JOURNAL OF THE

AERONAUTICAL SCIENCES
VOLUME 10 MAY/1943 NUMBER 5

Bending Strength in the Plastic Range

Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

FRANK P. COZZONE*
Lockheed Aircraft Corporation

ABSTRACT x, y = distances measured parallel to the X and Y axes

X, Y = mutually perpendicular axes
The conventional linear distribution of stress in a section
X',Y' = principal axes
subjected to bending in the plastic range, as usually represented z — width of section at distance y from the neutral axis
by the classical my/I formula, has been replaced by a simple
trapezoidal distribution t h a t varies with the applied bending Subscript
moment. This latter distribution approximates the distribution
— due to axial or direct load
based on the actual stress-strain characteristics of the material
= due to conventional bending
and t h e shape of the cross section.
= for a diamond type of cross section
A method has been developed for use throughout the elastic = for an hour-glass type of cross section
and plastic range of bending which covers the following points: = due to limit or applied load
(1) t h e resisting moment of any cross section corresponding to = maximum
any extreme fiber stress up to the ultimate stress of the material; = due to normal or conventional practice
(2) the corresponding stress distribution over the entire cross = minimum
section; (3) the corresponding shear distribution; (4) the effects = for a rectangular type of cross section; or resultant
of local crippling; (5) complex bending; and (6) the combined = tensile or total
effects of axial load and bending.
== ultimate
about X axis
SYMBOLS U S E D about Y axis; or yield; or fiber at distance y from the
Main (as distinguished from subscript) neutral axis
about X' axis
A = area between fiber in question and the extreme fiber about Y' axis
a = angle the neutral axis makes with respect to the
X axis
b — maximum width of the section INTRODUCTION
j8 = angle the moment vector makes with respect to

c
the X axis
— perpendicular distance from a principal axis to the
extreme fiber
T HE CLASSIC BENDING THEORY is based on the
assumption that plane sections before bending
remain plane after bending. Since, within the elastic
e = unit strain range, stress is proportional to strain, it follows that the
/ = applied stress stress distribution over the cross section will be linear,
F — allowable stress
varying from zero at the neutral axis to a maximum
v? = angle the principal X' axis makes with the X axis
j = an additional (multiplying) factor of safety at the extreme fiber. This distribution is expressed
k = section factor mathematically for a symmetric cross section subjected
m = applied moment to simple bending by the conventional formula
M — allowable moment
P — normal load on portion of cross section fb = my/I (1)
Q — static (first) moment, about the neutral axis, of
the area between the fiber in question and the where
extreme fiber
R =. rate of change of trapezoidal stress intercept with /» = bending stress at a fiber that is located at a
respect to the stress of the extreme fiber; or distance y from the neutral axis
stress ratio m = the applied moment (lower case used to differ-
S — shear load entiate applied moment m from allowable
Presented at the Structures Session, Eleventh Annual Meeting,
moment M)
LAe,S., New York, January 25-29, 1943. I = the moment of inertia about the axis of bend-
* Stress Methods Group Engineer. ing
137
 138 J O U R N A L OF THE A E R O N A U T I C A L S C I E N C E S — MAY, 1943

When, on the other hand, the extreme fiber stress ever, has approximated this relationship by expressing
exceeds the proportional limit stress of the material, stress / a s a function of some power of strain e, (f =
the stress distribution over the cross section is no longer aen). This method, however, was also limited by the
linear b u t tends to follow the stress-strain characteris- mathematical difficulties encountered in applying it
tics of t h e material. 1 Consequently, t h e conventional to typical problems.
formula for linear bending stress distribution no longer W i t h the idea of utilizing, whenever possible, the
applies. I t has been s t a n d a r d aircraft practice, in common physical properties of a cross section, a
such cases, to use a "form factor" to indicate the trapezoidal distribution approximating t h e actual
a m o u n t b y which t h e allowable ultimate stress of t h e stress-strain distribution was found to give a practical
material could be considered to be increased in solution. T h e trapezoidal distribution was varied, of
order to apply the conventional fb = my /I form- course, depending on the extreme fiber stress in-
ula. volved.
T h e use of this simple expedient of a form factor A formula was developed to determine t h e m o m e n t
necessitates having available t h e form factors for all of resistance offered by a cross section subjected to
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

the various shapes of cross sections commonly en- any bending m o m e n t m. T h e corresponding formula
countered in practice. Such form factors for round for shear distribution was also developed.
tubes have been incorporated in the "Bending Modulus In order to account for the effect of material proper-
of R u p t u r e Curves for R o u n d Chrome Molybdenum ties, curves enabling the designer to account for plastic
Steel T u b i n g . " 2 F o r m factors for sections other t h a n bending phenomena were developed for each of various
round are not available to t h e average structural materials commonly used in aircraft practice.
engineer. In view of this lack of material, it was A discussion of t h e proposed method of handling
proposed to m a k e a further s t u d y of form factors for local crippling, complex bending, and combined axial
various I-type, channel-type, and round tubular types load a n d bending is included, together with t h e re-
of cross sections. Simple bending tests were m a d e spective limitations of t h e proposed method.
on such sections, and t h e d a t a 3 pertinent to this article
are included in the Appendix. M O M E N T OF RESISTANCE OF A SECTION SUBJECTED
TO B E N D I N G
T h e current "form factor" method of accounting for
the effects of bending in the plastic range is subject Linear or Triangular Stress Distribution
to t h e following limitations: T h e stress-strain curve m a y be t h o u g h t of as a stress
(1) T h e physical significance of t h e plastic bending distribution curve if t h e vertical axis is assumed to be
phenomenon is lost b y t h e use of a fictitious stress the neutral axis.
greater t h a n the allowable ultimate stress Ftu of t h e Fig. 1 (a) shows a typical stress-strain curve in which
material. t h e strain of t h e extreme fiber ec corresponds to a stress
(2) I n m a n y cases the full a d v a n t a g e of the form fb which is below the proportional limit of the material.
factor effect a t ultimate loads cannot be realized since T h e equivalent stress distribution is shown on section
the yield strength criteria m u s t also be met, particularly A - A o f Fig. 1(b).
in aircraft work. Therefore, yield form factors as
well as ultimate form factors should be m a d e avail-
able.
(3) T h e stress distribution across t h e entire cross
section is fictitious and, therefore, local crippling
effects cannot be t a k e n care of unless a sufficient number
of tests are r u n on various sections of varying wall
thickness. (The bending modulus of r u p t u r e curves
for round tubes of various D/t values include this
"local crippling" effect. 2 )
(4) In order to include t h e effect of material proper-
ties, a complete set of form factors would have to be
m a d e available for each material. F I G . 1. Triangular stress distribution.
T o remove the above limitations it was decided to
abandon the conventional form factor theory and T h e moment of resistance mb of the section subjected
evolve another method for handling the plastic bending to bending m a y be expressed as
phenomenon. T h e use of a bending stress distribution
mb = MI/c) (2)
based on t h e exact stress-strain characteristics of t h e
material was considered b u t was found to be imprac- Subscript b designates a triangular distribution;
tical because of t h e difficulty of expressing t h e stress- fb is the corresponding extreme fiber stress; and I/c
strain relationship mathematically. Beilschmidt, 4 how- is the conventional section modulus.
 B E N D I N G S T R E N G T H IN THE PLASTIC RANGE 139

