INSTANTANEOUS AND TIME-DEPENDENT DEFLECTIONS OF

SIMPLE AND CONTJ]fQOJJ~ REINFORCED CONCRETE BEAMS

PART I

by

Dan E. Branson
Associate Professor of Civi l Engineering
Auburn University

DEPARTMENT OF CIVIL ENGINEERING

AND
AUBURN RESEARCH FOUNDATION
AUBURN UNIVERSITY
1963

federal Highway Admin.

Technical Reference_ Center
6300 Georgetown Pike
McLean, VA 22101-2296
 ii

FOREWORD

This is a report of research performed under Project

No. 5O26C-l, Auburn Research Foundation and sponsored by
the State of Alabama Highway Department in cooperation with
the U. S. Department of Commerce, Bureau of Public Roads .
The project was conducted by personnel of the Department of
Civil Engineering, Auburn University .

Grateful acknowledgment is ma.de to Mr . J. F. Tribble,

Chief 1 Bureau of Research and Development and Mr. F . L.
Holman, Assistant Research Engineer, State of Alabama Highway
Department for their cooperation, and to Mr . D. R. Luger,
Graduate Assistant, for conducting the laboratory tests of
the investigation . Appreciation is also expressed to the
American Cast Iron Pipe Company, Birmingham, Alabama for
donating the iron bricks used as loads in the study.
 iv

TABLE OF CONTENTS
Page

FOREWORD t ••••••••••••••••• ••• •••••••••• ••• • • t • t • •••••• • •••• ii

ABSTRACT lii

TABLE OF CONTENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iV

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V

LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi

I. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1 .1 Object and Scope of the Study ... .......... . ... . . 1
1 . 2 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . l

II. NATURE OF THE DEFLECTION PROBLEM FOR REINFORCED

CONCRETE FLEXURAL MEMBERS .... .... . .....••. ,... . .. . ... . 4
2.1 Primary Factors Involved in Deflection Prediction
and Control of Reinforced Concrete Flexural
Members . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2. 2 Review and Discussion of Existing Methods, Guides
and Rules of Thumb f or Predicting Deflections .... 5
III . DESCRIPTION OF EXPERIMENTAL INVESTIGATION .... . . ....... 19
3.1 Specimens and Instrumentation . .. .. . . ... . ........ 19
3. 2 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

IV . EFFECTS OF CRACKING ON INSTANTANEOUS DEFLECTIONS OF

SIMPLE AND CONTINUOUS REINFORCED CONCRETE BE&'1S . . . . . . . 23
4.1 Development of an Analytical Method for Including
the Effects of Cracking in the Prediction of
Instantaneous Deflections .. ...... ....... .. . . ... . 24
4.2 Outline of Computational Procedures . ........ . . .. 27
V. DISCUSSION OF TEST RESULTS ... . . ... .. .. ... ..... . .... • . . 46
5.1 Shrinkage Warping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.2 Deformational Behavior of Test Beams ... . .. ... ... 49
VI. CONCLUDING REMARKS ... . . . .. .. .. .. . ..... ..... ..... . ..... 63
VII. REFERENCES CITED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
VIII . APPENDIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . 67
8.1 Specimen Details and Experimental Data Obtained
in the Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
 V

LIST OF TABLES
Table

1. Creep -Coefficients 7
2. Computed Deflections Compared with Test Data . . . . ..... . 35
3. Loads, Beam Details and Section Properties for
Test Beams .. . .. .. . .. .. . ..... .. .. . .. .. . . . . . . , . . . . . . . . 37

4. Concrete Properties and Parameters for Test Beams .. . .. 43

5. Beam Details and Concrete Properties for Shrinkage
Speci.Jnens . . . . . . . • . . . . . . • . . . . . . . • . . . . . . . . . • . . . . . . . . . . 56
6. Computed Shrinkage Warping Compared with Test Data . . . . .58
7. Beam Moments and Experimental Curvatures for the Test
Beams of t he Current Investigation ....•.. .. ... .. ... . 61
A. 1. Design Details for the Test Beams of the Current
Investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
 vi

LIST OF FIGURES
Figure

1. Creep strains by the rate of creep method 15

2. Creep strains by the superposition method 15

3. Geometry and details of test beams .. . . .. . .. ... .. . .. .. . 20

4. View of t est beams, shrinkage specimens and

instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

5. View showing close- up of Whittemore gage and dial gage. 21

6. Example of Newmark numeri cal solution for computi ng
deflections of simple beams (Beam SB- 3) using an
effective moment of inertia at the individual
sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

7. Example of Newmark numerical solution for computing

deflections of continuous beams (Beam LB- 3) using
an effective moment of inertia at the individual
sections . Effects of moment redistribution due
to cracking are incorporated in the numer ical
solution .. .. ..... . .... . . .. .... . ... . ...... .. . . . , . . . . . 33

8. Comparison of shrinkage strains at the top fiber for

t he specimens with different steel percentages
(strains proportioned to extreme fiber using a
linear distribution with the top and bottom gages) • .. 52
9. Average rate of increase for shrinkage and creep
strains . .. . . . . .. ... .. .. ........ . ..... . , . . .. . . . . . . . . . . 53
10. Compression and tension gage creep coefficients versus
time curves for four test beams . .. ...... . . . . . . .. •.. . 54
11 . Time- dependent deflection coefficient versus time
curves for four test beams . . . . . . . . . . . . . . . . . . . . . . . . . . 55
A. l . Concrete stress- strain curve at age 28- days .... .. . .. . . 69
A. 2 . Concrete shrinkage versus time curves for specimens
containing different steel percentages (duplicate
shrinkage specimens were used) .......... . ..........• 70

A. 3 . Average shrinkage curvature along members versus time

curves . . ....... . ... . .. . .. . ... . .. ... . . .. . . ..... . . . _ . . 71
 vii

A.4. Total (instantaneous plus time-dependent) concrete

strain versus time curves for two simple beams
with different steel percentages and loading, but
the same computed elastic concret,e stresses . . . . . . . . . 72

A.5 Total (instantaneous plus time- dependent) concrete strain

versus time curves for two continuous beams with
different steel percentages and loading, but the same
computed elastic concrete stresses . ..... ... .... . .. . . 73
A.6 . Instantaneous plus creep strain versus time curves for
two simple beams with different steel percentages
and loading, but the same computed elastic concrete
stresses .. . .................... . . .... .. . .. ..... .. , • . 74
A.7 . Instantaneous plus creep strain versus time curves for
two continuous beams with different steel percentages
and loading, but. the same computed elastic concrete
stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
A. 8 . Total (instantaneous plus time-dependent) curvature
versus time curves for four test beams .. . . . . .. .. ... . 76
A. 9 . Instantaneous plus creep curvature versus time curves
for four test beams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

A.10 . Time -dependent deflection versus time curves for four

test beams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
 l

I. INTRODUCTION

1.1 Object and Scope of the Study

With the present-day tendency toward the use of higher

strength concrete and reinforcing steel, and shallower aect-
ions, the problem of deflections ia aasuming greater and
greater importance . The purpose of thia inveetigation 11 to
consolidate information on deflections as much as possible
and to study the complex defo~mat1onal behavior of reirtforced
concrete beams as influenced by the interrelated effects of
cracking, shrinkage warping, creep, teneile and compreosive
steel percentage , continuity, moment redistribution in stat-
ically indeterminate beams, etc.

The experimental phase of the program waa designed to

elucidate certain aspects of the deflection problem not here-
tofore clearly defined, such ae the relative effects of high
quality concrete, effects of sustained loads sufficient to
cause 1110derate cracking, and the effects of special combina-
tions of singly-reinforced steel percentages in companion
simple and continuous beams.

Particular emphasis is placed on a study of the effects

of random cracking on deflections; especially with regard to
moment redistribution in continuous beams resulting from
cracking. Shrinkage warping and creep deflection are also
analyzed from both theoretical and empiri cal points of view.
Analytical procedures for predicting the various aspects of
the deflection problem are discussed and. in certain cases,
new procedures advanced. Comparieons are made with test
data to show the nature of the agreement that can be expected.

1.2 Notation

Avg. 1 eff average effective moment of inertia for simple

spans (Eq. 24)
As area of tensi l e steel
Ns area of compressive steel
a ~- cremental length of beam
b width of beam at the compression side
b' width of beam at the tension side
C constant, also used to denote compresaive force
ct creep coefficient defined as ratio of creep strain
to initial strain
D total depth of beam
d effective depth of concrete section
d' distance from centroid of compressive steel to
extreme compressive fiber
EI flexural rigidity
 2

--aodulua of elasticity of concrete, short duration of

loading
Ect --reduced or sustained modulus of elasticity of con-
crete, long duration of loading
--modulus of elaeticity of ateel
--average effective modulu1 of elasticity of steel
when participation of tensile concrete is taken in-
to account (see Eq. (9))
e --distance between the centroids of the uncracked trans-
formed aection (u1ing net> and the steel area
--distance between the centroids of tbe gross concrete
section and the • teel area
--compressive stress in concrete
--concrete compre• sive strength at age 28 daya 1 or
other age if specified
--modulu• of rupture of concrete
--steel stress
--yield point of steel
--relative humidity (H • 70 for 70l herein)
--average effective moment of inertia for continuous
beam, (Rq1. 25 and 26)
--moment of inertia of the cracked transformed section
--moment of inertia of the uncracked transformed sec-
tion using net
--effective moment of inertia at an individual section
(Eqs. 21, 22, 23)
--moment of inertia of the grosa concrete section (neg-
lecting all steel)
--moment of inertia of the uncracked tranaformed sect-
ion
kd --distance from extre• e compression fiber to neutral
axia for cracked transformed section
L --span length
M --bending moment of beam
Mer --moment corresponding to flexural cracking
m --a conatant power
max --aubacript denoting maximum value
n --IQOdular ratio defined aa E8 /Ec
net --increased modular ratio defined as Es/Ect
p --tensile steel percentage defined herein as (Aa/bd)
(100) 1
p' --compre•sive steel percentage defined herein as
(A~/bd)(lOO) ~
Pw -•steel percentage in T-beama defined as (A 8 /b'd)
Pf --•teel percentage in T-beams defined as (A 8 f/b'd).
where As£= (0.85)(fb)(b - b')(t)/fy
--equivalent concentrated load
--tensile force
--total compressive force induced in steel by shrirulage
--flange thickness for T-beama
 3

t - - denotes time interval, also used as subscr i pt

denoting t ime-dependent
u --subscript denoting ultimat e value
V --beam shear
w - - uniformly distributed load, a lso unit weight. of
concrete in Eq. (1)
- -uniform dead- load
--uniform superimposed- l oad
--beam deflection
-- - distance from neutral ax.is to the extreme fiber in
tension
--maximum deflection
- - comput ed maximum deflection using t he cracked
transformed section moment of inertia
-- specific creep or unit creep strain defined as cr eep
strain per unit stress
- - unit strain
-- steel strain
- - free shrinkage strain
--berun slope
- - unit stress
- -curvature or angle change per unib length of beam
- - curvature due to shrinkage warping
--equivalent concentrated angle change
--coefficient taking into account the participation of
concrete in tension (see Eq . 9)
 4

II . NATURE OF THE DEFLECTION PROBLEM FOR REI WORCED cm CRETE

FLEXURAL MEMBERS

2. 1 Primary Factors I volved in De lection Prediction and

Control o Reinforced Concrete Flexural Members

The problem of predicting and controlling deflections of

reinforced cone ete flexural members under wo king loads is
extremely complex as a result of the large number of sig ifi-
can yet uncertain £actors involved. A partial list and brief
di cussion of t he more important factor follows :

1 . I.ack of accura e knowledge in advance of pertinent

concrete properties· such 2s modulus of rupture and c.ompressive
trength moduius of elasticity, and shrinkage and creep charac-
teri stics. Knowing m:i11imum specified strengths is not enough
~ince this does not provide sufficient information of for
example shrinkage and creep behavior. Higher strength con-
cretes may or may not shrinK and creep less than lower strength
concretes. It can obv-iously be aid, however, that when minimum
strength and modulus values and maximum shrinkage and creep
values are used computed deflections will tend toward the
high side.

2. Ambient emperatures and humidities which arfect the

i~ems in l , The priJllary influence here i uso.ally the effect
of humidity on shrinkage and creep .

J.Concrete age when sustained loads are applied which

primarily affects creep behavior .

u. The ef_ective section proper t ies under instantaneous

load along the beam, including primarily the effect of "extent
of cracking"_ The cracked and uncracked transformed section
properties are the two theoretical extremes and t.hen only for
linear- elas ic materials . Differences in the gross and uncracked
transformed section properties are seldom worth con idering and
the gross section is much more convenient to use for design pur-
poses. I nvolved in the determination of the e_fective flexural
rigidity is the contribution of concrete in tension between
cracks . Also involved is the effect of steel percentage, varying
depths and the flanges of T- be8.JTlw (especially for continuous
beams) on the effective section properties along the beam.

5. Difficulty in determining shrinkage warping and creep

deflections including the effects of a given crack pattern as
well as the phenomenon of progressive cracking under sustained
loads. Involved is a movement of the neutral axis with tilTle
as a result of the time-dependent deformations in the non-
homogeneous composite concrete- steel structural member , Also of
 s
importance is the effect of compression steel in reducing
shrinkage and creep deflections. This i s especially important
with regard to ultimate strength designs where it is usually
more economical, from a s t rength standpoint, to place additional
steel in tension rat her than use compression steel .

6. The determination of what constitutes critical

deflections; that i s the difficult question of serviceability.

7. Other factors include . the increase (above the 28- day

values used in design) in concrete strength and modulus of
elast icity with time the effects of bond creep, member size,
slab act ion etc.

The difficulties involved in rationally analyzing the above

effects are virtually insurmountable in the average design
office if not in the research office . The problem appears to be
primarily one of a statisLical nature involving statistically
optimum designs and confidence intervals for computed deflections.
The large number of variables involved, the variabilit y of these
parameter s and the interdependence of most of the variables
strongly supports this point of view. Nevertheless a determin -
istic formula or formulas however approximate, which incorporates
all of t he factors that may be pertinent in a given design
si tua t.ion would be of benefit to both the designer and t he
researcher. It is to thi s t ask that the report herein addresses
i t.self particularly with regard to the effects of cracking)
warping, continuity and steel percentage .

2 . 2 Review and Discussion of Existing Methods Guides and Rules

of Thumb for Predicting Deflections

Presented in the following paragraphs i s a brief discussion

of existing methods. guides and rules of thumb for determining
deflection parameters and deflections themselves of reinforced
concrete flexural ;members. Items 1 through 6 of Section 2 .1 are
considered in that order :

1 . , 2 . and 3. Concrete Proper ties :

Values of .modulus or rup ture and modulus of elasticity of

concrete are not accurate function of compressive strength
alone. evertheless for mot practical applications, the
:'allowing approximat e _ornnD.a- are usua11y sali sfa.c uOr:'r :

l 2 Ec = 33 ~ (1)

EC = 57 . 700 VT fo concrete weighing 145 pcf (2)

3f~b= 7.5 l{rf (3)

 6

where Ee is the ins Lan taneous modulus of elasticity, 1r1 is the

unit we1.ght, of concrete . f~ is t he comp!'essive strength and f~b
is the modulus of rupture .

Coucret.e streng Lh , modulus of elasticity, shrinkage and

creep cont.:..nue to increase .for very long periods of tiffie. I n
the case of shrinkage and creep properties it. is only poss ible
to generalize wit.hin ra~her broad limiLs , and accurate test dat.a
whi ch incorporates the effects of local conditions should be
used when availabl e . In the absence of test data: the following
shri nk.:ige and creep inf orma t.i on is of ten useful :

Schoret' ' s 4 formula is probably adequate for calculating

shrinkage strains for 1·11ost desig!'I purposes :

€sh = 12 . 5 x 10 - t' ( yo - H) (Lt.)

where €. sh is the free sl1rj rikage strain in inche:;; per inch and
H i s relative: humidity (H = 70 fo1· 70% rel. hurn . ) . This formula
gives an ultimar,e ur design Lotal ::;ltrinkage strain as a fw1cti('ln
of relative humidit,y bu!, olher variables accoru1t f or rather
wide variations W1der cer-r;ain conditions . Howevec, most,
Ehrinkage data agree with Eq . (4) within 25%.

In cons:..dering the effects of creep on the deflection 01'

concrete members, the use of a unit creep strain dt (creep per
w1it. stress) or a creep coefficient Ct. (ratio of creep strain to
ini tial strain) amom.ts to t.i:.e same tni!lg, since the concrete
modulus Ee must be brought in in either case and

C t. == J t.Ee (5)
This is seen from the relati.on

Creep Strain = ( 0-cunstan 1) rft, = ( €.initial) l\ (6)

where i,::
C
= a-constant) I ( t. initial) (7)

Which to use is a ma Lter o.:· c:onvenience depending un whether it

i.,.; desired to apply the creep factor t-.o applied stress or str ain
1-,hen compu Ling creep strain in Eq . ( 6) .

Approxil!late ultimate values .for Lhe creep coefficient fot·

·1ormal weight concrete tinder average design conch tions a r e shown
in '!'able 1, where. in each case , the larger of t he values
corresponds Lo an earlier loading age.
 7

Table 1 . Creep Coefficients

l.Jl. tima te C = C . (Ratio of lil tima te Creep Strain to

u Initial 0trai n)

Concre te Avera e Relative Hurni dit

f - - -----,---:r--""-"""T'""- -,:-:,--;: r - - - - - r - - - - '":rc-~- - - - - ;
Stren th 100•~ 70
Ordinary 1 - 2 1.5 - 3 2 -
Hi h 0.7 - l .S 1 - 2.S 1.S - 3.5
h. Effective Section Properties Under Short- Term Loading

The stress distribution and effective moment of inertia of

reinforced concrete beams vary considerably along the length of
tne beam. In region of small moment the concreLe works in
Len .. inn, and the unci·acked transformed section properties are
effective ~n determining stresses and deflections under short-
term loads . In !·egions where the bending moment is greater than
he moment corresponding to flexural cracking) Mer, the concrete
cracks al though tensile concrete oet\•Teen cracks sti 11 cont1 ibutes
signi_icantly to the flexural rjgidity of the beam.

