Beruflich Dokumente
Kultur Dokumente
253
I. GENERALINTRODUCTION.
The use of iron in the form of chain probably represents one of
the earliest applications of this metal t o the service of man and very
early examples, still in existence, show what a high degree of skill
in the technique of forging and welding had been acquired even in
the Middle Ages ; many of these examples were used, of course, for
purely decorative purposes or as fastenings. The use of chains
definitely for lifting purposes, with the consequent employment of
hooks, rings, and other fittings, probably became general only in
later times : but even in this application, the use of chain must be
of very great antiquity. It is not surprising to find, therefore, that
the forms and dimensions of these components, in relation to the
loads they are required to bear, were decided largely from accumulated
experience acquired before scientific methods of design had been, or
indeed could be, attempted. As a result, a very wide diversity of
types and designs are in current. use. In some cases, regulations
concerning such components as chains and hooks appear in certain
official codes or regulations, or even form the subject of legislation ;
mention may be made of the following : Anchors and Chain Cables
Act, 1899; Factory and Workshop Acts, 1901 to 1923; Docks
Regulations, 1925 ; Building Regulations, 1926 ; Shipbuilding
Regulations, 1931 ; Coal Mines Act, 1911, etc. The majority of
these Codes and Regulations relate chiefly to the specification of
the material, treatment, testing, etc., and the dimensions of the
components in relation to the loads are not usually specified.
With regard to the general analytical methods of stress estimation
in chains, hooks, and rings, these are, of course, treated in standard
textbooks,* whilst very valuable notes on current practice are given
in Mr. Stevenson Taylor’s memorandum.? In spite of this available
information it appears that the possibilities of the application of
scientific principles to the design of lifting gear components have
not yet been sufficiently recognized; although the majority of
components in use may be regarded as representing safe practice,
many are certainly far from being economical.
~ __- -- __ .~
* e.g. Strength
“ of Materials,” Morley, 7th ed., I928 (Longmans,
Green).
t ‘’ Memorandum on Chains and other Lifting Appliances,”
G , Stevenson Taylor, M.I.Mech.E., 1929 (H.M. Stationery Office).
11. GENERALNATURE
OF THE PXOBLEM.
The majority of lifting gear components consist essentially of
curved beams of very ductile material ; in many cases the radii of
curvature of portions of the components are relatively small and
many are necessarily of such shape or design that the induced forces
and bending moments cannot be derived directly from the values
FIG. 1 .
Hence
r
and if we put = k = sin 2w
R
where n is the ratio of the width of the section a t the intrados to the
width a t the extrados and z is the ratio of the radius of curvature
-_- ~ ~~~~ -. .~. -- ~ - ....
*
SCHUSTER, L. W. ’‘ Comparison of the Stresses in Curved Beams
with a Circular and a. Trapezoidal Section,” Technical Report, British
Engine Boiler and Electrical Insuresce Company, 1926, p, 162.
95
K
/
I-
2Y
w
2
t
24
0
2
I c 5
rule or some other system) and the area A' from derived ordinates
calculated from the measured ordinates of the actual section. In
this way, although the area of section may be affected by errors in
drawing or measuring the section, the ratio A'/A will be affected
only by error in the numerical calculations. The latter errors may,
of course, be made as small as necessary by increasing the accuracy
of the calculations.
IV. STRESSES
IN HOOKS.
and fz(compression) = -
.
stresses are then - .- - y1 at theintrados and - - 1 Yz w
A A R-yyl A ' )I ' c+Tz
at the extrados. The tensile stress at the intrados will be greater
numerically than the compressive stress at the extrados if yl/(R - yl)
+
> yz/(R yz) i.e. if R <2y1yz/(yz - yl). Or, if C/2 = R - y1
= radius of curvature of intrados, the condition is C < 2Yl(Y2-- Y1)-
+
Y2 - Y1
FIG.4.
C
_-
- 2Yld or yl/d >tGc. Normally this condition will be fulfilled ;
d-2Y1
but the greatest economy of material would be effected by making
the two stresses equal, i.e. by making yl/d = &C/(d C). For +
normal values of the ratio d/C from 0.84 to 1.00, the required value
of y,/d is 0.27 to 0.25. A triangular section would give yl/d = 6,
so that for this range of values of dlC the condition cannot be fulfilled
unless a re-entrant section be used. (It should be noticed that equal
stresses would be developed in the triangular section if d/G = 0.5 ;
this value of the ratio lies well outside normal practice.)
The above analysis shows that the most efficient section for a
hook would normally be an I, T, or bull-head section ; but that, if
for other reasons such sections are impracticable, the triangular form
should be approached as closely as possible.
The form of section (Pig. 4) adopted in the designs of hooks
recently prepared a t the National Physical Laboratory for the
British Standards Institution represent a reasonable approach to
the ideal triangular form consistent with the limitations of the
manufacturing process. The ratio of the maximum compressive to
the maximum tensile stress is, however, less than 0.5 even in the
best case.
The form of section having been decided, the next step in the
design of hooks consists in the determination of one dimension (say
the depth) of the section in relation to the radius of the intrados.
Variation of the value of the ratio of the depth of section to the
inside diameter of the hook will influence both the area A and the
1
stress factor - -&- .
R-Y1
and the object of design is to make the stress
w 1
.
fi = - - . --'L
A h R-yyl
equal to some specified value. For this purpose
curves such as those plotted by Schuster * for the simple trapezoidal
section form an invaluable guide to the necessary proportions of
hooks having practical forms of section, such as that of P i g 4.
I n the case of hooks of circular section, the stress factor
K=-.
1 r
-=Cot2w
sin 2 0 - - -- --2 cos3w. This
h R-r 1 - sin zw sin ~ ( C O Sw - sin w)2'
has a minimum value equal to 13.5 when 3 sin w - cos w = 0, which
gives k(= sin 2w) = $. A curve representing the variation of this
stress factor with the ratio k is shown in Fig. 5 .
In the case of other practical forms of section, analytical expressions
for the stress factor are not practicable, although curves similar to
that shown in Pig. 5 can, of course, be constructed. However, for the
practical purposes of hook design, other forms of expression of the
relation between load, stress, radius of curvature, and area of section
are more useful.
The dimension 2(R - yl) = C (Fig. 3) is usually termed the
" bed diameter " of the hook and is, in effect, the dimension governing
(5.0-
14.5-
K
0
0
2
ln
w
ln
a
34.0.
13.5
f3.0
( 4 5 0.6 0.7 8
R =r/R
The design problem can then be postulated thus : given that the
“ bed diameter,” C = 2(R - yl) = pd/uT, where p is a constant,
* Some designs of hook (cf. Liverpool hooks, p. 269) constitute a partial
exception to this rule ; but the bed diameter is always the ruling factor
in the accommodation the hook provides.
1.40-
& 1.30
1.20
1-10
2E
or p?hdlC =-= constant
af
. . . . (44
* The value of c i a is for the circular section 21.. and for the standard
trapezoidal section (Fig. 4) 0.95.
Since h is a function simply of dlC, this equation, for any given values
of the factors making up t8heconst,ant on the right-hand side, can be
used to define d/C in terms of p and thus t.0 provide a direct answer
to the design problem. Actually, in practice it is ensiegt t o assume
values of d/C, to calculate the corresponding values of A and hence
to deduce the values of p. Curves in which d/d/w(= pd/C) is plotted
against p, for practical ranges of the latter va.riable, are shown in
1.5
- _-
W ASSUMED LINEAR RELATION
d =1.25+0.10/L
1.4
L A T E 0 VALUES
1.3
5
Figs. 6 and 7 for hooks of both circular section and the trapezoidal
section shown in Fig. 4. I n these Figures curves have been drawn
for a maximum stress at the working load of 12 tons per sq. in. ;
erimilar curves for other maximum working stresses may readily be
obtained since for any given value of d/C, p.2 is inversely proportional
t o the working stress. I n the case of the circular section X = tan2 w
and ajc =
*I# = %sin 2w so that
R--r l-sinh'
2~ cosw (coso - sinw)' -~~-. . . . .
