100% (10) 100% found this document useful (10 votes) 4K views 332 pages Principles of Naval Architecture Vol 2 (Sname)
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content,
claim it here .
Available Formats
Download as PDF or read online on Scribd
Go to previous items Go to next items
Save Principles of Naval Architecture Vol 2 (Sname) For Later nd Revision
SecoAcknowledgments
‘The authors of Chapters V and VI, J.D. van Manen and P. van Oossanen, wish
to acknowledge their indebtedness to the author of Chapter V in the preceding
edition, Frederick H. Todd. Extensive use has been made of the original text and
Rgures, The authors also wish to recognize the assistance provided by U. Nienhuis
iP tne Maritime Institute Netherlands in working through the entire text a second
time, making additions and corrections whenever necessary. And valuable ideas
cad suggestions regarding high-speed displacement and planing hulls in Section
9 of Chapter V were provided by ‘Daniel Savitsky, Director of the Davidson. Lab-
oratory and are acknowledged with thanks.
"The author of Chapter VII, William 8. Vorus, expresses his appreciation of the
pioneering work of Frank M. Lewis, as distilled in Chapter X of the preceding
Piition of this book, which provided a foundation for the new chapter. He appre”
ciates the reveiw and comments on early drafts by Edward F, Noonan, of NFK
Engineering Associates, Inc., and John P, Breslin of Stevens Institute ‘of Technology.
Bie Control Committee provided essential guidance, as well as valuable assist-
ance in the early stages. Members are:
Jobn J. Nachtsheim, Chairman
‘Thomas M. Buermenn
William A. Cleary, Jr.
Richard B. Couch
Jerome L. Goldman
Jacques B. Hadier
Ronald K. Kiss
Donald P. Roseman
Stanley G. Stiansen
Charles Zeien
Finally, the Editor wishes to thank all of the authors for their fine work: and
for thelr full cooperation in making suggested revisions. He acknowledges the
indispensible efforts of Trevor LewisJones in doing detailed editing and preparing
text and figures in proper format for publication.
‘May 1988, E. V. Lewis
Editornd Revision
I
SecoOGG IIGEsJ. D. van Manen
P. van Oossanen
CHAPTER V
Resistance
Section 1
Introduction
1.1 The Problem. A ship differs from any other
large engineering structure in that-—in addition to all
its other functions—it must be designed to move ef-
ficiencly through the water with a minimun of external
assistance. In Chapters LITT of Vol, It has been shown
how the naval architect can ensure adequate buoyancy
y for a ship, even if damaged by collision,
grounding, or other eause. In Chapter IV the problem
of providing adequate structure for the support of the
ship and its contents, both in calm water and rough
seas, was discussed.
In'this chapter we are concerned with how to make
it possible for a structure displacing up_to 500,000
tonnes or more to move efficiently across any of the
world’s oceans in both good and bad weather, The
problem of moving the ship involves the proportions
and shape—or form—of the hull, the size and type of
propulsion plant to provide motive power, and the de-
vice or system to transform the power inta effective
thrust, The design of power plants is beyond the scope
of this' book (see Marine Engineering, by RL, Har-
Hngton, Ed, SNAME 1971). The nine sections of this
chapter wil deal in some detail with the relationship
between hull form and resistance to forward motion
(or drag). Chapter VI discusses propulsion devices and
their interaction with flow around the hull.
‘Phe task of the naval architect is to ensure that,
within the limits of other design requirements, the bull
form and propulsion arrangement will be the most
efficient in the hydrodynamic sense. The ultimate test
is thet the ship shall perform at the required speed
with the minimum of shaft power, and the problem is
to attain the best combination of low resistance.and
high propulsive efficiency. In general this can only be
attained by a proper matching of hull and propeller.
‘Another factor that influences the hydrodynamic de-
sign of a ship is the need to ensure not only good
* Complete references are listed at end of chaptar.
smooth-water performance but also that under aver
age service conditions at sea the ship shall not suffer
from excessive motions, wetness of decks, or lose more
speed than necessary in bad weather. ‘The assumption
‘that a hull form that is optimum in calm water will
also be optimum in rough seas is not neeessarily valid,
Recent rosearch progress in oceanography and the
seakeeping qualities of ships has made it possible to
predict the relative perfonmance of designs of varying
hull proportions and form under different realistic sea
conditions, using both mode! test and computing tech-
niques. The problem of ship motions, attainable speed
and added power requirements in waves are discussed
in Chepter Vil, Vol. IIL This ehapter is concerned
essentially with designing for good smooth-water per-
formance.
“Another consideration in powering is the effect of
deterioration. in hull surface condition in service as the
result of fouting and corrosion and of propeller rough-
ress on resistance and propulsion. This subject is dis-
cussed in this chapter
Asin the case of stebility, subdivision, and structure,
criteria are needed in design for determining accept-
able levels of powering. In general, the basie eontrac-
tual obligation laid on the shipbuilder is that the;ship
shall achieve a certain speed with a specified power in
good weather on trial, and for this reason smooth-
water performance is of great importance. As previ:
ously noted, good sea performance, particularly the
maintenance of sea speed, is often 2 more important
requirement, but one that is much more difficult to
define. The effect of sea condition is customarily al
lowed for by the provision of a service power margin
‘above the power required in smooth water, an allow-
‘ance which depends on the type of ship and the average
weather on the sea routes on which the ship is designed
to operate. The determination of this service allowance
depends on the accumulation of seaperformance data
on similar ships in similar trades. Powering eriteria in
the form of conventional service allowances for both2 PRINCIPLES OF NAVAL ARCHITECTURE
sea conditions and surface deterioration are considered
in this chapter.
“The problem of controlling and maneuvering the
ship will be covered in Chapter IX, Vol. IIT.
1.2 Types of Resistance, The resistance of a ship
av given speed is the force required to tow the ship
at that speed in smooth water, assuming no interfer-
ence from the towing ship. If the hull has no appen-
dages, this is ealled the bare-hull resistance. The power
necessary to overcome this resistance is called the tow-
rope or effective power and is given by
Po= RV (ta)
where Py = effective power in kWatt (kW)
Ry.= total resistance in kNewton (kN)
speed in m/sec
or ehp = RrV./826 (2)
where chp = effective power in English horsepower
Ry = total resistance in Ib
‘Y= speed in knots
‘To convert from horsepower to S.L. units there is
only a slight difference between English and metrie
horsepower:
hp (Bnglish) x 0.746 = kW
hp (metric) x 0.785 = KW
Speod in knots x 0.5144 = m/see
‘This total resistance is made up of a number of
diferent components, which are caused by 2 variety
of factors and which interact one with the other in an
extremely complicated way. In order to deal with the
question more simply, itis usual to consider the total
elm water resistance as heing made up of four main
components,
(a) The frictional resistance, due to the motion of
the hull through 2 viscous uid. :
(2) The wavemaking resistance, due to the energy
that must be supplied continzously by the ship to the
waye system created on the surface of the water.
(c) Hddy resistance, due to the energy carried away
by eddies shed from’ the hull or appendages. Local
‘eddying will occur behind appendages such as boss-
ings, shafts and shaft struts, and from stern frames
‘and rudders if these items are not properly streamlined
and aligned with the flow. Also, if the after end of the
ship is too blunt, the water may be unable to follow
the curvature and will break away from the hull, again
giving rise to eddies and separation resistance.
(q) Air resistance experienced by the above-water
part of the main hull and the superstructares due to
the motion of the ship through the ai.
‘The resistances under (t) and (2) are commonly
taken together under the name residuary resistance.
Further analysis of the resistance has led to zhe iden
tification of other subcomponents, as discussed sub-
sequently.
‘The importance of the different components depends
upon the particular conditions of design, and much
of the skill of naval architeets lies in their ability to
choose the shape and proportions of hull which will
result in a combination leading to the minimum total
power, compatible with other design eonstraints.
Tn this task, knowledge derived from resistanee and
propulsion tests on small-scale models in a model basin
or towing tank will be used. The details of such tests,
‘and the way the results are applied to the ship will be
described in a later section. Much of our knowledge
chip resistance has been learned from such tests,
and it is virtually impossible to discuss the various
types of ship resistance without reference to model
work,
1.3 Submerged Bodies. A streamlined body moving.
in a straight horizontal line at constant speed, deeply
jinmersed in an unlimited ocean, presents the simplest
case of resistance. Since there is no free surface, there
fs no wave formation and therefore no wavemaking
resistance. If in addition the fluid is assumed to be
without viscosity (a “perfect” fluid), there will be no
frictional or eddymaking resistance. The pressure éis-
tribution around such a body ean be determined the-
oretically from considerations of the potential ow and
has the general characteristics shown in Fig, 1(a)
‘Near the nose, the pressure is increased above the
hydrostatic pressure, along the middle of the body the
pressure is decreased below it and at the stern it is,
‘again inereased, The velocity distribution past the hull,
by Bernoulli's Lave, wil be the inverse of the pressure»
tribution—along the midportion it will be greater
than the syeed of advance V and in the region of bow
and stern it will be less.
‘Since the fluid has been assumed to be without vis-
cosity, the pressure forces will everywhere be normal
to the hull (Fig. 1(0)). Over the forward part of the
hhull, these will have components acting towards the
stern and therefore resisting the motion. Over the
after part, the reverse is the ease, end these compo-
nents are assisting the motion. Tt cen be shown that
the resultant total forees on the fore and after bodies
are equal, and the body therefore experiences no re-
sistance.’
Tn a real fiuid the boundary layer alters the virtual
shape and length of the stern, the pressure distribution
there is changed and its forward componentiis reduced.
‘The pressure distribution over the forward pertion is,
but little changed from that in a perfect fluid. There
fs therefore a net force on the body acting against the
motion, giving rise to a resistance which is variously
referred to as form drag or viscous pressure drag.
in a res! fluid, too, the body frictional
resistance and perhaps eddy resistance also. ‘The fluid
immediately in contact with the surface of the body is
“Th
1144, and is Known a8 Alember:’s parador.
first noted by the French mathematician d’Alembert in
DectRESISTANCE
Sane counan eave
POTENTIAL. FLW.
