Professional Documents
Culture Documents
125-150
In the operation and design of high-speed ships, a greater emphasis is placed on good seakeeping
performance because it is found that large motions and high accelerations can significantly degrade
the operational capabilities. The need for better hull forms and the increased interest in seakeeping
performance call for more and better data to be available at the design stage to obtain a right balance between seakeeping and other, often conflicting, requirements. In this paper the genesis of
a systematic series of model experiments is given, illustrated with results. Attention is focused on
the general thoughts underlying the series, the selection of the characteristic section shape, the selection of the basic hull shape, the choice of the parameters to be varied in the series, and the parameters to be fixed from the outset. The choice of the parent hull form and the seakeeping aspects
associated with this choice are discussed, and the amalgamation of the data in the form of design
charts is shown.
125
Introduction
RECENT YEARShave seen an increased interest in seakceping
for naval ships. This has occurred in part because of operators
having become aware of significant and quantifiable differences between ships designed for the same task, and in part
because researchers have been able to come up with tools to
predict ship motions, accelerations and extreme effects and
have put together the framework required to estimate the
overall performance and its degradation as a result of environmental severity in statistical terms.
In the naval architectural field an inherent assumption is
often intuitively made when sufficient data are lacking, that
is, "small changes have small consecluence~" This may be true
in many fields of engineering; however, it can prove utterly
false in fluid dynamics. The first rule the hydrodynamicist has
to learn is that small changes to flow boundary geometry can
dramatically alter the flow field [1,2]. a Ship designers live up
to this rule in the design of the underwater hull form inasmuch
as they squabble over the lines in a degree unparallelled in any
other field of the design.
Seakeeping has long held the stigma of being hard to improve
upon and subject to statements like "ships do move anyway.
Yet now that the tools, experimental as well as computational,
have attained a state of maturity equaling those of other naval
architectural fields, it is possible to investigate the seakeeping
aspects of a ship at the design stage.
The increased interest in seakeeping for naval ships stems
from the full-scale observations, backed up by research findings,
that ship's behavior in seakeeping can indeed be improved upon
if one is willing to look into it, to handle the elusive tradeoff
between uneven and conflicting requirements, and is ultimately
willing to pay the price and accept the penalty.
The importance of seakeeping to a naval ship is of a diversifted nature. For the ship type we are concerned with in this
p a p e r - - t h e fast frigate--seakeeping pervades the operability
of almost any system and subsystem of the ship, [8-7].
Ship motions lead to extreme effects like slamming under the
bow or green water on deck and attending high loads that can
cause damage to the hull and the equipment topside. Excessive
a Numbers in brackets designate References at end of paper.
Nomenclature
aa ffi acceleration amplitude (single)
2~al/a ffi double significant acceleration
a19 acceleration at station 19
B = ship's breadth
bt ffi transom stern width
CA ffi incremental resistance coefficient
for model-ship correlation
Ca = block coefficient
Caa ffiblock coefficient aftbody
CB~ = block coefficient forebody
C~s = specific frictional resistance coefficient of ship according to ITTC57
CM = midship section coefficient
CG ffi center of gravity
Ct, -- horizontal prismatic coefficient
Ct,A = horizontal prismatic coefficient
aftbody
CpF -- horizontal prismatic coefficient
forebody
CaM = specific residuary resistance coefficient of model
--
126
Design c r i t e r i a
In setting up the Series it was important to adopt design criteria first. A systematic series hinges on the choice of the
characteristic hull shape--the parent hull--which can be selected only in the light of a set of design criteria. Furthermore,
for the scope of the series and its parameter selection and parameter space as .well as the subsequent test program details,
a set of design criteria would be an imperative departure point.
The criteria formulated for this hull-form series were:
good calm-water resistance properties, and
good seakeeping characteristics.
Concerning the first point, attention focused on resistance; for
the time being, propulsion was left out of the picture. The
second point, seakeeping, was elaborated to cover
low motion and acceleration levels,
low wave added resistance, and
small probability of incurring extreme effects like
slamming under the bow and shipping of water on
deck.
These criteria can all be narrowed down to one and the
same--low motion levels--because usually if the motions decrease, the accelerations and the wave added drag will also go
down. Extreme effects may not always follow suit, but can be
remedied easily by ensuring a sufficient draft forward and an
appreciable freeboard and flare on the forebody.
For the subject high-speed ships it was further considered that
the most critical motion and acceleration levels would occur
in head seas. Wave added resistance is also largest in that case.
It goes without saying that other wave headings may pose other
problems on the ship; for instance, rolling in beam seas and in
stern quartering seas. For the purpose of comparison of hull
forms, however, it was deemed sufficient to study the hulls only
in head seas. The restriction following from this design requirement has to be imposed on the series before the relevant
parameters are selected.
v / ~
127
Table 1
Cwp
Parametersto befixedandvaried
L/T
B/T
Main parameters
} -L/vl/3
CB
Cvp
LCB }
LCF
bt[B
iE
MODEL1
Subseries1
It follows as a logical consequence from the foregoing that
a limited number of parameters will have to be fixed from the
outset:
Cp
Cwe or Cvp
MODEL2
LCB
LCF
bt/B
iE
MODEL3
84
Fig. 1
128
20//
I
Body plans of models
SystematicSeriesHullForms
MODEL 1
MODEL 2.
