(Reapproved 2000)
(Appendix B added 2003)
*
Editor
Acknowledgment is due to Robert F. Mast for his major contributions to the Report, and to Dr. Ward R. Malisch for his extensive input to the various chapters.
The Committee also acknowledges the processing, checking, and editorial work done by Kristi A. Latimer of Rutgers University.
This report presents a consolidated treatment of initial and timedependent shrinkage; slabs; strains; stresses; tendons; tensile strength; timedepen
deflection of reinforced and prestressed concrete elements such as simple and dent deflection.
continuous beams and oneway and twoway slab systems. It presents the
state of the art in practice on deflection as well as analytical methods for CONTENTS
computer use in deflection evaluation. The introductory chapter and four
main chapters are relatively independent in content. Topics include Deflec Chapter 1Introduction, p. 435R2
tion of Reinforced Concrete Oneway Flexural Members, Deflection of
Twoway Slab Systems, and Reducing Deflection of Concrete Members.
Chapter 2Deflection of reinforced concrete oneway
One or two detailed computational examples for evaluating the deflec flexural members, p. 435R3
tion of beams and twoway action slabs and plates are given at the end of 2.1Notation
Chapters 2, 3, and 4. These computations are in accordance with the current 2.2General
ACI or PCIaccepted methods of design for deflection.
2.3Material properties
2.4Control of deflection
Keywords: beams; camber; code; concrete; compressive strength; cracking;
creep; curvature; deflection; highstrength concrete; loss of prestress;
2.5Shortterm deflection
modulus of rupture; moments of inertia; plates; prestressing; preten 2.6Longterm deflection
sioned; posttensioned; reducing deflection; reinforcement; serviceability; 2.7Temperatureinduced deflections
435R1
435R2 ACI COMMITTEE REPORT
3.5General approach to deformation considerations strength concrete with reinforcing bars and prestressed rein
Curvature and deflection forcement, coupled with more precise computeraided limit
3.6Shortterm deflection and camber evaluation in state serviceability designs, has resulted in lighter and more
prestressed beams materialefficient structural elements and systems. This in
3.7Longterm deflection and camber evaluation in turn has necessitated better control of shortterm and long
prestressed beams term behavior of concrete structures at service loads.
This report presents consolidated treatment of initial and
Appendix A3, p. 435R42 timedependent deflection of reinforced and prestressed
Example A3.1Short and longterm singletee beam concrete elements such as simple and continuous beams and
deflections oneway and twoway slab systems. It presents current engi
Example A3.2Composite doubletee cracked beam neering practice in design for control of deformation and
deflections deflection of concrete elements and includes methods
presented in Building Code Requirements for Reinforced
Chapter 4Deflection of twoway slab systems, Concrete (ACI 318) plus selected other published approaches
p. 435R50 suitable for computer use in deflection computation. Design
4.1Notation examples are given at the end of each chapter showing how to
4.2Introduction evaluate deflection (mainly under static loading) and thus
4.3Deflection calculation method for twoway slab control it through adequate design for serviceability. These
systems stepbystep examples as well as the general thrust of the report
4.4Minimum thickness requirements are intended for the nonseasoned practitioner who can, in a
single document, be familiarized with the major state of prac
4.5Prestressed twoway slab systems
tice approaches in buildings as well as additional condensed
4.6Loads for deflection calculation
coverage of analytical methods suitable for computer use in
4.7Variability of deflections deflection evaluation. The examples apply AC1 318 require
4.8Allowable deflections ments in conjunction with PCI methods where applicable.
The report replaces several reports of this committee in
Appendix A4, p. 435R62 order to reflect more recent state of the art in design. These
Example A4.1Deflection design example for longterm reports include ACI 435.2R, Deflection of Reinforced
deflection of a twoway slab Concrete Flexural Members, ACI 435.1R, Deflection of
Example A4.2Deflection calculation for a flat plate Prestressed Concrete Members, ACI 435.3R, Allowable
using the crossing beam method Deflections, ACI 435.6R, Deflection of TwoWay Rein
forced Concrete Floor Systems, and 435.5R, Deflection of
Chapter 5Reducing deflection of concrete members, Continuous Concrete Beams.
p. 435R66 The principal causes of deflections taken into account in
5.lIntroduction
this report are those due to elastic deformation, flexural
5.2Design techniques
cracking, creep, shrinkage, temperature and their longterm
5.3Construction techniques effects. This document is composed of four main chapters,
5.4Materials selection two to five, which are relatively independent in content.
5.5Summary There is some repetition of information among the chapters
in order to present to the design engineer a selfcontained
References, p. 435R70 treatment on a particular design aspect of interest.
Chapter 2, Deflection of Reinforced Concrete OneWay
Appendix BDetails of the section curvature method Flexural Members, discusses material properties and their
for calculating deflections, p. 435R77
effect on deflection, behavior of cracked and uncracked
B1Introduction
members, and timedependent effects. It also includes the
B2Background
relevant code procedures and expressions for deflection
B3Crosssectional analysis outline computation in reinforced concrete beams. Numerical
B4Material properties examples are included to illustrate the standard calculation
B5Sectional analysis methods for continuous concrete beams.
B6Calculation when cracking occurs Chapter 3, Deflection of Prestressed Concrete OneWay
B7Tensionstiffening Members, presents aspects of material behavior pertinent to
B8Deflection and change in length of a frame member pretensioned and posttensioned members mainly for
B9Summary and conclusions building structures and not for bridges where more precise
B10Examples and detailed computer evaluations of longterm deflection
B11References behavior is necessary, such as in segmental and cablestayed
bridges. It also covers shortterm and timedependent deflection
CHAPTER 1INTRODUCTION behavior and presents in detail the Branson effective
Design for serviceability is central to the work of struc moment of inertia approach (Ie) used in ACI 318. It gives in
tural engineers and codewriting bodies. It is also essential to detail the PCI Multipliers Method for evaluating time
users of the structures designed. Increased use of high dependent effects on deflection and presents a summary of
DEFLECTION IN CONCRETE STRUCTURES 435R3
various other methods for longterm deflection calculations CHAPTER 2DEFLECTION OF REINFORCED
as affected by loss of prestressing. Numerical examples are CONCRETE ONEWAY FLEXURAL MEMBERS*
2.1Notation
given to evaluate shortterm and longterm deflection in A = area of concrete section
typical prestressed teebeams. Ac = effective concrete cross section after cracking, or
Chapter 4, Deflection of Twoway Slab Systems, covers area of concrete in compression
the deflection behavior of both reinforced and prestressed As = area of nonprestressed steel
twowayaction slabs and plates. It is a condensation of ACI Ash = shrinkage deflection multiplier
Document 435.9R, StateoftheArt Report on Control of b = width of the section
c = depth of neutral axis
Twoway Slab Deflections, of this Committee. This chapter
Cc ,(CT)= resultant concrete compression (tension) force
gives an overview of classical and other methods of deflection Ct = creep coefficient of concrete at time t days
evaluation, such as the finite element method for immediate Cu = ultimate creep coefficient of concrete
deflection computation. It also discusses approaches for d = distance from the extreme compression fiber to
determining the minimum thickness requirements for two centroid of tension reinforcement
way slabs and plates and gives a detailed computational D = dead load effect
example for evaluating the longterm deflection of a two Ec = modulus of elasticity of concrete
way reinforced concrete slab. Ec = ageadjusted modulus of elasticity of concrete at
time t
Chapter 5, Reducing Deflection of Concrete Members,
Es = modulus of elasticity of nonprestressed reinforcing
gives practical and remedial guidelines for improving and steel
controlling the deflection of reinforced and prestressed concrete EI = flexural stiffness of a compression member
elements, hence enhancing their overall longterm serviceability. fc = specified compressive strength of concrete
Appendix B presents a general method for calculating the fct, ft = splitting tensile strength of concrete
strain distribution at a section considering the effects of a fr = modulus of rupture of concrete
normal force and a moment caused by applied loads, fs = stress in nonprestressed steel
fy = specified yield strength of nonprestressed reinforc
prestressing forces, creep, and shrinkage of concrete, and ing steel
relaxation of prestressing steel. The axial strain and the h = overall thickness of a member
curvature calculated at various sections can be used to calculate I = moment of inertia of the transformed section
displacements. This comprehensive analysis procedure is for Icr = moment of inertia of the cracked section trans
use when the deflections are critical, when maximum formed to concrete
accuracy in calculation is desired, or both. Ie = effective moment of inertia for computation of
deflection
The curvatures and the axial strains at sections of a
Ig = moment of inertia for gross concrete section about
continuous or simply supported member can be used to centroidal axis, neglecting reinforcement
calculate the deflections and the change of length of the K = factor to account for support fixity and load
member using virtual work. The equations that can be used conditions
for this purpose are given in Appendix B. The appendix Ke = factor to compute effective moment of inertia for
includes examples of the calculations and a flowchart that continuous spans
can be used to automate the analytical procedure. ksh = shrinkage deflection constant
K(subscript)= modification factors for creep and shrinkage
It should be emphasized that the magnitude of actual effects
deflection in concrete structural elements, particularly in l = span length
buildings, which are the emphasis and the intent of this L = live load effect
Report, can only be estimated within a range of 2040 percent M(subscript)= bending moment
accuracy. This is because of the large variability in the prop Ma = maximum service load moment (unfactored) at
erties of the constituent materials of these elements and the stage deflection is completed
Mcr = cracking moment
quality control exercised in their construction. Therefore, for
Mn = nominal moment strength
practical considerations, the computed deflection values in Mo = midspan moment of a simply supported beam
the illustrative examples at the end of each chapter ought to P = axial force
be interpreted within this variability. t = time
In summary, this single umbrella document gives design Ts = force in steel reinforcement
engineers the major tools for estimating and thereby controlling wc = specified density of concrete
through design the expected deflection in concrete building yt = distance from centroidal axis of gross section,
neglecting reinforcement, to extreme fiber in tension
structures. The material presented, the extensive reference lists = thermal coefficient
at the end of the Report, and the design examples will help to c = creep modification factor for nonstandard
enhance serviceability when used judiciously by the engineer. conditions
Designers, constructors, and codifying bodies can draw on the sh = shrinkage modification factor for nonstandard
material presented in this document to achieve serviceable
deflection of constructed facilities. *
Principal authors: A. S. Ezeldin and E. G. Nawy.
435R4 ACI COMMlTTEE REPORT
as recommended by ACI 318 and its Commentary Normal strength concretes are those with compressive
(1989) and ACI Committees 435 (1978), 363 (1984), and strengths up to 6,000 psi (42 MPa) while higher strength
209 (1982). concretes achieve strength values beyond 6,000 and up to
2.3.1 Concrete modulus of ruptureAC1 318 (1989) 20,000 psi (138 MPa) at this time.
recommends Eq. 2.1 for computing the modulus of rup ACI 435 (1963) recommended the following expres
ture of concrete with different densities: sion for computing the modulus of elasticity of concretes
with densities in the range of 90 pcf (1445 kg/m3) to 155
fr = 7.5 X K, psi (2.1) pcf (2325 kg/m3) based on the secant modulus at 0.45 fc
(0.623 X g, MPa) intercept
where X = 1.0 for normal density concrete [145 to 150 E = 33 MQ*~ K, psi (2.4)
pcf (2325 to 2400 kg/m3)] )$) 1.5
.
(ocO43 c g9 MPa)
= 0.85 for semi lowdensity [ll0145 pcf
(1765 to 2325 kg/m3)] For concretes in the strength range up to 6000 psi (42
= 0.75 for lowdensity concrete [90 to 110 pcf MPa), the ACI 318 empirical equation for the secant
(1445 to 1765 kg/m3)] modulus of concrete EC of Eq. 2.4 is reasonably appli
Eq. 2.1 is to be used for lowdensity concrete when cable. However, as the strength of concrete increases, the
the tensile splitting strength, fct, is not specified. value of EC could increase at a faster rate than that
Otherwise, it should be modified by substituting fc_ t/6.7 for
generated by Eq. 2.4 (EC = wclo5 K), thereby under
fl, but the value of fct/6.7 should not exceed \/fc'. estimating the true EC value. Some expressions for E,
ACI Committee 435 (1978) recommended using Eq. applicable to concrete strength up to 12,000 psi (83 MPa)
2.2 for computing the modulus of rupture of concrete are available. The equation developed by Nilson (Carra
with densities (wc) in the range of 90 pcf (1445 kg/m3) to squillo, Martinez, Ngab, et al, 1981, 1982) for normal
145 pcf (2325 kg/m3). This equation yields higher values weight concrete of strengths up to 12,000 psi (83 MPa)
of fro and lightweight concrete up to 9000 psi (62 MPa) is:
1.5
Table 2.1  Creep and shrinkage ratios from age 60 days to the indicated concrete age (Branson, 1977)
Concrete age
Creep, shrinkage ratios
2 months 3 months 6 months 1 year 2 years > 5 years
available to the designer (Nawy, 1990). Each K coefficient is a correction factor for conditions
2.3.3 Steel reinforcement modulus of elasticityAC1 318 other than standard as follows:
specifies using the value Es = 29 x 106 psi (200 x 106 Khc = relative humidity factor
MPa) for the modulus of elasticity of nonprestressed re K/ = minimum member thickness factor
inforcing steel. KS = concrete consistency factor
2.3.4 Concrete creep and shrinkageDeflections are KC = fine aggregate content factor
also a function of the age of concrete at the time of C
= air content factor
loading due to the longterm effects of shrinkage and K;: = age of concrete at load applications factor
creep which significantly increase with time. ACI 31889
does not recommend values for concrete ultimate creep Graphic representations and general equations for the
coefficient Cu and ultimate shrinkage strain (E&. modification factors (Kvalues) for nonstandard condi
However, they can be evaluated from several equations tions are given in Fig. 2.1 (Meyers et al, 1983).
available in the literature (ACI 209, 1982; Bazant et al, For moistcured concrete, the free shrinkage strain
1980; Branson, 1977).ACI 435 (1978) suggested that the which occurs at any time t in days, after 7 days from
average values for C, and (QU can be estimated as 1.60 placing the concrete
and 400 x 106, respectively. These values correspond to
the following conditions: (2.8)
 70 percent average relative humidity
 age of loading, 20 days for both moist and steam and for steam cured concrete, the shrinkage strain at any
cured concrete time t in days, after l3 days from placing the concrete
 minimum thickness of component, 6 in. (152 mm)
Table 2.1 includes creep and shrinkage ratios at dif (2.9)
ferent times after loading.
