4
A Free Surface
as an A
Vice for Fishing Vessels
by J. J.
vaiz den Bosch
Utilisatioii d'une citerne a carne liquide comme amortisseur de
roufls a bord des bateaux de pciie
L'auteur passe en revue divers systèmes d'amortissement du roulis pour voir s'ils satisfont aux besoins des bateaux de péche. I.e choi se porte stir une citerne constituée par un paralkIépipède rectangle, dans lequel l'énergie est fournie par un simple efTet de "mascaret".
Des equations mathématiques thCoriques de mouvement sont
formulCes pour un bâtiment roulant librernent dans un fluide, et
i'on étudie VinfJuence de divers paramétres sur Ic niouvernent.
L'équation est ensuite modifiCe pour tenir compte de linstallation a bord du bãtiinent d'une citerne simple a carCne liquide. La section qui suit est consacrée aux donnCesrelatives a un reservoir parallélC-pipédique et a i'effet des paramètres du reservoir sur Ic roulis.
L'auteur dCcrit une citerne concue pour un hãtiment particulier,
pour trois conditions de déplacement, et construit des courbes
théoriques qui sont comparées avec les résullats de quatre series d'essais sur modèle. Ii Ctudie les variations par rapport aux résultats théoriques, et termine en présentant diversessuggestions.
TJHERE is no doubt that
many fishing vessels would benefit from the installation of sonic meansIt of roll damping to ease deck working conditions and increase effective fishing time. The major types of
roll damping devicesnow in use are:
o Active fins
o Passive tanks based. on the U-tube principle
o Active tanks based on the U-tube principle
Passive free-surface tanks
This paper deals with the application of the passive free-surface tank in its simplest form.
Bilge keels are excluded from this discussion because other roll damping devices are considered as supple-mentary rather than as an alternative to bilge keels.
This is discussed more fully at the end of the paper. A fishing vessel's damping system should be:
o Effective even at low or zero speeds o Efficient for many conditions
o Inexpensive to install and maintain
Trouble-free in normal use
Active fins are not effective at low speed. Initial costs are relatively high.
Passive tanks based on the U-tube principle are often rather sensitive to differences in the natural roll period of the ship, and if designed to cover a wider frequency
range the overall efficiency falls.
Activated U-tanks are efficient over a wide range of
conditions. They seem good for large boats, but are complicated and may be too costly for small boats.
Although passive free-surface tanks are less efficient
than activated U-tanks, they are simple to install and
perform satisfactorily in most conditions.
[159]
El tanque de balance con superficies libres corno cstabilizador para barcos de pesca
Se estudian y comparan varios tipos de estabilizadores, en orden a
los requisitos dc UflO adecuadopara los barcos de pesca. Se elige uii
simple tanque rectangular, de forma de caja, que aprovecha Ia energIa de Ia ola de marea. Se desarrollan ecuaciones
teórico-matemáticas de movimiento de una embarcación que se balancea
lihrementc en un IIquido, cstudiándose el efecto de los diversos
parámetros sobre ci movimiento. La ecuación se modifica despuCs para instalar en Ia emharcaciOn un simple tanque con superficies libres. La sección siguiente trata de los datos de un tanque
rec-tangular y Ia intluencia de sus parãrnctros en el moviniiento de
balanceo.
Se expone Un ejemplo de diseño detanque para un tipo deter-rninado de emharcaciOn. con tres condiciones de desplazamiento, y se presentan cubiertas teOricamente calculadas, comparándolas con
cuatro series de ensayos de nodelos. Se discuten las variaciones
de los resultados teóricos y se hacen, por ültimo, aigunas suge-rencias.
= &siflwt
The solution of the above equation is expressed by:
Ka
and:
ROLLING MOTION ACCORDING TO SIMPLIFIED TFIEORY
Roiling without tank in operation
The equation of motion. In recent years the theoretical
approach to ship motion
has been improved, and
frequently a simplified mathematical model is used, such as the damped linear mass-spring system (Vossers, 1960).
For the calculation of the influence of the tank on the ship's rolling motion this same method is used only considering the rolling motion. The equation of motion
is given by the expression:
I+N+R.,q=K
if the exciting moment K varies sinusoidally with time, the resulting motion would also be sinusoidal and of the
same frequency. The moment, however, will always be in advance of the motion. The lihase angle between the moment and the motion varies from zero, for very low
frequencies, to 1 80°, for high frequencies. If the moment is expressed by:
K = KCsin(a)t+CKO)
and the motion by:
/
\ (RI.w2)2+No2
Nco
tang CKI
Often the amplitude is written in the form ofa
magni-fication factor, that is, the ratio of the motion amplitude at a certain frequency to the static angle of heel under influence of a heeling moment of the same magnitude. At the natural or resonance frequency:
IR
(O°
the phase angle becomes 900 and the magnification factor can become very large if the damping is relatively small.