Cwe and Cve. The resulting body plans of these three models,
Fig. 1, show that for the same displacement and in fact on the
same curve of sectional areas a gradual transition was-m~de:
from a narrow-forebody/wide-aftbody shape (Model 1) to a
wide-forebody/narrow-aftbody shape (Model 3). In order to
keep to realistic models it was felt that for this great a shift in
LCF, it was not entirely possible to keep Cwe the same. We
would have liked to, but the waterplane width forward would
have been too much and the waterline entrance angle would
have become too large for Model 3. Therefore, of necessity,
Cwe was reduced from Model 1 to Model 3 and Cve likewise
went up for the same CB; see Fig. 2. As a consequence, the
results to be found from these models, perhaps not so much on
resistance, but all the more so on seakeeping, cannot be attributed to LCF only. Yet the differences in seakeeping between
the three models, if only attributed to narrow-wide and widenarrow configurations, are unmistaken.
This was perhaps one of those points in which "scientific
rigor" (whatever that may be), or the wish to vary only one
parameter at a time, which is elegant when dealing with systematic series, had to give way to the one and only rule that
pervades this whole series, "the hull forms have to be realistic
and practical."
When the results on these three models became available and
Fig,- 2 Comparison of
;u;t[oOrnmSby Salient
~
~
.
.
! ~ V ~ -----'''~
were collated with other data, it was found that still more improvements could be made. Therefore the models were cut
up amidships and the fore and aft bodies combined in a different way. This resulted in models with a wider waterplane
area for the same displacement. Three models thus obtained,
denoted Models 4, 5, and 6, were also tested. Figure 8 gives
"narrow"
"medium"
"wide"
MODEL 1
MODEL 4
=+6
MODEL 5
@=
+13
]
"wide"
"medium" ~ u m "
a~
om~~
I.- IJJ
LL~
<:
MODEL 2
-6
-0
MODEL 6
:+6
"wide"
122
rr"
<
Z
MODEL 3
=
-14
"NARROW"
= -7
=-i
"WIDE"
129
I
I
0.5
CB
sis.////
0.4
~ 4 - - -i /-/ 4 - ~ -~ - ~ .t-"
z
//
//
...... ;~-,5
I
0.35-8"
.4 /
12/(J-
;~5
L/B
Fig. 4
L/B = 8
B/T = 4
CB = 0.4
The location of the parent hull and the width of the parameter
space were derived on the basis of the data files of MARIN, in
which all data of models tested at MARIN are stored. All
high-speed hulls in the data set were closely investigated and
plotted in diagrams like those shown in Figs. 5-8. These show
at a glance that the extent of the variation chosen for the series
is by no means extraordinary and that the particulars of the
parent hull are right in the middle of the data set.
It should be kept in mind that the data pertain to ships actually built; thus the variation in parameters would be quite
wide. L/B ranging from as low as 4 to a high as 12 should include anything from a short beamy patrol vessel to a slender
cruiser. The BIT ranges from 2.5 to 5.5; anything lower than
2.5 would face serious stability problems while 5.5 is so much
more than current coefficients that only patrol craft might
approach it. The range of block coefficients was chosen with
an eye to determining the limits of applicability of the highspeed surface ship, so it is somewhat lower than for current
surfaee combatants: CB = 0.35 to 0.50. For some time it had
been the idea to adopt a bottom figure for CB of 0.30, but it was
found that the characteristic hull shape could not be realistically
transformed so far down the CB scale without seriously distorting the section shapes or adopting hollow sections, or sections
with a hollow deadrise, for the greater part of the length.
Characteristic shape of parent hull
For the main particulars selected for the parent hull form a
characteristic shape had to be selected that would be optimum
from the aspect of calm-water resistance and seakeeping. The
body plans and test results of the models related to Figs. 5, 6,
and 7 were closely investigated and the characteristics of the
very best models compared. They all showed triangular Vshape sections forward and an almost flat ahbody.
Most of them were round bilge, some of them were hard130
MODEL DATA
MODEL DATA
15
1.0
10
nnffIo
0.5
..j
."
J~ o uo~oeo
.i;,~;;:..". _. i-.
Fig. 5
15
10
Correlation of BIT to L I V
Fig. 7
I/3
MODEL DATA
10
15
Correlation of Cs to L I V 113
MODEL DATA
15
/J
10
I
oe"
#.
...~m
S
f'#
.. :.,"
. c~
0
5
10
15
0.4
03
0.4
Q5
0.6
CB
Fig. 6
Correlation of U B to L I V 1/3
Fig. 8
Systematic Series Hull Forms
Correlation of CB to CM
131
Table 2
Designation
Length/breadth
Breadth/draft
Block coefficient:
total
forebody
aftbody
Midship section
coefficient
Waterplane coefficient:
total
forebody
aftbody
Horizontal prismatic
coefficient:
total
fore
aft
Vertical prismatic
coefficient:
total
fore
aft
Longitudinal center of
buoyancy in % L
Longitudinal center of
flotation in % L
Angle of waterline
entrance, deg
Breadth transom/breadth
Notation
Model 2
LIB
B/T
8
4
8
4
CB
CBF
CBA
0.396
0.327
0.465
0.396
0.327
0.465
0.396
0.327
0.465
0.396
0.327
0.465
0.396
0.327
0.465
0.396
0.327
0.465
CM
0.633
0.633
0.633
0.633
0.633
0.633
Cwp
CWPF
CWPA
0.785
0.577
0.992
0.768
0.588
0.947
0.749
0.600
0.898
0.790
0.588
0.992
0.796
0.600
0.992
0.774
0.600
0.947
Cp
CpF
CpA
0.626
0.517
0.735
0.626
0.517
0.735
0.626
0.517
' 0.735
0.626
0.517
0.735
0.626
0.517
0.735
0.626
0.517
0.735
Cvp
CvpF
0.505
0.567
0.469
0.516
0.556
0.491
0.528
0.545
0.518
0.501
0.556
0.469
0.497
0.545
0.469
0.512
0.545
0.491
CVpA
8
4
8
4
Model 5
Model 6
8
4
8
4
LCB
-4.97
-5.12
-5.16
-5.02
-5.11
-5.22
LCF
-9.23
-8.11
-6.77
-9.01
-8.68
-7.77
iE
bt/B
6.5
0.89
9.5
0.74
11.0
0.59
9.5
0.89
11.0
0.89
11.0
0.74
132
Model 1
Calm-water experiments
The resistance tests in calm water in the Deep Water Towing
Tank were carried out for the whole speed range Fn = 0.1 to
1.2 with a large number of measurement points to determine
the humps and hollows of the resistance curve with sufficient
accuracy The wetted surface of all models of Subseries 1 was
very nearly the same.
