Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Design of RC Members Under Flexure

Download as pdf or txt
Download as pdf or txt
You are on page 1of 16

Vol.

33 PROCEEDINGS OF THE AMERICAN CONCRETE INSTITUTE

JOURNAL
of the
AMERICAN CONCRETE
INSTITUTE
7400 SECOND BOULEVARD, DETROIT, MICHIGAN . . . . . MARCH-APRIL 1937

Design of Reinforced Concrete Members Under Flexure or


Combined Flexure and Direct Compression*

BY CHARLES s. WHITNEYt
MEMBER AMERICAN CONCRETE INSTITUTE

It is the purpose of this paper to suggest a complete revision of the


method of designing reinforced concrete members subjected to bend-
ing and to present a rational method for the proportioning of arch
ribs, rectangular columns under eccentric load, and rectangular
beams. This method may be extended to cover the design of T
beams, round columns and any other form of concrete members.
Simple formulas are given which predict the ultimate strength
with remarkable accuracy based on the cylinder strength of the
concrete and the yield point of the steel independent of the ratio
of their moduli of elasticity.
It is believed that these equations will greatly simplify the work
of the designer and result in the more efficient and economical use
of reinforced concrete. The suggestions are intended to place struc-
tural concrete design on a solid practical foundation which is now
justified by the advance in construction practices, the availability
of better materials, and the information gained by a great amount
of research work.
The present method of designing members under bending and
combined bending and direct stress is unsatisfactory for several
reasons.
The usual formulas based on assumption of a cracked section and
straight line variation of stress are far from correct both under work-
ing loads when the concrete is not materially cracked and under
ultimate loads when the stress in the concrete is not even approxi-
mately proportional to the distance from the neutral axis. They
*Received by the Institute Feb. 11, 1937.
tConsulting Engineer, Milwaukee.
(483)
484 JoURNAL OF THE AMERICAN CoNCRETE INSTITUTE Mar.-Apr. 1937

do not give a good indication of the conditions under working loads


and cannot be used to predict the ultimate strength with any degree
of accuracy.
They do not properly recognize the strength of the concrete in
compression, and provide a much higher factor of safety against
failure of the concrete than of the steel. <2><6> That may have been
justified during the early days of reinforced concrete on the grounds
of unreliability but it is certainly not justified today. Properly
balanced design will permit the use of a considerably higher percent-
age of tensile steel with a consequent reduction in the size and weight
of members.
The usual flexure formulas are complicated by the use of the value
of n which is quite unpredictable under high loads and actually has
little effect on the ultimate strength of the beam. <2>< 9>< 10>
The effective value of n is widely different for dead and live loads.
In the case of arch ribs, the dead and live load stresses cannot be
computed separately with different values of n and added together,
because dead load may produce compression only, while the live
load moment alone would require assumption of a cracked section.
The calculation in this case would be very complicated with two
values of n.
The present method makes no prediction of the loading which
will cause cracking and does not give accurate control over the factors
of safety under dead loads and live loads.
The method here proposed for the design of members under bend-
ing and direct stress, particularly for arches, is direct and simple.
First, an analysis of tensile stress in the arch rib should be made
assuming the critical combination of live and dead loads, together
with temperature, shrinkage and plastic flow effects. This can best
be done by using the complete transformed section without cracks.
The effective values of n for dead and live loads can be predicted
with sufficient accuracy for this purpose at allowable working stresses
and the dead and live load stresses for corresponding value8 of n
can be computed separately and added together. The tensile stress
in the concrete so determined should not exceed the modulus of
rupture of the concrete in an unreinforced beam in order that the
arch rib shall not be cracked in service. The compressive stress for
this condition is relatively unimportant because the strength of the
rib will be controlled by an ultimate strength calculation.
In arch ribs whose center lines follow the dead load pressure lines,
the tensile stress can easily be kept somewhat below the modulus
.of rupture. In the rigid frame type of arch, it may be possible to
Reinforced Concrete Members Under Flexure 485