The racked transfo.1.med sec t,ion proper Lies (neglecting all

concrete on the ten ..ion side of Lhe neutral axis) aTe not
unreasonable for u c in calculating stresses in cracked regions
unoer worldrig load , because the governing stresses usually
refer prilnarily to maximu.rr! moment sections. Also, any di..,crep-
ancies c _countered in con1pu ting stresses using the cracked
section propertie are on the high or safe side, and are rs-
flec1.ed, a:. least in part in well tested safety factors . The
ques ion with regard to deflections is serviceability J not
safety: and here it is not generally possible to provide lim:its
of serviceability i' r all t.ypes f structure"'. I n other words,
th ere is more of a premium on bejng able to predict de:lections
accnra,.,e]y, than to compute f ictitous nuJnbers call ed stresses.
Also, defleci.,ion::: are seen ana felt .

The e.ffec l,ive .:lexural r'igidity can vary grea vly along a
c·einl'o!·ced con ere te beam in r· egions of cracld..ng . The ratio of
uncracked o cracked transformed moment of inertia for "low 11
5teel - percen Lage beams i.s often of the order of five and larger.
The effective 10menL o.f ir.erti.a at any section that is cracked
has some value be t,ween the uncracked and cracked r1oments of
lner1;ia) which depends primarily on the magnitude of the moment
or a given l>eam and materials _

An acceptable method in many cases js to stmply use an

average of l,he uncracked and cracked t.rans.:ormed moments of
5
inertia .r r t.he enLire length of beam _ An European Concrete
Committee recommends that the gross - section flexural rigidity
 8

be used for tha t. part of the load that produces firs t cracking
and a modif i ed cracked- t ran sformed- section flexural rigidity for
~he remainder of the load, with the computed deflection not to
exceed the 11 crackecl t ransformed section'' deflection . This
provides a consi deration of l oading stage s but does not account.
for variation s in f l exural rigidity along t he beam. With the
question of loading stages , however, arises the thought t hat the
por tion of t he beam t ha t cracks under maximum load no longer is
uncracked (even under the first i ncrements of reload) upon
reloading .

Since the secti ons being discussed are gross and transformed
concrete sec t i ons, the concrete modulus of elasticity i s , of
course, used in any flexural rigidity (EI) expression .
6
Yu and Winter developed an expression for an average
effective moment of i nertia to t ake into account the participa-
t i on of tensile concrete in resisti ng defl ections . Their
results were stated in t.he fo:lowing form : Multiply (and t hus
reduc e) deflections , computed us ing the cracked transf ormed
secti on propel' ties by the factor

(1 - b ' I\ ) (8)
M
where Ml : 0 . 1 (r • )2/3 (D) (D kd)
C

M = momen t under wor king load::;

b' = width of beam at the tension s i de

D = total depth of the beam

Tl:e ciel'i va tion or t.hi s cx.pressio11 followed an elastic- theory

c:.pproad1 witn Lhe f actor 0 . 1 having been determined eJ1pirical:.v
from heam tests.

The m,Jrrtent. Jvl wa,'": a pu.rr:; bendin~ mon:ent in the deriv,3.t,ion,

5.nd r.he fact,01 0 . ~ ;,ra:'" dete!'lllined 011 the basis that H i : the
rnaxi.rit1111 moment. in the sp~n for t.he bc9.ms teste,J . It doe..:
suffice Lo suggest that Ll1e ef'fective m01nent of inertia at, a
given section might. be olJt,afr1ed by dividing th,:, crac~eo
tran:3formed n:omenL of inerL.;d b~- 3ome J·ac,,•n· ·imilar to ~'l • (13) ,
vihere i' is 1;.h,: man.en-: :at the glven .::ecU,·,n .

l'he modiffr.atinn fact.or given hy Eq . (8) has a similar

efl'ee · 011 ~omputed de.:-=.ec ti,.>n~ a2 the meil10d nf' Murashev7 _·or
Laki 11g i.n to arcou11 c. the part..icipa t::.,:-n <..• f t,8nsi.le concre-;r; in
"i'e.::i-;LiHE; je~'le(: t.,llli!;:. . Thi ~, mt:r,l!od u:-cs the r·raclced 1..?'!J.n::l'urmed
 9

moment, o: iner-cia and an :_ncr eased effectjve steel modulus of

elastici ~y . E. given by Eq. (9) .

E= E/ If ) f ~ 1.0 (9)

wher e }V = l - C (.Mcr/M) 2 and C is a constanL . This method is

based on t,he consideration tha t be Lween cracks the steel stress
and hence deformation is less than right, at the cracks; there -
fore, r,he average effecr,ive steel modulus of elasticity, S,
should be greater than the actual steel modulus, E. at the cracks .
A value for the consLant, C, of 2/3 was recommended .
Specific locations of sections o: first cracking can be
determined by Eq . (10) ,
.1.

Jvicr ::: f ~b I~cr (10)

Yt
where Mer is the moment correspo~ding to flexural cracking,
i'tb i s the modul us of ruptur e I~cr is the moment of inerr,ia
01' the uncracked transformed se ction and Yt i s the distance
f rorn t he neutral axis of the u.n cracked transformed sec-r;i on to
the extreme fiber in tension . For most purposes and most cases!
Eq . (10) can be replaced by -r;he simpler Eq . (11),

~er == ftb Ig (11)

Yt
where Ig i s the moment of i nertia of t he gross concrer.e secti on
alone (neglecting all steel) and Yt refers to t he same gross
concrete section .

There ",Iould be 2 of these Mer- sections in a t:.ypical

-reinforced concrete simple beam W1der service loads . Where
cracking occurs in both positive and negative moment regions,
4 such Mer-sect i ons would exist in ful'ly conti.nuous beams and
3 in beams wiLh only one end continuous . Consideration of the
effects of con Linuous T-beam flanges and beams of varing depths
would affect t he above only in detai :.s . Also, t he effect of
varying r,ensile and compressive s teel percentages along t he
heam would usually be a minor factor in l 0catin~ a given Mer-
.section and would not be involved at all when Eq. (ll) is used .

At. a time when low worldng stre sses were used, it was
deemed satisfactory to use t,he cracked transformed sec t.i on pro-
perties in computing Jeflections . An American Concrete Institute
Deflection Comnd.t tee Ileport8 in 1931 recommended t his for general
use . However , in the last, twenty- f l ve years or so it has become
common practice to use the gross s ection propert i es in com-
put ing deflections under working load3 . The Portland Cement
 10

Associat,ion has recommended this practice for many years .

The new ACI Code 2 contains the same gross- section pro-
vision but modifies it slightly to provide for the use of the
cracked transformed secti on properties when pfy is greater than
500. Tlti8 i::; an attempt to guard cJ.gainst W1derestimating
deflections (using the gross moment of inertia) when high steel
stresses exist, such as where high working steel stresses are
used . or whe11 high yield- point steel is used in ul t:i.ma.te strength
design .

In ultimate strength designs by Whitney's method 9 , a

balanced steel percentage is given by Eq . (12) .

Tu = Cu

As fy = 0 . 85 .r-1
J. C b (0 . 537d)
f Cl
Pbal :: 0.46 (12)
fy
Investigators lO, 11 have felt that a deflection warning should
be sounded when the rati o p for singly-reinforced beams, (p - p')
for doubly-reinforced beams and (pw - Pf) for 1'- beams exceeds
0 .18 fb/f,. This ratio is close to the balanced steel ratio by
elastiic tKeory and less i:.han one- half the balanced design ratio
by ultimate strength theory.

For singly-reinforced beams the marginal steel percentage i s

p = 0 . 18 f~/fy (13)
and ~
P.:...y = 0 .18 f'C = 5ho when fc = 3000 psi .

Hence the ACI value of pfy = 500 is selected for ordinary

strength concrete .

For the cases where pfy is less than 0 .18 fe, , the previous
reasoning calls f~r
the use of the gross section properties.
nowe:ver 1 tile PCA showed that the use of gross- section proper-
ties could be dangerous when steel percentages are low and where
working stresses are relatively high . It follows from the
previous observation that the effect of ::-teel percentage alone
on effective f1 exural rigidity tends to be contradictory .

Tne AASHO3 and others have for a long Lime advocated the use
of the gross- section properties to determine the flexural rigidity
of continuous beams for purposes of L~determinate analysis as
well as for computing deflections . This, admittedly, has been a
rather vague compromise, but one that was dictated by the nature
 11

of the problem. In the case of continuous T-beams (flange

usually cracked in negati ve moment regions) and beams of
va rying depth, an average of the positive and negative moment
section properties is often used in estimating deflections
using conventional formulas for prismatic members .

Since the use of the cracked transformed moment of inertia

tends to overestimate deflections, a reduced modular ratio (such
as n = 8 for all strength concretes recommended by the AASHo1 3
for computing deflections under short-term loads) i s often used
in an attempt to offset the high computed deflec t ions . This
reduced modular ratio has the same effect as t hat provided by
an increased effective concrete mopulus of elasticity . Another
technique that has been suggestedl4 is to reduce the deflectionsj
computed using the cracked transformed moment of inertia, oy t.he
following empirical factors:

Deflection, 6 = 0 .9 A~r for simple beams

= 0.8 ..6.gr for one end continuous (lu)
-= 0.7 A~r for both ends continuous
where Ll ~r is the computed deflection using the cracked trans-
formed moment of i nertia . For continuous beams, the section
properties corresponding to the points of maximum positive and
negative moments are usually used in this method as constant 1 1 s
throughout the regions of positive and negative momentJ
respectively .

The misuse of the cracked transformed section properties

Lends to be more pronounced in conti.riuous beams than in sil'lple
beams, as indicated by the factors in Eq_s, (lL) . A greater
length of beam will norrri.ally be uncracked in continuous beam~
as compared to simple beams (moment gradients are greater in
continuous beams and hence maximum moments drop off faster).
For example, consider the following extreme case : if a uniformly-
loaded, continuous, prismatic reinfo:rced concrete beam with the
same positive and neg~tive moment reinforcement has a crac.t{"'ng
moment. capacity of wL'"/2L, 0.821 or 82% of the span will be un-
cracked. For the same simple beam, but with the load multiplied
by 2/3 to account for the smaller allowable load on the simple
beam (the ratio of the maximum moment3 .:or the two cases), only
0 . 291 or 29% (18% if l..he load were not i-educed) of the span will
be uncracked. However, certain factors such as distribution of
load s, varying section depth, steel percen"tage, etc., can cause
the use of these factors to lead to e1-r0neous results.
 12

S. Shrinkage Warping and Creep Deflection

Concrete s.J,.rinkage induces stre ses in both statically

determinate and indeterminate reinforced concrete struct.ure~ .
In deter.mina.te member,., the shortening of the beam resulting
from shrinkage is resisted by the reinforcing steel, inducing
compressive stresses in the steel aid tensile stresses in the
concrete. The tensile concrete tresse are maximum in the
vicinity of ~1e reinl'orcement and thus combine with ~ensile
stresses re~ulting _rem transverse load~ to cause additional
cracking . Shrinkage of he girders in redundant frames also
induces additional bending moments which are subject to direct
analy""is .

\~1en rein orce~ent is unsymmetrical shrinkage causes a

nommirorm strain distribut,ion which results in warping of the
cross-section . Although shrinkage and creep are undoubtedly
interdependent, the coe ficients defining the magnitude of
these effects are usually expressed separately for practical
p poses . The.e are eY.ceptions to his that are discussed
later in this ,3ection. Eve 'though the effects o_ hrinkage
might be considered (in an approxima e manner) apart from those
of transverse load hrinkage warping is obviou ly affected by
cracking and t herefore by transverse load.

Shrinkage warping formulas have be~n developed for both

uncracked and era.eked sectionsl2t 1.5, 16, 17, in which an
equivalent elastic analysis is employed . In consi dering craclced
sec -ions. however 1 t e effect of load and sh ' inkage must be
co sidered simultaneou lyJ vince the extent of cracking is a
direct .function of the tramn erse load . Since shrinkage warping
rreq ently has only a secondarJ effect and seldom a predominan
effect on total deflections the simpler uncracked section
met.hod L p obably just as adequate as the other method and can
be used wi thou I, regard to effects of Lransverse load .

Co siderin an uncracJ.ced tran ... formed section (either singly

0

or doutly- reinforced beams with or without flanges) tho

warping curvature at any cros3- e~t.i.on due t.o shrinkage is given
by
TSe -~
= i"I =
(15)
4>sh EI Ect Ic
where = warping curvature resulting from ohrinka.ge
(/)sh

~:- Note that Ferguson16 did not include the effects f creep in
t,he expression for EI a doe.., Eq . (15).
 l3

e = distance between the centroids of the uncracked

transformed sectioD (using net= E5 /Ect) and the
steel area

E
ct = sustained modulus of elasticity as defined by Eq.
(19)
= moment of inertia (using net= E5 /Ect) of the
uncracked transfor med secti on

and Ts = (As + A'3 ) Esh Es (16)

where Ts = total compressive force induced in the steel
As tensile steel area

A•s = compressive steel area

t sh = free shrinkage strain

Es = modulus of elasticity of steel

For singly-reinforced beams. A5 = 0. When As, As and e are

essentially constant along the span. the maximum shrinkage
deflection for a simple beam becomes,
2 = Ts e 12 (17)
'F' sh 1ti 8 Ect 1 ct
where 6. is the midspan deflection and L is the span length .

In considering the distribut.ion of shrinkage strains and

corresponding shrinkage warping, creep effects should be :t,nclndedJ
because shrinkage st.resses are sustained stresses . However, the
use of the usual creep factors, for concrete under constant
compressive stress, are rather nebulous, since shrinkage stresses
are variable (increasing at a decreasing rate with time), and
are tensile in nature. Also, the effective concrete modulus of
elasti city of interest here should refer to concrete in tension .
It. is obvious from this discussion that the solutions of
shrinkage warping using quasi- elastic concepts leave much to be
desired. The_y, never theless do provide rough estimates of
shrinkage deflections that can be compared with experimental
data with partial success .

Mi1ler18 has presented an interesting and different approach

to the shrinkage warping problem for singly- reinforced beams
only . His basic assumption is that the extreme fiber of the beam
on the side a.way from the reinforcing steel shrinks the same
amount as the plain concrete (Fergusonl6 disagrees with thi s) .
 l4

Foll owing t his a ssumption, the b eam curvature i s given by

cf>sh = €sh
d
(1 _~i
€ sh
(18)

where fs rus the steel strain and dis the usual eff ective depth
measured from the center of gravity of the steel to the opposite
extreme fiber . Miller suggests empirical values of ~s/c,~ = 0 .1
f or heavily reini'orced members and O. 3 for moderately reinforced
members . This type of simplified empirical approa ch seems to have
merit, and is discussed further in Section S. l .

Time- dependent deflections of reinforced conc~ete flexural

members, resulting solely from effects of sustained load (creep
deflections), are usually greater than, and often two to three
times as great as deflections resulting from all other effects
combined during the life of a structure subjected predominantly
to sustained loads . Thus, creep deflections are of primary
interest and should always be considered in addition to those
re sulting from instantaneous loads and shrinkage .

In addition to the difficulty of computing the creep- time

history of a particular concrete under constant, unifonnly-
distributed sustained stress, a reinforced concrete flexural
member is suoject to a nonuniform stress distribution and very
often a variable- load hisLory . An accurate analysis of the
effects of a variable stress history even for uniformly loaded
specimens , requires creep - time curves and a 'knowledge of the
l oadL~g history The rate-of- creep method1 9 or the super-
2
position method 0 can t hen be used when detailed creep and
loading information are available .

The rate-of-creep method , illustrated in Fig. 1, is s traight

forward . Consider an extreme case i:1 which a concrete specimen
is s ubjected to a compressive stress a- f or a time int erval t1 .
At the end of thi s interval . t he stress is removed completely .

Ac ... ording t·r1 the r a T,e - o f-cree:p tne Lhocl, r.he creep strain a L
time t 1 is (T ~ •,1 . the p.roifoct, of t,he sust.ai..1:ed sl;ress and the
w,it, creep 51.,rai n for the Lime r-on2idered. Once Lhe stress is
removed. there is no further change in c:?:eep s train and at a
Lj_me , saJ 2t1 . the creep strain L: s ti 11 rr J t .
1
The 3Uperpo::ii t.ion method . i:..l 1:.:.t1·atecl :..n Fig. 2 , preciicts
1

!_,Ji& same creep s ....ruin at. c:..me L1 ,.:: f CT cft . However, ra 1,her than
as sumi ng directly t,har, the coinprr!ssi ve st ress is removed c1.t time
Li . it, is assu-rn~d :;hat i:,he specimen is <;1Jbjec tea to an adcii ti onal
st.r ess o.f CT in tensi on and creeps undc"" l.w0. opr,osi:ig fictitious
<:: tresse s , For example, assuming that. Lhc creep characteristi cs
of t-he concrete are I.he same ii: tension a:1d compre ssion and are
indeper:dent. of :r.e concrete age when l,:,aded , the compressive c.~·eep
 15

Time

Fig. 1--Creep strains by the rate of creep method

+ - •
---
r;-
--- - - -
---

Time

Fig. 2--Creep strains by the superposition method

 16

strain a L time 2 t 1 is r;r J 2 t while the tensile creep strain

is q-- of L , sj nee the tensile stress is a new stress applied for
a t ilne interval t 1 . The l;,o Lal compressive creep strain at time
2 t is thus O'" ( r:f 2 t - rJ t,, ) and represents a reduction with
1
respect to Lhe creep sL:rA.in at tilne t1, since ( r:f 2t1 - f t ) is
1
less Lhan J t, (pri1na1·y creep curve i.ncreases at a decreasing
rate with time) .