Measurements
Concerning the measurements, he~/ge was measured at t h e
CG using a potentiometer fitted on the air-lubricated cylinder;
pitch was measured with a potentiometer on one of the axes of
the universal joint in the model; vertical acceleration was
measured at 5station 19 on the forebody by means of an accelerometer; and the vertical relati~ee motion between water surface and hull was measured at station 17 using a resistance-type
5 By European custom the transom stern has been denoted station
0 and the fore perpendicular station 20.
wave probe. Themodel resistance was recorded in a straingage cell in the universal joint so that the resistance force
measured remained horizontal at all times.
For the correction of the phase angles to the wave crest it was
necessary to have a measure of the wave; this was taken 3.5 m
(11.5 ft) ahead of the model. As the very high speed would
render ordinary wave probes useless because of the wave system
of the wires themselves, a servo-controlled wave follower device
was used that could cope with the very high velocities and accelerations.
Measurements were all recorded on fiber optics recorder strip
charts for quick-look inspection and on magnetic tape for
proper analysis.
Results on seakeeping behavior
Figures 9-12 serve to illustrate the influence of model geometry on characteristic seakeeping transfer functions for one
.speed, Fn = 0.570. In Fig. 9 for heave the differences between
the models amount to.some 10 percent in the region of wavelength, which is of most interest to this kind of ship at such
speeds. It is not possible to attribute the differences to one
single cause, as many influence factors change.
The wave exciting force for heave is directly proportional
to the waterplane area, which was different for all six models.
The one with the largest waterplane area, Model 5 (Cwe =
0.796), would have the largest wave exciting force. The added
mass is to leading order proportional to beam squared, then to
B/T ratio, and finally to section shape. Model 5 would also
have the largest added mass, which would suppress the motions
in the high-frequency regime.
As to the damping, also proportional to beam squared, the
same can be said, and it may be expected that the model with
the largest Cwe (No. 5) would have the highest damping and
the lowest transfer function in the X/L = 1 region. The difference in spring rate is proportional to the waterplane area,
hence the Cwe, and in the long wave range the model with the
largest Cwe would have the lowest transfer function in heave.
The differences due to change in heave-pitch coupling can be
understood only through a direct computation, for instance with
strip theory. With the foregoing in mind the qualitative trend
of the data in Figs. 9-12 can be explained to a fair degree.
Model 5 has the largest Cwe and thus the greatest waterplane
area for the same displacement, resulting in very low heaving
for all speeds. Model 3 has the lowest Cwv and thus the
smallest waterplane area for the same displacement, resulting
in rather large heave motions for most speeds. Because the
block coefficient is kept constant, the heave motion exhibits the
same trend on the basis of (Cve) -1. The differences in heave
between the six models are largest around X/L = 1.2 for low
speed and around X/L = 2.4 for the highest speed, which
demonstrates that the greatest differences are to be' found
around the peak of the magnification factor Za/~a. The influence of Cvv on heave will be noted in Fig. 9; it is even more
pronounced for higher speeds where it is found that the models
with the lowest Cve (Models 1, 4, and 5) have the smallest
heave, whereas the other three with the highest Cve have the
greatest heave. Because of other effects such as LCB-LCF
separation, which greatly influences the heave-pitch coupling,
the heaving may not always be directly in line with the Cvv,
but a strong dependency certainly exists in this case. When we
relate Cve to hull shape we find that for this family of models
the great breadth of the transom stern, and as a direct result the
Width of the whole aftbody, is of direct influence on Cve, so that
we may qualitatively conclude that a wide aftbody results in
ahigh Cve, which in turn results in low heaving. This reasoning cannot be detached from the present family of models
in which the displacement and the block coefficient were kept
the same, and in which the CweF was not allowed to vary to the
same extent as Cwva because the waterline entrance angle had
to remain fine.
Influence of g e o m e t r y on pitch
133
Fn = 0 . 5 7 0
1.5
......
.....
-----.........
.....
MODEL
MODEL
MODEL
MODEL
MODEL
MODEL
1
2
3
4
5
6
Fn : 0 . 5 7 0
- ......
.....
-----........
----
MODEL
MODEL
MODEL
MODEL
MODEL
MODEL
1
2
3
4
5
6
/-,,
//'~,
1.0
0.5
25
X/L
Fig. 9
F'n :
~'/L
Fig. 11
0.570
- - ......
.....
-----........
.....
MODEL
MODEL
MODEL
MODEL
MODEL
MODEL
1
2
3
4
5
6
Fn = 0 . 5 7 0
......
. . . .
-----.........