SS SSS
WERE ig =
ERINATION FONT
(a)
ig, 1 Examples of flow obact 9 aonerped body
ied clong with the surface, and that in the close
rrcinity is set in motion in the same direction as thet
in which the Body is moving. This results in a layer of
water, which gets gradually thieker from the bow to
the stera, and in which the velocity varies. from that
of the body at its surface to’that appropriate to the
potential flow pattem (almost zero for a slender body)
at the outer edge of the layer (Fig. 1(c)). This layer 3
called the boundary layer, and the momentum supplied
to the water in it by the hull is a measure of the
frictional resistance. Since the body leaves behind it a
frictional wake moving in the same direction as the
body (whieh ean be detected far astern) and is contin-
ually entering undistarbed water end accelerating ic
to maintain the boundary layer, this represents a con
tinual drain of energy. Indeed, in wind-tunnel work
the measurement of the velocities of the fluid behind
a streamlined model is a common means of measuring
the frictional drag,
TE the body is rather blunt at the after end t
may loave the form at some point—called a separation
point—thus reducing the total pressure on the. af
terbody and adding to the resistance. This separation,
resistance is evidenced by 2 pattern of eddies which
is a drain of energy (Fg, Ud).
1.4 Surface Ships. A ship moving on the surface of
the’ sea experiences frictional resistance and eddy-
making, separation, and viscous pressure drag in the
same way as does the submerge body. However, the
pifesenie of the free surface adds a further component.
‘The movement of the hull through the-water erestes
a pressure distribution similar to that around the sub-
merged body; i¢,, areas of increased pressure at bow
and stern and of decreased pressure over the middle
part of the length,
But there are important differences in the pressure
distribution over the hull of a surface ship because of
the surface wave disturbance ereated by the ship’s
forward motion. There 1a yreater pressure acting
over the bow, 5 indicated by the usually prominent
bow wave buildup, and the pressure increase at the
stern, in and just below the free surface, is always
Jess than around a submerged body. The resulting
added resistance corresponds to the drain of energy
into the wave system, which spreads out astern of the
ship and has to be continuously recreated. (See Section
4.8), Henee, it has been called wave-making resistance.
‘The result of the interference of the wave systems
originating at bow, shoulders (if any) and stern is to
produce a series of divergent waves spreading out-
wards from the ship ata relatively sharp angle to the
centerline and a series of transverse waves along the
hhull on each side and behind in the wake (Fig. 7).
‘The presence of the wave systems modifies the skin
friction and other resistances, and there is a very com-
plicated interaction among all the different campo-
nents.
Section 2
Dimensional Analysis
2.1 General. Dimensional analysis is essentially a
means of utilizing a partial knowledge of a problem
when the details are.too obscure te permit an exact
analysis. See Taylor, E. S. (1974). Tt has the enormous
advantage ef requiring for its application a imowledge
only of the variables whieh govern the result. To apply.
it to the Zow around ships and the corresponding re-
sistance, if is necessary to know only upon what var~
iables the latter depends. This makes it a powerful
‘tool, because the correctness of a dimensional solution,
does not depend upon the soundness of detailed anal-
yyses, but only upon the choice of the basic variables.
Dimensional solutions do not yield numerical answers,
but they provide the form of the answer so that every4 PRINCIPLES OF NAVAL ARCHITECTURE
experiment can be used to the fullest advantage in
determining # general empirieal solution.
2.2 Dimensional Homogeneity. Dimensional snal-
ysis rests on the basic principle that every equation
which expresses a physical relationship must be di-
mensionally homogeneous. There are three basie quan-
tities in meehanies—mass, lengthand time—which are
represented by the symbols M,Z, and T. Other quan-
tities, such as force, density, and pressure, have di-
mensions made up from these three basic ones.
‘Velocity is found by dividing a length or distance
bya time, and sohas the dimensions L/T: Acceleration,
whieh is the change in velocity in a certain time, thus
has dimensions of (L/T)7, or L/T*.
‘Force, which isthe product of mass and acceleration,
has dimansions of Mx L/T? or ML/T*.
"AS simple case to illustrate the principle of di-
mensional analysis, suppose we wish to dotermine an
oxpression for the time of swing of a simple pendulum.
If Tis the period of such a pendulum in vacuo (80
that there is no frictional damping), it could depend
‘upon eertain physical quantities such as the mass of
the pendulum bob, m, the length of the eord, 2, (sup-
posed to be weightless) and the are of swing, s. The
foree which operates to restore the pendulum to its
original position when itis disturbed is its weight, mg,
and 80 the acceleration due to gravity, g, must be
involved in the problem.
‘We can write this in symbols as
T= f(r, 13,9)
where fis a symbol meaning “ls some finetion of.”
If we assume that this function takes the form of
a power law, then
To mi Pg
If this equation is to fulfil the principle of dimen-
sional homogeneity, then the dimensions on each side
must be the same. Since the lefthand side has the
dimension of time only, s0 must the righthand side.
Writing the variables in terms of the fundamental
linits, we have
Ts WLI LT
Equating the exponents of each unit from each side
of the equation, we have
btotd
°
°
—2d =1
Hence
d= -12
a=0
bre
‘The expression for the period of oscillation 7 seconds
is therefore
i
T = constant x Da+ x st X grt
= constant x Vi7g x (6/0)
‘The solution indicates that the period does not de-
pend on the mass of the bob, but only on the length,
the acceleration due to gravity, and the ratio of length
of are to length of pendulum. ‘The principle of dimen-
sions does not supply the constant of proportionality,
which must be determined experimentally.
‘Phe term (¢/f) is a mere number, each quantity being
of dimension L, and dimensionally there is no restric-
tion on the value of ¢. We can therefore write
‘T = constant x Ve x f(s/) @
Although the form of the function fis undetermined,
it is explicitly indicated by this equation that it is not
the are ¢ itself which is important, but its ratio to /
ie. the maximum angle of swing, s/Z radiar
‘The function fean be found by experiment, and must
approach the value unity for small swings, so 2s to
lead to the usual formula for a simple pendulum under
‘such conditions:
P= constant x 79
‘The most important question regarding any dimen-
sional solution is whether or not physical reasoning
has led to 2 proper selection of the variables which
govern the result,
“Applying dimensional analysis to the ship resistance
problem, the resistance could depend upon the fol-
lowing:
(a) Speed, V.
@) Size of body, which may be rerresented by the
linear dimension, L.-
(©) Mass density of fluid, p (mass per unit volume)
(2 Viscosity of fuid, »
(@) Acceleration due to gravity, 9
() Pressure per unit area in fluid, p
Itis assumed that the resistance & can now be writ:
ten in terms of unknown powers of these variables:
@)
Since B is a force, or a product of mess times ec-
celeration, its dimensions are ML/T?
‘The density p is expressed as mass per unit volume,
or A/T
in a viseous fluid in motion the force between ad:
jacent layers depends upon the area A in contact, the
‘oefcient of viscosity of the liquid and upon the rate
at which one layer of @uid is moving relative to the
ext one. If w is the velocity at a distance y from the
boundary of the fluid, this rate or velocity gradient is
given by the expression du/du.
‘The total foree is “hus
F = pAdu/dy
du/dy boing a velocity divided by 2 distance has di
mensions of (L/TWL, or 1/T, and the dimensional
equation becomes
Ra pVLp oy!RESISTAN
ML/T =
or
n= MAT
pis a force per unit area, and its dimensions are
(ML/TIVLA, oF M/LT™
‘The ratio /p is called the kinematic viscosity of the
Jiquid, v, and has dimensions given by
pip = W/ET)MLIM) = LIT
Introducing these dimensional quantities into Bqua-
tion (8), we have
ML/T* = (U/L (L/T) (Ly (MALTY
x (L/T?S (MALT'Y (A)
whence
or
oe Clee
and
Hea bt dots
48-90 ~3f- 24d
2
‘Then from Equation (3)
zener) OF] ©
‘All three expressions within the brackets are non-
dimensional, and are similar in this respect to the s/2
term in Equation (2), There is therefore no restriction
dimensionally on the exponents ¢, ¢, and f The form
of the function f must be found by experiment, and
may be different for each of the three terms.
‘Writing v for je/p and remembering that for similar
shapes the wotted surface § is proportional to ZL’,
Equation (5) may be written
©
where the left-hand side of the equation is a not
dimensional resistance coefficient. Generally in thi
chapter 2 will be given in kN and p in kg/L (or t/m,
although N and kg/m? are often used (as here) in the
eases of model resistance and ship air/wind resistance.
A term first suggested by Dr. EV, Telfer.
Equation (6) states in effect that if all the parameters
on the righthand side have the same vaiues for o¥o
geometrically similar but different sized bodies, the
fow patterns will be similar and the value of
2/4 pSV" will be the same for each
28° Corresponding Speeds. Equation (6) showed
how the total resistance of 2 ship depends upon the
various physieal quantities involved, and that these are
associated in three groups, VL/v, gL/V? and p/pV*
Considering first the case of a nonviscous liquid in
which there is no fietional or other viscous drag, and
neglecting for the moment the last group, there is left
the parameter gL/V* controlling the surface wave sys-
tem, which depends on gravity. Writing the wavemak-
ing ‘or residuary resistance as Rp and the cor-
responding coefficient as Ce, Cp can be expressed as
AAV gL) mM
G= tn
‘This means that geasimas* (geometrically similar bod-
ies) of different sizes will have the same speeific ro-
siduary resistance coefficient Cy if they are moving at
the same value of the parameter V'/gL.
‘According to Froude's Law of Comparison': "The
(cesiduary) resistance of geometrically similar ships is
Jn the ratio of the cube of their linear dimensions if
their speeds are in the ratio of the square roots of
their linear dimensions.” Such speeds he called cor-
responding speeds.* Tt wil be noted that these cor
responding speeds require V//L to be the same for
model end ship, whieh is the same condition as ox-
pressed in Dquation (1), The ratio Vx/JZ, commonly
‘with V, in knots and Z in feet, is called the speed-
length ratio, This ratio is often used in presenting
resistance data because of the ease of evaluating it
arithmetical, but it has the drawback of not being
nondimensional. The value of V/Jgb, on the other
hhand, is nondimensional and has the same numerical
value in any consistent system of units. Because of
Froude’s clase association with the coneept of speed-
length ratio, the parameter V/GL is called the Froude
number, with the symbol Fn.
Whea Y, is expressed in kmots, Z in feet, and g in
ft/sec, the valation between V/A/Z and Froude number
Fa = 0.298 YATE
or
V,AfE = 3.355Fn
“Stated in 1858 by Willa Proude (1955) who drut recognized
the practical necessity of separating the total resistance inta com
‘onents, sed on the genera lav of mechanical cimlituda, rom
Sbservalnns of the wave patarne of modals of the seme form but
of different szr.6 PRINCIPLES OF NAVAL ARCHITECTURE
‘The residuary resistances of ship (Res) and of model
(Rex) from Equation (7) will be in the ratio
where subscripts sand wrefer to ship and model, re-
spectively.