permit the tensile stress for the extreme case to slightly exceed the
28-day modulus of rupture without introducing objectionable crack-
ing; or if the critical combination of loads includes only a portion of
the full live load, the calculated tensile stress can probably be kept
below the modulus of rupture without difficulty. This method of
design would have the advantage of directing attention to the tensile
stresses and would encourage the use of concrete with higher tensile
strength.
This calculation using the full strength of the uncracked section
might re&ult in the use of too little reinforcing steel and provide an
inadequate factor of safety unless an additional calculation is made
of the ultimate strength of the rib. Because of the large strains
occurring before failure, the effects of temperature change, shrink-
age, and plastic flow have no practical effect on the ultimate
strength. (t)(a)( 4l It is therefore proposed that these effects be neg-
lected and that the ultimate strength of the member be computed
by a new type of formula based on the ultimate strength of the con-
crete, and on the yield point strength of the steel.
It has been pointed out recently by several investigators< 4l<~l(to)
that while the stress variation in the concrete is approximately linear
under very light loads, and parabolic under intermediate loads, as
the ultimate load is approached it assumes a shape about as shown
in Fig. 1.
The stress increases very rapidly near the neutral axis and is nearly
uniform for the greater part of the depth of the compression section,
probably decreasing slightly toward the edge of the beam. Saliger< 10>
reports the ultimate strain in the concrete at the outer edge of the
beam to be from .003 to .007 while the limit reached by concrete
prisms at failure was .002 to .004. The usual "parabolic" formulas
have been based on the theory that the stress variation in the beam

/
/
I
II
I· Mo.r. Strain
.OOZ-.004-
·I
prjsm
FIG. 1-STRESS-STRAIN CURVE FOR PRISM AND FOR BEAM AT ULTIMATE
LOAD
486 JOURNAL OF THE AMERICAN CONCRETE INSTITUTE Mar.-Apr. 1937

followed the shape of the first part of the stress-strain curve for
the concrete cylinder up to the point of maximum load.
This is evidently considerably in error because of the greater ultimate
strains and the different behavior of the concrete in the beam. On
this account, the writer questions the value of any refinements of
beam theory based on an attempt to estimate the ultimate stresses
in a beam by comparison with the stress-strain curve for the cylinder
or other standards.
It is therefore proposed that a rectangular block of uniform stress
as indicated by the dotted lines on Fig. 1 be used to represent what-
ever stress may exist in the concrete. Whatever it actually is, it
must have an average intensity, fc, and an effective depth, a. The
resultant is assumed at the middle of the rectangle. Under ultimate
load, Hooke's Law and the theory of elasticity have no significance as
far as the internal stresses are concerned. The materials are more
nearly in a plastic state but the imperfect and variable action of
concrete makes a rigid solution according to the theory of plasticity
impractical. No further theoretical justification of the assumption
a rectangular compressive stress block is necessary if the formulas
derived therefrom accurately predict the ultimate strength of the
member. This they appear to do.
This assumption of uniform compressive stress on the concrete
in a beam has been suggested by Gebauer and by Copee but their
formulas do not appear entirely satisfactory because of their other
assumptions. Since the development of the following treatment
by the writer, von Empergm·<9l and Saliger(JO); in excellent discussions
of this subject have recommended the use of simplified ultimate
strength formulas for beams independent of n. Their formulas are
somewhat less simple and do not appear to check the results of Amer-
ican beam tests quite as well as those presented herein.
SIMPLE FLEXURE
The assumed relations for a rectangular beam under simple flexure
are shown in Fig. 2.
It is assumed that in an under-reinforced beam, that is, one which
will fail in the tensile steel, the concrete will crack as the steel
stretches and the depth of the beam in compression a will be reduced
until the concrete unit stress reaches the ultimate or
a = A,f, ........................................................ Eq. (!)
bf,
in which, A, = area of tensile steel
f, = yield point stress in steel
b = width of beam
f, = ultimate strength of concrete
Reinforced Concrete Members Under Flexure 487

