You are on page 1of 38

On the Turning and Steering of Ships

BY K E N N E T H S. IX{. DAVIDSON, / ~ E M B E R 1

Ships have been turned and steered with reasonable success for centuries. Yet, as recently as
1923, Sir Arthur Johns referred to the study of
turning and steering as "a generally neglected
portion of naval architecture" [1]. 2 T o be sure,
the literature on these subjects has been augmented in the twenty years since 1923, [2] [3] [6]
[11] and others. The fact remains, however, t h a t
it has not yet become possible to predict turning
and steering characteristics with assurance in the
preliminary design stage.
The situation is not unlike that which would
exist with regard to resistance and effective horsepower estimates, if codified experimental results
like those for Taylor's Standard Series of hull
forms [4] had never been worked out. The Standard Series data provide a coordinated framework
of quantitative information on resistance which
brings out clearly the relative influence of the principal governing variables, and from which reliable
estimates can be made for new designs. Lacking
a corresponding framework of quantitative information on turning and steering, the designer
finds himself in a position where, as one naval
architect put it recently, " I t is one thing to progress from one design to the next, on the basis of
accumulated experience, and quite another to 'hit
it on the nose' every time."

This paper makes no pretense of providing design information. I t is essentially a progress report on methods of attack, dealing first with
broad concepts in an a t t e m p t to gain perspective
on the scope of the problems involved, and
second with analytical and experimental procedures aimed at simplifying a comprehensive
attack.
Attention is centered largely on the influence
of the hull. This is done, not with any thought
that the rudder is of secondary importance in fixing the turning and steering characteristics, but
because (1) the rudder is basically an appendage,
the design of which is necessarily governed b y the
job it has to do, that is, b y the characteristics of
the hull to which it is applied; (2) a good deal of
reliable information on rudder characteristics is
already available, [2 ] [ 10] [ 11] and others,, whereas .
relatively little systematic work has been done on
hull characteristics. In effect, the point of view is
adopted that a better knowledge of hull action is a
prerequisite to a better understanding of rudder
action and, ultimately, of interference effects between hull, rudder and propellers.
GENERAL CHARACTERISTICS OF THE TURNING
MANEUVER

The general features of the turning maneuver


are well known. Fig. 1 refers to the conventional
maneuver ordinarily used for reference purposes,
in which (1) the ship approaches on a straight
course at a uniform speed, with fixed throttle settings which are not altered during the maneuver,
(2) the maneuver is considered to begin with execution of the order to lay the rudder, (3) the rudder
is laid to a prescribed angle at the maximum rate
provided b y the steering engine and is not subsequently altered during the maneuver. With these
restrictions, it has been found t h a t for practically
all types of seagoing vessels:
(a) A substantially steady state of turning
motion can be expected to establish itself, roughly
within the first quadrant of the turn, the ship
287

1 Professor of M e c h a n i c a l E n g i n e e r i n g a n d D i r e c t o r of Experim e n t a l T o w i n g T a n k , S t e v e n s I n s t i t u t e of Technology, Hoboken,


N . J . ; C o n s u l t a n t w i t h D a v i d W. T a y l o r M o d e l Basin, U n i t e d States
N a v y D e p a r t m e n t ; M e m b e r , S u b c o m m i t t e e on Seaplanes, N a t i o n a l
A d v i s o r y C o m m i t t e e for Aeronautics.
Dr. D a v i d s o n was born in Buffalo, N. Y., on F e b r u a r y 9, 1898.
He received t h e degree of Bachelor of Science in M e c h a n i c a l Engineering from t h e M a s s a c h u s e t t s I n s t i t u t e of T e c h n o l o g y in 1920.
I n 1918, he was a Second L i e u t e n a n t , P i l o t , in t h e A r m y Air Service.
F r o m 1923 t o 1929, he was engaged by t h e Crosby S t e a m Gage a n d
Valve C o m p a n y as a research engineer a n d t h e n as chief engineer.
I n 1929 he came t o Stevens, a n d in 1935 became the D i r e c t o r of t h e
E x p e r i m e n t a l T o w i n g T a n k . I n 1944, he was awarded a n h o n o r a r y
degree of Doctor of Science b y S t e v e n s I n s t i t u t e of Technology.
Dr. D a v i d s o n is t h e a u t h o r of n u m e r o u s technical papers including,
"'Some E x p e r i m e n t a l S t u d i e s of t h e S a i l i n g Y a c h t , " Transactions of
T h e Society of N a v a l Architects a n d M a r i n e Engineers, 1936, for
which he won t h e first award of t h e C a p t a i n Joseph H. L i n n a r d Prize
a n d he c o n t r i b u t e d t h e c h a p t e r on " R e s i s t a n c e a n d P o w e r i n g " t o
the book " P r i n c i p l e s of N a v a l A r c h i t e c t u r e . " H e is a m e m b e r of T h e
Society of N a v a l A r c h i t e c t s a n d M a r i n e Engineers, t h e I n s t i t u t e of
the A e r o n a u t i c a l Sciences, a n d T h e A m e r i c a n Society of M e c h a n i c a l
Engineers.
2 N u m b e r s in brackets designate references l i s t e d a t t h e end of t h e
paper. N o a t t e m p t has been made to prepare a complete bibliogr a p h y on t h e subject, t h e references being essentially i n d i c a t i v e of
a v a i l a b l e material.

288

ON

THE

_ .........

TURNING

AND

STEERING

OF SHIPS

To_d2~~ i Di22 ~__+~


$Z s.__o_o~ ~_( T~c =s.3o ?

Trat~s%r : L6~ xL(T/-~D~

I
I
I
i

"E'I

/ x"

?
.-.i I
I
col
@

7\.

?I

\
'\
\,

~I

<I

-~Jr

I
J
I

-olT>

./

,/

\~o//

I
]
Execue

/- . /
/ //

5+ec~dy- Turrli~g
Approoch

Speed

Smeed

V/
= ~/~'o

T
J

_jl
7t

~J
31

Note:
(1) Successive positions s h o w n are for equal t i m e intervals.
(2) A d v a n c e , t r a n s f e r a n d tactical d i a m e t e r are defined with r e s p e c t to
t u r n i n g p a t h . I n o r d i n a r y n a v a l u s a g e t h e y are defined w i t h r e s p e c t to t h e
s h i p ' s h e a d i n g . T h e principal effect of this difference is Oll t h e transfer.

l
l

FIG.

1,--TYPICAL

M I N I M U M T U R N FOR A N O R M A L SEAGOING ~rBSSEL

Proportions are consistent with m e a n curves of Fig. 3.

ON T H E T U R N I N G AND S T E E R I N G OF SHIPS

289

Ficjhfer

Tramspor+

AirpI~ne

Airplane

Sh!ip

5Bamk

80Bcmk

Ce.n-~rif:uojal Accel.l.09 ~

O.04oj-

t.)

Ca)

(b)

No+e:
Forces Propo~+~om(xl +o Accelerc~fions

L =40F~.

L =I00 Ft.

L =400F+,

:4ooM.~.H.

: Z O O M.~.H.

: I~ K+s.

v/TE : ss (v i~ K~o+~)

7E: ~4 (v i~ K~o+~)

7V[ = 0.8

%-9o

y~s4

o/~=3

FIG. 2.--~OMPARATIVE 1V~INIMUM TURNING CIRCLES OF AIRPLANES AND SHIPS, SHOWING THE GREAT EFFECT OF SPEED
ON THE I~ESULTANT ACCELERATION IN TURNING

swinging thereafter at constant angular velocity stitutes really good turning, or whether it repreabout a fixed center, at a uniform turning speed sents simply a mean of what has been accomplished
less than the approach speed.
so far. This is rather a basic question, which de(b) A transient phase occurs at the start, dur- serves some attention.
ing which the instantaneous turning radius diminThe possibility suggests itself of getting a someishes from infinity to the steady-turning value and what broader view of the question by comparing the
the instantaneous speed diminishes from the ap- minimum turning circles of ships with those of airproach to the steady-turning value.
planes (Fig. 2). Airplanes and ships differ radically
Fig. 1 has been drawn with some care to show to in their overall configurations. Yet both are free
scale a typical minimum-diameter turn (with bodies moving through fluids, and, in general, have
maximum rudder angle) for a normal seagoing to be turned as expeditiously as possible. Without
vessel having good turning characteristics by looking into the matter, it could easily be imapresent-day standards. The term "good" is ap- gined that airplanes might be found to have better
plied to these characteristics primarily because the turning characteristics than ships. This, however,
tactical diameter is only three times the length is not the case. Airplanes have to turn in much
of the ship (TD/L = 3), and this figure is rarely larger circles than ships--measured in terms of
bettered in practice at the present time. Mini- length--simply because of the limitation imposed
mum tactical diameter is a simple criterion by by the permissible centrifugal force. In airplanes,
which to compare the turning characteristics of the centrifugal force caused by the high speed may
different vessels, and has been used widely for the easily exceed the force of gravity by a large marpurpose. But the question naturally arises of gin, making heavy banking necessary--as also in
whether a tactical diameter of three lengths con- racing automobiles or even racing bicycles. In

t.O

0.9
15

E ~ 0.s

...,,.,.L.~
~dP'

E
L

--...~

~"

;-

" e ~

p
t

C0
"4CO

f:

4-

/0

fZ

12
I
I
[
~L

:Z

1.0

~ O.g

~0

ogo

__. I L

0 ii~|

- " V . . ~ -

.roy

"l

i,......I

o
0.5
2

I
9

"

]l

,I

10

11

'~ ] . . . .

12

1S

IZ

t3

F-

"~. 0.7
!

I
[

L) 1_5

:I.o
8

p
];If;..~.

Yc~c~]c~I D]~me+er
Lengf~
GI{OMETR1C 1)ROP()RIIONS ~1~" r i t e "I'I'RNIN(; I'A'I'H.

r/L

l0

II

I/ROI~I "I'EST I)ATA FOR NORMAL ~t,;AGOIN(; StlIP5 (.AND ~I{~DIgLS) AT VARI()L'S ~PEIFDS s~.ND t)~UI)DI~R ANC*I.ES

>-~
e)
@,

u)
A2
4--

0
0
0
O:Z

to
r"
4-

09

..A

tD

r'o
.4-0

p')

)0

'D

I~
0
"+-

0_~.

x:
L~

:Z

,C
I

'~)B

t
~B

IB

~z

=].

6',

TWL=3

~J

11

-'2d';1

TD/L= 2

Execufe

7_

TD/L= 4~

~o~
/30
,,

\
s

,/

~
4

~z

\\Zl\\\

3
.

(/Z

-rD/t :e3~4

~Z
5
TD/L= IO "8

F I G . ~ : . - - T U R N I N G ~ H A R A C T E R I S T I C S AND ~ P E E D RECOVERY

Note:
(1)
(2)
(3)
(4)
(5)
lines).

T h e t u r n i n g p l o t s h a v e p r o p o r t i o n s in accordance w i t h t h e m e a n c u r v e s of Fig. 3; t h e speed recovery p l o t s are c o m p u t e d .


T h e t u r n s shown represent a geometric progression in t a c t i c a l diameter, each t u r n being half a g a i n as large as t h e preceding t u r n .
Spots show successive positions of t h e ship a t u n i f o r m t i m e i n t e r v a l s equal to t h e t i m e required t o t r a v e l 2 ship l e n g t h s a t V a / ' ~ / L = 1.
S t e a d y - t u r n i n g speed is a s s u m e d a f t e r t h e t h i r d t i m e i n t e r v a l .
C e n t r a l angles show r e l a t i v e rates of change of h e a d i n g in s t e a d y t u r n i n g .
A u x i l i a r y speed-recovery p l o t s show r e l a t i v e distances required for regain of speed ( d o t t e d portions), a n d r e l a t i v e distances r u n in four t i m e i n t e r v a l s after centering rudder (total l e n g t h of

t~

292

ON THE TURNING

AND STEERING

ships, on the other hand, which b y comparison


travel at low speeds for their lengths, the centrifugal force does not exceed a few per cent of the
weight, as m a y be seen in Fig. 2. Thus the airplane designer need not strive for very short turning, since centrifugal force will govern in any case,
whereas the ship designer is not governed b y this
limitation and still holds the initiative.
A more promising approach to the question of
what constitutes really good turning in a ship can
be made, beginning with a somewhat more detailed consideration of Fig. 1. The tactical diameter there shown is three lengths (TD,/L = 3),
as previously noted. But, the speed reduction is
large (I~/Va = 0.5S), and the transient phase of
the turning p a t h is long (A/TD = 0.99). Obviously, a lesser speed reduction and a shorter advance would be desirable i m p r o v e m e n t s - - i f they
could be obtained. I t is therefore of interest to
look into the variation of these two additional
characteristics of the turn, and into their functional connection (if any) with the ratio of tactical
diameter to length.
T h e charts of Fig. 3 were prepared with these
ends in view. T h e y show d a t a deduced from the
tactical trials of a number of ships and models representing a variety of types of actual seagoing
vessels of generally normal design. There is a
separate chart for each of the two ratios mentioned in the preceding paragraph,
speed reduction . . . . . . . . . . . . . . . .
advance to tactical diameter . . . . .

VJ V~,
A/TD,

and one for the less i m p o r t a n t ratio


transfer to tactical diameter . . . . .

T/TD.

All of the charts are plotted to the same scale of


abscissae; namely, the ratio
tactical diameter to length . . . . . .

TD/L.

T h e wide lateral spread of the charts is due mainly


to the fact t h a t d a t a are given for three rudder
angles, usually 35, 25 and 15 degrees, and these are
naturally distributed across the charts since, in
a n y given case, a reduction of rudder angle results
in an increase of tactical diameter. Also, the
different designs shown have different minimum
ratios of tactical diameter to.length (at m a x i m u m
rudder angle), and in m a n y cases data are given
for two or more speed-length ratios. Thus the
charts do not in any way indicate the relationship
between tactical diameter and rudder angle for
individual designs; they merely compare the
various other factors involved in turning, on the
basis of tactical diameter.
I t will be seen t h a t the form of plotting involves

OF SHIPS

only dimensionless ratios, so t h a t differences of


absolute size, whether of ship or turn, are eliminated, and the several characteristics are compared on a purely geometric basis. A speed ratio
m a y not ordinarily be thought of as a m a t t e r of
geometry; in this instance, however, interest
centers not so much in the relative speeds themselves as in the relative distances traversed along
the p a t h in equal time intervals (the relative distances between the successive positions in Fig. 1,
for example), and these are clearly a m a t t e r of
geometry.
I t is apparent at once t h a t Fig. 3 effects reasonably good correlation of the test data, and thus indicates general functional relationships to a fairly
good first approximation. I t is true, of course,
t h a t the scatter of individual points from the
mean curves drawn through them is often as much
as 10 per cent, and in some cases more. B u t this
is scarcely surprising in view of the variety of the
designs included, and the fact t h a t the points represent tests at various rudder angles and speedlength ratios, plotted without regard to the values
of these variables. If there is anything surprising
about the charts, it would seem to be t h a t the
scatter is not greater than it is. As a m a t t e r of
fact, however, discrepancies in turning characteristics of the order of 10 per cent are well within
tolerable limits for most practical purposes, so t h a t
the mean curves m a y be said to provide a reasonably good picture of the interrelationship of the
several ratios for the average seagoing vessel of
generally normal design.
I t appears, then, t h a t for normal seagoing vessels, the speed reduction in turning (l/~/V~) and
the relative length of the advance (A/TD) are
governed primarily b y the ratio of tactical diameter to length (TD/L), and are not materially
affected b y differences of design within the limits
of what m a y be called normal designing. Thus
the large speed reduction shown b y Fig. 1 (to
about 60 per cent of the approach speed) and the
considerable length of the advance (about equal to
the tactical diameter) have to be considered the
more or less inevitable consequences of reducing
the tactical diameter to three lengths, while Fig.
3 shows clearly t h a t further reduction of the tactical diameter becomes increasingly costly in respect to both speed reduction and advance. I t is
therefore natural to enquire whether, under these
circumstances, further reduction of the tactical
diameter would be worth while, even assuming it
possible.
Fig. 4 helps to throw additional light on this
m a t t e r b y giving a visual impression of a series of
turns corresponding in their proportions to the