Constant or Rectangular Stress Distribution m u m fiber stress of fr and a triangular distribution with
If a stress-strain curve of a material were as shown in a maximum fiber stress of fb.
Fig. 2(a), the shaded area could be t h o u g h t of as the T h e values of mb and mr when substituted in Eq.
stress distribution. T h e equivalent stress distribution (4) will give
is shown on section A-A of Fig. 2(b).
m f„(I/c) + 2frQ„ (5)
T h e m o m e n t of resistance mr of t h e section subjected
to bending m a y be expressed approximately as However, it m a y be seen from Fig. 3(a) t h a t fr — f0
t h e trapezoidal intercept stress and t h a t fb = fm — / 0 .
mr = 2frQ„ (3) Therefore, these values when substituted in Eq. (5)
Subscript r designates a rectangular distribution; will give
fr is the corresponding extreme fiber stress; and Qm is m = {fm -fo)(I/c) +2f0Qn (6)
the maximum static m o m e n t of area (i.e., t h e first
moment about t h e neutral axis, of the entire area By regrouping the terms, this expression m a y be
between the neutral axis and t h e extreme fiber). further reduced to
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

m = fm(I/c) +fo[2Qm - (I/c)] (7)

In order to utilize the conventional expression tnc/I,
divide through by I/c to obtain

#-0
mc
— fm + fo (8)

T h e term mc/I m a y be t h o u g h t of as a measure of

t h e m o m e n t of resistance t h a t could be developed b y
the section, for when multiplied b y the section modulus
I/c it will give t h e m o m e n t of resistance.
FIG. 2. Rectangular stress distribution. T h e term 2Qm/(I/c) which depends only on t h e shape
of the cross section is designated by k and will be re-
Trapezoidal Stress Distribution ferred to as the ''section factor."
Consider the condition where the extreme fiber stress In its final form the expression reduces to
fm of a section subjected to bending lies between the mc/I = fm + f0(k - 1) (9)
allowable yield stress Fty and the allowable u l t i m a t e
stress FtU} as shown in Fig. 3(a). T h e stress-strain Basic Assumptions
distribution has been approximated b y a trapezoidal
In order to apply the expression for trapezoidal dis-
distribution as indicated b y t h e shaded area. T h e
tribution to t h e general problem of plastic bending, the
stress indicated b y t h e intercept of t h e trapezoid with
following assumptions are m a d e :
the vertical axis through t h e origin is designated as t h e
(1) T h e tensile and compressive stress-strain proper-
trapezoidal intercept stress / 0 . T h e equivalent trape-
ties of t h e material are identical. This gives a simple
zoidal distribution is shown on section A-A of Fig.
distribution and avoids the difficulty of a complex
3(b).
distribution. These difficulties will be discussed in
the latter p a r t of this article.
(2) T h e trapezoidal intercept stress / 0 is to be
determined so t h a t t h e first moments about the neutral
axis of t h e area under the actual stress distribution
curve are equal to the first moments about t h e neutral
axis of the area under the equivalent trapezoidal stress
distribution curve. I t is true t h a t such an approxi-
mation is exact for a rectangular cross section only.
T h e variety of cross sections encountered in practice
A-J fb)
precluded an investigation on all of them. Therefore,
FIG. 3. Trapezoidal stress distribution. it was decided to consider t h a t most aircraft sections
could be broken down into three types.
T h e m o m e n t of resistance m of the section subjected T y p e (a) is the rectangular type, in which the
to bending m a y be expressed as material is distributed uniformly from the neutral axis
to the extreme fiber.
m — mb + mr (4) T y p e (b) is the diamond type in which the a m o u n t
since the trapezoidal distribution m a y be considered of material decreases from a maximum a t t h e neutral
as m a d e u p of a rectangular distribution with a maxi- axis to a minimum a t t h e extreme fiber. Under this
 140 JOURNAL OF THE A E R O N A U T I C A L S C I E N C E S - M A Y , 1943

classification would come the solid round and relatively TABLE 1

thick-walled tubes. Values of Section Factor, k = 2Qm/(I/c)
Type (c) is the more common hour-glass type, in
which the amount of material increases from a mini- Section Value of k
mum at the neutral axis to a maximum at the extreme
fiber. Under this classification would come the / ,
channel, and most of the tubular sections.
The relative values of trapezoidal stress intercept
!-l- 1.0

- | - 1.5 |
for these three types have been shown, in the Appendix,
to be approximately equal. This implies that the
resisting moments for all three types can be predicted
by the use of the equivalent trapezoidal distribution
-£- 1.0 to 1.5

based on a rectangular cross section. -1~ 1.0 to 1.5

-®-
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

1.25 to 1.70

- •A- -— 2.0

— 1.5 to 2.0*

-i-
F I G . 4. Type (a). Rectangular type.
4^3
* Any section having a k value larger than 2.0 should be used
with caution as it indicates a relatively thin outstanding leg.

Since it is more convenient to work with shear

flow, Eq. (10) will be replaced by the equation
q = SQ/I (10a)
where q equals the shear load per inch.
The new expression for q is derived in the Appendix
F I G . 5. Type (b). Diamond type. and may be written as
- SQjl + R[(Ac/Q) -1]\ (n)
2
l\ 1 + R(k - 1) i
where Q equals the first moment about the neutral
axis, of the area A between the fiber in question and
the extreme fiber and R equals df0/dfm or the rate of
F I G . 6. Type (c). Hour-glass type.
change of the trapezoidal intercept stress with respect
to the maximum stress.
Limits of Factors
In the elastic range where the trapezoidal distribution
The range of values for the section factor k for various becomes triangular (/0 = 0), the value of R is zero and
sections is shown in Table 1. the expression for shear flow reduces to the conventional
It may be interesting to note that Table 1 will be form.
found useful to check preliminary section properties.
If the value for a section factor lies outside of the range DEVELOPMENT OF CURVES OF PLASTIC BENDING AND
shown in Table 1, it is fairly certain that an error has SHEAR
been made. This simple check has been used
extensively in the Lockheed Engineering Depart- In order to make use of the relations discussed in the
ment. previous pages, certain factors and properties of
individual materials will have to be made available.
SHEAR DISTRIBUTION The basic need is for a complete stress-strain diagram
for each material. By complete is meant a diagram
The shear stress distribution corresponding to the
that shows the stress-strain properties up to and
trapezoidal bending stress distribution will show a
including the ultimate stress. As the aeronautical
marked change over that obtained from the con-
structural engineer is faced with more and more re-
ventional formula:
finement of design, the need for complete stress-strain
fs = SQ/bI (10) diagrams becomes more urgent. In fact, much of the
 BENDING STRENGTH IN THE PLASTIC RANGE 141

future development in structural analysis will be

TABLE 2 3
dependent on t h e plastic stress-strain characteristics Comparison of Test and Predicted Moments
of materials.
Stress-strain curves have been assumed for some rr~ F2~ p*~• |1 16^
Ratio 1"
common aircraft materials so as to satisfy t h e minimum
guaranteed values of: (1) modulus of elasticity, (2) rH Actual Average

Section
proportional limit, (3) yield stress, (4) ultimate stress,
•H To Pre- X-Ray
*-< g d i c t e d Rating
and (5) ultimate elongation. <D
Moment for
B y means of a tabular form, a copy of which is shown
in t h e Appendix, t h e stress-strain curve for each of
,1 <D
(15) Casting
CO
several materials was integrated numerically t o give (10)
an equivalent trapezoidal distribution with a minimum
stress a t t h e neutral axis designated as f0 and a maxi- 31 1.104
m u m stress a t t h e extreme fiber designated as fm. 80 1.105
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

Values of fm were assumed a t various points on t h e 33 1.079

stress-strain curve, a n d t h e corresponding f0 was found 13 1.064 mm 1

so t h a t t h e first moments about t h e vertical axis through 14 1.070

the origin were equal for (1) t h e area under t h e trapezoid •H
bO 35 1.045
and (2) t h e area under t h e stress-strain curve. This *-*
37 1.039 mm 1

value of /o was substituted into t h e expression: £ 1 1.054

EH
mc/I =fm+fo(k- 1) (12) CO 2 1.028
rH
3 1.020
I-Sections
for values of k = 1.0, 1.2, 1.4, 1.6, 1.8 and 2.0. *** 1
T h e values of intercept stress f0 were plotted versus 4 1.059
fm and the slope of this curve (df0/dfm) was in t u r n 5 1.030
computed numerically to give values of R. 6 1.040 «* 1