TE both model and ship are run in water of the same
density and at the same value of V*/gL, as required
by Equation (7, ie.
(Mobs = Vill
then Cy will be the Same for each, and
Res/ Ray, = Ss (Vs}/Su (Vu? = Lega? L/L
= (Lys) = Ald B)
where A; and Ay are the displacements of ship and
model, respectively.
"This is in agreement with Froude's law of compar-
ison,
Tt should be noted from Equation (8) that at eorre-
sponding speeds, i.e, at the same value of V/ YL
Ras/As = Ray /Ay o
i.e, the residuary resistance per unit of displacement,
is the same for model and ship, Taylor made use of
this in presenting his contours of residuary resistance
in terms of pounds resistance per long ton of displace-
ment (Section 86).
1f the linear scale ratio of ship to model is A, then
the following relations hold:
Lgfly =>
Ve/Vy = lig hfig = R= 10)
Rgs/Ray = (Ls) Mba)? = As/ Ay =
‘The “corresponding speed” for 2 small model is much
iower than that of the parent ship. In the ease of a 5
‘m model of a 125 m ship (linear scale ratio d= 25),
the medel speed corresponding to 25 knots for the ship
is 25/K", or 25/5, or 5 knots, This is a singularly
fortanate circumstance, since it enables ship models
to be built to reasonable seales and ran at speeds which:
are easily attainable in the basin.
Returning to Equation (6), consider the lest term,
n/pV* If the atmospheric pressure above the water
Surface is ignored and p refers only to the water head,
then for corresponding points in model and ship p will
vary directly with the linear seale ratio). At corre:
sponding speeds ¥* varies with 4 in the same way so
that p/pV” will be the same ior model and ship. Sineo
This cua law had previously boon put forward by the French
Naval Constructor Reseh in 1852, Dut be had not pursued it or
strated how it could be applied tothe practical problem of
predicting ship resistance (Reesh, 1852}.
the atmospherie pressure is usually the same in model
and ship, when itis included in p, 30 thet the latter is
the total pressure at a given point, the value of
p/p¥" will be mach greater for model than for shij
Fortunately, most of the hydrodynamic forces arise
from differences in local pressures, and these are pro-
portional to V%, so that the forees are not affected by
the atmospheric pressure so long as the fluid remains
in contact with the model and ship surfaces. When the
pressure approaches very low values, however, the
‘water is unable to follow surfaces where there is some
curvature and cavities form in the water, giving rise
to cavitation. The similarity conditions are then no
longer fulfilled. Since the absolute or total pressure is
greater in the model than in the ship, the former gives
no warning of such behavior. For tests in which this
danger is known to be present, special facilities have
bean devised, such as varlablepressure water tunnels,
channels or towing basins, where the correctly sealed
down total pressure can be attained at the same time
that the Froude condition is met
‘In the ease of a deeply submerged body, where there
is no wavemaking, the first term in Equation (6) gov-
ferns the frictional resistance, Rp The Zrietional re-
sistance cooflicient is then
R
ipSv®
Cr = AVL) ay
and Cr wil be the same for model and ship provided
that the parameter VL/v is the same. This follows
essentially from the work of Osborne Reynolds (1889),
for which reason the product VL/v is known a8 Rey-
nolds number, with the symbol Ra,
‘If both model and ship are run in water at the same
density and temperatare, so that v has the same value,
it follows from (11) that, VeLs = Vir Ly. This condition
js quite different trom the requirement for wave-male
ing resistance similarity. As the model is made smaller,
the speed of test must increase. In the case slready
‘used as an illustration, the S-m model of a 125m, 25-
not ship would have to be run at a speed of 625 kn
‘The conditions of mechanical similitude for both tric-
tion and wavemaking cannot be satisfied in a single
test. Tt might be possible to overcome this difficulty
by running the model in some other fluid than water,
50 that the change in value of v would take account
of the differences in the VL product, In the foregoing
example, in order to run the model at the correct wave-
making corresponding speed, and yet keep the value
of VLA: the same for both model and ship, a fluid
would have to be found for use with the model which
had a kinematic viscosity coefficient only 2/125 that of
water. No such fiuid is known, In wind-tunnel work,
similitude can be attained by using compressed ai: in
the model tests, so decreasing v and increasing VL/v
to the required value.
"The practical method of overcoming this fundaraen-
tal difficulty in the use of ship models is to deal withRESISTANCE
the frictional and the wave-making resistances sepa-
rately, by writing
Cy + Cp 2)
This is equivalent to expressing Equation (6) in the
form
&,
Co gin = AVG) + HAVE) (08)
Froude recognized this necessity, and so made ship-
model testing & practical tool, He realized that the
frictional and residuary resistances do not obey the
same law, although he was unaware of the relationship
expressed by Equation (11).
2.4 Extension of Mode! Results te Ship. ‘To extend
the model results to the ship, Froude proposed the
following method, which fs based on Equation (12)
Since the method is fundamental to the use of models
for predicting ship resistance, it must be stated at
length:
Fronde noted:
tg) Tho model is made to lines ele aio of
and run over_a range of “corresponding” speeds suc!
that Ve/yLs = Ve / WL
(@) The total model resistance is measured, equal to
"%6, The frictional resistance of the model Rr is eal
stance to be the same as
goth flat plank of the same length and
ace as the mate!
(@) Tho residuary resistance of the model Ryy is
found by subtraction:
ey — Rew
(e) The residuary resistance of the ship Reg is eal-
culated by the Jaw of comparison, Equation (10)
Ras = Ray X 0°
‘Vhis applies to the ship at the corresponding speed
given by the expression
Vg = Vy x 0%
() The frictional resistance of the ship Rrs is cal-
culated on the same sssnmotion as in footnote (4),
using a fictional cooficiont appropriate to the ship
length.
@ The total ship resistance (smooth hull) sis then
given by
Rg = Bos + Res
‘This principle of extrapolation from model to ship is
still used in all towing tanks, with certain refinements
to be discussed subsequently.
‘Bach component of reristance will now be dealt with
in greater detail
Section 3
Frictional Resistance
3.1 General One has only to look down from the
deel of a ship at sea and observe the turbulent motion
in the water near the hull, increasing in extent from
hhow to stem, to realize that energy is being ebsorbed
in frictional resistance. Experiments have shown that
even in smooth, new chips it accounts for 80 to 85
percent of the total resistance in slow-speed ships and
fs mach as 50 percent in high-speed ships. Any roagh-
ness of the suriaco will increase the frictional resist~
ance appreciably over that of a smooth surface, and
swith subsequent corrosion and fouling still greater
increases will occur. Not only does the nature of the
surface affect the drag, but the wake and propulsive
performance are also changed. Frictional resistance is
thus the largest single component of the total resist-
ance of a ship, and this accounts for the theoretical
and experimental research that has been devoted to it
over the years. The ealeulstion of wetted surface area
which is required for the caleulation of the frictional
resistance, Equation (11), is discussed in Chapter I
3.2 Froude's Experiments onFriction, Froude, know-
ing the law governing residuary resistance and having
concluded that the model-ship extrapolation problem
could only be solved by dividing the resistance into
‘two components, undertook a basie investigation into
the frictional resistance of smooth planks in his tank
at Torquay, England, the results of which he gave to
the British Association (Froude, W., 1872, 1874).
"The planks varied in lengths from 0.61 m @ #) to
15.2 m (60 i) and the speed range éovered was from
0.8 m/sec (1.61 fps) to 4.1 m/sec (18.3 fps), the miax-
imum for the 15.2 m plank being 8.8 m/sec (10.8 £p3).
Froude found that at any given speed the specifie re-
sistance per unit of surface area was less for.a long
plank than for a shorter one, which he attributed. to
the fact that towards the after end of the long plank
the water had acquired a forward motfon end so had
a lower relative velocity.
He gave an empirical formula for the resistance in
the form
R= sv a
where Hl
B= resistance, KN of Ib
‘S = total area of surface, m* or ft*
V = speed, m/sec or ft/sec8 PRINCIPLES OF NAVAL ARCHITECTURE
Toble 1—froude's Skin-Fiction Coefficients?
Ra fS-V¥ R= resistance, Ib S = area of plank, sq ft speed, fps
‘Length of surface, or distance from cutwater, £
2 8 20 0
fon ok f o2 fon ok fon ok
.00d10 2,00 0,00890 0.00460 1.83 0.00874 0.00300 1.85 0.90897 0.00870 0.00885
Cioodes 198 00044 0.00860 1.94, 0.00300 0.00318 198 0.0028 —
Calico... Grand 193 0.00830 0.00750 1.92" 0.00600 0.00580 1.89 0.00870 o.0ge4n 6.00570
Fine sand aoe) 200 0.00600 0.00880 2100 0.00450 0.00480 200 0.90884 0.00400 9.00880
Medium sind 6.00900 2.90 0.00780 0.00630 ¥ 6.00880 200 0.00450 @.00490 200 0.00460
Goarse sand......, 0.01000 2.00 0.00889 0.00710
0.00890 2.00 0.00490 =
“Ww, Fronde’s resulls for planks in fresh water at Torquay (British Association 1872 and 1874)
Note: The values of ie represent the valu
at the head of the column,
_fand n depended upon length and nature of surface,
and are given in Table 1.
‘For the smooth varnished surface, the value of the
exponent n decreased from 2.0 for the short plank to
1588 for the 15.2:m (60 #t) plank. For the planks rough
‘ened by sand, the exponent had e constant value of
20.
For a given type of surface, the fvalue decreased
swith increasing length, and for 2 given length it in-
creased with surface roughness.
Tn ordor to apply the results to ships, the derived
skin-friction coefficients had to be extrapolated to much
greater lengths and speeds. W. Froude did-not give
these extrapolated figures in his reports, but sug-
ested two methods which might be used for their
derivation, In his own words, “It is at onee seen tha:,
fat a length of 80 feet, the decrease, with increasing
Tength, of the friction per square foot of every addi-
tional length is so small that it will make no very great
difforonce in our estimate of the total resistance of a
‘surface 900 ft long whether we assume such decreas
to continue at the Same rate throughout the last 250
fect of the surface, or to cease entirely after 50 feet;
‘while it is perfectly certain that the truth rust lie
Somewhere between these assumptions.” Payne,
(1936) has reproduced the curve Froude used at Tor-
quay in 1876 for ships up to 152.4 m (600 ft) i length.