Fro. 2
This determines the lever arm of the steel reinforcement since
a
c = d--
2
It will be assumed that the ultimate compressive strength of the
concrete in the beam is equal to 85 per cent of the cylinder strength
in order to be consistent with the results of tests on concentrically
loaded columns and to have a treatment which can be applied to the
full range of cases from simple flexure to direct load.
The values of a and c are derived as follows for any particular
bending moment:
M = ( d- ; ) abf, . ..........................................••.. Eq. (2)

from which
a = d- / d' - 2M or _!!___ = l - /1 - ZM · · · .. · .............. Eq. (3)
'\} bf, ' ' d \ bd'f,
and

c = d- + +( + ~
= d d'- 2: . ) .............................. Eq. (4)

These expressions are independent of the area of steel and the

value of E • .
Ec
The required steel area is simply,
A,= - M ......................................................... Eq. (5)
c.f,
Since the assumed compressive stress distribution has no exact
theoretical basis, the limiting value of the depth of compression a
for equal concrete and steel strengths in flexure must be determined
experimentally. If the beam has at least sufficient steel to fully
i
.....
0
TABLE 1-BEAM TESTS REPORTED BY SLATER AND LYSE
~
Actual Maximum Moment Divided by ~
!:"
Cylinder a Maximum Moment as Given by
Group d b p strength Maximum Value of Maximum 0
No. (in.) (in.)
, ,
(lbs. per
sq. in.)
Moment
(in lbs.)
a
from Eq (3)
d d Mument
from Eq (6)
Equation
(6)
A. C. I.
Formula
Parabolic
Formula
..,
"'J

ti:
i:'l
1 10.2 8.2 1 .021 1 139o 597,ooo >d > 1.o* 396,000 1.508* 2.06* 1.50*
2 10.3 8.2 I .028 I 2790 818,000 5.6 0.544 .728
----
811,000 1.007 1.51 1.03 :.::>
i:'l
:;)
3 1o.3 ~.2 .o37 1 4o7o 1,152,ooo 5.3 o.515 .742 1,18.1,000 0.973 1.50 0.98 8
~
>
4 10.1 8.2 4800 1,360,000 5.6 0.554 .723 1,342,000 1.013 1.54 1.00 z
5 10.2 8.3 .056 1 5740 1,584,ooo 4.9 0.481 . 759 1.660.000 0.955 1.46 0.95 0
0
z
6 14.2 -s.2j~ 2590 1,8o5,ooo 12.9 o.910* .545* I 1.430.000 I 1.252• 1.85* 1.27* n
~
6A 14.1 8.2 .039 4130 2,196,000 7.3 0.518 .741 l
2.250.000
1 -- 1
0.976 1.50 0.97
i:'l
~
7 12.2 8.3 .028 2950 1,240,000 6. 75 0. 553 . 723 1,220,000 1.015 1.56 1.03 .....
---- z
8 8.0 8.1 .031 2760 1 543,ooo 5.36 0.670* ~- 478,000 1.135* 1.66* 1.13*
~
I I
~
9 5.9 7.9 .032 2900 310,000 4.17 0.708* .646* 266,000 1.166* 1.69* 1.14*
10 4.1 8.0 ~030 2820 129,000 2.26 0.552 .724 127,000 1.016 1.48 1.00
-----
lOA 4.1 8.0 I .040 3810 I 179,000 2.37 II 0.576 • 712 171,000 1.046 1.55 1.03

Average .537 = .024 . 732= .od I 1.00= .02411.51= .028/ 1.00= .024
~
*These values not included in average.
~
~:"
......
~
-1
Reinforced Concrete Members Under Flexure 489

develop the strength of the concrete, additional steel does not mate-
rially increase the strength of the beam.
The limiting value of a as computed from tests reported by Slater
and Lyse< 2l is given in Table 1. The value off, assumed in Eq.(3) is
85 per cent of the corresponding cylinder strength, of, f, = 0.85 fc'.
Eliminating four groups of tests which appear to be erratic, the
average value of__!!__ is 0.537. This is independent of p, d, and f,.
d
The flexural strength of a fully reinforced rectangular beam with
tensile steel only is then, from Eq. (2).