T h e n for each material, there appears on one sheet: 7 1.013

(1) a plot of stress / versus strain e, (2) a plot of stress
•H
8 1.040
/ versus mc/I for various values of k, (3) a plot of
stress / versus R, and (4) values of fo/Fty and u 9 1.057
fo/Ftu-
10 1.062
11 1.050
All t h e available curves of this t y p e have been
included as Figs. 9 through 17 of this article in t h e form i 12 1.013 •» I

in which t h e y were used in t h e Lockheed Engineering 48 1.114 Fair- 1

Department. 49 1.135 Poor
• 43
18 75 1.225
G E N E R A L DISCUSSION OF CURVES 2rH CW 6o
78 1.113
Bending Stress Distributions
Round Sections

15 1.123
If t h e complete stress-strain curve of a material IH.5 79 1.063
CO feD
were available, a n y value of extreme fiber stress (fm) ^* U
•HO 17 1.125 -
within t h e limits of t h e material could be assumed 19 1.002
and a complete picture of t h e actual bending stress feO
39 1.336 Pair 1
distribution over t h e remainder of t h e cross section
could be had. On t h e assumption t h a t plane cross
40 1.131 Fair-
sections remain plane after bending, t h e stress a t any •H o 41 1.440 Fair J
intermediate point (y) corresponding t o t h e propor- 29 1.078
tionate strain (ey) of t h a t point relative to t h e maximum to
27 1.037 -
Channel Sections

strain (em) of t h e extreme fiber (i.e., ey = (y/c)em) CO tiO 21 0.965

could be determined. <# u 25
rH O
1.095
Moments of Resistance Corresponding to an Extreme 23 1.024
Fiber Stress (fm)
62 1.054 Poor- J
Corresponding t o a n y extreme fiber stress (fm) t h e to 60 1.080 Poor*
CO S*
m o m e n t of resistance m m a y be found b y entering t h e EH « 61 1.235 Fair-
stress- (mc/I) curve with fm and determining t h e 1 43
mc/I value for t h e proper "section factor" k. 2£ *»l 59 1.075 Poor+
Then w = (mc/I)(I/c). 6 3-A 1.117 | Poor I
 142 JOURNAL OF THE A E R O N A U T I C A L S C I E N C E S — MAY, 1943

When t h e extreme fiber stress is chosen as the TABLE 3

allowable yield stress Ftyy then t h e resisting moment Average
will be t h e allowable yield moment, My. Section Material Ratio
When t h e extreme fiber stress is chosen as the / 14ST forging 1.057
allowable ultimate stress Ftut then t h e resisting mo- Mg. forging 1.039
m e n t will be t h e allowable ultimate moment, Mu. 195-T6 casting 1.147
For a n y applied m o m e n t m t h e corresponding Round 14ST forging 1.078
195-T6 casting 1.302
extreme fiber stress and t h e stress distribution over Channel 14ST forging 1.040
the entire cross section m a y be found b y entering t h e 195-T6 casting 1.112
curve with mc/I a n d determining t h e extreme fiber
stress fm for t h e proper k value of t h e section. After Shear Flow
the extreme fiber stress has been obtained, t h e stress
distribution over t h e remainder of t h e section m a y be T h e shear flow (g) for a n y section m a y be found b y
found for a n y intermediate point as described under entering the curve with the mc/I value and determining
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

the heading "Bending Stress Distributions." t h e corresponding value of R for t h e proper k value of
the section. This value of R when substituted in t h e
Comparison of Test Results expression for shear flow will give t h e desired result.

Simple bending tests 3 h a d been r u n a t Lockheed on Shift in Neutral Axis

various i", channel, and round sections for 14ST forging,
T h e actual neutral axis of t h e section will n o t always
magnesium forging, and 195-T6 castings. Ratios of
pass through t h e geometric centroid of t h e section. F o r
actual test values to values computed b y t h e proposed
instance, consider t h e case in which, although t h e
method are given in Table 2. A more complete table,
tensile a n d compressive properties of t h e material are
showing t h e computations, is included in t h e Appendix.
identical, t h e section is n o t symmetric a b o u t its cen-
T o reduce test results from actual physical properties
troidal axis. F o r such a case t h e neutral axis will no
to minimum guaranteed properties, a correction factor
longer coincide with the centroidal axis in t h e plastic
based on the relative first moments of t h e respective
bending range. Although this effect is usually neg-
trapezoidal distributions, about t h e neutral axis, h a d
ligible, t h e designer should be aware of its presence.
to be determined. T h e derivation of this correction
factor is also included in t h e Appendix.
Local Crippling
Correction factors were applied to all b u t t h e casting
tests, since, in t h e case of casting design, additional A discussion of modulus of r u p t u r e effects would not
factors of safety are included to t a k e care of possible be complete without mention of local crippling phenom-
defects in castings. Physical properties of castings ena. This is especially t r u e in connection with
determined from a test billet do n o t necessarily predict structural aircraft practice, since t h e "Bending Modulus
the physical properties of t h e casting proper; therefore, of R u p t u r e Curves for R o u n d Chrome M o l y b d e n u m
average X - r a y ratings for t h e castings are given in Steel T u b i n g " 2 has been in use for m a n y years.
Column 17, Table 2. An application of t h e proposed m e t h o d t o "chrome-
Column 16, Table 2, shows t h e ratio of actual m o l y " steel tubes heat-treated t o 150,000 lbs. per sq.in.,
allowable test m o m e n t s (reduced t o minimum guaran- for various ratios of B/T gave results shown b y curve
teed values except for t h e castings) t o allowable com- A F B C in Fig. 7. Curve D B G E has been reproduced
p u t e d moments based on minimum guaranteed proper- from Figs. 4-20 of ANC-5. 2
ties. There is only one case in which this ratio is less On t h e surface it would seem t h a t t h e two methods
t h a n unity. Specimen 2 1 , which is a channel section are inconsistent. T h e differences, however, in t h e
of 14ST forged material, shows a ratio of 0.965, which ranges of B/t larger t h a n 12 m a y be a t t r i b u t e d to the
is equivalent t o a negative margin of 3.5 p e r cent.
Other specimens from t h e same billet show positive 260
margins.
0)
A
*? Predicted (does
£3 240 I not include
T h e average ratio for t h e various t y p e sections and U T *

<H P<
MM
?zo fcj M i l l A] 1
"looal crippling")

materials is given in Table 3.

MNU
L MUKI
o*i
r M
200
TO44-U r T T T
modu
usand

r H CO
Reproduced from
Margins of Safety
Sf5
180
°NM IT ANC-5 (Ref. 2)
(includes "local
In t h e plastic range, t h e margins of safety m a y no
•H +>

1.3
/t>0 Ml \^vA\ i crippling")
.O
longer be found from t h e ratio of t h e allowable stress t 140 Tn^-4-B
to t h e actual stress, since actual stress is no longer tf
1ZO
directly proportional t o loading. T h e margin of safety 0 20 30 40 50 60 70

m u s t of necessity be found from the ratio of allowable

F I G . 7. Bending modulus of rupture "chrome-moly" steel heat-
moment to actual moment. treated to 150,000 lbs. per sq.in.
 BENDING STRENGTH IN T H E PLASTIC RANGE 143

effects of local crippling. There is a definite need,

therefore, for actual crippling data which, if available, +Y
could be used as allowable fiber stresses in the proposed
method to predict allowable bending moments. For ^„ I i
the present it will be conservative to use the proposed #1 'm
method and ensure that no fiber will be stressed beyond
its ''crippling stress in pure compression," whenever Pr B j4A
\
such data are available. For round tubes, however, T <£/?
the ANC-5 bending modulus of rupture curves should h +X
be used, since they already include the crippling ?n>
effects.