ON THE TURNING

AND STEERING OF SHIPS

mean curves of Fig. 3. The series comprises a


geometric progression in the value of the ratio of
tactical diameter to length beginning with a minimum of two--in other words, with a value well
below the minimum ordinarily reached in actual
ships at the present time. I t will be seen that as
this ratio diminishes there is progressive improvement, all the way down to the two-length turn, not
only in the tactical diameter itself, but also in
(a) the advance (itself, not in the ratio A/TD),
(b) the total angle swung through in a given
total time from "execute,"
(c) the time rate of change of angle (heading) in
steady turning.
In these respects, then, a minimum turn of two
lengths would obviously be superior to .one of
three lengths (assuming, of course, that incidental
questions such as excessive heel or vibration did
not become prominent).
In respect to the speed reduction, the twolength turn appears at first sight to be distinctly
inferior to the three-length turn. Here the difference is large when expressed in the form used in
Fig. 3; namely, VJ Va = 0.39 for a two-length
turn, as opposed to Vt/I~ = 0.58 for a threelength turn--practically in the ratio of 2 to 3.
But, as has already been emphasized, relative
speeds are not in themselves of much interest; of
much more importance are the relative overall
distances covered in equal time intervals. In the
present instance, a good index should be the relative distances run on straight course during the
period of speed recovery, supposing the rudder to
be centered after the ship has turned far enough
to bring about the entire reduction to steadyturning speed, and the throttle settings to remain
unaltered. These relative distances can be estimated with sufficient accuracy for the present purpose b y assuming that the portion of the thrust
that has been working against the increase of
drag due to turning becomes available to accelerate
.the ship to its original approach speed, and going
through much the same process as is discussed
later for the computation of speed reduction.
The results of estimates made in this way, on
the basis of the mean curve of relative resistances
in Fig. 7, are shown in Fig. 4 by auxiliary plots
placed adjacent to th.a corresponding turning path
plots. The auxiliary plots indicate distances run
in successive time intervals equal to the time intervals used on the turning path plots, and bring out
the important fact t h a t differences in the overall
distance run in a given overall time are not much
more than one ship-length even when the two
extreme turns are compared, and less than half a
ship-length when three-length and two-length
turns are compared. I t is clear, then, that while

293

the greater speed reduction of a two-length turn is


admittedly a disadvantage, its ultimate effect is
not very great because of the rapidity with which
speed is regained.
One other point m a y be mentioned in comparing three-length and two-length turns; namely,
the relative centrifugal forces developed. If the
speed reductions were the same, the centrifugal
force would obviously be about 50 per cent
greater in the smaller turn. On account of the
much greater speed reduction, however, the
smaller turn is actually found to develop less ~
centrifugal force with the same approach speed.
On the whole, therefore, the characteristics of
a two-length turn appear to be superior to the
characteristics of a three-length turn. And it
should not be forgotten that the mere availability
of a two-length minimum turn would not make its
use m a n d a t o r y - - t h e maximum rudder angle is
ordinarily employed only in emergencies.
It might be argued that, if a two-length turn is
superior to a three-length turn, a turn of less than
two lengths might be even more desirable. However, in spite of the general rapidity of speed recovery already noted, the speed reduction itself
must obviously become prohibitively great if the
tactical diameter is reduced far enough. Hence,
bearing in mind that few existing vessels can turn
in less than about three lengths, a tactical diameter of two lengths certainly seems acceptable
as a practical minimum to work toward, at least
for the present. Parenthetically, it may be mentioned that the minimum turning circles even of
automobiles are not ordinarily less than about
twice the overall length (or three times the wheelbase).
I t has been emphasized throughout the foregoing discussion of generalizations regarding the
geometric characteristics of the conventional
turn that the charts of Fig. 3 are based on the
experimental evidence of normal seagoing vessels
only. This means that special types, such as tugs,
trawlers and the like, have not been included.
The term "normal seagoing vessel" is used, however, in a fairly broad sense, to include any vessel
which would obviously be so classified, without
regard to differences of design details. Thus
Fig. 3 includes data for single, twin and quadruple-screw vessels, having various displacementlength ratios and differing designs of stern, with
and without deadwood, and fitted with balanced
or unbalanced rudders. Under these circumstances, the relatively good correlation of the
geometric data afforded by Fig. 3 implies a certain
broad similarity among all normal seagoing vessels.
I t does not require much stretch of the ima-

294

ON T H E T U R N I N G A N D S T E E R I N G OF S H I P S
o,
d

_3

(a)
5hip

J-fitc, ss Y a c h f

i
i

-<

.....

Lengfh,L

<- . . . . . . .

. . . . .

-/-/i

-jI~-

Lengfh, L

(c)
S f a r Cle~ssYoch+
L
)

" x
"~

I
I

(d)

!),i

Transpo-+

Fit;.
Showing diagrammatically

,~.-

Mrplane

-T&'PICAL HULL

FORMS

a p r o g r e s s i o n f r o n t t h e t m i f o r m d r a f t of t h e n o r m a l s e a g o i n g v e s s e l to t h e s e p a r a t e w i n g a n d t a i l of t h e a i r p l a n e

gination to explain this broad sinfilarity. When


viewed in sufficient perspective, the designs of all
normal seagoing vessels are, after all, very much
alike. Fig. 5(a) is a typical example of the general
type of design to which practically all seagoing
vessels conform. The keel is straight and generally parallel with the water surface, the load waterline is sharply pointed at the bow and usually
pointed at both ends, and the sides are roughly
vertical in way of the load waterline. The draft is
around ~ 0 to }~0 of the length, and the beam
around ]/{ to }'~0 of the length. It will be seen that
the extreme variations of these dimension ratios
result in variations of form which, in a broad
sense, are only trifling. There are no differences
between various ship forms which remotely approach in magnitude the differences between an
average ship form and the forms of the J-class
sailing y a c h t in Fig. 5(b), the Star-class sailing
yacht in Fig. 5(c), or the transport airplane in
Fig. 5(d).
It is not contended that seagoing vessels should
be made to look like sailboats or airplanes. Their

present form is dictated b y m a n y considerations,


most of which are of necessity weighted more
heavily than the turning characteristics. It is
merely pointed out that, given their present
general similarity of form, certain consequences
with respect to the turning characteristics, such as
those shown b y Figs. 3 and 4, are more or less
inevitable.
I t may be shown, for instance, that the consist-"
ency in the speed reduction data of Fig. 8 refleets directly a uniformity of pivot point location
which seems to be characteristic of ship forms.
The pivot point is associated with steady turning,
and is that point on the axis of symmetry of the
vessel at which a line drawn horizontally and at
right angles to the axis will pa~s through the center
of the steady-turning circle. Various observers
have noted the marked tendency for the pivot
point to fall at something like ]/~ of the length
ahead of the center of gravity [2], and this average location is fairly well confirlmed b y the tests on
which Fig. 3 is based.
Now, referring to Fig. 6, mad noting that in

ON T H E T U R N I N G A N D S T E E R I N G OF SHIPS

DrcLcjCornponenf

, ~Cenfr,fu~ol Force,Mv~/r

(MvVr) si= ~ \
Propeller

Pafh of C.G.

M= Mass

--pivo+

poin+

FIG. 6.--DRAG COMPONENT OF CENTRIFUGAL FORCE IN


~TI~ADY TURNING

steady turning there is no unbalanced rotating


couple (the ship not being subject to angular
acceleration), it will be seen that, other things being equal, the location of the pivot point determines the magnitude of the drag component of
the centrifugal force, b y determining the magnitude of the drift angle at the center of gravity.
Since this drag component constitutes a deduction from the available propeller thrust, it reduces the speed of the ship, and any change in its
magnitude will alter the a m o u n t of the speed reduction.
The possibility suggests itself, then, of calculating the speed reduction as a function of the diameter-length ratio. T h e speed reduction should
not be materially affected b y differences of displacement-length ratio since the centrifugal force
is directly proportional to displacement and both
resistance and thrust ordinarily v a r y approximately in proportion to the displacement. As a
m a t t e r of fact, the curve on the speed-reduction
chart of Fig. 3 is a computed curve, and not simply
a mean curve drawn through the points. I t is
based on (1) centrifugal force components determined as in Fig. 6, with average pivot point locations from test data, (2) the mean curve of relative resistances per ton shown in Fig. 7, and (3)
one or other of the assumptions, t h a t any increased thrust attributable to greater propeller
slip at reduced ship speeds (the throttles remaining
unchanged) is approximately offset b y increased
drag of the hull and rudder caused b y the turning,

295

or t h a t these two secondary forces are negligible


in magnitude. T h e general agreement between
the computed curve and the observed points is
evidence of the reasonableness of this procedure,
and of the roughly fixed position of the pivot
point.
I t seems probable t h a t some roughly parallel analytical t r e a t m e n t could be "worked out to account for the shape of the curve of advance to
tactical diameter in Fig. 3; this has not y e t been
attempted.
LATERAL FORCES IN

STEADY T U R N I N G

I t m u s t be emphasized once more t h a t nothing


which has been said so far is to be construed as
implying t h a t all normal seagoing vessels can turn
in minimum circles corresponding to the same
ratio of tactical diameter to length. F a r from it.
M a n y present-day ships have nfinimum tactical
diameters as large as four or five lengths or even
larger, so t h a t there is plenty of room for improvem e n t if all ships are to be put on a par with the
better current ships, and still more room for improvement if minimum turns of the order of two
lengths are to be realized.
A striking illustration of the differences which
exist today between actual designs is afforded b y
Fig. 8, which is an extract from Fig. 3. Here are
two models of existing ships, designated A and B,
the second of which, with the same m a x i m u m
rudder angle as the first (35 degrees), has a minim u m diameter-length ratio more t h a n twice as
large--6.0 as opposed to 2.4. When both are
turned in circles corresponding to the same diameter-length ratio, the geometric aspects of the
turns are seen to be very nearly the same, as would
be expected from previous discussion; very different rudder angles are required, however, to produce the same diameter-length ratios.
The difference between the m i n i m u m diameterlength ratios of these two ships obviously cannot
be accounted for b y a difference in the m a x i m u m
rudder angles, since these are the same (35 degrees). Nor can it be attributed to a difference
in the relative rudder areas alone, since, expressed
as fractions of the lateral plane areas, the rudder
areas are not far from the same. These fractions
are 0.025 for ship A and 0.023 for ship B, a difference of 9 per cent in relative rudder area to be sure,
b u t a difference of only 0.2 per cent of the respective lateral plane areas. Evidently, then, other
factors m u s t contribute heavily to the overall
result. T h e m o s t obvious are the hull form, together with a n y m u t u a l interactions between the
hull, the rudder and the propellers.
There is nothing v e r y surprising, or essentially

296

ON THE

TURNING

AND STEERING

OF SHIPS

I00
I

I co#

:
i "~I

-0

!~

-f-

80

Dracj ornpo~nen}of\
-o

Cen+ri#ugal Forca
P e r Ton at TD/L=

ok.
u9

-6 6o

od
~6 4O

Typ'mal
Resisfance Curv

EL
20

cO

i~.."
c~

4
- - '

OL-" ~ ' ~ ' ~

I
20

)'I

40

60
Per CenJ- of Normca', Speed

FIG, 7.--I%{ETHOD OF C O M P U T I N G SPEED R E D U C T I O N IN TURNING;


CURVE FOR FIG. 3

80

Io0

U S E D FOR M E A N

Typical curves of straight-course hull resistance and drag component of centrifugal force
Resistance curves are for Standard Series hulls, as follows:
Ships
I
II
III

DisplacementLength Ratio
60
100
170

BeamDraft Ratio
2.5
2.5
2.5

Prismatic
Coefficient
0.60
0.65
0.70

Normal SpeedLength Ratio


1.2
1. (I
0.8

Drag curves are based on pivot point locations from test data. Drag curves are parabolas,
since centrifugal force varies as I:% If total propeller thrust is assumed to remain constant
during turning, and straight-course hull resistances are assumed to apply, then the drag
component of centrifugal force reduces the thrust available for forward speed. As an example, a ship having the mean resistance curve indicated has a steady-turning speed 71
per cent of its normal speed, when turning in a steady circle with T D / L = 4.

new, in this conclusion; 35 degrees has come to


be g e n e r a l l y a d o p t e d as a b o u t the m a x i m u m
p r a c t i c a b l e r u d d e r angle, a n d the r u d d e r areas of
n o r m a l seagoing vessels are rarely m a d e as small
as 0.015 or as large as 0.030 of the lateral p l a n e
a r e a s . ~ Hence, the c o m p a r a t i v e l y large differences
in the m i n i m u m r a t i o of tactical d i a m e t e r to l e n g t h
which are f r e q u e n t l y found m u s t necessarily be
p u t d o w n to the hull design, or to the a r r a n g e m e n t
of t h e a p p e n d a g e s (i.e., i n t e r f e r e n c e effects),
r a t h e r t h a n to the r u d d e r alone.

T h e c o m p a r a t i v e analysis in Fig. 9 of c e r t a i n
tests of the same models used for Fig. 8 brings o u t
forcefully the relative influence of the hulls in
this instance. T h e s e models are a d m i t t e d l y
s o m e w h a t extreme examples, i n a s m u c h as model
A t u r n s very well a n d model B q u i t e poorly.
Both, however, are models of existing ships;
t h e y are properly referred to as n o r m a l seagoing
types, a n d t h e y are in no sense freaks. As a
m a t t e r of fact, b o t h h a v e t h e same t y p e of stern,
m o d e r a t e deadwood a n d s p a d e - t y p e r n d d e r s well

ON THE TURNING AND STEERING OF SHIPS

297

...,..,,~ ~.....-.-~ - ~ "


/-,.,

./,o

> 0.6
0.5 /
O.4
I.I

Rudder
Angle
~
Model A

1.0
c~ 0.9

I0o

~. 0.8
0.'7

15

0.6

Angle HodetB

0.5
O.1
,.-., 0.6

0.4

8
TD/L

lO

11

IZ

IS

Fla. 8.--TURNING PATH PROPORTIONS FOR MODELS A AND ]3 (EXTRACT FROM FIG. 3)

separated from the deadwood. In general, they


may be said to be quite similar in respect to parent
hull form: the most obvious differences are in
the displacement-length and beam-draft ratios;
that is, in matters of hull proportions. The
analysis is made at the same ratio of steadyturning diameter to length, for the same approach
speed-length ratio, and on the basis of models of
very nearly the same length. Direct comparison
is therefore possible at each step.
From the particulars of the two models, listed
below the force diagrams, it will be seen t h a t
model A has roughly two and one-half times the
displacement (hence displacement-length ratio)
of model B, but, because of its greater beam-draft
ratio, relatively little more lateral plane area of
hull (27 per cent); the rudder areas differ, as
previously noted, b y only 0.2 per cent of the
respective lateral plane areas.
From the test data given, it will be seen that
widely differing rudder angles (6 degrees for model
A and 35 degrees for model B) are needed to produce the common ratio of turning diameter to
length (5.9) selected for study, and that widely
differing lateral rudder-force components were
observed under these conditions, this force component being practically zero for model A (actu-

ally slightly negative). The lateral rudder-force


component is the lateral force transmitted through
the rudder stock while the self-propelled model
is turning freely in steady motion, as measured b y
a special dynamometer mounted in the model.
The pivot point positions, as deduced from the
kinematic test data, are in both cases somewhat
forward of the previously mentioned mean position of 1/~ of the length ahead of the center of
gravity, but do not differ greatly from each Other.
The analysis as a whole is confined to lateral
(athwartships) force components. The first step
is to determine the hull force component, or "lateral resistance" of the hull. This is readily
done, since, in steady turning, the hull force component must be equal and opposite to the resultant
of the rudder force component (in this case measured) and the centrifugal force component" (exactly determined b y the speed, turning radius and
displacement); the resultant moment about the
center of gravity is zero. The procedure is the
same as that adopted by Hovgaard some years
ago [5], except that measured values of the rudder
force component are substituted for estimated
values, and the computed hull force components
should therefore be correspondingly more reliable.
Referring to the results of this step, as shown

298

ON THE TURNING AND STEERING OF SHIPS

0.70 Lb.

9 Lb.

~11-~- o.oos L

0.4z kb.[

t0.s0Cb.

0.0~L ~ ~ _ ~
----0.43
Model A

5.63
1.03
0.27
53.50
1. 405
O. 0348
138
5.5
3.81
0. 512
0. 466
0.02,5
3.19
2.65
0.83
0.80
0.66
33.2
5.9
6
--0.01
0.70
O. 69
0. 005L, forward
0.43L
8.3

O. 0726
0. 0088

FIG

9.--COMPARATIVE

k--~-

- - 8 . 5 % - _~

PARTICULARS
Dimensions
L e n g t h , L, ft
B e a m , B, ft
Draft, H , ft
D i s p l a c e m e n t , W, lb
Area of lateral plane, L A , sq ft
R u d d e r area, RA, sq ft
Ratios
D i s p l a c e m e n t - l e n g t h r a t i o , A/(L.'100) a
L e n g t h - b e a m ratio, L / B
B e a m - d r a f t ratio, B / H
e.g. aft of b o ~ < l e n g t h , G.'L
R u d d e r post aft of e . g . / l e n g t h , b
R u d d e r a r e a / l a t e r a l area, R A / L A
TUST DATA
Ap'proach speed, I;~, ft per sec
T u r n i n g speed, I't, ft per see
Ratio, I ' t / l ' ,
A p p r o a c h s p e e d - l e n g t h ratio, V , / ' V ~ T u r n i n g speed-length ratio, 7v~,/x/ZSteady t u r n i n g diameter, D, ft
Ratio, D / L
R u d d e r angle, deg
R u d d e r force (measured), lb
(Mv2,,"r) sin 0, lh
Hull force, lb
Hull force location, from e.g.
P i v o t point, a h e a d of e.g.
Drift angle, 4,, at e.g.
HULL COEFFICII~NTS
" L i f t " coefficient of hull force,
hull force
(CLIH = (LA) (p/'2)vt ~
Lift coefficient/drift angle,
(CL)H/d)~.~.