A criterion of damping is the non-dimensional damping
coefficient:
N,
\7IR
At resonance the magnification factor amounts to:
The influence of the separate coefficients
Using the above expressions a qualitative analysis of the influence of the separate coefficients can be made:
An increase of I is accompanied by a decrease of
the natural frequency. At the
same time the magnification factor increases because of the in-fluence of I ono An increase of R results in a shift of co towards
a larger value and also in an increase of the
magnification factor at resonanceo An increase of N results in smaller amplitudes
over the entire frequency range
These tendencies are illustrated in fig 1 Starting from
an amplitude characteristic with v0.l (a probable
155
Rft3INA
Fig 1. Influenceof coefficients ofequation of motion
[160]
value for rolling) and co= 1, the effect is shown ofa
doubling of one of the coefficients while the other two
remain unaltered.
Influence of the tank on rolling motion Fundamental behaviour of the tank
When a tank, partially filled with a liquid, say water, is forced to oscillate about a fixed axis, the water
move-ment creates a momove-ment acting about the same axis. When the motion of the tank is sinusoidal themoment. appears to be mainly sinusoidal of the same frequency as the motion, with a phase lag ranging from zero to
180° depending on the frequency. The natural frequency
of the system is defined as the frequency at which the
phase angle equals 90°.
The moment can be resolved in a component which is in phase with the motion and the quadrature coin-ponent which has a phase lag of 90° with the motion.
Expressed mathematically:
the motion is:
aS1110)t and the moment:
M =MaSfl(Wt+Cz)
= MaSin WI COS C + Ma COS Cot Sin E
The first term is the in-phase term and the second is the term with a 90° phase difference. In fig 2 the amplitude and the phase angle are shown as functions of the fre-quency, and in fig 3 the corresponding components are given. l3oth figures serve only as examples to illustrate
the tendencies. 180 -Et degrees go U Ci) sec
M3COSEt
W sec1
Fig 3. Componentsoftankmoment
The equation of motion with the tank in operation
Consider the tank moment as an external moment
acting on the ship on a fore and aft axis through the centre of gravity of the ship. The equation of motion is:
I+NS+R
\Vith 4) = & sin wi this expression becomes:
/
- j
M+
M K
+
(N
co - -- sin cos cot = --- sin (cot + eK)The reduced ui-phase component @fa/4)n)
cos e can
be considered as a reduction of RIw2. The reduced quadrature component (ilIi/4)a) sin e can he considered as an augmentation of the damping when sin r is
negative, i.e. throughout the entire frequency range
con-sidered.
Influence on amplitude characteristic
Utilizing the above, the influence of the components of
the tank moment (fig 3) can easily be combined with the curves in fig 1. Assuming that the natural frequencies of the tank and the ship differ little, it follows:
From fig 3 it appears that the in-phase component
of the tank moment is positive for frequencies below the natural frequency of the tank. For this range the amplitude characteristic of the ship. including the tank, tends to the curve on the left
in fig 1. The positive value of (Ma/4)a) cos i has
the effect that the vessel seems to have a longer natural period. (Reduction of R, or augmentation
of J.)
The quadrature component can be considerable
if compared to the ship's own damping and
because of this the reduction of the amplitude in the range around the combined natural frequency
can be very large
o For the range beyond the natural frequency of the tank the vessel obtains the character of a stiffer
ship due to the negative in-phase component
15 0 F-U Li 10 7 .W1T1.TANLJN.
'
OPERTION\
.' Fig 5 I/
[161]r
0.5 L)se'
1.5 20Fig 4. Schematic pe,centation of tank iiiflueizce
Fig 4 shows a tentative curve of the amplitude versus frequency for the ship-pius-tank system. Notable is the
occurrence of tile two secondary peaks in tile curve. This
)
freedom, which are here the rolling of the ship, and the motion of the tank water. The flatter the phase charac-teristic of the tank, the wider the frequencyrange which is covered by the quadrature component, and the more these two secondary peaks are smoothed.If the natural frequency of the tank is higher than the natural roll frequency of the ship, the secondary peak in
the lower range is more accentuated, while the other one
can disappear completely.
RECTANGULAR TANK DATA
General
The free surface tank owes its damping characteristics to the development of a bore, a typical shallow water
wave. Fig 5 shows two photographs of this phenomenon
in two consecutive stages. Jt is evident that with every
roll work is done in raising the tank water, thus reducing
0020 0015 0005 0 -180 d.gr.es -50 0
the energy of motion. The theoretical natural frequency for small amplitudes can be derived as follows:
The velocity of propagation of this type of wave is:
C =\/gh
The distance travelled in one period is twice the breadth of the tank so the natural period is:
2b
=
'.,tgli
and the natural frequency is:
2ir
rr
-0)1 = - =A series of tests shows that an increase of the amplitude
of the motion induces the actual natural frequency to
increase.