Complex Bending
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

The phenomenon of complex bending, which occurs

when the resultant applied moment vector is not
parallel to one of the principal axes, is quite common
in aircraft practice. The term "complex" bending has FIG. 8.
been used instead of "unsymmetric" bending, since,
even though the word unsymmetric is usually asso- implies, therefore, that the moment vector is the axis
ciated with the shape of the section, it is possible for of the moment.
the resultant moment vector to be nonparallel to a
principal axis of a simple symmetric shape. The Let an = the angle the neutral axis makes with the
classic beam theory gives a formula for complex bending X axis in the elastic range
stress as follows:6 _x myIx + mxIXi
= tan (14)
{myIx + mxIxy)x — (mxIy + myIxy)y mzIy + myIX]
h = (13)
±%l-y
(/„)* for the general case but, due to the
X and Y represent two mutually perpendicular symmetry of this case, it will reduce
centroidal axes; x and y represent the coordinate dis- to:
tances from the center of gravity measured positive to tan-1 {myIJmxIy) (15)
the right and upward, respectively; mx equals the ap-
plied moment about the X axis, the moment being Let /3r = the angle the resultant applied moment
positive when it tends to put compression in the vector makes with the X axis
"upper" fibers; mv equals the applied moment about tdLn~l{my/rnx) (16)
the Y axis, the moment being positive when it tends
to put compression in the "left" fibers, and Ix, Iy and If the ratio of applied moments my/mx is assumed to
Ixy are the respective moments and product moment be approximately 5 and the ratio of moments of inertia
of inertia about the X and Y axes. IJIy is assumed to be x/u when b/h = 4, then
This formula is applicable only in the elastic range; an = t a n - W ^ ) - 17.4°
therefore, it needs to be modified for plastic bending. pT = tan" 1 5 - 78.7°
However, since the development of such a formula is
beyond the scope of this article an alternative plan The section and relative moments shown in Fig. 8
is proposed. In order to clarify the relationship be- represent the conditions of an actual lug that came up
tween section properties, applied moments, location of in practice. The large difference in relative position
the neutral axis, and principal axes, a review of complex of neutral axis and moment axis as indicated by the
bending in the elastic range will be necessary for com- angle /3r — an seems unusual. However, it has pur-
plete understanding of the alternative plan. posely been included in this discussion so as to familiar-
Consider a rectangular section as shown in Fig. 8. ize the structural engineer with this phenomenon. The
The respective applied moment vectors about the physical significance of this case indicates that, even
centroidal X and Y axes, which in this case are principal though mx is small relative to my, it is still capable of
axes, combine to give a resultant moment vector m at developing a relatively high compressive stress at
an angle of /3r with the X axis. Moment vectors are point A on the section. In fact, the compressive stress
designated by double-headed arrows. They are to be in the upper fibers due to mx is not completely counter-
interpreted by the left-hand rule (i.e., point thumb of acted by the tension stress in the upper right fibers
left hand as indicated by the double-headed arrow and due to my until point B is reached.
the natural curl of the remaining fingers of the left As long as the maximum stress/ w of the fiber fartherest
hand will designate the direction of moment). This away from the neutral axis (i.e., the upper left-hand
 144 JOURNAL OF T H E A E R O N A U T I C A L S C I E N C E S — MAY, 1943

corner) is below the proportional limit of the material, SUMMARY

the neutral axis location will remain as indicated
The following summary constitutes a presentation
above.
of the proposed method suitable for direct engineering
On the other hand, as soon as fm exceeds the propor-
use in aircraft work.
tional limit and the triangular distribution of stress
relative to the neutral axis changes to a trapezoidal General
distribution, the neutral axis will rotate toward the
resultant moment vector. This rotation will be a The conventional beam theory (based on the as-
function of the relative distribution area on either side sumption that a plane section before bending remains
of the neutral axis as it rotates about the center of plane after bending) gives a linear distribution of strain
gravity of the section. Pending a further study of this and stress in the elastic range (i.e., up to the propor-
phenomenon, it is proposed that the stress-ratio tional limit). In the plastic range, however, although
method 6 be applied with a straight-line interaction the strain distribution is assumed to remain linear,
curve between the following stress 'ratios: the stress distribution corresponds with the stress-
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

strain relationship for the material.

Rx = (mx/Mx) and Ry = (my/My) An approximation of this distribution has been
where Mx equals the allowable moment about the obtained which enables the prediction of the effects
principal axis X and My equals the allowable moment of shape and material properties on bending in the
about the principal axis F. plastic range.
The margin of safety may then be expressed as Graphs are included (Figs. 9 through 17), each of
which is applicable to a particular material. These
M.S. = [1/(RX + Rv)] - 1
graphs are to be used in connection with the following
Difference in Compressive and Tensile Properties of a procedures.
Material
Simple Bending
All the preceding derivations and conclusions have
been based on the assumption of identical tensile and (A) Determine the "section factor" k = 2Qm/(I/c).
compressive properties of the material. Although this (If k > 2.0 use k — 2.0.) Qm is the static (first) moment
may not always be true, the means of evaluating a about the neutral axis of the area between the neutral
new effective center of gravity of the cross section, axis and the extreme fiber. I/c is the conventional
based on a different distribution for the tensile and section modulus.
compressive bending stresses, are beyond the scope of (B) Bending stress at the extreme fiber due to any
this article. Since this situation also occurs in the applied moment, m. (1) Determine mc/I. (2) Enter
present usage of the classic beam theory, it is proposed the graph with mc/I and obtain the maximum stress,
to neglect it, at least until the more important factors fm, at the extreme fiber for the proper "section factor"
are taken care of. k.
(C) Bending stress at an intermediate fiber due to
Combined Bending and Direct Load any applied moment, m.
Many complications enter into a practical means of (1) Obtain the stress at the extreme fiber from
finding the position of the effective center of gravity. item (B).
Even the simplest bending case in the plastic range will (2) Enter the stress-strain curve with this stress
be subject to this effect when a direct axial load is and obtain the maximum strain em at the extreme
included. Further study is definitely required with fiber.
respect to this phenomenon. (3) At the proper proportional strain ev of the
Here again it is recommended that the stress-ratio intermediate fiber, which is at a distance y from the
method 6 be used with an interaction between the three neutral axis, read the corresponding stress, where
stress ratios represented by a plane surface: ev = (y/c)em.
Rx + Ry + Ra = 1.0 (D) Margins of safety in bending. (The margins
must be determined from the bending moments since
where Rx equals mx/Mx, Ry equals mv/Myy and Ra stress is no longer proportional to moment.)
equals fa/Fa, and where fa is the applied axial stress, (1) Enter the graph with the allowable ultimate
Fa is the allowable axial stress, and x and y subscripts tensile stress Ftu to obtain an mc/I value for the proper
indicate "about the principal X and Y axes," re- "section factor" k.
spectively. (2) Multiply this value by I/c to obtain the allow-
The margin of safety in this case may be expressed able ultimate moment Mu.
as
(3) Repeat (1) using the allowable yield stress
M.S. = [l/(R* + Ry + Ra)} ~ 1 Fty to obtain the allowable yield moment My.
 BENDING STRENGTH IN THE PLASTIC RA N G E 145

OC 'o a: i f
frl
*
^L a.
3 •
h\
"> rM
C
o
~ .5= *T o
\\\
cc
. X

1
1 •
r\
<n
i\\
cu Js 2'
5 a 1 / \u
-
ql_
J\ Z
i-
o
a - {2
U
\
to

SI*
n^-\ \\ ^
o o
\ vj
11 >. D

< o
&

i \ \ ^ 0a
NJf\ \ \
o tf>M 3
is u- <0 2 M
en
_
n - ~ X* o jl
1\ \ oc
3 § 5 2 "
u\i \\ \^A
1

<n
/ \
l\ \[ \ \

The Effects of Shape and Material Properties

/ \\1 1 ° S
- o oL
\T\ o _

\K\ \ k ^

ALUMINUM ALLOY 'SAND CASTING '

- 0-

K\ /
1
—j- 3 O O
s.
f £ j — — ~
\ > \ \

Bending in the Plastic Range

4> o\ V
ft\
. \ Si* \
y
—
•nL K\ \
"5 e «) ——
v
v 1 f°l
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

o v0 U O - \ *\ \\\ 1III°l 1 ° °. 1
w
c-/Ar
1\ \ in |

\\
I i
—r w

\ \ 1
^ A\ ?
\ \ V
V
H CO

\ -^ \
^V u?
IT r j_4§-
's $^ \ \
-a oU
^5T
1 o .