‘This curve is almost an arithmetic mean between those
which would be obtained by the two methods sug-
ested. W. Froude (1874) also obtained some fullseale
Information in on attempt to confirm his law of com-
parison and to assist in the extrapolation of the frie-
Table 2—Results of Towing Triels on HM'S Greyhound
Speed V, formes seen ‘0 800 1000 1200
Beseaee iby fram ship. "111" 3100 5400 8800 19100
Barb prediied from model......, 2900 4300 8750 17500
Pétoont dierence si as an a
(B79 10 for ship 3 #6 084 0.99 138
GEA) X torsmodel peedidion’.. et 070 ose 122
ilerence in (R/V C10 a nnensss 0.92 014 0.12 0.2
6
0.
aay Ot To Last equare foot of e surface whose lengih is equal to thet given
tional coefficients to ship lengths by carrying out
tovring tests on the sloop HMS Greykownd, a wooden
ship 62.58 m (172 ft 6 in.) in length, with copper
sheathing over the bottom. The results of the towing
tests and the predictions made from the model are
given in Table 2.
‘The actual ship resistaneo was everywhere higher
than that predicted from the model, the percentage
inerease becoming less with inerensing speed. The dit
ference in R/V?, however, is almost the same et all
speeds, except the lowest, and decreases only slowly
wrth increasing speed, as might occur if this additional
Tosistance were of viscous type and varying at some
wer ‘ess than the second. Froude pointed out that
The additional resistance could be accounted.for by
Eesuming that the eopper-sheathed hull was equivalent
fo smooth varnish over 2/8 of the wetted surface and
to ealico over the rest, This he considered reasonable,
and the tivo resistance curves were then almost ider
eal, which he took as a visible demonstration of the
correctness of his law of comparison.
Tn his paper on the Greyhound trials, Froude states
quite clearly how he applied his idea of the “equivalent
lank" resistance: “Foe this calculation the immersed
Tin was carefully measured, and the resistance due
to if determined tipon the hypothesis that it is equiv-
Hlont to that of a rectangular surface of equal area,
and of length (in the line of motioa) equal to that of
the model, moving at the same speed.”
‘the 1876 values of frictional coefficients were stated
to apply to new, clean, freshly painted steel surfaces,
but they lie considerably above those now generally
accepted for smooth surfaces. The original curves have
‘een modified and extended from time to time by R.E.
Froude, ap to a length of 966 m (1200 ft), but these
extended curves had no experimental basis beyond the
15.2 m (50 £4) plank tests made im 1872, (Froude, R. E.
11826). Novertheless, they are still used today in some
iowing tanks.
3.3 Two-dimensional Fricional Resistance Formula~
tions, In the experiments referred to in Section 2.3,
Osborne Reynolds made water flow through @ glassRESISTANCE
tube, introducing a thin stream of dye on the centerline
at the entrance to the tube. When the velocity was
all, che dye remaiaed as a straight filament parallel
tp cbs axis of the tabe. At a certain velocity, which
Reynolds called the critieal velocity V., the filament
began to waver, became sinuous and finally lost all
definiteness of outline, the dye filing the whole tube.
‘The resistance experienced by the fluid over @ given
iongth of pipe was measured by finding the lose of
pressure head, Various diameters of the tube, D, were
tised, and the kinematic viscosity was varied by heating
the water. Reynolds found that the laws of resistance
exuetly corresponded for velocities in the ratio v/D,
fand when the results were plotted logarithmically
YV, = 2000v/D
Below the critical velocity the resistance to flow in
the pipe varied directly as the speed, while for higher
velocities it varied at a power of the speed somewhat
Jess than 2
“When the foregoing relationship is written in the
form
V.Div = 2000
the resemblance to Equation (11) is obviow
‘Stanton, et al. (1952) showed that Reynolds’ findings
applied to both water and air flowing in pipes, and also
that the resistance coefficients for models of an airship
on diffecent scales were sensibly the same at the same
°
value of VE/v. Baker (1925) plotted the results of much
Of the available data oa planks in the form of the
resistance coefficient
ff
Om
to a base of VE/v, and found that a mean curve could
be drawn passing closely through Froude’s results ex:
cept at low values of VL/v.
Experiments such as those performed by Reynolds
suggested that there were two separate flow regimes
possible, cach associated with a different resistance
Taw. At low values of VD/v, when the dye filament
retained its own identity, the fluid was evidently flow.
ing in layers which did not mix transversely but slid
‘over one another atzelative speeds which varied across
the pipe section. Such ‘iow was called laminar and was
associated with a relatively low resistance, When the
Reynolds number VD/v increased, either by increasing
YD or by decreasing v, the laminar flow broke down,
the fluid mixed transversely in eddying motion, and
the resistance increased. This flow is called turbulent.
‘In modern skin-friction formulations the specific fric-
tional resistance coeficient Cis assumed to be 2 func:
tion of the Reynolds number Rn or VE/v. As early a8
1904 Blasius had noted that at low Reynolds numbers
the flow pattern in the boundary layer of a panic was
laminar (Blssius, 1908). He succeeded in caleulating
0009,
coy
pation. VOM Karena ,=0072
masts
cpsiser(
‘a0ce|
&
0003) re Sg ae
Curent
aod
1
oat, Ls 5
ng
cauinad ow
vou pol ee
- ca ow oF o
a
Foc? shin etna nen, ebolent ond foinar fom10 PRINCIPLES OF NAVAL ARCHITECTURE
the total resistance of a plank in laminar flow by in-
tegrating across the boundary layer to find the mo-
mentum transferred to the water, and gave the
formula for Cy in laminar flow in terms of Rn:
Reo ee
pave 7 LO (rLey as)
C=
‘This lin is plotted in Fig. 2. Blasius found good agree-
ment between his calculated resistances and direct ex-
periment, but found that the laminar flow became
unstable at Reynolds numbers of the order of 4.5.
10%, beyoud which the resistance coefficients increased
rapidly above those calculated from his equation.
Prandél and von Karman (1921) separately published
the equation
= 007K VLIvy"* (16)
for turbulent flow, which is also shown in Fig. 2. This
equation was based on an analytical and experimental
investigation of the characteristies of the boundary
layer, as well as on the available measurements of
overall plank resistance, principally those of Froude
and further experiments run by Gebers in the Vienna
tank (Gebers, 1919).
‘At low values of Reynolds mumber, and with quiet
water, the resistanee of a smooth plank elosely Zollows
the Blasius line, the flow being laminar, and from
Equation i its seen thatthe resistance varies 2s
For turbulent flow, the value of the resistance coef-
ficient is considerably higher than for laminar flow,
and varies as a higher power of the speed; according
to Equation (16) a5
‘The transition from laminar flow to turbulent fiow
docs not cecur simultaneously over the whole plank.
‘Transition begins when the Reynolds number reaches
a.critical value R.. As the velocity Vincreases beyond
this value, the transition point moves forward so that
the local value of the Reynolds nuraber, Ve/v, remains
equal to the critical value, « being the distance of the
transition point from the leading edge of the plank.
‘This is called the “local Reynolds number,” and for
the constant value of this local Rn at which transition
takes place, x will decrease as V increases, and more
and more of the plank surface will be in turbulent flow
and so experience a higher resistance. The value of C
will thus increase along a transition line of the type
shown in Fig. 2, and finally approach the turbulent line
asymptotiealiy. It should be noted that there is no
unique transition line, the actual one followed in a
given ease depending upon the initial state of turbu-
fence in the fiuid, the character of the plank surface,
the shape of the leading edge, and the aspect ratio,
‘These transition fines for smaoth planks oceur at
values of Reynolds number within the range over
which most plank-friction tests have been run, and if
such plank results are to be used to predict the values
of Cpat Reynolds numbers appropriste to a ship—100
times or so larger than the highest plank values—only
those results for fully turbulent flow can properly be
used,
3.4 Development of Frictional Resistance Formulations
in the United States, With the completion of the Ex-
perimental Modal Basin (EMB) in Washington in 1900,
new experiments were made oa planks and new model
coefficients were derived from these tests. For the ship
coeflcients, those published by Tideman (1876) were
adopted, These did not representany new experiments,
being simply a re-analysis of Froude’s results by a
Duteh naval constructor. This combination of friction
coeficients—EMB plank results for model, Tideman’s
coefficients for ship—was in use at EMB from 1901 to
1928 (Taylor, D. W., 1943).
‘By this time the dependence of frictional resistance
on Reynolds number was well established, and a for-
mulation was desired which was in accord with known
physical laws, In 1923, therefore, EMIB changed to the
use of frictional coefficients given by Gebers for both
the model and ship range of Reynolds namber (Gebers,
1919). This practice continued at that establishment
and at the new David Taylor Model Basin (DTMB) until
1947 (aow DTRC, David Taylor Research Center)
Schoenherr (1922) collected most of the results of
plank tests then available, and plotted them as ordi-
nates of C,-to a base of Rn as is shown in Fig. 3. He
included the results of experiments on 6,1 m (20 ft)
and 9.1 m (80 ft) planks towed at Washington, and at
the lower Reynolds numlers some original work on
1.8 m (6 ft) eatamarans with artificially induced tur-
ibulent flow. At the higher Reynolds numbers he was
guided largely by the results given by Kempf (1923)
for smooth varnished plates. Kempf’s measurements
were made on small plates inserted at intervals along
476.8 m (282 ft) pontoon, towed in the Hamburg tank,
‘The local specific resistances so measured were inte:
grated by Schoenherr te obtain the total resistance for
surfaces of different lengths. In order to present these
data in conformity with rational physieal principles,
Schoenherr examined his results in the light of the
theoretical formuls of Prandtl and von Karman, which
was of the form
AISEp = logy (Rn Cx) + At
Ho found he could get a good fit to the experimental
data by making M zero.and A oqual to 0.242, so arriving
at the well-known Schoenkerr formulation
0.242/JC = logy (Rm Cr)
‘The Schoenherr coefficients as extended bs
ula to the ship range «! Reynolds numbers apply to
a perfect'y smooth bull surface. For actual ship hulls
with structural roughnesses such as plate seams,
welds ot rivets, and paint roughness, some allowance,
ac.RESISTANCE
E osume pester
Soy piewssoies 6g
tt
i
Frown rcutegPRINCIPLES OF NAVAL ARCHITECTURE
”
4008 PTT TT TPT
; ors
coos irre. une r= 0028 4
N 9288 -L06,5 (Rn XCF)
coop i =
Ys HUGHES LINE Cpp2——2-28®
a. Se 0-4
-— owanute cro eB + $2
0 S
QL ape une
oe E SK
004
arent
9003} = ta
wre. une
2.00% a =
ach wuswes une? T
o ao foi satin pa pos be
108 Sx108 108 ‘Sx10® 107 ‘Sx107 10% ‘x10! 10? ‘sxto? 10"
Fa Shin bn os
the magnitude of which is discussed later, is necessary 0.000418 + goes) sve (at
EB + $2 if
tp give 2 realictie prediction.
