UNIVERSITY OF LONDON: IMPERIAL COLLEGE

A plastic theory of design for ordinary reinforced

and prestressed concrete including moment
re-distribution in continuous members
by Professor A. L. L. Baker, B.Sc. (Tech.), M.I.C.E., M.I.Struct.E.

SUMMARY: when (b) the steel reaches a stress at which a

A general expression is derived for the ultimate small number of further repetitions of load
strength in bending of ordinary reinforced concrete would cause increased yield and fracture.
and two types of prestressed concrete beams, taking
NOT A TION. (See Figure 1.)
into account plasticity. A method of calculating
b breadth of beam.
elasto-plastic deformations is developed and applied
d = effective depth of beam.
to the problem of re-distributing bending moments
n1d depth of neutral axis.
in continuous members and frames.
yn1d depth of centre of compression.
Introduction p. percentage of steel.
c concrete stress at the extremity of the
SUFFICIENT IS ALREADY known of the stress section.
strain relations of concrete in beams to justify tensile stress in the reinforcement.
presenting a tentative theory of ultimate strength, ocn1dc area of concrete stress diagram.
deformation and moment re-distribution in regard AB
to beams and continuous frames. Further research E, modulus of plasticity of steel =OB
is needed to determine more closely the values of
some of the factors involved and this may lead to (see Figure 2.)
some modifications. The present theory, however, = AB (see Figure 2) = modulus of plas-
may have some value in guiding further research. OB ticity of concrete
For simplicity rectangular sections have been at the extremity of
investigated, but using the same assumptions the the cross section.
theory can be extended to T and I sections.
E.
Ultimate strength of ordinary reinforced and m
Ee
prestressed beams in bending M moment of resistance or applied bending
ASSUMPTIONS moment.
(I) Strain is proportional to distance from the Steel stress at loading condition. Fig. 1 (c)
neutral axis, in accordance with a straight line f3 ultimate stress
law. (Test results show that this is very (in prestressed beams).
approximately true.) Average concrete stress adjacent to steel
F
(2) Prestressed beams crack before failure and Maximum concrete stress adjacent to
concrete has no tensile strength. steel
(3) A beam is loaded to its ultimate strength: when (in end anchored prestressed beams)
(a) the concrete begins to crush, either (i) without
the steel yielding or (ii) on account of the steel NOTES ON THE ABOVE FACTORS
yielding and causing the neutral axis to rise or The value of y may be determined from beam

57
 Magazine of Concrete Research: June 1949

f. _~~t-'~

.!!------JC_-"-A _ _ _ _ _
1Jnrd
I
_
~ --
_
I
I

H.T.Stul

...
a. Strain Dil\ribution. b. Stre .. Distribution. Mild
ORDINARY REINFORCED ~:"---:-?i---IcStce I
CONCRETE.
L

Jj

Strain

c. Suen Diltributiol't.

PRESTREsseD CONCRETE.
a.STEEL
Figure 1: Diagrams showing strain and stress dis-
tributions in ordinary reinforced and pre- In the case 01 prestressed concrete stress
stressed beams. at the origin 0 is not zero. but corrcsilonds
to that given by the load conditions in Fig.le
tests by measuring the concrete and steel strains
from which values of nl and T (the total tension)
can then be deduced. Then ~ =d-ynldwhereMis
the externally applied bending moment. The value
of !:! is probably '33 at low stresses and between
·4 and '5 at high stress, approaching failure of the
concrete.
The value of 0( can also be deauced from
beam tests and by applying the expression
M =0( bn1dc (d -yn1d) and is probably ·5 at low
stresses, and at high stress varies between approxi- A
mately ·75 for dense high grade concrete and 1 for
ordinary grade concrete. A value of c equal to ..... ./
the concrete prism strength must be assumed.
In relating the steel to the concrete stress in end
..
L

\I) ,/
,/
./

,/
anchored prestressed concrete beams it is necessary ,,/
to equate the total strain of the steel throughout ,/
the length of the beam to the total strain in the ./
./
adjacent concrete since there is no bond. A ./
factor F must therefore be introduced and its value ./
may be assumed to be about ·66 for beams o 8
Strain
supporting uniformly distributed load, and not
less than ·5 for beams supporting central point
load.
The value of f1 depends on the value of the pre- b.CONCRETE.
tension applied to the steel and the steel stress at
which the beam reaches its ultimate strength.
Since the value of the steel pre-tension is usually
approximately '66 of the yield stress. lmd since Figure 2: Stress-strain curves for steel and concrete.