\
NW\T _i •> _ « °-
v>X 1
r •^$ d wl
^ i\i\t 1
f ft) -r-
> q
\ V CK V) \
,NM
L
|
O
3
Lh

Figure 12
f
>s >>Vfc; - >
. 3
\ in: \MI 1 h *>
0)

220 T4
. o o
V 41 -
T T"C

on
O

u - ^ v ^ io \V 1 1 °
3 V 1
d d \ H E
li

|
.5 ct£ k:^ -s tl II
~ li!
3 J T>\ 1—h§
a ^
^ u? M?I uT
u.
• 2

u
A $-°2 1 iI^ < 7 ^v
X
•83-
^
\
\ ^-R5
<
I I I -ol
I I I ^l
s to

• o ~ lu> ^ V ,
o
\{ k\l I col
O 1 o o
* o o o M
sf
lf>
12 ? 3
r * I
v> 'o
Qu 3 S. , 2 j
X
u. in .$ u. 1

II
M
1 1 1II 1 1
_
CC od
A

Properties on Bending in the Plastic

>

Material
- / ~\ c ]c fl\ -
o
a
on Bending in the Plastic Re

• * •

/ \ K
f~ " DC

\
Material

1
EXTRUSION

\ o X \
The Effects of Shape and
z
/ \ S '
a: 1
{
Figure 10
o
.. / \ u.
/ ^
w
Effects of Shape and

\ / \
'o .
1 1\\
ST
'

i|"
9

i\ \] < / V
\ / \
24

1\
Figure

o."
\ \ / \ st
/1
c o_
c <vJ-
O 3 _ |
l 5trqi ?> e 1 1. I I -
w /
Range

c
\ \
1- n^ ' \ < \
erties

1-
^ K \
| \ |c
|\ \
100

I"
IPO

" > \
c
\ ,
Jl L
M\ \
V N r1 \ ^s
r
,00 1

\ I 1 o
i\H ^ N\ 1 ss^S
|c
o -
00 p
M i

[S % K ^ \s k\ 9 ^

N ^ *\ A t^ \
\
"* 1
o_ H'X * ^f \ \4
\\ - «n
> o
o .
| J 3 p •|!
N~ \ \\ 4sNrK O
ranteed Prop

v
IV % \ \ \\ v\ ^—
;
a | >
r °^>St4 ^ m
i • •

N
#5«
"° J '>k \ N \ T

4 > *** \, \ ^ d
3 Q
tfl O
8- l
_ O _
^
^ \ ^ \ £ o
" -7 \ ^ li II

\ % \^ % ^1 M 6 d
\ c J\
J?
*\ UT 3
^ ^ - ° —1 5
\ i
E = \
\ ^ ^1^ .0 - J u.
\V ?f- ^ \ o . E3
o

\ Ji - <o -

)
LL"

- iJs
O -

N
u
p - -e-
, 2c
\ / - s -j \ V
*m
V ~ e H o
— 3 _ — c —
H —
\ o

• I f* r
|c
o
o o
l o - L o.
? ® \ S
CO J
(O
lO S o
<
v> 5
- g o
\ * o 1
S o-
^ ^Jr V »-s U-|

C .0I x j 'iffd u| sspjts

 25 j isjfe in psi x I0~ 3
0 5 | 10 15 20 30 35
0
"40
0.2 0.4 R
I I
1-t.i" 1 120 160 200 140 280 •^ in psi x 10-*
/ / / / 0 40 60
0 2 4 6 8 R
A/ / / / 'Y !— w o p r
tu
/ / — 7/~ ^ ^C
/'•AX, / h/w
/ £ Stress - S t r a i n & hr V»-

y. tAfrft k/
— I40f-
C u rve 1X - S$r N
rw^ -
&rAr
WK*r
^>
r
t M

f.'%#&
/ ty&r /I_^-Stress - R Curve
A
/ / A A,
/ A^
U - 0.S30
^ "A
r i
\
\
— I20U k
^
/
/
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

F
ty~
A A7y fo
|\ 1?
tr
)
tress - M£ Curves
s.-
0i69 --S
= 100k
^y Su LL.
/ _ 0.445
s ^ ^ '<6y'iW
fo

A N, .- / Fty

/ ^ yA y
/ -
«*•

* )\ fo
O.i 61
Curve - ^ / /' A%^ 13 ALUMINUM ALLOY 01E CAST1N6
\ - / •hi

^ ^ <y Stress - R .•
s_ 2 /
A ^ The E f f e c t s of S h a p e a n d M a t e r i a l Properties, UrV
^~^' lO / 4330XX STEEL CASTING 150,000 HT

/ on Bending in the P\ast-\c Range /

/ Figure 13 — 60h / The Effects of S h a p e and M a t e r i a l Properties

/ '-Stress - M t Curves
/
/S I | / i t i . i i . l , , ! . , ,

1/// *> . / Fi gure 15.

0w .001
1 '..II
""'6
/
.02 .04 .06
1 i
.06 .10 .12 .14 .16 .IS
.007 .008 .009 .01
1 1 * Based on Minimum Guaranteed Properties I I I I i i *i i I I -I M I I I I I 1 I I * * Based on Minimum Guaranteed P r o p e r t y . | Strain,, e j j | J j | 1 1 1 1 1 1 1 1 1 1 ?*

1 1 1 1
1 1 1 '1 1 1 11 1
0 40 80 120 160 200 240 280 M^ inps i x 10 " 3
1
" 0 2 4 6 8 R
150 k- 1
-Ftu
1 o/ 1 ' K D ^ X "S
S^ / 0 10 20 30 40 50 60 70 -M£. in DSI* A I0"3
1 X ^\J
sgf > ^ a - * r^SB-
I
b*tress-Strain 4
/ • Fty <Jurve y
r | 1 0 0.4 0.0 1.2 K
42,
r 7/ i— F
tu:
7,^ ~Y /
kw ^
'?
/
^= =""-
^-Stress-R Curve
*5tress-Strain ^,
\A V*-

/
/
Stress
r 7

•^/ "^^ i>

y^
V-^

^r
in p s i , \y>

F x 10"3 F
/ / — Stress
Curve-
(
fp
A- S t r e s s - * ^ Curves
*f / / y
V\
in psi,
/ 1—' ' ' *0
_ 0.. i.94
Fx 1(T3
mJ_ / Ft 9
A „._,/^^fef^
—r V
^t>
/ fo
_ O f 165 f
/ ^ //~;V ^> ^-. -
"+o o.r ao