ACI 209 (1971, 1982,1992) recommended a timede where (E&, Mar = 780 x 106 ysh
pendent model for creep and shrinkage under standard x sh = Kh Kds K; Kbs K,,
conditions as developed by Branson, Christianson, and = 1 for standard conditions
Kripanarayanan (1971,1977). The term standard condi
tions is defined for a number of variables related to Each K coefficient is a correction factor for other than
material properties, the ambient temperature, humidity, standard conditions. All coefficients are the same as de
and size of members. Except for age of concrete at load fined for creep except K,9, which is a coefficient for
application, the standard conditions for both creep and cement content. Graphic representation and general
shrinkage are equations for the modification factors for nonstandard
a) Age of concrete at load applications = 3 days conditions are given in Fig. 2.2 (Meyers et al, 1983). The
(steam), 7 days (moist) above procedure, using standard and correction equations
b) Ambient relative humidity = 40 percent and extensive experimental comparisons, is detailed in
c ) Minimum member thickness = 6 in. (150 mm) Branson (1977).
d) Concrete consistency = 3 in. (75 mm) Limited information is available on the shrinkage be
e ) Fine aggregate content = 50 percent havior of highstrength concrete [higher than 6,000 psi
f ) Air content = 6 percent (41 MPa)], but a relatively high initial rate of shrinkage
has been reported (Swamy et al, 1973). However, after
The coefficient for creep at time t (days) after load drying for 180 days the difference between the shrinkage
application, is given by the following expression: of highstrength concrete and lowerstrength concrete
/ CO.6 \ seems to become minor. Nagataki (1978) reported that
Ct = IlO+ to.6J cu (2.7)
the shrinkage of highstrength concrete containing high
range water reducers was less than for lowerstrength
where Cu, = 2.35 YCR concrete.
yCR = Khc Kdc K KF K,, KIOc = 1 for stan On the other hand, a significant difference was re
dard conditions. ported for the ultimate creep coefficient between high
DEFLECTION IN CONCRETE STRUCTURES 435R7
0
0.90
K t0 Kch
0.85
0.80
0.5 l k 1 1 1 1 1
0 10 20 30 40 50 60 0 4 0 5 0 6 0 70 80 90 100
Age at loading days (b) Relative humidity, kf o/o
(a)
0.8
061
W l 0.6
0 10 20 30 40 50 60 0 5 10 15 20
cm cm
l
0 5 10 15 20 25 0 2 4 6 8
Minimum thickness, d, in. Slump, s, in.
(c) (d)
Fig. 2.1Creep correction factors for nonstandard conditions, ACI 209 method (Meyers, 1983)
DEFLECTION IN CONCRETE STRUCTURES 435R9
Table 2.2Recommended tension reinforcement ratios for nonprestressed oneway members so that deflections will
normally be within acceptable limits (ACI 435, 1978)
For continuous members, the positive region steel ratios only may be used. pl: Refers to the balanced steel ratio based on ultimate strength.
Table 2.3Minimum thickness of nonprestressed beams and oneway slabs unless deflections are computed (ACI
318, 1989)
Minimum thickness, h
Member Members not supporting or attached to partitions or other construction likely to be damaged by large
deflections.
e = Span length
Values given shall be used directly for members with normal weight concrete (w, = 145 pcf) and grade 60 reinforcement. For other conditions. the values
shall be modified as follows:
a) For structural lightweight concrete having unit weights in the range 90120 lb per cu ft. the values shall be multiplied by (1.65  0.005 WJ but not less than
1.09, where wC is the unit weight in lb per cu ft.
b) Forf, other than 60,000 psi, the values shall be multiplied by (0.4 + fJlOO,oOO).
strength concrete and its normal strength counterpart. have been modified by ACI 435 (1978) and expanded in
The ratio of creep strain to initial elastic strain under Table 2.4 to include members that are supporting or at
sustained axial compression, for highstrength concrete, tached to nonstructural elements likely to be damaged
may be as low as one half that generally associated with by excessive deflections. The thickness may be decreased
lowstrength concrete (Ngab et al, 1981; Nilson, 1985). when computed deflections are shown to be satisfactory.
Based on a large number of computer studies, Grossman
2.4Control of deflection (1981, 1987) developed a simplified expression for the
Deflection of oneway nonprestressed concrete flex minimum thickness to satisfy serviceability requirements
ural members is controlled by reinforcement ratio limita (Eq. 4.17, Chapter 4).
tions, minimum thickness requirements, and span/deflec 2.4.3 Computed deflection limitationsThe allowable
tion ratio limitations. computed deflections specified in ACI 318 for oneway
2.4.1 Tension steel reinforcement ratio limitationsOne systems are given in Table 2.5, where the spandeflection
method to minimize deflection of a concrete member in ratios provide for a simple set of allowable deflections.
flexure is by using a relatively small reinforcement ratio. Where excessive deflection may cause damage to non
Limiting values of ratio p, ranging from to structural or other structural elements, only that part of
are recommended by ACI 435 (1978), as shown in Table the deflection occurring after the construction of the
2.2. Other methods of deflection reduction are presented nonstructural elements, such as partitions, needs to be
in Chapter 5 of this report. considered. The most stringent spandeflection limit of
2.4.2 Minimum thickness limitationsDeflections of l/480 in Table 2.5 is an example of such a case. Where
beams and one way slabs supporting usual loads in build excessive deflection may result in a functional problem,
ings, where deflections are not of concern, are normally such as visual sagging or ponding of water, the total
satisfactory when the minimum thickness provisions in deflection should be considered.
Table 2.3 are met or exceeded. This table (ACI 318,
1989) applies only to members that are not supporting or 2.5Shortterm deflection
not attached to partitions or other construction likely to 2.5.1 Untracked membersGross moment of inertia Ig
be damaged by excessive deflections. Values in Table 2.3 When the maximum flexural moment at service load in
435R10 ACI COMMITTEE REPORT
Table 2.4Minimum thickness of beams and oneway slabs used in roof and floor construction (ACI 435, 1978)
Members not supporting or not attached to nonstructural Members supporting or attached to nonstructural elements
elements likely to be damaged by large deflections likely to be damaged by large deflection
Simply One end Both ends Simply One end Both ends
Member supported continuous continuous Cantilever supported continuous continuous Cantilever
Roof slab l/22 l/28 1135 U9 l/14 VI8 l/22 115.5
Floor slab, and l/18 V23 l/28 l/7 1112 l/15 l/19 US
roof beam or
ribbed roof
slab
Floor beam or l/14 1118 l/21 l/5.5 l/10 lfl3 l/16 114
ribbed floor
slab
a beam or a slab causes a tensile stress less than the loading conditions. M is the maximum flexural moment
modulus of rupture,f, no flexural tension cracks develop along the span. The modulus of elasticity EC can be ob
at the tension side of the concrete element if the member tained from Eq. 2.4 for normalstrength concrete or Eq.
is not restrained or the shrinkage and temperature tensile 2.5 for highstrength concrete.
stresses are negligible. In such a case, the effective 2.5.2 Cracked membersEffective moment of inertia Ie
moment of inertia of the uncracked transformed section, Tension cracks occur when the imposed loads cause
II, is applicable for deflection computations. However, for bending moments in excess of the cracking moment, thus
design purposes, the gross moment of inertia, I@ resulting in tensile stresses in the concrete that are higher
neglecting the reinforcement contribution, can be used than its modulus of rupture. The cracking moment, MC,.,
with negligible loss of accuracy. The combination of ser may be computed as follows:
vice loads with shrinkage and temperature effects due to
end restraint may cause cracking if the tensile stress in (2.11)
the concrete exceeds the modulus of rupture. In such
cases, Section 2.5.2 applies. where yt is the distance from the neutral axis to the
The elastic deflection for noncracked members can tension face of the beam, and f, is the modulus of
thus be expressed in the following general form rupture of the concrete, as expressed by Eq. 2.1.
Cracks develop at several sections along the member
6=KMIZ (2.10) length. While the cracked moment of inertia, Ic,., applies
EcI, to the cracked sections, the gross moment of inertia, Ig,
where K is a factor that depends on support fixity and applies to the uncracked concrete between these sections.
DEFLECTION IN CONCRETE STRUCTURES 435R11
Several methods have been developed to estimate the Eq. 2.12 can also be simplified to the following form:
variations in stiffness caused by cracking along the span,
These methods provide modification factors for the flex
ural rigidity E I (Yu et al, 1960), identify an effective
moment of inertia (Branson, 1963), make adjustments to
the curvature along the span and at critical sections
(Beeby, 1968), alter the M / I ratio (CEB, 1968), or use a Heavily reinforced members wiIl have an Z, approx
sectioncurvature incremental evaluation (Ghali, et al, imately equal to Icr, which may in some cases (flanged
1986, 1989). members) be larger than Zg of the concrete section alone.
The extensively documented studies by Branson (1977, For most practical cases, the calculated Z, will be less
1982, 1985) have shown that the initial deflections q than Zg and should be taken as such in the design for
occurring in a beam or a slab after the maximum deflection control, unless a justification can be made for
moment M, has exceeded the cracking moment M,, can rigorous transformed section computations.
be evaluated using an effective moment of inertia Z, 2.5.2.2 Continuous beamsFor continuous mem
instead of I in Eq. 2.10. bers, ACI 31889 stipulates that Z, may be taken as the
2.5.2.1 Simply supported beamsACI 31889 r e  average values obtained from 2.12 for the critical
quires using the effective moment of inertia Z, proposed positive and negative moment sections. For prismatic
by Branson. This approach was selected as being suffi members, Z, may be taken as the value obtained at mid
ciently accurate to control deflections in reinforced and span for continuous spans. The use of midspan section
prestressed concrete structural elements. Bransons properties for continuous prismatic members is con
equation for the effective moment of inertia Z,, for short sidered satisfactory in approximate calculations primarily
term deflections is as follows because the midspan rigidity including the effect of
cracking has the dominant effect on deflections (ACI
435, 1978).
If the designer chooses to average the effective
moment of inertia Z,, then according to ACI 31889, the
following expression should be used:
where
%, = Cracking moment I, = 0.5 4(m) + 0.25 (G(1) + h(2)) (2.14)
Ma = Maximum service load moment (unfactored)
at the stage for which deflections are being where the subscripts m, 1, and 2 refer to midspan, and
considered the two beam ends, respectively.
Gross moment of inertia of section Improved results for continuous prismatic members
Moment of inertia of cracked transformed can, however, be obtained using a weighted average as
section presented in the following equations:
The two moments of inertia Zg and Z,, are based on For beams continuous on both ends,
the assumption of bilinear loaddeflection behavior (Fig.
3.19, Chapter 3) of cracked section. Z, provides a trans 4 = 0.70 Ze@) + 0.15 (I,(,) + h(2)) G95a)
ition between the upper and the lower bounds of Z and
I,,., respectively, as a function of the level of cracking, For beams continuous on one end only,
expressed as i&/Ma. Use of Z, as the resultant of the
other two moments of inertia should essentially give Z, = 0.85 I+) + 0.15 (I,(,)) (2.15b)
deflection values close to those obtained using the bi
linear approach. The cracking moment of inertia, I,, can When Z, is calculated as indiuated in the previous dis
be obtained from Fig. 2.3 (PCA, 1984). Deflections cussion, the deflection can be obtained using the mo
should be computed for each load level using Eq. 2.12, mentarea method (Fig. 3.9, Chapter 3) taking the mo
such as dead load and dead load plus live load. Thus, the mentcurvature (rotation) into consideration or using
incremental deflection such as that due to live load numerical incremental procedures. It should be stated
alone, is computed as the difference between these values that the Z, value can also be affected by the type of
at the two load levels. Z, may be determined using M,, at loading on the member (AlZaid, 1991), i.e. whether the
the support for cantilevers, and at the midspan for simple load is concentrated or distributed.
spans. Eq. 2.12 shows that I, is an interpolation between 2.5.2.3 Approximate Ie estimationAn approximation
the welldefined limits of Z and I,,. This equation has of the !8 value (Grossman, 1981) without the need for
been recommended by ACI Committee 435 since 1966 calculating Z,, which requires a priori determination of
and has been used in ACI 318 since 1971, the PCI Hand the area of flexural reinforcement, is defined by Eq. 2.16.
book since 1971, and the AASHTO Highway Bridge Speci It gives Z, values within 20 percent of those obtained
fications since 1973. Detailed numerical examples using from the ACI 318 Eq. (Eq. 2.12 and could be useful for
this method for simple and continuous beams, unshored a trial check of the Z, needed lor deflection control of
and shored composite beams are available in Branson the cracked sections with minimum reinforcement 200/fy,
(1977). The textbooks by Wang and Salmon (1992), and For MJM, I 1.6: .m
by Nawy (1990) also have an extensive treatment of the
subject. (2.16a)
435R12 ACI COMMlTTEE REPORT
n.0.