M = (d- 0.2685 d) 0.537 d b j,


= 0.393 b d 2 f,
M
or bd• = 0.393 j, = +
!' ··················........................... Eq. (6)

in which f, is 85 per cent of the compressive strength of standard


cylinders, j/. Column 10 of Table 1 gives the maximum moment
for the test beams as predicted by Eq.(6). It may be noted that the
percentage of error is practically the same for Eq.(6) as for the para-
bolic formula given by Slater and Lyse using the measured value
of n.
M
Fig. 3 shows the value of given by the 36 beams tested
bd 2j/
by Slater and Lyse and by 33 beams tested by Humphrey<11 l which
had cylinder strengths between 1000 and 2500 p.s.i. The Humphrey
beams with higher strength concrete were not included because it
appears that there was not sufficient reinforcement to fully develop
such higher concrete strength even though they may have been re-
ported as compression failures. Some confusion no doubt exists
because it is difficult to differentiate between primary steel or con-
crete failure unless the amount of steel is considerably below the
critical percentage required to develop the full concrete strength. A
slight stretching of the steel may reduce the concrete compression
area and cause what looks like a concrete failure in an under-rein-
forced beam.
The value of ~' shown by Fig. 3 is unaffected by the concrete
bd j,
2

strength between 3000 and 6000 p.s.i. although it rises about 50 per
cent for weaker concretes from 3000 to 1000 p.s.i. With the present
methods and materials, it is not probable that much structural con-
490 JOURNAL OF THE AMERICAN CONCRETE INSTITUTE Mar.-Apr. 1937

0.7
0 l It ...1 Gt;.o! Su ve: Tttsfs
0
(.0 oc. A.c . I. I~Z p 25- rz7;
0.6 )(

><,1 Ia I er one L< se Te Is


0
(.A ur1. o/ ~.c.I. '.930 pt. 50)
as "CO• 0
0 )(

0
~0 )(
0
a4 "coo A

x"><
M 'Oo ~c:P )(
"" )(

~ ):c )(
)(
03 , )(
)(

02

0./

0 /000 ZOOO .3000 ;/000 .5'000 6tXYO


CYlinder Sfrenqfh-0bs. pe:rsq. in.)

FIG. 3- VALUE OF b~~e' FROM TESTS OF BEAMS

crete will be used with a strength of less than 2500 or 3000 p.s.i.,
and if it is used, a greater factor of safety is not inappropriate. There-
M 1 can be used generally.
fore it appears that a value of 0.333 for
bd 2/c
For beams with less steel than required to develop the full com-
pressive strength the allowable bending moment is given by the
formula:
,; = pf, ( 1 - 1.;;:, ) ············. ·· .·..... ·... ·................ Eq. (7)
The critical percentage of steel required to develop the full com-
pressive strength of the concrete is:
j/
p. = 0.456 - - · ................................. . .................. Eq. (8)
f,
Table 2 shows the results of the application of Eq.(5) to 72 beams
of the series tested by Humphrey<•> except beams 489-490-491 where
Eq.(6) controls on account of the low strength concrete. The average
:..