A Ffu
20 / A^ ^ V
^
/ X4I30 STEEL F0R6ING 150,000 H T _
— / '0 - ift 0.i75
60\- / 1 Jf Ftu •Stress- R Curve
1 1
/ The Effects of Shape and Material '
'Ftp
//
/ Properties on Bending \r\ the Plastic Rang e
10
/ \ - S t r e s s - M°. Curves
AM 5 8 5 MAGNESIUM F0RGIN6

60 U / , Figure (4 The Effecfs of Shape and Material

/ / 1 Properties on Bending in the Plastic
/ — 0 V 1
/ 3 .010 .020 .030 .040 ' ' .050

/
1
Fi
Strain, e ?" . . 9ure l6
.
0 .02 .04 .06 .08
n B
.10 .12 .145
U Jon! 1inimu1 1 1jaran-
J _Based 1 1 1 1 1 1 1 -1 ~ -
1* Bo s e d on M iniru u r n Ck i a r a n t e e d Pro p e r t i e s St rai
» ( *u
eed f ties IV m Gi
3
ro ser .
 BENDING STRENGTH IN THE PLASTIC RANGE 147

AM 265 HTA MAGNESIUM SAND CASTING

The Effects of Shape and M a t e r i a l Properties

u or\ Bendfng in the Piastre Range

Ffgure 17

1 • 1
1 1 \ 11
1 11
0 10 1 1 1 1
20 30 40 1 50 ^ in psi x 10"3

0 0.2 0.4 0.6 , 0.8 , i_ 1.0 R

34 Ftu / y /\
s
30
)
<y<W\)y/<os*
5^v
$&&& 2v
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

IStress
in psi,
i
20
. Stress - 5 f rain f\ ?, ^ '/
<r
/ - S t res.s - R Curve

Curve-A y
F X 10~3
1 F
tu X
**
£ % 't>
A 'A ^s'/
rS. 'V*
*0
rty
= a: 96
^
S
^£
^ - o.<£91
/ r
y
J &
^ &^
til
r'tu
10 ^
— — 2
¥1 z^ S t r e stP
n- s - Mt CurVes
/ / A

/ /
J
°()1/ .004 .008 .012 .016 .020 .024 .028 .030
. , Strain, e i i i U

* Based on Minimum Guaranteed Properties

(4) The margin of safety on an ultimate basis: Ac

1+ R
M . S . „ = (Mu/jumu) - 1 q (shear flow) =
SQ
I 1 + R(k - 1)
j u is an additional (multiplying) ultimate factor of
safety, if any, otherwise use7M = 1. mu is the applied where 5 equals shear load and Q equals static (first)
ultimate moment. moment about the neutral axis of the area A between
(5) The margin of safety on a yield basis: the fiber in question and the extreme fiber.

M.S., {My/jymL) 1 Local Crippling

j y is an additional (multiplying) yield factor of safety, (A) Assume critical portion of section is developing
if any, otherwise use jy = 1.0. mL is the applied its crippling stress.
limit moment and is usually = mu/\.b. (B) Enter stress-strain curve with this crippling
This method of determining bending allowables has stress and determine the strain of the critical portion, ey,
been substantiated by tests on aluminum-alloy castings (C) Determine the proportionate strain of the ex-
and forgings, and on magnesium forgings of relatively treme fiber em, where em = (c/y)ey.
thick-walled sections only. (D) Enter the graph with this strain em to obtain
(E) Shear flow at any fiber due to any applied an mc/I value for the proper "section factor'' k.
moment, mf and the corresponding shear, S. The (E) Multiply this value by I/c to obtain the allow-
formula that follows is applicable only in the special able moment, M, which will subject the critical fiber
case where both principal axes are symmetric axes. to its crippling stress.
This method of obtaining shear flow is included in For round tubes use the bending modulus of rupture
order to indicate how the classic formula SQ/I is modi- curves available in ANC-5 since they include local
fied by the nonlinear stress distribution. crippling effects.
(1) Determine mc/I.
(2) Enter the graph with mc/I and obtain an R Complex Bending
value for the proper ''section factor'' k> where R = This condition occurs when the resultant applied
df0/dfmy the rate of change of intercept stress fo with moment vector is not parallel to a principal axis.
respect to the maximum stress fm. Let X and Y represent two mutually perpendicular
(3) Substitute in the following: centroidal axes; let x and y represent the coordinate
 148 JOURNAL OF THE AERONAUTICAL S C I E N C E S — MAY, 1943

distances from the center of gravity, measured positive (B) Determine allowable moment My* as described
to the right and upward, respectively; and let X' and under simple bending, where
Y' represent the principal axes.
Qmy' = 0A25y/lyrA t (approximate formula)
Moment vectors are designated by double-headed
cyt = x cos <p + y sin <p
arrows and are to be interpreted by the left-hand rule
(i.e.; point left thumb in direction of vector and natural where x and y in this case are the coordinates relative
curl of fingers will designate the direction of moment). to the original X and Y axes of the fiber f artherest away
from the principal Y' axis.
mx = the applied moment about the X axis, the
moment being positive when it tends to Ky> = Z>\lrny> J \ly'/Cy>)
put compression in the ' 'upper" fibers
(C) Determine applied moments about the principal
mv — the applied moment about the Y axis, the
axes.
moment being positive when it tends to
put compression in the "left" fibers mX' — mx cos cp + my sin <p
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

<p = the angle the Xf principal axis makes with the mv> — —mx sin <p + mv cos cp
original X axis, the angle being positive
(D) Determine stress ratios.
when measured in a counterclockwise
direction. Rx, = mx>IMx> and Ry> = m,y>/My>

(E) Determine margin of safety.

M.S. = {l/[j(Rx> + Ry)}} - 1
where j is the proper additional factor of safety de-
pending on whether a limit or ultimate margin of
safety is involved.

Combined Bending and Axial Load

(A) In a manner similar to that described under
"Complex Bending'' determine the stress ratios.
Rx, = mx>/Mx> and Ry> =• my<]Mv>

(B) Determine Ra — fa/Fa, where fa is the applied

direct stress and Fa is the allowable direct stress.
F I G . 18. (C) Determine margin of safety.
M.S. = {l/[j(Rx, + Ry, + Ra))} - 1
Any case of complex bending may be resolved into
two cases of simple bending about the principal axes where j is the proper additional factor of safety depend-
of the section. The principal axes are defined as ing on whether a limit or ultimate margin of safety is
mutually perpendicular centroidal axes about which involved.
the moments of inertia are a maximum and minimum, (D) Secondary bending should be included when
respectively, and about which the product moment of beam columns are being investigated. In such cases
inertia is zero. it is necessary to determine only the ultimate margin
of safety.
Imax. = " =*= SeC 2<p
min. ^ ^
REFERENCES
where t a n 2<p = 2Ixy/{Iy — Ix)< 1
Timoshenko, S., and MacCullough, G. H., Elements of
(A) Determine allowable moment Mx» as described Strength of Materials, p. 226; D. Van Nostrand Company, Inc.,
New York, 1935.
under simple bending. 2
ANC-5, Strength of Aircraft Elements, Figs. 4-20, p. 4-35,
Qmx' = 0A2b\/lxAt (approximate formula) October, 1940.
3
Willett, N . L., and Bonza, L., The Bending Modulus of Cast
where At = total area, and Forged Sections, Lockheed Report No. 2772.
4
Beilschmidt, J. L., The Stresses Developed in Sections Sub-
Ox' = — x sin <p + y cos <p jected to Bunding Moment, Journal of Royal Aeronautical Society,
Vol. XLVI, No. 379, pp. 161-182, July, 1942.
where x and y are the coordinates relative to the original 6
Niles, A. S., and Newell, J. S., Airplane Structures, 2nd
X and Y axes of the fiber farther est away from the prin- Edition, Vol. 1, p. 164; John Wiley & Sons, Inc., New York,
cipal X' axis. 1938.
6
Shanley, F . R., and Ryder, E. I., Stress Ratios, Aviation,
**' = %Qmx'/(Ix>/cx') Vol. 35, No. 6, June, 1937.
 B E N D I N G S T R E N G T H IN T H E P L A S T I C RANGE 149