Fig 6. Non-dinien3iolzal amplitude and phase of tank Inonle/il for S/b =0
[162]
S/b=O
(i.O10j
T
-/
'I-
hla' .f..-4 OS I0 Cs to wV7 is
O.O2 aois aoio a a - 1zo t-9
The data shown in fig 6,
7, 8 and 9 are results of
experiments with a model tank excited by an oci1lating mechanism. While the tank performed a swinging
motion the moment about the axis of rotation was
measured. In fig 6 and 7 the phase and the reduced moment amplitude are shown versus the reduced fre-quency c'/h/g and in fig 8 and 9 the sine and cosine components are given. The amplitude is made
non-dimensional by dividing by p1gb3 1.
Influence of parameters
There are six parameters which control the tank moment: The motion amplitude
11 & increases the moment amplitude does not increase at the same rate and so the tank is less
effective when the motions are large
[163 1
o The influence of the frequency of the motion co is evident from fig 6, 7, 8 and 9
The tank breadth b
The moment amplitude varies with the third
power of the tank breadth provided that the ratio of the water depth and the breadth is kept con-stant; (the breadth is measured across the ship)
o The tank length 1
The moment is directly proportional to the tank length (measured in fore and aft direction) o The water depth Ii
As is shown, the natural tank frequency depends on the breadth and the water depth. in addition to this influence on the phase relation, the depth of the water influences the total weight and
there-fore the moment
o The height of position s
The vertical height of the position of the tank is F2
s/b=O.20
of b.\
-
44
-r Os to t5 as to wV isU
U Fig 8. Non-dimensional components of tank moment for S/b=0
measured from the axis of rotation to the tank bottom. A negative value means that the tank bottom is situated below the axis of rotation A comparison of fIg 6 and 7 or $ and 9 indicates that a more highly situated tank produces a larger
stabilizing moment
Considerations for application
The presented data can he used for ships with
values ranging from 0.03 to 0.18 B
The static reduction of GAi due to the free sur-face of the tank must be acceptable. The loss of static stability can demand a restriction of the tank dimensions. The reduction of GM, should not be taken into account while the ship's own
natural period is being calculated
@ A roll damping tank should, if possible, extend over the full breadth of the vessel because of the large influence of the breadth on the moment
amplitude
o The tank should be situated as high as possible
[164]
s/b= 020
Fig 9. Non-dimensional components of tank mnonzent for S1b = 0.20
o From experience it is known that the minimum depth of the tank should be approximately three times the water depth in the tank. This influences
the position of the tank in height. As a preliminary estimation for the tank depth the value D 2 GAI can be used
o The data which are shown in fig 6, 7, 8 and 9 are results of measurements with çf= 0.10 radians or about 6°. The calculation of the influence of the
tank on the rolling motion is based on the assumed
linearity of the system. When the results are
interpreted, it has to be borne in mind that this assumption is not strictly true
The diversity of conditions imder which a vessel has to fulfil its task makes it very difficult to
suggest an optimal design for the anti-rolling tank, and for that reason a compromise has to he found.
EXAIPLE OF APPLICATION
Ship and tank data
As an example the design of an anti-rolling tank for a
small trawler is discussed. The
7i/B values of this
s/b=0
---Lt
S/b=-0.20-I
4Ii
\\
'ii-i'
---//i ---
'-'
--t()
ship are fairly high in order to comply with Rahola's stability criteria. Such a vessel, operating on the North
Sea or in comparable areas, often meets conditions which may cause it to roll heavily.
The main dimensions are: Loa=91.87 ft (28.00 m) Lpp=78.42 ft (23.90 m)
B =20.28 ft (6.18 m)
D = 11.36 ft (3.46 m)
From a variety of possibilities three tentative loading conditions were selected for further investigation. The
conditions are summarized in table 1.
TABLE 1: Considered loading conditions
The "fishing condition" (B in the table) was rather extreme. For the calculation of GM it was assumed that the consumption of fuel and stores was about 22 tons, that the fish hold contained about 15 tons catch and that
a new catch weighing about 30 tons lay on deck.
The values ofM, for the conditions A and C satisfied the criteria of Rahola, also if the reduction of GM due
to the free surface in the tank was accounted for.