Reinforced Concrete Members Under Flexure 491

ratio of actual maximum moment to that given by the formulas is


1.071 to 1, and, eliminating the low strength cinder concrete beams,
the ratio is 106.24 to 1. Classified according to age, the ratio is 1.052
to 1 at 4 weeks and 1.09 to 1 at 52 weeks. According to percentage
of steel, it is 111.7 for p = 0.0049, 105.44 for p = 0.0098, and 104.02
for p = 0.0196. There appears to be no correlation between modulus
of elasticity of the concrete and the accuracy of the formula.
The value of the steel lever arm, _c_, from Eq.(4) is given in Table 2
d
for comparison with the standard value, j. The value given by Eq.(4)
is greater and appears to be more satisfactory than j by the usual
method.
Equations (5) (6) and (7) appear to check the results of well con-
trolled tests within the limits of the variability of the materials.
They appear to be as accurate as any treatment which can be de-
vised from information available although more complete tests may
make possible a slight improvement in the empirical constant.
The effect of these formulas is shown in Fig. 4 based on 3000 p.s.i.
concrete cylinder strength and 50,000 p.s.i. steel yield point with a
factor of safety of 27;:l for both concrete and steel.. The allowable
value of M is plotted against the percentage of steel and compared
bd2
with the value by the usual formula with fc = 1200 p.s.i., j. = 20,000
p.s.i. and n = 10. The new formulas raise the critical percentage
of steel from 0.0113 to 0.0273 and the allowable M from 197.5 to
bd2
400. For under-reinforced beams, they give a higher value of the
steel lever arm and a closer approximation to the true steel stresses.
It may be noted that the same method can be applied to T beams
if it is determined from tests what the limiting effective proportions
should be. The writer has not as yet made any examination of test
data but there are probably enough on record to indicate how widely
the uniform compressive stress in the concrete can be assumed to
be distributed.
BEAMS REINFORCED FOR COMPRESSION
If there is steel in the compression side of the beam as in the case
of a symmetrically reinforced arch rib, the compression steel will
be stressed up to its elastic limit if the beam fails in compression.
If it fails in the tensile steel, the presence of the compressive steel will
have comparatively little effect on its ultimate strength. In the latter
~
""
TABLE 2----cOMPARISON OF CALCULATED AND ACTUAL ULTIMATE STRENGTH OF BEAMS
TESTED BY HUMPHREY AND LOSSE '-<
0

----
Concrete
-- - - - -
Steel
Max. Moment c
Max. Moment Maximum
Calculated MDment
~
>
Beam Numbers Cylinder Yield Actual - FromEq (5) Actual t<
Aggregate Age Strength A, d Point (in.lbs.) d (in.lbs.) Divided by 0
(lbs. per (sq. in.) (in.) p (lbs. per Calculated "l
weeks sq. in.) sq. in.)
~
162-163-164 Granite 4 3241 .393 10 .49 42,490 177,000 .9621 160,500 1.103 1.':1
186-187-188 4 2807 .785 10 .98 43,043 321,000
~
.9115 307,500 1.043
237-238-239 4 3158 1.571 9)4 1.96 41,857 492,000 .8350 507,000 .970
171-172-173 52 5589 .393 10 .49 42,573 193,600 .978 163,500 1.183 1.':1
195-196-197 52
52
5086
4957
. 785
1.571
10
9)4
.98
1.96
42,320
41,877
340,000 .952 319,000 1.066 ::l
234-235--236 543,000 .895 545,000 .996 &:
255--256-257 Limestone 4 2087 .393 10 .49 41,653 164,000 .9425 154,000 1.064
z
273--274-275 4 2600 .785 10 .98 40,797 291,000 .9095 289,000 1.006 0
321-322-323 4 2725 1.571 9)4 1.96 38,037 471,000 .815 451,000 1.044 0
246-247-248
282-283--284
52
52
4069
4360
.393
. 785
10
10
.49
.98
41,003
41,413
183,000
317,000
.9708
.9451
157,000
307,000
1.165 ~
1.033
33Q--331-332 52 4230 1.571 9)4 1.96 37,673 532,000 .890 487,000 1.092 ~
~
333-334-335 Gravel 4 3441 .393 10 .49 43,447 169,000 .9635 164,000 1.030 ,_.
357-358-359 4 3440 . 785 10 .98 42,600 321,000 .9285 310,000 1.035 ztil
405--406-407 4 3376 1.571 9)4 1.96 38,633 504,000 .857 482,000 1.045