~~.~-- - 0.40 L-- -->


tdodC B

-/.6%-->

5.88
0.63
0.21
23.50
1. 105
0. 0253
54
9 4
3.0
0..522
t). 451
II 023
3.26
2.67
0.82
0.80
O. 66
34.6
5.9
35
0.12
O. 30
I). 42
I). 13L, aft
0.40L
7.8

t). 0548
~). ()~)71)

ANALYSIS (IF TURNING TESTS MODELS A AND B AT SAME D / L


STRICTED TO LATERAL COMPONENTS OF FORCES

AND Vt/%'/L.

ANALYSIS RE-

ON THE TURNING AND STEERING OF SHIPS

299

(b')
Yaw

(a)
S+ectdy- Turn[no3
D/L=3
FIG. 1 0 . - - M O T I O N OF H U L L IN STI~AD TURNING AND IN Y A W

in the diagrams, it is obviously no exaggeration to


say that:
Model A turns in spite of its rudder,
Model B is forced around b y its rudder.
The difference is clearly attributable to the differing positions (i.e., centers of pressure) of the hullforce components, and not at all to their differing
magnitudes. The fact that the hull force center is
some 13 per cent of the length abaft the center of
gravity in model ]3 means t h a t the rudder force
has to be about 40 per cent of the centrifugal
force; the fact that the hull force is practically
at the center of gravity in model A (actually
slightly ahead) means that the rudder force can
be practically zero (actually slightly negative),
even though the centrifugal force is well over
twice as large as in model ]3.
From this evidence, the general conclusion is
inescapable that the hull form, through its control
over the hull-force center, plays an extremely important part in fixing the steady-turning characteristics, and can in no sense be considered
secondary in importance to the rudder. This
conclusion is emphasized because, on the whole,
in the study of turning over a period of years more
attention seems to have been devoted to rudders
as such than to hulls as such, even though the
rudder is basically an appendage.
It is interesting to carry the analysis of the hull
forces somewhat farther. In particular, the
question arises of accounting for the observed fact

that model A develops some 65 per cent more


lateral hull-force component than model B, with
much the same pivot point location and hence
with much the same drift angles at all corresponding longitudinal points. One factor is, of course,
the greater lateral plane area of model A; this
suggests dividing the hull-force components b y
the respective lateral plane areas, for comparative
purposes. From this it is a natural step to divide
them also b y the dynamic pressure (p/2)v 2, to
produce a dimensionless coefficient for the hull
force which has the same form as the lift coefficient of an airfoil.
The use of "lift coefficients" to define cross
forces on hydrodynamic or aerodynamic bodies
other than airfoils has become a commonplace in
recent years. Their use in the present instance
would not call for special comment, then, were
it not for the curvilinear motion of the hull.
Lift coefficients are ordinarily associated with
"angles of attack" which, in rectilinear motion,
are fixed throughout the length of the body, as in
Fig. 10 (b). For the case in point, because of the
curvilinear motion, the angle of attack varies
throughout the length, the variation being considerable when the ratio of turning diameter to
length is small, as in Fig. 10 (a). This difference
may cause a certain complication of thought, but
it is not actually very serious, first because it does
not affect the computation of the coefficient itself,
and second because the computed coefficient can

300

ON THE TURNING
Moctel A

O.IZ L--~

2.33 Lb.

I.] 3

Lb.

....

(ss),o.~o

D/L = 2.3

Lb.

iI

0.05 L-->- i<-I

I
l.lro Lb. I 1.02 Lb.
I

AND STEERING

OF SHIPS

subsequently be related to an arbitrarily selected


drift angle for comparative purposes. As a matter of fact, the use of lift coefficients for the present purpose has been suggested before now [6],
and is more or less implicit in H o v g a a r d ' s analysis [5].
T h e lift coefficients for the models under consideration are given at the b o t t o m of Fig. 9, together with the ratios of these coefficients to the
drift angles at the centers of gravity (practically
at the mid-lengths). F r o m these d a t a it will be
seen t h a t the additional lateral plane area of
model A, as compared with model /7, is insufficient to account for the very much greater hullforce component generated in this model at practically the same speed and with nearly the same
drift angle; the lift coefficient of model A, expressed as a ratio to drift angle, is some 25 per
cent in excess of t h a t of model B. Now the parent
forms are much alike, as previously noted, and the
difference of lateral plane area has been eliminated b y taking coefficients. T h e next most outstanding difference between the models lies in the
beams and the consequent fullness of the waterlines, model A having about 70 per cent more
beam than model B on the basis of equal lengths.

=o.o,oo

(15)~ 0.14 Lb.

4.

--~i<--0.005 L

MC'PC

(),-o.o~ Lb.
~

~':--;-~-f___g ~ ( c j l ~

....

D / L : 5.9

,~',

0.42 LO. I "t'0.30Lb.

;r--

"[ I

....

DzL=s.9

Note:
Tur'n~r~S Speed Sctme cts
in F;c3ure 9
0.I2 b ~

r~

D/L---Z8
Cs),s '~

o.g. = o.oo6~

(.~s),o.os u~.

FIG. I I . - - E x T E N S I O N

OF FIG. 9 TO OTHER TURNING DIAMETERS.

DATA FOR D / L

: D/L=I2.6
/

{.g-----.= o.oo67

= 5 . 9 A R E SAME AS ON FIG. 9

ON T H E T U R N I N G A N D S T E E R I N G O F S H I P S
The evidence of these two models is insufficient,
of course, to justify a general conclusion to the
effect that, other things being equal, an increase
of beam-length ratio produces an increase of hullforce coefficient. This seems a good working
hypothesis for the present, however; if for no
other reason, then simply on the ground t h a t first
order differences ought to be considered first.
Fig. 11 broadens the scope of the analysis of
Fig. 9, by showing what happens at other ratios
of steady-turning diameter to length, resulting
from the use of other rudder angles. The data
for each model are for the same steady-turning
speed as in Fig. 9, hence for somewhat differing
approach speeds. Use of the same speed in turning eliminates any possible influence of speed on
the characteristics shown, even if the relative force
magnitudes are somewhat distorted in comparison
with those which would occur with constant approach speed.
The ultimate points of interest in Fig. 11 are
the indications that, with decrease of diameterlength ratio (increase of rudder angle) :
(1) The ratio of hull lift coefficient to drift angle
at the center of gravity increases.
(2) The center of pressure moves aft.
These indications are consistent for the two
models; they are rather more marked in model A
than in model B, but this may be only because the
range of diameter-length ratios for this model
embraces smaller values.
STEERING
It is an old adage that a good-turning ship is
hard to steer and that a good-steering ship is hard
to turn [1]. The implication is clear that good
turning and good steering are essentially antagonistic qualities. The definitions often used,
that turning is the process of altering course and
steering the process of maintaining course, tend
also to suggest a certain antagonism between the
two. Under these circumstances, it becomes important to consider steering characteristics along
with turning characteristics, before reaching
general conclusions regarding either one. I t is
particularly important to do this as additional
background for judging whether or not tactical
diameters of the order of, say, two lengths are
practicable.
It is somewhat more accurate to define steering
as the process of "correcting" course than as the
process of "maintaining" course. In practice,
steering is a m a t t e r of bringing the ship back to
the desired course, after a departure from this

301

course has been noted. The actual path traversed


is then, in reality, a succession of the initial portions of turns of very large diameter, involving
small rudder angles. From this point of view,
turning and steering are seen to be much more
similar in nature.
For the ordinary case in steering, where departures from the desired course are reasonably
small, the actual path traversed b y the center of
gravity will usually be very nearly a straight line;
in other words, changes of heading will usually
be a much more important feature of the motion
than departures of the center of gravity from a
rectilinear path. On this account, it is not unreasonable to explore the practicability of going at
once to the limit and of studying steering as a
problem involving purely rectilinear motion of the
center of gravity, in combination with appreciable
angles of yaw; a relatively simple model-test
procedure is thereby made possible. T h e same
point of view is ordinarily adopted in tests of the
directional stability of airplane models; it is
obviously applicable, also, to studies of the initial
instants of turning, before an appreciable curvature of the path has developed.
Fig. 12 shows the results of tests along these
lines, carried out on the same two models, A and
B, already considered in the discussion of steady
turning. Moments tending to cause rotation are
given as functions of yaw angle, for two series of
rudder angles--one applying to steering and the
other to the initiation of turning. I t should be
understood that in these tests the self-propelled
model was constrained to move in a straight line
under the carriage of the towing tank, and was
held at predetermined angles of yaw b y the dynamometer equipment which measured the yawing
moment. The angles of yaw investigated a r e
much larger than occur in practice; they were so
chosen in order to develop unmistakable trends.
The speed for the tests shown was the same as the
approach speed in the turning tests of the same
models, in Figs. 9 and 11.
I t has been seen that model A has a much
smaller minimum tactical diameter (with maximum rudder angle) than model B. Hence, if the
rule held that poorer steering goes with better
turning, model A would be expected to have
poorer steering characteristics than model B.
In the face of this, however, Fig. 12 implies a
surprising similarity in the steering characteristic of the two models. Both are seen to be
directionally unstable with their rudders amidships, which means that unless correcting (left)
rudder is applied, any accidental (starboard) yaw
introduces a moment (to starboard) which tends
to further increase the yaw angle. Thus, neither

HodeI A

Mod~l B

tomgifud{~ol Momem+ of inertia = 3.42 ~lug F+ a

bO

tomgiacudlno/ Morneu~o~ Iner~i~ = I.$8 Slucj F~r~

(K=o.2s L3

+t+
!S

So

..A

0 o

~-+,3

D,

X5

5 ii.

~o

0c -2!'--

V9 (

4-

2
g

9o

on
c

20

+ 2 _ _
RighRudder
om
( ,Irni+ia+[om o Turning)

30
25 ~
20 o

+-

b~
i0 o ed
I5

5~

+~

>

Oo

L
@

~J

9o

12

15 o

c
a~
m
,)
.I2

x~
<

ff
r~
<.
~c

Rudder

te++

(S}eer;ng)
{

0~

8o

i~o

a~o

'*'~w ,Angle t o S f G r b o o r d
V = 3 3 9 F%. Per S e c .

(I)

(2)

"r'cxw

V -5.2_6

6-degree correcting rudder offsets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


12-degree correcting rudder at 2-degree y a w causes correcting a n g u l a r acceleration of . . . . . . . . . . .
25-degree rudder at 0-degree y a w causes a n g u l a r acceleration to initiate t u r n i n g of . . . . . . . . . . . .
Fro,

12.

YAW I~IoMENT

Angl~

VFRSUS

YAW

ANGLE

FOR VARIOUS

RUDDER

ANGLES,

8o
+o

Far, P e r

Sec.

~iodel A
1.8 deg y a w
0.09 rad./sec."0.33 rad./secY
~IODELS

17o

S+c~rboc~rd

Z AND B

NIodel B
2.2 deg yaw
0.I0 rad./sec. 2
0.30 r a d . / s e c 2

ON T H E T U R N I N G A N D S T E E R I N G OF S H I P S

505

model has an inherent tendency to steer itself. steady-turning circle. Model A is so extremely
But, as will be seen from the data under the dia- sensitive in its response to the rudder at small
grams, a correcting rudder angle of just about angles t h a t the actual process of steering in openthree times the amount of the yaw angle will sea conditions might well be more difficult.
neutralize the unstable moment in both models.
There is thus some question as to whether the
Poin[s 1 show, for example, that the application results of straight-line tests like those of Fig. 12 are
of 6 degrees correcting rudder in either model will a sufficient index of the steering qualities of ships.
offset any tendency for a yaw of about 2 degrees The whole matter wilt bear further study. The
to increase. This rough three-to-one ratio holds implication is clear that, to completely evaluate
over a wide range; it is a rather startling indica- the steering characteristics, it may become necestion of uniformity in the steering characteristics sary to take into account the turning characterin view of the big difference in the relative turning istics at very large diameter-length ratios, in
characteristics.
addition t o - - o r even, conceivably, instead o f - Nor is this all. If the same over-correction in the directional stability characteristics on straight
rudder angle is applied in both cases, say 12 de- course. For the present, however, it seems better
grees left rudder at 2 degrees starboard yaw, sub- to concentrate attention on the directional stastantially the same angular accelerations tending bility, and to explore fully this relatively simple
to return the models to the original course can be approach. This point of view is adopted in the
expected, Points 2, because the restoring moments remainder of the present discussion.
introduced are nearly in proportion to the probReverting, then, to Fig. 12, the fact must not
able yawing moments of inertia (to a first approxi- be lost sight of that for any given combination of
mation, all normal seagoing vessels have longi- yaw and rudder angles the moment created by the
tudinal radii of gyration, k, of roughly one-quarter rudder of model A is much greater than the moof their lengths). Further, the indications re- ment created by the rudder of model B. The ratio
garding the initiation of turning are entirely con- between the moments is, in fact, much the same
sistent with the indications regarding over-correc- as the ratios between the displacements and the
tion in steering. Supposing, for instance, that, to moments of inertia, and this is evidently the
start a turn, 25 degrees of right rudder is applied fundamental reason for the indicated similarity
with no yaw, it will be seen that the angular ac- of steering qualities in the two models. Thus, a
celerations of the two models are again much the conclusion that good turning is not necessarily
accompanied b y poor steering, stated in just those
same, Points 3.
Fig. 12, then, does not indicate that model A is words, m a y easily be misinterpreted. A better
inferior to model B with respect to either steering statement is perhaps, that:
or the initiation of turning. On the basis of this
(1) A good-turning hull needs powerful rudder
analysis, the conclusion is therefore strongly sug- action for steering, but not for turning.
gested that good turning qualities (small tactical
(2) A poor-turning hull needs powerful rudder
diameters) are not invariably purchased at the action for turning, but not for steering.
expense of good steering qualities, since model A
Further light is thrown on the matter by the
has been seen to have the better turning qualities. data in Fig. 13. The yawing tests represented in
This is not an entirely satisfying conclusion. Fig. 12 included measurements of resultant lateral
It is true that the full-size prototypes of both force components and lateral rudder-force commodels A and B are being successfully operated ponents, as well as yawing moments. With these
as designed, so that neither one of them can very additional data, it is possible to make an analysis
well be considered markedly deficient in steering of lateral hull-force components similar to that
qualities. On the other hand, some difficulty made for the case of steady-turning in Fig. 9.
seems to have been experienced at first in steering This is done in Fig. 13 for the models with rudders
the prototype of model A, until her crew got used amidships. No problem arises in this instance
to her, and the fact remains that in the models it regarding the angle of attack, because the motion
was very easy to make model B run on straight is rectilinear and the geometric angle of attack
course (without restraints) b y accurately center- (identical with the angle of yaw) is therefore coning the rudder, whereas it was nearly impossible stant throughout the length, as in Fig. 10 (b).
to accomplish this with model A. This last dif- However, for the sake of consistency, the analysis
ference is no doubt directly associated with the is carried out for yaw angles equal to the drift
much greater sensitiveness of the turning diameter angles at the centers of gravity in Fig. 9. These
to the rudder angle in model A; as shown b y Fig. are, of course, rather larger yaw angles than would
9, a 6-degree rudder angle in model A and a 35- ordinarily occur in the normal course of events,
degree rudder angle in model B produce the same either in steering or in the initial instants of turn-

504

ON THE TURNING

AND STEERING

OF SHIPS
Model B

Model A

/O~#iLFOrce

Rudder Force

-.
~
uctaer force

* =~-s ~ - -

V = 3.19

F*.Per Sec.

/ Hull F o r c e
/ O ~3 Lb

/"

~-~-_-~-~'~-4~

V=3.?.6 F.

Per Sec.