The position of the tank was chosen approximately amidships taking up part of the bunker space between the engine room and the fish hold. It extended over the
full breadth of the vessel, so b=B=20.28 ft (6.18 m). The
length of the tank was restricted to 4.40 ft (1.34 m) in accordance with the mentioned values and the criteria of Rahola. The values of s/b in table 1 show
rounded-off values following from an assumed height of
the tank 6.56 ft (2.00 m) above the base line.
Determination of water depth and discussion of the tank
effect
The choice of depth of water in the tank is governed by the requirements of easy operation. it is important that the captain of a small fishing vessel is not burdened by such matters as adjusting the water level in the tank to the momentary value. Once it is filled to its pre-scribed level in port it should be unnecessary to give it
any attention, except in emergencies.
In fig 10 the curves of/2a sin e, and ia cos are given
for the vertical tank position s/b= -0.150. These curves
Were obtained by interpolation from the diagrams in the
figures. Although the ratio s/b varies slightly for the three loading conditions, the mean value was sufficient
15 10 n 10 5 -2 -2 1.0 x 13 - 1a NE l5aCt sE 0.5 x 10 sfb=-0 150 0.10
Fig 10. Components of tank inoineizt
CONDITION A
h/bOO8
9
WITH TANK EMPT
WtTBJA1JK IN OPERATION Conaition Quantity A Leaving port B .E'uring fishing C A verage of other conditions 4 tons (ton) 161.4 (164) 184.1 (187) 172.2 (175) GM ft (m) 2.75 (0.84) 1.87 (0.57) 2.36 (0.72) GM 0.136 0.092 0.117 B K=O-4 B ft (rn) 8.11 (2.46) 8.11 (2.46) 8.11 (2.46) Tsec 5.39 6.55 5.83 c sec' 1.165 0.960 1.079 0.07 0.07 0.07 KG ft (m). 9.12 (2.78) 9.97 (3.04) 9.55 (2.91) s/b --0.125 -0.175 -0.150 0 0.5 10 1.5 20 C.) sec Fig 11. Results of calculation
to determine the desired water depth for all conditions. The appropriate quantities for these conditions are listed
in table 2.
The reduced frequency wv"b/g corresponding to the natural frequency of the ship with an empty
tank
165]
-0,lxlO
p
P 0 or tL 10 z 0 I-or 0 Li z 0 or 5 0 CONDITION BWITH TANK EMPTY
,WIIH..IANKIN
hA0.08
,1
OPERATION0.5 1.0 15
L) Secl
The water depth ratio (h/b)th which was derived
from the consideration that the theoretical
natural frequency of the tank and the ship should
be equal
The water depth (/iJb)0 which was derived from fig 10 giving the largest sine component of the
moment at the stated reduced frequency
Evidently, as table 2 shows, the water depth for which
the largest damping effect at a given frequency was obtained, was somewhat larger than the theoretical value. To comply with the demand for simplicity, one value of h/h must be chosen. In order to make a correct choice, the effect of the tank on the amplitude charac-teristic was calculated for all three loading conditions, for their respective s/b ratios and for two water depths,
namely h/b=O.06 and h/h == 0.08. The tank was assumed to be filled with fresh water.
TABLE 2: Reduced natural frequencies of the ship and water-depth
ratios under consideration Condition
h/b 0.06
L_L_L I _LL.L_ I
20
The actual calculation is omitted from the paper. The
results are shown in fig 11, 12 and 1 3. Fig 11 represents
condition A. It was evident that with both depths the [166J 15 5 CONDITION C YIITH.JANK E1PTY 0 05 Jb0008 WITh TANIcJri QRERAJ ION 0O6 1.0 1.5 C.) sec 20
tank gave considerable damping. Although the peak
amplitudes of both curves were approximately equal, the largest water depth was to be preferred because, with the
peak occurring at a lower value of w, the accelerations of the vessel will be less. In fig 12 the results of the calculation are shown for condition B. In this condition the vessel was less stiff. The natural frequency of the ship was considerably lower than that of the tank for /i/b=0.OS, which accounted for a marked peak in the
lower frequency range. Although the curve for /z/b=0.08
was certainly an improvement in comparison with the original characteristic, the aniplitude characteristic for
h/b=0.06 was better. For the third condition fig 13 shows
a result which was somewhere between the other two
conditions as was to be anticipated.
The main conclusion from these figures was that, in spite of the different water depths in the tank and the different loading conditions, the calculated amplitude
characteristics were all very similar and all show a rather
large improvement over the amplitude characteristics of the ship without the tank operating. A sound choice was li/b = 0 07, which gave a water depth of h= 0.07 x 20.28
= 1 42 ft (0.433 m). The total amount of water was Lxbxh= 126.71 ft3 (3.586 m3). This was about 2 per
cent of the average displacement.