AS
0 1
Without compression steel With compression steel
Ig = (bbJh;/l2 2
+ b,,h3/12 + (bb,)hf(hhf/2yt)2 + b,,h(yth/2)
Fig. 2.3Moments of inertia of uncracked and cracked transformed sections (PCA, 1984)
DEFLECTION IN CONCRETE STRUCTURES 435R13
82
1 I H
f
b ta
For 1.6 5 MJM, I 10: The stresses, f,r, fs2 ,..., corresponding to the strains, cSl,
Q,***, may be obtained from the stressstrain curves.
(2.16b) Then, the reinforcing steel forces, TSl, TS2,..., may be
calculated from the steel stresses and areas. For example:
where
Tsl = f,l * 41 (2.18)
145/w,
&= d The distribution of concrete stress, over the com
pressed and tensioned parts of the section, may be ob
O*9h 0.4 + [&+A, (2*16c) tained from the concrete stressstrain curves. For any
given extreme compression fiber concrete strain, cc, the
resultant concrete compression and tension forces, C,
but, Ie computed by Eq. 2.16a and 2.16b should not be and C, are calculated by numerically integrating the
less than stresses over their respective areas.
Eq. 2.19 to 2.21 represent the force equilibrium, the
I, = 0.35 Ke I (2.16d) moment, and the curvature equations of a cracked sec
tion, respectively:
nor less than the value from Eq. 2.16b, 2.16c, and 2.16d,
where Ma is the maximum service moment capacity, com T,, + TS2 + . . . + c, + c, = 0 (2.19)
puted for the provided reinforcement.
2.5.3 Incremental momentcurvature methodToday A4 = C (A& cf,)i [c  (d)J + C, XT + C, A, (2.20)
with the easy availability of personal computers, more
accurate analytical procedures such as the incremental and
momentcurvature method become effective tools for +> (2.21)
computing deflections in structural concrete members
[Park et al, 1975]. With known material parameters, a The complete momentcurvature relationship may be
theoretical momentcurvature curve model for the determined by incrementally adjusting the concrete
cracked section can be derived (see Fig. 2.4). For a given strain, cc, at the extreme compression fiber. For each
concrete strain in the extreme compression fiber, E,, and value of ec the neutral axis depth, c, is determined by
neutral axis depth, c, the steel strains, cSl, eS2,..., can be satisfying Eq. 2.19.
determined from the properties of similar triangles in the Analytical models to compute both the ascending and
strain diagram. For example: descending branches of momentcurvature and loadde
cd. flection curves of reinforced concrete beams are pre
Cl = 2 EC (2.17) sented in Hsu (1974, 1983).
c
435R14 ACI COMMITTEE REPORT
where
g = elastic deformation calculated with the rigidity EcIg of
the gross cross section (neglecting the reinforcement)
= correction factor (see Fig. 2.6), which includes the
effects of cracking and creep
cm = geometrical mean percentage of the compressive
reinforcement
where
L, R = percentage of tensile reinforcement at the
left and right support, respectively
C = percentage of tensile reinforcement at the
maximum positive moment section
lL,lC, and lR = length of inflection point segments as indi
cated in Fig. 2.6 (an estimate of lengths is
generally sufficient)
The net restraining axial force and moment are obtained where k = 8 for simply supported beams and 2 for cantilever
by integrating over the depth: beams.
2.7.2 Effect of restraint on thermal movementIf a
h member is restrained from deforming under the action of
P =
A f dA = [ Ec t ( y )b ( y ) ] dy (2.35) temperature changes, internal stresses are developed.
0 Cracking that occurs when tensile stresses exceed the
concrete tensile strength reduces the flexural stiffness of the
h member and results in increased deflections under subse
M = A f ( y n ) dA = [ Ec t ( y )b ( y ) ( y n ) ] dy (2.36) quent loading. Consequently, significant temperature effects
0 should be taken into account in determining member stiff
ness for deflection calculation. The calculation of the effec
In order to obtain the total strains on the unrestrained cross tive moment of inertia should be based on maximum
section, P and M are applied in the opposite direction to the
moment conditions.
restraining force and moment. Assuming plane sections
In cases where stresses are developed in the member due
remain plane, axial strain a and curvature are given by:
to restrain of axial deformations, the induced stress due to
h
axial restraint has to be included in the calculation of the
P
a =  =  [ t ( y )b ( y ) ] dy (2.37) cracking moment in a manner analogous to that for including
AE c A the prestressing force in prestressed concrete beams.
0
h
APPENDIX A2
M
=  =  [ t ( y )b ( y ) ( y n ) ] dy (2.38) Example A2.1: Deflection of a fourspan beam
Ec I I A reinforced concrete beam supporting a 4in. (100
0
mm) slab is continuous over four equal spans 1 = 36 ft
The net stress distribution on the cross section is given by: (10.97 m) as shown in Fig. A2.1 (Nawy, 1990). It is
subjected to a uniformly distributed load wD = 700 lb/ft
P M(y n) (10.22 kN/m), including its selfweight and a service load
f n ( y ) =   E c t ( y ) (2.39)
A I wL = 1200 lb/ft (17.52 kN/m). The beam has the dimen
sions b = 14 in. (355.6 mm), d = 18.25 in. (463.6 mm) at
For a linear temperature gradient varying from 0 to t, the midspan, and a total thickness h = 21.0 in. (533.4 mm).
curvature is given by: The first interior span is reinforced with four No. 9 bars
435R17
D = 700 Ib/ft
l
A
<36 tMb36 ftA 36 ft+36 ft+j
t3 C D E
(a)
d = 3f in.
4 in.
I
f
 
t
4 in, (10.4 mm)    .
.c
1J
I
ll
2
C14 in.
I
k)
Fig. A2.1Details of continuous beam in Ex. A2.1 (Nawy, 1990, courtesy Prentiss Hall)
__
b = 6, + 16h, = 78 in.
(a)
(b)
Ratio MJMO:
D ratio = 0.41
1,170,ooo Deflection requirements (Table 2.5)
D + 50 percent L ratio = 36 x 12
= 2.4 in. >~SL = 1.0 in., O.K
483,000 180 180
= 0.22
1,170,000 + 0.5 x 2,000,000
483,000 l = 1.2 in. > 6, = 1.0 ., O.K.
D + L ratio =
3,170,000 360
Effective moment of inertia for support section:
Z, for dead load = 0.07 x 10,800 + 0.93 x 6900 1 = 0.9 in. < sLT = 2.4 in., N.G.
= 7170 in.4 480
Z, for D + 0.5L = 0.01 x 10,800 + 0.99 x 6900
= 6940 in.4  1 = 1.8 in. < S,, = 2.4 in., N.G.
Z, for D + L = 0.003 x 10,800 + 0.99 x 6900 240
= 6910 in.4 Hence, the continuous beam is limited to floors or
roofs not supporting or attached to nonstructural ele
Average effective Ie for continuous span ments such as partitions.
average Z, = 0.85 I,,, + 0.15 Z,
dead load: Z, = 0.85 x 15,200 + 0.15 x 7170 Application of CEBFIP method to obtain longterm
= 14,000 in.4 deflection due to sustained loads:
435R20 ACI COMMITTEE REPORT
+A
midspan Q = z = 4 x 1.0 = 0.0028 =pc = 0.40 in. (10 mm), say 0.5 in.
bd 78 x 18.25 Example (c): Simply supported tee section  Constant
A temperature over flange depth
6 x 1.0
support p = 2 = = 0.0235 = QL = pR I = 69319 in4 (2.88 x 10 mm4)
W
14 x18.25 n = 26.86 in (682 mm)
_t = 40 F (4.4 C)
^
o( = 0.0000055 in./in.p?
($ b; h = 36 in. (914 mm)
L = 60 ft. (18.4 m)
36
+=
Assuming that the location of the inflection points as
defined by Ir, and ZR for negative moment region, and Zc (a/O po x 96) OI  2~JwtY
for the positive moment region in Figure 2.16 are as fol
lows: (88,O;; x 0.0000055)/69319
L//L = L,/L = 0.21 and L&L = (l0.21 x 2) = 0.58 0.00000698
Also, assume pL = pR s = (+L2)/8 = (0.00000698 x 7202)/8
Hence, 0.45 in. (11.4 mm), say 0.5 in.
lom = 2(0.0235 x 0.2) + 0.0028 x 0.58
= 0.0094 + 0.0017 = 0.0111 = 1.11 percent CHAPTER 3DEFLECTION OF PRESTRESSED
CONCRETE ONEWAY FLEXURAL MEMBERS*
From Fig. 2.6, 7 = 2.4
From ACI Method Solution: Ig = 21,000 in.4 3.1Notation
Shortterm deflection, Ac = area of section
Ag = gross area of concrete section
6 = o.oo69wz4 = 5240 w As  area of nonprestressed reinforcement
EcZ < AP = area of prestressed reinforcement in tension
l
zone
+ 5.240(700 + 1200) = 0.47 in., say 0.5 in.
b width of compression face of member
21,000 bw = web width
LongTerm increase in deflection due to sustained load: c = depth of compression zone in a fullycracked
section
h3
6 LT = 2 Ml  2op,>a cgc = center of gravity of concrete section
0 cgs = center of gravity of reinforcement
= 1.52 x 2.4(1  20 x 0.0111)0.5 C = creep coefficient, defined as creep strain divided
= 1.35 in., say 1.4 in. (35 mm). by initial strain due to constant sustained stress
(1.41 in. by the ACI procedure solution) c, = PCI multiplier for partially prestressed section
c, = PCI multiplier for partially prestressed section
Example A2.2: Temperatureinduced deflections Ct = creep coefficient at a specific age
Cu = ultimate creep coefficient for concrete at loading
These design examples illustrate the calculation pro equal to time of release of prestressing
cedures for temperature induced deflections. d = distance from extreme compression fiber to cen
troid of prestressing steel
Example (a): Simply supported vertical wall panel  d' = distance from extreme compression fiber to cen
Linear temperature gradient troid of compression reinforcement
_t = 40 F (4.4 C) dp = distance from extreme compression fiber to cen
^
o( = 0.0000055 in./iIl./F troid of prestressed reinforcement
h = 4 in. (101 mm) e, = eccentricity of prestress force from centroid of
a) Single story span: L = 12 ft. (3.66 m) section at center of span
8 = (0.0000055 x 40 x 1442)/(4 x 8) ecr = eccentricity of prestress from centroid of
= 0.14 in. (3.6 mm), say 0.2 in. cracked section
b) Two story span: L = 24 ft. (7.32 m) e, = eccentricity of prestress force from centroid of
s = (0.0000055 x 40 x 2sS2)/(4 x 8) section at end of span
= 0.57 in. (14.5 mm), say 0.6 in. Ec = modulus of elasticity of concrete
Example (b): Simply supported tee section  Linear E ci = modulus of elasticity of concrete at time of ini
Temperature gradient over depth tial prestress
^o_( t == 0.0000055
40 F (4.4 C)
in./in./F
Es = modulus of elasticity of nonprestressed rein
forcement
h = 36 in. (914 mm)
L = 60 ft. (18.4 m)  simply supported
S = (0.0000055 x 40 x 7202)/(36x 8) * Principal authors: A Aswad, D. R. Buettner and E. G. Nawy.