P2 =.f f sin% (abl
2
OB PLASTIC
V. EFEECT DEFORMATIONUPON THE DISTRIBUTION
OF STRESS.
In discussion of the stresses in hooks, attention has so far been
confined to the stresses set up by the working load. The maximum
permissible values of these stresses, as calculated by elastic theory,
are usually of the order of 9 to 12 tons per sq. in.* and, since the
yield stress of the material (wrought iron or mild steel) will not
normally be less than 14 tons per sq. in., the calculated stress
distribution may be regarded as an accurate representation of the
actual stress system set up. On the other hand, if the same analysis
is used to estimate the maximum stresses a t the proof load, usually
equal to twice the working load, the values obtained (18 to 24 tons
m
u)
w
K
i7l
STRAIN
per sq. in.) are considerably greater than the yield stress of the
material. It is clear, therefore, that these stresses will not actually
be induced, for plastic deformation will intervene, and the stress
distribution will be thereby modified.
In the case of some of the chain components discussed below,
calculation by elastic theory of the stresses set up even by the
working loads indicates maximum stresses greater than the normal
yield stress of the material. In these cases, therefore, even if attention
is confined to the working range of load, the modification of the
stress system by plastic deformation must be considered.
Although the redistribution of stress due to yield is necessarily
dependent to a large extent upon the elastic and plastic properties
~~
*
The values adopted in the British Standard Specifications are 9 tons
per aq. in. and 12 tons per sq. in. for two different types of hook (see
Section VII, p. 307).
of t,he material, the fact that the materials, wrought iron and mild
steel, usually used for lifting gear are characterized in the annealed
condition by a well-marked yield (see, for example, Pig. 9) may be
used to develop a general description of the behaviour of beams when
the strain of the outside fibres exceeds the elastic range. In the
following discussion of the effect of yield, it will be assumed (1) that
the material exhibits linear elasticity up to the yield and that
thereafter any further strain leaves the stress unaffected, ( 2 ) that
the yield stresses in tension and compression are equal (unless
otherwise stated), and (3) that plane sections remain plane throughout
the deformation.
Consider the case of a straight beam having a section symmetrical
about two diameters a t right-angles and bent in the plane containing
one of these diameters. Referring to Pig. 10, let the distribution of
strain across the depth A A of the section be represented by the
ordinates from AA’ t o the line PP‘, the tangent of the angle AOP
being inversely proportional to the radius to which the beam is bent.
Let AQ(= A’Q’) represent the strain a t yield. Then, so long as
AP < AQ, the deformation is entirely elastic and the stress
distribution may also be represented by the line POP’; but, if
AP > AQ, plastic deformation must occur in the fibres covered by
the lines QR and Q’R’ and the stress distribution is then represented
by the line QROR’Q’. The bending moment corresponding to this
stress distribution may easily be calculated (account being taken of
the section of the beam), and the relation established between bending
moment and curvature of the beam. It is, however, sufficient to
remark that the apparent proportional limit of the beam as a whole
coincides with the occurrence of yield in the outermost fibres and that
yield of the whole beam does not occur until the angle AOP becomes
very nearly equal to ~ / 2 .If the “ apparent ” fibre stress is calculated
merely from the value of the bending moment, values greater than
the yield stress will be obtained. The ratio of the “ apparent ”
fibre stress a t yield of the beam to the actual yield stress of the
material depends upon the section of the beam; in the case of
rectangular section this ratio is 3/2, for circular section 16/37, and for
diamond section 2.
If, after subjection to the bending moment that sets up a stress‘
distribution such as QROR’Q’, the beam is again unloaded, it will
not completely recover its original form. The amount of the elastic
recovery of any fibre will, in fact, be the same as if the bending moment
had set up not the actual stress distribution QROR‘Q‘ but the
“ apparent ” distribution NO”. After release the beam will be
self-stressed by the self-equilibrating system STOT’S‘, the outer
18
fibres on the tension side being left in compression and the inner
fibres on this side in tension. with similar but reversed stresses on
the compression side. It will be seen that repeated application of
the same load will not cause further yield : but that the fibres will
FIG. 10.
yield, of course in the opposite sense, when the load is removed ; but
in actual materials, it is possible that this conclusion may be vitiated
by failure to realize the assumed condition that the stress does not
rise above the yield value.
From this brief sketch of the nature of the effects of plastic
deformation, it is apparent that the values of the maximum stress,
calculated by elastic theory from the value of the bending moment,
considerably exceeding the yield stress are quite in accordance with
the assumed condition that the actual stress can in no circumstances
exceed the yield. In addition, however, there are two reasons why
the true upper limit for the maximum " apparent " stress may be
considerably higher than the analysis above would indicate. First,
the yield in compression may be considerably greater than the yield
in tension ; such a difference would cause the neutral axis (in Pig. 10
for instance) to move towards the compression side as the load was
increased and would necessitate a greater bending moment t o cause
complete yielding of the beam. Second, in all practical materials,
although the yield may be very well marked and result in considerable
extension, nevertheless considerable increase of stress above the
yield is usually necessary before failure occurs. Considering, for pur-
poses of illustration, the case of a bar of circular section, the change
in the curvature ratio k( = r/R)is approximately equal to the strain
of the outside fibres, so that even a very marked yield (say 5 per cent
extension) would not result in any very considerable distortion of
the beam before the stress in the outside fibres began to rise above
the yield stress, while, if the initial curvature of the beam were
considerable, the change of curvature occurring before the stresses
in the outside fibres exceeded the yield might be almost imperceptible.
The wide variations and complex forms of stress-strain relations
above the yield render any more particular discussion of this effect
unprofitable ; but it is clear that this increase of stress must result
in increased moment of resistance and hence lead to still higher
values of the " apparent " stresses.
In the case of several of the multiply-connected components to be
dealt with in Section VI, p. 279, the elastic theory indicates that very
high compressive stresses may occur a t the point of application of
the load. As, however, in addition to the above considerations of
stress redistribution, any yielding due to these high compressive
stresses results in the distribution of the load over a larger area,
these stresses are not of great importance. This conclusion is
confirmed by the fact that when such components fail without marked
distortion of the component as a whole, and thus without consequent
alteration of the bending moments and direct loads applied (e.g. in
the correct expression for the strain energy in a curved beam in place
of the approximate one above, will be shown to be so small as to
be negligible (see Appendix I, p, 321).
2EIU = /R-’Mo2RldB
- 0
+ / M&z + r M + 2 R p d $
.1
.’ 0 . 0
Since by symmetry, no change in slope occurs a t the ends,
aU/aMO = 0 ; also, since no relative displacement between the ends
occurs in the direction of force H, HJjaH = 0
and 2EI
au
-- ;= R l ~ / ~ - i M o (-l cos 0)dO + R, M,&R, + Z cosec +
8H
Rz 1/1 (R,
++ + + +
MO[Rl(.r - +) Rz+ 11 H[Rl*(.rr - - sin +)
+ I cosec +) RIZ(l cos $) (R,, +
+ + + t Z 2 j sin $1
W
= ~-[R12(1
2
+ +
cos $) R,z(l -. cos 4)
__ W[Rl3 (I +
- -- -t- R2“l - C O S + ) ( R+~ ZCOSW$)
COSI/J)~
2 2
R
sin2+ + R121 sin +(I + cos $1
3
+ 2-
2
+ J R ~ P ( +I cos +)(I - 2 cos+) - -6 sin 241 . . . 18
(5b)
These equations are too unwieldy to be solved in their present
form. In any given case, however, substitution of the dimensions
R,, R,,?, and + = tan--* -___
R, - Rz
)
gives two simultaneous equations
with numerical coefficients for Mo and €1.