~
342-343-344 52 5186 .393 10 .49 44,140 173,000 .9754 169,000 1.022
366-36 7-368 52 5242 . 785 10 .98 40,107 356,000 .9559 300,000 1.186
414-415-416 52 5467 1.571 9)4 1.96 38,643 533,000 .9120 512,000 1. 0-!0
417-418-419 Cinders 4 1629 .393 10 .49 37,357 163,200 .9337 138,000 1.182
441-442-443 4 1555 . 785 10 . 98 41,170 281,000 .8475 273,000 1.030
489-490-491 4 1643 1.571 9)4 1.96 37,887 403,000 . 7315 376,000* 1.071
426-427-428 52 2944 .393 10 .49 38,343 172,600 .9624 145,000 1.188
450-451-452 52 2619 .785 10 .98 41,103 306,000 .9090 293,000 1.044
498-499-500 52 2763 1.571 9)4 1.96 38,047 488,000 .827 458,000 1.064 s::
1. 071 = . 05
~
*Determined by Concrete Strength, Eq. (6)
Average
>
~
......
~
-1
Reinforced Concrete Members Under Flexure 493

p 0/) se II or mu a
400 \ .h_ ·1 00
)

- -
/
/
w
' ,...o v .- ~~
(]0
b~ k"
,.;l' \...._
.PJ" ~se '(21 A ~I ;:; /"17. u!a

J00
I
J
7
0 I 2 .3 4 5
Percenlo9e of Slee/
FIG.4-COMPARISON OF RESISTING MOMENT OF RECTANGULAR BEAMS
AS GIVEN BY PROPOSED FORMULA AND PRESENT A. C. l. FORMULA

case, the compressive steel can still be figl!red at its elastic limit if it
comes within the compression zone and the lever arm of the tensile
steel can be computed on that basis.
With compressive steel, the ultimate strength in compression is
computed by adding the moment of the steel compressive stress to
that of the concrete stress computed before. This condition is shown
in Fig. 5.

FIG. 5

The ultimate compressive moment will be


M, = 0.333 bd2fc' + d' A',f, .... ....................................... Eq. (9)
The ultimate tensile moment will be
M, = A.j,
cabf, + d'A',f, = X A,f, ................................. Eq. (10)
abf, + A',f,
494 JoURNAL OF THE AMERICAN CoNCRETE INSTITUTE Mar.-Apr. 1937

in which the value of c is obtained from Equation (4) using


(M - d' A' .j.) in place of M, and x is the distance from the center of
tensile steel to the center of gravity of the compressive stresses in the
beam. Also note that fc = 0.85f'co Equations (9) and (10) have also
been checked against tests with satisfactory results provided the
failure is not through shear or bond.
The results of tests of beams with compressive steel made by
Bach and Graf at Stuttgart and reported by von Emperger <7• page 225>
are of interest because they show the danger of bond failure. The
full yield point stress of 35000 p.s.i. was developed in 20 mm. round
bars but the beams with bars of steel with a 60,000 lb. yield point
failed in bond before the full strength was developed. Glanville< 12>
advises against counting on the compression reinforcement but his
tests were made on very small beams which evidently failed in shear
before the full moment was developed. He says, "In the simple
beams, serious cracking developed towards the end of the tests, and
it is possible that the higher shear stresses in the continuous beams
with compression reinforcement may have been the reason for the
low moment carried over the central support."
FLEXURE AND DIRECT LOAD

The case of bending and direct stress can be treated in the same
manner as bending alone. The strength of the compressive side will
be the same as before and the steel tension will be reduced by the
amount of the direct compression, P, as shown in Fig. 6.