Usuall y such a detailed anal ysis is not feasible, and a

shorter, more approxilllate method is used . One such method i s
t he sustained-modulus method which refers to concrete under a
constant sustai ned stress . In this case a reduced or effective
modulus cal led the sustained modulus of elasticity is used for
computing initi al-plus- creep deflections .
CTconstant trconstant (19)
Et= € + E = e . .• . l ll + Gt)
c initial creep 1niu1a
where Ect = sustained concrete modulus of elasticity

EC = ordinary concrete modulus of elasticity under

instantaneous load

ct = creep coefficient defined as the ratio of creep

strain to ini t ial strain

When the sustained modulus of elasticity is used with, say

the gross section properties in computing deflections, the
resulting creep deflecti ons are silnply equal to the initial
deflections multi plied by the creep coefficient . It seems
inappropriate however, to use the term flexural rigidity (EI)
or beam stiffness in connection with the sustained modulus of
elasticity, since the effect of creep is to increase deflections
bu t not Lo decrease the bendi ng stiffness of the beam (such as
for additional short-term loads, eLc . ).

Most recommended methods for computing creep deflections

follow some ramification of this approach . Usually the deflec-
tions computed using the gross- section properties are obtained
and creep factors (or deflection factors), which include com-
pressive steel effects, specified . Both shrinkage and creep
deflections tend to be drastically reduced when compressive
steel is used . Only the quasi- elastic method (Eq . 17)~ and not
t he method of Miller (Eq. 18) refer to shrinkage warping for
doubly- reinforced beams .

The CRSI 21 suggests the following method for computing

combined shrinkage and creep deflections : Use the gross
concrete section proper t ies and a shrinkage- plus- creep factor
of J ; that is, the toLal deflection is h times the initial
deflection or Ect = Ec/h . For a compression J teel area equal
 17

to. the tension steel area, use one- half the usual shrinkage-
plus - creep factor or 1.5 for simple beams and one- third the
usual factor or 1 . 0 for continuous beams.

Yu and Winter6 presented an empirical table of such

shrinkage-plus-creep factors for dirferent durations of loading
up to five years . The new ACI Code 2 adopted their 5-year or
11 ultima.te 11 values as follows : "The additional long- time deflec-
tions may be obtained by multiplying the :immediate deflections
caused by the sustained part of the load by 2 . 0 when AJ:, = O;
1. 2 when A!:, = 0. 5 As; and O. 8 when A:l, = As· 11 Typical differ-
ences are seen for such recommended factors by comparing the
CRSI and ACI values of 3 with 2 and 1.5 or 1.0 with· 0 . 8. The
reason for such variation is that o~her factors, such as con-
crete quality, age ,-1hen loaded, loading duration> relative
humidity, etc . . significantly influence time- dependent concrete
defonnations .

Total time-dependent (combined shrinkage and creep) deflec-

tions might be computed simultaneously, with the use of some
combined shrinkage-plus- creep factor, using any method advocated
for computing creep deflections alone . The combination of these
two effects is probabl y satisfactory for broad-approximate design
procedures, but leaves much ~o be desired in analytical work
where reasonably precise results are desired in unusual as well
as typical structures .

In addition to the fact that the strain distribution is

nonuniform i n any flexural member, even though linear> creep
of the reinforced concrete beam seems to have the effect of
moving the neutral axis toward the tension zone . This effect
can be obtained by the use of a cracked transformed section
method where an increased modular ratio (resulting in an
increased effective steel area), is defined by
:: Es (20)
Ect

where n = E 5 /Ec , However> in regions where cracking is limited

or nonexistent, this method tends to lead to computed deflections
that are too large, as does t he use of the cracked transformed
section for short- term loads with the usual modular ration.

6. Serviceability

Deflections of reinforced concrete flexural members should

be controlled so as not to affect adversely t he appearance and
serviceabili~y of a structure . This statement is completely
general but is of primary concern to the design engineer .
 l9

III . DESCRIPTION OF EXPERIMENTAL INVESTIGATION

3.1 Specimens and Instrumentation
The experimental phase of this investigation included
primartly the measurement of instantaneous deflections; time-
dependent deflections; and concrete strains resulting from
elastic shortening, shrinkage and creep . Two silnple- span beams
and two continuous beams (each continuous over two spans) were
the pri ncipal test specimens. One simple (SB- 1) and one
continuous beam (LB-1) were reinforced wi th one #3 bar and the
other simple (SB- 3) and continuous beam (:B- 3) were reinforced
with three #3 bars. All spans were 9 feet (continuous beams,
18 feet long) . Duplicate shri nkage specimens containing one #J
bar, three #3 bars, and also containing no steel were used.
These were placed on their sides on a smooth surface in order
to minimize frictional effects .

The geometry and de~ails of the test beams are shown in

Fig . 3. ~o stirrups were required in the beams of this investi-
gation . The shrinkage specimens were the same size as the
simple beams . The design details of the test beams are shown
in Table A. l.

The slump of the concrete was 1.5 in . , and the 28- day
concrete cylinder sgrength and modulus of elasticity were 5130
p. s.i. and 4.4 x 10· p . s . i . , 1·espectively . The concrete mix
design, per cubic yard of concreteJ was as follows:

Cen1ent (Type I) l.J23 lb

Sand 1335 lb
Stone 1930 lb
Water 20 gal

The ~ensile yield point of Ghe hard grade billet steel rein-
forcement averaged 52 ,000 p.s.i .

A Whi ttemure mechani cal strain gage, shoi,m in Fig . 5,

(ten- inch gage length providing direct readings to 10 x 10- 6
inches per inch) was useo to measure the concrete strains. The
gage points were stainless steel inserts imbedded in the concrete .
Each beam had one gage near the top and one near the bottom on
both sides and at three different locations along the beam, as
shown in Fig . 3. The strain gage points on the shrinkage
specimens were placed in the same locations as those of the
simple beams except on one si de only, since these shrinkage
specimens were placed on their sides. A total of 12 strain
gages ( 2u gage poi..11ts) were used on each simple and continuous
beam and 6 strain gages (12 gage points) used on each shrinkage
specimen . Strai ns resulting .from temperatm·e changes were
 20

..1 ,.

4" 5,,
1 - #3 bar 3 - #3 bar~
. ,..
As 0.11 in( A9 =0.33 inl
0 0 0 O

Notes: 1. These sections inverted (9ame section) in negative

moment regions.
2. No web reinforcement was used.
J. All main reinforcement in continuous beams was cut
off one foot beyond the elastic inflection points
(quarter-points). No bent-up bars were used.

(a) One-bar and three-bar cross-sections

311
4 Dial-Gage Location

3
l"
"2"
1011
}·
- 311 2' - JI' 2' - )I' 2' - 11

L' - 6" 4, - 6n

(b) Simple beam

Dial-Gage Locations
Strain-G J.1"
}?11
T1 11

9' - 0'' 9' - 011

II

t1
'2"
11

( C) Continuous beam

Fig. J--Geometry and details of test beams

 21

Fig. 4-~View of test beams, shrinkage

specimens and instrumentation

Fig. 5--View showing close-up of

Whittemore gage and dial gage
 22

eliminated from all shrinkage and creep data by means of a

control gage having the same thermal coefficient as the concrete .
The inner bar of the Whittemore gage is made of invar metal .

Dial gages were used on both sides of each simple beam at

midspan and at the point of maximum elastic deflection for the
continuous beams . The accuracy of the dial gages (0 . 0001 in.)
for measuring deflect ions provided excellent data for this par t
of the study .

3. 2 Experimental Results

All beams were loaded at age 28 days with the beam dead -
load plus a uniformly distributed superimposed- load . Iron
bricks were used for the additi onal loading. The br icks were
placed continuousl y along t he 3- bar beams and spaced 11-~iformly
along t he 1- bar beams (in the latter case the difference between
t he deflections computed for t he int ermittent - l oad and the equiv-
alent continuous- load was of the order of 1% and was ignor ed in
the study). A superimposed- load to dead- load ratio of 2 . 0 was
used for t he 1- bar beams and 5.5 for t he 3-bar beams. The total
loads resulted i n computed maximum concrete compressive stre sses
tha t were the same for t he corresponding simple and continuous
beams (the 1- bar beams - -also the 3-bar beams) 3 also re s ulted in
computed maximtm1 concrete compressive stresses that were the
same at all points along t he 1 - bar and 3- bar simple beams--also
the same at all points along the 1-bar and 3- bar continuous
beams .
A comprehensive schedule of deflection and strai n measure-
ments was maintained throughout t he test period of 60 days.
Each deflection and strain value reported is an average of the
readings on both sides of t he beam in the same posi t i on . ThusJ
any small effects resulting from warping or accidental eccentri -
cities of loading were compensated for . Also , onl y the average
of the corresponding strain readings on the duplicate shrinkage
speci.mens, t he quarter-point strain gages for the simple beams
and the strain gages located at the points of maximum el astic
deflection for the continuous beams were reported . This pro-
vided a statistical approach for determining eA--perimental values .
The variations were random and not significant . The basic
strain curvature and deflecti on data are s hovm in Figs . A. l
through A.10 .

Additional daLa obtained ir.clude temperatm·e and relative

humidity data. The average ambient tempera t·..rre was 81.i degrees
F. with e):tremes recorded of 79 and 91 degrees F . The aver age
ambien t relative humidity was 59% with extremes recoTded of 32
and 72% . Pict1.rres of the test specimens and instrurnem,ation are
shown ·in Figs. 4 and S.
 23

IV . EFFECTS OF CRACKING ON INSTANTANEOUS DEFLECTIONS OF SIMPLE

AND CONTINUOUS REINFORCED CONCRETE BEAMS

As discussed in Section II, a relatively large number of

methods, guides and rules of thumb have been recommended from
time to time for computing instantaneous and time - dependent
deflections of reinforced concrete flexural members with varying
degrees of success . Conflicting aspects of the existence of a
complex problem and the need for quick, practical design methods
have resulted in an over-emphasis on the latt er. It now seems
e1rlden t that it is probably not p ossible to describe an accept-
able method for predicting deflecti ons that is as brief as
desirable and still includes provisions for all eventualities.

Irrespective of the difficultie s of nol:. lmowing, in advance,

the material properties and time- dependent characteristics of the
particular concrete to be used, it is, nevertheless) of utmost
desirability to prescribe designJT1ethods that incorporate all of
t.he pertinent aspects of the problem. The business of gettmg
concrete that meets specified conditions is largely one of
quality controlj an area that is subject to improvement in
keeping with the demand for such improvement,

Instantaneous deflections are of primary importance in con-

sideri ng deformational behavior of ~einforced concrete beams
under transient live-loads as wel l as in determining initial
deflections under sustained loads . Most practical methods for
c0mputing creep deflections are based on the initial computed
deflections .

Considered in this section are the effecLs of cracking on

deflections of reinforced concrete beams under short- term loads .
fhis requires an evaluation of the effective section properties
along t he beam as influenced by effects of cracking and partici-
pation of tensile concrete between cracks . Since behavior under
repeated loading (not necessarily in the sense of fatigue loading)
shoul d generally be considered, the effective sections a l ong t he
beam under all increments of loading should be taken as those
under the maximum load, or neglecting the effect of loading stages .
That is; t.he portions o.f the beam that have cracked under maximum
load. can no longer be uncracked w1der smaller loads, if heal ing
effects are neglected. Overloa ds would affec t this consideration
but would tend to be offset by t he continued increase in concrete
strength with time. A dis~inction might be made between short-
term live-load deflections, where reloading occurs, and initial
sustained- load deflections such as under dead-load, which may be
applied only once . However, this distinction is probably not
 24

jusLif" ed in most case s and is con~idered of ~econdary impor -

t a nce in the analyses to e discussed. Also oi interec- t i s a
practical me ,hod for in egrating the effects of cracking long
the l ength of the beam in the case of both simple and continuou..,
beams .

u .l Development of an Analytical Met.hod .for Including t he Effec t

of Crack· t1g in the Prediction of Instantaneous De "l ections

In regions of cracking the effective moment o_ ine c.ia ,

Ieff under ins tan.taneous load is l ess than the uncracked trans -
Iormed moment or inertia, I~cr but greater than the cracked
transformed momen•~ o.f i.I:ertia ~ rgr, due to the participa t.ion o.f
-Len ile concrete be ween crack;:; . The actual value of Ieff' at. a
gi ven section depends primarily on the extent. of cracking or the
magnitude of the bending moment M, i n addition to the section
de1.,ails and conc1~ete strength .

One logical form of an expression for Ieff a a given sec -

ti on . that sa tisfi e ~ thE: boundary conditions t..rhen M =- Mc r,
Ief.f = 1-~cr · and when M-;;,,-""? Mer Ieff - I~r) > i s give1, by Eq . (21) .

hlhen M 7 Mer,

Ief:f • I/icr - [ I/icr - I~r ] [ l - ( M:,.)m] (21)

where m i'"' an unlmown power, A pre ede t for a power ..:\mer.ion

rela ,ion relative to the distribution o:' cracking effects was
e t abli,,.hed by Murashev ' s Eq . (9) in a totally different form ,
However, a considerably different va l ue for the power is deter-
mined herein alt.houeh initially ii:. was thought tha t a secon
degree fw1ction was reasonable. as int e case of Eq . (9) .

SL11ce the w1cracked transfo1med moment o.f inertia is u.,ually

only J ightly larger c.han the gros sect:_o. moment of j_nertia ,
th e latter is used in the rema i nder of the discus-ion . ln case
involving heavily reinforced members, it might be desirable to
u -e the uncracked ran s orn1ed section value.

(22)

It is seen that, l,he um of th e two ur·acketed terms is always

equal Lo unity, and . hence . Ieff in Eq . (22) always has :.:;ome
value between Ig and I~l' when M ;::, vJc..r ,
 25

If an acceptable evaluation can be made of the

appropriate value form, Eq. (22) should provide an effec-
tive means for determining the severity of cracking at a
given section under applied moment in a form directly
applicable to the computation of deflections. A study of
Eq. (22) reveals the following weighted values for the two
section properties corresponding to different magnitudes
of moment greater than Mer:

leff = C1Ig + C2I~_i:

M = 1.2 Mer M = 2.0 Mer M = 4.0 Mer
C1 c2 C1 c2 C1 c2

m =l 0.83 0.17 0.50 0.50 0.25 0.75

m =2 0.69 0.31 0.25 0. 75 0.06 0.94
m =3 0.58 0.42 0.13 0.87 0,02 0.98
m 4
;:
0.48 0.52 0.06 0.94 0.00 1.00
rn =5 0.40 0.60 0.03 0.97 0.00 1.00

An exhaustive study was made of the current and other

experimental data involving statically determinate rectangu-
lar and T- beams to determine the appropriate value or values
form, corresponding to the effective moment of inertia at
the individual sections. The Newmark 2 2 numerical procedure
(illustrated in ~ig. 6) was used for this purpose . Results
using m: 4 for both rectangular beams and T-beams are seen
in Table 2, Col. F to agree with test data in all cases
within± 25% and in 65% of the cases within± 10%. Twenty-
three test results were used in the comparison.

In addition, test data for eleven continuous rectangu-

lar beams were compared with the calculated results using
m = 4 . The Ne~mark procedure, as used in these solutions
(illustrated in Fig. 7), provides a method for incorporat -
ing the effects of moment redistribution due to cracking
in statically indeterminate beams. As shown in Table 2,
Col. F, the computed results agree with the test data in
all cases within t 17% and in 70% of the cases within t 10%.

All of the test beams, concrete properties and computa-

tion details referred to are summarized in Tables 3 and 4.

Thus, for determining the effective moments of inertia

at individual sect ions, Eq. (23) is suggested:
 26

For Rectangular Beams and T-Beams

leff = t
Icr (23)

Fol lowing the above evaluation , it was deemed

desirable to attempt to obtain appropriate values for the
power min an expression that could be used as an average
effective moment of inertia for the entire length of a
beam. The general expression provided by Eq . (22) is of
a form that should accommodate such an evaluation, since
it includes both extremes of moment-of-inertia values
along the beam as well as appropriate moment variables.
Since all of the test data invo lved uniformly distributed
loads, other distributions of moment might be expected to
resu lt i n a different evaluat ion of m. In cases involving
heavy concentrated loads, for example, the more general
solution such as that provided in the Newmark numerical
solution with the use of Eq. (23), should be emp loyed.

In effect, the use of Yu and Winter's Eq. (8) along

with the cracked transformed moment of inertia provides
an average effective moment of inertia for an entire length
of beam. However, the empirical constant of 0.1 was based
on test beams that were all rather severely cracked. The
res ults in Table 3, Col. X for beam LB-3 suggest that
Eq. (8) may not apply general l y in cases where beams are
only moderately cracked; a condition that was included in
the evaluations herein .

For determining an average effect ive moment of inertia

over the entire length of a simple reinforced concrete
beam, Eq . (24) was found to be appropriate (see Table 2).

For Rectangular Beams and T-Beams

Avg. Ieff = [ (::J J] lg + [ 1 _ (::: .)

3
J r&r (24)
 27

Because of the way in which these equations are bounded by

reasonably well-established limits (lg and I~r) in addition
to the experimental verifications herein, the use of Eq.(24)
should be acceptable for general use with a considerable de-
gree of confidence. The results of the experimental evaluation
of the powers in Eq. (24) is shown in Table 2, Col. H. These
solutions using Eq. (24) differed from those using the 1>0re ia-
volved mume:c:i.a.ld solutions and Eq. (23) by a maximum of 3'L.
This comparison is shown in Table 2, Col. I.

This short-cut approach for obtaining average effective

moments of inertia for simple beams was found to be applicable
to beams continuous at one end using the following weighted
average for the positive and negative moment regions (see Table
2):

: (25)

Although, the experimental data did not include beams continu-

ous at both ends, it is believed that an acceptable solution
for obtaining an average effective moment of inertia for beams
continuous at both ends is as follows:

Left
End
(26)

I ¾ [Neg.Mom.Avg. Ieff] Right

End
In either case (involving Eqs.(25) or (26) the positive
moment section properties have the dominant influence on deflec-
tions. Results using Eqs.(24) and (25) are shown in Table 2,
Col. H to agree with test data in all cases within l 151..
Eleven test results were used in the comparison. The redundant
moments were determined on the basis of elastic analysis for
prismatic members in these solutions.

4.2 Outline of Computational Procedures

The following procedures are outlin~d for computing instan-

taneous deflections using the previous equations and Bq.(ll);

Simple Beam {Constant Concrete Dimensions)

1. Computed the cracking moment, Mer, using Eq.(11).

 28

2. If the inaximum bending mo111ent under service loads i

less than Mer, use Elg for the flexural rigidity at all point
along the beam in computing the beam deflections.