DEFLECTION IN CONCRETE STRUCTURES 435R21
values to which a variability of k 20 percent or more in wire and tendon prestressing stegl reinforcement,
the deflection values must be considered. Deflection 3.3.1.2 Hightensilestrength prestressing barsHigh
calculations cannot then be expected to be calculated tensilestrength alloy steel bars for prestressing are either
with great precision. smooth or deformed to satisfy A S T M A 722 require
3.2.3 ScopeBoth shortterm and longterm transverse ments and are available in nominal diameters from J/e in.
deflections of beams and slabs involving prestressing with (16 mm) to 13/8 in. (35 mm). Cold drawn in order to raise
highstrength steel reinforcement are considered. Specific their yield strength, these b a r s are stress relieved to
values of material properties given in this chapter, such increase their ductility. Stress relieving is achieved by
as modulus of elasticity, creep coefficients, and shrinkage heating the bar to an appropriate temperature, generally
coefficients, generally refer to normal weight concrete al below 500 C. Though essentially the same stressrelieving
though the same calculation procedures apply to light process is employed for bars a s for strands, the tensile
weight concrete as well. This chapter is intended to be strength of prestressing bars has to be a minimum of
selfcontained. 150,000 psi (1034 MPa), with a minimum yield strength
Finally several of the methods described in this chap of 85 percent of the ultimate strength for smooth bars
ter rely solely on computer use for analysis. They do not and 80 percent for deformed bars
lend themselves to any form of hand calculation or ap 3.3.2 Modulus of elasticityIn computing shortterm
proximate solutions. The reader should not be deluded deflections, the crosssectional area of the reinforcing
into concluding that such computer generated solutions tendons in a beam is usually small enough that the
from complex mathematical models incorporating use of deflections may be based on the gross area of the con
concrete properties, member stiffness, extent of cracking crete. In this case, accurate determination of the modulus
and effective level of prestress somehow generate results of elasticity of the prestressing reinforcement is not
with significantly greater accuracy than some of the other needed. However, in considering timedependent deflec
methods presented. This is because of the range of varia tions resulting from shrinkage and creep at the level of
bility in these parameters and the difficulty in predicting the prestressing steel, it is important to have a reasonably
their precise values at the various loading stages and load good estimate of the modulus of elasticity of the pre
history. Hence, experience in evaluating variability of stressing reinforcement.
deflections leads to the conclusion that satisfying basic In calculating deflections under working loads, it is
requirments of detailed computer solutions using various sufficient to use the modulus o f elasticity of the pre
values of assumed data can give upper and lower bounds stressing reinforcement rather than to be concerned with
that are not necessarily more rational than present code the characteristics of the entire stressstrain curve since
procedures. the reinforcement is seldom stressed into the inelastic
range. In most calculations, the assumption of the modu
3.3Prestressing reinforcement lus value as 28.5 x lo6 psi (PCI Design Handbook, Fourth
3.3.1 Types of reinforcementBecause of the creep and Edition) can be of sufficient accuracy considering the fact
shrinkage which occurs in concrete, effective prestressing that the properties of the concrete which are more criti
can be achieved only by using highstrength steels with cal in the calculation of deflections are not known with
strength in the range of 150,000 to 270,000 psi (1862 great precision. The ACI Code tates that the modulus
MPa) or more. Reinforcement used for prestressed con of elasticity shall be established by the manufacturer of
crete members is therefore in the form of stressrelieved the tendon, as it could be less than 28.5 x lo6 psi.
or lowrelaxation tendons and highstrength steel bars. When the tendon is embedded in concrete, the free
Such highstrength reinforcement can be stressed to ade dom to twist (unwind) is lessened considerably and it
quate prestress levels so that even after creep and thus is unnecessary to differentiate between the modulus
shrinkage of the concrete has occurred, the prestress of elasticity of the tendon and that of singlewire rein
reinforcement retains adequate remaining stress to pro forcement (AC1 Committee 435, 1979).
vide the required prestressing force. The magnitude of 3.3.3 Steel relaxationStress relaxation in prestressing
normal prestress losses can be expected to be in the steel is the loss of prestress that occurs when the wires or
range of 25,000 to 50,000 psi (172 MPa to 345 MPa). strands are subjected to essentially constant strain over a
Wires or strands that are not stressrelieved, such as period of time. Fig. 3.2 relates stress relaxation to time
straightened wires or oiltempered wires, are often used t in hours for both stressrelieved and lowrelaxation ten
in countries outside North America. dons.
3.3.1.1 Stressrelieved wires and strandsStress The magnitude of the decrease in the prestress de
relieved strands are colddrawn single wires conforming pends not only on the duration of the sustained pre
to ASTM A 421 and stressrelieved tendons conform to stressing force, but also on the ratio fpilfw of the initial
ASTM A 416. The tendons are made from seven wires by prestress to the yield strength of the remforcement. Such
twisting six of them on a pitch of 12 to 16 wire diameters a loss in stress is termed intrinsic stress relaxation.
around a slightly larger, straight control wire. Stress If fpR is the remaining prestressing stress in the steel
relieving is done after the wires are twisted into the tendon after relaxation, the following expression defines
strand. Fig. 3.1 gives a typical stressstrain diagram for fPR for stressrelieved steel:
DEFLECTION IN CONCRETE STRUCTURES
435R23
250
t 
00
;;
._
9 150
H Grade 160 alloy bar
&
100
,l% Elongation
I I I I I I II 
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 in/in
Strain
Fig. 3.2Relaxation loss versus time for stressrelieved lowrelaxation strands at 70 percent of the ultimate (PostTensioning
Institute Manual, fourth edition)
435R24 ACI COMMITTEE REPORT
6
Stress
Temperature (*C)
100 C 60 C 40 C
11  7
18  lo
6 
16 g
8 5
14 7 
12  6  4
5 
l0  3
4 
* 3  2 
6 
4  2 
l
2 l
1 2 3 4 5 6
1 10 10 10 10 10 10
Time in Hours
Fig. 3.4Stress relaxation of stabilized strand at various tensions and temperatures (courtesy STELCO Inc., Canada)
stress for bonded prestressed members due to creep can fcsd = stress in concrete at the cgs level of the re
be defined as inforcement due to all superimposed dead
loads applied after prestressing is accom
plished
AfpCR = C (3.6) KCR should be reduced by 20 percent for lightweight
concrete.
where fcs is the stress in the concrete at the level of the Fig. 3.5 shows normalized creep strain plots versus
centroid of the prestressing tendon. In general, this loss time for different loading ages while Fig. 3.6 illustrates in
is a function of the stress in the concrete at the section a threedimensional surface the influence of age at load
being analyzed. In posttensioned, nonbonded draped ing on instantaneous and creep deformations. Fig. 3.7
tendon members, the loss can be considered essentially gives a schematic relationship of total strain with time
uniform along the whole span. Hence, an average value excluding shrinkage strain for a specimen loaded at a one
of the concrete stress between the anchorage points can day age.
be used for calculating the creep in posttensioned mem 3.4.3 Loss of prestress due to shrinkage of concreteAs
bers. A modified ACIASCE expression for creep loss with concrete creep, the magnitude of the shrinkage of
can be used as follows: concrete is affected by several factors. They include mix
proportions, type of aggregate, type of cement, curing
time, time between the end of external curing and the
AfPCR (3.7)
l application of prestressing, and the environmental condi
tions. Size and shape of the member also affect shrink
age. Approximately 80 percent of shrinkage takes place
where in the first year of life of the structure. The average value
KCR = 2.0 for pretensioned members of ultimate shrinkage strain in both moistcured and
= 1.60 for posttensioned members streamcured concrete is given as 780 x lo4 in./in. in the
(both for normal weight concrete) ACI 209R92 Report. This average value is affected by
 = stress in concrete at the cgs level of the re the duration of initial moist curing, ambient relative
f cs
inforcement immediately after transfer humidity, volumesurface ratio, temperature; and con
435R26 ACI COMMITTEE REPORT
TIME  DAYS
Fig. 3.5Creep curves for different loading ages at same stress level
Fig. 3.6Influence of age at loading on instantaneous and creep deformations (3D surface)
crete composition. To take such effects into account, the adjusting for relative humidity at volumetosurface ratio
average value of shrinkage strain should be multiplied by V/S, the loss in prestressing in pretensioned members is
a correction factor ysH as follows
Fig. 3.7Typical concrete strain versus time curve for constant stress applied at release time
l l
1 28 365 /
Time t (days)
3.5General approach to deformation considerations  centricity of the prestress, the length of the span, the size
Curvature and deflections and configuration of the cross section, boundary condi
In beamlike structures the curvature at any section, tions and the properties of the concrete. More specifical
defined as + = l/R, is the key element in calculating ly, the effect of critical variables may be summarized by
rotations and deflections. Based on geometry of the de the magnitude of the strain or stress gradient or the
flected shape, the two following expressions are derived, curvature at a section and its variation along the span.
see Fig. 3.9. The initial curvature at a particular section (Fig. 3.10)
is defined by
(3.12a) bi  %i = bf
= (3.13)
4i h EC1
A
6 BA = qmi!x (3.12b) in which tensile strains are considered positive, and M is
I the moment at the section.
In most cases, the amount of prestressing steel rein
where 4 is the curvature, 8 is the rotation and S is the forcement has a negligible effect on section properties
tangential deviation (deflection). These two expressions for shortterm deflections due to gravity loads.
are generalizations of the familiar Mohr or momentareas 3.5.1 Beams subjected to prestressing onlyStress and
theorems, and are applicable whether sections are strain distributions over the depth of a cross section of a
cracked or uncracked. When the material is linearly elas rectangular bonded beam immediately after application
tic, 4 can be replaced by M/EI. Software programs use of the prestressing force are shown in Fig. 3.10. It is
Simpsons rule to approximate the above integrals. Fig. assumed that there is a linear relationship between con
3.9(c) shows how the deflection y at any section can be crete stress and strain. Under normal conditions both of
calculated. these assumptions are reasonably correct. The stress at
Based on these fundamental principles, the designer any level is given by the wellknown relationships:
can calculate the curvature and rotation incrementally at
P
f =  A
any section and hence the deflection or camber of the MY (3.14)
prestressed beam at the critical sections. 7
Shortterm deflections are defined as those occurring
instantaneously under the application of any internal or
external force. The time element is assumed to be unim and A4 = Pe (3.15)
portant, no matter what the rate of loading, provided the
load is applied within a matter of hours.
In general, the principal variables affecting shortterm for use in Eq. 3.13 and where P is the prestressing force.
deflections of a prestressed concrete beam are the magni The stress and strain distributions in Fig. 3.11 depict
tude and distribution of the load, the magnitude and ec the conditions existing after a given time. The normal
DEFLECTION IN CONCRETE STRUCTURES 435R29
(c) Beam Elastic Curve Deflection y and Tangential Deviation s,, 6 Bc,
n d
(a)
I 1
C
STRESS
(b)
.
4
liii
$1
laJ
& sl
STRAIN
(c)
H 
pbt_l IEbt,._1
STRESS STRAIN
(a) w w
Fig. 3.11Stress and strain distribution at a time t after initial application of prestress
Aft %
h H
stresses on the section decrease as a result of a reduction ment strain which corresponds to a reduction in the pre
in the prestressing force while there is a general shift to stress. The loss in prestress causes a change in the stress
the right in the strain distribution accompanied by an in distribution over the depth of the section as indicated in
crease in the strain gradient. Fig. 3.12c and the corresponding change in the strain dis
These changes are caused by an interaction between tribution, Fig. 3.12d. Thus, the change in curvature is
creep and shrinkage of the concrete and relaxation of the
reinforcement. All of these effects progress with time and
(3.16)
continuously impact on each other. However, to simplify h
the calculation, it is preferable to treat these three types
of strains separately. The effect of the relaxation losses in the steel rein
Consider first the effect of shrinkage strains. It is forcement is quite similar to that of shrinkage. At a time
assumed that each element of concrete in the crosssec t there is a finite loss in the prestressing force which
tion shrinks equally. Thus, the shrinkage strain distribu creates a change in the curvature as explained above. The
tion after a time t is given in Fig. 3.12b. The distribution effects of the creep of the concrete are not as simple,
of shrinkage strain causes a reduction in the reinforce since the reduction in steel stress causes changes in the
DEFLECTION IN CONCRETE STRUCTURES 435R31
Curvature
A
TimeDependent Curvature
lnstanteneous Curvature
I Time
rate of creep strain. It is assumed that the amount of If the beam considered in t h e preceding paragraph is
creep strain at a given time is proportional to the stress. subjected to gravity load, the stre s distribution across the
Thus, the change in strain caused by creep is directly pro section at a given point along t e span may be as indi
portional to the instantaneous strain distribution (Fig. cated in Fig. 3.14d. Provided neither the concrete nor the
3.10c), which is directly related to the stress distribution. reinforcement is strained into the inelastic range, the
This change in the strain distribution involves a contrac stress distribution caused by the prestressing force (Fig.
tion at the level of the steel, hence, a reduction in pre 3.14b) can be superimposed on the stress distribution
stress. The reduction in prestress caused by creep, shrink caused by the transverse load on the uncracked trans
age, and relaxation decreases the normal stress, which in
formed section (Fig. 3.14c) to obtain the total stress
turn reduces the rate of creep.
A qualitative curvature versus time curve is shown in distribution shown in Fig. 3.14d.
Fig. 3.13. As in the case of shortterm deflections, the The strain distribution shown in Fig. 3.15b corresponds
magnitude of the deflection may be estimated by the to the stress distribution in Fig. 3.14c. It depicts the
magnitude of the stress gradient over the depth of the strains that would occur in an uncracked section under
section after release of prestress. If the stress gradient is the influence of only the transverse load. The shortterm
very small, then shrinkage and relaxation are bound to curvature is
dominate, in which case the beam may deflect downward.