FIG. 6

The ultimate compressive moment is given by Eq. (9)


M, = Pe = 0.333bd 2f', + d'A',J,
Reinforced Concrete Members Under Flexure 495

The ultimate tensile moment is


_ x _ .......... ................................... Eq. (11}
M, = Pe = A.j, x ·
1--

When there is no steel on the compressive side Equation (11) becomes

_ c _ ............................................ Eq. (12)


M, = Pe = A.j, c
1-_
e
and the value of c can be obtained from the formula

c= +[d + e- ~ (d + e)•- 4de + 2Ye]. ........................ Eq. (13}


in which
y = AJ, = AJ.
bfc 0.85 bj' c

When there is steel on the compression side but the load is not
sufficient to develop the full compressive strength of both the steel
and concrete, the required tensile steel area can best be computed
by considering the effect of the compression steel and concrete sepa-
rately. First compute the area of tensile steel required to develop
the moment (d' A'.!.) of the compressive stress with the formula:

A, = A', ( 1 - : ) .............................................. Eq. (14}

Then deduct the moment (d' A',f.) from the external moment, Pe,
and compute the additional area of tensile steel required to balance
the compression stress in the concrete using Eq. (12).
When the eccentricity is small compared with d, the ultimate
strength of the member can be computed as twice the strength of the
weaker side of the section from the formula (see Fig. 7):

t d

FIG. 7
496 JOURNAL OF THE AMERICAN CoNCRETE INSTITUTE Mar.-Apr. 1937

P = 2A',f, + 1.7b (d-e) f'c ................ ........................ Eq. (15)


which becomes for a concentric load on a symmetrical member:
P = 2A',f, + 0.85btf'•............................................... Eq. (16)
The value of e for which the Equation (9) and (15) will give equal
values of P is given by Equation (17) and for smaller eccentricities,
Equation (15) should be used.

e = ~[d + Y' + ~ (d + Y')2-0.786d 2 -2Y'd'] ................. Eq. (17)

in which
A',f, .
0.85bf'c

The formulas for ultimate strength of members under flexure and


direct load have been verified by comparison with the results of tests
on 46 columns made by Bach and Graf <7><8>. The ratio of actual
ultimate load to calculated ultimate strength is given in Table 3
which shows an average error of three per cent for the series. The
columns were about 8 ft. 3 in. long and 16 in. square and were loaded
with eccentricities varying from zero to about 20 in. The average
30 em. cube strength of the concrete was 225 kg. per sq. em. and

TABLE 3-COMPARISON OF CALCULATED AND ACTUAL ULTIMATE LOADS FOR COLUMNS


TESTED UNDER ECCENTRIC LOADS BY BACH AND GRAF
(See JouRNAL Amer. Concrete Inst. Proceedings Vol. 26, p. 661. April, 1930.)

Calculated Ultimate Load Actual Actual Load


Column Eccentricity Ultimate Divided by
Formula No. Load-Kg. Load-Kg. Calculated Load
Type I 0 16 271,000 280,333 1.035
.50 12 94,000 93,000 .990
.75 12 57,000 60,333 1.057
1.25 12 29,000 29,967 1.016
Type II 0 16 331,600 338,333 1.021
.25 15 196,100 202,500 1.032
.50 9 128,000 124,000 .968
.75 12 & 14 69,800 69,600 .998
1.25 12 & 14 32,800 32,350 .986
Type III 0 16 379,100 404,700 1.067
.25 15 243,800 225,000 .923*
.50 9 150,000 157,500 1.050
.75 12 & 14 106,500 105,000 .987
1.25 12 & 14 56,000 53,500 .955
Plain Columns 0 16 270,400 276,167 1.020
.25 15 135,200 136,000 1.005

Average 1. 007
Average Error • 031
*Average of two tests only, 0.882 and 0.965.
Reinforced Concrete Members Under Flexure 497

the cylinder strength was taken as 225 or 199 kg. per sq. em. The
1.13

yield point of the steel averaged 3773 kg. per sq. em. for Types I and
II and 3572 for Type III.
The maximum variation in the strength of the cubes was about
±9Y2 per cent. The errors shown by the ultimate strength formulas
are well within the limits of the variability of the materials and it
would appear that the use of more elaborate formulas is not war-
ranted.
For small eccentricities, Equation 15 agrees with test results much
better than the usual formula based on straight line stress variation.
Such agreement is more important than theoretical verification but
it can be supported on the basis of actual stress variation. It is
further supported by the results of tests on two plain concrete prisms
reported by Slater and Lyse <2 • page 859 >. The 8 x 8 x 12 in. prisms

were loaded with e = _t and the ultimate load averaged 155,520


6

lbs. The cylinder strength of the concrete was 4060 p.s.i. Eq. (15)
gives
8
P = 1.7 X 8 X S X 4060 = 147,000 lbs.

or about 94.6 per cent of the actual load.