To demonstrate the general type of stress distribution in a fitting
of this form and t o show the method of calculation for any case of
practical design, we may consider a link which is very similar to
the proposed British standard egg link (non-reevable). This has t,he
form and dimensions shown in Fig. 12, the egg link being suitable for
use with 1-inch chain, i.e. for a maximum safe working load of 6 tons.
Here R, = 2.18 inches, R, = 1-43 inches, I = 2.12 inches,
+ = 1.231 radians, and W = 6 tons.
Equations (5a) and (5b) give :-
8-05Mo 2262 H = 33.90 +
22.62 Mo 95.49 H = 91.20+
FIG.12.
p+ = -2
sin4 - H cosq5 = 3.00 sin4 + 0.13 c o s 4
The bending stress a.t any section will he :-
iM
J% = p rA (see equation (3), p. 260)
Here Y = 0.68 inch. A = 1.453 sq. in.
r
At large end, k = - = 0.312 and, from Pig. 2, p. 261, p o = 3-31,
R
Po-
rA - 3.25, p1 = - 5.22, and = - 5.29
rA
A t small end, k =; 0.476 and = 2.87, Po 2.90, p1 = - 6.26,
rA
and -PI = - 6.34
rA
1 Part. I Extrados.
I Intrados.
I
13.08 - 19.20 sin 0 - 22.76 + 36.67 sin 0
+ 0.79 COY 6 - 1.52 c o e~
Straight side . 3 . 5 7 ~- 5.66 9.62 - 3,572
11.51 - 10.25 sin 4 +
- 25.40 29.27 sin 4
-0.43 COY 4 +
1-23 cos 4
These stresses have been evaluated and are plotted in Fig. 13,
the base of which represents the straightened centre line of a half-
link. At the junctions of the ends with the straight side, the stress
will be seen to have two values ; one of these represents the stress
just to one side of the junction, e.g. the straight side, where the
radius of curvature is infinite, the other that on the curved end, where
the stress due to a given bending moment will be less a t the extrados
and greater a t the intrados than in the straight beam. In practice,
of course, the transition will be gradual and not discontinuous as
shown.
The maximum tensile stress at the safe working load a t any point
in the link is 14.8 tons per sq. in. a t the extrados a t the large crown.
The tensile stress at the intmdos of the large end a t 8 = 92.4 deg.
is only very slightly lower, namely, 13.9 tons per sq. in. The only
other point of high tensile stress is the extrados a t the small crown,
a value of 11.1 tons per sq. in.
In the above example only the egg link for 1-inch chain has been
considered. But in any range (such as the British standard links)
which forms a geometrical series, having dimensions proportional
W
/
n I
%lo.- I
I
ALARCE
END- ----++-STRAIGHT SIDE-
I
I
I
25-
to the square root of the load, the stresses calculated apply throughout
to the whole range.
The effect of varying the length of an egg link on the stresses
induced may be of interest and is shown in Pig. 14, in which the two
principal values of the stress are plotted against cos$. The other
dimensions of the link (Rl, R,, and r ) are as in Pig. 12. This diagram
shows the effect of length between the two extremes cos $ = 1 (ring)
and COS$ = 0 (infinite length). The shapes of link corresponding
to four given values of cos+ are also indicated. In the practical
20-
I
0
zd15- 1NTRADOS. MAXIMUM STRESS--
v)
K
W EXTRAOOS. CROWN
a
v)
z
P
I
v)
2 10
[r
t-
v)
0 ______c__
0-25
will affect the relative positions, though not the general shapes, of
the two curves.
Mx -- WR2 .I_-
2-77
2 nR+Z
Po = Wl.2 sin 0
P, = w/2
and the stresses are
P, 4Mx W
f x = = ~ + 2A
4R 2-7s
~ = - ( lTR
&f ~
l - ->
where k = rlR as before, and p = curvature factor = p,, a t extrados
or p1 at intrados.
At the extrados, the maximum tensile stress will occur when
0 = 0 deg., i.e. a t the crown and will be
WPO 2 + m
f o == (-
2A Jc T-+ m )
m=-
E
where
R
At the intrados, the maximum tensile stress will be at 0 = 90 deg.
and will be
2A k n+m
As W/2A is equal to the apparent tensile stresa in each side of
the link, it will be convenient to call fIWI2A the '' stress factor ' I
for the liuk ; this will be denoted by KO and K1 for the extrados
and intrados respectively.
h ’1 -- 1 - - P I T - 2 = 1 -+ SIT1 . . (7h)
k . ?r 111 + *
50
-
I\
I
J.H H
&lo= M, - WR
~ sin e + H R ( ~
- cos e;
2
97 1
3
2EIU = j0Me2RdB + jfMs2dx
By symmetry, no change in slope occurs between the ends
of the quadrant, hence aUjaMo = 0. Since, as stated previously,
compressive strains may be neglected in comparison with
displacements due t o bending, no relative displacement of the ends
in the direction of H occurs, hence aU/aH = 0, whence
Mo = F . WRj2 ; H = GWj2
where
(m +
F - ___--____
2){d +
6m2 + 12(4 - n ) m 48(n --3))
.-
+
rn4 + 4nm3 + 48m2 24nm 24(n2 - 8)+ .
+
and +
G =.-- 1 2 ( + ~ a)((, - 2)qn + 2(4 - T ) )
m4 +4nm3 48m2 + +
24nm 2 4 ( d - 8) +
i!
(ni = --, as before)
R
Then
WR
Me= --(I!
2
+ G -sine - Gcos8)
M,=- WR - - ( F + G - ~ + G - -2)
2 R
W
+
Y,y = ---(sin 0 G cos 0)
2
P, = w j 2
and the stress factors are :-
K e = @k ( F + G ) + ( l - P o ) Gk = @ F +kG . . (Sa)
and for all practical cases this will be the greatest stress a t any point
of the extrados.
The stress fact,ors in the various parts of the link, open and
studded, are as given in the table below.
_.____
* Loo. cit.
END
J
-- -< STRAIGHT SIDE
These stress factors are plotted in Fig. 20, the base again being
the straightened centre line of a quadrant of the link.
The maximum tensile stress factors are, for the open link, 5.47
at the crown extrados, and for the studded link, 4.17 a t 8 = 70.9 deg.
great practical interest that the case of three loads may be usefully
considered.
Referring to the half-ring shown in Fig. 21, half the main load
W/2 will be balanced by one of the other two loads Wj2 sec a inclined
a t an angle c( as shown. I n addition, forces H and (H - W/2 tan a ) ,
perpendicular to the line of load, will be induced. Hence, Me = hir,
+ HR(1 - cos 0) - WR/2 sin (0 - a ) sec a (the last term applying
only when 0 > a).
Again aU/aMo = 0 and aU/aH = 0
TT-U
Then H =W tana
27-5
WR
M, = --{I
277
+ see a - ( x - a) tan a)
FIG.21.
... Me =WR
-(1 + sec a - n sin 0 + a t a n g cos 6), when 6 > a
2?7
WR
or ---(1 + seca - - a) tanacos0}, when 6 < x
(TT
277
and Po = H cos 0 + W/2 sin (0 - a ) sec a
w
= -(T sin 0 - a tan a cos O), when 0 > M
2n
W
or - (T - a ) tan u cos.0, when 0 < v.
277
Hence, the stress factors are
K = l/.rr{p/k(l+seca)+(l-p/k)(x sin 0-a tancr cos 6)),when B>a
or ~ / - / i { p / k ( l + s c c a ) + ( 1 - ~ p l k ) ( n -t~a)n a cos B},whenB<a
The variation of stress round the ring (k = 0.25) for the cases
u = 0 (two loads only) and u = 45 deg. is shown in Fig. 22.
FIG.22.--Streaaes in Ring under different conditions of loading.