However, under usual circumstances the stress gradient (3.17)
is large and creep dominates the deflection thus causing
the beam to move upward causing increased camber in
a simply supported case (ACI 435, 1979). where the subscripts b and t define the bottom and top
3.5.2 Beams subjected to prestressing and external loads fibers respectively.
1
435R32 ACI COMMITTEE REPORT
&
bi bt
STRAlN
(a) (b)
UP
A (Due To Prestress)
t
m
^
 + So
6 (Due To Load)
The changes in the curvature or in the deflection of vice load level or at a fraction of the service load level.
the beam caused by the combined prestress and the 3.5.3 Momentcurvature relationshipThe instantaneous
transverse load are henceforth determined by superposi momentcurvature relationship r a prestressed cross
tion. Both of these curvature distributions will change section is illustrated in Fig. 3.1 . Concrete can sustain
with time. The deflections corresponding to these two tensile stresses and contribute t o the carrying capacity of
imaginary systems are shown in Fig. 3.16. a member until cracking occurs at a moment MCP A mo
To get the net deflection, the deflections caused by the ment M, larger than the moment M,, produces
prestress and transverse load can be added as indicated curvature that can be defined a
by the (AG) curve. It is seen that the magnitude of the
beam deflection (and whether it deflects upward or
(3.18)
downward) depends on the relative effect of the prestress
and of the transverse loads. Ideally, a beam can be
designed to have a small camber at midspan at the ser where P is the prestressing force, and eC, is its eccentricity
435l?33
CRACKING
d is eccentricity of
prestress considering
tension stiffening,
"effective 0
M 0
0 0. 0CI
CURVATURE
relative to the centroid of the cracked section. The drop or camber, due to the effects of initial prestressing Pi and
in rigidity due to cracking is represented by the horizon member selfweight is generally in the elastic uncracked
tal line at the M,, level. For the prestressed section, both range. Therefore, the elastic formulas presented in Table
Icr and ycr (and in turn e,,) are dependent on the loading 3.2 could be used to calculate the instantaneous deflec
level, with the M4 becoming nonlinear after cracking. It tion of the members. The value of Pi is equal to the
is important to note that the shift in the centroid of the jacking force less the initial prestress loss due to an
cross section upon cracking results in larger prestressing chorage set, elastic shortening, and the relatively small
force eccentricity, ecr than the uncracked member eccen relaxation loss occurring between jacking and release
tricity. This fact is particularly significant in flanged time. Since Pi varies from section to section a weighted
members, such as double tees which are characterized by average may be used. An average initial loss of 410 per
the relatively low steel area ratio pf) and because con cent can be reasonably used in order to get fpi.
crete tensile strength is not zero, cracking does not Unless test results are available, the modulus of elas
extend to the neutral axis. In addition, uncracked con ticity of concrete can be estimated from the expression
crete which exists between cracks in the tension zone, recommended in ACI 318 (See Chapter 2, Section 2.3).
contributes to the stiffness of the member (tension For uncracked sections, it is customary to use the gross
stiffening). Taking this into account, the M4 diagram moment of inertia Ig for pretensioned members and the
becomes continuous, as indicated by line A in Fig. 3.17 net moment of inertia Z, for posttensioned members
and as is usually accepted in engineering practice (ACI with unbonded tendons.
318, 1989) and verified by numerous tests (Aswad, 1992). 3.6.2 Cracked members  Effective Ie methodIn pre
stressed concrete members, cracks can develop at several
3.6Shortterm deflection and camber evaluation in sections along the span under maximum load. The
prestressed beams cracked moment of inertia I,, applies at cracked sections
Several methods to estimate shortterm and longterm while the gross moment of inertia applies in between
deflections of prestressed concrete structural members cracks. ACI 318 (Section 18.4.2c) requires that a bilinear
are presented in this article. Included are procedures for momentdeflection relationship be used to calculate in
uncracked members and cracked members. stantaneous deflections when the magnitude of tensile
3.6.1 Uncracked membersWhen a concrete section is
subjected to a flexural stress which is lower than the stress in service exceeds 6&. A value of 12 & is per
modulus of rupture of concretef,, the section is assumed mitted when the immediate and longterm deflections are
to be uncracked and thus its behavior is linear. Under within the allowable limits. 1s is used for the portion of
this condition, the deflection is calculated by the basic moment not producing such tensile stress, while for the
principles of mechanics of elastic structures. In pre remaining portion of moment, Icr is used.
stressed concrete construction, the immediate deflection, The effective moment of inertia Ie for simply sup
435R34 ACI COMMITTEE REPORT
Table 3.2Shortterm deflection in prestressed concrete beams (subscript c indicates midspan, subscript e, support)
.
J1 w 
e&J       I     
P ec P
1
b
;~#
. I .a*
I
wb
1
6  (31 2  4b2)
=24EI s =
PI2
. 2
+ 9
W
P
 c9c         
c9s 1
a P
I  t8 J?
. l
Pec12 12
6 = SWI @ 512 6fV
384EI c48 8EI % 8
DEFLECTION IN CONCRETE STRUCTURES 435R35
Load b,
I
I Postcracking i Postserviceability I
Deflection A
Fig. 3.18Loaddeflection relationship in prestressed beam: Region precracking stage; Region postcracking stage;
Region III, postserviceability stage
ported beams, cantilevers, and continuous beams between The effective moment of inertia I, in Eq. 3.19a and b
inflection points is given in ACI 31889, Section 9.5.2.3, thus depends on the maximum moment M, due to live
but with the modified definitions of M,, and A4, for pre load along the span in relation to the cracking moment
stressed concrete as follows: capacity M,, of the section due to that portion of the live
load that caused cracking.
In the case of beams with two continuous ends ACI
(3.19a) 31889 allows using the midspam Z,. However, more ac
curate values can be obtaine d when the section is
uncracked using the following expressions as discussed in
or Chapter 2, Section 2.5.
Deflection 6
Fig. 3.21Strain distribution and curvature at controlling loading stages (Nawy, 1989): a) initial prestress; b) effective
prestress after losses; c) service load; d) failure. If section is cracked at service load, Fig. 3.21c changes to reflect tensile
strain at the bottom fibers (see Fig. A3.2)
the cracking load (see Eq. 3.18), including the pre 4) Failure:
stressing primary moment M,, about the centroid (center
ofgravity of the concrete) of the section under consid
eration. Eq. 3.21 can be rewritten to give (3.226)
4= E* h Ecb (3.22c) refined. This chapter presents in detail the simplified PCI
multipliers method even though it is sometimes more
conservative, since it is the most commonly used for
DEFLECTION IN CONCRETE STRUCTURES 435R39
deflection and camber calculation in normal size and When such reinforcement is used, a reduced multiplier
span prestressed beams such as double tees, hollow core C, can be used as follows, to reduce the values in Table
slabs and AASHTO type beams. Numerical examples on 3.4,
its use are given in the appendix.
It is worthwhile noting that prestressed building pro
(3.23a)
ducts generally comply with the deflection limits in Table
9.5(b) of ACI 31889. Industry and local practices, how
ever, may be more stringent, such as requiring that and
double tee or hollow core slabs should have a slight
camber under half of the design live load. It is also good 6* = qa, (3.23b)
practice to never allow a calculated bottom tension stress
due to sustained loads. 3.7.2 Incremental timesteps methodThe incremental
Other selected methods are briefly described and ref timesteps method is based on combining the computa
erence made to existing literature for details on camber tions of deflections with those of prestress losses due to
and deflection design examples in those references, that timedependent creep, shrinkage, and relaxation. The
the designer can choose for refined solutions. design life of the structure is divided into several in
3.7.1 PCI multipliers methodThe determination of creasingly larger time intervals. The strain distributions,
longterm camber and deflection in prestressed members curvatures, and prestressing forces are calculated for each
is more complex than for nonprestressed members due to interval together with the incremental shrinkage, creep,
the following factors: and relaxation losses during the particular time interval.
1. The longterm effect of the variation in pre The procedure is repeated for all subsequent incre
stressing force resulting from the prestress losses. mental intervals, and an integration or summation of the
2. The increase in strength of the concrete after re incremental curvatures is made to give the total time
lease of prestress and because the camber and deflection dependent curvature at the particular section along the
are required to be evaluated at time of erection. The PCI span. These calculations should be made for a sufficient
Design Handbook, fourth edition, provides a procedure number of points along the span to be able to determine
wherein the short term deflections (calculated using con with reasonable accuracy the form of the momentcur
ventional procedures) are multiplied by factors (multi vature diagram.
pliers) for various stages of the deflection (erection, The general expression for the total curvature at the
final), for deflections due to prestress dead and applied end of a time interval can be expressed as
loads and for composite and noncomposite sections to
obtain long term deflections. These multipliers vary from (3.24)
1.80 to 3.00, as shown in Table 3.4 (PCI Design Hand
book, 4th ed., 1993).
Shaikh and Branson (1970) propose that substantial
reduction can be achieved in longterm deflections by the
addition of nonprestressed mild steel reinforcement.
I
435R40 ACI COMMITTEE REPORT
where where
Pi = initial prestress (at transfer) before losses E,,(t) = time adjusted modulus
eX = eccentricity of tendon at any section along the
span
Subscript nl = beginning of a particular time step (3.29)
Subscript n = end of the aforementioned time step
cn_l, c, = creep coefficients at beginning and
end, respectively, of a particular time in which
step E,(tl) = modulus of elasticity of concrete at start
= prestress loss at a particular time in of interval and x is an aging coefficient
ClCl1
terval from all causes c,(t) = creep coefficient at end of time interval
Obviously, this elaborate procedure is usually justified 3.7.3 Approximate timesteps methodThe approximate
only in the evaluation of deflection and camber of timesteps method is based on a simplified form of sum
slender beams or very longspan bridge systems such as mation of constituent deflections due to the various time
segmental bridges, where the erection and assembly of dependent factors. If C, is the longterm creep coeffi
the segments require a relatively accurate estimate of cient, the curvature at effective prestress P, can be
deflections. From Eq. 3.24, the total deflection at a defined as
particular section and at a particular time t is
PJ$
s, = c#+ ke2 (3.25)
@$ = g +
cc
(pi 
cc
(3.30)
easily evaluated for deflection and camber using this The final deflection under PG is
incremental numerical summation method. Detailed ex
ai +
amples are given in the textbooks by Nawy, 1989 and
Nilson, 1987.
The total camber (t) or deflection (4) due to the pre
stressing force can be obtained from the expression & =
6, = ai + (a,  Se) 
t
c
2
6
1 C, (3.31a)
&_1 + #n so that
where
tjT = 6,  
. l 1
isi + 6,
2 I
cu + (a, + b&l + CJ + 6,
(3.32b)
$1(t) = midspan curvature at time t
= The approximate timesteps method originally pre
$26) support curvature at time in which
sented by Branson and Ozell, 1961, and ACI 435, 1963,
tends to yield in most cases comparable results to the
PCI multiplier method. Detailed examples are given in
the text books by Branson, 1977; Libby, 1984; Nilson,
DEFLECTION IN CONCRETE STRUCTURES 435R41
1987; and Nawy, 1989. to 0.8 and C, is the creep coefficient. Values of x and C,
3.7.4 Axial strain and curvature method (GhaliFavre) are given as functions of to and t in ACI 20992. After
This approach gives a procedure for the analysis of cracking, the concrete in tension is ignored and only the
instantaneous and longterm stresses and strains in rein area of concrete in the compression zone of depth, c, is
forced concrete crosssections, with or without pre included in calculating the properties of the trasformed
stressing but considering cracking. Slope of the strain sections. This method is detailed in Ghali and Favre,
diagram is set equal to curvature (see Section 2.6.3.2, 1986; Ghali (1986); and Elbadry and Ghali (1989).