58
Plastic theory for reinforced concrete
losses due to shrinkage creep and elastic yield are
small and bending before cracking only slightly
increases the steel tension, the value of fJ will
generally be approximately' ,75. Its precise value o·a ~----+-----~-----+------~--~
can in any given case be determined without great n _ __ p.+/p.Z+~
difficulty. ,
100
E.. sometimes called the effective or secant 0-6 ~____+-____~~z~m~+-____~~~~~
modulus of elasticity of steel, may drop from ~t1\.l!l
30x 106 to 20x 10 6 or even lOx 106 lb. per sq. in. n,
with some steels without appreciably reducing the
strength for repeated loading. Generally it is not
necessary to consider values below lOx 106 lb.
per sq. in. as when such a value is reached, the
concrete begins to crush on account of the high
position of the neutral axis. For high tensile steel
wires as used in prestressed concrete the value of
E, may drop to 20 x 106 lb. per l1q. in. on account
of straightening. o 0'2 0-4 0·6 o·a 1·0
Ps per cen t
E., sometimes called the effective or secant
modulus of elasticity of concrete, varies with the
stress, quality, water-cement ratio and form of Figure 3: Curves relating depth of neutral axis to per-
cenrage of steel for various values of zm.
the concrete (e.g. test cube, prism or column values
are lower than beam values). It is essential that
the values used should be those at the extremity of the value of Ec in regard to the concrete at the
the cross section of a beam. The value increases extremity of the section lies between 500c and 820e,
towards the neutral axis. C. S. Whitney (1) found where e is the cylinder crushing strength of the
from tests of a large number of beams that for concrete. Because of this relationship it is possible
balanced designs M='33 bd2 e. (where e is the to omit the term m in the expression for the
cylinder crushing strength). Professor R. H. mOIllent of resistance of a concrete beam of
Evans(2) has also obtained similar results. balanced design.
Since M=(1.bd 2c (nl-yn12)
EXPRESSIONS FOR ULTIMA TE STRENGTH IN
Assuming (1.=1 and y= '5, M=bd 2c (nl_n;) BENDING

n1 2 Since strain is proportional to the distance from

.. nl - 2 =t or nl =0,425 the neutral axis
Since strain is proportional to distance from the (=e (l-nl) Fm " .. (1)
neutral axis ndl-fJ)
E, e(l-nl)
(=- ---
Ec nl F
Sttbstituting nl =0,425 Let z = 2(1. (l-fJ)
E,=30 X 106 and (=50,000
Then substituting in (1)
Then E c =820c
1
Similarly, assuming y =·8 and y= '45 then nl=------
Ec=500c.
The value of c at the edge of the section is 1+ 2 (1. ( (2)
zmc
slightly less than the cylinder strength, but the
error involved in assuming them to be the same Expression (2) gives the value of nl for the
is not important. ultimate load condition for a balanced design,
Since the limits chosen above for (1. and yare provided the values of mt and c are appropriate
reasonable, it can be assumed that at the stress at for this condition.
which the concrete in a beam commences to crush, By equating total tension to total compression

(1) c. S. WHITNEY Application of plastic theory to the design of modern reinforced concrete structures, Journal of the
Boston Society of Civil Engineers, January. 1948.
(2) R. H. EVANS The plastic theories for the ultimate strength of reinforced concrete bpams, Journal of the Institution
of Civil Engineers, December, 1943.

59
 Magazine of Concrete Research: June 1949
r ABLE 1 : Values of R for ultimate load conditions in balanced design for various types of beam.
M=rJ. bd2 c (nl-ynl2)
=Rbd2