.....
MODEL
MODEL
MODEL
MODEL
MODEL
MODEL
1
2
3
4
5
6
ul
0
0
X/L
Fig. 10
134
0
)-/L
Fig. 12
S y s t e m a t i c S e r i e s Hull F o r m s
Fn= 0.285
----
MODEL
.....
~
.....
m---..........
..... ~
MODEL
MODEL
MODEL
MODEL
MODEL
2
3
4
.5
6
.....
. . . .
-----...........
.....
Fn = O. 8 5 5
MODEL
MODEL
MODEL
MODEL
MODEL.
MODEL
1
2 "
3
4
5
6
1.5
1.5
1.0
1.0
....
--___.______--_--_.
tO
C)
0.5
0.5
/,(J
0
0
2
X/L
X/L
Fig. 13
---......
Fn :, 0 . 5 7 0
Fig. 15
.....
.
.........
......
MODEL
MODEL
MODE L
MODEL
MODEL
MODEL
1
2
3
.....
.....
-----..........
.....
Fn = 1 . 1 4 0
4
5
6
1.5
MODEL
MODEL
MODE L
MODEL
MODEL
MODE L
1
2
3
4
5
6
1,
1.0
t~
(!)
(D
.///~,
/ zZl//"
O.
0 . ,~
)'/L
Fig. 14
2
X/L
Fig. 16
MODEL
- - - -
Fn =,0.285
.......
F'n = O. 5 7 0
......
Fn = 0 . 5 7 0
. . . .
Fn = 0 . 8 5 5
MODEL
, Fn = 0.855
Fn = 1.1 4 0
= 0.285
------Fn
= 1.140
1.5
1.0
\\\\
f0
,/'//!,,
0.5
.,
/
0
0.5
Fig. 17
1.5
1.0
----Fn
MODEL
......
0.5
Fig. 19
1.0
Fn = 0.285
= 0.285
i
Fn = 0.570
......
Fn = 0.855
. . . .
Fn=
1.5
MODEL
......
Fn = 0.570
.....
Fn = 0 . 8 5 5
------Fn
1.140
= 1.140
300
200
- - .
/$I-.,, '\
/'-~ \
\
/
\,\\ ',,
r0
100
!! A
/7/ Y
\
X
\
0
0
0
0.5
1.0
1.5
0.5
1.0
1.5
L,/-c-5:
Fig. 18
Fig. 20
137
used for the transformation of the ship section to the unit circle.
The results are also presented in Figs. 21-28.
Figures 29 and 30 show the results of computations and form
the counterpart of Figs. 14 and 16.
For the same irregular seas as used for the experiments, calculations have been performed to determine the dimensionless
values of the significant motions, accelerations and added resistance. Some of these results are shown in Figs. 33 to :36, again
for Model 5 but for Fn = 0.570 only, including results based on
w = a { ~ + n=0
"~" a2n+a~-(2n+ 1)}
(1) Versions i and 2 and the close-fit transformation.
It is generally stated that the following assumptions restrict
where ~ represents a point on the unit circle and w represents application of the strip theory:
the corresponding point on the ship section (w and ~ are com1. The ship form should be slender with gradual change
plex coordinates).
of this form in the longitudinal direction.
For n = 1 the so-called Lewis transformation is obtained
2. The frequency of motion should not be too low or too
which in general provides sufficient accuracy for most ship
high.
sections. This Lewis transformation determines the ship section
:3. The forward speed of the ship should be low.
by the breadth-draft ratio and the sectional area coefficient.
In the past, special experiments have been performed to
A procedure of this transformation for shiplike cross-sections gather more information about the aforementioned limits.
has been presented by Tasai [22] and Grim [2:3]. A more ac- With respect to the first assumption, reference is made to [24],
curate description of the ship section may be obtained by the in which investigations are described related to the ship's
"close-fit" transformation, for which case in equation (1), n > slenderness. The surprising result was that even for L / B = 4
1.
the calculated responses in head waves show good agreement
After determination of the hydrodynamic coefficients for with the measured model experiments.
each section the values for the whole ship are found by inteIn the present work, very high forward ship speeds are congration over the ship's length according to the method presented sidered in order to investigate also the limits mentioned in the
in [24,25].
third assumption. It may be concluded that even for the
For different frequencies or wavelengths and the speeds highest speed, Fn = 1.140, the agreement between experiment
considered, the response functions have been calculated for the and calculation is satisfactory as far as the heaving and pitching
vertical motions, accelerations, and added resistance in waves. motions and the vertical accelerations are concerned.
These response functions have been determined for two versions
Up to Fn = 0.570, this agreement is also reasonable for the
related to the speed influence as mentioned in [24] and [25]:
relative motions and added resistance in waves. Above this
Version 1, sometimes called the ordinary strip method speed, there is a considerable lack of agreement with respect
(OSM), leads to a set of motion equations which lack some of to these phenomena. This might be due to the significant inthe symmetry relations in the mass coupling coefficients and fluence of the ship's own wave profile and dynamic swell-up,
some additional terms in other hydrodynamic coefficients. For which is appreciable at these speeds. The influence of these
this version 1 the speed influence is taken into account only for phenomena has not been taken into account for the calculations,
the derivative of the sectional added mass with respect to the as yet further investigations in this respect are required.