COEFFICIENTS

Lift Coefficients of Hull Force


0.057(;

(CL)~' =

hull force
(LA) (p/2) Y 2

0.0465

Lift Coefficients/Drift Angle


(C~)H'/

0.0069

F I G . 1 3 . - - L A T E R A L FORCES IN ~YA~,V, ~ [ O D E L S

0. 0059
A AND

J3

Y a w angles ~b are equal to t h e drift angles at the center of g r a v i t y , c.g., in Fig. 9.

ing. B u t t h a t is of no special importance for the


present purpose, which is to gain an insight into
the behavior of the hull forces.
T h e following table compares the positions of
the hull-force centers from Fig. 13 with those from
Fig. 9 for the two models, in terms of their distanees from the center of gravity. I t will be seen
t h a t both positions of the center are farther ahead
in model A than in model B, the "shift" from one
position to the other being substantially the same
for both models.
Center position in
straight-line motion (Fig. 13)
Center position in
steady-turning
(Fig. 9)
Shift of center

Model A

Model B

0.390L (ahead)

(1.230L (ahead)

0.005L (ahead)

0. 130L (aft)

0. 385L"

0. 360L

This shift of the center position is evidently


the consequence of passing from straight-line
motion to steady-turning motion at a given diameter-length ratio. Thus, in general, the position
of the center in steady turning m a y be considered
to be determined b y its position in straight-line
motion and b y the magnitude of the shift. F r o m
this point of view, the equality in the magnitudes
of the shift found for models A and B, which have
such differing center positions in straight-line
motion, is of especial interest. T h e fact t h a t the
shift is practically independent of design differences in this instance suggests t h a t it m a y be
similarly independent for all normal seagoing
vessels (its magnitude presumably being dependent on the diameter-length ratio, as implied b y

Fig. 11). If so, and if the two center positions


are in fact related to each other in this simple
fashion, it follows t h a t a relatively far forward
center in straight-line motion (which works
against the steering qualities) will always be accompanied b y a relatively far forward center in
steady turning at a given diameter-length ratio
(which works in favor of the turning qualities),
and vice versa.
This thesis cannot, of course, be accepted as
proved, on the evidence of only two models. I t
is considered, however, at least as good a working
hypothesis for the present as the hypothesis discussed earlier in the paper, t h a t an increase of
b e a m (decrease of length-beam ratio) causes an
increase of the hull-force coefficient in steady
turning. T h e latter hypothesis, based on Fig. 9,
m a y be said to be strengthened b y Fig. 13. T h e
ratios of hull-force coefficient to drift angle are:
Model A Model B
(CL)lt'/ in straight-line
motion (Fig. 13) . . . . . . . . . . . . . . . . .
(CL)H/4~.g. in steady turning
(Fig. 9) . . . . . . . . . . . . . . . . . . . . . . . . .

0. 0069

0. 0059

0.0088

0:0070

from which it is seen t h a t model A has the higher


coefficient in straight-line motion as well as in
steady turning.
With this additional piece of information, it
becomes possible to combine the two hypotheses
regarding the influence of beam. Specifically, it
becomes possible to say of model A , as compared
with model B, t h a t under comparable conditions
in both straight-line and steady-turning motion,
(1) the hull-force coefficient is greater,

ON T H E T U R N I N G A N D S T E E R I N G O F S H I P S

505

(2) the hull-force center is farther forward.


(a) tends to prevent smaller turns than are
Then, since the greater beam of model A is the
safe
largest difference between the models to which
(b) facilitates steering
these differences in characteristics might be laid, it (2) the directional instability of ships
is reasonable to suppose it to be mainly respon(a) facilitates quick turning
sible.
(b) does" not complicate steering unduly beIt should be understood that the apparent incause of the large size and the greater
fluence of beam is not stressed here with any
time intervals involved.
thought that beam is necessarily a more important
factor in the overall picture than other factors-- I t should be added, for the sake of completethe amount of deadwood, for instance--but only ness, that a sufficiently large rudder in combinabecause a difference of beam happens to be the tion with very little fixed fin or deadwood can, in
most outstanding difference between A and B, principle, provide directional stability--for either
the two models for which fairly complete informa- ships or airplanes--in combination with good
tion is currently available. The influence of dead- turning qualities. T h e rudder sizes required are,
however, larger than have been found practicable
wood, and of other factors, awaits investigation.
In concluding this discussion of steering, and of for either ships or airplanes, although they are ofthe relation of steering to turning, it seems worth ten used for motorboats.
while from the point of view of gaining overall
CONCLUDING RBMARKS
perspective to consider one more broad concept.
The ultimate objective of a serious study of
This is perhaps best brought out b y reverting to a
comparison between ship and airplane character- turning and steering should be to provide a coordinated body of design information, covering
istics.
Considerable pains are taken to give airplanes the major variables. There are at least three
positive directional stability in straight flight, or important prerequisites:
at least neutral stability. This requires the very
(A) The development of suitable test equiplarge vertical tail surfaces which are a conspicu- ment and procedures for experiments in model
ous feature of current airplane designs, and which size.
are inherently costly in weight and in complica(B) The development of simple analytical
tion of structure. On the other hand, there is methods for handling the experimental data and
good reason to believe that practically all normal ~orrdating the results.
seagoing vessels are like the two models here con(C) The identification of the major variables or
sidered in that they lack directional stability on parameters.
straight course. W h y the difference? Airplanes This paper deals principally with the second and
and ships have much the same ranges of "dis- third of these, taking the first more or less for
placement-length ratio" and relative longitudinal granted.
moment of inertia in yaw. And it m a y be acThe fact t h a t the first can be taken more or
cepted as an established fact that positive direc- less for granted should not be passed over too
tional stability makes for easy steering, and that, lightly, perhaps, because it means that unother things being equal, it is a desirable charac- pleasant surprises in the turning or steering charteristic. Why, then, if directional stability is so acteristics of new vessels can be avoided, to a
important in airplanes, is it not equally important large extent, b y making model tests in advance.
This is by no means an unimportant matter.
in ships ?
The difference between the two lies in the short- However, since it is beside the main issue of the
ness of the time intervals required for individual present paper, it is not dwelt upon.
The other two prerequisites, which constitute
events to take place in airplanes, both because of
the main issue, are interrelated and more easily
their greater relative speed (speed-length ratio)
treated together than separately.
and because of their smaller size, and in the enorThe elements involved in turning and steering
mously larger inertia forces which occur with any
are
change of direction. These make it imperative
(1) the hull form,
that airplanes be directionally stable or nearly so.
(2)
the mass and the position of its center,
In substance, and at the risk of some over(3)
the appendages, including the rudder(s),
simplication for the sake of emphasis, it may be
primarily; the propeller(s), secondarily.
said that
If these did not influence each other, and if
(l) the directional stability of airplanes
small turning circles did not have to be considered,

306

ON T H E T U R N I N G A N D S T E E R I N G OF S t t l P S

a very simple analysis would suffice, and systematic experiments could readily be organized to explore each of the elements separately. Generally suitable analyses have in fact been suggested,
Fig. 14 showing the basic diagram for one of them
[6] in which the hull is replaced b y an "equivalent" airfoil. These, or even simpler, analyses
work quite well for airplanes [12], where an assumption that the elements act independently is
close to the facts.

/3
.

F I G . 1 4 . - - D I A G R A M FROM REFERENCE ~ 6 ] SttOWING A


COMBINATION OF AIRFOIL SECTIONS CORRESPONDING TO A
SHIP H U L L AND RUDDER

The difficulty in dealing with ships is that hulls,


rudders and propellers are ordinarily so jammed
together that they affect each other materially.
This gives rise to the same sort of complication
as is encountered in the analysis of the propulsion
characteristics of ships, where the interference
effects between hulls and propellers (wake fraction, thrust deduction, etc.) are usually much too
large to be ignored, and are even now too incompletely understood to be handled easily. Furthermore, the hull, being long and shallow, is subject to large longitudinal shifts of the lateral center
of pressure as the turning circle is diminished to
small multiples of the length.
Nevertheless, just as separate studies of hull
resistance and of open-water propeller performance are vital to progress on the propulsion problem, so separate studies of hull-force characteristics and of rudder-force characteristics must be
equally vital to progress on the turning and steering problem. T h a t this has been recognized in
the past, at least in a general way, is evidenced by
the comparatively large amount of work which
has been done on rudder characteristics as such,
[2] [10] [11] and others. Corresponding work
on hull characteristics is largely lacking, however,
and this is the principal burden of the present
paper. When the hull characteristics are better
understood, the interference effects can then be

concentrated upon, and these will almost certainly


require a good deal of attention.
The method for handling the lateral component
of the hull force which has been described b y
means of examples, models A and B, is basically
sound, and is believed to be suitable for an extended exploration of the characteristics of hull
forms. It is no more than an extension to the
hull of the method already in common use for
airfoils and other aerodynamic surfaces, as well
as for rudders; namely, the description of the
force component in terms of its dimensionsless
"lift" coefficient and of its center-of-pressure
position. About the only complication, in applying it to hulls, arises through the curvature of the
path in turning and the consequent variation of
the "angle of attack" throughout the length.
This complication is more apparent than real,
however, because the whole variation of the drift
angle throughout the length is itself a function of
the ratio of turning diameter to length; thus the
lift coefficient and center of pressure can be treated
either as functions of this ratio or as functions of
an arbitrarily defined angle of attack. The method
is more or less implicit in the analysis proposed by Hovgaard in 1919 [5], and does not
differ in basic principle; it is, however, somewhat
less pretentious, which is thought to be an advantage at the present time.
In the overall picture, the function of the rudder
is to adjust the orientation of the hull to the flow,
just as the function of the elevators of an airplane
is to adjust the angle of attack of the wing. Thus
it is natural and proper, in dealing with the hull,
to consider the rudder largely in terms of the
forces and moments produced, or needed to effect
a given result. Direct measurement of the lateral
rudder-force component does just this, and, with
the readily calculated centrifugal force, provides
the necessary information for exact determination
of the hull force in steady turning.
The two models for which analyses are given
in detail, models A and B, are not especially significant in themselves. T h e y were selected largely
because, with generally similar types of hulls,
they had quite different displacement-length ratios
and had been found to have very different turning
characteristics. The analyses show the hulls to
differ considerably in respect to hull-force coefficients and centers of pressure, both in steadyturning and in straight-line motion, and suggest
two generalizations :
(1) That, under corresponding conditions, the
force coefficient increases and its center moves
forward with increase of beam (decrease of lengthbeam ratio),
(2) T h a t the shift of the center which aceom-

ON T H E T U R N I N G AND S T E E R I N G OF SHIPS
panics a change of diameter-length ratio, or a
change from straight-line to steady-turning motion, is independent of design differences.
To be sure, these are as yet nothing more than
indications--guideposts for directing further experiments on a systematic basis. They are important at this time primarily because, unless full
advantage is taken of indications of this sort which
come to light, and reasonable luck is had in selecting for emphasis those indications which bear most
directly on the controlling variables, any comprehensive program of systematic experiments becomes too lengthy and involved to be practicable.

507

ACKNOWLEDGMENT
This paper represents a distillation of certain
ideas which have developed in the course of recent
work on the turning problem sponsored by the
Bureau of Ships, Navy Department, and by the
National Defense Research Committee. Acknowledgment is made to these agencies for permission to publish.
Acknowledgment is made also of help in the
preparation of the paper given by the staff of
the Experimental Towing Tank, Stevens Institute
of Technology, in particular by Messrs. John B.
Drisko and William H. Sutherland, and by Miss
Ruth Magor.

APPENDIX
NOTES ON TEST METHODS

The turning data used to illustrate various aspects of the discussion in the text were derived
from two sources--tactical trials of ships, and
turning tests of models.
Tactical trials of full-size naval vessels have
been made periodically by the Navy, over a
period of years, by methods developed by the
U. S. Experimental Model Basin and the David
Taylor Model Basin [7]. The trials are made
near shore, where deep water is available and observation from two shore stations is practicable.
The path of the ship is determined by triangulation from the shore stations. The heading of the
ship is determined from simultaneously recorded
gyro-compass readings on the ship. All data are
photographically recorded to provide a permanent
record.
Turning tests of models have been made at the
two Navy Model Basins and also, during the past
four or ~five years, at the Experimental Towing
Tank at Stevens Institute of Technology. The
following notes relate principally to the work at
the Experimental Towing Tank.
Early tests were made in the Stevens Institute
swimming pool, and in the somewhat larger swimming pool at Columbia University. In 1942, a
maneuvering basin was constructed at Stevens
Institute by the National Defense Research Committee. This tank is 75 feet square, by 41/~ feet
deep, and has a 25-foot extension on one side to
provide scope for an approach run preceding a
turn. The models tested are generally from 5 to
7 feet in length, and weigh from 15 to 75 pounds.
The path of the model is determined from polar

coordinates recorded by a single motion picture


camera set at one corner of the tank so that it can
sweep the test area and follow the model as it
maneuvers. The method has been described in an
earlier paper [8]; azimuth is given by a degree
scale photographed through an auxiliary lens, and
distance by an adaptation of the "stadia" surveying method. Data ordinarily derived from the
test, in addition to the path, are the headings of
the model, the approach and steady-turning
speeds, and the heel angle in steady turning.
Recently a special dynamometer has been put
into use for measuring the lateral component of
the rudder force either during turning or on
straight course. The dynamometer is self-contained, and is mounted in the model. The values
of lateral rudder force referred to in the text were
determined with this piece of equipment.
The lateral force and moment on the hull, when
moving on straight course with combinations of
yaw and rudder angles, are measured in the towing tank with a pair of laterally acting dynamometers originally developed for testing sailing
yacht models [9]. These dynamometers are
mounted on the towing carriage, and restrain the
self-propelled model to move on straight course at
the prescribed yaw angle, besides measuring the
lateral force and moment. Data in the text on the
lateral force and moment in yaw were determined
in this way. The lateral rudder-force component
is separately measured during the course of the
test by the special self-contained dynamometer,
above described.
A three-component rudder dynamometer to

308

ON T H E T U R N I N G AND S T E E R I N G OF SHIPS

measure lift, drag and torque is now being developed, for use on large models at the David
Taylor Model Basin. This piece of equipment
is a step in the direction of getting a more detailed

picture of the forces acting during turning, which


appears to be necessary in a complete study of
interference effects between hulls, rudders and
propellers.

REFERENCES

[1] Johns, Sir Arthur, "Ship Form and


Steering," The Engineer, Volume 126, 1923, pages
551-552 and 576-577.
[2] Schoenherr, Karl E., "Steering," "Principles of Naval Architecture," Volume 2, 1939,
Chapter IV, pages 197-233. (A good bibliography, as of the date of publication, is included.)
[3] Minorsky, Nicholas, "Note on the Angular Motions of Ships," Transactions of The American Society of Mechanical Engineers, Volume 63,
1941, pages A-Ill-A-120.
[4] Taylor, D. W., "The Speed and Power
of Ships," 1943 Edition.
[5] Hovgaard, William, "Turning Circles,"
Transactions of the Institution of Naval Architects, Volume 54, 1912, pages 23-24.
[6] Kucharski, W., "Zur Theorie des Steuervorganges bei Schiffen," Werft, Reederei, Hafen,
page 35, 1932.
[7] Farrin, Lieutenant J. M., and Landweber, L., "Tactical Trial Methods and Equipment," Taylor Model Basin Report No. 476,
March 1941.

[8] Davidson, Kenneth S. M., "The Growing Importance of Small Models for Studies in
Naval Architecture," Transactions of The Society
of Naval Architects and Marine Engineers, Volume 49, 1941, pages 91-121.
[9] Davidson, Kenneth S. M., "Some Experimental Studies of the Sailing Yacht," Transactions of The Society of Naval Architects and
Marine Engineers, Volume 44, 1936, pages 288303.
[10] Baker, G. S., and Bottomley, G. H.,
"Maneuvering of Ships--Part I--Unbalanced
Rudders of Single Screw Ships," Institution of
Engineers and Shipbuilders in Scotland 1921-22,
page 522 ft. Parts II, lII, and IV, by G. H. Bottomley appeared subsequently, see "Bibliography"
of reference [2].
[11]. Gawn, R. W. L., "Steering Experiments," Transactions of the Institution of Naval
Architects, Volume 85, 1943, pages 35-73.
[12] Diehl, Walter S., "Engineering Aerodynamics," Revised Edition, 1936.

DISCUSSION

REAR ADMIRALE. L. COfiHRANE,U.S.N., Council ~[ernber: Professor Davidson has approached


the subject of turning and steering characteristics
of ships from an exceptionally broad point of
view. He has stimulated thought on these questions along lines which should lead to further experimentation, to new techniques for such experimentation, to a clearer understanding of the problems involved and, it is hoped, to eventual improvement in ship performance. He has certainly
put forth an excellent case for continued research
and study and the new techniques to which he re-

fers are very promising for future w.~)rk. It is


hoped and expected that Professor Davidson will
carry on a large part of this work, using facilities
and techniques he has developed at Stevens Institute and data from other sources, including fullscale trials. The David W. Taylor 5iodel Basin
will also continue to have a large share in the investigation, primarily in collaborating with Professor Davidson in determining the course of the
investigation, and in experimentation with its own
facilities which permit the use of somewhat larger
models than those used at Stevens Institute.

