupled pitch and roll motions in the presence of internal
resonance; part 1, theory. International Shipbuilding Progress 42, 295–324.
Nayfeh, A.H., Mook, D.T., Marshall, L.R., 1973. Nonlinear coupling of pitch and roll modes in ship
motion. Journal of Hydronautics 7, 145–152.
Oh, I.G., Nayfeh, A.H., Mook, D.T., 2000. A theoretical and experimental investigation of indirectly
excited roll motion in ships. Philosophical Transactions of the Royal Society, Mathematical, Physical
and Engineering Sciences 358, 1853–1881.
Science Applications International Corporation, 1998. Large Amplitude Motions Program (LAMP) for
Ship Motions and Wave Loads Predictions, Version 2.9.2 User’s Guide.
Stigter, C., 1966. The performance of U-tanks as a passive anti-rolling device. The Royal Institution of
Naval Architects 13, 249–275.
Van Den, B., Vugts, J.H., 1964. In: Some Notes on the Performance of Free Surface Tanks as Passive
Anti-rolling Devices, The Report of the Shipbuilding Laboratory No. 119. Technological University
of Delft, Delft, pp. 3–52.
Webster, W.C., Dalzell, J.F., Barr, R.A., 1988. Prediction and measurement of the performance of free-
flooding ship anti-rolling tanks. SNAME Transactions 96, 333–364.