I E, t
Type of beam F rJ.
f1 z lOS E. m zm 1,000 nl Y R
I
-- --- - - --- -- --- - - - - - - --- - - -- --
60,000 38,000
Ordinary reinforced 1·0. 0.·8 0. 0.·63 30. 500c 50. 0.'55 0.·4 o.'33c
concrete, plastic c c
theory
60,000 38,000
1·0. 0.·8 0. 0'63 30. 500c 40 0.·60 0.'4 o.'38c
c c
36,500 18,250.
1·0. 1·0. 0. 0.·50. 30. 82o.c 50. 0.·42 0.·5 o.'34c
c c
36,500 18,250. _
1·0. 1·0. 0. 0.'50. 30. 82o.c 40 0.'47 0.·5 o.'37c
c c
--- -- - - --- "-- -- --- --- -- -- -- -
End anchored 60,000 75,000
0.·5 0.·8 0.'75 1·25 30. 500c 200 0.·375 0.·4 o.'25c
pre-stressed reinforced I c c
concrete beam. Central
point load. Straight 24,300 10,000
wires. 0.·5 1·0. 0.·4 0.·415 20. 82o.c 175 0.·102 0.·5 a'lOe
e e
-- -- -- -- -- -- --- --- -- -- -- --
End anchored 60,000 98,400
0.·66 0.·8 0.·75 1-64 30. 500e 200 0.'44 0.·4 o.·2ge
pre-stressed reinforced c e
concrete beam.
Uniformly distributed 24,300 13,400
load. Straight wires. 0.·66 1'0. 0.·4 0.·55 20. 82o.e 175 0.·133 0.·5 o.'12e
e e
"- -- -- -- -- -- -- --- --- -- -- -- --
60,000 150.,000
Concrete gripped 1·0. 0.'8 0.'75 2'5 30. 500c 200 0.·54 0.·4 o.'34c
pre-stressed concrete c c
60,000 150.,000
1·0. 0.·8 0.'75 2'5 30. 500c 175 0.'51 0.·4 o.'36c
c c
24,300 20.,000
1'0. 1·0. 0.·40. 0.·83 20. 820e 200 0.·17 0.'5 o.'16c
c c
24,300 20.,000
1·0. 1·0. 0.·40 0.·83 20. 82o.c 175 0.·18 0.·5 o.'17c
c c

the following expression for nl is obtained: and the curves can be used in the usual way with
the appropriate value of m (e.g. m=15 in 1948
_p+ ~p2+ 200p,
__
, • zm British Standard code of practice).
(3) M =rJ.bd2c (nl - ynl2) .. (4)
100 = say Rbd 2 (5)
zm
Expression (3) gives values of nl in terms of Table 1, calculated by applying the above
p, and zm and may be used for "unbalanced" expression, gives values of R for ultimate load
designs, at any stage of loading to find the value of conditions in balanced designs for various types of
nl for appropriate values of p, and zm. Figure 3 beam, for extreme values of rJ. and zm and for
shows curves relating nl to p, for various values of suitable values of t. Concrete moments of
zm and may be used to find values of p, for a resistance in .. unbalanced" designs can be
balanced design (i.e. for the general case nl having obtained in a similar way by using the curves in
been found from expression (2)). In ordinary Figure 3 to obtain the value of nl for the appro-
reinforced concrete design (standard theory) z = 1 priate conditions. Whitney's results appear as a

60.
Plastic theory for reinforced concrete
~ TENSION STRAIN x IO~ COMPRESSION - . .
-12 -II -10 -9 -8 -7 -6 -5 -4 -3 -2 -I 0 I .2 3 '" 5 6 7 8 9 10 11 12 13 14 15 16 17 18 ,<

Figure 4 : Curves showing the strain distribution measured at the centre of an ordinary reinforced concrete beam.

~TENSION STRAIN X 10 4 COMPRESSION ~

.E, 25 20 15 10 5 o 5 10 15 20 25 30
0
T//~
~ ~
c(
2

.
W
CO

0
Q. 6
4
~H ~~ ~
~

tr
~ ~V-/ ' ~~
2101 iD-o 22-1
~ 1---'
____

-
8
~ V
V
~ //J
.
IX

W
10

12
17-0 I ·0

.,.-
14·0

././
.,.- //
.,.-

/ /} Numb(~r ncar loch curve indic;:atls total

lood P in tons .
U 14
.,.- ...-
Z

~ 16
./
~...- / / } --
25 18
IN /'
./
10'0/ 7·,L 4-Oi
(a) Strain d j stri but ion at various stagu of load j·ng .

TOTAL LOAD (p) TONS.

10 15 20 25
0
p p

r
"2 "2

r-
E:raCkS

+ + 5'-1
2 opening
s' s'
C
I
4
'0_

f f
C 6 80ttom Of flon e

10
(b) Mov~m~nt of n~utral axis. (c) LoadinCj system.

Figure 5 : Test on an end anchored prestressed concrete beam (Magnel system).