2.5 The Work of the Towing Tenk Conferences. The
International Conference of Ship Tank Supexintend-
ents (ICSTS) was a European organization founded in
41982 to provide a mecting place for towing-tank staffs
‘9 discuss problems peculiar to their field, In 1085, the
ICSTS agreed to adopt the Froude method of model
extrapolation, among the decisions recorded being the
following:
sg Pn the determination of ength and wetted sur-
ice:
(@)° For every kind of vessel, the length on the
waterline shoul be ued
(®) The mean girth multiplied by the length is
adopted as the wetted surface.
Vi—Froude’s method of ealeutation:”
(a). ‘The Committee adheres to the skin friction de-
duced from Froude’s 0 values," and takes these to be
represented by the formula below, since this gives the
same values of friction for model and ship within the
limits of experimental errors:
Epa is, no “obliquity” correction
‘hese were the Froude fricsonalcoeffsients prs
ular notation cee Froude (888)
where
‘Rp = resistance in kNewton;
LL’ = length in meters;
'S. = wetted surface in square meters;
Vie = speed in knots,
(6) All model results should be corrected to a stan-
dard temperature of 15 deg C (= 59 deg F) by &
correction of —0.43 pereent of the frictional resistance
per + 1 deg C or ~0.24 percent per + 1 deg F."
‘In 1946 the American Towing Tank Conference
(ATTO) began considering the establishment of a uni-
form practice for the calewlation of skin friction and
the expansion of mode} data to full size, In 1947 the
following two resolutions were adopted (SNAME,
1948):
"1, Analysis of model tests will be based on the
Schoenherr mean line. Any correction allowances 2p-
plied to the Schoenherr mean line are to be clearly
stated in the report.”
As poineed out by Nordstrom (ITC Proceedings, Wassingtor,
196i} this forma applies to salt Water. For trrsh water ti
responding formula
Re = [0.000807 + 200218/(88 = B21LIS.VRESISTANCE 13
r calculations will be based
on the Schoenherr mean line with an allowanee that
is ordinarily to be +0.0004 for clean, new vessels, to
be modified as desired for special cases end in any
event to be stated in the report.”
No decision was made as regards a standard tem
perature for ship predictions, but this has subse-
quently been taken as 15 deg C (69 deg F) in conformity
swith the ICSTS figure (ATT, 1953). It was also agreed
that the Schoenherr line shall be known as the "1947
ATTC line” (ATTC, 1956). This line, both with and
without the 0.0004 allowance, is shown in Fig. 4, The
method of applying the coefficients has been deseribed
in detail by Gertler (1947), He also gave tables of their
values for 2 wide range of Reynolds numbers, together
with values of p and v for fresh and salt water.
New values of v were adopted by the IPTC? (1963)
at the 10th Conference in London in 1968. These are
also reproduced together with the C. coefficients in a
SNAME Bulletin (1976)
‘he allowances referred to in the second resolution
of the ATTC was orjginally considered necessary be-
cause of the effect of hull roughness upon resistance.
However, the difference between the ship resistance
as deduced from full-scale trials and that predicted
from the model depends upon other factors also, as is
discussed in Section 6.4 and-at the ITTC meeting in
1963 it was agreed to refer to it as a “modelship
correlation allowance” and to give it the symbol Cy,
ITC, 1969).
‘The 5th Conference of the ICSTS was held in London
in 1948, and was.attended for the first time by dele-
gates from the United States and Canada, There was
much discussion on the model-extrapolation problem,
and unanimous agreement was reached “in favor of
departing from Froude's coeficents and selecting a
substitute in line with modem concepts of skin fric-
tion.” However, the delegates were unable to agree
uupon any such alternative, largely because it was felt
that the progress in knowledge might in the near fa-
tare demand a further change. The Conference there-
fore agreed that in published work either the Froude
or Schoenherr coefficients could be used, and at the
same time set up a Skin Friction Committee to rec-
ommend further research to establish a minimum tur-
bulentfriction line for both model and ship use.
The Committe was instructed that any proposed
{friction formulation should be in keeping with modern
concepts of physies, and the coefficient C, should be
2 function of Reynolds number Rn. The Schoenherr
(ATC) line already fulfilled this requirement, but the
slope was not considered sufficiently steep at the low
Reynolds numbers appropriate to small models, so that
"2, Ship effective pow:
‘The Internttional Conference of Ship ‘afk Superintendents
(GOSTS) became the Interasional Towing Tank Conference (TTC)
mist.
it did not give
small and larg:
ing, ships
long, all-welded si
rrelation allowance nec
essam th reconcile the ship resistance with the pre
diction from the model using the ATTC line was
sometimes zero or negative. Also, Schoenherr had used
data from many sources, and the planks were in no
sense geosims, so that the experimental figures in-
cluded aspect ratio or edge effects (the same applied
to Froude's results). Telfer (1927, 1950, 1951, 1952)
suggested methods for taking edge effects into ac-
count and developed an “extrapolator” for predicting
ship resistance from model results which was an int
verse function of Reynolds number. Hughes (1952),
(1954) carried out many resistance experiments on
planks and pontoons, in the latter ease up to 77.7 m
(255 ft) in length, and so attained Reynolds numbers,
as high as 8X 10, These plane surfaces covered a
wide range of aspect ratios, and Hughes extrapolated
the resistance coefficients to infinite aspeet ratio, ob-
taining what he considered to be a curve of minimum
turbulent resistance for plane, smooth surfaces in two
dimensional flow. This curve had the equation
Cro = 0,086/ logyeRn — 2.08)" (ay
and is shown in Fig. 4. Cro denotes the frictional re-
sistance coefficient in two-dimensional flow.”
‘The ITTC Friction Committee, with the knowledge
of so much new work in progress, did not feel able in
1957 to recommend 2 final solution to the problem of
predicting ship resistance from model results, Instead,
it proposed two alternative singleline, interim engi-
neering solutions. Ono was to use the ATTIC line for
values of Rn above 10', and below this to use a new
line which was steeper than the ATTC line. The latter
would, in the Committee's opinion, help to reconcile
the results between large and small models, while us-
ing the ATTC line above Rn = 10° would meke no
difference in ship predictions from large models. ‘The
second proposal was to use an entirely new line, cross-
ing the ATPC line at about Ra ~ 10, and being slightly
steeper throughout, This would result in lower ship
predictions, and so would tend to increase the carre-
lation allowance C, and avoid negative allowances
Jong ships.
‘The Conferenee in Madrid in 1957 adopted a slight
variation of the second proposal, and agreed to
Co = 0.0T5/ (logis — 2)* 20)
‘This line is also shown in Fig. 4
‘The Conference adopted this as the “ITTC 1957
modelship correlation line,” and was careful to label
HITC Presentation Committes Report, Ottawa 1975, leo pub-
lished by the Bmtigh Ship Reseaveh Atsoeaton, now British Mar
tine Technology (BMT), as Technical Memorandum No, 500cr
CURVE OF Cry (MODEL
(6K) Cfo
PRINCIPLES OF NAVAL ARCHITECTURE
CURVE OF Crs (SHIP)
Cw Cws
Rag Re
Fin, § Exopalaton of m
it as “only an interim solution to this problem for
practical engineering purposes,” {ITTC 1957). Equa-
ion (20) was called a model-ship correlation line, and
not a frictional resistance line; it was not meant to
represent the frictional resistance of plane or curved
Surfaces, nor was it intended to be used for such a
purpose.
‘The Hughes proposal in Equation (19) is of the same
general type as the ITC line but gives much lower
values of Cy than either the IT'TC 1957 formulation o
the ATC 1947 line. On the other hand, the Hughes
fine docs claim to be a true friction line for smooth
plates in fully turbulent, two-dimensional flow, but its
low values have been eriticized by many other workers
inthis field, The 1957 ITTC line, in fact, gives numerical
values of Cy which are almost the same 28 those of
the Hughes line with a constant addition of 12 pereent.
Granville (1977) showed that the 1957 IPTC model:
ship correlation line can also be considered as a tur
bulent flat plate (two-dimensional) frictional resistance
line, From fundamental considerations involving the
velocity distribution in the boundary layer, he derived
the general formula
Cro = a/(og,Bn = bY +
1.88 and ¢
he form of the 1957 IPTC line
aad
with a = 0.0776,
is a generalization
as given by Equation (20), with a = 0.075, b
ve
a= Yb (Los BASED
Eee
results ship sing ho Form Hostor method
= 0. Good agreement of Equation (21) with the 1957
{ETC tine is obtained for values of Ra less than 5 x
30%. At values of Rn above 1 x 10%, the 1957 ITTC
the 1947 ATTIC, and the Granville lines are all in good
agreement, as shown in Fig. 4.
53.6 Three-Dimansional Viscous Rosistonce Formula~
ions, In association with his two-dimensional line,
Hughes proposed a new method of extrapolation from
model to ship. He assumed that the total model re-
sistance eoeficient Cry could be divided into two parts,
Cr and Cry, Tepresenting the viscous and wavemak-
ing resistance, respectively. At low Froude numbers,
Cray wil become very small, and at & point where
wavemaking ean be neglected, the eurve of Cry, will
become approximately paralle to the two-dimensional
friction line. Hughes called this point the run-in point,
The value of Cpy at this point can then be identified
with the total viscous resistance coefficient Cyy at the
same point Rn,
‘The form resistance coefficient, due at least in part
to the curvature of the holl (see Fig. 5), is defined by
‘The three-dimensional model viscous resistance for ar
w= +b) Cro
(Rn) where Cro is the equiv: te resistance
coaficient. The factor k accounts for the three-dimen-
Sonal form, and is appropriately termed the form fee-tor, The form factor (L + A) is assumed to be invariant
with Rn and the line (1+ &) Gry is now taken as the
repoiator for the hull form concerned, and the ship
urve of Cs can be drawn above the (1 + &)Crp curve
at the appropriate values of the Reynolds number. In
the Froude method the whole of the model residuary-
resistance coefficient C, is transferred to the ship un-
changed, while in the form factor method only that
part of Cy attributed to viscous effects (Cromya in Fig.