ON THE T U R N I N G A N D STEERING OF SHIPS


The program now envisioned by Professor
Davidson, with the full endorsement of the Bureau of Ships, is in such an early stage of development t h a t it is difficult, if not impossible, to project its outline very far into the future. In any
long-range research program of this scope and
magnitude, the objectives and the course to be
followed will stand out more clearly as the program develops. The first steps will necessarily be
broad, general, and of an exploratory nature. The
findings, as Professor Davidson has been careful
to point out, will not be immediately useful as
specific design data, but they should surely lead to
a determination of factors upon which to concentrate attention, and investigation of these factors
should, in turn, give results which will furnish a
powerful tool to the designer. Although it is too
early to comment very specifically on the course of
the program, it does seem pertinent to raise certain questions in connection with the present
paper in the hope of future stimulation of thought
along the lines suggested by Professor Davidson.
With respect t o t h e turning of ships rather than
steering, Professor Davidson's paper is broady
divided into two parts. The first is a presentation
of the thesis that, once the tactical diameter is
known, the other elements of turning can be readily estimated b y means of the curves shown in
Fig. 3. This is an important conclusion which appears to be justified in a general way b y the data
presented. The horizontal dispersion of spots,
however, is so great t h a t one cannot help but feel
t h a t there are a number of factors involved which
need further exploration.
The second phase of the turning problem presented b y Professor Davidson is considered to be
of greater significance than the first phase. Here
he sets out to determine the effects of hull form
factors rather than rudder factors on turning. I t
would certainly be desirable for the designer to
know which factors of hull form affect turning
and, hence, afford a means of adjusting such elements of a design as can be modified to improve
maneuvering characteristics.
In the present
paper, the author has cited an example which indicates t h a t beam-length ratio is a predominant
factor. I t certainly appeals to one's common
sense t h a t this should be the case, with the maxim m n beam-length ratio being represented b y the
circular form such as a saucer, which has practically no resistance to turning. On the other hand,
in the examples cited b y Professor Davidson, displacement-length ratio ~tands out as a factor of
possible importance since its value in model A is
more than twice its value in model B. Whether
this has any effect on turning or not is certainly
not clear at present. Be t h a t as it may, it is

309

fortunate from the designer's point of view t h a t


beam-length ratio is not the only major hull form
factor affecting turning because this ratio is
usually of necessity determined b y other considerations than turning. W h a t the designer would
like to find is some hull form factor upon which he
can operate to improve turning without serious
adverse effects on other characteristics of the design such as speed, endurance, etc.
There are a number of other questions which
might be raised concerning specific points brought
out in this paper, but only one more appears to
require comment in the present discussion. The
paper minimizes somewhat the danger of excessive
angle of heel in very short turns. For high-speed
craft, or moderate-speed craft with unusually high
center of gravity, great care m u s t be taken in the
design to avoid excessive angle of heel in turning
and this m a y well prove to be a limiting factor in
reducing turning-circle diameters of ships in these
categories.
Professor Davidson deserves not only congratulations for his splendid paper, b u t also the full encouragement of this Society for further work he is
undertaking in this i m p o r t a n t field.
REAR ADMIRAL HERBERT S. HOWARD, U.S.N.,
_/].'[ember: I wish to congratulate Professor Davidson on this paper, "On the Turning and Steering
of Ships," since all too little is known about this
very vital subject. ;~1ore t h a n ever before, I believe, the present war with the extensive bombing
of ships by airplanes and the need for rapid evasive tactics, as well as the frequent and rapid
changes of course necessitated in modern attack,
have made it of the greatest importance that
naval fighting ships should possess rapid turning
and maneuvering characteristics. On the other
hand, these characteristics must not be attained
at the expense of good gteering qualities.
Little has been known in the past as to the
fundamental reasons for certain turning or steering qualities, and it has simply been assumed t h a t
one ship, generally like another in hull form,
speed, etc., would turn and steer about like its prototype. I t is definitely of importance t h a t we
should be able in a given design to forecast what
the steering and turning qualities of the ship will
be. Professor Davidson in his work of the past
few years at the new maneuvering basin at Stevens Institute, and through his theoretical studies
as evidenced b y this paper, is gathering the knowledge which should lead to the results which we
are seeking.
We at the David Taylor Model Basin are very
closely associated with Professor Davidson in his
work on turning and maneuvering at Stevens, and

310

ON T H E T U R N I N G A N D S T E E R I N G OF SHIPS

we look to him for much of our present knowledge


on this important subject, while we shall look for
even more in the future.
Steering and turning are very complicated subjects, and from actual observation of steering and
handling of ships in the service over a good many
years past, there are two or three points which I
might add to Professor Davidson's comments, as
to why one ship turns or steers differently from another as outlined in his paper.
Looking at his models A and B, and forgetting
everything but the fact that model A is broader
in the beam for the same length, m y own experience would lead me to believe that in general a
broad beam means a Chip which is hard to steer
and keep on her course. As an ultimate in this
direction, all of us who learned our naval construction from Captain Hovgaard will remember
his description of the Russian vessel l%e Admiral
Popoff, which, as I remember it, was completely
circular and could not be steered at all but simply
went completely around to starboard or port
whichever way she started first. I mention this
simply as the ultimate of a broad beamed ship.
As another point, speed has a great deal to do
with steering and turning as must have been
noted by anyone who has been aboard destroyers
a good deal. When these ships get up to high
speed, they steady down and hold themselves to a
fixed course, except in a heavy seaway, so that
practically no steering at all is required. This
condition is undoubtedly brought about by the
stability of motion and probably as well bv some
effect of the wake behind the ship.
I mention these two points only to add to what
Professor Davidson has already said as to the complication of this subject and the many elements
that enter into it.
In this paper, Professor" Davidson has taken a
very sound step in his study of steering and turning, in that he has tried to break down these complicated maneuvers into their basic parts, and has
endeavored to separate out one element, the effect of hull form on turning and steering, and then
to study and analyze this one element. I feel that
this is a real step in the right direction, amt we will
follow his future work with much interest.
It will be of interest to those attending this
meeting to know that at the present time we have
under consideration further and closer contract
arrangements with the Stevens Institute, which
will insure that Stevens and Professor Davidson
have the backing of the Bureau of Ships of the
N a v y Department and of the Taylor Model Basin
in the continuation of this work on which he is now
engaged.

DR. L. hi. OOLDSMITII, J[ember: Professor


Davidson's paper "On the Turning and Steering
of Ships" presents a discussion of some of the factors entering into this subject and some methods
of analyzing the results of turning and steering
tests. It is pointed out that there is a lack of
quantitative data on the subject which prevents
any sure prediction of steering and turning qualities in the earl)" stages of design preparation. The
author states that the paper makes no a t t e m p t to
provide such quantitative design information and
is intended only to suggest the most promising
lines of attack for further experiments.
Particular attention is given to the influence of
hull form oi1 turning and steering, and experiments
are described in which the turning quality appears
to be considerably affected by the length-beam
ratio or the displacement-length ratio, or both.
This is sad news for the i)o~)r naval architect who,
it seems, is about to be confronted with another
antagonistic variable in hull design, as if he
didn't have enough of these already which have to
be appeased and reconciled with compromises. It
is hoped that Professor Davidson is not going to
bring us eventually to the point where we shall
have to change tile hull instead of tile rudder in
order to improve steering on a completed ship!
The paper is divided broadly into two parts, the
first dealing with turning and the second with
steering. I believe it will be generally conceded
that the latter is the more important of the two
qualities in actual ship operation and that the
1S0-degree turn, for non-combat, ocean-going vessels, is a seldom-required maneuver. I t m a y be
important for certain coastwise or inland waterway craft where docking conditions or restricted
channels make a short turning radius desirable
The turning diameter of any ship has been generally considered as a measure of steering quality
but in the experiments with models A and ]3, described in the paper, doubt is cast on this assumption by results showing only a very small superiority in steering in model A which had a much
smaller turning radius.
Professor Davidson states, however, that this is
not a satisfying conclusion in view of the further
information given on page 303 which indicates that
model A was much more sensitive to small rudder
movement than model B, even to the extent that
it was difficult to keep on a straight course. This
tends to introduce a contradiction or at least an
uncertainty which may depend on the definition
of good steering and on the personal opinion or experience of the helmsman. Ordinarily sensitivity
of helm or quick response to small rudder movement would be considered one of the most important and valuable attributes of good steerin K

ON T H E T U R N I N G A N D S T E E R I N G OF S H I P S
and the difficulty is to determine when it ceases to
be a virtue and becomes a fault in which the directional instability is violent and the ship is subject to wild yawing requiring excessive and continual application of helm to keep her on a straight
course. From the statement made in the paper
regarding the full-size prototype of model A it
probably does not fall into this category although
there appeared to be some difficulty in keeping it
on a course at first until the crew got used to it.
tt can doubtless be assumed, therefore, that the
hyper-sensitive helm on model A is not a liability
and may be a great asset in an emergency where
quickness of response means everything. This is a
different picture, however, with automatic steering by gyro-pilot and may cause steering gear
trouble.
It would be of interest and give a clearer picture
of the difference between the steering qualities of
models A and B if turning diagrams for each were
plotted to scale and superimposed on a drawing
similar to Fig. 1 and which would show the relative positions and headings of each model at equal
time intervals during a 1S0-degree turn. It is
hoped that the data are at hand for plotting such
a diagram or can be obtained without too much
additional work.
Although Professor Davidson states that the
present discussion is not concerned with appendages, it is suggested that the interest and value of
the paper would be increased if another figure were
added showing to a reasonably large scale the stern
profiles of models A and B with dimensions and
relative positions of rudders, propellers, deadwood, etc.
This paper is of much interest and we hope Professor Davidson can follow it up with another one
at each succeeding meeting of the Society until
we have the quantitative information which he
states is lacking on this subject.
The discusser would not consider these comments complete if he did not record here that the
effect of rudder design will undoubtedly influence
all of this work of Professor Davidson's. Professor Davidson particularly set forth that he was
going to hold that in abeyance for future experiment. The writer's experience has disclosed that
there is a tremendous effect on steering, particularly from the standpoint of quick answering, as
indicated by rudder design, and not the least of
this effect is from the fact that a properly designed
streamline rudder will reduce the load on the steering gear so as to speed up the hard-over to hardover action and thus improve steering.
We look forward to Professor Davidson's future
experiments on rudder and appendages as a very
necessary and worth-while effort.

311

CAPTAIN H. E. SAUNDERS, U.S.N., Council


Member: This excellent paper by Dr. Davidson is
one of the first, at least in this profession, to
emerge from the confines of the barbed wire
fence which has necessarily been erected around
the war activities of the National Defense Research Committee. I t will give to our membership
an idea of the useful work which has been done
during the war along one particular line, and I
think I m a y safely say that it is only a forerunner
of m a n y such papers which will in time turn the
weapons of war into useful tools of peace.
Commander Wright of the David Taylor Model
Basin sta/t will comment on certain basic features
of the problem so ably discussed here; t shall
therefore confine m y comment to a few particular
points.
I think that the rate at which steering and turning phenomena take place in service deserves considerable emphasis in any consideration of the
ease of handling a ship, whether in straight route
or in turns. Dr. Davidson makes a comparison on
page 305 between the acceptable directional stability and maneuverability characteristics of a
ship and an airplane and shows how the single element of time is a powerful factor in determining
these characteristics.
Ships and airplanes are, after all, not too similar,
and it might be useful to make a comparison between large and small ships of the same type,
such as that afforded by the experience of our
N a v y with large and small submarines.
When the first 3000-ton submarines were built
in this country, with lengths exceeding those of
the most modern destroyer, there was considerable uneasiness about operating these huge and
supposedly unwieldy craft submerged. To the
surprise of everyone, they handled amazingly well.
Although over twice as massive as anything our
naval personnel had previously taken under the
water, they seemed to behave with a corresponding degree of ease.
I t was puzzling for a while, to those of us who
had served our apprenticeships in the skittish
craft of one-tenth their size, to understand why
these huge creatures should be so tractable. It
was perhaps only by accident that we realized that
the time element was favoring the big craft, in a
sort of competition between their movements and
our own mental processes and physical reactions.
The big submarine was following diving curves
generally similar to the small one, but the speedlength ratio was much less, with a corresponding
reduction in the rate at which things were happening, so to speak.
Although these large submarines had a high degree of maneuverability, sufficient to permit them

512

ON T t I E T U R N I N G A N D S T E E R I N G OF S H I P S

to turn inside of a destroyer, the fact that thev


were large led to the successful use on them of
hand steering, which is rather out of the ordinary
for a 3000-ton, high-speed vessel. Naturally hand
steering was too slow for close quarters but in the
open sea, on the surface, the vessels could with
some experience be steered by hand almost as well
as b y a gyro pilot. The secret lay in the use of the
minimum possible rudder angle, and when the
steersman had to apply it by hand, the angle
automatically became small.
These large submarines, with good form, oversize rudders, and plenty of power, proved that
ships can be good steerers as well as good turners.
.X[R. S. A. VINCENT, Co.ncil Member: W e are
fortunate in having this little tmdcrstood subject
under study by" such an able man as the author.
Although, to use his own words, he makes no pretense of providing design information at this time,
it seems reasonable to assume that a continuation
of his tests and analyses might, and I hope will,
make clear and evaluate, at least apl)roxirnately,
the wtriables that contr()l the steering vnd turning
of ships.
He menti~)us the impracticability of providing
directional stability'. While this may be true, it is
nevertheless well established that st~me vessels
continue on course with only occasional smallangle use of the rudder and but little m~ticeablc
yawing. Others do not. The exact turning circle
of merchant ships is rarely a specified requirement
and for most seagoing vessels is of sec~mdary importance compared with directional stability and
minimum loss of speed due to yawing and rudder

angle.
I have never taken the idea seriously, but it
m a y be of interest to mention that to improve directional stability one well-known naval architect
makes a practice of adjusting the forefoot and
skeg so that the center of area of the underwater
lateral plane is at least slightly, abaft the center of
gravity of the vessel. In contrast to this, it is well
known that sailing yachts having a deep forefoot
stay o n course better than those having the modern cut-away forefoot. The author's ideas on
means to minimize directional instability would
be welcomed.
Shallow water steering and astern steering
should not be overlooked. Some merchant vessels easily maintain an astern course and can be
put through a figure eight astern maneuver at will.
Others cannot. One important modern class of
twin-screw seagoing vessels is completely out of
control going astern. These vessels' propellers
turn inboard going astern. They promptly start
to turn and regardless of rudder angle, varying

revolutions---even one engine ahead and the other


astern--will not stop turning until the ship is dead
in the water. They cannot be held on a straight
astern course. These ships normally trim considerably by the stern, but reduced trim does not
appear to affect noticeably the performance.
Knowledge of the cause of and cure for lack of control going astern seems important and I am sure
that m a n y of us would also welcome any light that
the author m a y throw on this subject, eventually
if not at this time.
;~IR. JMES L. BATES, Member: It would have
been difficult to suggest a subject of more importance to the successful ship than its furning and
steering; nevertheless, the investigation of the
subject has been generally neglected. It is true
there has been some investigation of the action of
inclined planes and with the aid of mysterious coefficients the data derived therefrom have been
utilized in approaches to rudder design.
This paper presents in an interesting and lucid
manner some of the results of a long and carefully
planned investigation of turning and steering.
To the knowledge of the writer, there has not been
another investigation1 comparable in range and
accuracy. It is believed that the work on which
this pal)er is based may well f(~rm the point of departure for what may ultinmtelv become a really
satisfactory approach t(, the problem of rudder
action behind ships' hulls so that the turning and
steering characteristics ()f ships may be intelligently forecast during the design stage.
The problem p~sed m a y be summarized in the
following statements quoted from the paper:
"The general eonclusi()n is inescapable that the
hull forln, through its control over the hull-force
center, plays an extremely important part in fixing
the steady-turning characteristics, and can in no
sense be considered secondary in importance to
the rudder.
"The ditticultv in dealing with shii)s is that
hulls, rudders and l)ropellers are ordinarily s~)
jammed together that they affect each other materially.
"In the overall picture, the function of the rudder is to adjust the orientation of the hull to the
flow, just as the function of the elevators of an airplane is to adjust the angle of attack of the wing.
" I t is an old adage t h a t a good-turning ship is
hard to steer and that a good-steering ship is hard
to turn. The implication is clear that good turning and good steering are essenti~dly antagonistic
qualities. The definitions often used, that turning is the process of altering course and steering
the t)rocess of maintaining course, tend also to suggest a certain antagonism between the two."