(4
p1 1. +rrsec a + ( 1 -;:)d’l+--
--
tan2 a a2
. .
712
at e = cot-l(-a tan tc ) 77
Forf= 24 tons per sq. in, (a value based on rings which have
given satisfactory performance in service),
\ y = -2=4 -7 T ~ 59-2a2
Po Po
x; k
Over the range of values of k(k = Q to k = i)generally used for
rings, the relation between po and k can be represented almost exactly
by the formula po = 4 - 2 4 3 .
14.8d2
Hence w=-=59.2d2
4
- - 2.6 R -00.65
k r
But Di = 2(R - T )
..
For simplicity, the factor 0.35 is taken as 0.3, the error thus
introduced never exceeding 3 per cent in the range considered.
It should be noted that, in the case of thick rings ( k > 0.27) this
formula, which is based on the stress a t the extrados in the line of
* Assuming elastic conditions.
wsin a-H
\1
V
4
wcosu-v
experiment and this course has been adopted. Some of this work
may be worthy of record, not only because some of the analytical
work is original, but because the experiments subsequently carried
out verified, in a striking manner, the stress distribution indicated
by the theory of elasticity applied to a component of complex shape.
The eyebolt consists essentially of three parts : eye, collar, and
shank, of which probably the greatest interest attaches to the eye.
In what follows, it will be assumed * that the line of the applied
external load lies in the plane containing the axis of the eye.
(a) Eye.-The eye of the bolt springs from the collar, which is
normally so rigid that any deflexion in it due to bending may be
neglected. The eye will, therefore, be considered as fixed in position
and slope at two points, radial lines from which subtend an angle
2/3 and are equally inclined to the axis, as shown in Fig. 34. The
forces on the ring are the load W, the horizontal and vertical reactions,
H and V, at the left-hand abutment, and (W s i n a - H) and
(W cos M - V) a t the right-hand abutment as shown. The weight
lifted is W cos M, and the bending moment a t the left-hand abutment
is M,.
+
Then Me = M,, - HR( cos /3 cos 0) VR(sin p - sin 0) +
WR sin (e a) + +
(the last term applying only when 0 > T - ct).
Since no relative displacement or change of slope occur at the
au au au
abutments, - = - = - = 0 (U being the strain energy taken
a% a~ av
from abutment to abutment). These give three equations for
the determination of the three unknowns M,, H, and V, from which
the following values are obtained :-
v = W_y .cos a- - sin y cosp
2 /3 - s i n p c o s p
=
2 1 -_
sin /?(~ cos_y ) ~- p(sin p sin y y sin a)
_ _ _ _ _ _ _ _ ~
+ .
+
~
(sin -
- __ p p cos__
/3){2
...__
sin y + y sin a)}
sin /3( 1 - cos y ) - /?(sinp___-__ .
Z sin2 p - /3(p+ sin /3 cos p)
y cos u - cos /? sin y
where y=.+P.
- /? sin /3--
/3 - sin p cos /? -1
-_ . _ _ _ _ _ ._.______
_ ~ ~ ~ ~ _ . _ ~ . ~
FIa: BB.-British
Standard Eyebolt.
+
Dimension X varies from &.inch for #-inch t o inch for 3-inch eyebolt.
Dimension Y varies from A inch for 3-inch t o -&inch for 3-inch eyebolt.
\J'
2D
' K
i
which the approximate value of /3 is 135 deg. For this case the
expressions given above simplify t o
+
H/W = (0.500 0.349a) sin a 0.026 cos cc - 0-210 +
+
V/W = (0.500 0.175~)cos a - 0.088 sin a
+
M/,WR = 0.423 - (0.232 0.124~)cos a - (0.292 0.352~)sin a +
Also, Me = M, + +
0.707R(H V) - R(H cos 6' V sin 6') ___
+
-~- WR sin 8 Q + +
+
and, Po = (H cos 6' V sin 6') - W sin 6' a +
__-
(In these expressions for Me, Pg, and KO,the term sin B + a applies
only when 8 > T - a.)
In the original design, k = 4,so that Pok = 9.48, @k = - 15.99.
Possible positions of maximum tensile stress, with the corresponding
stzess factors, are as follows :-
(1) Extrados at point of application of load
Mo H
where K = 18.96---
WR
+ (13.40 + 16.96 cos a)W-
+ (13.40 - 16-96sin a)-.WV . . (12a)
(2) Extrados at support remote from load
where
MO H v
K = 18.96-
WR
+ 14----+
w 25.38- w + 11.98 (sin ct - cos a). (12b)
( 3 ) Iritrados at tan 6' = V/H
K -31.98-
M, - 22.60( --
,H V
+ -w-- )
where =
WR .W
0 0
ANGLE OF LOAD c(I
for the British standard 1-inch eyebolt are 2.75 tons at a = 0 (direct
lift) and 0.675 ton at u = 45 deg. The latter condition gives a value
of W = 0.6752/2 = 0.955 ton. The sectional area of the ring
(3 inch diameter) being 0.442 sq. in. and the respective stress factors
4-82 and 7.85, the maximum stresses in the ring have the values
(a)15.0 tons per sq. in. for direct lift, and (b) 8.5 tons per sq. in. for
45 deg. lift.
In presenting these values, however, attention must be drawn
_ _ . _ -~ - _. - ____
* Actual vertical load lifted, not value of W,
Nominal Size
of Bolt D,
inches. Diameter of Material in Eye.
4D ;kD __.___
QD PD
_____ -
~. - _______
* A striking experimental verification of the equality of these three
stresses is described in Section VII, pp. 319-20.
*These were made with eyebolts screwed into mild steel and cast iron
blocks.
Referring to Fig. 27, where P is the load lifted and W the lifting
force (inclined a t an angle a to the axis of the bolt), the shank is
subjected to tensile, shearing, and bending actions due to the
components W cos u, W sin u, and the moment (2DW sin a ) ,
W
=.p
Owing to the fact that in the Whitworth thread form dlD is not
constant over the required ranges of sizes (D = 8 inch to 3 inches),
the factor g is not a constant, but it will be found that if the average
value of d/D, equal to 0.84, is employed throughout, then the values
of g thus obtained, over the whole ranges of D = 8 inch to 3 inches
and a = 0 to 45 deg., are never in error by more than 1 per cent.
Hence the variable g in equation (13) can be replaced by a constant
(2.62) and the relative values of P, for various angles of lift, are given
directly by the equation. (The relation between P and a is, of
course, independent of the actual size of the bolt.)
The relative values of P, the load lifted vertically by the bolt,
given in the following table shows the inanner in which this quantity
varies with the angle of lift :--
RESULTS.
VII. EXPERIMENTAL
The experimental work involved in the preparation of British
Standard Specifications for lifting gear components on behalf of
the British Standards Institution falls into three main groups :
(a) preliminary tests on existing commercial types, t o discover their
strength and other properties, (b) tests made in the development
stage of a proposed standard, to investigate the effect of some
variable, and (c) tests made on the proposed standard, to ascertain
its resistance to various types of loading and thus to furnish practical
data regarding the efficiency of the design and material used. The
results of the first two groups of tests were very useful for their
immediate purpose but are not of general interest. Of the third
group of tests, a brief account will be given of the tests made a t the
National Physical Laboratory on (1) hooks, and (2) eyebolts, a.s the
results not only give data which may be of pmctical interest to
manufacturers and users, but also are of some theoretical importance
in connexion with the theory of elasticity on which the designs were
based. Also, they are believed to represent the first attempts a t
complete series of tests made on full-size lifting gear components of
these kinds.
Rated 1 Estimated
Propor- 1I
Per-
Safe Stress at 1 Maximum manent
Refer- Working safe tional Ratio, Fibre
ence
Mark
Type of
Hook
Material Load
(S.W.L.),
working
load,
Limit
(P.L.), 1
1 P.L.
s,w.L. 1 stress
at P.L.,
Set at
proof
load,
tons jons per tons tons per in. x 10-s
sq. in. 1 sq. in.