3. If the maximum moment (including overloads if desired),

Mmax> is greater t:har. Mer, compute values for Ieff uaing Bq. (23)
~ta sufficient number of sections in the cracked regions and
compute the service-load deflections using the moments of iner-
tia thus determined. The conjugate beam method or, preferably,
the Newmark numerical procedure (illuatr ted in Fig. 6) are
well suited for this purpose.

3(a). Sufficient accuracy can usually be obtained with

the use of a constant moment of inertia value deter~ined by
Eq. (24).

Continuouu Beam (Con tant Concrete Dimensions, Including

T-Beama)

1. Compute the cracking moment, Mer, for both po itive

and negative 010ment regions (same value for both except for
T-bea111e, in which case the flange overhangs should be neglected
in computing the negative-moment value) using Eq. (11).

2 . If the maximum bending moment (determined from a prie-

matic beam analysis) under service load • ia less than Her in
both positive and negative wment regions, use Elg for the
flexural rigidity at all point& along the beam in computing
the beam deflection .

3. If the maximum neg tive moment using prismatic beam

analysis (including overloads if desired).~~• is greater
tban Mer, computed values for Ieff using Eq. {23) at a auffi-
cent number of sections in the negative moment region or region.
Do the same thing for the positive moment region. If the maxi-
mum ~oment is less than Mer in only one of the region•, use lg
in that region. Compute the service-load deflectiona u ing
the moments of inertia thus determined and the Newmark numerical
procedure (illustrated in Fig. 7 for a beam continuous at one
end only) whieb includes the effect of moment rodi tribution
due to cracking.

3(a). Sufficient accuracy can u • ually be obtained witb

the uae of a constant moment of inertia value determined by
Eq. (24) and Sqs. (25) or (26).
 29

Continuous Beam (Witb Variable Depthsl

l. Determine values for M, Mer• I , I~, and Ieff in

the Newmark solution (also using 2qs. (fl) a~d (23)) and com-
pute the deflections at the same time. A unique solution can
be found which incorporateB the effects of moment redistribu-
tion resulting from cracking, although a number of trials will
usually be required. A shorter method in this case can easily
lead to erroneous results. However, a ve~y rough short-cut
estimate could be obtained by following the procedure outlined
for constant-dimensioned beams ueing Eqa. (24) and Eqs. (25)
or (26).

In many cases computed deflections using the ordinary

gross-section method will not be greatly different fro111 de-
flections using the numerical procedure. However, the more
extensive method is needed to take into account unusual con-
dition• of proportioning, loading, ete.

The following is a eummary of the boundary conditions,

aaeociated with different cases of statically indeterminate
beams, required in the numerical solution to incorporate tbe
effects of moment redistribution resulting from cracking in
computing deflections of continuous reinforced concrete beame:

l. Single Span Bea~, One End Fixed, One End Pinned

The eolution of tbie problem ia illustrated in Fig. 7.

The procedure applies equally well to uniform and nonuniform
beams (symmetrical or unsy~metrical), with variations in I
properly taken into account for nonuniform beams. The trial
shear diatribution ia required since no boundary condition is
known for shear. '
Boundary Conditions:
V • 1
M: 0 at pinned end
8 • 0 at fixed end
y = 0 at both ends

2. Single Span Beam, Both Buds Fixed

A. Symmetrical Beam (uniform or nonuniform)

Consideration of half of the beam would be con-
venient in the numerical procedure. A trial
DOment distribution is required since no boundary
condition ia known for IDOment, in general. The
procedure would be similar to that of Fig. 7 for
Caee l above, except that the diatribution check
would be made for alope inetead of for deflection.
 Simple Beam
"' L/2
- a I a I_ a I a
De scription I -I I MultiElier
Moment, M 0 o. 4375 0 .7500 0. 9375 1.0000 wL2 /8
Curvature ,t, 0 0. 4375 o. 7500 0. 9375 1.0000 ~,1L2 /8EI
aEquiv. Conc . Value, cp 5 .1250 8.8750 11 .1250 Start 11. 8750 w12a/96EI
,,
Avg. Slope, e 31.0580 25 .9330 17.0625 5.9375 (~r 80)
Deflection y 0 31.0580 56 . 9910 74 .0535 79. 9910 vL 2a 2/96EI
A = 80 wL2 (L/8) 2 = 5wLL. , Exact Answer
96 EI 3Bt""ITT
(a) Example of Ideal Solution for Constant EI Beam Under Uniform Load

0. 4375 0. 7500 0. 9375 1.0000 16 L.00 11 #

l+ 0.732 0. 586 0.549 -:--4
41. 7 25 .0 21. 0 20 . 3 1.Il
0.010, 0. 0300 0.0446 0.0493 16,400/Ec
0 . 135 0.355 0.525 Start 0.582 16 1 400a/12Ec
e 1. 306 1.171 0. 816 0.291 II

y 0 1 . 306 2.L.77 3. 293 3 .584 16,40oa 2/12Ec

? 2
= 3.584 (16,400)( 9 ~8) - (12) = 0 . 20311 as compared to 0 ,206!', computed by t he
(12)C 4 .!.i (10)6
approximate method recommended, and 0.153'' determined experimentally. Thi s example demonstrate s the
worst agreement between computed and measured deflections of thirty-four test results . The other
extreme is illustrated in Fig. 7 . The comparisons are shown in Table 2 .

(b) Example of Solution in Whi ch Cracking is Considered for Beam SB- 3.

(Continued on next page)
a -
<P1 =12 ( o/0 + 10 ~l + ~2 ), etc .

Example above : f5.125 =i;[ 0 + (10)(.4375) + • 7500] = ~(5.2150)

1

b Mer computed using Mer= (f~b I g~/(D/2)

c Using Ieff • [( M~r)4 ] Ig + [ 1 - ( M~r)4 ] I~r

Fig . 6--Example of Newmark22 ntUnerical solution for computing deDections of simple beams
(Beam SB- 3) using an effective moment of inertia at the individual sections.
 Continuous Beam

DescriEtion
1 L
J.
M l tiolier
a .,k a a .1 .. a a a
Distributed Load l.raO 1~00 itoa liOO 1100 1toa 1~00 135 . 2# /f-'-

aEquiv . Conc . Load, Q 1.00 1 . 00 1.00 1.00 1.00 13.~ . 2 a

Start
bAssumed Avg . Shear V 3 . 25 2 . 25 1.25 o. 25 - 0 . 75 -1 . 7>
Follow Procedure Belo~
Trial Deflection . Yt 0 0 .878 2.254 3 . 366 J . 863 3.757 bJ . 180 lJ5 . 2aL./12E~I.;

bThe shear dist ibution must be adjusted until the required linear
deflection- corrections are zero (or small); in which case the
deflection curve has been determined . Also the effects of moment
redistribution due to cracking in statically indeterminate beams
are included in the solution . w
w

aEquiv. Conc . Load . Q 1.00 1.00 1.00 1.00 1.00 135 . 2a

bNew V 3 .16 2 . 16 1.16 - 0 . 84 -l . 8h "
Moment, M -3 , 96 - 0.80 1. 36 2.52 2 .68 1.84 0 1}5. a2
cMcr/M -0 . 623 - 1+ l+ 0 . 979 0 . 920 l+ .-r;
d1 6.L.,f 21.7 41.7 41.7 39 . 8 35 . 0 41.7 - l.:1 2 .
Curvature qi 0 . 1825 0 . 0192 - 0 . 0326 - 0.0633 - 0 . 0766 - 0 . 0441 0 -35.2a /Ee
eEquiv . Conc . Value, <P 0 , 713 0 . 342 - 0 . 370 - 0 . 742 - 0 . 873 - 0 .518 1 5 . 2a3/-23c
A.verage Slope, e 0 - 0,713 -1 . 055 - 0 . 685 0 . 057 0 . 930 1.44 II

(Small)b
Trial Deflection , Yr, 0 0 . 713 1. 768 2.453 2 . 396 1.46 0 . ,)18 135,, . 2a 4/12Ec
Linear Correction 0 - 0 . 00} - 0.006 - 0 . 002 - 0 . 012 - 0, 072 0 .018
Deflection y 0 0 . 710 1 . 762 2 . 444 2.38h 1 . 451 0 II

(Continued on next pa ge)

 b. = 2./.41.i4 ((i~5(t:E{ti~j~ 2) 3 = 0.0548 11 as compared to 0.0550' computed by the
approximate method recommended) and 0 . 056" determined experiJnentally . This example demonstrates
one of the best agreement- between computed and measured deflections of _thirty-four test results.
The comparisons are shown in Table 2 .

aValues for Q can also e easily computed for other than nniform loads.

1.:: er computed using : Mer= (f 6t Ig)/ (D/2)

dUsing Ieff a [ ( M~r)" ] Ig + [ 1 -( ~r) 4] rgr

e l
_ a
o/o -12 (1 ~ + 6 ip
2 ° 2'
- 12 q>2 ) Example above: f . 713 -~~ (.1825) + 6
-2 (. 0192) -
2
(- .0326)] = ~(0 .713)
12
\,,.)
a 4!'
i = - (
l2
~2. + 10 <p1 + ~'t ) etc . Example above: <P 0.742 = 1~ [ .0326 + (10)( .0633) + .0766] = 1~(0.742)

Fig . ?--Example of Newmark 22 numerical solution for computing deflections of continuous beams
(Beam LB - 3) using an effective rno.m ent of inertia .at the individual sections. Effects
of moment redistribution due to cracking are incorporated in the numerical solution.
 TABLE 2, COMPUTED DEFLECTIONS COMPARED WITH TEST DATA

Concrete bcomputed ccomputecl

Compressi ve aMeasured Deflections De lections
Refer- Desig- Strength At Instan- Using Col. D Using Col. D Col. E
ence nation Age-When- taneous Newmark Col. E Short-cut Col. G Col. G
Loaded Deflections Procedure Procedure

psi in. in. in .

A B C D E F G H I

SIMPLE SPAN RECTANGULAR BEAMS

\.,J

Current SB-1 5130 0.041 o.oso 0.82 0 .050 0 . 82 1. 00

V,

Investi- SB-3 5130 0.153 0.203 0. 75 0.206 0 . 74 0. 99

~ation
asha Al,A.4 3630 0 . 53 0 . 61 0.87 0.62 0.85 0 . 98
and ') 3 Bl Bu 3020 0 . 92 0.99 0.93 0 . 99 0.93 1.00
FluckL Cl,CL 291.io L58 1.75 0.90 1.75 0.90 1.00
Dl DL. 2920 0 .47 0.63 0 . 75 o.63 o. 75 1.00
El,El, 2990 2.34 2.07 1.13 2.06 1.14 0.99
A2,A5 3630 0,62 o.63 0.98 0 .64 0.97 0.98
B2 B5 3020 0 . 98 1.02 0 . 96 1 .02 0.96 1.00
C2 C5 2940 1. 71 1.76 0,97 l. 77 0 . 97 0.99
D2,D5 2920 o.56 0 . 64 0.88 0 . 65 0 . 86 0 . 98
E2 ES 2990 2 . 20 2 . 09 1.05 2. 08 1.07 1.01
A3,A6 3630 0.67 0.66 1.02 0 . 66 1.02 1.00
B3,B6 3020 1.04 1.02 1.02 1 . 02 1.02 1.00
CJ,C6 2940 1.88 1.83 1.03 1.82 1.03 1.01
D3,D6 2920 0.70 0 , 66 1.06 0 . 66 1.06 1.00
E3,E6 2990 2. 48 2 .10 1.18 2. 09 1.19 1.01
TABLE 2. (Continued)

A B C D E F G H I

SIMPLE SPAN T-BEAMS

Yu and A-1 3680 1. 34 1.25 1.07 1.24 1.08 1.00
Winter 6 B-1 3880 1.24 1.24 1.00 1. 24 1.00 1.00
C-1 3530 1.19 1.24 0.96 1. 23 o. 97 1.00 C
D-1 3680 1.27 1. 36 0.93 1.36 0.93 1.00
E-1 4260 0.51 0 . 61 0.84 0.60 0.85 1.02
F-1 4260 2.20 2.28 0.97 2.25 0.98 1.01

RECTANGULAR BEAMS CONTINUOUS OVER SINGLE SUPPORT (TWO SPANS)

Current LB-1 5130 0.021 0.021 1.00 0.021 1.00 1.00

Investi- LB-3 5130 0 . 056 0.055 1.02 0.055 1.02 1.00
ation
Washa Xl,X4 3230 0.56 0.63 0.89 0.65 0.86 0.97
and Yl, Y4 3360 0.89 0.97 0.92 0.99 0.90 0 . 98
Fluck 24 Zl,Z4 3300 1.04 1.02 1.02 1.04 1.00 o. 98
X2,XS 3230 0.57 0.65 0,88 0.65 0.88 1.00
Y2 , Y5 3360 0.93 1.01 0.92 1.00 0.93 1. 01
Z2,Z5 3300 1.13 1.03 1.10 1.04 1.09 0 . 99
X3,X6 3230 0 . 62 0 . 64 0 . 97 0.65 0 . 95 0. 98
Y3,Y6 3360 1.00 0.99 1.01 1.01 0. 99 1.00
Z3.Z6 3300 1.20 1.03 1.17 1 .04 1.15 0.99
8Both measured and computed deflections refer t o combined dead-load and superimposed-load.
bsee Figs. 6 and 7 for examples of Newmark22 numerical solution for computing deflections
using effective moments of inertia at individual sections obtained in Eqs. (23)and(24)
ccomputed using Eqs. (25) ~ (26), and (27).
 TABLE 3. LOADS, BEAM DETAILS AND SECTION PROPERTIES FOR TEST BEAMSa
(This table is composed of 6 pages; the 2nd, 4th, and 6th
pages being lateral extensions of the 1st 3rd and 5th
pages, respectively . )

Beam Super- Crack-

Refer - Desig- imposed Dead Span ing Max . Rati o Section De~ails and Properties
ence nation Load Load Mom. Mom .

WSL WDL L M t\ax

M
er b b' t d d' D ar g As p = A'
er-
~ax 12~
;:,
X
~

bd
. 2 in2
#/ft #/ft ft . in- k in-k in in in in in in inL lll % ......,
-....J
A B C D E F G H I J K L M N 0 p Q R

SIMPLE SPAN RECTANUULAR BEAMS

Current SB- 1 h1.6 20.8 9 9 ,0 7.6 1.18 4 4 4 .00 5 hl. 7 0 .11 0. 69

Investi- SB-3 114.4 20.8 9 9.0 16.h .5h9 4 4 h.oo 5 hl. 7 0 . 33 2.07
ation
Wasna Al,A 2 1 97 20 . 227 .J 1 -
.
lJ .L.: -...:.. • ·-- L•
\
l._ ll~C -J • ":.,~':. 1. -,., 1. ;,..:
and 23 Bl,B4 59 48 20 26.h 64 . 2 . 411 6 6 6.19 1.81 8 256 0.62 0.62 1 . 67
Fluck Cl,C4 22 60 20.8 20.3 53 , 2 .382 12 12 4.00 l .00 5 125 0,80 O 80 1.67
Dl , D4 169 60 12 .5 20 . 3 53. 7 ,378 12 12 4 . 00 1.00 5 125 0 . 80 0.80 l. 67
El,E4 0 38 17 . 5 7.4 17 .5 .423 12 12 2.31 o . 69 3 27.0 0.44 1 .59 0 . )44
A2,A5 281 97 20 86.5 227 .381 8 8 10.l2 1 .88 12 1150 1.32 1. 63 0 . 62
B2,B5 59 48 20 26.4 64.2 .411 6 6 6.19 1.81 8 256 0 . 62
1 .67 0.31
C2,C5 22 60 20.8 20 . 3 53. 2 .382 12 12 w.•oo 1.00 5 125 a.Bo 1 .67 0 . 40
D2,D5 169 60 12.5 20 . 3 53 .? .378 12 12 4.00 1.00 5 125 0,80 1 .67 o.4o
E2,E5 0 38 17.5 7.4 17 .5 . 423 12 12 2. 31 o . 69 3 27.0 o.44 1. 59 0 . 22
TABLE 3. (Continued)

Ieff = I~r/(1-b I Ml )
~
Beam where M1 =
Refer- Desig- Section details
ence nation and properties O. l(ft) 2f 3n(D-kd) Ratio
Using- n = 29 x 10b /Ee Ieff by Method B
of Ref. 6 Col. U
kd Tt
-er bAvg . l!rav Col .W
1 eff
in in4 . 4
lil in . in4
A B s T u V w X w
CX)

SrnPLE SPAN RECTANGULAR BEAMS

Current SB-l
Investi- SB-3 1.65 18. 2 22.0 20.9 1.05
o-ation
Washa - . , .~ I
3 .6h 600 630 65;, 0 . 96
and 23 Bl B4 2.13 111 121 122 0 . 99
Fluck Cl,C4 1.54 60 . 0 63.6 65.1 0,98
Dl,D4 1.54 60 . 0 63 .6 65 .1 0.98
El,E4 0 . 90 10 . 9 12 . 2 12 . 0 1.02
A2,A5 3.82 583 614 636 0. 97
B2,B5 2 . .50 108 118 U8 1.00
c2,c5 1.60 59.0 62 . 7 64.0 0 . 98
D2 D5 1.60 59.0 62 . 6 63 . 9 0 ,98
E2,E.5 0 . 93 10 . 8 12 .1 ll.9 1.02
TABLE 3. (Continued)

A B C D E F G H I J K. L M N 0 p g R

AJ.A6 281 97 20 86 . 5 227 .381 8 8 - 10.12 12 1150 1.32 1.63

B3,B6 59 48 20 26.4 64 . 2 .411 6 6 6.19 8 256 0.62 1.67
C3,C6 22 60 20.8 20.3 53 . 2 . 382 12 12 4.00 s 125 0 . 80 1.67
D3,D6 169 60 12.5 20 . 3 53.7 .378 12 12 4.00 s 125 0.80 1.67
E3,E6 0 38 17 .5 7 .4 17 .5 .423 12 12 2.31 3 27 .o 0.44 1.59

SIMPLE SPAN T-BEAMS

Yu A-1 349 91 20 100 264 .379 12 6 2.5 10.2 12 6.82 0.62 0.51
and 1153
Winter 6 B-1 350 91 20 103 265 .389 12. 6 2.5 10.2 l.6 12 6.82 0.62 0.51 9.31
v-l
1153 \0
C-1 348 91 20 98 263 .373 12 6 2.5 10.2 1.6 12 6.82 0.62 0.51 0.62
1153
D-1 682 122 20 92 483 . 190 24 6 2.5 9.7 12 7.83 1.20 0.52
1513
E-1 752 90 14 108 248 .436 12 6 .2 . 5 9.8 12 6.82 0.62 0.51
1153
F-1 198 62 20 35.9 156 .230 12 6 2.0 6.2 8 4.60 0.62 0.84
347

RECTANGULAR BEAMS CONTINUOUS OVER SINGLE CENTER SUPPORT CTWO SPANS)

Current LB-1 41.6 20.8 2- 9.0 7.6 1.18 4 4 - 4 5 41.7 0.11 o. 69
Inveeti- 9 4.3 2. 10 0.11 0.69
gation LB-3 114 .. 4 20.8 2- 9.0 16.4 .549 4 4 - 4 s 41.7 0.33 2.07
9.2 . 935 0 •.33 2.07
TABLE 3 . (Continued)

A B s 'l: u V w X

AJ , A6 4 . 01 566 589 615 0 . 96

B3 B6 2. 54 107 118 117 J . 01
CJ C6 1.68 57 .5 61.3 62 . 2 0 . 99
D3 .D6 1.68 57 .5 61.2 62 . 2 0 . 98
EJ~E6 0 . 95 10 . 7 12 .0 17 . 7 l . 03

SIMPLE SPAN T- BEAMS

Yu A- 1 2.66 392 415 414 1.00
and r..
.f:-
~vinter 0 B-1 2.59 395 420 421 . 00 o.