ON THE TURNING

AND STEERING

I t appears impossible to present a satisfactory


definition of ship turning and steering which would
be applicable to all sizes and types. The reason is
that, as pointed out in the paper, turning or
maneuvering is essentially a different phenomenon
from steering or maintenance of course in free
route. For example, an Atlantic passenger liner
requires a m a x i m u m of steadiness on her intended
course. For her, turning is of lesser importance.
On the other hand, a change in direction is of first
importance to a harbor tug, whereas her maintenance of a steady course is of less importance.
The deep end sections and long straight keel
line of the Atlantic liner are favorable to her performance, as above indicated, while the cut-away
forefoot and drag of the keel, common to most
tugs, are favorable to their desired behavior.
The comparison between the maneuvering characteristics of a surface ship and an airplane as
shown on page 289 is very interesting. I t is another
case where the progressive attitude taken b y the
airplane designer can be studied with advantage
by the naval architect.
The plot shown on page 290 giving the values of
advance divided b v tactical diameter is of interest
and shows the distinct trend in spite of the noticeable spread of the plotted spots.
The use of the term "speed recovery" is new to
the writer and forms an added element of interest.
I t has been the understanding t h a t the most important turning characteristic, particularly in time
of crisis, is the immediate response of the vessel to
rudder action. However, this response involves
both the element of change of direction of the vessel's axis, and the displacement sideways (toward
center of turning circle) of her center of gravity.
From the data given it is difficult to discover
whether these characteristics are more pronounced
in one model than in another.
I t would be desirable if Professor Davidson
could emphasize the effect of the shape of the after
underwater sections more than he appears to have
done in this paper. In vessels having broad shallow after bodies with after deadwoods cut away,
the easy flow of water transversely under the stern,
once turning is commenced, is of great importance.
The United States destroyers built just prior to
the first World W a r had deep V-sections fore and
and aft with long straight keels. Their turning
circle was verv large, the water piling up against
the topsides aft rather t h a n passing under with
facility. The contemporary British destroyer of
approximately the same length but built with a
broad flat stern and a minimum of after deadwood
turned in about two-thirds the length required for
the American type.

OF SHIPS

315

MR. J. P. COMSTOCK,Member: A considerable


p a r t of this interesting paper is devoted to an analysis of the difference in turning behavior of models A and B as shown in Figs. 8 and 9. I t is shown
t h a t the easy turning of model A is due in large
p a r t to the fact t h a t both when starting to turn
and when yawed the center of transverse hull force
is much farther forward on model A than on
model B, and this in turn is attributed to the relative fatness of model A.
This suggests two questions: Can we, in basic
design, choose hull characteristics conducive to
easy turning, and, second, if we could, would it be
desirable to do so ?
T h e answer to the first seems to be in general
"no." The b e a m - d r a f t ratio, beam-length ratio
and displacement-length ratio will be settled b y
other considerations such as cargo cubic, passenger
area, stability, etc., and in general we will accept
the resulting turning characteristics as far as the
hull is concerned.
The answer to the second m a y be " y e s " for
naval vessels, where fast maneuverability in action
m a y be essential, and " n o " for merchant vessels
where ability to hold a course m a y be more desirable than maneuverability. I think the author is
correct, on page 303, in declining to be entirely
satisfied with his own reasoning leading to the conclusion t h a t good turning m a y be associated with
good steering. Course-holding ability, typified by
a plank-on-edge and to a less degree b y model B,
would seem to be inherently antagonistic to easy
turning, and for a merchant vessel course-holding
m a y be the more desirable attribute. However, as
just noted, we will probably never design our
principal hull characteristics primarily to get
either one.
The author has opened up a wide field, which
both Stevens Institute and the Taylor Model
Basin are now well equipped to pursue, and which
probably will not be exhausted for some time.
PROFESSOR G. C. ]~IANNING, Council Member:
Dr. K. E. Schoenherr's equation for the radius of
the turning circle at constant rudder angle [2] is
p =

2 V' cos

If we assume that, up to the angle of rudder breakdown, the rudder force varies as the sine of the
rudder angle, as given b y Joessel and verified
reasonably well in rudder experiments, this equation m a y be re-written as follows:
p

2a V' cos
Ca sin a cos aAk'( L -- a)

314

ON T H E T U R N I N G A N D S T E E R I N G OF S H I P S

Further, since sin a cos a = 1/',2 sin 2a, we may


further simplify the foregoing formula to

C a A k ' ( ~ -- a)sin 2~
This last formula indicates the generally accepted
facts t h a t : (1) the radii of the turning circles of a
model and its prototype vary as the ratio of their
linear dimensions; (2) the radius of the turning
circle varies inversely as the rudder area; (3) the
radius of the turning circle varies inversely as the
sine of twice the rudder angle, provided no breakdown in flow occurs; (4) the radius of the turning
circle varies directly as the distance between the
center of gravity of the ship and the point of application of the hull force component. This formula also indicates that the radius of the turning
circle is independent of speed. While not absolutely correct, speed variation does have a relatively minor effect on radius.
Reference t2t does not give any precise means of
determining the location of the point of at)plication of the lateral resistance of the hull. This
paper suggests that the distance of this point from
the ship's center of gravity depends principally on
the length-beam ratio for hull forms which are
otherwise similar. The author points out that
this probability is based on experiments with only
two models and that no general conclusion can
safely rest on such a narrow base. Nevertheless,
it would appear that a promising field for further
research has been opened by this author. It is to
be hot)ed that he will run a series of models based
on the same parent lines but differing ill lengthbeam ratio to ascertain if, at constant speed and
rudder angle, there is a consistent relation between length-beam ratio and the distance of the
hull force from the center of gravity. This would
make possible k much more accurate estimate of
the maneuverability of a projected ship in the
early stages of design. If we express the tactical
diameter as a multiple of the length (i.e., let o =
nL), we have

nL =

4a V' cos ,
Cm4k ( L -- ,,) sin 2,~"

Solving for a gives

KL
2K 4- 8 V' cos 4,
where

K = ComLAk' sin 2a
From this we can plot a curve of values of a
against rudder area and thereby, for assumed hull

characteristics, estimate the rudder area retluired


for a desired turning radius.
it is hardly likely that the length-beam ratio of
ships will be chosen for their maneuvering characteristics. This is more likely to be selected from
its effect on l)r()pulsivc resistance and general suitability of the hull form for a specific use. Knowledge of the l~mati,')n of the hull force will be ,)f
assistance in estimating the rudder area required
to obtain a desired turning radius.
One conclusion that must follow, if the indications of the two m o d d s investigated prove to be of
general apt)lication , is that long slender ships like
destroyers and cruisers cannot be given turning
radii relatively as small as battleships and carriers
unless they have relatively larger rudders.
It is noted in Fig. 1 that advance, transfer and
tactical diameter are defined with respect to the
path of the ship's center of gravity, rather than
with respect to the ship's heading. When the
former convention is used, the change of ship's
heading is 90 or t80 degrees plus the drift angle.
Since most information about the maneuvering of
ships is based (m the latter practice, I cannot refrain from asking the author why the change.

XIR. C. S. L. ROBINS()N, Associate M e m b e r : It is


true, as Professor Davidson states, that a thorough analysis of the forces which turn and steer
ships is yet to be made. It would certainly be desirable to be able to evaluate the hull forces and
the rudder forces, and their points of application,
and then from the centrifugal forces and the inertial forces to predict the turning paths and the
steering qualities.
The geometric similarity of various turning
circles is well established by Fig. 3. The bottom
graph shows how closely the path approaches a
true circle. In the middle graph, however, the advance may be divided into two parts:
(t) A transient initial period.
(2) The first quadrant of the turning circle.
While the second part is obviously 0.5 X T D ,
the first part is not a function of the tactical diameter and should not be plotted as such. From
the author's curve it appears to be about 1.4 X L.
In the to I) graph of Fig. 3, the correlation is excellent but the underlying assumption of Fig. 7,
upon which the curve is based, is very broad.
Quoting from the text, this assumption states that
" . . . a n y increased thrust attributable to greater
propeller slip at reduced ship speeds (the throttles
remaining unchanged) is approximately offset by,
increased drag of the hull and rudder caused by

"

ON THE TURNING

AND STEERING

turning, or t h a t these two secondary forces are


negligible in magnitude." Is there any other
justification for this assumption besides the correlation of this particular curve?
In using the terms "lift" and "drag" a large
amount of confusion m a y be averted if lift is always normal to the fluid velocity. Then the hull
lift will be in just the opposite direction to the
whole centrifugal force vector. T h e propeller
thrust would then largely oppose the hull drag,
with a small component opposing the centrifugal
force. Does not this resolution of forces appear
more logical? in the same terminology, what the
author calls the hull "lift" force should be called
the " n o r m a l " force.
The difficulty which the author is trying to
avoid is the fact t h a t the angle between the water
velocity and the ship eenterline varies as much as
30 degrees a from the bow to the stern. This angle,
which m a y be called the "angle of attack,"
changes thusly:
(1)
(2)
(3)
(4)
angle.

At
At
At
At

bow-- angle is negative.


pivot p o i n t - - a n g l e is zero.
e.g.--angle is defined as drift angle.
stern--angle is almost three times drift

Hence, dividing C~ b y the angle at the c.g., the


drift angle is purely arbitrary. Here it is desired
to find the angle which best corresponds to an
average angle of attack, for the lift coefficient
usually varies linearly with this quantity. Comparing (CL)H/'4)c.~ for turning with (CL)xt'flP for
yawing it is evident t h a t the average angle of attack is best represented b y an angle somewhat
larger than the drift angle.
The value at the stern should always be estimated, since aside from the propeller race and the
hull interference effects it should be subtracted
from the rudder angle to determine the rudder
attack angle.
Here
= tan-1 2L1/D
where
L 1 = distance from the pivot point to the rudder.
D = diameter of path of pivot point.
For instance, in Fig. 9 this correction to rudder
angle is as much as 18 degrees. Thus the rudder
angle with respect to the ship has little meaning
unless data are supplied for calculating the foregoing correction.
For small

TD/L

ratios;

less for larger ones.

OF SHIPS

315

In regard to steering, the directional instability


might be evaluated quantitatively. Possibly another concept from aerodynamics could be used;
t h a t is, the m o m e n t coefficient plotted against the
angle which the ship eenterline makes with the
flow direction. I t is usually indicative of the mom e n t about the leading edge. F r o m this the hydrodynamic center of pressure could be determined and the turning m o m e n t on the ship for
various yaw angles could be taken as the product
of the force component normal to the ship centerline and the distance from the center of pressure to
the center of gravity.
In calculations for actual ships, the mass and
the m o m e n t of inertia should be increased to allow
for entrained water. These values are used when
estimating linear and angular accelerations, respectively.
Although the contents of this paper are not intended to supply data for more accuratelY estimating turning and steering characteristics, it is
useful in synthesizing the ship's course from the
component forces and moments. Such is the only
possible general t r e a t m e n t of the problem.

MR. W. W. SMITH, I[onorary I~ce-President:


This paper gives a very thorough and excellent
analysis of needed and useful information on this
subject.
Work of this kind consists of two parts; namely,
(1) making tests and finding the facts, or research;
and (2) analyzing and organizing the facts in a
systematic manner for practical use in the design
of ships, or development. Far more time is expended on the former, and it seems to be the feeling t h a t the job is complete when the tests have
been made. But, unless the tests are systematically analyzed and codified, as the author puts it,
they are not useful for practical design. There is
a great mass of undigested test data in model basins t h a t could be and should be analyzed and
codified to make these data available for practical
use. The time required would be a small part of
the large amount of time currently expended in
making tests.
The author is commended on
making a step in this direction, and it is hoped
t h a t the Taylor Model Basin will follow his lead.
The curves of geometric proportions on page 290
appear to be well-established fundamental values,
and should be very useful in solving turning problems. Specific data of this kind are just what designers need.
The form of presentation on page 291 shows very
clearly, and at a glance, the comparative characteristics of various sizes of turning circles. The
advantages of small turning circles are apparent.

"--'4

516

ON T H E T U R N I N G A N D S T E E R I N G O F S H I P S

These advantages are less advance, less time and


quicker turn-around; but they must be balanced
against the design provisions to obtain them.
Referring to page 29S, specific data for determining the diameter-length ratio would also be desirable; but, since this involves both the hull and
the rudder, complications may be encountered.
Referring to page 301, the use of contra-guide
rudders on single-screw ships has improved both
the turning and the steering qualities, especially
the latter. With this design of rudder, the axial
streamline flow of the water, without discontinuity, causes the water t() hug the rudder closely on
both sides, so that the smallest change in rudder
angle gives an immediate response in steering.
On twin-screw ships, and especially those with
full forms, it is necessary for a single rudder to project into the propeller slip stream in order to obtain good turning and steering qualities. In the
case of one full-form tanker, the ship could not be
kept within two or three points of the course until
the width of the rudder was increased so as to project into the slip stream. There was another case,
considerably more ancient, in 1902, of the old gunboat Ilelena, which was unable to keep on a course
within two or three points, and was therefore unable to dodge the samtbanks in the Yangtse River.
The width of the rud(ler had to be increased and
this corrected that difficulty. After that, the
ship steered quite well.
The steering and turning qualities of twin-screw
ships could be greatly" improved by using twin
eontra-guide rudders h)eated behind the screws;
and, in addition, the prot)ulsive efficiency would be
increased by changing the flow in the slip stream
from a spiral to an axial direction.
The higher
velocity of the water in the slip stream and the
axial streamline flow of the water acting on the
rudder would greatly increase the rudder forces
and the sensitivity of response to steering. In
ordinary ships, the wake velocity acting on a
single rudder would be from 0.7 to 0.S of the ship
speed, whereas the veh)eity of the slip stream acting on the twin rudder would be about the same
as the ship speed. Since the forces vary with the
squares of the velocities, the unit forces acting
on the twin rudder would be from 1.5 to 2 times
the unit forces acting on the single rudder. For
the same total rudder force, the total area of two
twin rudders would be from 0.5 to 0.7 times the
area of a single rudder; and the area of each twin
rudder would be from 0.25 to 0.35 times the area
of a single rudder. Therefore, the size and weight
of the two twin rudders would be considerably
smaller (and of each twin rudder, very much
smaller) than that of a single rudder. Thus, apparently, twin rudders have marked advantages

and, consequently, their use is indicated as preferable.


I wish to congratulate the author on his very interesting and valuable paper, for which he should
receive the thanks of the Society.
I_)R. KARL E. SCHOENIIERR, Member: In this
interesting paper Professor Davidson discusses
the broad concepts of the stability of ships and
presents conclusions based on the work performed
by the Stevens Institute Experimental Tank during the past two vears.
The most important conclusion reached by Dr.
I)avidson appears to be that the hull has been the
neglected factor in previous discussions of the
turning problem and that progress will be made
only if this fault is corrected. It is agreed that
there is some justification for this contention, in
that up to now no systematic experiments have
been carried out to determine the lift forces on the
hull such as the numerous experiments that have
been made to determine the resistance of the hull
(,n a straightaway course. However, it is believed
that this neglect is not due to an oversight or t,)
faulty analysis of the problem by previous investigators but to the simple fact that for surface vessels turning and steering as a whole have not been
considered of sufficient importance to warrant
the expenditure of large amounts ()f time, effort
and money.
To support this statement, it may be mentioned
that several previous investigators of the subject
have proposed to measure the lift and moment on
the hull by towing models in yaw at a steady
speed, such as is proposed by I'rofessor Davidson
in the present paper. (For example, see references [45], [49], [50] and [7)3] in the Bibliography
following Chapter IV in "Principles of Naval
Architecture," Volume II.) A research program
of this nature was also seriously considered at the
U. S. Experimental Model Basin about 1928, but
after some deliberation was rejected as not likely
to produce results commensurable to the ext)ense,
time and effort involved. Apparently, then, there
are difficulties in this proposition that require
clarification.
It has been stated that up to now directional
stability of surface vessels has been considered of
secondary importance. This, however, has not
been the ease for other automotive bodies such as
submarines, torpedoes, airships, airplanes and so
forth and quite an extensive literature on the directional stability of such bodies has been built up.
In this literature, one frequently meets two terms;
namely, static directional stability and dynamic
directional stability. Static stability is said to
exist if the center of pressure on a body being