MODEL EXPERIIENTS
Purpose and performance
The large number of simplifications which had been introduced in the course of the calculation procedure
required checking. Item lb U)
/-A 0.925 B 0.762 C 0.856 Ag (/z'b)h 0.087 0.059 0.074 (h/b)0 0.094 0.070 0.082Fig 12. Results ofcalculatioiz Fig 13. Results ofcalculation
The tests were carried out with a 1 : 1 5 scale ship
model in which a roll oscillator and a recording gyro-scope had been installed. In this model a tank was installed with dimensions according to those mentioned previously. Next, the model was ballasted and trimmed so its stability and natural roll period corresponded to
the condition A specified in table 1. By means of the roll
oscillator the model was subjected to a moment with a
sufficiently constant amplitude having any desired
frequency within the considered range. The gyroscope served to measure the roll angles.
The following test series were carried out with the model with the tank empty, then filled to the level
corresponding with li/b = 0.08:
An oscillation test with the model subjected to roll moments with a constant amplitude of 0.289 lb.ft (004 kg.m) but with different frequencies The object was firstly to determine an acceptable
value of v, for the ship with empty tank (the value
of i'=0.07 which was introduced in table
1), and secondly, to furnish a comparison with thecalculated curve for the stabilized ship
o A similar test but with a larger moment amplitude,
namely, M0=0.434 lh.ft (0.06 kg.m). The aim of this was to obtain an impression of the degree of the non-linearity of the system, both with and without the tank in operation
o The third series consisted of experiments in
regular waves. The model was held at zero for-ward speed and the waves on the beam, butwas
free to roll, drift and heave. The wave dimensions
and roll
angles were measured. From these measurements the ratio between the rolling amplitude and the wave slope amplitude has been calculated as a function of the frequency [a/k(W)]The purpose of this
test series was twofold. Firstly, it should provide a comparison with the results of the oscillation tests and with the cal-culation from which it could be determined if the motions of drift and heave did have a significant influence on the performance of the stabilized ship. Secondly, these measurements should formthe basis for comparison with the results obtained in an irregular wave pattern
o The object of the fourth series was the measure-ment and analysis of the model rolling motion
with zero forward sl)eed in irregular beam seas. A
check on the correctness of the mathematical assumptions in this particular case was provided by a comparison between the experimentally
determined spectrum of the rolling motion and the
spectrum calculated from both the results of tests
in regular waves and the measured spectrum of the
waves
The results
In fig 14 the results of the first series of tests are shown.
The measured and calculated amplitude characteristics
for the ship without the tank closely resemble each other.
The curves with the tank in operation differ. The cal-culated curve appears to be a mean of the experimental
curve. The curve of measured roll amplitudes, presented
[167] 15 cc 0 U U-10 5 15 CONDITION A CALCULATED CURVE FOhLb0 1YtItLiTM4k EMPTY fLEXPERIMZNT WLTK MO289f1tbs. JLCAL CLLA I ED ACCORtIN T. LINE.AR 4. EQUATION OF lMQT ION IANKJt OPERATION E.X.PERIMENT WITH Ma
-
0299ft.tbs N 05 10 i15 20 (ii sec.Fig 14. Results of oscillation tests compared with calculated cuimes
CONDITION A ALCULATED CVE FOR hLb0.08 LW1TJ±IANKMPLY \E XP EM EN I WITH I MQ289ftWS 0
12.
EXFEfJ11ENLWJIH Mz0.434ftjbs. (a)mx 16.1.1-\\
\j
I TANK IN [OPERATION EXPERIMENT WITH.: MQ2.8i.tt1bs. Ma043Left. Lbs. 0 0.5 1.0W sec 11.5 20Fig 15. Results of oscillation tests for different moment amplitudes in dimensionless form, shows two pronounced secondary
peaks. The reason for this discrepancy is the
non-linearity of the tank moment. The dimensionless pre-sentation obscures the fact that the measured values
were all considerably lower than the 0.1 radians (570) which was the basis of the calculation. For such small amplitudes the curve of e versus the frequency is much
steeper than for an amplitude of
0.1 radians, whichresults in a greater damping in the neighbourhood of the tank's natural frequency and a less damping
else-where. These differences are not of practical importance as the measured values are small.
In fig 15 the results for the largest moment amplitudes
are shown. The amplitude characteristic for the ship without the tank is lower than in the previous case, indicating that the damping coefficient increases with in-creasing amplitude. The curve for the stabilized ship shows the same tendency as has been described in the
-15 5 - çALc'JLATEO 1AGNFICATION -fACTORfQft b/bQ.O8 0 0.5 ItQDELWJTkE TANXMPJY amax2i0 OPERAT)ON \SURED \ALUES 10 U) sec
Fig 16. Results of tests in regularit'ares
former section hut to a lesser degree. For the larger (but still small) rolling angles, the measured values for the largest moment amplitudes show the least deviation
from the calculated curve.