ship length. In such a way, only terms are introduced with
Another conclusion which may be derived from the comparison between experiments and calculations is that for this
din'
V-ship form, hardly any improvement can be achieved with a
dx
close-fit transformation of the ship section. Moreover, it has
Version 2 also takes into account the speed influence re- been shown that the speed influence according to Version 2
lated to the derivative of the damping coefficient with respect yields less reliable results especially for the highest speeds in
to the ship length, and so terms are introduced with
spite of the presence of the required symmetry relations in the
hydrodynamic coupling coefficients.
dm'
dN'
V ~ and V d---xC h o i c e of p a r e n t h u l l f o r m
In this way the mentioned symmetry relations are present in
the hydrodynamic coupling coefficients and so the presentation
In choosing the parent hull form, the work on Subseries 1,
of the hydrodynamic coefficients agrees with [26].
incorporating LCF variation and C w v variation, was brought
The added resistance in waves has been calculated according to a conclusion and the work on the true series as shown in Fig.
to the method of Gerritsma and Beukelman as developed in 4 could start in earnest. The most important aspects of this
[27}. Their method is based on the relation between the ra- choice are related in the following.
diated damping energy as calculated by the strip method and
the added resistance in waves accounting for the relative motion Calm-water resistance aspects
of the ship with respect to the water surface.
The resistance measured in calm water was split up into a
The response functions of the motions, accelerations and frictional resistance part, calculated according to the ITTC-57
added "resistance in waves have been calculated for the two extrapolation method, and a residuary part. Because all six
mentioned versions by the computer program "TRIAL" of the models had virtually the same wetted surface--their disShip Hydromechanics Laboratory of the Delft University of placement was the same--the comparison of the six models
Technology. As an example, the calculated values of the ver- could just as well be done on the basis of residuary resistance.
tical absolute motions, accelerations and added resistance in The values thus obtained were normalized on the basis of the
waves are shown in Figs. 21-28 for Model 5 at two speeds: Fn residuary resistance of Model 2, which was arbitrarily set at 100
= 0.570 and Fn = 1.140.
percent. The percentages are given in Table :3. It is shown
These response functions have also been calculated by a Delft that Model 5--the parent to be--is not the best model on recomputer program based on a close-fit transformation with n sistance. No. 6 would do better at high speed. But model No.
= 9 in equation (1); this means that nine coefficients have been 5 holds an attractive position:
138
MODEL
---. . . .
.....
Fn = 0 . 5 7 0
VERSION 1
"VERSION 2
C L O S E FIT
MARIN EXR
MODEL.
VERSION
------
VERSION
.....
CLOSE
MARIN
Fn = 0 . 5 7 0
1.5
75
1.0
5C
ll}
FIT
EXR
0.5
25
1.5
Fig. 23
0
Fig. 21
0.5
MODEL
5 .
1.0
__-....
Fn = 0 . 5 7 0
J
r
0.5
1,0
15
MODEL
VERSION 1
VERSION 2
CLOSE
FIT
MARIN EXR
Fn = 0 . 5 7 0
VERSION 1
VERSION 2
C L O S E FIT
MARIN EXP
-----.....
o
1.5
1.0
1u
oY
0.5
,\
.J/
0
0
Fig. 22
0.5
1.5
Fig. 24
1.0
0.5
1.0
1.5
S y s t e m a t i c Series Hull F o r m s
139
.oo~, ~
Fn = 1.140
v~#s..o~ ~
.oo~. ~
CLOSE FIT
MARIN EXP
Fn = 1,140
VE#S',oN 2'
~-~
/b \
CLOSE FIT
MARIN EXP.
1.5
/
o/./
"
,.o
ii
/1
o P/ ~~!
\i
"
o
0
Fig. 25
0.5
1.0
1.5
MODEL
Fn = 1,140
....
.....
o
0
Fig. 27
0.5
1.0
1.5
VERSION 1
VERSION 2
CLOSE FIT
MARIN EXP.
MODEL
- -----.....
o
Fn = 1.140
VERSION 1
VERSION 2
CLOSE FIT
MARIN EXP.
1.5
f-,
I\
\
1.0
o? /
\
io~
ool
~n
~ - ~ ~ - - ~ ~.o Oo \
C
c
N~\ i
Q5
0
0
Fig. 26
140
/o
0.5
1.0
1.5
Fig. 28
0.5
1.0
1.5
- ......
.....
-----........
. . . .
Fn = 0.570
MODEL
MODEL
MODEL
MODEL
MODEL
MODEL
1
2
3
4
5
6
Fn = 0.570
300
1.0
200
(D
._1
0.5
~Q
IN
100
1
2
3
4
5
6
"h\
/
/
0
El{2
)'/L
Computed
MODEl_
MODEl_
MODEl_
MODEl_
MODE[_
MODE[_
Fig; 29
------
.........
.....
1.5
tl
- ......
.....
pitch
transfer function
models
comparison
- ......
.....
-----...........
. . . .
1
2
3
4
5
6
F n = 1.140
MODEL
MODEL
MODEL
MODEL
MODEL
MODEL
for
Fig. 31
six
300
1.0
200
F n = 1.140
1.5
......
.....
-----.........
.....
MODEL. 1
MODEL. 2
MODEL 3
MODEL. 4
MODEL. 5
MODEL. 6
~:~:.~-~
/" /~.
tO
0
_J
/I/!/
0.5
.,{/:,y
IO0
Fig. 30
C
1
0
Computed
X/L
3
comparison
for six
0
Fig.
L/~2
32
S y s t e m a t i c S e r i e s Hull F o r m s
7"
J/
MODEL
~f.
,r - /
Fn = 0.570
- . . . .
.....
.......
VERSION 1
VERSION 2
C L O S E FIT
MARIN EXR
MODEL
--VERSION
1
....