ON THE TURNING

AND STEERING

towed at an angle of yaw in a straight line lies aft


of the center of gravity of the body. Dynamic
stability is said to exist if a body moving in any
p a t h with its long axis tangent to this p a t h has the
tendency to return to the initial course when temporarily deflected from this course b y the action
of extraneous forces. Casually considered, the two
statements appear to mean the same thing; actually they do not. Static stability gives us the
location of the center of pressure with reference to
the center of gravity for a fixed yaw angle; dynamic stability tells us what happens next when
the y a w angle changes.
If the dynamic effects are overlooked, erroneous
conclusions m a y be drawn from experimental resuits. For instance, we know of several cases
where towing tests in y a w have indicated directional instability.
Nevertheless, these same
bodies when propelled b y their own power plant
have proved to be stable. T h e answer is this:
When a body, say a ship, is deflected off the course
under the influence of wind and waves, it begins
to travel in a curved path. T h e curvature of this
p a t h changes the flow about the ship in such a way
that the directional stability is increased or, as
some writer has expressed it, curvature of the p a t h
acts like an invisible rudder.
Dynamic stability can be measured in the model
basin as well as static stability. However, we
must know first the quantities to be measured.
To obtain these, one usually starts with the fundamental equations of motion and makes simplifying
assumptions to arrive at a differential equation
t h a t can be handled by simple mathematics. This
process is outlined in "Principles of N a v a l Architecture," Volume II, pages 217-218. One then
expresses the forces and moments t h a t come into
play as simple functions of the drift angle and the
angular velocity. This leads to an expression in
which constants appear t h a t can be determined b y
experiments.
An a t t e m p t to draw far-reaching conclusions
from static stability measurements alone, such as
has been c~one by Professor Davidson on pages 302
and 303 of the paper, is likely to lead to error. For
instance, on pages 301 and 302 Professor Davidson
states t h a t although model A has a much smaller
minimum tactical diameter than model B and
hence would be expected to have poorer steering,
Fig. 12 indicates t h a t this is not the case. From
this he draws the conclusion t h a t good turning
qualities are not invariably purchased at the expense of steering qualities. Now, when the problem is considered from the standpoint of dynamic
stability, the governing factors are not alone the
actual values of the moments and forces t h a t come
into play, but also their rates of change with yaw

OF SHIPS

517

angle. Looking at Fig. 12 of the paper, it will be


seen t h a t the slopes of the m o m e n t curves at zero
y a w are much steeper for model A than those for
model B. This indicates t h a t the yawing moment on model A builds up much more quickly
than on model B and more powerful rudder action
is required to correct this yaw. I t seems not surprising to us, therefore, that the prototype of
model A has proved to be a difficult vessel to steer
in a seaway as reported. The remedy would seem
to be enlargement of the deadwood aft so as to introduce damping action and improve the inherent
instability of the hull.
The method devised b y Professor Davidson for
expressing the speed reduction in a turn as a function of the centrifugal force and t[ae hull resistance
is believed to be a good one. This necessitates
estimating only one unknown, the radius of the
turning circle, from the geometrical properties of
the hull and the rudder, rather t h a n two.
I t is stated in the paper that, next to length, the
beam-length ratio of a vessel has a powerful effect
on the turning radius. This conclusion agrees
with t h a t reached from a statistical analysis of
ship trial data as given in "Principles of Naval
Architecture," Volume II, page 217.
Reverting to the main theme of the paper,
probably every investigator will agree with Professor Davidson's concluding remarks t h a t the
prerequisites for a comprehensive s t u d y of the
problem comprises (a) the development of suitable equipment and procedures for experiments
in model size, (b) the development of simple
analytical methods for handling the experimental
data and correlating the results, and (c) the
identification of the major variables or parameters. I t is believed, however, as indicated in the
preceding discussion, t h a t the major variables are
already well known and t h a t suitable analytical
methods have already been developed. W h a t is
needed now is the equipment and the personnel to
set up and carry out a comprehensive research
program. The staff at the T a y l o r Model Basin
has been intensely interested in such a program
but as y e t has not had the time to give it much
attention. I t is hoped t h a t Professor Davidson's
timely and interesting paper will be a starting
lever to get such a program under way.
I t is felt t h a t the author is to be congratulated
on this very interesting and valuable paper.
MR. A. J. DAWSON, Member: Professor Davidson's studies offer a valuable contribution to this
generally neglected portion of naval architecture
and suggest methods for continuance of studies
desirable for further clarification of this prime
problem of directional control.

518

ON T H E T U R N I N G A N D S T E E R I N G OF SHIPS

Push towing of barge flotillas as established on


the western rivers introduces unusual problems
along these lines. To be specific, I will offer the
circumstances o( two existing fleet combinations:
Fleet A has an overall length of towboat and
barges of 9;55 feet. The lateral plane for barges
alone equals 570(I square feet when loaded, and
1990 square feet when light. The towboat itself
has a lateral plane of 950 square feet. The steering rudders total an area of 140 square feet. We
then find the ratio of rudder area to total lateral
plane for the loaded fleet is 2.1 per cent. When
the barges are empty" this ratio increases to approximately 7 per cent. The steering gear is designed to operate from hard-over to hard-over in
about 12 s e c o n d ' u n d e r full-load conditions. When
such a towboat is running light (that is, without
barges), we then find the ratio of rudder area to
lateral plane to be about 14a~" per cent, and under
these conditions the rudder will run from hardover to hard-over in about 9 seconds.
Fleet B shows an overall length of 74,5 feet, a
ratio of rudder area to total lateral plane of 2.;{
per cent; ratio with barges light of 7.5 per cent,
and rudder area ratio of 15 per cent of the lateral
plane of towboat only. Here again the steering
gear will run the rudder from hard-over to hardover in about 12 seconds under full-load conditions and in about 9 seconds with the towboat
running free.
Each of these towboats, I feel certain, will meet
Professor Davidson's suggested two lengths turning radius although no accurate check has been
made. I do know that, when one of these towboats is running full speed light and the rudders
are run hard down in 4} ~ seconds, it is advisable
for a person standing on the stern to keep himself
well braced or he m a y find that the boat has
moved out from under him.
Needless to say these towboats when operated
without barges are tender on the helm.
An example of commercial push towing where
steering requirements approach the maximum is
included in a current survey for the development
of a feeder river where some of the bends have
limiting bank radius as low as 750 feet. It is
planned to use fleets totaling 450 feet in length.
The distance from the bank to mid-channel adds
75 feet to the radius or we are manuevering a 450foot fleet around the bend on a diameter of 1650
feet giving a ratio of ;3/2.~to 1. This requirement of
near ultimate in manuevering is further complicated b y stream restrictions showing a 150-foot
wide channel on the straight ways increased to a
2204oot wide channel on the bends. At certain
seasons of the year there will be substantial currents. Fortunately, the total length of arc of the

minimum radius will probably be under 4,5 degrees


and the lead-in radius on both sides will be about
twice the minimum, nevertheless it is a neat navigation problem and for this fleet we are proposing
twin-screw towboats with total rudder areas not
less than 21 !~ per cent of total lateral plane, possibly more.
The simple evaluation of lateral area ratio,
while informative, is not conclusive. Professor
Davidson refers to the lack of directional stability
frequently encountered. This lack of directional
stability or tendency to yaw is, I believe, more
pronounced with barge forms than with normal
ship-shaped hulls, such condition maintaining
whether we are discussing a single barge or a flotilla of barges tightly lashed to form a single unit.
When push towing barges or fleets of barges in
restricted waters, the forces in action, specially
those produced by water being forced around the
relatively hard corners at the bow, are very powerful and while basically symmetrical are easily disturbed into instability by comparatively minor
exterior forces. Once they are disturbed, the unbalance is accumulative and constant corrective
steering control of some sort is imposed. Lengthto-beam ratio appears to have some influence.
The magnitude of these unbalanced forces is
difficult to measure but we have learned ways and
means of modifying the various influencing characteristics to offer some reduction in the degree of
corrective steering required.
Two hulls may have the same profile of lateral
plane, but widely different values of directional
instability in motion. We have found t h a t an
easing off of the shoulders either by drawing in
the deck line or increasing the radius of the bilge
turn at the rake ends, or both, offers material relief along these lines. Barges having large bilge
radius, as for instance a radius equal to the draft,
are much more difficult to control than those offering a vertical or nearly vertical side and a bilge
radius of, say, one-eighth the draft. In very
shallow water there appears to be some favorable
return, after accepting the penalty o} reduction
in displacement, for a moderate rise of bottom;
this rise of bottom m a y take the form of a reversed
camber.
The well-known sternwheel towboat with a
shallow, flat-bottomed hull and simple scow ends
frequently required the application of a tripping
skeg forward and manuevered largely by flanking
the stern around so that in effect the boat was reaimed at a new target before proceeding ahead.
There remains much work to be done along the
lines as developed b y Professor Davidson and I
sincerely trust he will pursue the problem further,
reporting to the Society so that we m a y all benefit.

ON THE TURNING

AND STEERING

I wish to congratulate Professor Davidson for his


very excellent presentation on these important
problems.
)IR. JOHN B. DRISKO,4 I Tsitor: Dr. Davidson's
paper provides a great deal of food for thought.
Fig. 3, based on tests of numerous ships and models, relates speed loss, advance and transfer to tactical diameter in such fashion as to show that, except for the minimum diameter attainable with
full rudder, all ships turn very much alike. This
provides a great simplification for further studies;
it shows, for instance, t h a t preliminary comparative studies can be made on the basis of steady
turning circles, which are somewhat easier to deal
with than initial transient phases of turns.
The relation of steering to turning is a puzzling
one. The curves of Fig. 12 show models A and B
to be directionally unstable in straight line motion. Fig. 11 shows, however, t h a t when the models are turning in diameters of 5.9 lengths, model B
needs 35 degrees of rudder to force it to turn,
whereas model A turns with only 6 degrees of rudder and, since the hull force acts ahead of the center of gravity, it shows a negative rudder force;
i.e., the model wants to turn shorter t h a n the rudder will permiL
At a larger diameter, of 12.~i lengths, model B
still requires 15 degrees of rudder, while model A,
to hazard a guess, would probably need little or no
rudder. This assumption regarding model A was
substantiated b y recent tests.
The two models m a y perhaps be characterized
as follows:
Model A is unstable in straight running, in fact
so unstable that, with rudder amidships, it wants
to turn in a diameter of around 12 lengths. A
diameter of, say, 15 or 20 lengths would be attainable only with a negative rudder setting. Thus,
for an "indeterminate zone" of diameters larger
than 12 lengths, and on up to infinity, the ship
would seem to be controllable only b y the use of
negative rudder, or by ceaseless adjustment of the
rudder.
Model B is also unstable in straight running,
but it requires as much as 15 degrees of rudder to
turn in a diameter of 12 lengths. I t seems safe to
guess that, with its rudder amidships, it will turn
in a very large diameter, say 50 or 100 lengths.
The helmsman thus has adequate control over diameters smaller than this 50 or 100 lengths; and
in the indeterminate zone of diameters larger than
this the course is so near to straight running t h a t
there is no pronounced steering problem.
The difference between the two models there4 S t e v e n s E x p e r i m e n t a l T o w i n g T a n k , Hoboken, N. J

OF SHIPS

519

fore seems to be not in the fact of their directional


instability in straight running, b u t rather in the
extent of this indeterminate zone of large diameters t h a t are not directly attainable with a single
fixed rudder setting.
Increasing the size of the rudder on model A
would result in smaller minimum turning circles,
and would also increase the size of turning circles
at small rudder angles, where the hull force acts
ahead of the center of gravity and the rudder
force is negative. This would cut down the indeterminate zone and improve the steering behavior
of the ship.
Increasing the size of the rudder on model B
would also give smaller minimum turning circles,
but it would probably affect the steering v e r y little
since the indeterminate zone is already acceptably
small.

CAPTAIN CIIRISTIAN BLOM, Member: The author should be complimented on attacking a subject so complicated as the turning and steering of
ships, and on dealing with it in a masterly way.
I t is particularly the influence of the hull t h a t
makes the m a t t e r involved, and I quite agree with
the remark t h a t a better knowledge of hull action
is a prerequisite to the whole subject.
The way in which the work on the problem of
turning is tackled seems to me quite convincing.
I t brings forth i m p o r t a n t factors that, to m y
knowledge, have as yet been passed unnoticed b y
the profession and will be of great help in figuring
out probable turning qualities in future designs.
I think most naval architects will bear the
author out in his r e m a r k on steering, t h a t "the
whole m a t t e r will bear further study," and it is
again the knowledge of hull action t h a t is found
lacking.
The question of good steering is of importance
primarily in a seaway, when the ship is incessantly rolling more or less. Experience tells us
t h a t some ships, which could be steered all right
when in sheltered waters, in a rough sea start pronounced yawings t h a t the rudder cannot take
charge of, while other ships are noticeably easy to
steer under all circumstances. This difference in
behavior m u s t arise from action of their hulls alone.
In this connection, I beg to call attention to two
papers read before the Institution of N a v a l Architects, London, (1937 and 1942) b y Engineer Admiral Alfred Turner, "On the Balance of Heeled
Ships." These papers give evidence t h a t ships
which are not "balanced" will deviate from their
course when inclined. ( " N o t balanced" is, in Admiral Turner's terms, a ship where the listed centers of gravity o f the different underwater section areas projected on the listed waterplane do

320

ON THE TURNING

AND STEERING

not come in line with or in line parallel to the intersection of this waterplane and a vertical plane
drawn parallel to the ship's course through the
ship's listed center of buoyancy, b u t follow an undulating line, the axis of which the ship will follow
despite the counter-action of the rudder.)
A ship which is set rolling will, if not balanced,
for each heel to starboard swing out of its course
to one side or the other (according to its special
form of hull) and to the opposite side when heeled
to port. As the center of buoyancy of the ship, at
times, also shifts in a longitudinal direction, when
the ship heels over there will at the same time
be a change of trim accompanying the heelings,
and so the ship is constantly p e n d u l a t i n g - - t o starboard and port, up and down.
These are the m o v e m e n t s when rolling in still
water, to say nothing of the movements in a rough
sea.

The problem of designing a "balanced," normal


seagoing vessel t h a t steers well under all circumstances is rather involved and requires much accurate drawing and calculation. But for a modern experimental towing t a n k with an experienced
staff, where trials could follow calculations
through the different steps of the research, it appears to me to be a feasible although not an easy
job. If undertaken, it will likely shed useful light
on steering of ships and other important sides of
the o p t i m u m hull form.
COMMANDER E. A. WRIGHT, U.S.N., Associate
.llember: In his new maneuvering basin at the
Stevens Institute of Technology, Dr. Davidson is
doing a masterful job of defining the boundaries of
turning investigations and of spotlighting the
significant variables. However, if " t h e principal
burden of this paper" is to pave the way for research on the influence of hull characteristics on
the turning of surface vessels, the author has not
carried it nearly far enough.
Several of us at the T a y l o r Model Basin have
believed strongly for a number of years t h a t
knowledge of the turning behavior of surface vessels will never be systematized in a form suitable
for design purposes until the theory of turning is
attacked vigorously. T h e basis of such an investigation must be the differential equations of motion, which can be found in most textbooks on
naval architecture. These equations are general
in form and can be applied to transient conditions
such as zigzagging and steering as well as to steady
turning. However, in a broad and fundamental
paper of this nature which presumably is intended
to deal with the scope of the turning problem, the
equations of motion do not appear and no mention
is made of them.