61
 Magazine of Concrete Research: JUlIe 1949
·8 r---.,---.,-----r----,----,

y ---x

".
~·4r------r----~------+_~r7~~--~
,
N.
c:
----
n,d
-- --t -strain crost
distribution 'Scction
o ·8
",
Figure 7 : Relationship between values of n1 and
o x (1Il 2-YIll3)
for various value of y.

short length ds,

r=radius of curvature of the member at P
1I 1d=depth of neutral axis

Figure 6: Diagram showing the change of curvature of Eo = modulus of plasticity of concrete at

the neutral axis plane. . extremity of section.
c=maximum stress of concrete in segment
PQ
particular application of a more general theory. M=applied bending moment at P (which is a
The value of the lever arm for determining the area function of x, hence M and curvature
of reinforcement increase in direction x)
=d-yn1d.
In a very much under reinforced beam of high Slope of beam at point P =~ 'fix
grade concrete y might be as low as 0'33 and 111
high on account of the value of <X being about dx=ds very approximately as long as the
slope of the beam is small.
0'50. Eo however would be high, say 4 x 106
lb. per sq. in., and E, migRt drop to 20 X 10 6
d2 y
Rate of change of slope at P=--
lb. per sq. in. before the steel appreciably weakened.
dx2
Re-distribution of moments in continuous beams I
=degree of curvature = .-
and frames, and elasto-piastic deformation of re- r
inforced concrete members. dy /1,dx
SIope=(jx=
ASSUMPTIONS
(I) Strain is proportional to distance from the QR is paranel to P'O, .'. ~~ =: ~~
neutral axis. alsods=dx
(Tests indicate that this is approximately true ds cdx
even after plasticity has fully developed. r
Figures 4 and 5 show typical results from tests 1 c
carried out on ordinary reinforced and pre-
stressed beams respectively).
(2) All changes of slope or radius of curvature of
members can be related to and expressed in
terms of the deformation of the concrete on the
compression side of the neutral axis.
Referring to Figure 6 which shows the side
elevation of the neutral axis plane OPQ of a
reinforced concrete member having zero slope
at x=l, PQ is a segment of the member of infinitely

62
Plastic theory for reinforced concrete

lb. It is also assumed that the concrete section of

the beam is constant and the value of EI = 1 before

~."
plasticity develops .
.. sIope at B = T
. h A as ongm
WIt 1 JMXd
EI x
A
~
C~ 60ft.
B
~
Equating/times the slope at B due to free and
support moments (EI being constant and equal
to 1 is omitted).
Figure 8: Bending moment diagram for two-span con- }MFq· = MBi·tl
tinuous beam. 24,000
=
.. M B =MF =20 lb. ft.
Then slope of beam = f~dX. Therefore the
If no plasticity develops then the bending
usual methods of calculating deformations accord- moment at the support =20 lb. ft.
ing to elastic theory can be used, provided values and the mid-span bending moment = 10 lb. ft.
of EJ due to plasticity are taken into acc·ount. If the beam is slightly over-reinforced (Ps = 1 per
Elastic theory methods which are convenient for cent.) so that the concrete commences to crush at
dealing with problems in which the value of I varies B before the steel commences to yield, the value
are also convenient for taking into account the of I might be reduced before failure over the end
effects of plasticity on moment distribution. The 4 ft. of the beam adjacent to B (referring to
curves in Figure 3 can be used to determine varia- Figure 7) from '58 IX bd3 to 0·5 IX bd3 causing an
tions in the value of ni caused by changes in the increase in the
value of zm when plasticity brings about reductions MX ·58
in the value of Ee or E,. J EI dx value of 20 x 4 x 58(
=740
--:s
-1)