5) is reduced in the transfer. Accordingly, the three-
dimensional method gives substantially lower ship pre-
dictions and so calls for larger values of the correlation
allowance C,. This procedure avoids the negative al-
lowances sometimes found when using the Froude
method, It should also be noted that in the case of the
Froude method only the slope of the two-dimensional
friction line matters while in the ease of the form factor
approach the vertical position of the line also affects
the ship prediction. Phe choiee of the basic line becomes
an essential factor in the case of the three-dimensional
approach.
‘The study carried out by the ITTC Performance
Committee has shown that the introduction of the form
factor philosophy has led to significant improvements
in model-ship correlation (TTC, 1978), The ITTC has
recommended that for all practical purposes, for con-
ventional ship forms, a form factor determined on an
experimental basis, similar to Prohaska's method, is
SISTANCE
advisable; ie,
=k) = 6 Pa Coe
ea
Crel Cro =
where n is some power of Fh, 4_<¢ m ¢ 6, and cand
A are coefficients, choson so as vo fit the measured C,
Fn data points a5 well as possitle (Prohaska, 1966). (4
rummerieal example of how Prohaska’s method is used
is given in Section 6.4). This requires that tho resist
anee of the todel be measured at very low speeds,
generally at Fn ¢ 0.1, This is a drawback because
unwanted Reynolds scale effects are then often intro-
duced, For this reason sometimes empirically-derived
form factors values are adopted, However, no satis-
factory method to derive appropriate values of such
form factors has as yet beon found. The ITTC Per-
formance Committee, which reviews, collates and tests
the various proposed methods, states in its 1978 report:
"With regard to the influence of form on the various
components of the viseous resistance no clear conela-
sion can be drawn, Results reported by Tagano (1973)
and Wieghardt (1978) show that the form mainly in-
fluences the viscous pressure drag, while Dyne (1977)
stated that the pressure drag is low and its influence
on k is practically negligible. Furthermore, the inter-
action between different resistance components is hin
dering the isolation of a single signifieant facto
Section 4
Wave-Making Resistance
41 General. ‘The wavemaking resistance of a ship
is the net fore-and-aft force upon the ship due to the
fluid pressures acting normally on all parts of the hull,
just as the frictional resistance is the result of the
tangential fluid forces, In the case of a deeply sab-
merged body, travelling horizontally at a steady speed
far below the surface, no waves are formed, but the
normal pressures will vary along the length. In & non-
viscous fluid the net foreand-aft foree due to this var-
fatfon would be zero, as previously noted.
If the body is travelling on or near the surface,
however, this variation in pressure causes waves which
alter the distribution of pressure over the hull, and the
resultant. net fore-andaft foree is the wave-making
resistance. Over some parts of the hull the changes in
pressure will increase the net sternward fore, in oth-
ers decrease it, but the overall effect must be a re-
sistance of such magnitude that the energy expended
in moving the body against it is equal to the energy
necessary to maintain the wave system. The: wave-
making resistance depends in large measure on the
shapes adopted for the atea curve, waterlines and
transverse sections, and its determination and the
methods by which it can bé reduced are among: the
‘main goals of the study of ships’ resistance: Two paths
have been Zollowed in this study—experiments with
models in towing tanks and theoretical research into
wave making phenomena, Nefther has yet led toa com-
plete solution, but both have contributed greatly to 2
better understanding of what is 2 very complicated
problem. At present, model tests remain the most im-
portant tool available for reducing the resistance of
specific ship designs, but theory lends invaluable help
in thterpreting model results and in guiding model
research,
4.2 Ship Weve Systoms. The earliest aecount of the
way in which ship waves are formed is believed to be
that due to Lord Kelvin (1887, 1904). He considered
single pressure point travelling in a straight line over
the surface of the water, sending out waves which
combine to form a characteristic pattern, This consists
of a system of transverse waves following behind the
point, together with a series of divergent waves ra-
diating from the point, the whole pattern being eon-
tained within two straight lines starting from the
Pressure point and making angles of 19 deg 28 min6 PRINCIPLES OF NAVAL ARCHITECTURE
est
wave, 7ROUGH
ig & Kelsie wove parm
‘on each side of the line of motion, Fig. 6. The heights
bi successive transverse-wave erests along the middle
line behind the pressure point diminish going aft. The
waves are carved back some distance out from the
centerline and meet the diverging waves in cusps,
‘which are the highest points in the system, The heights
of these cusps decrease less rapidly with distance from.
the point then do those of the transverse waves, so
that eventually well astern of the point the divergent
Fig Mo) Pato of dharing waves,
waves become the more prominent (see Fig. 7).
‘Tho Kelvin wave pattern illustrates and explains
many of the features of the ship-wave system. Near
the bow of a ship the most noticeable waves are a
series of divergent waves, starting with a large wave
at the bow, followed by others arranged on each side
along a diagonal line in such a way that each wave is
stepped back bebind the one in front in echelon (Fig.
8) and is of quite short length along its crest line
Between the divergent waves on each side of the ship,
transverse waves are formed having their crest lines
normal tothe direction of motion near the hull, bending
‘back as they approach the divergent-system waves and
finally coalescing with them. These transverse waves
are most easily seen along the middle portion of a ship
or model with parallel body or just behind a ship run-
ning at high speed. Itis easy to see the general Kelvin
pattern in such a bow system.
‘Similar wave systems are formed at the shoulders,
if any, and at the stern, with separate divergent and
transverse patterns, but these are not always so
clearly distinguishable because of the general distur-
bance alzeady present from the bow system.
Since the wave pattern as a whole moves with the
ship, the transverse waves are moving in the same
direction as the ship at the same speed V, and might
Fla, 7b) Type! shin wove posteRESISTANCE
Fig, Schematic diagram of how ond stim wove systems
be expected. to have the Tength appropriate to free
waves running on surface at that speed
Ly = 2m V9
Actually, the waves in the immodiate vieinity of
model are found to be a little shorter, but they attain
the longth Ly about two wave lengths astern,
"The divergent waves will havo a different speed
along the line normal to their crests (Fig. 9). In this
case, the component of speed parallel to the line of the
ship's motion must be equal to the ships speed in order
to retain the fixed pattorn relative to the ship. Tf the
‘ line normal to the exest of a divergent wave makes an
7 angle @ with the ship’s course, the speed in that di-
rection will be Veo 6, and the corresponding wave
‘ length
: Liy = (@n¥2/g) cost 0
4.3 Wavo-Making Resistance of Surface Ships. At
low speeds, the waves made by the ship are very small,
and the resistance is almost wholly viseous in char-
+ acter, Since the frietional resistance varies at power
‘of the speed a litle less than the square, when the
oefieian of total resistance G- = Ry/ipSV2 is plotted
to a base of Froude number Fn (or of Vg/Z), at first
the value of Cr decreases with increase of speed (Fig.
10). With further inerease in speed, the value of C;
‘ begins to increase more and more rapidly, end at
Froude nambers approaching 0.45 (Vx/yZ = 1.5)
the resistanee may vary at a power of of 6 or more.
However, this general increase in Cris usvally eecom-
panied by a mumber of humps and hollows in the re-
sistance curve. As the speed of the ship inereases, the
‘ wave patters must change, for the length of the waves
vill inerease and the relative positions of their crests
and troughs will alter. In this process there will be
c succession of speeds when the crests of the two eys-
in. Spand ond ong of divergent waves
c
Fig. 10-Typlalrestnee curve showing Inoence eens
‘tems reinforce one another, separated by other speeds
at which crests and troughs tend to eancel one another.
‘The former condition leads to higher wave heights, the
latter to lower ones, and as the energy of the systemsPRINCIPLES OF NAVAL ARCHITECTURE
18
=depends upon the square of the wave heights, this
ane alternating speeds of higher and lower than
average resistance, The humps and hollows in the
C,curve are dae to these interference effects between
the wave systems, and it is obviously good design
.ctige to enstire whenever possible that the sip wil
running under service conditions at a favorable
speed. As will be seen later, it is the dependence of
these humps and hollows on the Froude number that
accounts for the elose relationship between economic
speeds and ship lengths.
‘The mechanism by Which wavemaking resistance is
generated is well illustrated by experiments made
Eqgort (1989). He measured the normal pressure dis:
‘eibution over the ends of 2 model and plotted resulting
pressure contours on a body plan (Fig. 11). By int=
rating the longitudinal components of these pressure
forees over the length, he showed that the resulting
resistance agreed fairly well with that measured on
the model after the estimated frietional resistance had
been subtracted. Fig. 12a shows curves of longitudinal
force per meter iength; Fig. 12b shows form resistance
derived from pressure experiments and residuary re-
Sistance model tests. One important point brought out
by these experiments is that a large proportion of the
vwavermalking resistances is generated by the upper part
of the hull near the still waterline.
‘44 Theoretical Calculation of Wave-Moking Resist-
ence. Much research has been devoted to theoretical
tethods of caleulating wave-making resistance and to
their experimental verification (Lunde, 1957),
‘One method is to determine the flow around the lull
and hence the normal pressure distribution, and then
to integrate the fore-and-aft components of these pres
sitres over the hull surface. This method was developed
by Michell (1898) for a slender hull moving over the
surface of a nonviscous fluid. It corresponds to the
experimental technigue employed by Eggert. The pi
oneer work of Michell was unfortunately overlooked
and negieeted ior many years until rescued from ob-
scurity by Havelock (1962).
‘A second method is to caleulate the wave pateorn
generated by the ship at 2 great distance aster, as
done by Havelock, the wave-making resistance then
being measured by the flow of energy necessary to
maintain the wave system, This method has been used
experimentally by Gadd, et al (1962), Eggers (3962),
Ware (1982) and many others.
‘Both methods iead to the same final mathematical
expression, and in each case the solution is for a non-
viseous and incompressible fiuid, so that the ship will
making resistance (Timman. et
ematical expression for
‘siendeo” ship of nerrow beam whe
yrm stream. From the resultant velocity
‘and pressure distribucion aver
ined, and by integrating the fore
the now aroun
placed in 3 u
Potential the reice
the hull can ®:
SISTANCE
and.agt components of the pressure an expression can
be derived for the total wave-making resistance."
‘The theory as deveioned by Micheli is valid only for
cervain restrictive conditions:
fa). ‘The fuid is assumed to be nonviseous and the
flow irrotational. Under these circumstances the mo-
tion can be specified by a velocity potential &, which
in addition must satisfy the necessary boundary con
ditions.