In fig 1 6 the results of the tests in regular waves are
given. Tile non-dimensional roll amplitude curve for
the unstabilized ship seems to be shifted somewhat when
compared with the previously deterniined amplitude characteristics, and the peak is considerably reduced. This last point is undoubtedly partly due to tile larger rolling angles, for the deck entered the water; and probably there is also a decrease of the wave moment due to the orbital motion mentioned earlier.
It is possible that part of this
larger damping andprobably the slight shift of the curve are caused by the coupling of the rolling motion and other motions; i.e. sway and heave. This, however, is not yet clear. Tn any case, this appears to be of no practical significance as is shown by the good agreement between the measured
20
[ 168)
and the originally calculated values, for tile stabilized
ship.
Before considering tile next figures some quantities have to be defined. The roll angles in regular waves are given as ratios to thewave slope. Therefore, the spectral density of the irregular wave pattern is presented bya
wave slope spectrum which is defined by:
S(w) do) E
S:(W) dco
g
-The average value of the highest
third part of tile
observed amplitudes of a random fluctuating quantity, say the roll angle, is often called the significantampli-240 80 0 COND!TION A EXCLUSIVE TANK, 4 S,MEAURED
ii
05 -t cALcjJLA.T.Ep J FROMRESPONSE WAYES Il -fi \ A \xSkck \ 1.5 20 sec.Fig .17. comparison of calculated and measured roll speclra,for the
s/zip alone
tude. Tile significant roil amplitudecan be regarded as a
measure of the impression the amplitude of the rolling motion makes on the observer. For a stationary time series and a spectrum of small width tills quantitycan be calculated from the expression:
Pa1/3 2s/io
where:
?flo = JS(w)dw
0
In fig 17 the wave slope spectrum is shown with two roil spectra, one measured, one calculated from the experimental results in regular waves with the tank not operational. The agreement is
good, as regards the
significant roll amplitudes and the frequency range.
In fig 18 the spectra for the model with the tank in
operation are shown. One spectrum is measured, another
COND!flON A
LCALCVLALED
1 11AGNiF1CAiiLQN
(
is calculated from the results in regular waves, and athird is calculated from the originally computed response shown in fig 14. The agreement is good.
A comparison of the significant roll angles for the
stabilized and the unstabihized ship reveals that an overall reduction of 50 per cent is achieved in this wave spectrum.
20 160 -x (.0 c '-I) 80 C0NDJT!0N A INCLUSIVE TANK. s,,cAcvj.ArEDfiwNi R$fQ1tE L TO R UA8W1E a i/3-4-LLY''/ CQMftUTED AME1JTJDE HARACTEPISTJC
///
<' /_-'/
a :656° S,J1EASURE_±i:
0 05 10 1.1 U) sec:1Fig 18. Gomparison of calculated and measured roll spectra for the S//ij) lilt/i tank
SUGGESTIONS Bilge keels
As has been mentioned, the omission of bilge keels is
not advised. There are two reasons for this. The first and most important is that there can be emergency situations
(i.e. icing up) when it will be necessary to empty the tank because of its negative influence on tile statical stability. If no bilge keels are fitted it will leave the ship
with extrenlely low roll damping.
The other reason is that the non-linear effect of the
tank is somewhat neutralized by tile bilge keels. When the
motion amplitudes become large because of bad sea conditions tile tank moment is not so effective. The damping of bilge keels, on tile contrary, increases
con-siderably with increasing motions. Dimensioning the tank
In the foregoing pages little is said about the amount of water a roll damping tank of this type should contain. In this case it was about 2 per cent of the displacement which is certainly not an insignificant amount. An. overall reduction of 50 per cent was achieved but it depends on the sea conditions what this reduction will
he. it is dependent on many factors, which roll amplitudes 2.0
[169]
and roll accelerations are found acceptable. and it is very difficult to define a basic criterion. The ultimate
answer can only be found by experience. It15tile author's
opinion that it does not pay to economize too much on
tile dimensions of the anti-rolling tank, especially ifone
has a free hand during the design stage of the vessel.
Posilion of the tank
If an anti-rolling tank is wanted, one has to provide space for it where it will work efficiently, and the same remarks hold as for tile dimensioning of the tank.
Obstructions in the tank
It is not always possible to avoid placing stiffeners in
the tank side. if the obstruction is snlall, say the stiffener
height is not more than about 10 per cent of the tank length, the influence, as has been shown by tests, is not
serious.