Chapter 2), which can be used to calculate the change in 3.7.5 Prestress loss methodIt is assumed in this
deflection. The method does not require determination method that sustained dead load due to self weight does
of prestress losses. It introduces, as in the Naaman not produce cracking such that the effects of creep,
approach, an aging coefficient that adjusts the modulus shrinkage, and relaxation are considered only for un
of concrete E, between time limits t, and t. After cracked cross sections. Additional stress in the concrete
cracking, the concrete in tension is ignored and only the caused by live load may result in cracking when the ten
concrete in the compression zone of depth c is included sile strength of concrete is exceeded. Whether cracking
in calculating the properties of the transformed section. occurs, and the extent to which it occurs when the live
A crosssection provided with prestressed and non load is applied depend upon the magnitude of the pre
prestressed reinforcement of areas Ap and As respec stress losses.
tively, is subjected at time fa to a flexural moment M and The method recommends stress loss coefficients due
to normal force N. Analysis is required for the stresses to creep, shrinkage and relaxation such that the change
and strains which occur at the initial time to and at t > in the prestressing force AP, is given by the following:
to after development of creep and shrinkage in the con
crete and relaxation in the prestressed steel reinforce
ment. M and N are taken as the internal moments and &P, = A&~ + AfpcR  ApAfpRl (3.34a)
forces due to all external forces plus the prestressing
introduced at time to. The transformed section is com A set of multipliers, as listed in Table 3.5, are applied
posed of the area of concrete and the areas+$ andA, of to the deflections due to initial prestress, member self
the reinforcement multiplied by the respective modular weight, superimposed dead load, and timedependent
ratios, npr or n; where prestress loss in a similar fashion to the multipliers used
in the PCI multipliers method. Thus, total deflection
after prestress loss and before application of live load
(3.33a) becomes
2 48
W
% y
Fig. A3.1Noncomposite beam geomety Example A3.1 (Nawy, 1989. Courtesy Prentice Hall)
= +500  392 = 108 psi (T) < 184 psi, OK Total moment M, = MD + MSD + ML
= 5,170,000 + 7,610,000
= 12,800,000 inlb. (1443 kNm)
fb =
Liveloadf, = 6~~l\~ = 530 psi (C) = 2740 + 2670 = 70 psi (C), OK
t
Hence, the section is uncracked and the gross moment
of inertia Ig = 169,020 in.4 has to be used for deflection
Liveload fb = 6,970,000 = 1450 psi (T)
4803
435R44 ACI COMMITTEE REPORT
Midspan Support
fl fa fl fb
Prestress Pi only +500 3310 +100 2200
At transfer and WD +108 2230 +100 2200
At service load 560 70 +80 1810
1 psi = 6.895 MPa
6, = 5 x 67.9(65 x 12)2/384 x 3.49 x lo6 x 169,020 Hence, camber only at erection (30 days) can be rea
= 0.55 in.4 sonably assumed
= 1.50  0.26 = 1.24 in. t
Thus, the net camber at transfer is 1.50 t + 0.55
Solution (b)
= 0.95 in.? (24 mm).
Alternate solution by incremental moment curvature
4Total immediate deflection at service load of method
uncracked beam
a) Superimposed dead load deflection: P, at 30 days after transfer is 335,000 lb. So 30 days
435R45
%b = 949 x 10d in./in. Superimposing the strain at transfer on the strain due
to prestress loss gives the strain distributions at service
Support: f, = +lOO psi load after prestress loss due to prestress
_ only, as shown
fb = 2200 psi inn Fig. A3.2.
% = +28 x lo4 in./in. From Fig. A3.2:
%b = 631 x 10 in./in.
Midspan curvature
(1 psi = 6.895 MPa)
785  118 x lo_6
(ii) Due to prestressing and selfweight (Pi + W,) , = = 18.8 x 10m6 rad/in.
48
Midspan: fi = +108 psi E: = 31 x 106 in./in.
fb = 2230 %b = 640 x 1Oa in./in. Support curvature
Support: same as in (i)
Strain change due to prestress loss , = 522  23 x 106 = 11.4 x lo6 rad/in.
48
 hp = 70,000 lb
From Table 3.2, for a = t/2, the beam camber after
E,i = 3.49 x 10s6 psi losses due only to P, is
0+
Midspan section:
P
8, t = 4, $ (4,  4J 24
\+i=
949 +I64
jly 785
5 +23
Fig. A3.2Strain distribution across section depth at prestress transfer in Ex. A3.1)
Multiplier
Load Transfer S,, (in.) Multiplier Erection 6, (in.) (noncomposite) Final S,, (in.)
(1) (2) (3)
PCI section
lOLDT32 + 2
(128 D1)
Noncomposite Composite
Scf for 2 in. slab top fibers = 8152 in.3 from data.
s& for 2 in. slab bottom fibers = 10,337 in.3 from
before for top of precast section.
=
Stress f, at top slab fibers = ns
= +433  1163 = 730 psi < 0.45f, < 2250 psi, OK s:
f, = 7.5 K = 7.54%ilRj = 530 p s i = +1.7 in. (45 mm) 4 (as an average value)
fb at service load = 814 psi (5.4 MPa) in tension (from When the concrete 2 in. topping is placed on the pre
before). cast section, the resulting topping deflection with Ig =
Hence, the section is cracked and the effective Ie from 59,700 iIL4:
Eq. 3.19(a) or (b) should be used.
5x20.8(72~12)~ =063 i n . 4
SD =
dP = 18.73 + 10.02 + 2 (topping) = 30.8 in. (780 mm) 384(4.03 x 10,59,700  l
PCI multiplier
Load Transfer S p , in. PCI multipliers S
o30, in. (composite) a,,,, in.
WL
+1.88J +1.881
Final S
o 0.74t +1.27.J +2.071
S
oCT
JwJ4
= = 5 x 9.6(72 x 12)4 2 1 yield strength of reinforcement
plate or slab thickness
384 E,Z, 384(4.03 x 107 11,110 = moment of inertia
=
I
k coefficient
= 1.56 in. 4 L = live load
M = bending moment
Total live load deflection prior to prestress losses = N = number of shored and reshored levels
SLl + %r = 0.32 + 1.56 = 1.88 in. 4 R = applied load/slab dead load ratio
From before, ai = 0.70 t =
thickness
Net shortterm deflection prior to prestress loss is w, w, = intensity of transverse load per unit area
6 T&d = 0.70 + 1.88 = 1.2 in. & 4Y = coordinate axes
Y, = distance from neutral axis to extreme tension
5Longterm deflection (camber) by PCI multipliers fiber
When the 2 in. concrete topping is placed on the pre v Poissons ratio
cast section, the resulting topping deflection with I8 = CY = coefficient
59,720 in.4 is = deflection
; = curvature
5 x 20.8(72 x 12)4 A = longtime multiplier
SD =
= +0.63 in. (16 mm) ultimate shrinkage
384(4.03 x 106) 59,720 Eshm =
P = reinforcement ratio
Using PCI multipliers at slab topping completion stage
(30 days) and at the final service load (5 years), the tabu
lated deflection values can be gotten from the above
table.
l Principle authors: A. Scanlon and C. T. Hsu.
Hence, final deflection = 2.1 in (56 mm) 4.
II
DEFLECTION IN CONCRETE STRUCTURES 435R51
mined for factored loads according to ACI 318 multiplied Ml> M2 = end moments per unit width
by the ratio of service load to factored load. Mm = midspan moment per unit width
Fig. 4.1 shows a rectangular panel in a column (Positive M,, A4r, or M2 produce tension at bottom
supported twoway slab system. The dotted areas re fiber.)
present a set of crossing beams from which column strip Using this procedure the deflection of each column
deflection, S,, and middle strip deflection, &, can be strip (8,) and of each middle strip (6,) can be calcu
obtained. Each beam can be treated as a strip of unit lated. The midpanel deflection, a,, is obtained by
width for which end moments, midspan moment, and adding the column and middle strip deflections.
flexural rigidity properties can be obtained. Note that, by
definition, end moments are those at the faces of sup
ports, such as column or column capital faces, and that
the beam span is the clear span between the faces of For cantilever slabs the rotation at the support must
such supports. be included.
Once the end moments and midspan moment have An earlier version of the equivalent frame method for
been obtained for a column or middle strip, the de calculating deflections proposed by Vanderbilt, Sozen
flection for the strip can be calculated, using the elastic and Siess (1965) considers the midpanel deflection as
beam deflection equation: the sum of a column strip deflection, cantilever deflection
of a portion of the middle strip extending from the
5 5 column strip, and the midpanel deflection of a simply
6 = W = O.l(M, = M2)] (4.4) supported rectangular plate. The procedure developed by
48 EI m
Nilson and Walters (1975), based on the equivalent
where frame method, is similar to the method outlined above
en = clear span except that a reference deflection is calculated for the
DEFLECTION IN CONCRETE STRUCTURES 435R53
total panel width. Deflections for column and middle application of classical anisotropic theory to analysis of
strips are then obtained from this reference deflection twoway reinforced concrete slabs is given in the text by
using lateral distribution factors based on relative M/EI Timoshenko and WoinoswkyKrieger (1959). More re
values. A numerical example (Nawy, 1990) calculating the cently, procedures have been proposed for including
expected deflection limits using this procedure is given in cracking in finite element analysis and in the crossing
Appendix A4.1. The resulting values are applicable in beam analogies for twoway slabs.
lieu of Table 4.2. The effective moment of inertia, Ie, concept devel
Ghali (1989) calculates the deflection at midspan of oped originally by Branson (1963) for beams can be ap
a column or middle strip from values of curvature cal plied directly to the column and middle strips in the
culated on the basis of compatibility and equilibrium at crossing beam analogies described in Section 2.2.2 for
the midspan and supports using the relationship: elastic uncracked plates. In Eq. 4.4 the crosssection
stiffness, El, becomes E$,, using the usual averaging
procedures given in ACI 318 for I,, calculated at both
6 = $ (4L = 104, = 4J (4.6) positive and negative moment locations. Kripanarayanan
and Branson (1976) presented an extension of the Nilson
where +L, c&,, 4R are, respectively, the curvatures calcu and Walters equivalent frame procedure to include crack
lated from analysis of sections at the left end, center, and ing using the Ie procedure.
right end of column and middle strips and e is the dis A review of the more sophisticated cracking models
tance between the two ends. This relationship is based on proposed for finite element analysis of slabs is given in
the assumption that variation of curvature over the the report of an ASCE Task Committee (1982). A simple
length .! is parabolic, The effects of cracking, creep, and generalization of Bransons effective moment of inertia
shrinkage are accounted for in the analysis for + at each concept to twoway systems has been suggested by Scan
section. In the absence of prestressing, simplification can lon and Murray (1982) and implemented in a modified
be made by use of multipliers and graphs (Ghali, 1989; version of a linear elastic plate bending finite element by
GhaliFavre, 1986) that also account for the cracking, Graham and Scanlon [1986(a)].
creep and shrinkage effects. 4.3.3 Restraint crackingIn twoway reinforced con
crete slabs built monolithically with supporting column
4.3.1.3 Finite element methodThe finite element and wall elements, inplane shortening due to shrinkage
method can be used to analyze plates with irregular sup and thermal effects is restrained. The restraint is pro
port and loading conditions. Effects of beams and col vided by a combination of factors, including embedded
umns can be included and a number of general purpose reinforcement, attachment to structural supports, and
computer programs are available for elastic analysis of lower shrinkage rates of previously placed adjacent
plate systems. panels when slab panels are placed at different times.
The plate is divided into a number of subregions or Nonlinear distribution of free shrinkage strains across the
elements. Within each element the transverse displace crosssection may also be a contributing factor.
ment is expressed in terms of a finite number of degrees Service load moments in twoway slabs are often of
of freedom (displacements, slopes, etc.) specified at the same order of magnitude as the codespecified crack
element nodal points. In other words, the continuous ing moment, i%fcr Deflection calculations made using the
displacement function, a&y), is approximated by another codespecified modulus of rupture will often result in an
function with a finite number of degrees of freedom. uncracked section being used when cracking may actually
Based on the assumed displacement function and the be present due to a combination of flexural stress and
given stressstrain, or momentcurvature relationships restraint stress.
(such as Eq. 4.2 for elastic plates), the element stiffness
ACI 318 specifies the modulus of rupture for deflec
matrix can be derived. The stiffness matrix of the entire
slab is then assembled. The solution for displacements tion calculations as 7.5 K psi (0.62 K MPa). Labora
and internal moments proceeds using the standard matrix tory test data, summarized in ACI 209R, indicate values
analysis techniques applicable for solving equilibrium ranging from 6 to 12 E psi (0.5 to 1.0 fl MPa).
equations, as outlined in a number of textbooks (e.g. For slab sections with low reinforcement ratios,
Cook 1974; Gallagher 1975; Zienkiewicz 1977). Although approaching minimum reinforcement, the difference be
the method is becoming increasingly popular in engineer tween cracked and uncracked flexural stiffness is signi
ing practice, some skill is required in selecting an ficant. It is important, therefore, to account for effects of
appropriate finite element model, developing an appro any restraint cracking that may be present. Unfortunate
priate mesh, preparing computer input data, and inter ly, the extent of restraint cracking is difficult to predict.
preting the results. To account for restraint cracking in twoway slabs, Ran
4.3.2 Effect of crackingThe procedures outlined gan (1976) suggested that column strip deflections be
above are applicable to linear elastic isotropic plate based on the moment of inertia of a fully cracked sec
systems and must be modified for concrete slabs to in tion, Zcp and that middle strip deflections be based on (Ig
clude the effect of cracking on flexural stiffness. An early + Z&2. Good agreement was reported between calcu
435R54 ACI COMMITTEE REPORT
lated and field measured deflections. longterm deflections; namely, by detailed computations
A more general approach was proposed by Scanlon and by the ACI multiplier methods.
and Murray (1982). They suggested that the effect of re 4.3.4.1 Detailed calculationsEffects of creep
straint cracking be included by introducing a restraint deflection and shrinkage warping may be considered
stress,f,, that effectively reduces the modulus of rupture separately using procedures outlined in ACI 209R (82)
for calculating A&+ i.e. (1986), based on the work of Branson, Meyers and Kri
panapayanan (1970), and Branson and Christiason
(1971).