'
___ ___ ~ -.
-1
____
FKN. 1
FEN. 2
FKN. 3
Crane Mild steel
Wrought, iron
Mild steel
4
4
4
12
12
10
I
1
5.7,
4.6
6.8
'I
j
1
1.44
1.15
1.70
16.1
13.7
18.7
2
7
0.5
20.7
27.3
I
5.2
6.8
I
FKN. 4 Wrought iron 4 10
___-
j 5.6 j
-I
1.40 13.7 2 22.5 5.6
___.
GWG. 3 Livcrpnol Nild steel 1.5 9 ' 2.8 1.93 17.2 0.5 134 8.9
I
~
~
* Some small allowance t o be made for departure of actual shape of hook from standard form (see last column of Table 1).
Esti- Esti-
I mated t mated $
Reference
Mark
Type i &raterial
Safe Rated Actual
Stress Stress
Working at safe at safe
?$' working working
load, load,
tons per tons per
sq. in. sq. in.
FKN.
FKN.
2
3
,, Wrought iron
Mild steel
4
4
12
10
11.9
11.0 '
i
above the safe working load, but only in one case did this load exceed
the proof load ; the permanent set a t the proof load, although of
definite amount (with one exception) was small in magnitude. The
loads a t which the hook opened out to such an extent as to be
incapable of retaining the load were equivalent to from 5 to 7 times
the safe working load for the crane hooks and 74 t o 9 times for the
Liverpool hooks, being in each case rather higher for the hooks
made of mild steel.
In one case with each type of hook, the load was reduced before
final slipping occurred in order to ascertain what load the hook
would support indefinitely when considerably deformed. In the case
of FKN. 4 the rate of deformation under a load of 20 tons (= 5 x safe
working load) was practically zero, and in the case of GWG. 3 the
load supported, after the point of the hook had become horizontal,
was 10 tons (= 6.7 x safe working load). It is, therefore, concluded
that the crane hooks would support indefinitely a load equal to
4 times, and the Liverpool hooks a load equal to 6 times, the safe
working load. The deepening of the section of the crane hooks a t
45 deg. below the principal section had therefore very satisfactorily
achieved the desired object of increasing safety in cases of heavy
overloading.
In two cases only did fracture occur ; hooks GWE. 3 and 4
(Liverpool, wrought iron) being about one-third and three-quarters
severed, respectively, a t the point of maximum stress just before
the load was released. The wrought iron crane hooks showed
longitudinal cracks along the pilings, but did not fracture. The
mild steel hooks showed no signs of fracture.
From these results, it was decided to adopt the same stress rating
(at safe working load) of 12 tons per sq. in. for hooks of wrought iron
and mild steel.
(ii) Tensile Impact Tests to Destructiolz.-These were made on
samples of standard sling hooks (Fig. 8b) and Liverpool hooks
(Fig. 8c) of the sizes specified for a safe working load of 1& tons.
Particulars of the hooks used and the results obtained are given in
Table 3.
As a convenient means of comparison the energies absorbed are
also stated in terms of the weights of the hooks. This figure should
be constant for geometrically similar hooks of the same material ;
it is therefore of value in affording an index of the combined efficiency
of the material and design of a hook or other fitting. All the hooks
were tested in the normalized condition.
These results compare very favourably with those of hooks of
other designs previously tested a t the Laboratory, a value of 0-6
II I Safe R.ated
Work- Stress
Energy
Absorbed, ,
Mark 1
teference Type of
Hook 1 hIaterial ing at
Load, 3.W.L.
tons tons
Weight -
lb.
ft.-
l l Pe: sq.
in.
_ _
tons
_
lb.
~ -
GFV. 1
GFV. 8 I
GWG. 1 Liverpool
Sling
., 1 Mild steel
.,
,,
1.5
1.5
1.5
12
12
9
~-
3.80
3.73
6.61
2.66
3.03
5.58
0.70
0.81
0.84
GWG.2 ?, 1.5 9 6.57 5.46 0.83
GWE. 11 ,, Wrought iroI 1.5 9 6.50 4.43 0.68
GWE. 1 1-5 0 6.83 4.42 0-65
GWE.2 ,, 1.5 9 6.48 4.85 0.75
ft.-ton per lb. being usually regarded as satisfactory. All the hooks
opened out very considerably, the four mild steel hooks showing no
cracks or other sign of actual fracture. Of the three wrought iron
hooks, two, after undergoing considerable deformation, fractured
a t the section of maximum stress (when opened out) ; the other
exhibited a large crack a t the same position.
(iii) Patigue Tests.-Fatigue tests were carried out on sling hooks
(Fig. 8b) and Liverpool hooks (Fig. 8c), all of 1.5 tons safe working
load, using a Haigh electromagnetic fatigue testing machine of 6 tons
capacity. All the hooks were normalized before test. In order t o
avoid impact, the minimum load of each cycle was maintained at a
positive value, namely, ton ; the frequency of loading was 2,200
cycles per minute. The results of the tests are plotted in Fig. 28
(the number of reversals to fracture being plotted to a logarithmic
scaIe) and are summarized in Table 4.
Estimated
Stresses at Safe
Range, tons
per sq. in.
15.1
Liverpool , 19.6
18.4
The fact,or o f safety of 1.2 against fatigue failure for the sling
hooks may seem unduly low, but it may be remarked that these safe
ranges represent t'hose under which fracture should not occur in an
indefinitely great number of applications of load, while the service
FIG.28.-ResuEts of Fatigue Test8 on 1.5-ton hooks.
every precaution to avoid the hooks being subjected to ill usage. The
reason for the higher " factor of safety " in the Liverpool hooks is
the heavier section employed, which was considered desirable in view
of the severe character of the service to which this type of hook is
often subjected.
The hooks failed by the development of cracks a t the points of
maximum stress, no appreciable distortion of the hook taking place.
The position of the crack in t,he sling hooks varied between the
principal section and a position about 35 deg. above that section.
The stresses a t sections above the principal section, in the hook as
designed, are somewhat higher than that a t the principal sect.ion ; the
variation in the position of the crack is probably due to slight
variations in shape of the samples, also, possibly, to variations in
the condition of the material due to varying amounts of reduction
during forging a t various points. The Liverpool hooks all failed
in the neighbourhood of the horizontal section where the variation
of stress along the intrados is very gradual.
In order to ascertain whether repeated application of a safe range
of stress produced any deterioration in the impact resistance of the
hooks, impact tests were subsequently carried out on some of the
hooks which remained unbroken by fatigue after ten million cycles of
loading. The results are given in Tablc 5.
1 Range of Energy
Refer- Load in Absorbed,
ence Type Of Ma,terial fatigue
Mark Hook test, lb.
tons
1~
GFV. 6 Crane
GWG. 10 Liverpool
Mild steel
GWE. ti Liverpool wrou&t iron
~-
1.6
3.25
3
1
I-
3.73
6.56
2.65
4.93
4.14
0.71
0.75
0.62
of $ inch nominal diameter were used for the vertical loading tests
and li-inch bolts for inclined loading.
All eyebolts were first subjected to the proof loads, applied
axially, of 3 tons and 9 tons respectively, for &inch and la-inch bolts.
per sq. in. for the normalized material ; it is evident, however, that
the safe range is well above the safe working load.
Considerable interest attaches to the positiolz of the fractures.
In the case of the 14-inch eyebolts tested under loading inclined a t
45 deg., the estimated stresses in the shank are only slightly below
those in the eye ; as there is in the shank an initial stress due to
tightening up, and a stress concentration due to the presence of the
screw thread, the strengths of the shank and eye should, under
fatigue conditions, not be widely different. This expectation was
fulfilled, one of the eyebolts breaking in the thread and three in the
eye. The latter all broke near the support nearer the load, the
fracture starting from the intrados; this is in agreement with
the stress distribution according to the theory, given previously,
of the stresses in eyebolt rings.