Tabular Form 77
TABULAR FORM 7 7 PLASTIC BENDING FACTORS. BASED ON _ MATERIAL PROPERTIES • AND SHAPE OF SECTION DATE .
9fc " (Ftp)fetp) "T 3 * (D| Z ®| "r 3 • (where. e t p • strain at proportional limit)
O I © l © l © | © | © | © | ® l @ l @ | @ | © | @
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

n_ L

APPENDIX—ANALYTIC DERIVATIONS OF FORMULAS Shear Flow Distribution

Moment of Resistance mc
y =U+Mk- 1)
m = 2 I zfy dy c dm = dfm , dfo(k _ ^
I dx dx dx
where z is width of section at point y (see Fig. 1 superimposed
on Tabular Form 77). But dfv
(3)
dx
/ = fo + (U - My/c)
since 5 = dm/dx, from the relation P = /«
fm(Q/c) -h folA — (Q/c) ]:
ff» = 2 1 I fozy + fm / o — ) dy
^P dfn Q
q (shear flow)
= MI/c)+fo[2Qm- (I/c)]
dx dx c +
or Let R = rate of change of trapezoidal intercept s t r e s s / 0 with
respect to maximum stress = dfo/dfm.
mc/I =fm+ Mk - 1) (1)
Therefore,
where k = 2Qm/(I/c).
Resisting
P
Load P over a Portion of the Section
= load on that portion of the cross section between the
ftfi+*-?)] (4)

extreme fiber and the fiber t h a t is at a distance y from But from Eq. (3):
the neutral axis dfm/dx = (Sc/I)/[1 + R(k - 1)]
Substitute in Eq. (4):
£ zfdy
(Sc/DUQ/c) + R[A - (Q/e)]\
1 + R(k - 1)
zy dy
= I /02 dy + (fm fo) Factor out Q/c

H)
5- ^( 26/ /11 ++ ^RKAc/Q) - 1\\
= M + <U^MQ_fmQ+f( (2) q
=
r (5)
c c M * '1 + 2?(ft - 1) /
^4 == area between the extreme fiber and the fiber that is at a Correction Factor to Reduce Test Data to Minimum Guaranteed
distance y from the neutral axis Values
Q — static (first) moment of area A about the neutral axis In Fig. 2 (superimposed on Tabular Form 77):
 150 J O U R N A L OF THE A E R O N A U T I C A L S C I E N C E S —MAY, 1943

/ = Ftv+ l(Ftu - Fty)/c]y /o = (6Qs/c2) - 2fm (8)

ff = Ftv + [(Ftu' ~ Fty')/c]y (Column 8 on Tabular Form 77)
The general expression for equivalent first moments about
I yfdy \y\ Ft„ + y~^~c—~ U? the neutral axis for a stress-strain distribution and a trapezoidal
Correction
factor distribution should include the variation of the cross-sectional
yfdy width z from the neutral axis to the extreme fiber. This may be
written as

Fty~ + {Fu

C
F,y' - + (Ftu> ~
Fty)~

Fty')~
-'f. yzf dy

When z is a constant as in the rectangular type of cross section,

0)

the above reduces to

= (Fty + 2Ftu)/(Ftu' + 2Ftuf) (6)

Approximation to Qm
m = 2z I yfdy = 2zQ, (10)
Jo
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

Let 2Qm = KpAt. K is a relationship factor, p is radius of which indicates that the resisting moment m is directly propor-
gyration, and A t is total area. tional to the stress-strain static moment Qs and, therefore, to
Type (a) rectangular cross section (Fig. 3 superimposed on its equivalent trapezoid as represented by OABE in Fig. 6 super-
Tabular Form 77): imposed on Tabular Form 77.
However, in the case of the diamond-type cross section where
712) h
- VI - V ^ pAt = bh2/SA6
bh
=

3.46
b is the width at the neutral axis:
z = b - y(l/c) (11)

2Qm = (2bh/2)(h/4) = bh*/4 yzfdy = 2 y[f - y{b/c)]fdy

Therefore, ./«
= 2bQs - (b/c)IB (12)
K = 2Qm/At = 0.865
since I8, the moment of inertia about the neutral axis for area
Type (b) diamond cross section (Fig. 4 superimposed on OABCD and a symmetric area below the axis, is
Tabular Form^77):

(2W/12)

pAt =
V (2bh/2)
bh*/2A5
= A/2.45 Is = 2\
J/ y*fdy

This expression for moment may be written in terms of a true

(13)

equivalent trapezoidal stress intercept designated as f0d and the

2Qm = {2bh/2){h/Z) = b¥/S
maximum fiber stress fm:
2
(bh /S)
K = = 0.817 m = -\fm + Mk - 1)1 (see Eq. 9)
(W/2.45)
c

Type (c) hour-glass cross section (Fig. 5 superimposed on Tabu- Since k — 2 for a diamond section (see Table 1):
lar Form 77): m = (I/c)[fm + M (14)
Substitute I/c for a diamond section = bc2/6 and equate to
P = J ^ m = V ^ 2 - Vl-414 . Eq. (12):
V
M (2bh/2) ' be2 N b
pAt = bh2/lAU m = — {fm + M = 2bQ~ ' ~ s - c- /' . (15)
6
2Qm = (2bh/2)(2h/S) = bh2/1.5 12QS 6IS
fod =
K =
(bh*/1.5)
0.943
~J~ ~ ~* ~ fm (16)
h2/lA14)
But Q8 = / 0 ( c 2 / 6 ) +fm(cyS) (exact value) and Is = (/ 0 c 3 /6) +
A general value of K — 0.85 has been assumed, which in most (fmCz/2) (approximate value). Therefore,
cases is conservative. 12/ c2 cA 6 / c3 c>\
/o. fm
Relative Values of Trapezoidal Stress Intercepts

In pig. 6 (superimposed on Tabular Form 77): = fo (approximately within 2 per cent) (17)
Let Qs designate the actual static moment about the N.A. In the case of an hour-glass section where b is the width at the
of the area OABCD. extreme fiber:
The equivalent trapezoid OABE has a trapezoidal stress inter- « = (f>/c)y (18)
cept QE = fo and a maximum stress AB = fm at t h e extreme fiber.
T h e equivalence of the trapezoid lies in the fact that f0 has
yzfdy - 2 I y[(b/c)y]fdy = (b/c)I, (19)
been determined so t h a t the static moment of the trapezoidal
area about the neutral axis is equivalent to Qs.
This may be written as This expression for moment may be written in terms of a true
equivalent trapezoidal stress intercept designated as foh and the
Qs = /„ c»/2 + (fm - fo) cyz (7) maximum fiber stress fm;
= (/ocV6) + (/»cV3)
m = - [fm+M(k - 1)] (see Eq. 9) (20)
from which c
 05 co

i U
O O E « iH O -P i t i i i i i i i i i i i i i i i i i I O
a> I -P <H w u u u u u u u u
,2, • H * H «H O O -H O O
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.11021

o5 v-i o3 03 0i
£-, f t . [x<
O o aJ o o
O rH OXi -P ^ W O > ^ O W « 5 H CO O C> O Q lO O M M O ^ •^ I D lO tO ft, f^ fe OH p^
• H «j t« a> d O O b- CO b- *H tO to CM CM lO tO %£ H ^ l O ( O l O H rr H
H t O CM r H
r H CM r H tO tO ID CM CO r H O C O b l O l O ^ N * O W W b
43 3 fl«.p © CM CC CM O to to ^ t-tO^OaiCMltDOOtOb-rH
H H O O O O O O O O O O O O O O O O O ^ O r-i O O O O i O O O O C M O r H
ert-P O 6
£
£ d <VH
pi O O «H O • ••• •••• • • « • • r-i• r~» ^
• • • • •
r H - H pH
si 3
o3 G
4> "8
rHiHiHiH i-trHiHiH rHiHiHiHiH
0 O W W O O S
O O Q O O O O O O O O O
O W O » ^ H O O O O ^ CO CO O O O O O O O CM CD O
0 1 O W H CM O lO QD CM l 5 O lO CDU5 C- O O lO to CM CM I D CM CO I D CO I D <# CM I D t O CM
I XJ O
0)
8fc»- *5M 00 O) W H
% * «» •. « «
H O C O O W ^ l W O O o i o> o> o> t o «o CO t - I D I D
O b- ^i ID
CO iH to r-i ^ to ^ b - b - b - r H 0>
(appr oxim atelj