VERSION 2
.....
C L O S E FIT
.......
M A R I N EXR
Fn= 0.570
1.!
1.0
it,..A
~__.._~
iN
0.5
0
0
Fig. 33
Fig. 35
MODEL
Fn = 0.570
~
. . . .
.....
.......
VERSION 1
VERSION 2
CLOSE FIT
MARIN EXR
MODEL
Fn = 0 . 5 7 0
- - -----.....
......
VERSION 1
VERSION 2
C L O S E FIT
MARIN EXR
300
200
|~,.n
-J
0'1
=O
IO0
S''/' /
ss"
~
0
0
Fig. 34
142
LI{
Flg, 36
S y s t e m a t i c Series Hull F o r m s
\
L
MODEL 5
MODEL 5
0.08
RISE C.G
/
I
iI
/i
i
0.06
//
14
/'
/
,/
s
0.04
s
/
/s
/
/
0.02
/
/
i"
/
s
TRIM
ANGLE
/
ii !
.~s Si /
~s
0
0
FnV=
Fig. 37
2
Vs x 0.514447
FnV=
2
Vs x 0 . 5 1 4 4 4 7
Fig. 38
number of hull shapes. In this model a great number of seakeeping parameters such as heave, pitch and acceleration are
amalgamated to arrive at one single figure of merit indicative
of the overall seakeeping performance of a hull.
The validation of this model is also shown in [28] and stems
from back-designing a hull shape which has seakeeping characteristics superior to any ship in the data set. This was accomplished with the predictor model that uses only the ship's
main particulars and associated quantities.
At first, the predictor model was tried on our Models 1, 2 and
3 which, as shownin Fig. 3, constitute the diag, onal, incorporating a shift in LCF. Earlier the work of Bales and Cummins
[29] had indicated a sizable improvement of heave and pitch
if the forebody CweF were to be increased. The improvement
found from the experiments was not as great as had been expected, perhaps because CweF had not been increased enough.
In the series, CWeFwas not taken any higher than 0.6 as it was
felt that this would result in an important calm-water resistance
increase of more than 10 percent at Froude numbers in excess
of 0.5. The restriction on Cwev is not made in current frigate
Table 3
Speed,
F.v
1.0
1.5
2.0
2.5
3.0
3.5
95.2
97.6
100.3
101.2
104.1
103.5
100
100
100
100
100
100
Model No.
3
4
100.0
91.0
93.8
93.2
89.9
89.8
120.8
96.0
99.2
99.5
100.0
101.1
96.2
91.6
89.4
94.9
98.0
98.0
120.6
88.4
91.1
90.9
88.7
90.4
143
/1
./.-"
Table 4
Calm-Water Resistance
Model
NO. '
1
2
3
4
5
6
Seakeeping (~haracteristics
CFS + CA
103
CRM
103
CTS
X 103
RTS
(tonne)
1.428
1.428
1.428
1.428
1.428
1.428
1.168
1.651
1.553
1.593
1.511
1.470
3.046
3.079
2.981
3.021
2.939
2.898
88.39
88.44
83.87
88.75
86.85
84.44
1
2
3
4
5
6
+0.1
0
+5.2
-0.4
+1.8
+4.5
Z
(m)
rms
0.36
0.39
0.39
0.39
0.36
0.37
+7.7
0
0
+2.6
+7.7
+5.2
Fn = 0.855
Fnv = 2.49
Sea state 4 : ~ w l / 3 = 2.50 m (8.20 ft)
7~1 = 7.00 s
For this nominal .point the calm-water resistance, the wave
added resistance, and the total resistance were determined as
well as the seakeeping determinants: heave, pitch, relative
motion at station 17 and vertical acceleration at station 19.
The values are given in Table 4 for all six models, in the upper
figures. In order to weigh the different quantities and to select
the best model, the central model (No. 2) was arbitrarily set to
144
0
()
rms
0.66
0.66
0.73
0.61
0.59
0.63
+7.6
0
-10.6
+7.6
+10.6
+4.5
817
(m)
rms
1.66
1.77
1.94
1.56
1.69
1.63
+6.2
0"
-9.6
+11.9'
+4.5
+7.9
a19
(g)
rms
0.26
0.27
0.29
0.26
0.25
0.27
+3.7
0
-7.4
+3.7
+7.4
0
RT
RAW
(tonne)
(tonne)
6.77
6.84
7.77
6.77
7.11
7.50
96.15
95.28
91.64
95.52
93.96
91.94
+1.0
0
-13.6
+1.0
-3.9
-9.6
-i.0
0
+4.0
+0
+1.4
+3.6
0
0
-1
+6
+13
+6
145
/
Discussion
preliminary design of high-speed displacement-type hull forms.
Although today there is in the literature model series data It provides an important contribution to an aspect of the design
which cover the resistance qualities of high-speed hulls of the of ships which has not been covered by experiments.
The design criteria chosen have a significant influence on the
type described by the authors, little, if any, systematic seakeeping information exists for these forms and this paper is results. The design criteria adopted in this paper are good
therefore most welcome. The authors should be congratulated calm-water resistance and good seakeeping characteristics.
on their thorough approach in testing a subseries of models in The application of the criteria led the authors to limit the
variation of the forebody waterplane coefficient to be no greater
search of a parent from which the series as a whole derives.