I n the case of the %-incheyebolts under direct loading, the stresses
in the eye are much higher than those in the shank. Also, as Yig. 26
shows, under direct loading, the stresses a t the extrados a t the load
and a t the intrados at, approximately, 0 = 110 deg. and 250 deg.
differ only very slightly, the two latter being slightly higher. In
three cases fracture occurred at approximately 0 = 110 deg. (or
250 deg.), but in one case three cracks were developed simultaneously.
This specimen is shown in Fig. 31 ; it will be seen that one of these
cracks occurred at the extrados a t the load and the others a t t h e
intrados a t about 8 = 110 deg. and 270 deg. This agreement with
the prediction of the theory is very striking.
As a result of these fatigue tests, it was decided to increase the
diameter of material in the eye from $ to D, thereby decreasing
the stress in the eye by 30 per cent at u = 0 and 32 per cent a t
u = 45 deg.
APPENDIX I.
MORE ACCURATE
THEORYOF BENDINGMOMENTS IN MULTIPLY-
CONNECTEDCOMPONENTS.
The method adopted in this paper for the calculation of bending
moments in multiply-connected components makes the assumption
that the strain energy in a curved beam can be expressed by the
formula
.I,.p(R + y)dA . 2E - 6U
60
-
= R(A' - A) = ARA
" I
For a straight beam, the total strain energy is
jR
2AN
2EI
The application t o the chain link will now be considered. Using
the same notation as previously,
WR W 13
7
Me = Mo - sin 8 ; Po = - sin 8 ; Qe = -- c,os 6.
2 2 2
aU -
_ - 0, U being taken round a quadrant of the link.
aM0
7r 1
._
:. ‘/’(&
0
- sin 8 cos2 w)R cos2 wd8 +]oz(ivIo- F ) d x = 0
M - WR
_ _ ~ - . 2 ~ 0 ~ 4 0m -+
whence
O- 2 TGOS~W m -+ ~
v
Morley * gives a refinement of the approximate method which
leads to the following results :-
WR 2R+Z - W R 2+m-
---
Mo =
TA‘R + 1
2 vsec2w vn +
A
WR
For rings W E = 0, M, =- c0s2 w
7r
I n the case of the chain link, the more accurate theory given by
Morley is carried only to a first approximation. The values of
Mo/WR given by the three methods for extreme and average cases
of links and rings are shown in the table beIow.
1 IFs$d1 I
Usual
Method
Morley’s
New
Method
Extreme case 1
I
I
0.318
(1.00)
1
I
0.297
(0.93)
1
I
0.297
(0-93)
1
I
,, ,, 1 0.362
(1.00)
1 0.343
(0.96)
1 0.348
(0.96)
1
It will be seen that the effect of the more exact method is to reduce
the value of Mo by about 4 per cent in the average cases-a correction
which is not worth the extra labour involved.
-- -
* ‘‘ Strength of Materials,” 6th ed., 1928, p, 402.
AYPENDIX 11.
LIST OF SYMBOLS.
A Cross-sectional area of material.
A' Modified area of cross-section.
C Bed diameter of hook.
D Diameter.
E Modulus of elasticity.
'
G j
)Constants in stress equations for studded chain links.
Load a t crown of link.
orizontal reaction a t collar of eyebolt.
I Moment of inertia.
I( Stress factor.
M Bending moment (positive when producing tension at
extrados).
N Modulus of rigidity.
P Direct load (positive when tensile).
Q Shearing force.
R Radius of curvature of centre line of material.
U Strain energy.
V Vertical reaction a t collar of eyebolt.
W Load.
* LOC.c i t .
$1
4
JZ
/,
Co-ordinates of section.
Angle of sides in egg link.
w 4 sin-%
Discussion.
Mr. J. P. BOWEN(Trinity House, London) wrote that he was
most interested in that portion of the paper dealing with the stresses
and strains in open-link and studded-link chain. As engineer-in-
chief to the Corporation of Trinity House, he was responsible for the
supply and testing of lightvessel mooring chains, which could not
be treated in the same category as ships’ cables or similar chains,
since they had special duties to perform and peculiar conditions to
withstand. Lightvessels were moored mainly in the open sea to
mark shoals and sandbanks and they must keep their station during
all weather conditions and the severest of gales, without dragging
or parting their moorings.
The moorings were subject to corrosion and abrasion and had
in the past been made of wrought iron of best quality (22-24-ton
cable iron grade); the chains were of the open link type with
dimensions corresponding t o what was generally known as long ((
link,” since in studded chain the studs had a tendency to rack and
loosen. The links were side welded, and the chains were heavy and
well in excess of Lloyd’s requirements for ships of similar tonnage.
Nevertheless, they were known t o part, not only in the weld but in
the iron, showing that very heavy stresses were at times imposed.
For some time he had been investigating the stresses set up in
these moorings, particularly with a view to ascertaining the cause
of corrosion since he had formed the opinion that constant variation
of stress was a contributory factor, an opinion which had so far not
been proved to be well founded. In carrying out the investigations,
various sizes of chain and independent links were subjected to tensile
stress, to determine the yield point of the iron when manufactured
into chain. It varied somewhat with the size of chain, but was
found t o be about 6-7 tons per sq. in. for chains of material
ranging between 1Q and 2 inches in diameter.
During these tests complete 15-fathom lengths were fractionally
subjected to tension by adding loads in regular increments. After
initial bedding of the links it was found that permanent stretch
occurred after a load equivalent t o about 6 tons per sq. in. of section
had been imposed ; above this loading the set was permanent. The
elongation of a 1%-inch chain under a load of 40 tons, which
approx imated t o Lloyd’s open link test, was 5 per cent ; with studs
inserte d, the extension was nearly 7 per cent under a load of 55 tons.
Of thi s amount, about 2 and 3 per cent respectively represented iron
FIG. 32.
(L b
materially altered. These trial lengths had been put back into
service again for extended test.
Elsewhere in the mooring, however, links which had been tested
to the full load, having legs which were necessarily straightened
under test, had been bent and twisted. One of the legs was bent
into convex form, instead of concave, as would normally be the case
under direct tensile stress. It might be that this distortion occurred
a t the hawse pipe or by foul links, though this was difficult t o say
as the chain was paid out and veered according to weather conditions,
the position of the mooring not being constant.
He would be glad, however, if the authors would show by a
diagram the stresses both in an open and a stud link in which the
pull was direct a t one end and a t 30 deg. to the longitudinal axis a t
the other, assuming the link to be supported a t the centre on one
side (see Fig. 32 u ) , as these, he thought, were the conditions under
which the links might have been deformed in the hawse pipe.
Another set of stresses which would be of considerable interest
to him were those set up in a foul link, where the link was jammed
transversely in the adjoining links on either side so as to set up a
semi-diagonal stress instead of a direct pull (Fig. 32 b). The size
of the link which he had under consideration had a length of 6
diameters and a width of 3.6 diameters. These investigations were
being continued and it was hoped some day to publish the results,
as well as some comparisons with welded forged steel stud-link chain
with which Trinity House was now experimenting on an extensive
scale .
that for loads of 40 tons and upwards hooks of other types were
recommended. In regard to material, class A steel in British
Standard Specification No. 24, Part 4,was specified. From inquiries
recently made, it appeared that none of the best-known hook makers
in the country were using class A material at the time of the inquiry,
and in fact only a very few of them were aware that a hook
specification had been issued about two years ago.