C- 1 2. 53 395 417 420 LOO

D-1 2.54 683 684 705 0. 97

E- 1 2.59 360 hOl 378 1.06

F- 1 1 . 99 130 131.i luO 0 . 96

RECTANGULAR BEAMS CO ITINUOUS OVER SINGLE CENTER. SUPPORT (TWO SPANS)

Current LB- 1
Investi -
gati on LB- 3 1 . 65 18 . 2 22 . 0 34 .0 20 . 9 L 0.5
1.65 18 . 2 40 . 0 23 . 3 1. 72
TABLE 3 . (Continued)

A B C D E F G H I J K L M N 0 p Q R

,,
Washa Xl XL. lh2 h8 2- 27 .3 117 .6 .232 0 6 - 6 ..19 l . 81 8 256 1.06 2, 86 0.93
and ? 4Y 20 66 . 2 . 4l2 6 .19 l.81 0.62 1.67 0 . 62
Fluck~ Yh 86 60 2- 21.a- 94 . 7 .230 12 12 - 4.00 1 .00 5 125 1.55 3. 22 1.55
20 . 8 53 . 3 . 409 4 . 00 1 . 00 0 . 80 1.67 0 . 80
Zl, Z4 30 38 2- 7 . 77 31. 2 . 2h9 12 12 - 2. 31 o . 69 3 27 . 0 0 . 80 2. 89 1 .00
17 .5 17 .6 . L41 2.31 0 .69 O.L4 1.59 0.44
X2,X5 11-12 h8 ::: - 27 . 3 117 .6 . 232 6 6 6 .19 l.81 8 256 1 . 06 2. 86 0 . 93
20 66 . 2 . 412 6 .19 l . 81 0 . 62 . 67 0 . 31
'£2 Y5 86 60 2- 21. 8 9L. . 7 . 230 12 12 - 4.00 l,00 5 125 1. 55 3. 22 1. 55 +"
20 . 8 53 . 3 .409 ti.oo 1 . 00 0 . 80 1.67 0 . 40 .....
Z2 Z5 30 38 2- 7. 77 31. 2 . 249 l_2 12 - 2 . 31 0 . 69 3 27 . 0 0 . 80 2.89 1.00
17 . 5 17 .6 .h41 2 . 31 0 .69 O.hl.t 1. 59 0 . 22
X3.X6 lu2 h8 2- 27 . 3 117 .6 . 232 6 6 - 6 . 19 l.81 8 256 1.06 2. 86 0 . 93
20 66 . 2 . 412 6.19 l.81 0 . 62 1. 67
YJ,Y6 86 60 2- 21.8 9h . 7 . 230 12 12 - /.i . 00 1.00 5 125 1.55 3.22 1.55
20 . 8 53 _3 . 409 h.oo 0 . 80 1 . 67
Z3,Z6 30 38 2- 7. 77 3 .2 . 249 12 12 - 2. 31 0 . 69 3 27 .0 0 . 80 2. 89 1.00
17 . 5 17 .6 .441 2. 31 0 . 44 1 .59

¾here two numbers appear the top number ref ers to t he maximum negative moment secLion value and the bottom
number to th e maxiJnum positi ve moment section value; excepu Col, 0 or T- beams . In Col . 0 for T- beams, the
top number refers to the distan e from the extreme tension fiber to the centroid of the gross concrete
secti on (neglecting all steel) .
TABLE 3. (Continued)

A B s T u V 'W X

Washa Xl,X4 2.85 160 161 133 168 0.96

and . 24 2.13 111 119 122 0.98
Fluck Yl ., Y4 1.81 98.5 99.0 760 103 0.96
1.54 60.0 64.4 65.7 0.98
Zl,Z4~ 1.03 27.5 27.5 17.4 28.9 0.95
0.90 10.9 12.3 12.0 1.03
X2,X5 2.85 160 161 132 168 0.96
2. SO 108 118 118 1.00
Y2,YS 1.81 98.5 99.0 75.3 103 0.96
1.60 59.0 63.5 64.6 0.98
Z2,Z5 1.03 27.5 27.5 17 .4 28.9 0.95
0.93 10.8 12.3 11.9 1.03
X.3,X6 2.85 160 161 132 168 0.96
2. 54 107 117 117 1.00
Y3,Y6 1.81 98.5 99.0 74.4 103 0.96
1.68 57.5 62.l 62.8 0.99
1.03 27.5 27.5 17 .4 28 .. 9 0.95
0.95 10. 7 12.3 11.8 l 04

bPor •i~le ::~t~::!:: Beans:

2
Avg.I 0 ff • [( ~~) : ] lg I [1 -( ~~) 3
] l~r

cFor Cont:inuous Bea.ma; lav : 3 (Pos.Mo111. Avg . Ieff) I 3 (Neg.Mo• .Avg. Ieff)
 TABLE 4. CO CRETE PROPERTIES AND PARAMETERS FOR TEST BEAlvJ.S

aConcrete Strength Moduli of Elasticiti

Beam Loading a t Age-When- d,b Tensile
b .easured Ee C
Refer - Desig- Sched- Loaded and Co - Mod - l Al Steel
s WSL
ence nation At 28 Days puted ular D Pe - ~
ule Rs
EC Rati o centage ..
WDL 1"1ma....

age when f Cl at f~bat f' a Ec at Ee at E Cat

loaded age age c28 age 28 age
w.en wher. days when days when n p
loaded loaded loaded loaded
Days psi psi psi psi,. psi,, psi %
X 10° X lQD X J.06 ...
l.,J
A B C D E F G H I J K L M N 0

SIMPLE SPAN RECTANGULAR BEAMS

Current SB-l 28 5130 539 5130 4.4 4.4 4 .1 7 22 0 0.69 2.0 1.18
Investi- SB-3 28 5130 539 5130 Li. 4 4.4 4.1 7 22 0 2 . 07 5.5 .549
Wat.ion
Washa Al.PL ih 3630 452 ho8o 3.0 3.3 3.5 8 20 1.0 1.63 -.
? 0; .381
and Bl:B4 14 3020 413 3420 2. 7 3.1 3. 2 9 30 1.0 1.67 1.2 .411
Fluc1c 23 Cl C4 14 2940 ho6 3290 2.7 2.9 3 .1 9 50 1.0 1.67 0.4 . 382
Dl D4 14 2920 405 3530 2.6 2. 8 3.1 9 30 1 ,0 1.67 2.8 . J78
El.Eu 14 2990 410 3660 2.7 3. 0 3.2 9 70 1 . 0~ 1.59 0 . l..i.23
A2 A5 14 3630 452 4080 3. 0 3.3 3. 5 8 20 o.s 1.63 2. 9 .381
B2,B,5 14 3020 413 3420 2.7 3.1 3.2 9 30 0 , 5 1.67 1,2 .411
c2.c5 14 2940 406 3290 2. 7 2.9 3.1 9 50 0.5 1.67 0.4 .382
D2,D5 14 2920 405 3530 2.6 2.8 3.1 9 30 0.5 1.67 2.8 . 378
E2 .E5 14 2990 hlo 3660 7,7 3.0 3. 2 Q
; 70 0.5 1.59 0 . 423
TABLE 4. (Continued)

.ri. 3 D E F Ci tl I J 1 '-'1 ,1

A3,A6 lh 3630 452 4080 3.0 3. 3 3. 5 8 20 0 1.63 2.9 . 381

BJ B6 14 3020 hl3 3h20 2.7 3.1 3. 2 0
., 30 0 1.67 1. 2 .411
CJ C6 14 2940 406 3290 2.7 2.9 3.1 9 50 0 1.67 0 .4 . 382
D3,D6 14 2520 405 3530 2. 7 2. 8 3.1 9 30 0 1.6? 2.8 . 378
E,J,E6 1h 2990 410 3660 2.6 3.0 3.2 9 70 0 1.59 0 . 423
SIMPLE SP AN T- BEAMS

Yu A-1 30· 3680 455 3680 3.1, 2.6 ) .1~2 .6 3.5 9 20 0 0. 51 J.8 . 379
and 6 B-1 29 3880 467 3880 J .1,2 .5 3 .1~2 .5 3.6 9 20 0.5 0. 51 3.8 .389 ~
l:'-
Winter C-1 28 3530 445 3530 3.1,2 .5 3.1 2.5 3 .u 9 20 LO 0 .51 3. 8 . 373
D-1 31 3680 455 3680 3.1,2 .6 3.1 2.6 3.5 a
,; 20 0 o .52 5.6 . 190
E-1 29 4260 h5o h260 3.1 2.6 3.1 2.6 3.8 9 lh 0 0 .51 8. 4 .l 36
F-1 34 4260 490 4260 3.1,2 .6 3.1,2.6 J.8 9 30 0 0 . 84 3. 2 . 230
RECTANG !ill BEAMS CON_ IN ·ous OVER SINGLE CENTER SUPPOR.T

Current LB-1 28 5130 539 5130 h.4 4.h 4.1 7 22 0 0.69 2. 0 1.18
Investi- 0 0.69 2.10
gation LB- 3 28 5130 539 5130 l.i.4 4.4 4.1 7 22 0 2.07 5.5 .549
0 2. 07 . 975
·, asha xi.xi 1L .;.2 30 2.6 ., .L -
.) .
. 9 30 0 . 5' 2.86 J.O . 232
and 1.0 1.67 . 412
Fl ck 2L. Yl Y4 14 3360 435 3990 2. 9 J .L 3. 4 9 50 1.0 3. 22 1.4 . 230
1.0 1.67 . 409
Zl ZG 14 3300 LJl 3760 2.9 .I • J 3.3 .,Q 70 1.3 ·2 .89 0 .8 . 249
1.0 l.~9 .441
TABLE h . (Continued)
,.,
.\ B u D E F G I ~1 K L N 0

X2:XS 14 3230 426 3680 2. 8 3.4 3 . -,.,) 9 30 O.? 2.86 3 .0 . 232

0.5 1. 67 _4- 2
Y2,Y5 lh 3360 LJ5 3990 2. 9 3. L 3.L 9 50 1.0 3 . 22 1. . 230
0 .5 1.67 .h09
Z2 . Z5 llt 3300 431 3760 2. 9 .3 , .,)") J.3 9 70 1 . 3 2.89 0.8 . 2h9
o.5 1.59 .441
XJ,X6 14 3l'.30 1.126 3680 2. 8 3 .h 3. 3 9 30 0 . 9 2. 86 3. 0 . 232
0 1.67 . 412
YJ . Y6 lu 3360 ltJS 3990 2.9 3. ~ 3.4 ,,0 so 1.0 3. 22 1.1.i . 230
0 1.67 . h09 .i:--
Z3 , Z6 lh 3300 uJ1 3760 2.9 3. 3 3.3 q 70 1.3 2.89 0.8 . 249 VI

0 .59 ,441
aAll con ret.e compressi e st.reng hs de ermined by 6 11 by - 2" cylinder t,e -t s .

b
The modulus of r p~ure, -cb
f'i •
was computed us1..ng _f I
cb = 7.5 Ji~
ValLLes in Cols . G and H efer to : Secant value at 0.h5 f~ for Re erences 23 and 2h ;
t l, e ini t ial t.angen t modulus f or " e current i nvestigation; and t he initial angen
modu1us a .d secan va1 e a .i.u 0 . 5 le:
·•t respect.:.vely ~ for Reference 6. The measured
init" a- tangent value for Ec at the age when Joaded was used in all cal cul at.ions,
except where tr.is ..ras not obtained 1 i.n whic! a""e the computed value for Ee at the
age when loaded was used .

cComputed values of Ee de t.ermined using Ee = 57 700 ~ , whe1 e f~ is the concrete

co:npressive strength a r.he age when loaded .
~

dModular ra io determined (and rounded off) using n = 29 x 10°/Ec,

 46
V. DISCU~SION OF T3ST RESULTS
The experimental phase of this investigation was undertaken
in order to evaluate the effects of certain variables heretofore
not clearly distjngnished . Relatively high quality concrete
beams of moderate span-depth ratios and loaded so that moderate
cracking occurred provided a useful disLinction from most of the
other deflection tests t.ha "L have been reported _; which were o.f
average concrete qua)ity average - to - large span- depth rat~os
(up to 70 which .:..s abnormally la:.ge) and severely cracked (see
Tables 3 ad 4) . Also, the test. beams herein were carefully
designed with different steel percentages so that the computed
maximum concrete compressive stre sses were the same for the
corresponding simple and continuous beam (the 1-bar beams --
also the 3- bar beams) · also that the computed maximum concrete
compressive stresses were the same at aJ1 points along the
_-bar and 3-bar simple beams -- also the same at all points
along the 1 - bar and 3-bar continuous beams. Compression steel
was not included as a variable in the current experimental
program.

5.1 Shrinkage Warping

Primary interest wi,:,h regard to analytical methods for com-

puting shrinkage warping centers around the basic assumptions and
hence t,he per t inent variables involved . For example, the quasi
ela_,tic "tensile force" method given b,v Eq . (16) includes a
flexural rigidity expression not found i n Mi.ller 1 s method given
by Eq. (18) .
T e
=-
s where T s = (As +AJ.)EhE
.:. 8 S (16)
</>sh Ee~ 1 ct

:=
c~h (1 - Es/ 6. sh) (18)
lp c'r. cl

The latter equation resul~s in a warping expression as a function

of t he free shrinkage effective depth and a constant (paren-
thesis) which was specified in a general way to be 0 . 9 for heavily
reinforced members and 0 . 7 .for moderat.ely rein.forced members.
The method is applicable to singly-reinforced beams only whereas
Eq . (16) is applicable to bot1 singly - and doubly- reinforced beams .
Basic to Miller's approach is the assump t ion that a concrete member
restrained at, some point outside the kern limit on one side~ will
not shrink more (but rather will undergo an equal shrinkage) than
the fr ee shrinkage on the opposite extreme fiber , as the tensile
force method of Eq. (16) predicts.
 47

The curvea of the currant inveatigation abown in Pig. 8

indic~te that tbe extra e fiber doea abrink more than th free
abrinkage of tba co• panion • pecimen, but not auch more. Hence
the effect • of the eccentric • teal re1i1tanc1, out 1d the kern
li11it of th section, do aee to produce "greater than fraa"
• brinkage of the oppo • lte extreme fiber. But Killer'• pproacb
would certainly appear to be a cloee approxieation. Of cour e,
in deeper beam• (greater eccentricity) the a• • u• ption would t ud
to be further i'o error, but in tbeae ca•• tbe increaa d depth
greatly reduce• the ahrinkaga-warping curvature• nyway.

The current and other ahrinkag~ data have been tabulated

in Tables 5 and 6 and the reeult compared with the following
procedure• for computing ahrlnkage w rping:

Eq. (16) ia modified to uae the simpler expreasiona

(Ec/2) (lg) in place of !er; let and eg which refer• to the
gross section. This Eq. (27) is applicable to both aingly-
and doub y-reinforced. bea e. Cloeer agreement with teat
re ulte was found aa a result of thi.a canvenient modiflcation.

{27)

Miller' e Eq. (18) ia applicable. on°ly to singly-reinforced

beam.

The following new empirical expreHions, which provide the

closest agreement with te8t re•ult•~ re introduced. Bq1. (28)
and (29) are applicable to both aingly- and doubly-reinforced
bea1111. (~
A-.. (O. 7) €0 ,h (p-p 1 ) l/ 3 P- '· 112 , for (p-p 1 ) '€ 3 .o~ (28)
'// •b :

and

</J b = f.!.g , for (p•p') > 3.0't (29)

D
For lingly-reinforced beams, p 1 = 0, and Eq. (28) reduce • to

(0.7) £..!h pl/3 (30)

D

With regard to comparisons with 16 test results, the fol-

lowing agreements were found and are sbown in Cola. K, N, and P
in Table 6 :
Using Eq. (27) Reaults agreed with test data in 25? of the
cases within 1O~.
 48

Using Eq. (18) Results agreed with test data lo 23~ of the
cases within 101.
Using Eqs. (28).(29),(30) Results agreed with te1t data in 69-X,
of the cases within lOt.
Keeping in mind the nature of the problem, the latter agreement
is thought to be reasonably good.

Eq. (28) is an adaption of Miller's approach. For example.

his method results in the following expreasiona for aingly-
reinforced beams only:

</) = 0.7 for moderately reinforced beams .

ah

A... • 0.9 €sh for heavily reinforced beams.