OF SHIPS

As a start in the right direction, it is gratifying


to see t h a t the Experimental Towing T a n k is beginning to work with hull and rudder forces or the
dynamics of turning, rather than simply the geometry or kinematics of the motion. The technique
of determining the lift on the rudder to arrive at
the lateral force and moment on the hull is a sound
procedure, b u t it is inadequate for a systematic
fundamental investigation of turning and steering.
Separation of the variables is the accepted approach to a n y complicated problem of this nature.
Dr. Davidson recognizes and emphasizes this consideration, but the paper leaves the impression
that the experimental technique described is sufficient to attain this objective. When Admiral
T a y l o r approached the problem of the speed and
power of ships, he treated the variables separately.
Resistance tests were run on hull forms alone and
propellers were characterized in open water. Not
until 1,q years after the Experimental Model Basin
was started were ship models self-propelled.
A similar procedure must be followed for study
ing ship maneuvers. In addition to rudder tests
in open water, a number of which have already
been made, a systematic series of tests must be
made on hull forms alone without rudder or propellers. The forms must be towed at different
drift angles in eurvilinear paths of different radii.
A thorough understanding of the effect of each of
the principal components must be obtained before
they are put together and their interaction studied.
To separate the variables as outlined, special
test facilities must be developed. One form of
apparatus to study hull forces alone during turning consists of a rotating arm which tows the
model in a circle. A simple form of this device was
used b y the Stevens Institute of Technology in
early investigations in this field. I t is understood
that Dr. Davidson is considering adding a rotating arm mechanism to his maneuvering basin,
which will certainly be another step in the right
direction.
Ultimately it will be necessary to have what
might be called a 2-dimensional towing carriage
for turning tests, exactly as we have a 1-dimensional carriage for resistance and propulsion tests.
The 2-dimensional towing carriage must permit
m o v e m e n t on any curvilinear p a t h in a horizontal
plane, in order to undertake experiments covering
initial turning movements and complicated attack
or evasive maneuvers of surface vessels, and to
s t u d y route stability in colmection with steering
problems. Fairly large models m a y be necessary
to carry the m a n y types of instruments t h a t will be
required. For example, the wake distribution in
way of the rudder will be measured while the

ON THE TURNING AND STEERING OF SHIPS


model is actually turning, exactly as is now the
practice for wake surveys at the propeller positions. The design and construction of a 2-dimensional towing carriage will be a difficult and costly
undertaking, but experience with m a n y problems
during the war has clearly demonstrated t h a t some
test facility of this type must come if we are to go
much beyond our present frontiers of knowledge
of turning and steering.
To obtain an understanding of the fundamentals
of turning and to design for maneuvering characteristics, it is necessary (I) to build special test
facilities for this purpose such as a rotating arm
and ultimately a 2-dimensional towing carriage,
(o) to t r e a t the variables separately before studying their interaction, and (3) to plan experimental
programs to obtain the coefficients in the equations of motion.
During the past few years, the Experimental
Towing T a n k at the Stevens Institute of Technology has been called upon for a great deal of research on the turning of surface vessels for the National Defense Research Committee and the N a v y
Department. I t is expected t h a t Dr. Davidson
will continue to enlarge the scope of his experiments until basic data on this problem are produced.
CAPTAIN C. D. WIIEELOCK, U.S.N., ~l[ember:
Dr. Davidson is to be respected for having attacked the baffling problem of steering fundamentals and for having discovered through analysis of model and trial data what appear to be some
of the i m p o r t a n t parameters t h a t must be used
eventually to describe a hull's behavior during
yaws and turns. This is pioneering research of
real importance.
The ship designer has so far been fortunate in
not being pressed for predictions as to turning
performance. Were it otherwise he would be in a
very difficult spot indeed. Usually he is in a position to provide a rudder of more or less conventional size and shape and then to let the tactical
diameter fall where it will. Turning experiments
with self-propelled models sometimes afford a fair
means for predicting full-scale results, b u t the design upon which the model is based is largely a
t h u m b rule product. The bigger the rudder the
shorter the turn is one of the rules. If his ship
should be found to have a larger rudder t h a n
necessary, and this has happened, the rudder can
be limited in angle of travel by resetting the hardover stops.
The factors of ship dimensions, coefficients and
hull form contributing to a minimum of resistance
to propulsion have been and probably always will
be of p a r a m o u n t importance. Changes to favor

321

turning and steering can be made only if they result in no increased resistance to propulsion.
This does not mean t h a t there is reduced need
for a better understanding of the behavior of ships
in a turn. On the contrary, improvements of the
hull and the rudder can be advanced only through
better understanding. This work begins to throw
light on the sea of conjecture in which we have
been floundering. Dr. Davidson has his foot in
the door. I have faith he can push it open.
PROFESSOR DAVIDSON: I had naturally hoped
t h a t this paper would provoke discussion. T h a t
it would provoke so much highly instructive discussion, in these busy times, I had not dared hope.
I t seems clear t h a t the subject is one of considerable interest, and t h a t there is plenty left to
be learned a b o u t it.
The paper a t t e m p t s to do just three things:
First, to present in broad perspective a picture
of the turning characteristics of existing vessels.
This is based on actual turning-test data; its objective is to "see where we've been and to t r y to
find out where we want to go."
Second, to bring out the need for a systematic
experimental s t u d y of the variables governing
turning characteristics, and to suggest a few first
steps in a general attack.
Third, to emphasize the need for considering
steering characteristics along with turning characteristics, particularly in conjunction with any
a t t e m p t to improve on the best turning characteristics of the present day.
The very first sentence in the paper is "Ships
have been turned and steered with reasonable success for centuries." F r o m this, it would not have
been unreasonable to expect t h a t a n y group, like
ourselves, undertaking serious work on these subjects at the present time would have found a fairly
well-established framework of knowledge on which
to build. We might have expected to find ourselves concerned, not so much with broad exploration, as with clarification and elaboration of detail. This, however, was not found to be the case,
and the overall survey of the turning characteristics of existing vessels, in the first portion of the
paper, was really undertaken on this account. I
think t h a t the discussions which have been presented, when taken as a whole, tend to confirm
the fact t h a t clarification was needed, and t h a t
even more would be in order.
A number of the discussions dwell on the importance of good steering (course-keeping) qualifies rather than of good turning qualities. I t is
particularly interesting to find, however, t h a t
most of these discussions have come from individuals whose p r i m a r y concern in peacetime is

522

ON THE TURI~ING AND STEERING

with merchant types. I am sure t h a t this should


not be taken to mean t h a t merchant ship operators
are in any sense opposed to good turning qualities
in their vessels; I think it means simply t h a t they
are more directly concerned with having good
course-keeping qualities, and t h a t they m a y be a
little pessimistic--through hard experience--of
getting both. The moral would seem to be t h a t
steering qualities ought to be given at least as
much emphasis as turning quahties in future
studies.
An overall survey of the steering characteristics
of existing vessels, generally parallel to the survey
in the p a p e r for turning characteristics, would
naturally suggest itself to an individual of m y
t e m p e r a m e n t as a logical first step. Such a survey
would be difficult to m a k e a t the present time, however, for the reason t h a t there is no generally accepted criterion of steering ability. The minim u m tactical diameter (rudder hard-over), expressed as a multiple of length, is a simple criterion
of turning ability and one which has been found
quite satisfactory. There is no criterion of steering ability, t h a t I know of, which is as simple or as
generally acceptable. Until a satisfactory criterion can be decided upon (that is, until we can
state in s!mple terms just what we mean b y good
or bad steering qualities), there really is no way of
making an adequate overall survey for existing
vessels.
I w a n t to m a k e myself perfectly clear on this
matter, because I think it quite important. The
paper m a y imply, although I did not so mean it,
t h a t a criterion of steering might be the number
of degrees of rudder angle necessary to correct a
given n u m b e r of degrees departure from the desired course in cahn water. This, however, is obviously an insufficient criterion; it does not state
how quickly the course shall be corrected, nor how
quickly or often the rudder m u s t be readjusted
to prevent over-correction. In fact, it disregards
the whole m a t t e r of time intervals. I am very
much aware of time intervals, and Captain Saunders has drawn attention to t h e m in his very interesting remarks on the handling of large and
small submarines.
Of course, if the thesis were true t h a t good-turning ships are hard to steer, and vice versa, then
t h a t in itself might perhaps be said to settle more
or less the problem of a steering criterion. I am
convinced, however, t h a t this thesis is not necessarily true, at least in principle, however true it
m a y or m a y not be for what the paper refers to as
" n o r m a l seagoing ships." And the last sentence of
Captain Saunders' discussion makes it clear t h a t
he concurs with m y view. Thus, the question of a
suitable steering criterion remains; I believe we

OF StlIPS

should try to settle on one as quickly as we can.


Several of the discussions--notably Commander Wright's and Dr. Sehoenherr's and, to a
lesser extent, 5{r. Robinson's--call attention to
the basic differential equations of motion, and to
the fact t h a t they were not mentioned in the
paper. I am very well aware of this theoretical
approach; it is fundamentally the same as t h a t
made b y Hunsaker and others to the dynamic
stability of an airplane in flight, along about 191.5,
and later b y Glauert and others to the porpoising
of a seaplane during take-off and landing. There
is no question but that it is basically sound. The
only question has been whether or not it was necessary, since it does not lead to rmmerical solutions
without the introduction of empirical constants.
In the case of seaplane porpoising, where I have
had some experience with it, the thne and labor
involved in determiIiing the necessary empirical
constants (slopes) has, so far, appeared to outweigh any advantage t h a t might be derived from
the theory, and direct experimental determinations
of stability have been found satisfactory. The
same thing might conceivably be true in the
present instance. However, an increased emphasis on steering, and, in particular, the problem
of establishing an adequate criterion for steering,
certainly focus attention on the theory as the
best means of clarifying fundamental concepts
and of establishing an understanding of both
steering and turning characteristics on a firm
basis. Steering and turning m u s t go hand in hand
since we are in some doubt about one of them
steering. If we could confine our attention to
steady-turning alone, I do not think t h a t we would
need to pursue the theory very far. Steering,
however, brings up the whole m a t t e r of directional stability, probably in b o t h its static and
dynamic aspects, and this is a m a t t e r for the
theory.
As a m a t t e r of fact, the experimentally derived
forces and moments which are referred to in the
later portions of the paper are precisely those
needed to determine the constants in the theoretical equations. Thus, a trial application of the
theory can be undertaken on the basis of the data
given for models A and B. Work in this direction has been started, at the present writing, usinK
somewhat extended data of the same sort given.
Such special problems in controllability as steering in rough water, in shallow water, and when
going astern, have been brought up. To these I
might add one which has a special appeal to me;
namely, the question of broaching in h e a v y quartering seas. All of these " a b n o r m a l " problems are
obviously important, and some of them m a y even
be crucial. In the case of airplanes, abnormal

ON T H E T U R N I N G A N D S T E E R I N G O F S H I P S
flight situations, rather than normal, often determine control characteristics in many respects,
but with ships I do not think the special problems
can be attacked intelligently until we have a clear
understanding of turning and steering in calm
water.
,
A number of specific questions have been
raised.
Admiral Cochrane calls attention to the spread
of the test points in Fig. 3, and suggests that
there are " . . . a number of factors involved which
need further exploration." This is undoubtedly
the case. The purpose of Fig. 3, and of the
reasoning based upon it, is to get a broad perspective. I am prepared to defend it on that
ground, and, as long as it is understood for what
it is, I think it may well serve as a good springboard for the next jump.
Admiral Coehrane mentions also the question
of excessive angle of heel in very short turns at
high speed. Having recently spent a good deal
of time on this particular question in certain specific cases, I am not likely to underestimate its
importance. A reasonable accumulation of general information regarding it is already available
from tests; the paper does not dwell upon it only
because I could see no possibility of covering all
of the m a n y detailed aspects of turning and steering at one time.
Dr. Goldsmith discusses the question of
whether the hyper-sensitive helm of model A is
good or bad. Observations of the actual steering
characteristics of the corresponding full-size vessel itself, under ordinary service conditions, have
been made since the paper was written. It seems
quite clear from these that this design is somewhat deficient" in steering ability. The ship is
hyper-sensitive on her helm in the sense that large
changes of turning circle result from very small
changes of rudder angle; but in steering on
straight course her response to rudder applications is at first sluggish and then rather extreme.
I would like to emphasize, however, that I consider Dr. Goldsmith's discussion to be of precisely
the type which is required if we are to reach a
satisfactory definition of good steering.
Mr. Vincent's very interesting remarks regarding directional stability are noted. I t is
hoped that the more concerted attack on the steering problem as a whole, which has now been
started through the theory, will help to clarify this
matter.
Mr. Bates has brought out the important influence on turning of the after underwater sections, and has suggested t h a t the paper does not
give enough emphasis to these. Our experience
at the Experimental Towing T a n k amply bears

323

out his statements regarding the effect of broad,


shallow afterbodies and of the amount of deadwood. It should be noted t h a t these are matters
of hull form, and that the paper emphasizes, as a
general proposition, the importance of hull form.
It could not deal broadly with the ramified aspects of hull form without departing from its main
purpose. But I think Mr. Bates is right in what I
take to be his contention, that in emphasizing the
influence of the most outstanding difference of
hull form between models A and B (namely, the
greater relative beam of model A) I m a y have
tended to imply that this difference is more significant than others. The point is well taken;
I had no such intention.
Commander Manning points to certain wellknown equations [2], and to the conclusions to
which they lead. I think it has been generally
agreed for some time that the trouble with these
equations is not so much that they are wrong as
that they are incomplete; hence, conclusions
based on them will naturally suffer from the same
cause. T h e y constitute, in effect, an over-simplification of the problem.
Commander Manning asks why the values of
advance, transfer and tactical diameter are defined with respect to the turning path rather than
to the heading. This was done entirely as a matter of convenience; in particular, to avoid the influence of drift angle on these quantities and
thereby to eliminate one uncertainty in their determination. As pointed out in a note under Fig.
1, this departure from the definitions of ordinary
naval usage affects principally the value of the
transfer, which is really the least important of the
three quantities. A return to the more customary
definitions is contemplated when the behavior of
the drift angle is better understood.
Mr. Robinson brings up several pertinent matters. First, with respect to Figs. 3 and 4, he points
out that the advance appears to be made up of
two parts, a transient portion which is roughly
constant, plus one-half of the steady-turning diameter. Recognition of this fact may well lead
to further simplification; this is now under study.
The speed reduction curve in Fig. 3 is, as Mr.
Robinson says, based upon a very broad assumpt i o n - a c t u a l l y one or both of two equally broad
assumptions. The only justifications I have for
these assumptions at the present time, beyond the
correlation shown by the chart, are that the two
additional force components to which they relate
are obviously compensating in sign, and that one
of them, the increased drag of the hull, was
estimated as small b y Hovgaard [5], and would
be expected to be reasonably small in view of the
comparatively small magnitudes of the dimension-

524

ON THE TURNING

AND STEERING

less hull force eoellicients found from experiment.


A general tightening up with respect to these assumptions is a logical step for the future.
I agree heartily with Mr. Robinson t h a t the
terms "lift" and "drag" should be reserved to
denote forces at right angles and parallel, respectively, to the flow. T h e y were used otherwise
in the present instance for the sake of vividness,
but perhaps that is an insufficient reason. I do
not agree entirely with Mr. Robinson t h a t the
drift angle at the stern (rudder stock) should always be used in dealing with the rudder, because,
in a great m a n y cases, the interference effects of
hull and propellers, on the rudder, have been found
to be so large t h a t the actual angle of attack of the
rudder bears little relation to this geometric drift
angle. The mass and m o m e n t of inertia of the
entrained water will have to be considered, of
course, in undertaking more exact analyses than
those given here.
Mr. W. W. Smith's colmnents, with respect to
twin rudders and contra-rudders, are of particular
interest in connection with rudder design, and,
ultimately, deserve attention.
I have already referred to Dr. Sehoenherr's
very interesting comments in discussing the general proposition of attacking the whole m a t t e r
of steering through the theoretical equations.
At the end of his discussion, however, I believe
t h a t Dr. Schoenherr creates an impression t h a t
work on the three prerequisites mentioned in the
paper, as necessary for a comprehensive s t u d y of
turning and steering, is further advanced than it
really is. The theory is well known, to be sure,
but there are m a n y problems in its application
t h a t have not been worked out. Nor do I agree
t h a t " . . . the major variables are already well
known," if b y "major variables" he refers to the
ordinary differences of design.
C o m m a n d e r Wright argues t h a t the paper does
not go nearly far enough in paving the way for
research on the influence of hull characteristics
on turning. No one is more fully aware of this
than I. The p r i m a r y purpose of the paper, as I
have said before, was to gain perspective. I agree
with C o m m a n d e r Wright t h a t a comprehensive
program is called for to clarify fully all the ramifications of the turning and steering problems as
these apply to ships. I sincerely hope, however,
t h a t it will not require eighteen years to get at the
fundamentals. As things appear to me at the
present writing, too close an analogy to the propulsion problem m a y be a little dangerous. Without in the least meaning to minimize the impor-

OF SHIPS

tance of thrust deduction, etc. (that is, of the interference effects in the case of propulsion) I
strongly suspect t h a t interference effects in the
case of steering and turning---between hulls, rudders, and propellers--are essentially larger. If
this turns out to be so, it would be dangerous to
proceed too far without giving t h e m a prominent
place in the studies.
C o m m a n d e r Wright mentions the use of a
rotating a r m to provide for towing models in a
uniform curvilinear p a t h in the same way t h a t
they are towed in uniform rectilinear p a t h in an
ordinary towing tank, and has pointed out t h a t a
small-scale version of this device was used b y the
Experimental Towing T a n k in early investigations. A larger one is now under construction for
use in the new maneuvering tank.
In reviewing the discussions as a whole, for the
purpose of these closing remarks, I have been impressed anew b y the extraordinarily wide range
which they cover, and b y the v e r y considerable
interest in the subject which they reveal. Each
one adds to the overall picture, and it is an education to read t h e m through consecutively. However, notwithstanding their breadth and the
nmnber of points of view they represent, I think if
one reads them carefully he will be impressed, as
I have been, with the extent to which they tend
to focus on one central theme. This is, as I see it:
(1) T h a t there is a definite need for a fuller
knowledge of the whole subject.
(2) That, in further work, steering should receive at least equal emphasis with turning.
C3) T h a t the real problem is to improve the
turning and steering characteristics of vessels
without material detriment to the resistance and
other i m p o r t a n t characteristics; f6r t h a t matter,
without having to alter designing practices in any
radical way on their account.
(4) That, in general, the s t a r t already made is
in the right direction, but t h a t greater concentration on fundamentals would be desirable for the
future.
I t is very stimulating to find t h a t the work discussed in the paper has m e t with general approval,
and to find such universal interest in carrying it
farther.
REAR ADMIRAL OEORGE H. ROCK, Chairman:
On behalf of the Society, I t h a n k Dr. Davidson for
the very excellent paper he has prepared and also
those who have taken such an interest in discussing it, bringing out so much t h a t all of us wanted
to hear.

You might also like