3 107 80
(alloy steel)
',
____
4 75 60
(carbon steel)
IW
d
dC ---+-4----t-
9 h s2-E2.9
SY 2
The difference between the endurance ranges did not exceed 5 per
cent, yet theory indicated a difference of not less than 25 per cent
in the ranges of stress. A possible explanation was that plastic
flow modified the mean stresses on the inside of the coils in such a
way that the safe range of stress in the inner fibres of the springs
with a ratio of coil diameter to wire diameter of 5.14 was greater
than in the springs with a ratio of coil diameter to wire diameter
of 17.19. It would appear from the results that unnecessary emphasis
had been laid on the danger of neglecting curvature factors when
calculating the working stresses in helical springs.
The following analysis of the distribution of torsional stresses
was developed in connexion with the foregoing observations.
Consider a part of a round bar having given the radius of curvature
and loaded by a force along the axis of curvature. Fig. 34 showed
part of such a bar. The loading was the same for each cross-section,
80 that the faces AB and CD were similarly stressed. P(x, y, I) and
+
Q(x, y, $ 8$) were points similarly situated in the cross-sections.
The small element P and Q, of cross-section 6z x 6y and length
(R - x)8$, had six faces. For equilibrium, it was necessary to
consider (1) the forces due to the normal stresses on the six faces,
and (2) the forces due to two components of shear stress on each
of the six faces.
The normal stresses (Fig. 35) had no moment about the axis of
curvature, because :-
(1) Normal forces on the faces abfe and dcg?b were parallel to
the axis of curvature.
(2) Normal forces on the faces aehd and bfgc acted through the
axis of curvature.
(3) Cross-sections AB and CD were similarly stressed so that
moments about the axis of curvature of any normal
forces on the face abcd were balanced by moments of
similar forces on face efgh.
Fig. 36 indicated the shear stress components which had no
moment about the axis of curvature. These were accounted for as
follows :--
(1) Shear stresses on the faces abcd and efgh were in planes
containing the axis of curvature.
(2) Radial components of shear stresses on the faces a@ and
dcgh acted through the axis of curvature.
(3) Axial components of shear stresses on the faces bfgc and
aehd were parallel to the axis of curvature.
The remaining stresses were the circumferential components of
shear stresses on the faces aehd, abfe, bfgc, and dcgh. Taking the
( S1 f as1
62
6s
- . -)(R
2
- x ---@YSyv
2
+ (S, +6S2
6y 2
6y
- . -)(R - x)28~8$
= (S1 --
62
.-)p
as, sx
2
-x
6x 2
+y) 6y6$ + (& - s . & ) ( R
6Y 2
- ~)2S&,4
s, = sel . . . . . .
2 2
(1 - E)
For a straight bar of circular section the shear stress components
due to torsion were
S1' = - Ky . . . . . . . (4Q)
So'= Kz . . . . . . (4b)
where K mas a constant determined by the torque and the diameter
of the bar.
For the same bar curved to radius R and stressed by a load along
the axis of curvature,
If a waR the radius of the section, the torque about the bar axis
was
a 2s
0 0
KR2r3 cos2 8
rd8 + (1 (RK R V
0 0
sin28
- r cos 8)zd
rd8
+
KR2r3(R r cos $1 al.d9
(R2 - $)%
00
a
=hKR31
r3
0 (R2 - rZ)%
dr . . . . . (8)
.. 2rr
n
The upthrust was KRz/,’----- 72 cos 0
drd0
b ,j (R - 7 C O 0~) 2
R WR 1
Let a = . Then Q =-
a 2a2 - I
WR
The expression -7T
was the stress due to torque when curvature
-a3
2
was neglected.
1
The factor C= . . . (14)
2 6 -1
C was the " curvature factor " for the stress in the inner fibres.
For purposes of calculation it was more convenient to express C as a
series. After expanding by the binomial theorem,
2 2 2 33 17 35
C=1+-+-+++-+~5+-+. (15)
u2 u3 1 b 4 16d
mean coil diameter
where u = - ~ ~- for a spring wound from round bar ;
rod diameter
from equation (13) the maximum stress in the inner fibres
WR
&=-C
rr
-US
2
It should be understood that though the values of S, and Sz
v
given in equations (3a) and (36) were shown t o satisfy the differential
equation (l),these were probably not the only solutions. Another
differential equation satisfying the strain conditions was required,
but these conditions would appear t o be indeterminate. One
justifiable assumption was that if initial strain was to be independent
of the number of coils in the spring, a fibre forming a helix before
loading would form a helix after loading. The assumption that
cross-sections which were plane before loading remained plane after
loading, though justified in the case of a straight bar of cylindrical
section, was not justified when the same bar was wound to form a
spring.
the latter process the meaning of all the quantities would be clear
to any student ; moreover, this would be using much more straight-
forward principles than the energy principle. Further, it would
show that the essential assumption in this problem was that plane
cross-sections remained plane.
The authors rightly pointed out that Morley's suggestion for
finding A' by a graphical method would in practice be far from
correct. But they did not themselves push their correction of
Morley's method far enough. If A' were calculated graphically it
would, of course, be known to the same accuracy as A itself. Now
A' - A was always small in practice. A typical result in such
examples would be A' = 1-04A, to three-figure accuracy. Thus
A' - A = 0.04A, and so A, the quantity really needed in addition
to A, was obtainable only to one-figure accuracy. But equation (1)
showed that A should be calculated at least as accurately as A
because, as a rule, the 1/X term in the expression forf was the largest
term having the factor M,. Thus the graphical method should
be applied to find A and (A' - A) directly. The latter was just as
easy to find as A' itself, for
his own book * a very similar problem was worked out; but there the
depth was small compared with the thickness, yet the analysis was
the same as for Filon’s case of a thick beam. In this book a case
was worked out in which the radius of curvature was twice the
thickness, corresponding to the value 6 for the ratio r @ in the paper.
He had there calculated the value of (R y’)/(R - yl), where y‘ +
referred to the neutral axis, and had found that this ratio was 0-693,
to three-figure accuracy. He had never before noticed that this
number agreed with log,2 to these three figures. Now log,2 mas
exactly what the beam theory used in the paper gave for this identical
ratio, and consequently it must be taken as strong justification of
the usual beam theory ; that mas to say, justification of the process
of calculating stresses from the assumption that plane sections
remained plane, although it was well known that they did not
actually remain plane.
Mr. J. LUMSDEN RAEwrote that the formula for the proof load
of a ring, in tons,
where d was the diameter of the material in inches and D, the internal
diameter in inches, was a modification of the formula used up to the
present by reliable makers, and might give more satisfactory results.
The limits to its application did not appear to be clearly defined in
the paper, and this might be rectified.
A well-known chain maker recently carried out, a t his suggestion,
two experiments on rings. In one experiment the ring was of
chain iron 1 inch in diameter, i.e. l& inches actual diameter, with
an average internal diameter of 11%inches ; it was found that a
permanent deformation of =&inch occurred with a load of 12 cwt.,
which would probably represent about 10.8 cwt. as a safe proof
load. By the suggested formula, the proof load would be 1-3 tons ;
by the old formula, about 1.5 tons.
A second experiment was with a ring of chain iron 146 inches in
diameter, having an internal diameter of 6% inches. It deformed
& inch a t a load of 8 tons, the safe proof load being about 7.2 tons ;
by the suggested formula this should be 7.1 tons.
A third ring, made by a well-known firm, had an internal diameter
of 24 inches, the material being 64 inches nominal diameter, and
indicated a proof load of 162 tons. The second and third rings, it
~~
transverse load was applied to the ends and on the other a bending
moment of value PR was applied, the difference between the
respective stresses was not constant a t each radius of any given
cross-section. The change in value of this difference, which
corresponded to what was usually termed the direct stress, was
exemplified by the curves in Big. 39. Curve A showed that for a
beam of very small curvature the variation in the stress was
inappreciable. Curve B, however, showed that when the curvature
.t-
2.91 00 I
:I
I
:I I -
:I
:I I
:I
:I
I
‘ I
I\
I
Ln:1
m . 1
I
: : I I
11.0
10'0 USUAL RA
CRANE ;/~~?