CONCLUSION
A free surface tank of the type presented here, that isa
rectangular tank with flush front and rear bulkheads and bottom, can provide an efficient means of roil damping. its simplicity of installation, ease of maintellance and
reliability nlakes it especially attractive for small vessels. Nomenclature
b Breadth of the tank measured athwartships
D Depth of the tank
f
Magnification factorMagnification factor at the natural frequency of
roll
Ii Depth of water in tile tank measured from the
water surface at rest to the bottom of the tank
J,
Virtual mass moment of inertia of rollk Virtual radius of gyration of roll
/ Length of the tank measured in fore and aft
direction
Al Moment produced by the anti-rolling tank M Moment amplitude produced by tile anti-rolling
tank
in
The integral of the roll spectrum over the
frequency from zero to infinityDamping coefficient of roiling
R Stiffness coefficient of rolling S1(w) Spectral density of wave slope
S4,(w) Spectral density of roll
s Vertical distance of the axis of rotation to tile bottom of the anti-rolling tank
T1 Theoretical natural period of the water motion in the anti-rolling tank
Phase angle between the rolling motion and the tank moment
&113 Significant roll amplitude Non-dimensional lank moment
Pa Non-dimensional tank moment amplitude
Non-dimensional damping coefficient of roll
w1 Theoretical natural frequency
of the water
motion in tile tank
c
V
U)
4(e)
11h1
lid
.A
II, !II__
II!iIlh
JIL'L'L
0
-I
BL1L
II
$
I,
A
15
2xJ0
L.0 -1 1,5sec
FIGURE 1.
INFLUENCE OF COEFFICIENTS
OF EaUATION
W sec
FIGURE 2.
AMPLITUDE AND PHASE OF TANK MOMENT.
I
180
Ct
degrees
15
0
I-U
LI-10
z
0
U
U-z
CD5
10
0
0,5
_11,0 (.4)sec
FIGURE 4.
1,5SCHEMATIC PRESENTATION OF TANK INFLUENCE.
Il
ii
Il
If
Il
I I'1
III
iWITHOIJJT
ILIANK
I I 1 II
'I
/
I
\
\
WITH TANK IN
//
ION
\
%PER
--2
1.4 x 10-2
1.OxlO
-.L SINC
a
t
+L COSE
a
t
-2
0.5x10
0
0.25
0.5
Wj
FIGURE 10.
COMPONENTS OF TANK MOMENT.
-s/b -0.150
(t)a0'0
-
Ia
C OS C
h'bO.lO
= =".7
A,
...b/b=O.1O-
4
SINC
t
..0Q6.0.08
I I I I - I I0.75
1.015
10
0.5
1.0 1.5Ci sec
FIGURE 11.
RESULTS OF CALCULATION.
2.0
CONDITION 1
WITH TANK EMPTY
h/b-0.08
WITH TANK IN
P ER A I I 0 N
i
I I Ih/b= 0.06
I I I I I I ct:0
C-) 1<IL
10
z
0
I-1<(J
IL
z
CDI
I
5
15
4-0
I-U
LL10
z
0
F-U
LLz
0
15
10.5
1.0 1.5L) sec
FIGURE 12.
RESULTS OF CALCULATION.
2.0
CONDITION 2
311TH TANK EMPTY
h/b-008
WITH TANK IN
OPERATION
I I I Ih/b =0.06
,15
'4-1
0
2.0
CONDITION 3
..WITH TANK EMPTY
h/b0.08
I IWITH TANK iN
OPERATION
I I I0.06
I I I I I I I I L_.._____i 0.5 1.0 1.5sec
FIGURE 13.
RESULTS OF CALCULATION.
A5
0
I-C-) LL 10z
0
I-0
U-z
CD15
10
z
0
F-U
Liz
CDI
I
5
10
0.5 1.0 1 1.5W sec.
FIGURE 14. RESULTS OF OSCILLATION TESTS
COMPARED WITH CALCULATED CURVES.
2.0
CONDITION
-1
VITH TANK EMPTY
EXPER MENT WITH
I
M0.289ft.Lbs.
CALCULATED
ACCORDING
TO
-/\
110
EÜUATION OF
MOTION
CALCULATED CURVE
I I IAJK
IN
FOR h/b0.08
IOP-RATION
-I I I IPERIMENT
ITH Ma
-.O.289ft.Lbs.
I I I I I I I I I15
4-0
U
10I
5
1CONDITION
-1
tWITH TANK EMPTY
\\I\EXPERIMENT
WITH
\I
\
I\Mo2e9ftlbs
1-amax.
EXPERAMENT
12.3°
WITH
M -0434ft.lbs.
I Ia)max.16.1°
-CALCULATED CURVE
\
\/
I ITANK IN
A .OPERATION
FOR h/b0.08
EXPERFMENT
-
/
/
-
,i
I I I I I I I IWiTH:
,1=s.2:ft.lbs.