@)_ ‘The hull is narrow compared with its length, so
thatthe slope of the surface relative tothe centerline
plane is small.
(2) The waves generated by the ship have heights
small compared with their lengths, so that the squares
of the particle velocities can be neglected compared
with the ship speed.
(@) The ship does not experience any sinkage or
tim,
‘The boundary con
locity potential @ are:
(@) At all points on the surface of the full, the
normal velocity relative to the hull must be zero.
(G). The pressure everywhere on the free surface
of the water must be constant and equal to the at-
‘moapherie pressure.
‘To make the problem amenable to existing mathe-
matical methods, Michell assumed thatthe first bound-
ary condition could be applied to the centerline plane
rather than to the actual hull surface, so that the
results applied strictly to a vanishingly thin ship, and
thatthe eondition of constant pressure could be applied
to the original flat, free suriace of the water, the dis-
tortion of the surface due to the wave pattem being
neglected. 1
‘The alternative method developed by Havelock, in
which the wavemaking resistance is measured by the
snergy in the wave system, makes use of the idea of
sourees and sinks. This is a powerful tool with whieh
to simulate the fiow around diferent body shapes and
50 to find the Wave patzern, pressure distribution, and
resistance. A “thin” ship, for example, can be simu-
lated by @ distribution of sources on ‘the centerline
plane of the Zorebody and of sinks in the afterbody,
the sum of their total strength being zero. The re.
tions to be satisfied by the ve
T ihe velocey potential & has the propersr that the velocity of
ie flow In any given cireadon ie te pareal deivasie of @- mith
rte that erection, Thos lor 3 anor Seam 9f yell
the negate eauretin, the Veloety ponds Wil be defines bssPeE0 OF ODEL
Ree IK POUNDS
re
“ive FoRces, i 80H IEREASE
oT NESSIE oF MODEL
BEETS a ree tenets
‘Carve show longitudinal free per inch length
o
‘
rel
PRINCIPLES OF NAVAL ARCHITECTURE
to be everywhere proportions] to the slope of the hull
Surface, this will result in a total strength of zero, and
the total velocity potential wil be the sum of those
gusto theta dual sources and sinks and the uniform
oe “
‘Each source and siak when in motion in a uid; om
or near the surface, gives rise to a wave system, and
‘bys ‘these up the total system for the ship can
be obt Havoloc by tis ‘method foand the wave
pattern far astern, and from considerations of energy
Sbtained the wavemaking resistance,
‘Much of the research into wavemaking resistance
has been done on models of mathematical form, having
seetions and waterlines defined by sine, cosine, oF par
abolie functions. When the calewations are applied to
aebual ship forms, the shape of the latter, must, be
sed approximately by the use of polynomials
(Weinbium, 1950, Wehausen, 1913); or by considering
the hull as being made up of s number of elementary
wedges (Guilloton, 1951).
nvrecent years, a great deal of work on the caleu-
FORM RESISTANCE ——Lo/7
' fi
‘ Siouary
fi — tee
Fim | 7
= T
Ife RESIOUARY) iE
ere eC =
Fig i) Curves of Sar rsiance ond cea rete
Form rsstance derived ftom presure experiments
REIRAISy telistance Genved from vowing experiments
striction to 2 “thin” ship can be removed if the sources
and sinks are distributed over the hull eurface itself
Tf the strengths of the sourees and sinks are assumed
Tp source may be looked upon as point in 2 fu ax which new
‘uid'e being continuously intradueed. and a sik isthe reverse,
ae eens i beme comambously aostractad, The RoW o¥t cE
punt whe ea sink wi copsstof radial straight swear
[Buide ust une coud be tonioed oy a 80" 000,
lation of wave making resistance has been carried out
in Japan by Professor Inui and his colleagues (Inui,
Tea0) They used 2 combination of mathematical and
experimental worl and stressed the importance of ob-
Serving the wave pattern in detail as well.as simply
measuring the resistance. Instead of starting with
given bull geometry, Professor Inai began with an
Eeumed sourcesinit distribution, with 2 view to ob-
faining better agreoment between the measured and
Calealated wave systems, both of which would refer
$ethe same hull shape. The wave pattern and the waver
tnaicing resistance were then calculated irom the any
HFlitudes of the elementary waves by using Haveloci’s
concopt.
Professor Inui tried various distributions of sourees
and sinks (singularities) by volume over the curved
and, sinks (Ciogrtfzntal plane and over the vertical
niddleline plane, For displacement ships at Froude
Timbers from 0.1 t0 0.85, he found the geometry of
Tretnds to be most important, and these could be
Jepresented quite aceurately by singularities on the
Tadldleline plane. For higher Froude numbers, the dis
\Uy, He
WZ : UY
WE OSNIXe
A A,SISTANCE :
tribution of sources along the whole length becomes
important. In summary, the method is to choose 2
singularity distribution which will give good resistance
‘qualities, obtain the corresponding null geometry,
carry out resistance and wave-observation tests and
modify the hull to give a more shipshape form araid-
ships. In this way Inui has been able to obtain forms
with considerably reduced wavemaking resistance,
‘usually associated with a bulb at the stem and some-
times at the ster also,
"Recent developments in wavemaking resistance the-
ory ean be divided into four main categories. The first
concerns applications of linearized potential fow the-
ory, either with empirical correetfons to make it more
accurate, or uncorrected for special cases where the
errors due to linearization are not serious. The second
concerns attempts to improve on linearized potential
flow theory, by analysis of non-linear effects on the
free-surface condition, or by an assessment of the ef
fects of viscosity. Thirdly, attempts have been made
to apply wave resistance theory to hull form desiga.
Fourthly there has been an increase in the number of
primarily numerical approaches to ship wave resist:
ance estimation. In the second eategory (nonlinear
caleulations) the work of Daube (1980), (1981) must be
mentioned. He uses an iterative procedure where at
each step 2 linear problem is solved, To this end an
initial guess of the location of the free surface is made
which is subsequently changed to full a tree surface
condition. in the computation of the free surface ele-
vation the assumption is made that the projection of
the free surface streamlines on the horizontal plane,
0, always agree with the double model streamlines.
‘This is in Zact a low-speed assumption. The norrlinear
calculation method has been applied to a Wigley hull
and a Seriés 60 ship. Comparison with messurements
show a qualitatively satisfuctory agreement, and quam-
titatively the caloulations are better than with linear
theory for these cases. Part of the discrepancies be-
freen measurements and calevlations (at least for the
higher speed range) can be ascribed to trim and sin-
d kage effects which have not been property included in
Daube’s method.
‘An interesting development has been the determi
nation of pure wave-making resistance from measure
ments of model wave patterns. The attempts to
improve hull forms using the data of wave pattern
measurement combined with linearized theory are par-
ticularly interesting, For example, Baba (1972) mea
sured the difference in wave pattern when a given hull,
was modified according to the insight gained from
wave resistance theory and thereby gained an im-
provement.
Toa certain degree, the hull forms of relatively high-
speed merchant ships have improved because of the
application of wave resistance theory. Pien etal (1972)
proposed a hullorm design procedure for high-speed
displacement ships sided by wave resistance thedry.
Inui and co-workers have applied the streamline trac-
ing method to practical hull forms with fat bottoms
and a design method for highspeed ships with the aid
of minimum wave resistance theory has been proposed
by Maruo etal (1971). The developmentof special types
of hull forms for drastically redueed wave making
have also been guided to a certain extent by wave
resistance theory. One of these is the small waterplane
area tvin-hull (SWATH) ship, discussed in Section 9,
‘The accuracy and usefulness of wave resistance theory
was recently demonstrated at a workshop organized
by the DIRC ‘Bai, et al, 1979).
‘The results of theoretical york would therefore
seem at present to be most useful in giving guidance
in the choice of the secondary features of hull shape
for given proportions and fullness, such as the detail2
Fn = 0.238
PRINCIPLES OF NAVAL ARCHITECTURE
Digecrion oF MOTION
oan 5)
PLAN OF WATERLINE _ z
SIDES VERTICAL
ICAL 7
bigtunoance
pow sysreM Pees,
SEAIE OF WAVE HEIGHT,
eae
Fore SuguLDER
‘Sesto
arr sHouLoen ~
svstew
STERN SYSTEM
‘TorAL WAVE
PROFILE
——— catcaren PROFLE
TI neasure proFiLe
sremn
Fos15. Wave nyse for simple wedgecinoned fore
shapes of waterlines and sections, and choice of size
and location of bulbs. The ealeulation of resistance
cannot yet be doie with suficient accuracy to replace
model experiments, but it is a most valuable guide and
interpreter of model work. The advent of the computer
has placed new power in the hands of the naval ar-
chitect, however, and has brought mueh closer the time
ohea theory can overcome its present limitations and
begin to give meaningful numerical answers to
resistance problem.
43° Interference Effects, ‘The results of mathemat-
jeal research have been most valuable in providing an
insight into the effects of mutual wave interference
upon wavemaking resistance, A most interesting ex-
ample is that of 2 double-wedgeshaped body with par-
tlle! inserted in the middie, investigated by Wigley
(1991). The form of the hull and the calculated and
measured wave profiles are shown in Fig. 15. He
Showed that the expression for the wave profile along
the hall contained five terms:
fa) A symmetrical disturbance of the surface,
which has a peak at bow and stern ané a trough along
the center, dying out quickly ahead and astern of the
hull Tt travels with the hull and because of its sym
metry does not absorb any energy at constant speed,
and four wave systems, generated at
Fig 1S Rosen
Dimensions: 48 x 06 X 0.89 = prismasi coe
a crest;
(®) the bow, beginning wiRESISTANCE:
(aoe i
Le
|
"ue Yo wiTeRreRENCE
! |
Torak Gy cuRve.couat TO
Sun oF Scouranents
T x
\WAVE-MAKING RESICTANCE COEFFICIENT,
Gy Due 10 wrERFERENCE—
g Seriya eons
Pore
| |
rao
fa
Fa. 17 Arey of wovemaking essonce ito components fr wade
“hoped del shown above
(@. the forward shoulder, starting with a trough;
(@) the after shoulder also starting with a trough
(@) the stern, beginning with a crest,
‘These five systems are shown in Fig. 15. Consider-
ably aft of the form, all four systems become sine
curves of continuously diminishing amplitude, of a
length appropriate to a free wave travelling at the
speed of the model, this length being reached after
about to waves.