'I" sh d

Eq. (30) for singly- reinforced beams re1ults in the following:

¢;sh = 0.56 Csh/D when (p - p') = 0 .5

= • 70 II
1.0
.. . 80 II
l.5
= .88 It
2.0
= • 96 2.5
• 1.01
,," 3.0

The use of the 110re convenient overall depth D instead

of the effective depth d was found to provide cloaer agreement
with the data. The difference is negligible for all but shallow
beams and for these, the use of D seemed to provide the beat
fit. It is , of course, assumed that abnormal covers (abnormal
differences in D ~nd d) are excluded from consideration.

Eqs. (28) and (29) refer to both singly- and doubly-

reinforced beams. The expression in the last parenthesis
of Eq. (28),

(31)

was found to be required in order to produce a somewhat smaller

curvature for doubly-reinforced members than for singly-reinforced
members when (p - p') for the doubly-reinforced members is equal
top for the singly-reinforced members; other conditions being
the same . It is seen that the modifier of Eq. (31) becomes unity
when p' • O. Eqs, (28) ,(29), and (30) provide very simple ex-
pressions for computing shrinkage warping in terms of only two
section properties (D and p or (p - p')) and the free shrinkage
~ sh· However, the data in Tables 5 and 6 tend to indicate
that the methods discussed should be used with caution when
dealing with high-strength concrete.
 49

It abould be aentioned that consideration has not been given

to effect• of cracking on shrinkage warping in either the experi-
lD8ntal atudiea of the current investigation and other• reported
in the literature or in the analytical methods di1cu11ed, At
leaat acco rding to the tensile force method, cracking would tend
to incTea1e the eccentricity of the te~•ile 1teel in aingly-rein-
forcad beam1 and would therefo re • eem to increase shrinkage warp-
ina. However, according to tbe other approach discussed, effecta
of cracking abquld play a minor role in producing shrinkage warp-
ing since the extre• e fiber is atill aasumed to shrink an amount
equal to the free shrinkage, and the resistance factor provided by
the eapirical constant (0.7) and the steel percentage term or term•
would not 1eem to be much different in the ca1e of warping of crack-
ed aec tions.

Witb regard to shrinkage deflection, of continuous beams, if

the effect of 110ment redistribution resulting from shrinkage cur-
vature• are neglected, the effecta of shrinkage on deflection• can
be deter• ined using any moment-ar ea technique or numerical proced-
ure and the curvature expre11iona discussed herein (by 1ubetituting
tbe curvature ¢> for M/EI). Rqs. (18), (27), (28), (29), and (30) all
define abrinkage cuntaturea at individual aections, although the1e
e.xprea1ions are usually con1tant for a considerable length of a
reinforced concrete beu.

3.2 Deformational Behavior of Test Beame

In addition to the shrinkage strain and curvature data for the

abrinkage specimens • hown in Figs. A.2 and A.3, the total (instan-
taneoua plu • time-dependent) and instantaneous plus creep strain
data are ahown in Fig1. A,4 through A,7 . Since the curves have
markedly "leveled off", and willh the additional information ehown
in Pig. 9 for projecting 2-inonth values to 20-yea-c or "ultimate"
values, certain quantitative aa well as qualitative concluaione
can be drawn with regard to ultimate deformational behavior.

In Pige, A.6 and A,7, the tension-gage strains are seen to

decrea1e with time in ca1e• where shrinkage etraine exceed the
creep atraina . The basic curvature and deflection data for the
te1t beam• are abown in Fige, A.8, A.9, and A. 10, and further
repreeented in Fig. 11 and Table 7. The te1ting period reported
for the beams of tbi1 investigation waa 2 monthe .

Average valuea for the creep coefficients (defined as rat io

of creep strain to i nitial atrain) shown in Fig. 10 were virtually
the 1ua f or the ten • ion and compresaion gages, although the great-
er variation wae obaerved for the tension gagea . Thia was probably
due to the random cracking at eome of the gage locations
 50

on the t en sion s i de of the beams. The average values for the

time-dependent (shrinkage plus creep) deflection coefficients
(defined as ratio of time- dependent deflection to initial
deflection) are shown in Fig . 11.

At 2 months the average tensile and compressive creep coef-

f icient was about 0 . 9 while the average time-dependent deflection
coefficient was about 1.5. Projecting these values to 20- year
values using Fig. 9 (multiplying by 2) results in corresponding
coefficients of 1. 8 and 3.0 re spectively , Results in Table 7
indicate that shrinkage curvatures varied from 11% to 19% of the
total time- dependent curvature, so that the corresponding average
creep deflection coel'ficient ( defined as ratio of creep deflection
to initial deflection) would be about 2 ,6 . By comparing t.he
ultimate creep strain coefficient of 1.8 with the ultimate creep
deflection coefficient of 2 . 6, it i s suggested that other effects
seem to have a definite influence on so called creep deflections
other than direct concrete creep strains . Undoubtedly one of t he
principal explanations is that of a shifting neutral axis and
time - dependent adjustments in the stress as well as strain dis-
tributions along the beam . Thi s is also discussed wi th regard
to ~he experimental curvatures obtained .

For relatively high strengt h concrete, l oads applied at age

28- days (considered an average loading age -- not particularly
early or late), and 59% average relat ive humidi ty t he value for
the ultimat e creep coefficient given in Table 1 is about 2 . 5.

Thus suggested in the previous paragraphs i s the nature of

the theoretical as well as empirical vagueness of the approaches
available for applying creep or shrinkage plus creep coefficients
to instantaneous deflections when computing creep or shrinkage
plus creep def l e ctions .

The effects of cracking on insta1:itaneous deflections were

studied in Sect.ion IV and are further evident with regard to
t ime-dependent deflection s in Fig . ~. 10. For example, the
maximum moment for the simple beam SB- J was about twice the
moment corresponding to .first cracking, while the simple beam
SB-1 was uncracked. However, the time-dependent deflection
coefficients at 2 months were 0.146/0.153 = 0 . 95 for SB- J and
0 . 0435/0.0410 = 1.06 f or SB- 1, indicating that extent of
cracking does not seen1 to mat erially affect one's choice of
time - dependent deflection coefficients .

Tabulated in Table 7 are t he instantaneous curvatures .

and curvatures at t he end of the testing period f or all of the
gage locati ons . These curvatures were obtained by dividing the
algebraic difference in the top and bottom gage readings by the
distance between them at each gage l ocation .
 51

From Table 7, Cols. D and E, it can be seen that even though

the desi gn stresses for the 1-bar beams were the same, the ratio
of experimental instantaneous curvatures to moment for the 3-bar
beams were of t he order of twice that of the l - bar beams, which
were 0ubjected to the correGpondingly smaller loads. This demon-
str ates ·the tendency for relat ively large steel-percentage beams
to undergo considerably greater curvatures and deflections when
designed for the same allowable stresses by elastic theory.
Similar behavio_r is seen for t he instantaneous pl us creep curva-
tures in Table 7, Cols. Land M but to a slightly lesser degree .

Interesting results are shown in Table 7, Cols.Hand I where

in every case t he ratios of time- dependent to initial curvatures
are larger in the smaller moment regi~ns. The same is true for
the c-:r eep rati.os (with one exception i n eight cases -- and it
thought to be insignificant) of Table 7, Cols . N and O. This
would suggest that in regions of higher moment (within working
stress ranges -- that is, below any high overload range) larger
early creep strains tend to cause greater reductions in concrete
stresses with accompanying greater reductions in creep curva-
tures with time. Involved is the phenomenon of the shifting
neutral axis with time as a result of the shrinkage and creep
behavior of a nonhomogeneous (particularl y so when cracked),
composite steel- concrete structural member,

The brief discussion of this section serves only to demon-

strate a number of fundamental phenomena-regarding instantaneous
and time -dependent characteristics of reinforced concrete beams
as observed in a limited munber of test results . Methods for
computing defl ecti ons that take into account most of these
effects have been discussed in thi-s report and in the case of
cracking effects and shrinka ge warp::j.ng, new procedures set forth.
It appears that the gap between fundamental answers related to
deformational behavior of such beams and empirical approaches
for controlling structural defl ections remains a formidable but
not impossible one to materially close in the not too distant
future.
 52

Q)
'°0
l

;a° rl 100
~ ><
A
•M ;t
l! . . s::.___
Cl)
200
A
Q)
+:i
I)
•r-t

Ol
.
~ ro
Cl
-~ 300
0 M
(.) +>
CJ)
4000 )

Concrete Age in Days (initial

readings taken at age 4 days)

0 A--Shrk. Spec. With No Steel (pcO), All Gages Used

0 B--Shrk. Spec. With One Bar {p•0.69%), All Gages Used
'P C--Shrk. Spec.' With Three Bars (p•2.07%), All Gages Used

Fig. 8--Comparison of shrinkage strains at the top fiber

for the specimens with different steel percentages
(strains proportioned to extreme fiber using a
linear distribution with the top and bottom gages)
 53

60 ~
~
so L---

40
_v v-
V
30 /
/v
20

10 /
0I
V ~u 4( C lJ
0
Ti.me in Days

Based. on "Long-Ti.me Creep and Shrinkage Tests of

Plain and Reinforced Concrete," by Troxell,
Rapttael and Davis~ Proceedings ASfM, V. SB, 1958

Fig. 9--Average rate of increase for

shrinkage and creep strains
 54

2.0

~
C> ~
bl) Q,)
cd 'rl
C) 0
..-1
§~ 1.0
"M Q)

:8
()
M p.
ii
o J-4
c.> c.,

0
0 20
Time in Days (time zero taken at
age 28 days--age beams were loaded)

2.0

~Q)
Q,) o.-1
:;f~
~ti
~ Q)
1.0
0 0
,-i
II)
u
C: p.
C> Q)
E-1 Q)
b Lower
0
0 20 40
Time in Days· ( time zero taken at
age 28 days--age beams were loaded)

Creep Coefficients Defined as Ratio of

Creep Strain to Initial Strain

Fig. 10--Compression and tension gage creep coefficient

versus time curves for four test beams
 55

0 20 40 60
Time in Days (time zero taken at
age 28 days--age beams were loaded)

Ti.me-Depen~ent Deflection Coefficient Defined as Ratio

of Time-Dependent Deflection to Initial Deflection

Fig. ll-- Time-dependent deflection coefficient

versus ti.me curves for four test beams
 TABLE 5. BEAM DETAILS AND CONCRETE PROPERTIES FOR SHRINKAGE SPECIMENS

Age
Beam Initial
Refer- Desig- Readings Duration
ence nation Section Details and Pro;eerties Sean Taken of ~est
eg I As Al aE,
d D ~ _,
T
b fj 5

in in4 in 2 2
in in in- in % % ft dazs months
A B C D E F G H I J K L M N

Current
Investi-
B- 1
B- 3
4
4
h
4
>
5
1.5
L.5
hl.7
41.7
0.11
0.33
0
0
0.69
2.07
0
0
9
9
L
h
3
3 V,

a.tion 0-,

3eries 1. 2
Mil er18 3n 3.25 2 . 25 3 o. 75 7. 3 0.22 0 3.01 0 3.5 1 3
u" 3.25 3.25 4 1 . 25 17 .4 0.22 0 2.08 0 3.5 1 3
5" 3 . 25 4. 25 5 1.75 33 . 9 0 . 22 0 1.60 0 3.5 1 3
6' 3.25 5.25 6 2.25 58 . 5 0 . 22 0 1.29 0 3.5 1 3
Series 3-~
"')fl
..J 3.25 2 . 25 3 0 . 75 7. 3 0 .11 0 1.50 0 3.5 1 3
h" 3.25 3 . 25 4 1. 25 17. h 0 .11 0 l . Oh 0 3 .5 l 3
511 3.25 4.25 5 1. 75 33 . 9 0 . 11 0 0,80 0 3 .5 1 3
'I I
3. _5 5.25 6 2.25 58 . 5 0.11 0 0 . 65 0 3.5 1 3
Wa ha c2 .c5 12 !.t. 00 5 0.50 125 . 80 1.,; .ho 1.67 '" .8£i 20 . " 1h 30
and D2.D.5 12 4. 00 s a.so 125 0 . 80 0 . 40 1.67 0.84 12 .5 14 JO
Fluck 2 3 E2,E.5 12 2 . 31 3 0 . 27 27 . 0 O.u4 0 . 22 1.59 0 . 80 17 ,.5 14 30
C3.C6 12 4.00 5 1 .5 125 0 . 80 0 1.67 0 20 . 8 14 30
D3,D6 12 4 . 00 5 1.5 125 0.80 0 1.67 0 12 .5 14 30
E3,E6 12 2 .;l 3 0 . 81 27 . 0 0 . 41.i 0 1.59 0 17 . S 14 30

a p , [ A~ ) 100, p' =( A~ ) 100

6d
TABLE r; _ (Continued)

Concrete Properties

Relative Humidi ty Measured

Free
Beam Strength Modulus Of Shrinkage
Refer - Desig- At Elasticity At End
ence nation Extremes Avg . 28 Da;y:s At 28 days Of Test
f 1c bE sh
C

{ % ESi psi x 10b inLin x 10 -

A B 0 p Q R s
Current B-1 32 -72 59 5130 4.1 245 1.11
.....,
Investi- B- 3 32- 72 59 5130 4.1 245
gation
3eri es 1. 2
Miller 18 3" 50 50 3500 3.4 650
4"
5,
so so 3500 3.4 650
50 50 3500 3. 4 650
6" so so 3500 3.4 650
Series 3.4
)1' 50 so 3500 3. 4 550
4" so so 3500 3.4 550
5" 50 so 3500 3.4 550
6" 50 50 3500 3.4 550
Washa C2,C5 20-80 50 3290 3.3 750
and D2,D5 20- 80 50 3530 3.4 750
Fluck23 E2,ES 20-80 so 3660 3.5 750
C3 C6 20-80 50 3290 3.3 750
DJ,D6 20-80 50 3530 3.4 750
EJ .E6 20- 80 50 3660 3.5 750
bcomputed using Ee = 57 700 {ff
 TABLE 6. COMPUTED SlffiINKAGE WARPING COMPARED WITH TEST DATA
Beam Concrete Over-
Refer - Desig - Strength Span all
ence nation At 28 Days . Depth
f C L D E P'
esi ft in % %
A B C D E F G

Curre. i:. B-1 5130 0

/ 5 o.69 0
Investi- B- 3 5130 9 5 2 . 07 0
gation
Series 1,2
Millerl8 ")!
.) 3500 3.5 3 3 . 01 0
411 3500 3.5 4 2 . 08 0
V,
c»
511 3500 3.5 5 1.60 0
6" 3500 3.5 6 1 . 29 0
Series 3 4
3' 3500 3. 5 3 1.50 0
4· 3500 3.5 4 1. 04 0
5' 3500 3.5 5 0.80 0
6" 3500 3.5 6 0 .65 0
Washa c2,c5 3290 20 .8 i:'
,/ 1.67 o.Bh
and D2,D5 3530 12.5 5 1.67 0 . 84
Fluck 23 E2,E5 3660 17 .S 3 1.59 0.80
CJ,C6 3290 20.8 5 1.67 0
DJ D6 3530 12,5 5 1 . 67 0
E3.E6 3660 17 . .5 3 1.59 0
TABLE 6. (Continued)
Experimental Valuee Coweuted Deflections And Conpariaona
Average els/ €eh
Beam Curvature aMidspan b,a Selected d,a e,a
Refer- Deaig- Along Deflec- Using Col.I For Use Using Col.I Using Col.I
ence nation Beam tioo Eq. (27) Col.J In Eq. (18) Eq. (18) Col.M Kq. (28), Col.O
(29)

1 X 10- 6 in in in in
in
A B H I J K L M N 0 p

Current B-1 9 0.013 0 .. 020 0 43 0.4 0.053 0.25 0.043 0.30

Investi- B-3 20 0.029 0.060 0.33 0.2 0.071 0.41 0.064 0.45
atio.n.
Series 1,2 V1
Millerl8 3'' 225 0.050 0.055 0.91 0.1 0.064 0.78 0.049 l.02 '°
4" 1.50 0.033 0.038 0.87 0 2 0 .039 0.85 0.032 1.03
5'' 108 0.024 0.028 0.86 0.3 0.026 0.92 0.024 1.00
6" 80 0.018 0.021 O.S6 0.3 0.019 0.95 0.018 1.00
Series 314
l" 130 0.029 0.023 1.26 0.3 0.042 0.69 0.032 o.91
4rr 88 0.019 0.016 1.19 0.3 0.026 0.73 0.022 0.86
5t• 66 0.015 0.012 1.25 0 3 0.023 0.65 0.016 0.94
6'' 57 0.013 0.009 1.44 0.4 0 .015 0.87 0.012 1.08
Wa•ha C2,C5 0. 50 0.49 1.02 0.54 0.93
.aud D2,D5 0.20 0-17 1.18 0.20 1.00
Plu.ck23 B2,R5 0.45 0.45 1.00 0 .. 63 0.71
C3,C6 1.20 0.98 1.22 0 .3 1.00 1.20 0.97 1.24
D3,D6 0.35 o.34 1.03 0.3 0.36 0.97 0.35 1.00
E3 1 B6 1.20 0.90 1.33 0.3 o. 72 1.67 1.13 1.06
TABLE 6. (Continued)

aoeflections determined using A = (/:> L2

when curvatures and not deflections were reported; also used to
8
colllpute deflections from curvatures in Cols. J, M, and 0.

bEq. (27), A, , where Ts - €. sh

'f' sh

cusing Miller's suggested value a of Es/ E sb - .3 for moderately reinforced members and f.. s/ Esh = .l for
heavily reinforced members.

CJ\
~q. (18) • Applies only to singly-reinforced beams. 0

eEq. (28) <P = (0. 7) ~h (p-p') l/3 ( ~ l / 2 , when (p - pf) <

:::. 3 • 07.
sh D

Eq. (29), 4> - €Dsh , when (p -p') '7 3.01.

sh
TABLE 7. BEAM MOMENTS AND EXPERIBENTAL CURVATURES AT ALL GAGE LOCATIONS FOR THE TEsr BEAMS OF THE
CURRENT INVESTIGA'r ION.