Deflection due to creep is obtained from
where f, = f,  f,
scp = kr ct si (4.8)
where
A value of 4 cpsi (0.33 & MPa), or about half of Ct = time dependent creep coefficient representing
the value specified in ACI 318, was proposed for the re creep strain at any time t in days after load
duced effective modulus of rupture. This approach was application
investigated by Tam and Scanlon (1986) and has pro k, = factor to account for compression reinforce
duced good correlation between calculated deflection and ment and neutral axis shift
reported mean fieldmeasured deflections [Jokinen and immediate deflection due to dead load plus
Scanlon 1985; Graham and Scanlon 1986(b)]. sustained live load, including effects of crack
Ghali (1989) has also used the idea of reduced mod ing
ulus of rupture and demonstrates the calculation of re
straint stress due to reinforcement in the presence of The general form of C, given by ACI 209 is
uniform shrinkage.
An additional consideration is that the moments used
in design for strength are based on some redistribution c, = (o~60 c
Flat roofs not supporting or attached to Immediate deflection due to live load L
nonstructural elements likely to be damaged
P*
by large deflections
ii6
Floors not supporting or attached to non Immediate deflection due to live load L
structural elements likely to be damaged by
P
large deflections
360
Roof or floor construction supporting or That part of the total deflection occurring
attached to nonstructural elements likely to after attachment of nonstructural elements
be damaged by large deflections fi
(sum of the longtime deflection due to all
sustained loads and the immediate deflec 480
tion due to any additional live load)7
Roof or floor construction supporting or
attached to nonstructural elements not
likely to be damaged by large deflections
* Limit not intended to safeguard against ponding. Ponding should be checked by suitable calculations of deflection, including added deflections due to
ponded water, and considering longterm effects of all sustained loads, camber, construction tolerances, and reliability of provisions for drainage.
t Limit may be exceeded if adequate measures are taken to prevent damage to supported or attached elements.
$ Long time deflection shall be determined in accordance with 9.5.2.5 or 9.5.4.2, but may be reduced by amount of deflection calculated to occur before at
tachment of nonstructural elements. This amount shall be determined on the basis of accepted engineering data relating to time deflection characteristics of
members similar to those being considered.
r But not greater than tolerance provided for nonstructural elements. Limit may be exceeded if camber is provided so that total deflection minus camber
does not exceed limit.
based on the ACI 318specified value of 7.5 K for sion developed by Rangan (1982) for maximum allowable
modulus of rupture, and ACI 209R Eq. 1517 for creep spantodepth ratio for beams. Rangans equation in
and shrinkage deflection. Construction loads due to volves rearranging the basic equation for beam deflection
shoring and reshoring were also considered. calculations,
2) When the calculations were based on a reduced
modulus of rupture to account for restraint cracking, the s = 6, + assus
ACI 318 limit of e/480 on incremental deflection was
exceeded for slab panels with aspect ratios less than 1.5. where
An increase of 10 percent over the current minimum St = deflection due to variable part of live load
thickness value for square panels was suggested to obtain = kw,t4fEJe
calculated deflections within the allowable limits. The 6 szls = total deflection due to sustained load in
suggested increase in minimum thickness decreases lin cluding sustained part of live load
early to zero for a panel with an aspect ratio equal to 1.5. = A (kw,@/EF,)
The results of this study suggest that the ACI mini A# = longterm multiplier
mum thickness equations will provide satisfactory service
Replacing Ie by cubd3, where the term a! gives an ap
ability in most cases, confirming the satisfactory perfor
proximation for I e as a function of the reinforcement
mance of slabs designed and built according to the re
ratio p, Eq. 4.13 can be rewritten as
quirements in ACI 318 prior to the 1989 edition. When
more stringent than normal deflection limits are required,
a thicker slab should be used. Other means to increase (4.14)
the slab stiffness, such as the addition of beams, can also
be considered.
Recently, attempts have been made to develop criter If S/e is given as the maximum permissible deflection
ia for spantodepth ratios or minimum thickness of slabs tospan ratio, the corresponding maximum span to effec
that explicitly include the effects of such parameters as tive depth ratio can be obtained from
livetodead load ratio, permissible deflectiontospan
ratio, effect of cracking, sustained load level, and time abE,
1P
between construction and installation of nonstructural (4.15)
elements. Two such approaches are described in the WC + Aw,
following paragraphs.
Gilbert (1985) extended to twoway slabs an expres where
DEFLECTION IN CONCRETE STRUCTURES 435R57
k, = a combination of factors to account for sup spans, and reduced construction time due to earlier
port conditions and effect of beam flanges removal of form work. In addition, the use of postten
sioning enables the engineer to better control deflections
Gilbert extended Eq. 4.15 by adding a slab system and cracking at service loads.
factor k, to account for twoway action, i.e. 4.5.2 Basic principle for deflection controlThe con
cept of load balancing [Lin (1963)] is often used to make
lb an appropriate choice of tendon profile, prestressing
(4.16) amounts and tendon distribution in twoway prestressed
and posttensioned floor systems. Service live loads,
rather than total dead plus live loads, should be used to
The factor k2 was developed for a variety of condi evaluate deflection of the slab. Load balancing from the
tions from parameter studies using a sophisticated finite transverse component of the prestressing force would
element model. Eq. 4.16 involves an iterative procedure have to be used to neutralize the deadload deflection or
since the reinforcement ratio required to determine LY, even induce camber if the live load is excessively high.
and the dead load are initially unknown. ACI 318 requires that both immediate deflection, due to
A somewhat simpler expression for beams was devel live load and longterm deflection due to sustained loads
oped by Grossman (1981, 1987); it was based on a large be investigated for all prestressed concrete.
number of computergenerated beam deflection calcula 4.5.3 Minimum stab thickness for deflection controlIn
tions. Grossmans minimum thickness equation is given choosing the slab thickness, the engineer must consider
by deflection control, shear resistance, fire resistance, and
corrosion protection for the reinforcement. While ACI
318 requires deflection calculations for a preliminary
(4.17) estimate of the twoway slab thickness, it is usual to
determine a minimum thickness for deflection control
Correction factors are given for variations in support based on traditional spandepth ratios as suggested by
conditions, d/h, fy, and concrete density. The term c was the PostTensioning Institute (1976). As an approximate
developed from the computer analyses and depends on guideline, a spantodepth ratio of 50 and 45 may be
the load levels and construction methods used. For heavi used for twoway continuous slabs with and without drop
lyloaded members, a limiting value of c = 4320 was pro panels, respectively. A minimum drop panel of l/6 span
posed by Grossman for heavily loaded members. Smaller length each way is recommended. A spantodepth ratio
thicknesses can be obtained if the required reinforcement of 55 for a twoway slab with twoway beams is reason
ratio for less heavily loaded members is known and is able. For waffle slabs, a lower value of 35 is recom
used to obtain a larger revised value of c from mended. Gilbert (1989) also gave a simple formula to
Grossmans data. express the maximum spantodepth ratio of twoway
The term C,, given by posttensioned floor systems. The expression provides an
initial estimate of the minimum slab thickness required
ID + L to limit deflections to some preselected maximum value.
c, = (4.18) 4.5.4 Methods for defection calculationsControl of
D + L
deflection in a twoway prestressed and posttensioned
accounts for both the liveloadtodeadload ratio, L/D, floor system is dealt with in Section 9.5.4 and Chapter 18
and the net longterm multiplier X, for deflections of ACI 318. However, unlike the twoway slab construc
occurring after installation of partitions in buildings. tion in a nonprestressed case, there are no provisions
Although developed for beams only, the equation containing requirements to determine a minimum thick
could be extended to twoway systems using a slab ness for twoway posttensioned slabs. To compute the
system factor similar to that given by Gilbert (1985). deflections, the engineer may apply the methods pro
posed for nonprestressed construction with appropriate
4.5Prestressed twoway slab systems treatment of the effects of prestressing.
4.5.1 IntroductionTwoway posttensioned concrete The accurate determination of deflections of twoway
slabs are widely used for the floor systems of office posttensioned slabs is a complex operation involving
buildings, parking garages, shopping centers, and lift slabs considerations of the boundary conditions, loading pat
in residential buildings. Due to its general economy and terns and history, changes of stiffness due to local
ability to satisfy architectural requirements, the post cracking, and loss of prestress due to creep, shrinkage,
tensioned concrete flat plate has been widely adopted in and relaxation. For practical design purposes, it is usually
the United States as a viable structural system. This type adequate to use simple approximate expressions to esti
of construction has grown over the past 25 years, despite mate the deflection, such as the crossing beam methods
competition from other floor systems. The popularity of including the equivalent frame approach described earl
this type of construction is primarily due to the econ ier, The midpanel deflection can be approximated as the
omies that result from reduced slab thickness, longer sum of the centerspan deflections of the column strip, in
435R58 ACI COMMITTEE REPORT
one direction, and that of the middle strip, in the ortho to total load minus that due to dead load. Under mono
gonal direction. A detailed numerical example is given in tonic loading, two effective moment of inertia values
Nawy (1989). The use of gross stiffness values to co should be used to calculate the deflections at the two
mpute deflections is justifiable only if the tensile stresses different load levels, as shown in Fig. 4.2.
in the concrete remain below the cracking stress. If For multistory slab construction, however, since the
cracking is predicted, then the effective moment of iner load imposed on the slab during construction often ex
tia may be used to estimate the influence of cracking on ceeds that due to the specified dead plus live load
the deflection, as discussed in Chapter 3, Section 3.6.2. (Grundy and Kabaila, 1963; etc.), the extent of cracking
Deflections of twoway prestressed systems can also is usually determined by the construction loads resulting
be computed by evaluating curvatures at sections based from shoring and reshoring procedures. Under these con
on compatibility and equilibrium as described in Section ditions all values of immediate deflection should be cal
4.3.1.2. The timedependent changes in strains in pre culated using the effective moment of inertia correspond
stressed sections are caused by relaxation of prestressed ing to the construction load level, as illustrated in Fig.
steel in addition to creep and shrinkage of concrete. The 4.3. This calculation procedure usually results in a
sections are subjected to normal force N and bending smaller live load deflection and larger dead load deflec
moment M, producing axial strain as well as curvature. tion, with correspondingly larger sustained load deflec
The position of the neutral axis after cracking is depen tion.
dent on the value (M/N) in addition to geometric proper A typical loadtime history is shown in Fig. 4.4 for a
ties. Analysis details and a computational example are slab in a multistory structure. During construction, the
given in Ghali,1990. load on the slab increases as new slabs are placed above.
When construction above is no longer supported by the
4.6Loads for deflection calculations slab under consideration, the load decreases to a value
ACI 318 stipulates that calculated deflections must corresponding to the slab selfweight plus an allowance
not exceed certain permissible values, expressed as frac for superimposed dead load and sustained portion of live
tions of span length. Components of deflections to be load (load level at tl in Fig. 4.4).
considered are immediate live load deflection and incre A simple procedure to determine slab loads during
mental deflection, including that due to live load, after construction was proposed by Grundy and Kabaila
installation of nonstructural elements. The live load com (1963). More refined analysis procedures reported subse
ponent of deflection is normally considered as that due quently [e.g., Liu et al (1985), AguinagaZapata and
DEFLECTION IN CONCRETE STRUCTURES 435R59
M
Mcmst
M
D+L
Fig. 4.3Definition of 6, 6, and 6,,,t when construction loads exceed specified dead plus live load
Load, W
j
Tf, V/L(var)
w sust
I I +.
Bazant (1986)] give results that are quite similar to the baila procedure
original Grundy and Kabaila procedure. The maximum Wdab = slab dead load
load during construction, including loads due to shoring WCL = construction live load
and reshoring plus an allowance for construction live N = number of shored and reshored levels
load, can be estimated using the following relationship:
Gardner (1985) recommends k, = k2 = 1.1. The con
struction live load may be taken as 50 psf (2.4 kPa) as
(4.19) recommended by ACI 347R. The factor k, accounts for
errors in computing R due to variations in stiffnesses
where between the stories in the supporting system. The factor
k, = allowance for error in theoretical load ra R has been shown to vary from 1.8 to 2.2, depending pri
tio R marily on the number of stories of shores and reshores in
k2 = allowance for weight of formwork the system. If the shoring system to be used is unknown,
R = applied load/slab dead load ratio a value of R = 2.0 can be used in the calculation. Instead
= load ratio calculated by Grundy and Ka of a factor k, for formwork weight, a value of 10 psf is
435R60 ACI COMMITTEE REPORT
Load
Time
eflection, g
S const
o 63
I
6 ^
'2
^
'3 Time
I
considered to be a reasonable allowance for most form tion load deflection, aconst.
work systems. 3. Calculate the live load deflection by scaling the
At time t, in Fig. 4.4, a slight increase in the sus construction load deflection.
tained load occurs as nonstructural elements are in
stalled. The variable portion of live load may be con
sidered as applied intermittently thereafter. One
application of live load is shown at time t3.
An analysis procedure based on this type of loading where E,(Const) and Ec(L) are modulus of elasticity val
history and ACI 209R creep and shrinkage functions was ues at application of construction load and live load,
developed by Graham and Scanlon (1986a), using the respectively.
principle of superposition. Effects of partial creep 4. Scale the construction load deflection to the sus
recovery were considered. Analyses were also made for tained load level. Sustained load includes dead load plus
the simplified loadtime history shown in Fig. 4.5 with the any portion of the live load assumed to be sustained.
corresponding displacementtime history. Longterm sus
tained load deflections were obtained using multipliers
61 = wmst 6
EJcom
calibrated with the more complex history of Fig. 4.4. wco~(max) l coMf l E&SW)
Resulting multipliers are included in Table 4.1.