The foregoing equations predict the untimate strength of a mem-
ber but cannot be used to compute the actual stresses below the
ultimate. Being based on 85 per cent of the cylinder strength, they
are consistent with the results of tests of columns under concentric
load. It is hoped that they will provide a satisfactory method which
is now lacking for the design of members of elastic frames and arches
after the external moments have been determined by the theory of
elasticity. They indicate that a higher percentage of tensile steel
may be used than is now general practice without overstressing the
concrete and they will permit lighter members where bond and shear
can be properly taken care of.
It is further proposed that the dead load moments and thrusts be
multiplied by a suitable factor of safety and added to the live load
moments and thrusts multiplied by an appropriate factor, probably
larger than that used for dead load, equating the total to the ultimate
.strength given by the equations. The factors of safety could also be
498 JOURNAL OF THE AMERICAN CONCRETE INSTITUTE Mar.-Apr. 1937

introduced by using reduced values of the stresses if it proves desir-


able. It may be that the ultimate strength formulas alone will be
sufficient for ordinary building design without the necessity for
making any calculation of the tensile stresses under working loads.
A thorough study should be made to determine what factors of safety
should be used for different types of structures.
REFERENCES
(')"Final Report of the Special Committee on Concrete and Reinforced Concrete
Arches," Transactions, Am. Soc. of Civil Engineers, 1935, p. 1427.
( 2)"Compressive Strength of Concrete in Flexure," Slater and Lyse, JouRNAL
Am. Concrete Inst., June 1930, Proceedings, Vol. 26, p. 831.
( 3)"Reinforced Concrete Column Investigation, Tentative Final Report of Com-
mittee 105," F. E. Richart, Chairman, JouRNAL Amer. Concrete Inst., Feb. 1933,
Proceedings, Vol. 29, p. 275.
('')"Considerations sur le Clacul et la Securite des Pieces Flechies," by R. Copee,
Publications of the International Association for Bridge and Structural Engineering,
Vol. 3, p. 19.
(')"Strength of Concrete Beams," by Richard L. Humphrey and Louis H. Losse,
Technologir Papers of the Bureau of Standards.
(•)"Flexural Resistance of Shallow Concrete Beams," by C. B. McCullough,
Engineering New-Record, Vol. 115, p. 406.
( 7 )"Handbuch fur Eisenbetonbau," von Emperger, Vol. 1, 3rd edition, p. 240.

( 8)Discussion by George H. Dell, JouRNAL Amer. Concrete Inst., Apr. 1930,


Proceedings, Vol. 26, p. 661.
( 9 )"Der Beiwert n," by Fritz v. Emperger, Beton und Eisen, Oct. 1936, p. 324.

( 10 )"Bruchzustand und Sicherheit im Eisenbetonbalken," by Rudolf Saliger,


Beton und Eisen, Oct. 1936, p. 317.
(")"Compressive Strength of Concrete in Flexure," by Slater and Zipprodt,
Proceedings, Amer. Concrete Inst., 1920, p. 120.
(1 2)"The Redistribution of Moments in Reinforced Concrete Beams and Frames,"
by W. H. Glanville and F. G. Thomas, JoURNAL, Institution of Civil Engineers,
.Tune 1936, p. 291.

Discussion of the foregoing paper will be welcome if received in tri-


plicate by the Secretary of the Institute by June 1, 1937. For
such discussion .as may develop readers are referred to a Supplement
to be issued with the JouRNAL for Sept.-Oct., 1937.

You might also like