~ax. Mom . (At ~om. At¼- c,drnstantaneo~s

Midspan of Point of Simple Beam Curvatures Under
Simple Beams Beams and At Dead-Load Plus Super-
a And At Center Point of Max, imposed-Load.
Beam Support of Elastic Defl. of
Cont . Beams) Cont. Beams
De signation Under Dead- Loa d Under Dead-Load Same Same
Plus Super- Plus Super- Poi nts Points
imposed-::Load im;Eosed-Load As Col. B As Col. C
1 -6 1 -6
in- kips in -kip s in X 10 In x 10
A B C D E
Ct•
......
Simple Beam, SB -1 7,6 S.7 31 20
4.1 J.S
Continuous Beam, LB-1 7.6 4,3 30 iL
4.0 3.3
Simple Beam, SB- 3 16.h 12,3 122 80
7.4 6.S
Continuous Beam, Ll3 - 3 16 .L 9.2 136 5L
8.J 5. 9
a
Note that t he cross-secti ons of the 1-bar beams (SB-1 and LB-1) were identical; also tAat the cross-
sections of the 3-bar beams (SB-3 and LB-3 ) were identical.

b Redundant moments were determined by elastic theory for prismatic members in cols.Band C.

c All beams were loaded at age 28- days. According to Fig. 9, 60-day test values can be projected t
20- year values by multiplying by a f actor of about 2.0.
d Bottom numbers are ratios of curvatures to moments; Col. D/ Col . B, Col. E/ Col. C, Col. L/ Col. B,
Col . M/ Col. C.
TABLE 7. (Continued)

Beam Curvatures Under Dead-Load Plus Superimpos ed-Load.Values At End of 2-Months Loading Period.

Total (Instan- e Time Dependent Shrinkage d Instantaneous e Creep

taneous Plus (Shrinkage Plus Creep
Time-Dependent) Plus Creep)
Same Same Same Same Same Same Same Same Same Same
Points Points Points Points Points Points Points Points Points Points
As Col. B As Col. C As Col. B As Col. C As Col. B As Col. C As Col. B As Col. C As Col. B As Col. C
!in -6 -1 1 1 1 1 1 1 1 1
X 10 in X 10-6 -in X 1 0 -6 -in X 10-6 -in X 10-6 In x 10- 6 -in X 10- 6 -in X 10- 6 -in X 10- 6 In x 10-6
A F G H I J K L M N 0
29 30 55 45 24 25 Q\
SB-1 60 50 0.9 1.5 5 5 7.2 7.9 0.8 1.2 N

0 3 0 21
LB-1 85 40 1.8 1.9 5 5 10.6 8.2 l.7 1.5
202 1 7 0 77
SB-3 210 165 0.7 l.l 8 8 12.J 12.8 0.7 1.0
212 97 7b 3
LB-3 220 105 o.6 0.9 8 8 12.9 10.6 o.6 0.8

e
Bottom numbers are ratios of curvatures at end of 2 months to initial curvatures; Col. H/ Col. D, Col. I/ Col. E,
Col. N/ Col. D, Col. 0/ Col. E.
 63

VI. CONCLUDING REMARKS

An attempt has been made to study the comp lex deforma-

tional behavior of reinforced concrete flexural members as
influenced by the interrelated effects of cracking~ shrink-
age warping, creep, tensile and compressive steel percent-
age, continuity, moment redistribution in statically
indeterminate beams, etc. Initially, a detailed review
and discussion of existing methods, guides and rules of
thumb for predicting deflections was presented for the
purpose of examining the nature of the deflection problem.

A new and practical method was presented for computing

shrinkage warping which agrees more closely with test data
than previous methods advanced. See Eqs. (28), (29), or
(30) for the appropriate curvature expressions to be inter-
grated acros~ the span. For example, the mid span deflec-
t ion A = q;, L /8 for a simple span. However, only shrinkage
warping of uncracked specimens has been investigated experi-
mentally to the writer ' s knowledge, and effects of cracking
on shrinkage curvature in uns ymmetrical sections represents
an area requiring further study . A number of interesting
observations related to effects of steel percentage, crack-
ing and the phenomenon of the shifting neutral axis with
time on deflections were made from the experimental
curvatures and deflections.

Consideration was given to the effects of cracking on

deflections and recommended design procedures presented for
predicting these effects. A method was demonstrated for
including the effect of moment redistribution due to crack-
ing in computing deflections of statically indeterminate
beams . Deflections computed by these procedures compared
reasonably well with the experimental data obtained in this
investigation and other data on deflections of simple and
continuous reinforced concrete beams. See Eqs. (23) through
(26). Comparisons are tabulated to show the nature of the
agreement that can be expected between analytical and
experimental deflections.

It appears that future studies should concentrate on

the effec t s of random cracking on deflections since both
instantaneous- load cracks and progressive cracking under
sustained l oads in many cases play a dominant role in
determining deflection behavior. In the case of statically
 64

indeterminate beams, moment redistribution effects

resulting from shrinkage, creep and cracking also
drastically inf luence deflections in many cases and
represent an area that has not been extensively explored .

The problem of deflection prediction and control of

reinforced concrete flexural members involves a number of
complex and interrelated influences herein discussed. In
addition to the largely empir ical approaches that constitute
the main tools for present-day prediction of deflections,
more attention should undoubtedly be given in the future to
the stat~stical aspects of the problem as related to
statist ically optimum designs, confidence intervals for
computed deflections, etc.
 65

VII . REFfi'__,RENCES CITED

1. Pauw, Adrian, 11 Static Modulus of Elasticity of Concrete

as Affected by Density, 11 ACI Journal, Proceedi ngs V. 57,
No. 6, Dec. 1960, pp. 679-687,

2.. Proposed Revision of "Building Cocie Requirements for

Reinforced Concrete II American Concrete Institute,
Apr . 1963.

3. ACI- ASCE Joint Connni1,tee 323, Tentative Reconnnendations

11

for Prestressed Concrete." ACI Journal., Proceedings V.

54, No. 7, Jan. 1958. pp. 545-578.
4. Scharer, Herman, ''Prestressed Concrete, Design Principles
and Rein.fo1•cing Units., 11 ACI Journal, Proceedings V . 39,
No . 4, pp. 493-528 .

5. =.evi, Franco, "Work of t he European Concrete Committee,''

ACI Journal, Proceedings V. 57, No. 9, Mar. 1961,
pp. 1041-1070 .

6. Yu, Wei-Wen and Winter, George, "Instantaneous and Long-

Time Deflections of Reinforced Concrete Beams Under
Working Loads, 11 ACI Journal, ProceedLT1gs V. 57., No . 1.,
July 1960, pp . 29- 50,

7. Murashev, V . E. , "Theory of Appearance and Opening of

Cracks, Computat ion of Rigidity of Reinforced Concrete
Members, 11 Stroitel naya Promishlenost, No . 11, Soviet
Union., 1940.

B. Myrlea, T. D;. Author- Chainnan, "Deflection of Reinforced

Concrete Members , 11 Progress Report of ACI Committee 307,
ACI Journal, Proceedings V. 27, 1931, p. 351 .

9. Wnitney, Charles S , , "Plastic Theory of Reinforced Concrete

Design," ASCE Transactions, 107, 1942, p. 251.
10. Report of ASCE-ACI Joint Committee on ''Ultimate Strength
Design," ASCE Proceedings-Separate 809, Oct . 1955.

11. Corning, Leo H., Chairman, Report of ACI- ASCE Commi 4tee 327,
11
Ul ti.mate Strength Design," ACI J rrnrnal, Proceedings V.~2,
No. 7, Jan. 1956, p. 505 .

12 . Po1~t1and Cement Association, "Deflection of Rei nforced

Concrete Members.," ST-70, 1952 .
 66

13. The American Association of State Highway Officials,

11 Standard Specifications for Highway Bridges," Eighth

Edition, 1961 .
14. Dunha.m, C. W., "Theory and Practice of Reinforced Concrete,"
McOraw-Hill Book Company, Inc., 1953.

15 . Large, G. E. , "Basi c Reinforced Concrete Design, 11 Second

Edition, The Ronald Press Company, 1957.

16 . Ferguson, P . M. , Discussion of 11Warping of Reinforced J

Concret e Due to Shrinkage, 11 !by A. L. Miller, ACI Journal,
Proceedings V. 54, No . 6, Dec. 1958, pp. 1393-1402.

17 . Private correspondence with M~ V. Pregnoff, San Francisco,

California, to be published as an ACI Committee 435 Report .
18. Miller. Alfred L., "Warping of Reinforced Concrete Due to
Shrinkage, " ACI J ournal, Proceedings V. 54, No . 11, May
1958 pp . 939- 950.
19. Ross, A. D., 11 Creep of Concrete Under Variabl e Stress,"
ACI Journal, Proceedings V. 54, No. 9, Mar. 1958,
pp. 739-758.

20 . McHenry, Douglas, 11 A New Aspect of Creep in Concrete and

Its Appl i cation to Design," Pr oceedings AS™, V. 43,
1943, pp . 1969-1984 .
21. Concrete Reinforcing Steel Institute, 11
CRSI Desi gn Hand-
book," Revised 1959 .

22 . Newmark, N. l-1 . , "Numerical Procedure for Comput.ing Deflec-

tions: Moments, and Buckling Loads, 11 Transactions, ASCE,
V. 108, pp . 1161-123L1.

23 . Washa G. W., and Fluck, P . G. , 11 The Effect of Compressive

Reinforcement on the Plastic Flow of Reinforced Concrete
Beams," ACI Journal, Proceedings V. lt9, No . 8, Oct . 1952,
pp . 89-108.

2u. Washa, G. W. , and Fluck, P . G.. ''Plastic Flow (Creep) of

Reinforced Concrete Continuous Beams," ACI Journal,
Pr oceedings V. 52, No. 5 Jan . 1956, pp . 549~561 .
VIII. APPENDIX

e,1 Specimen Details and Experimental Data Obtained in the Investigation

TABLE A. l. DESIGN DETAILS FOR THE TEST BEAMS OF THE CURRENT INVESTIGATION

Pne # 3 Bar, p = o.69%, w81/wn1 = 2.0 Three# 3 Bars, p = 2.07%, w8JwDL = 5.5
Beams 411 ' x 5n, b = 411 , d = .~", L = 9 1 , Beams 411 x 511 , b =- 411 , d = 411 , L = 9 1 ,
Descriptiou tf 1 : 500011 psit A~ = o.. ~ tn ,
kg - 1.00, Ifr - 7.27 in, 1 41.7 in,
0
n: 6,
-
. Li fl = 5000 psit A 5 = O.J~ tn 2, n =- 6, . h
led = 1. 56 , Icr = 24. 7 in , IJ = 41.. 7 1 n ,
11
~DL = 20.8 #/ t, WsL = 41.6 #;ft WDL = 20.8 #/ft, WsL = ll.4.4 /ft

Simple Beam Continuous Beam Simple Beam Continuous Beam

aMax
br-1aX:
Positive Mom.
Pos. Mom~ , in-lb
0.1250 wL2 at ~
2530 + 5060 = '590
0.0703
1420 +
wL2 at
0.1250 wL2 at
2840 =
.)751 i
0.0703 wL 2 at .J751
2530 + 13,920= 6,450 1420 + 78~0 = 9240
4260
a
~ax, Negative Mom. --- 0,1250 wL2 at
Suppt. 0.1250 wL at Suppt.
ax.
CMax.
Neg. Mom., in-lb
Pos. Mom. fc., psi
---
348 + 696 = 1044-:--
2530 +
196 + 391 = 58Y
5060 =
7590
160 + 879 = 1039~
---
2530 + 13,920=16,450
90 + 495,.. 585*
CMax.
Cfux.
Neg. Morn. fc, psi
Pos. Mom. fs, psi
--- 348 + 696 = 1044* --
160 * 879 = 1039*
)250 + 12500 ~ 18,750 3510 + 7020::: 10,,30 1500 + 8250 = 9,750 843 + L527 = 5,370
cMax:.
dMax.
Neg. Mom. f~p psi
v, psi
---
14
6250 + 12500 ~ 18,750
18 JO
---
1500 + 8250 = 9,7~0
38
~ax. u, psi 68 85 148 185
eMax. Pa.s .. Morn. ft, psi 149 + 297 = 4L6 84 + 167 = 251 142 + 778 = 920 80 + 437 = 527
9Max. Neg. Mom. ft, psi --- 149 * 297 = 446 ---
142 + 778 = 920

* Note that the computed maximum concrete compressive stresses are the &a.me for the -- cont.
 Table A.1--Continued

corresponding simple and continuous beams (t e 1-bar beams--also the 3-bar

beams); also that the computed ma.:ximum concrete compressive stresses are the
same at all points along the 1-bar and 3-bar simple beams--also the same at
all points along the 1-bar and 3-bar continuous beams.
8 In the case of the continuous beams, all moments are computed by elastic
theory for prismatic members.

bwhere 3 numbers appear, they refer to DL +SL= Total Load effects, re-
spectively . One number refers to total load effects.

~aximum stresses fc and f were computed using the cracked transformed

section properties and a ~odular ratio of 6, according to the A.ASHO
Specifications.

dcomputed using v = V/bd and u = V/2 0 j d .

6Maxi..mum concrete tensile stresses ft were computed using the uncracked

transformed section properties.
 69

4000 1
7
,/
l
1/
3000
Ekf
V
I
I
I

1000
V
I
I
0
7
0 2 4 6 8 10
I

12
Concrete Strain, in/in x 10-4

t6 • 5130 psi; E • 4.4 x 106 psi

Fig. A.1--Average 28-day concrete stress-strain

curve (6" x 12 11 cylinder tests)
 70
With No Steel
0

200

0 Shrinkage Specimen With One Bar, B-l,(p=0.69%)

20 0

300--------- ---------------
Three Bars B-3 ( =2.07%)
0

200

300 - - - - - - - - - - - - - - - - - - _ .
0 30 60 90
Concrete Age in Days (initial
readings taken at age u days)

0 A--Top Gages at Quarter-Point of Span

o B- -Bottom Gages at Quarter-Point of Span
A C--Top Gages at Midspan
t1 D--Bottom Gages at Midspan

Fig. A.2--Concrete shrinkage versus time curves for

specimens containing different steel percentages
(duplicate shrinkage specimens were used)
 71

0"'-4!!1-----------------i.------~
0 30 60 90
Concrete Age in Days (initial
readings taken at age 4 days)

0 A--Shrinkage Specimen With Three Bars, B-3, (p=2.07%)

0 B--Shrinkage Specimen With One Bar, B-1, (p=0.67t)

Fig. A.J--A 11erage shrinkage curvature along

members versus time curves
 72

600 Mi':1/J_p~ Gag~s ( Simple Beams)

oo .....,;;...,_____.,...Gages
Quarter-Point ________
(Simple..,..Beams)
_______,,
4

20 60
Time in Days (time zero taken at
age 28 days--age beams were l oaded)

0 --Bottom Ga~e (Tension) For Three-Bar Simple Beam,

SB-3, (p = 2.07%, wSL/wDL =5.5)
0 --Top Gage (Compression) For Three-Bar Simple Beam,
SB-J, (p = 2.07%, w81/wDL =5.5)
--Bottom Gage (Tension) For One- Bar Simple Beam,
SB-1, (p = 0.61%, w8 r!wn1 =2Ao)
--Top Gage (Compression) For One-Bar Simple Beam,
SB-1, (p = o.67%, ~81/wDL x2.0)

Fig. A.h--Total (instantaneous plus time-dependent) concrete

strain versus time curves for two simple beams
with different steel percentages and loading,
but the same computed elastic concrete stresses
 74

Midspan Gages (Simple Beams)

400
I
'O,()
c,j I

~~
S.. X
(l)
,:, ~
,;:::-,-i
p ........
(/) -~
.S.-0.,
c,j
H n1 0
tn s
P..'"O
Q.) Q.)
Cl) Ol
S.. o
<:) it le Beams
~
~ ·c
Q.)
400
Quarte
11.. P.
:;j
(f) U)
~
0 (/)

§~
~~
~ 0
o:i ....:I
i::
H 0
0 20 40 60
Time in Daya (time zero taken at
age 28 days--age beams were loaded)

0 --Bottom Gage (Tension) For Three-Bar Simple Beam,

SB-3, (p ~ 2. 07%, Ws1/WnL e 5.S)
0 --Top Gage (Compression) For Three-Bar Simple Beam,
SB-3, (pc 2.07%, w8L/wDL e 5.5)
8 --Bottom Gage (Tension) For One-Bar Simple Beam,
SB-1, (p: 0.67%, Wsr/Wn1 s 2.0)
tl --Top Gage (Compression) For One-Bar Simple Beam,
SB-1, (p == 0.67%, w81/wDL "' 2.0)

Fig. A.6--Instantaneous plus creep strain versus time curves for

two simple beams with different steel percentages and
loading,but the same computed elastic concrete stresses
 75

Middle Support Gages (Continuous Beams )

--Negative Moment Region
600- - - - - - - - - - - - - - - - - - - - - - -
I
"0'-0
"' '
I)
A~
0

~ ~
Q)

-g:::>'-....
-~
C:
(I) •r-1

-~
"''O
.
f-4 m
~ 0
Cl) t-1
P,'O 0
Q) a.>
Q) (I)
H 0
c.., a..
Cl)-~
:j f-4
~ ~ Point of Max. Elastic Defl. Gages (Cont.Beams)
~ J5 400 --Positive Moment Region
Q) ;:I
0(1) r:::~-~~~-(J--cr-
@~
+l
§~
~s
C:
H

0
0 20 40 60
Time in Days (time zero taken at
age 29 days--age beams were loaded)

0 --Bottom Gage (Tension-Pos. Mom., Compression-Neg.Mom.)