9.0
8'0
e I
o"7*0
I
a
W
I
I I
W
v) I
a
W6.0 --__c_-
0
5 I
I
I
5' 0
1
I
I
4'0
I
3.0
2.0
ii ___t___
I
I
I
I
4.0 4
2.64
=w
to the horizontal, a shear stress also ; the latter became considerable
in beams with B large curvature.
per sq. in. a t the crowns were not in fact realized. From the
foregoing consideration it would appear likely that the actual
maximum stresses in chain links might approximate to the actual
maximum stresses in rings designed for a maximum stress of 15 tons
per sq. in.
The curves in Pig. 16, p. 287, were very useful in determining the
maximum stresses in links, but from the practical aspect it would
appear that if they were used for calculations in the design of links
of exceptional length and width, the presence of the long weld and
the big curve a t the crown ought to be taken into account, and a
lower " apparent )' stress than that for short links should be used
for the calculation.
The fatigue test results of the hooks showed that good margins
of safety were provided by the designs, and that a hook so designed
should never fail by fatigue caused by the safe working load. The
impact tests showed the ample margin against shock loading possessed
by hooks designed according to the formula, and that any impact
to which a hook was likely to be subjected in service would not be
sufficient to cause fracture.
Before the publication of the paper the theory of the stresses
in eyebolts had not been satisfactorily set out ; the paper clarified the
position, and the tests showed that the margins provided by the
designs given were ample. The scientific and practical aspects of
loose lifting gear were now readily available to manufacturers and
should give them confidence in matters of design.
-8
-4"
LOAD W 1
the side load (open link) or at the extrados a t the end of the stud
opposite the side load (studded link) with factors of 8.38 (open) and
6.44 (studded). These factors were to be compared with the values
6-14 (open) and 4.20 (studded) for a link loaded axially (Figs. 41 G
and 42 c).
The type of jamming shown in Fig. 32 b could only occur if the
friction between the links were sufficient to cause the resultant
thrust of the foul link on either of the jammed links to pass through
the point of contact of the two jammed links ; for otherwise there
would be a resultant couple on the foul link which would cause it to
rotate and free the jammed links. Thus jamming could be rendered
impossible if the inside width of the link were made less than
2d/(l+ p2),where d was the diameter of the section of the links
and p was the coefficient of friction. It was important to notice
that the pressure between the jammed links was not the primary
cause of jamming. The essential condition was that the friction
between the links should be sufficient to bring the two resultant
thrusts on the foul link into the same line. Thus if jamming could
occur in position A (Fig. 43) it could also occur in any position
intermediate between A and B, although in the latter case there was
no pressure between the jammed links. The only effect of the
pressure between the jammed links was to render jamming possible
a t larger angles in position A than in position B. The load on the
foul link was, however, greatest in the extreme position A, when
both jammed links were parallel to the tangent a t the point of
contact. Neglecting friction between the jammed links, the ratio
of the reaction R on the foul link to the load W on the chain was
given by the formula R cos 8 = W cos u, where 6 and cc were the
angles between the directions of the loads R and W and the common
tangent a t the point of contact of the two jammed links. For the
type of link in question 6 was tan-1 2 and the least possible value of
0: was tan-1 &, so that the greatest possible value of R was 2.22W.
The foul link was subjected to the highest stresses if the line of the
loads R passed through the centre point of the link (Fig. 43). The
actual stresses produced were shown in Fig. 44. It should be
emphasized that this Figure represented the worst possible case, and
that in practice the stresses might be somewhat lower.
Mr. Broughton asked whether the results of tests on a number of
small hooks were regarded as sufficient to justify the specification
of hooks up to 75 tons capacity. Since all sizes of the standard
hooks were geometrically similar, there was no reason why large
hooks should behave differently from small ones, unless the material
were different. Differences in material were of course possible ;
recognized that the larger sizes of ordinary crane hooks had only
a limited application; it had been proposed, a t a later date, t o
prepare specifications for other types, such as the Ramshorn. The
issue of a specification could not absolve the user from the
responsibility of choosing the type of hook and material most
suitable to his purpose. 00 the other hand, if users specified hooks
to the British Standards Institution Specification, presumably hook
manufacturers would supply them.
23
sound, and the remarks on p. 262 of the paper were not intended to
preclude this method ; but in practice the calculation by slide rule
+
of successive values of b/(R y) was so much simpler than that of
+
successive values of by/(R y) that it was more convenient to
calculate A’ to six figures than to calculate (A’ - A) to three. They
regretted that they had unconsciously allowed this very practical
consideration to influence their statement of the best met,hod of
procedure.
Mr. Lumsden Rae asked what were the limits of application of
the simplified formula for the proof load of rings. The formula was
stated in the paper to be applicable for values of k > and < $
(p. 297). This range might be extended to include slightly lower
values of k ; but for values of k > 0.27, the maximum stress was
set up a t a different position and the analysis on which the formula
was based became inapplicable.
The formula for the proof load of a ring was based upon a specified
maximum stress of 24 tons per sq. in. under the proof load. An
annealed ring subjected t o the proof load would therefore certainly
yield, although the actual plastic deformation might be very small.
The comparison made by Mr. Rae between the theoretical proof
loads and the loads at which measurable permanent set was observed
was therefore not entirely logical and could scarcely be expected to
yield reliable results. In any case the load a t which measurable
permanent set was observed could only be expected to be a constant
proportion of the theoretical proof load if each ring were made of
the same material and were in the same state of initial stress. Such
differences as were quoted by Mr. Rae were probably due to
differences in the yield points of the different materials and to
imperfect annealing after previous proving or use. In general there
was no reason to believe that the elastic properties of normal
engineering materials varied with the size of the piece tested and
therefore there was no logical ground for using different values of
the allowable stress for different sizes of component. These points
had, in fact, received the careful consideration of the responsible
Technical Committee of the British Standards Institution before
the specification was issued in its present form.
That the defects of the Andrews-Pearson formula were discussed
by Mr. Schuster and others long before the publication of Winslow
and Edmonds’s paper was of course well known t o thein ; Winslow
and Edmonds were quoted in this context because that reference
appeared t o be one of the most accessible. They apologized for
making the reference in a way open t o misconstruction.
In calculations of the stresses in hooks and similar curved beams
i
w = 2 w0 cos cos oRae = 2 R n2
n
.2-1 w
W ~cos
R cc
-1
or wo = -
in seca
There would be in the crown of the link a tension due to the transverse
components of the radial pressure equal to
4
H = wocos n8 sin ORd8 = - (tan a
b
If Mo were the bending moment a t the crown
W
2
1
- - sec a)
n
WR
Mx --M'-- [If' = Mo + HR]
2
6U
Equating - to zero
SI&
..
..
p o = -W
- ( t a, n : r - - s e1c a ) ; F ~ =W= -
2 , 12 2
Whence maximum tensile stress factor a t extrados (crown)
. (14b)
k r4-m
I n the case of two similar links in contact, the load distributed
over the crown of one link could be balanced only by a load on the
other link concentrated at the mid-section of the crown. Hence
by symmetry one-half only of the load should be considered as
distributed, and the other half as concentrated a t the crown. The
actual stress factors would therefore be midway between those for
concentrated and distributed loads, namely
1 + seca(1 --&)1 + '117,
-"[
K O-k
77 + 97) n
. . (15a)
and K l = l - - - P1
k
T-1-secu
.rr+m
1--
( i2)
. . (15b)
lines represented the values for fully distribubed loads (equations (14a)
and (14b))and the full lines thosefor the practical case (equations (15a)
and (156)). It would be seen that the value of the maximum tensile
stress a t the crown fell much more rapidly than did that a t the end
of the straight side, the two becoming equal a t about Q = 14 deg.