\\
I I IM0.434ft.
lbs.
. 0.5 1.0 CL)sec15
2.0
FIGURE 15. RESULTS OF OSCILLATION TESTS
FOR DIFFERENT MOMENT AMPLITUDES.
0
0.5
1.01 1.5
(5*4)
sec
FIGURE 16.
RESULTS OF TESTS IN REGULAR WAVES.
2.0
CONDITION 1
I ICALCULATED
I'
it
I 'iMAGNIFICATION
FACTOR
i
it
/
/I.
I
/ ISI
I I I i I . IMODEL WITH
TANK EMPTY
a)max.210°
TANK IN
CALCULATED
MAGNIFIC1ATION
5\
IOPERATION
FACTOR FOR h/b0.08'
'V
'
,:
-I I I I\MEASURED
VALUES
-I I I15
a10
ka
I
5
1240
C-) a) (I) (44 a) ci)L
0)
a)160
80
0
0.5
1.0 1 1.5 Li)sec.
-2.0
FIGURE 17. COMPARISON OF CALCULATED AND
MEASURED ROLL SPECTRA FOR THE SHIP ALONE.
CONDITION
EXCLUSIVE
1.TANK.
I /CALCULATED
FROM RESPONSE
TO REGULAR
IWAVES
ai/313.28
SMEASURED
I I I / I I4OxSkk
ait.1372°
I I I I/
I/
I I-
II
I
I
I
I I I I I,'\.
\
I I2L0
e
cD
e
I
80
FIGURE 18.COMPARISON OF CALCULATED AND
MEASURED ROLL SPECTRA FOR THE SHIP WITH
TANK.
CONDITION
INCLUSIVE
1.TANK.
-.
S,CALCULATED
TO REGULAR
FR'i
S,J"1EASUR'
a1/3&84
ONSE
WA VS
°a1/366'
S ØCALCULAT a FsM
ICMPUTED AMPLIT DE
\
CHARACTERISTIC
/
656°
0.5 1.0 1.52.0
It)
sec
-
KtaSIN Ct
)
KtaCOSCt
U)
s/b =0
h/b=0.04
a°0333
0.50
0.25
s/b=0
h/b=0.04
(a0.10
a0.0667
7
0.25
0.50
0.75
1.00
1.00
0.75
1OLa
0.50
0.25
180
degrees
90
s/b=0
h/b =0.04
0.25
0.50
/
h
a0.10
a
=0.0333
075
1.00
0.25
0.50
0.75
1.00
2.0
1.6w
z
ml.2
x
cs.JD
0.8
01+7
0.75
1.00
w\/b/g
1.25
s/b=0.10
0.10h/b=0.08
h/b=0.06
h/b=0Q1.
'I
,=0.87m
(JJ0.75 W
=0.87m
V
DESIGN COND.
=1.30
6 m
0
0.25
0.50
1.6
01.
0
0.4
0.8
0.50
0.75
1.00
w\/b/g
1.25
s/b
= 0.10 0.10/
-h/b=0.08
h/b=0.0.
=,4
-'h/b
Co cDII CD (-)31
e-3
Liit.
c0
31TI
3
0
z
°E
ZD
(Ii"
wix
00
0.25
0
SHIP
TA N K
1'MEASURED
OSC.TEST
IWITHOUTCALCULATED
=0.06
9-I I I/
/
MEASURED
OSC.TEST
JTANK
SHIP
WITH
I
\v=o.o69
CALCULATED
\\
0
0.25
0.50
0.75
1.00
1.25
wVB/j
0.75
1.00
W\JB/g
1.25
I'I'
F
ISHIP
TANK
MEASURED
[OSC.TEST
WITHOUTCALCULATE
4w)-
IjI
/I
MEASURED
1SHIP
JTANK
v=vIw)
I
IOSC.TESTWITH
CALCULA
0
0.25
0.50
0.25
0.50
0.75
1.00
WB/g
1.25
EI'l
SHIP WITHOUT
TANK
3697
SCALCULATED
I IFROM OSC.TEST
I I II
I I / I IS1MEASURED
II
IlOxS
/
28
24
lOXSk
20
16
12
8
4
U0.25
0.50
0.75
1.00
W\[B/g
1.25
[c
SHIP WITHOUT
TANK
I JIA
SCALCULATED
FROM OSC.TEST
SMEASURED
I____
lOxSk_
0.75
1.1.00
wV-
1.25
SHIP WITH TANK
lOxSk
SPECTR.1
S,MEAS URED
,SCALCULATED
1"
FROM OSC.TEST
0
0.25
0.50
tLi