The choice of a parent model is very important and not easy, than 0.6 to limit expected increases in calm-water resistance
and one could of course argue indefinitely over its selection. for higher Froude numbers (in excess of 0.50). Previous studies
The authors have rightly balanced resistance and seakeeping have shown that increase in the waterplane coefficient above
qualities, but in considering the head sea case alone perhaps too 0.6 can lead to improved seakeeping performance. Table 5
limiting a view has been taken. Looking at the afterbody with this discussion compares other proposed seakeeping design
sections of the chosen parent I am not convinced that its slight procedures.
Research sponsored by the Society is currently in progress
immersion at rest is to be preferred in beam or following sea
to
develop a preliminary seakeeping design procedure for
conditions and it will be interesting to hear the authors' views.
merchant
ships. The design criterion used is minimization of
Also from resistance considerations increased transom immersion ratio, allowing as it does a gentler slope in the buttock freight rate through minimum initial cost and high average
lines, can show to advantage and I am not surprised to see Model sustained sea speed. One aspect considered in this work is seNos. 3 and 6 superior to the parent. It is also possible that the lection of proportions and hull form for subcritical tuning.
very flat and relatively wide transom will encourage slamming Subcrffical tuning refers to the ratio of ship resonant response
at the stern so that a few tests in following seas Would have been period to wave encounter periods being less than 0.75. This
reduces wave excitation and pitch response. Critical tuning
instructive.
As to the experiments themselves, although it appears per- refers to equal ship and wave encounter periods and is a region
fectly justifiable to use a towed model in regular waves for the of maximum motion. Supercritical tuning refers to the ratio
purposes of choosing a parent, tests on the complete series of ship to encounter periods being greater than 1.2. Reduced
should preferably be done with p~opelled models. The in- motions are expected in this region and lower pitch than in the
clusion of propeller action gives more realistic conditions par- subcritical region is possible. These effects of tuning on reticularly if the model is allowed to surge and also for faster craft sponse are an example of the "Doppler Shift" referred to by the
such as these where running trim influences performance. authors.
The results shown in Fig. 18 of the paper can be interpreted
Propeller action modifies running trim from that obtained in
towed models and this effect should be fully represented in the in terms of tuning. In head seas the encounter period is
experiments. Evidence as to how model behavior in waves
Tw
TE =
differs in propelled or towed tests is given in reference [80] ,
[1 + (kL)I/zFn]
(additional references follow some discussions). Here significantly different acceleration levels are shown over a range of and the tuning factor for pitch is
encounter frequency. There are also differences in pitch and
To = To[1 + (kL)I/2F,]
heave at resonant frequencies. Although these data relate to
Te
Tw
a planing craft at Fnv = 3.0, I suspect they would be reflected.
where
at lower Froude numbers as well.
High-speed displacement ships of this kind experience severe
To = pitch resonant period
motions which have in extreme cases led to structural damage
Tw = wave period
of the hull, and, since this is an important design consideration,
T = period of encounter
I am surprised that no plans appear to have been made to obtain
The model heave and pitch periods are not given in the paper
pressure and slamming measurements at the bow. Also, bridge
structures are vulnerable wit h respect to loss of visibility from and it would be useful if the authors could provide these data.
flying spray which is generated from the breakup of solid water I have estimated that in model scale the pitch period ranges
shipped over the bow. A technique for monitoring and mea- from around 0.95 sec at zero speed to less than 0.80 sec at the
suring impact and wetness on a bridge front is given by Lloyd higher Froude numbers. Applying this result to Fig. 18, the
following results for tuning were obtained:
[81].
This paper is of great interest and I hope that the authors will
FROUDE NO.
TUNING
be encouraged to publish the results of the whole model series.
0.285
Subcritical, To/T~ < 0.75 for all x/L/)~
If they can also be persuaded to add seakeeping data for wave
0.570
Critical, To/TE > 0.75 for v/L/)k > 1.0
directions other than head seas, then their findings will be very
0.855
Critical for 0.70 < v/L/)k < 1.00
keenly awaited.
Supercritical v/L/~ > 1.00
1.140
Supercritical for x/L/k > 0.75
Additional references
D. Bailey, 6 Visitor
Table 5
Proposed
Commercial Seakeeping
Design Procedure
Proportions
or Form.
Length
Beam
B/T
T
Cwp
CB
Frigates
Schmitke and Murdey (1980) a
increase to reduce
wave excitation and
thus reduce motion
increase to increase
damping
reduce to reduce
pitch period
reduce to reduce
pitch period
choose on basis of
arrangement and powering
increase to reduce
pitch period
has negligible effect
on seakeeping
.
choose on basis of
stability
...
Destroyers
Bales (1980), [28]
decrease T/L
.
.
...
decrease T/L to
improve
seakeeping
increase to improve
seakeeping
...
...
decrease (CvP)F
increase (CvP)a
a Schmitke, R. T. and Murdey, D. C., "Seakeeping and Resistance Trade Offs in Frigate Hull Design," 13th
ONR Symposium on Naval Hydrodynamics, Tokyo, Oct. 1980.
. [The views expressed herein are the opinions of the discusser and not
necessarily those of the Department of Defense or the Department of
the Navy.]
During recent years seakeeping considerations for naval ships
have received increased emphasis and have been incorporated
into the ship design cycle at an early stage. At the same time
energy considerations in terms of reduced powering requirements have also become increasingly important. It was shown
in reference [32] that good seakeeping and powering characteristics can be combined in one design and I am sure that the
information in this paper and results from the planned follow-on tests will aid designers to meet both the seakeeping and
powering goals simultaneously.
The paper selects a parent hull form for a new series and it
is important to compare resistance with that of other hull forms
of its type. Averaged residuary resistance to weight ratios for
other series of hull forms are presented in reference [88] and
the following table shows a comparison to the selected parent
hull No. 5.