Por Three-Bar Continuous Beam,LB-J,(p=2.07%,w81/wD1 =5.5)
0 --Top Gage (Compression- Pos .Mom., Tension-Neg.Mom.) For
Three-Bar Continuous Beam, LB-3,(p=2 . 07%,w51/w01 =5.5)
& --Bottom Gage (Tension- Pas.Mom., Compression-Neg.Morn.)
For One- Bar Continuous Bearn,LD-l,(p=0.67,w81/wD1 ~2.0)
O --Top Gage ( Compression-Pos .Mom., Tension - Neg.Mom.) r'or
One Bar-Continuous Beam,.LB-l,(p=0.67%,w51/w01 =2.0)

Fig. A.7--Instantaneous plus creep strain versus time curves for

two continuous beams with different steel percentages and
loadin~, but the same computed elastic concrete stresses
 76

Simple Beams
300 - .......- - - - - - - - - - - - - - - - - -
Midspan,
Midspan,

0
Quarter-Point,
Quarter-Point ,

Point of Max. Elastic Defl.,

Point of Max. Elastic Defl.,
Continuous Beams
300 - - - - - - - - - - - - - - - -..............-
Middle
Middle

100

0
0 20 uO 60
Days (t:iJne zero taken at
T:iJne in
age 28 days--age beams were loaded)

Fig . A. 8--Total (instantaneous plus time-dependent)

curvature versus time cur ves for four test beams
 77

Simple Beams
300 - - - - - - - - - - - - - - - - - - - - -
Midspan, SB-3 - ~
Midspan, SB-1

0
Quarter-Point,
Quarter-Point,

Point of Max. Elastic Defl.,

Point of Max. Elastic Defl.,

Continuous Beams
300 - - -- - -- -----,-----,;,-~

0
0 20 40 60
Time in Days (time zero taken at
age 28 days--age beruns were loaded)

Fig. A.9--Instantaneous plus creep curvature

versus time curves for four test beams
 78

Time i n Days (time zero taken at

age 28 days --age beams wer e loaded)

0 Simple Beam, SB-3, A initial = 0.153 in.

0 Simple Beam, SB-1, ~-in1.·t.1a1 = 0.041 in .
0' Continuous Beam, LB - 3, A initial = 0.056 in.
8 Continuous Beam, LB-1, b.. initial
:a: 0.021 in .

Fig. A.10 - -Time-dependent deflection versus

time curves for four test beams
 ii

TABLE OF CONTENTS
Page

LIST OF FIGURES••••••••••••••••••••••••••••••••••• i i i
INTRODUCTION ••••••••••••••••••••••••••••• ·• • • • • • • • • l

DESCRIPI'ION OF EXPERIMENT.AL INVESTIGATION......... 1

TESTING PROCEDURES •••••••••••••••••• • •• , • •• • ••••• • 2
COMPARISON OF TEST RESULTS•••••••••••••••••••••••• 2
CONCLUSIONS....................................... 3
 iii

LIST OF FIGURES
Fig!l,re
1. Comparison of shrinkage strains at the
top fiber for the specimens with
different steel percentages (strains
proportioned to extreme fiber using
a linear distribution with the top and
bottom gages)•• ••·••••••••••••••••••••••• 4
2. Compression and tension gage creep
coefficients versus time curves for
four test beams•••••••••••••••••••·•••••• 5
3. Time-dependent deflection coefficients
versus time curves for four test beams... 6
4. Concrete stress-strain curve at age 28-days 7

5. Concrete shrinkage versus time curves for

specimens containing different steel
percentages (duplicate shrinkage specimens
were used)•••••••••••••·•••••••••••••••• • 8
6. Average shrinkage curvature along members
versus time curves••••••••••••·•••••••••• 9
7• Total. (instantaneous plus time-dependent) ,
concrete strain versus time curves for two
simple beams with different steel percentages
and loading, but the same computed elastic
concrete s tresses••••·•••••••••·•·••••••• 10
B. Total (instantaneous plus time-dependent) con-
crete strain versus ti.me curves for two
continuous beams with different steel
percentages and loading, but the same
computed elastic concrete stresses...... 11
9. Instantaneous plus creep strain versus time
curves for two simple beams with different
steel percentages and loading, but with
same computed elastic concrete stresses. 12
 iv

Figure

10. Instantaneous plus creep strain versus

time curves for two continuous beams
with different steel percentages ·
and loading, but with same computed
elastic concrete stresses............... 13

ll .. Total (instantaneous plus time-dependent)

curvature versus time curves for four
test beams·••••••••••••·••••·•••••••••• 14
12. Instanteous plus creep curvature versus time
curves for four test beams.............. 15
13. Time-dependent deflection versus time
curves for four test bearns•••••••••••a•• 16
 1

L INffiODUCTION

Part II of this study consists of a rerun of tests

in Part I and an analysis of the resulting data.

The tests of Part I were rerun because some of the

beams were honeycombed and one of the beams (L-Bl) was
cracked while being moved into position for loading.
Concrete for the beams of Part II was vibrated
during pouring in order to minimize the honeycomb.

It was judged desirable to determine the ei'fect, if

any, of the condition of the beams of Part I upon the
results of the study.

II• DESCRIPTION OF EXPERIMENTAL INVESTIGATION

A total of four beams was tested, two simple beams

and two continuous beams ( each with two equal spans
contimious aver a center support). One simple beam
(SB - 1) and one continuous beam (LB - 1) were rein.forced
with one #3 bar. The other simple beam (SB - 3) and
contimious beam (LB - 3) were reinforced with three #3 bars.
ill spans were 9' long, the continuous beams having an
overall length of 18 1 • In addition to the four test
beams, six shrinkage specimens were tested. The shrinkage
specimens were the same size as the simple beams. Two
were reinforced with three #3 bars, two with one #3 bar,
and two were without reinforcement. The shrinkage specimens
were placed on one side on a smooth, oiled, plywood surface
in an attempt to eliminate any frictional effects which might
influence the shrinkage measurements. Details of the test
beams are shown in Fig. 3 of Part I of this study.

The properties of the materials were as follows:

Concrete SlUlllp • • • • • - • • • • • 21.n

2
28 day concrete cylinder strength • .44SO psi
Concrete modulus of elasticity •• •J.) X 106 psi
Tensile yield point of the steel. •L9 ,ooo psi
 2

The concrete strains were measured by using a Whittemore

mechanical strain gage with a 10 11 gage length. Gage point s
were imbedded near the top and bottom of each bearn at six
different l ocations giving a total of 12 gages and 24 gage
points for each beam. Six gages and 12 gage points were used
on each s hrinkage specimen. Temperature effects on strains
were eliminated through the use of a temperature bar made of
invar metal having the same coef fici ent of thermal expansi on
as the concrete.
III. TESTING PROCEDURES

All beams were loaded at age 28 days with iron bricks.

The bricks were spaced continuously in the 3 - bar beams and
unifornly in t he 1 - bar beams . The loading was the same as
in Part I of this study and can be seen i n Fig. 4 of Part I.

The deflection and strain readings reported were the

average of t hose on each side of the beam in t he same
position in order to eliminate any torsional effects. Also,
only the average of corresponding strain readings on the
shrinkage specimP.ns :uirl test beams were reportP.d.

IV. COMPARISON OF TEST RESULTS

Figures in Part II correspond to figures in Part I as

follows:

Part II Part I
Fig, 1 corresponds to Fig. 8
Fig. 2 II II Fig . 10
Fig. 3 II IT Fig. 11
Fig . 4 II II
Fig. A- 1
Fig. 5 II
Fig. A- 2
,,
II

Fig. 6 II
Fig. A-3
Fig. 7 It II
Fig. A-4
Fig .. 8 II
" Fig. A-5
Fig. 9 II II
Fig. A-6
Fig. 10 II II
Fig. A-7
Fig. 11 11 II
Fig . A-8
Fig. 12 II II
Fig. A-9
Fig. 13 ti II
Fig. A-10
A comparison of Fig . 4 of Part II with Fig . A-1 of
Part I shows that both fc and E were somewhat higher in
 3

tests conducted in Part I as opposed to those of Part II.

The modulus of elasticity was 26% higher in Part I as
compared to the modulus of elasticity of the concrete in
Part II.

Figures 1 and, of Part II and Figures 8 and A-2 of

Part I show that the shrinkage was about 20% greater in
Part I than in Part II. This was to be expected because
a rich concrete will tend to shrink more than a lean one.
In general, all other curves for strains and deflections
ran higher in Part II than in Part I by amounts ranging
from 15% t o about 40%. Since the modulus of elasticity
of the concrete in Part I was 26% higher than in Part II,
these larger strains and deflections appear quite reaaonable.
The only exceptions to this occur in the tension gage
creep coefficients of Fig. 2 and the concrete strains
in the positive moment region of Fig. 8. These were about
the same to slightly lower in Part II as compared to Part
I. In the writer ' s opinion, this was probably caused by
tension cracking of the concrete and a redistribution of
momenta in the continuous beams.

V. CONCLUSIONS

The test results in Part II agree quite well with

those of Part I. Strains and deflections are somewhat higher
in the second set of tests than in the first, but this is
caused by the lower modulus of elasticity of the concrete
in Part II. Because of the close agreement of the test
results, it is the writer 1 s opinion that neither the
honeycomb of the test beams in Pa.rt I or the hairline
crack of beam L - Bl had any effect on the data.
 4

t.O
I
Cl 0 100
c..c ......
«!
..!:.'. X
i:::
.,-; Q
,... •r1 200
..s:;'-..
I.I) C:
.,-;

"
+> " :ioo
... .,-;I=.
QI

()
[/)

~ «I
0 I,..
(.) +>
'.I) -100
() 30 GO 90

Concrete Age in Days (initial

read i ngs t ak en at a ge 4 days)

A--Shrk . :'pee . With o Steel (p=O), All Gages Used

13--Shr.!~ . Spec . With One Dar ( p=O. 69%), All Gages Us ed
C- -Shrk . .::>pe e. V{ith Three Bars (p=2 . 07%), All Gages Used

Fig . 1-- Comparison of shrinkage strains at the top fiber

for the pecimens with different steel percentages
( strains p ..·oportioncd to extreme Jiber using a
linear distribution with the top and bottom gages)
 ;i. 0 - - - - - - - - , - - - - - - - - - - - - - - - - - .
upper limit

0
0 20 40 60

Time in Days (time zero taken at

age 28 days--age beams were loaded)

2.0

limit

1.0

0
60

Time in Days (time zero taken at

age 28 days - -age beams were loaded)

Creep Coefficients Defined as Ratio of

Creep Strain to Initial Strain

Fig. 2--Compression and tension gage creep coefficient

versus time curves for four test beams
 6

3.0 ________________________

limit

~
A
a,
•rl
2.0 t-- -- -- -t -- ----::7""-==--=~----- --1
tJ
•rl
~
+:24-1
A a,
Q) 0
'OU
i::
Q) i::
n. 0
A~ i.01--___,,,___ _--::,,-"""--+-- - - - - --+- - -- - ---t
I tJ
Q) Q)
8 r-1
•rl 4-1
E--t Q)
p limit

0
0 20 40 GO
'l'imf' in Dayi:; ( time ze1·0 t ake:n at
a {!e 28 da ys-- agc heams were loader!)

Time-Dependent Deflection Coe ff i ci cnt Def inecJ a!'l ltF1{.io

of Time-Dependent Deflection to Initial Deflec ti on

Fig . 3-- Time-depend ent deflection coefficien

versus time curves for four test beam. ·
 7

4000

:1000
·.-4
1/J
D~
,._
1/)
rn
QJ
$,.,
+> 2000
en
Cl)
+>
...
QJ

(.)
i:=
0
C'

1000

0
0 2 4 G 8 10 12

Concrete Strain in/ in X 10- 4

fl
C
= 4 450 psi F. = 3.5 X 10 6 psi

li'i g . 4-Avcrage PS- day concrete stress-strain

curve ( 6'1 x 12 11 cylinder tests)
 8

Shrinkage Specimen With No Steel

:n
I
0
r-1

Shrinkage Svccirnen With One Bar, R-1, (p:-0.6f1%)

300----------------------

Concrete Age in Days (initial

readings taken at age 4 days)

A--Top Gages at Quarter- Point of Span

D- Dottom Gages at Quarter- Point of Spap
C--'rop Gag es at Midspan
D--Bottom Gages at Midspan

Fig. 5--Concrete shrinkage versus time curves for spe cimen$

containing different steel percentages (duplicate
shrinkage specimens were used)
 9

0
30 BO
Concrete Age in Days (initial
readings taken at a ge 4 days)

A--Shrinkage Specimen \Vi th Three Dars, B-3, ( p=2. 07%)

D--Shrinkage Specimen With One Bar, D-1, (p=0.67%)

fi t . 6--Average shrinkage curvature along mcu:bers versus

time curves
 10

Midspan Gages (Simple Beams)

0 20 40 60

'£ime in Days ( time zero taken at

age 28 days-- age beams were loaded)

O -- Dottom Ga?e iTension) For Thr~e:B~r Simple Beam,

Sfi--3, \P - 2.07,%, w JwDL - o.o)
8
•- Top Gage (Compression) For Thre: Bar Simple Beam,
SD-3 , (p = 2.07%, w5 JwDL = o.5)
A-- Bottom Gap;c ( 'l'nosion) For One-Bar Simple Beam,
SD-1 , (p = 0. (37%, , w
5
Jw 01 • 2 . 0)
~-- Top Gage ( Coi:ipre.~s i on) For One-Dar Simple Beam,
Sfi--1, (p = O.tn~:,, w.
SL
/wUL = 2.0)

Fig. 7- -Total (instantaneous plus tir.ie-clependent) concrete

strain versus time curves for two simple beams with
different steel percentages and loading, but the
same computed elastic concrete stresses
 11

Middle Support Gages (Continuous Reams)

--Negative Moment Region

Point of Max. Elastic Defl. Gages (Cont.Beams )

--Positive ~loment !legion

Time in Days (time zero taken at

a{?;e 28 days --age beams were load ed)

0 Bottom Ga ge (Tens ion-Pos . ~.lor.1. , Compression-Nc~ . ~10111 . )

For Three-Dar Continuous Deam,Lll-:IJ (p=2.07~; ,ws /wm,=;3.3)
1
[:J Top Gage (Co1>1pression-Pos . ~fom.) 'l'cns ion-Neg .~10111,) For
Three-Bar Continuous Berun, 1..B-a, ( p=-2. 07~ , wJw
5 1
Jl)=5 . 5 )
8- Bottom Gage (Tension- Pos .Mo11,., Compression-Ncg . ~io111.)
For One-De.r Continuous Ream, W-1, (p=O . fi7~ ,wsi/wDJ ,= 2.0)
):j_ -- Top Gage {Corupression-Pos.Mom. , Tension- Neg.Mom .) For
One-Dar Continuous Oeam, LD-1, LD-1, (p=0 . 67%,w 51 /wn1=2.0)
Fig. 8--'l'otal (instantaneous plus time-dependent) concrete
strain versus time curves for two continuous beams
with different steel percentages and loading, but
the same computed elastic concrete stresses
 12

1~dspan lrar.;es (Sfo1ple Deams)

fi OO.----------.-----'-----.-'--------

,...,
H
0 ::!O 110 60

Tin1c in Days ( time zero taker. at

age 28 days --a r,e beams were loaded

0 -- Botton! Gage 'Tension) For Tht ee- 13ar Simple Dcaru,

~11-3, (p = 2.07,~, ws1/wm1 ::,r 5 . 5)
~ - - 'rop Gage ( Comprcssi on) For Three- Dar Simple Dearr.,
SD-3, ( p == 2 . 075~, wsJwDJJ == 5 . 5)

8 -- Bottom Gage (Tension F or One-J1ar Simple Deam,

SD-1, (p = 0 . 67~ , w / w = 2 . 0)
SL DL
.'O'.. -- Top G-age ( Cor.1pressior.) For One-Dar Sir?:ple neam,
SH- 1, (p = 0.67~~ , w , /wn = 2.0)
SL l L

Fig. 0--Instantaneo-us plus creep str ain ver::rns time curves

for two simple beams with different steel percentages
and loading, but the same coitpt!tcd elastic concrete
stresses
 13

Middle Support Gages (Continuous Deams)

soo --N gative
-~.:.s:.......:..:.______ ~--""'T"""------,
Moment R gion

400 1--- - - - ---+-- -- - - ---- - - -- -------1

200

Point of Max. Elastic Defl. Gages ( Cont.Beams)

400 ~-=-~P~o~s~i~t~i~v~c:=.....:·~10~m~e~n~t~·~r~t.i.....~~-~-.--------,

0 20 40 60

Time in Days ( time zero talcen at

age 28 days--age beams were loaded)

0 Bottom Gage (Tension-Pos.Uom., Compression-Neg.11om.)

F or Three-Bar Continuous Beam, 113- 3, ( p=2. 071{ ,w 8 JwD 1=5. G)

• -- Top Gage (Compression-Pos. ~:om. Tension- g: . ~l om.) F r

Three-Bar Continuous Beal'!l , TJ1-;J, ( p=2. 07~,;, wsJwDL=5 . r.: )
A-- Bottom Gage (Tension-Pos . !,lorn., Co1:1pre sion- C:J.! lom.)
For One-Bar Continuous Deam, J3-l,(p=0.67, w8 JwnL=... . 0)
l::( -- Top Gage (Compression-Pos .Uom. 'l'ens .'on- eg.t,l om.) For
One Bar-Cont inuous Deam, Lll-1, (p=O.C7r ,w 81 0 L=2.0) /w
Fig. 10--Instantaneous plus creep strain versus ti me curves
for two continuous beams witli di f ferent steel pe r -
centages and loading, cut the same computed elastic
concrete stresses
 14

Midspan,
Midspan,
:~ oo

200

100

0
Quarter-Point, SB-3
Quarter-Point, SB-1
Point of Max. Elastic Defl., LB-3
Point of Max. Elastic Defl . ,
Middle Support, LB-3
Middle Support, LB-1
LB?i
~oa .--~~~~1~:-;_-:_-:__-::_-:.-=--=-~f--=--;-;..,--:..:t.~c...tn

Tifne in Days { time zero taken at

age 28 days--age beams were loaded)

Fig , 11--Total (instantaneous plus time-dependent)

curvature versus time curves for four test beams
 16

0.150

0.100

0 . 050

0
0 20 40 60

Time in Days (time zero taken at

a ge 28 days--age beams were loaded)

Simple Beam, SB-3, initial = 0.157 in.

Simple Deam, SB-1; initial = 0.036 in.

Oontinuous Deam, Ln-3, i nitial ::;

0.061 in.
Continuous Beam, LD-1, initial = 0.019 in.

Fig. 13 -- Time- dependent defl e ction versus time curves

for four test beams