CM o> b - b - ID I D ^H ^ CO CM
pezo idal

sec

t-t r-1 (J) W

3ns ar e wit hin

o '^ICM £ > O O Q O O O O O ) O O O O O Q
O CVJ < 0 CM OJ 3 * S
CM CO <
ID C- '
OH
£ )-i > CO _O 5
, __
CO - CM CO CO Q • O t O i H ^ i H CO H U ) O - rj« CO CO O O O O O O O CM CO O
O O O O
o o3 ^ oo o co co TJ< CM I D CO CM
<
OH
+ 03
*rt
a;
3
bfl
N M H H CO b - CO C - U> ID
CD 00 CO CO CM CM I D
-^ CM CO CD Tjt W ^ O O J N t - O ) b
tO CT> 0 0 b - I D I D ^
b b
TH
H O)
CO CM
- J CO U
3
Q< +-J
cu
A i.
4?irN OOOIO l O I O l O O O O O O O CM C\3 i O CsJ5 t O > 0
n •H U

H
+ r H r H r H <# ^ r H r H C M C M C M C M C M C M
05 O O O
O

•
O O •P
O > 0 > 0 0 > 0 > 0 > 0 0 > 0 > l"- t"- CO ^* <0 CO
• • •
O

•
Q

•
!£
ID O ID O
•<* i H ^ r H
Q> O) Ui G>
ID O O O lO
•<H r-i r-i r-i r-i
O 0> Oi Ct> CT>
-P O
O 4?
• ••• « • • • • • • • • O O O O O O Ctl « ••• O ? O
525 X» «S
m O O O O O O O O O O O O O O O ID ID
O O O O O O O* O O
< a » g S ID
— O b- t - O O O O O O O o o o o o ID I D I D I D O
ooo<Ro>^^o>o)(nojoo) ID ID ID O O O O
CD CD I D O O O «D CM CM CM CM O
PL, !* »COlOtDlDIDlOlOIDtOlOlO
o> o o> b- ID Oi ID O b- b- b - b - b - b - t - OJ
biO
ira ID ID 10 to w O O O O
W o3
S3' H H W H N N
t O t O CO CO
O O O O
O ID O ID O ID ID ID O ID ID ID ID ID
O ID O t-
W
ro O
S? t- t - I > t ~ t - I>-
b- O b- 05
O b - b - b - I D Oi CT> 0> OJ CO
b - 0> OS OJ CO CM CM CM CM O

t/3 CO fl)
U o
+J 01
co o o o o o w ^ o O I O I O H H ^ b - CVJ C M ^ ^ I D CM t O t O
O O t » t » N W l O O O W W t o W fc- t - i H r-i CM CVJ t O t D CO CO l> b- CM tO CM CM r H r H O l b - CO _ CO r H 0 0 I D ^«
C M C M t O ^ - s f l l O C M C O C O C M C M C M C M t O t O tO lO CVJ CVJ t O CO CO CO O O 0> tO r H t O O tO I D r H CO CM |M< b - Tj« r H r H
ID tD CM rH CO CO ^ ^ b - b - ^ CO ^ CO 0 0 0> CO
<N b~b-rHOOO>cOtD ID <sH «# O O O O CO CO « 0 CO
03 <"N r H 0 0 C*- CO I D I D - ^ t O CM CM
W cr
H »o Q Q o o o o o o
O O O O O O O O S S o O Q Q :88
O S
O O O O CO CO CM CM CVJ CVJ r H r H 3 * ^H
O O O O
O O O O § o o
ID O
O O O O O O
O
O O O O
O O O O
> b- b- O O rH ID O b- ^ O ^ * CO O CO

•8 a 3: l O ° 0 rH O
i 3« tO ^ tO
^ ^ I O t . t . t O I D Q O M > 0 0
cocococMcMcotoooo>o>b-b-
g oO t- b- o o
O* 0> b - fr-
i b- O O
» 04 b - b - O O I D CO O ID CO b - CO CM O t O ,0> b "<J« CM O
iDiotototocMtoiDiDtotocMCM IA US tO tO CM CVJ • t O CM CM O O t O o> O t o o> CO I D b - OJ CO rH tO CO b- rH
- |<o b- b - I D tO b- I D tO
H rH r H rH r H r H r H r H r H r-i r H
I D <tf< t O ^ CO ^ CO CM CO tO

to rH rH rH rH \rH r-i rH rH r-i

^Ji ^ p ( o I SS! O "* Q < O O O O O O O CO 00 CM O CO 00 rj« 0> I D T H
cocob=coco^cMiDi ID I D CM CM I D CM CD < b - b - rH r-i O b- r H CM CO CO 5ji I D b - CO
CM CM CM CM r H i H CM CM r H i CO CO t O CO CO CO CM O CO 00 ^ CM 0> 0 0
^ o r H r H r-i O rH rH rH rH O IrH rH rH O O
cti

£ ooooooooooooo • • • • • • O O O O O O O O
• • • • •
0) O O O O O
>—1

Q CO
CO
+ M | 0
O O ^ O O I O H N ^ M t O Q Q
O O H W l Q < O N © « 0
OOCOtONNHO><O<O(O(0NN
0 0 © « 5
r-lr-l»-lf-tiHcHOrir-Jr-liHr-lrH
C- t - tO tO Q Q
( O O CO CO CM CM
o ooooo S CO
O
> CM CM
CM CM * * O
CO CO I D CO
O O CO I D
CM -tf O
CO I D 0 0
OJ CO I D
CO O ^ I D I D CO CM b- rH O
b~ I D 0 0 0 0 O
r H OJ 0 0 CO CO OJ 00 CO rH CO
• • • • • • • • • • • • •
gla

O O O O O O O rH O O O O CM O Gt b- CO
J2J
o ooooo • • •• •
W V-)
o o o o o o o o o o o o o CM O O O O O H rl O O O
o ^ ^ ^ N C M W W H H H H H H CM r H CM r H r H

W .
3
o
•OH q.©TIT9
CM T * CM I D
• ••• •
CM I D I D I D t o CM CM CM CM ^
.a O O O O O
A *vi cq ***">
a C - CO CO O r H CM
•on necrjoeds

M C a t Q ^ ^ U X O H H ^ M t O I O r H r H CM CM CO t O r H CM t O t O
CM CO ^ I D CO CM CO - ^ I D CD
I I t I I I I I I !
A i . i • i i A i • i i • I I I I I I I I I
co
n* o
to to
II
3 ^ Shrr.2j0£ 20*3.8*80
X«TJ9q.^R 2XTCSJO^ igfri ranxeeuS-GR 3u-jSaoj SUJSJO^
91-961
9i-S6I 9I-96T
o no-jqx>eg SKOIiLOSS-I no-jq.o©s pnnoH uo^q-oos xamx^qo