7 David W. Taylor Naval Ship Research and Development Center,
Bethesda, Md.
RR/A
Hull model 5 (parent for prol~gsed series) 0.041
DTNSRDC Series 64
0.088
NPL Series L/B = 7.5
0.044
The above comparison is at the sarue nondimensional volume
Froude number Fnv = 2.49 and thesame length-to-displacement ratio L/AI/Z = 8.52 for all three hulls. The comparison
indicates that the smooth-water performance of Hull 5 is
comparable to that of the designs in the other two series.
The authors have concentrated on the seakeeping aspects and
have selected their parent hull form, No. 5, principally on its
low pitching characteristic and acceptable resistance in waves.
However, if some of the data in Table 4 are rearranged (see
Table 6 6f this discussion) in order of increasing resistance,
relative to that of Model 3, some other trends become evident.
The general trend is that excessive transom width, results in
poor resistance in calm water and in waves and that a wide
forebody is beneficial. The selection of Model 5 as parent was
based principally on the excellent pitching in wave characteristics as opposed to any resistance differences. Note that except
for the insignificant difference in the calm-water resistance of
Models 2, 4, and 1, the models with higher calm-water resistance tend to have higher resistance in waves. This similarity
in the ranking of the resistance in waves and the resistance in
calm water supports the validity of traditional naval architectural practice of basing ship resistance comparisons on calmwater data.
The authors state that "Because all six models had virtually
the same wetted surface--their displacement was the s a m e - the comparison of the six models could just as well be done on
the basis of residuary resistance." I have shown in Table 6 that
the wetted surface differs by as much as 5.0 percent for Models
8 and 5. The wetted surfaces that I have shown were derived
from the total ship resistance coefficient, CTS, and the total ship
resistance, RT:s, given in Table 4 of the paper. If one makes
a resistance comparison on residuary resistance as the authors
suggest, then one would assume that Model 5 is a significantly
better performer than Model 8 at a volume Froude number of
147
Table 6
Model
No.
Resistance
Ratio
Calm
Water
3
6
5
4
1
1.000
1..007
1.036
1.054
1.058
1.054
Resistance
Ratio
Waves
1.000
1.003
1.025
1.040.
1.040
1.049
Wetted
Surface
Ratio
1.000
1.035
1.050
1.021
1.044
1.044
[The views expressed herein are the opinions of the discusser and not
necessarily those of the Department of Transportation or the U.S. Coast
Guard.]
The authors are to be congratulated for offering a paper
which is relatively easy to follow on such a complex yet impertant subject. Seakeeping is perhaps the most important and
least understood subject of ship design today since ships are
expected to maintain a certain sea speed for operational or
economic purposes; courts are increasingly critical in responsibility for heavy weather damage; and the question of survivability of the entire ship in a seaway is matter of interest to
administrations deciding international standards.
International standards for commercial ships are largely those
in the International Convention on Load Lines (1966), establishing standards for intact reserve buoyancy; sheer; superstructure buoyancy in rolling action; bow height; clearance for
air pipes, ventilators, hatch coamings, etc.; strength of hull,
hatches, superstructure, etc.--all of which are affected by the
seakeeping ability of the ship. The International Maritime
Organization has recently authorized a long-term systematic
review of the convention with a,view to maintaining its applicability to present-day ships. Thus it is fortunate that the authors propose a system to investigate the seakeeping aspects of
a ship at the design stage.
My specific questions are submitted with full awareness that
the authors have used a series of ship.forms which are usually
148
Forebody
wide
wide
wide
medium
medium "
narrow
After
Body
Pitching
Ratio
narrow
medium
wide
medium
wide
wide
1.0
0.86
0.81
0.90
0.84
0.90
Authors' Closure
The discussions have brought forth a number of significant
points which require further amplification. We will respond
to each discussion separately.
We appreciate the comment of Mr. Bailey drawing our attention to the implications that the limited scope of experiments
may have had. We have limited ourselves to head seas, because
it was reasoned that this would constitute the most demanding
case. For the vertical motions and 'accelerations and associated
extreme effects like slamming on the bow and shipment of
water over the bow, this is indeed the most: critical wave
heading for the present ship type. Yet different wave headings
may pose different problems like rolling in beam seas, steering--i[ not broaching--in stern seas. Each of these could give
rise to an independent investigation and would constitute a
daunting subject to pursue. One has to restrict oneself.
As to the slight transom stern immersion at rest, it gave us the
best tradeoff between resistance and seakeeping although
viewed from other angles of interest an improvement may still
be possible. We do not think it will do much to rolling in beam
seas; the width of the transom stern would, but not the immersion. However, broaching of a hull with a wide and full
-stern with little, if any, submerged lateral skeg area i~ likely to
be more of a problem than on ships having a narrow stern and
more in the way of skeg. In considering the transom stern we
have to bear in mind that on a true design one would in most
,instances apply a stern wedge to improve the calm-water resistance. No such wedges were tried in the present study although they are high on the list of priorities for future work.
We would disagree with the discusser about tile risk of stern
slamming on such a wide and flat transom stern. It has been
observed in other experiments that on a frigate-type ship lying
hove-to in a rough sea, in particular when lying stern to the seas;
slamming under the stern does indeed occur. This is a rather
uncommon thing in the operational profile of a frigate, and it
has also been observed that just a few knots of forward speed
reduces dramatically the relative motions at the stern and the
associated encounter frequency, and slamming ceases to
occur.
149
150