‘The ealculated profile along the model is the sum of
these five systems, and the measured profile was in
general agreement with itso far as shape and positions
of the erests and troughs were concerned, but the
heights of the actual waves towards the starn were
considerably less than those caleulated (Fig. 18).
‘This simple wedge-shaped body ustrates clearly
the mechanism of wave interference and its effects
upon wavemaking resistance. Because of the definite
sharp corners at bow, stern, and shoulders, the four
freeware systems have their origins fixed at points
long the hull, As speed increases, the wave lengths
of each of the four systems increase, Since the primary
‘rests and troughs are fixed in position, the total wave
profile will. continuously change in shape with speed
as the erests and troughs of the different systems pass
through one another. At those speeds where the in-
terference is such that high waves result, the wave-
making resistance will be high, and vice-versa,
a
In this simple wedge-shaped form the two principal
types of interference are between two systems of the
same sign, e.g,, bow and stern, or the shoulder sys
tems, and between systems of opposite sign, o.g., bow
and forward shoulder. The second type is the most
important in this particular ease, because the primary
hollow of the first shoulder system can coincide with
the frst trough of the bow system before the latter
has been materially reduced by viscous effects
‘Wigley calculated the values of V/-VgL for minima
and maxima of the weve-making resistance coefficient.
Cy for this form, and found them to occur at the
following points:
Values of Fn
Minima Cy — 0.187 — 0.281 — 0,345
Maxima Cy 0.178 —~ 0.205 — 0.269 — 0.476
‘The mathematical expression for the wave-making
resistance Ry is of the form
Ry & V* (constant term + 4 ostillating terms)
so that the wavemaking resistance coefficient Cy is
Cy = Ry /MepSV* = V* (constant term + 4 oscillating
terms) @)
| mer ams
sure oF wi]
| eel!
wcrmpenpen |
|
ia ae lse tea
cary |
Ns| [rome ce wa I
ednavense waves
\WAVE-MAKING RESISTANCE COEFFICIENT Cy
car | {|
oa pow mubR A
Fo. 19 Contibuon made by ronane nd dieigent weve system
‘The Cy curve is thus made up of 2 steady increase
varying as V* due to the constant term and four os-
cillating curves due to the interference between the
diferent freo-wave systems (Fig. 17). These latter ul-
timately, at very high speeds, cancel both each other
and the Steady increase in Cy, and there is no further
hump beyond that occurring ata V gL value of about
0.45 after which the value of Ry, continuously de-
creases with further increase in speed. However, at
these high speeds the hull will sink bodily and change
trim so much that entirely new phenomena arise,
‘For more ship-shaped forms, where the waterlines
are curved and have no sharp discontinuities, the wave
pattern still consists of five components—a symmetri-
bal disturbance and four freewave systems (Wigley,
1924). Two eystems begin with crests, one ab the bow
and one at the stern, and are due to the change in the
angle of the flow at these points. The other two sys-
tame, like the shoulder systems in the straight line
form, begin with hollows, but are no longer tied to
definite points, since the change of slopeis now gradual
and spread over the whole entrance and run. They
commence at the bow and after shoulder, respectively,
1s shown in Fig. 18, much more gradually than in the
ease of the wedge-shaped form. The one due to en-
trance curvature, for example, may be looked upon as
2 progressive reduction of that due to the bow angle
as the slope of the waterline gradually becomes less
in going aft
Wigley also made caleulations to show the separate
contributions to the wavemaking resistance of the
transverse and divergent sysvems (Wigit™. 1942). Up
to. Froude number of 0.4 the transyer: waves are
‘mainly responsible for the positions of the numps and
hollows, Fig. 29. Above this speed the contribution
from the divergent waves becomes mere snd more
important, and the interserence of wae wansverse
raves tlone will not correctly determine the position
PRINCIPLES OF NAVAL ARCHITECTURE
of the higher humps, particularly the last one at
Pa = 05.
"The existence of interference effeets of this kind was,
known to naval architects long before such mathe
matieal analysis was developed. The Froudes demon-
strated them in a striking way by testing a number
of models consisting of the same, bow and stern sep-
arated by different lengths of parallel body (Froude,
W,, 1877 and Froude, RLE., 1881), W. Froude’s sketch
of the bow wave system is shown in Fig. 20. As the
ship advanees but the water does not, much of the
energy given to the water by the bow is carried out
Taterally and away from the ship. This outward spread-
ing of the energy results in a decrease in the height
fof each sucezeding wave of each system with no ap
preciable change in wave length, Fig. 21 shows a series
ff tests made at the EMB, Washington, and the cor.
responding curve of mode) residuary resistance plotted
‘against length of parallel body (Taylor, 1949). The tests
were not extended to such a length of parallel body
that tho bow system ceased to affect that at the stern,
This clear, however, that its effect is decreasing and
would eventually died ont, as suggested by the dotted
extension of the resistance curve.
Fig. 22 shows a series of curves for the same form
at various speeds. In this chert the change of paralle)
rniddle-body length which results in succossive bumps
fon any one curve is very nearly equal to the wave
Jength for the speed in question, as shown for speeds
of 26 and 8.2 knots. This indicates that chip waves do
have substantially the lengths of deep-sea waves of
the same speed.
Tf all the curves in Fig, 22 aré extended in the di-
rection of greater parallél-body length until the bow
system ceases to affect the stern system, as was done
in ig. 21, the mean residuary resistances for. this
form, shown by the dashed lines at the left of the
chart, are found to increase approximately as the sixth
power of the speed, They are, in fact, the actual re
Elstances stripped of interference effects and represent
the true residuary resistances of the two ends. This
rate of variation with speed is the same as that given
by theory for the basic wave-making resistance before
taking into account the interference effects (Fig. 17),
‘The mathematical theory indieates that the wave
resistance is generated largely by those parts of the
poll nest the surface, which is in agreement with the
experimental results obtained by Bggert. This sug-
gests that from the point of view of reducing wave-
aking resistance the displacement should be kept as,
low down as possible. The relatively small effect of the
lower part of the hull on the wave systems also means
that the wave-making resistance is not unduly sensi
tive to the midship section shape (Wigley, et sl, 1948)
“AS. Effects of Viscosity on Wave-Making Resis!-
once. Calculations of wave-making resistance have so
far been unable w take into account the effects of
viscosity. the role of which has been investigated
Havelock (1923), (1985) and Wigley (1998). One of theCroturat ay
fer 20 Knots :
: 23
: t oF
Ka a
igdle Body ngthvnches
fa 21 Quon eet fang length of porate de ody ish na
6
*Naturet’R be
for Various Speeds 3
we
:
2 ae
Zz «3
z 42
§ 1e
°
°
7% a
Parallel Middle Body: Length, inches
Fig. 22, Anahss eae af tering length of parle mide bea ah it)PRINCIPLES:
percent
‘or 3e Bs oo
‘FROUDE. HUNBER Fy
Fig. 23. Seale eet on moveemating eesiance
effects is to eroate a boundary layer close to the hull,
which separates the latter from the potentialflow pat
tem with which the theory deals. This layer grows
thieker from stem to stern, but outside of it the fuid
behaves very much in accordance with the potential
flow theory. Havelock (1926) stated that the direct in-
fluence of viscosity on the wave motioa is compara-
tively small, and the “indirect effect might possibly be
allowed for later by some adjustment of the effective
form of the ship.” He proposed to do this by assuming
that the after body was virtually lengthened and the
aft end waterlines thereby reduced in slope, 50 reduc
ing the afterbody wavemaking. Wigley (1962) Zol-
lowed up this suggestion by comparing ealeulated and
measured wave-making resistance for 14 models of
mathematical forms, and dariving empirical correction
factors. He found that the remaining differences in
resistance svere usually within 4 percent, and that the
virtual lengthening of the hull due to viseosity varied
bevween 2 and 8 percent.
‘The inclusion of a viseosity correction of this nature
also explains another feature of ealeulated wavemak-
ing resistance. For a ship model which is unsymmet-
rical fore and aft, the theoretical wavemaking
resistance in 2 nonviscons fiuid is the same for both
directions of motion, while the measured resistances
are different. With the viscosicy correction included,
the calculated resistance wil also be different,
‘Professor Inui (1980) in his wave-making resistance
work also allows for viscosity by means of two em-
pirical coefficients, one to take care of the virtual
lengthening of the form, the other to allow for the
effect of viscosity on wave height.
4,7 Scale EFlect_on Wave-Making Rosistancs.
Wigley (1962) has investigated the seale effeet on Cy
due to viscosity, pointing ont that the caleulated eurves
of Cy are usually higher then those measured in ex-
periments and also show greater oscillations. These
differences he assigned to three major causes:
fc) Errors due to simplifications introduced tw
make the mathematical work possible.
(Errors due to neglect ofthe effects of viseosity
(OF NAVAL ARCHITECTURE
(©) Brrors due to the effects of wave motion on Rp,
Brrots under (a) will decrease with increasing speed,
sinee they depend on the assumption that the velocities
due to the wave motion are small compared with the
speed of the model, whieh is more nearly falfiled at
high speed.
Excors under (6) will depend on Reynolds number,
and therefore on the size of the model, decreasing a8,
size inereases. From experiments on unsymmetrical
‘models tested moving in both directions, these errors
cease to be important for Fn greater than 0.45.
‘At low speeds errors under (¢) are negligible, but
become important when Fa exceeds 0.35, (7./VD
1.15) 28 evidenced by the sinkage aad trim, which
crease very rapidly ahove this speed.
‘A practital conclusion from this work is the effect
on the prediction of ship resistence from a modal. In
1 typieal model the actual wave resistance is less than
that ealeulated in a perfect fluid for Froude numbers
less than about 0.95. This difference is partly due to
viscosity, the effect of which will decrease with in-
erased size, and Cy will increase with seale instead
of being constant 25 assumed in extrapolation work
Wigley made estimates of the difference involved in
caleulating the resistance of a 121.9 m ship from that
of a 488 m model at a Froude number of 0.245 and
found that the resistance of the ship would be under-
estimated, using the usual caleulations, by about 9
pereent, the variation with speed being approximately
‘as shown in Fig. 23. The effect disappears atlow speeds
and for vahies of Fn above 0.48.
“4.8 Comparison Batweon Colevlctod and Observed
Wave-Making Resistance. Many comparisons have
‘been made between the ealeulated and measured wave-
making resistances of models. Such # comparison is
dificult to make, however. All that ean be measured
on the model is the total resistanee #, and the value
Gp AND 6