SPECIAL CASE OF ORDINARY REINFORCED

CONCRFTE BEAMS This reduction based on reasonable assumptions
Tests have shown that in the case of ordinary is small in comparison with 24,000. It is clear
reinforced concrete beams the position of the therefore that plasticity in the concrete can assist
neutral axis remains approximately the same as very little in equalising the moments at mid-span
long as the value of the modulus of elasticity of and support, unless plasticity extends over a wide
the steel does not alter. Referring to expression (3) support. It will be seen that yielding of the steel
above it will be seen that if nIP, and E, are constant is much more effective. From the shape of the
F=1 and {3=O, then Ee ~ must be constant. bending moment diagram it is also clear that
sufficient reinforcement must be provided at
Assume E/l. = K mid-span to prevent any yield occurring there
then EJ = Kbd 3 (n1 2 - y n l 3 ) .. (8) before sufficient yield has developed at the support
In a reinforced concrete beam, therefore, the to appreciably reduce the support moment. Yield
variation in value of EJ caused by plasticity over a considerable length at the centre of the beam
cannot be very great.
In Figure 7 the term (nI2_ ynI 3) is plotted for would increase theJ~;dx value for the free
various values of ni and for y = ·33 and '5, and it
moment at a faster rate than for the support
is seen that the value of EJ can only vary on
moments as these high moments only extend over
account of plasticity to a very small extent. Over
a short length of beam unless the support is very
a long period it is probable that owing to creep ~
wide. Yield due to plasticity of the concrete only
would approach 1 in value and Ee might be reduced
could have the same effect and cause an increase
to t of the low stress value, in which case the value
in the support moment.
of EJ obtained from expression (8) above would
be slightly reduced. Re-distribution of bending moment due to yield
of steel only
EXAMPLE OF MOMENT RE-DISTRIBUTION It is not practical to obtain a direct solution
IN A 2 SPAN CONTINUOUS BEAM since there are so many factors involved. A
Figure 8 shows a continuous beam having two solution within safe limits can, however, be quickly
equal spans. To avoid cumbersome expressions obtained by trial with the help of the curves in
numerical values are assumed for the equal spans Figures 3 and 7.
/=60 ft. and for the equal maximum bending A value of the ultimate load support moment
moments for uniformly distributed load M F =20 ft. must be assumed and checked.

63
 Magazine of Concrete Research: June 1949

TABLE 2

E, *M Increase in
x m nl n12-yn1 3 EI factor M factor*.
106
EI
II
59·5 10 4 0'2 0·035 2·5 0·5
58'5 17 6'8 0·24 0·055 1·64 0·64
57·5 24 9'6 0·28 0'075 1·18 0'18
56·5 30 12 0·32 0'090 1·0 0·0

plasticE'1values "Factors for the increase in ; values due to plasticity.

*Ratio of
elastic ~ values

TABLE 3

M
- x
Mx.
Yield increase I I ncrease In. EI
Mx
I EI I I EI I factor
Support moments. 16'75 59·5 1,000 1'5 1,500
14'50 58'5 850 0·64 545
13·0 57'5 750 0·18 135
I

EI Increase Iior support moments

IJMx. =2,180

Free moments. 0·5 59·5 30 1·5 45

1'5 58'5 88 0·64 66
2'5 57·5 144 0·18 26

I IJ ~x increase for free moments = 137

I I
n Mx
At ultimate load after yield ~ EI free moments = 24,000+ 137=24,137

IJMx
EI support moments=(l7x30x40)+2,180=22,480

ASSUMPTIONS (5) the beam (outside) value of Ec just previous to

(1) Maximum mid-span free moment=20 lb. ft. the concrete reaching crushing stress = 2· 5 X 106
(2) Support moment (no plasticity) =20 lb. ft. lb. per sq. in.
(3) The values of IX, h, d, Y and the crushing (6) Sufficient steel is provided at mid-span to
strength of the concrete are such that: prevent yield.
(a) the concrete crushes when nl is less than
As a first approximation try the support moment
0'20.
reduced' by plasticity to 17 lb. ft. The following
(h) the steel appreciably weakens and may be
conditions at the support section must be satisfied:
considered to have reached ultimate
(1) p, not less than 0·6 to avoid excessive steel
strength at a tension of 58,000 lb. per
tension when nl =0'2.
sq. in. for which a value of p, =·6 is
(2) nl not less than 0'2 which depends on the
required when nl =0'20, and M = 171b. ft.
value of p, and m which depends on the value
(4) the value of E, vanes as shown in Figure 9. of E, which again depends on the value of p,

64
Plastic theory for reinforced concrete

Tension.

30

L """ A

"'1°
w-
Concrete
Specimen. "
"
'I
I,
-'
10 Pocking Plate rli'I;~~~(----rrrr
Hydraulic cool
bursters for
applying non-uni form
load.
40 so 60
i Section A-A is identical with
10'
beam cross section.

Figure 9: Relationship between Plastic Figure 10: Diagram of bending assimulation machine.
Modulus of Steel and Steel
Stress.

which determines the tension in the steel. value between ·33 and ,5. The value of (n1 2 -yn13)
From Figure 9, E, =lOx 106 when 1=53,000 is proportional to the 1 value which for comparative

.. m=~s=4
Ec
purposes has been assumed = lover the un yielded
part ofthe beam, i.e. 1=1, when (n12-yn13)=·09.
From Figure 3, when Ps =·6 and m =4, nl =·2
approximately, which approximately satisfies the Assuming bending moment at B= 17 at the
above requirements. ultimate stage of loading, Table 3 gives elastic
It can therefore be assumed that under ultimate Mx
EI values for free and support moments and
load conditions the steel at the support develops a
stress of 53,000 lb. per sq. in. The bending increases due to plasticity_
moment diagrams to the left of B (see Figure 8)
must now be divided into strips of convenient It is seen that the ~; values for free and support
width, starting from B and extending towards A, moments are not equal and a further trial with,
until the bending moment is reduced to ~ of the say, support moment equal to 18 should be made
or, in practice, it would probably be found that by
maximum bending moment, i.e., until a section is taking into account yield extending over the width
reached at which the steel stress =40,000 and no
yield occurs. The effect of yield over this range in
of the support a sufficient increase in~; values for
. . Mx I . the support moments would be obtained to effect
mcreasmg EI va ues can now convemently be de- a balance.
termined by tabulation and summation. It is In the case of continuous prestressed members
assumed that the value of p, equals ·6 throughout, the effect on the value of nl at every section due
that no yield occurs at mid-span and that any to variation in the value of f3 would need to be
slight increase in p, to ensure this does not taken into account. Over the un cracked parts of
appreciably affect the value of 1 (see tables 2 and 3). bd 3
the beams Iwould be assumed = 12 approximately.
The value of E, varies with the steel tension in
accordance with Figure 9. The steel tension is Similarly in vertical members subject to bending and
assumed to vary with M (i.e., it is sufficiently bd3
accurate to assume that the lever arm remains compression over uncracked lengths 1 = - -
12
constant) : approximately and over cracked sections
E, 1=rf..bd3 (n12-yn13) approximately. Care must be
m = - - -6
2·5 x 10 taken to use appropriate values of Ec and 1 in
nl is obtained from Figure 3, and the terms conjunction with changes of stress due to bending.
(n12-yn13) from Figure 7, assuming y has a mean These approximations will probably serve to

65
 Magazine of Concrete Research: June 1949
demonstrate that plasticity in the concrete has a considerable influence on the moment re-
only a small influence on moment re-distribution distribution. Sway bending moments in frames,
in comparison with the yield of suitably arranged since they reduce less rapidly along the members
reinforcement and can often be neglected. than bending moments at supports in continuous
In frames subject to sway, yield of reinforcement beams supporting distributed load, are more
will extend over a much greater length of the subject to advantageous bending moment re-
members on account of the shape of the bending distribution. In prestressed concrete beams the
moment distribution diagrams and will therefore small amount of steel yield available may reduce
have greater influence in re-distributing bending the moment re-distribution possible as compared
moments. with ordinary reinforced concrete.
(4) More precise data are required in regard to
Conclusions the values of IX y and Ec for various grades of
(1) Plasticity of concrete considerably increases concrete at various stresses. It is impracticable
the ultimate concrete moment of resistance of both to test ·Iarge numbers of beams for this purpose
reinforced and prestressed concrete beams. and in any case the exact stress distribution cannot
(2) Plasticity in the concrete does not greatly be determined in this way. It ii therefore essential
influence the re-distribution of bending moments to test large numbers of specimens identical with
in continuous members, as the reduction in the short lengths of beams of various cross section in
value of Ec is small. In cases where high .. free" a testing machine which can apply a load similarly
bending moments extend over a considerable distributed to the stress in a beam subject to
proportion of the length of the beam plasticity may bending. Such a machine is shown diagram-
be a disadvantage in regard to bending moment matically in Figure 10. When specimens are
re-distribution. deformed so that the strains are identical with
(3) Yield of reinforcement can provide ad- those obtained in beam tests (generally a straight
vantageous moment re-distribution, but may be line distribution) it follows that the specimen
limited by weakening of the steel or by excessive must be subject to the same stress distribution as
rise of the neutral axis which causes the concrete similar concrete in the form of a beam. True
to crush. It is important to ensure that reinforce- stress-strain relations of concrete at all parts of a
ment cannot yield at mid-span sections of con- beam cross-section can then be determined from
tinuous beams until support sections have de- strain gauge readings on the specimen and pressure
veloped full yield. The width of the support has gauge reading on the hydraulic jacks.

66