Based on the procedures suggested by Sbarounis where E,(Sust) is the modulus of elasticity at the time
(1984) and Graham and Scanlon (1986a), the following sustained load is applied (i.e., at end of construction
approach based on the simplified loadtime history can period).
be used to estimate longterm deflections in multistory 5. Calculate sustained load deflection at time of
slab systems. installation of nonstructural elements.
1. Estimate the maximum construction load expected
[ %wd~crx)l based on usual procedures for multistory 62 = vz
construction.
2. Calculate the corresponding immediate construc where X, = multiplier corresponding to time interval tl
DEFLECTION IN CONCRETE STRUCTURES 435R61
to t2. (The time function given in Eq. 2.16 can be used to ity is evident, both during the construction period (first
determine 1, i.e., 35 days) and at approximately one year thereafter.
A histogram of oneyear deflections, shown in Fig.
4.7, indicates a coefficient of variation of 29.9 percent for
these slabs and a range of deflections from approximately
the mean minus 50 percent to the mean plus 70 percent.
Calculated deflections at one year based on three as
6. Calculate ultimate sustained load deflection. sumed values of modulus of rupture, and longterm mul
tipliers proposed by Graham and Scanlon [1986(b)], are
63 = A, s, shown in Fig. 4.6. These results indicate that the best
estimate of the mean deflection was obtained using an
where AI = longterm multiplier (Table 4.1). effective modulus of rupture of 4E psi (0.33 E MPa).
7. Calculate the deflection due to the variable portion
The calculated deflection based on the ACI 318 specified
of live load, i.e., that portion of live load not assumed as
sustained. value, 7.5E psi (.62g MPa) was found to lie at the
low end of the range of measured deflections. The cal
S&zr) = KSL (from step 3) culations incIuded effects of construction loads.
Sbarounis (1984) reported on deflection measure
where ments taken after one year on 175 bays of a multistory
building in Chicago. Measured deflections had a mean
value of 1.35 in. (34.3 mm) and a coefficient of variation
K variable live load w4w
= =  of 21.2 percent. The range in measured deflections was
total live load WL 0.53 in. to 2.16 in. (13.5 to 54.9 mm), i.e., from the mean
minus 60 percent to the mean plus 60 percent. Calcu
8. Calculate increment in deflection after installation lated values were close to the mean deflection.
of nonstructural elements. A number of case studies of large deflections re
ported in the literature has been summarized in ACI
6.mc S3
= o + S&w)  S
o2 435.8R. These case studies, including examples from
Australia, Scotland, Sweden, and the U.S., highlight the
9. Compare calculated deflections with appropriate large number of factors that can cause variability in
permissible values. insitu slab deflections.
LEGEND
50 l Measured Value
@I Mean Measured Value
. Calculated
_
/ \
\_/ Value
l
40 e l
.*
@< t=0.32&MP,
30 l
.0 04 fr =O.GJzMPc
20
l
7.5 f: psi
)
IO
Fig. 4.6Fieldmeasured deflections for 40 nominally identical slab panels in 28story building (Jokinen and Scanlon,
1985)
sign situations. Individual cases may require more strin 15 in. x 24 in. in the NS direction (381 mm x 610 mm).
gent requirements than the limited treatment given in The floor is subjected to a timedependent deflection due
ACI 318. Guidance on appropriate deflection limits for to an equivalent uniform working load intensity w = 450
a range of applications is given in ACI 435.3R (1984). psf (21.5 kPa). Material properties of the floor are:
FREQUENCY
20 
l 8
l6 
n = 40
l 4 x = 32.53 mm
* s = 9.72mm
l2  V = 29.9 %
I 0
8
6 ( 25.4 mm = 1 in. )
4 r +
2
l
0
5 l5 25 35 45 55 65
DEFLECTION (mm)
determine if its magnitude is acceptable if the floor EW direction deflections (span = 25 ft)
supports sensitive equipment which can be damaged by Longterm w, = 450 psf
large deflections. 450 x 20(25)4 x 123
3. Cracked moment of inertia: s2; = 0.069 in.
384 x 3.6 x lo6 x 63,600
EW: I,, = 45,500 in.4
NS: I,, = 32,500 in.4
sC = 0.069 x 0.81 63,600 = 0.066 in.
Solution: 53,700
Note: All calculations are rounded to three significant
figures. sS = 0.068 x 0.19 x !!?@!!f = 0.243 in.
3430
Calculate the gross moments of inertia (in.4) of the
sections in Fig. 8, namely, the total equivalent frame Its Rotation at end 1 is
in part (b), the column strip beam I, in part (c), and the
middle strip slab Is in part (d). These variables are: Ml 20 x ld x 12 = 1.67x 10m4 rad
8 zz
Ics Ic Is l &c 400 x 3.6 x lo6
ICR
EW 63,600 53,750 3430 45,500 and the rotation at end 2 is
NS 47,600 40,000 4290 32,500
,%.u
X
u3 
 4 
t
em 20 ft
(240 in.) *
support
support
Fig. A4.1Longterm deflection of twoway multipanel slab on beams in Ex. A4.1, equivalent frame calculation method
(Nawy, 1990 courtesy Prentice Hall)
So9) = 0.171 + 0.015 = 0.186 say 0.19 in. I e = 0.037 x 63,600 + (1  0.037)45,500 = 46,200 in.4
Ma unfactored load level due to dead plus live load are given
3.52 x lti l
in Table A4.1.
Cracking moment (A&,.)
Midpanel deflection
So = SC + Sm = 0.69 in.
16 x 16 Column (Typ)
I SL = (0.69) = 0.31 in.
= 4
= 251
360 360
Fig. A4.2Plan of flat plate edge panel in Ex. A4.2, beam = 0.70 in.
crossing calculation method > 0.31 in . . . OK for shortterm deflection
Incremental deflection
Mcr = 8
Use longterm multiplier = 2.5 applied to sustained
y* load deflection.
g = l/12 (12) (6)3 = 216 in.4
Assume sustained load = 75 + 20 = 95 psf
yt = 6/2 = 3 in.
~~~
I,, (in.4) 26.8 34.3 46.7   
Z, 56.0 50.6 50.1 216 (= Z,) 216(=Z) 216 (= Ig)
I, (average) 52.2 216
many cases determined by deflection response rather of deflection, and appropriate situations in which the
than by strength. This Chapter proposes design pro option should be considered. The options are arranged
cedures for reducing the expected deflection that will in three groups; Design techniques, Construction tech
enable design engineers to proportion building structures niques, and Materials selection.
to meet both strength and serviceability requirements.
The result could be more economical structures com 5.2Design techniques
pared to those designed with unnecessarily conservative 5.2.1 Increasing section depthIncreasing the depth
deflection response. The discussion assumes that a may not be possible after schematic design of the pos
competent design is prepared in accordance with Building sible after schematic design of the building has been
Code Requirements for Reinforced Concrete (ACI 318) and established because such dimensional changes may affect
construction follows good practices. the architectural and mechanical work. However, there
To properly evaluate options for reducing deflection, are many instances where beam depth can be increased.
a design engineer must know the level of stress in The reduction in deflection is approximately proportional
the member under consideration, that is, whether the to the square of the ratio of effective depth, d, for
member is uncracked, partially cracked or fully cracked. cracked sections and to the cube of the ratio of total
Heavily reinforced members tend to be fully cracked concrete depths for uncracked sections. This is based on
because of the heavy loads they are subjected to. In this the fact that the cracked moment of inertia, I,, is ex
Chapter, only two limiting conditions are considered, pressed as,
uncracked members and fully cracked members. If the I,, = nA,(lQd2 in reinforced concrete and
applied moment in the positive region is more than twice Icr = n#d2(11.6 \/npp,) in prestressed concrete.
the cracking moment, considering the effect of flanges, Hence,
the member may be considered as fully cracked. Fre 1,, = @d2 or Z,, = @dP2 and the gross moment of
quently, a member is only partially cracked (M, < M, < inertia Jg = bh3/12 for a rectangular section, namely lg =
2M,,) and the statements about both limiting conditions @)d3. For example, if an l8 in.deep, rectangular beam
are not strictly applicable. Engineering judgement and with an effective depth of 15.5 in. is increased to 20 in.
appropriate calculations should be made to assess the deep, and all other parameters are kept the same, the
actual serviceability conditions of the beam. Chapter 2 cracked stiffness will increase by 27 precent [(17.5/15.5)2
and 3 of this report outline methods for computing the = 1.27], and the uncracked stiffness will increase by 37
degree of cracking in a member. percent [(20/18)3 = 1.37]. For heavily reinforced mem
In addition to the stress conditions, there may be bers, if the amount of reinforcement is reduced when the
physical or nonstructural constraints on the use of some depth is increased, the cracked stiffness is increased only
options such as limits on increasing concrete dimensions. in proportion to the increase in depth or by 13 percent
All options must be evaluated in terms of cost since some for this example. This can be seen from substituting for
may increase the cost, and some may have offsetting con the reinforcement area its equivalent value Mfljd in the
siderations that reduce the cost, while still others may expression Icr= A#Qd2 giving I = f(d). The increase
have little effect on cost. For each option presented, in stiffness of an uncracked Tbeam when it is made
there is a discussion on the effect of implementation on deeper will be less than that for a rectangular beam
deflection, the approximate range of potential reduction because the flanges do not change. Flanges tend to have
435R68 ACI COMMITTEE REPORT
a fixed influence rather than a proportional influence on reinforcement. If the Tbeam has a thin slab subiect to
uncracked stiffness. higher than normal shrinkage because of its high surface
If, by increasing the depth, the concrete tensile stress tovolume ratio, then compression reinforcement will be
in a member is reduced sufficiently so that it changes more effective than for a rectangular beam. This will be
from a cracked, or partially cracked, to an uncracked true for ribbed slabs or joist systems as well.
member, the stiffness could increase dramatically. The 5.2.4 Addition of tension reinforcementFor un
uncracked stiffness can be as much as three times the cracked members, addition of tension reinforcement has
partially cracked stiffness (Grossman, 1981). hardly any effect on deflection. For fully cracked mem
5.2.2 Increasing section widthThis option is not bers, addition of tension reinforcement reduces both
applicable to slabs or other members with physical immediate and longterm deflection almost in proportion
constraints on their width. Where beams cannot be made to the increase in the steel reinforcement area. This can
deeper because of floor to floor height limitations, but be seen from the cracked moment of inertia, I,,, defined
can be made wider, the increase in stiffness is propor in Section 5.2.1. For all practical purposes IO = 0.9A,
tional to the increase in width if the member is un since the variation in the term (1k)j is usually small. For
cracked. If the member is cracked and remains cracked example, if the total deflection of a cracked member is
after increasing the width, the increase in stiffness is very 1.50 in. as in the previous example, increasing the tension
small. However, if a cracked member becomes uncracked reinforcement by 50 percent will reduce the deflection to
because the width is increased, its stiffness increases about 1.10 in. However, the increased reinforcement area
appreciably, possibly by as much as a factor of three should still be less than the maximum permitted by AC1
(Grossman, 1981). 318, namely a maximum of 0.75 times the balanced ratio
5.2.3 Addition of compression reinforcementUsing pb. This option is most useful for lightly reinforced solid
ACI 318 procedures, compression reinforcement has and ribbed slabs. The option of adding more tension re
some effect on immediate deflection as it can influence inforcement is not available or is limited for heavily
IC,; thus I, will be affected, as will the initial deflection, reinforced beams unless compression reinforcement is
however small the influence is. But it can reduce addi also added to balance the increase in tension bar area in
tional longterm or incremental deflection up to about 50 excess of 0.75 pb.
percent (ACI 318, 1989). The effect on total deflection 5.2.5 Prestressing applicationDead load deflection of
is somewhat less. The addition of compression rein reinforced concrete members may be reduced substantial
forcement reduces the additional longterm deflection in ly by the addition of prestressing. However, deflections in
the example to 0.50 in. or by 50 percent and the total prestressed concrete members due to live load and other
deflection to 1.00 in. or by 33 percent. transient loads are about the same as those in reinforced
Longterm deflection has two components, creep de concrete members of the same stiffness, EI. If prestress
flection and shrinkage warping. Compression reinforce ing keeps the member in an uncracked state, without
ment reduces deflection because concrete creep tends to which it would otherwise crack, the live load deflection
transfer the compression force to the compression rein would be considerably smaller. If, however, the pre
forcement which does not itself creep. The closer the stressed member size is reduced, as is usually the case in
reinforcement is to the compression face of the member, order to take advantage of prestressing, then the live
the more effective steel reinforcement is in reducing load deflection becomes larger.
longterm creep deflection. Thus, compression reinforce Consequently, the span/depth ratio in posttensioned
ment is more effective in deeper than in shallower beams twoway floors is normally limited to 48 in lower floor
or slabs if the concrete cover to the compression face of slabs with light live load and 52 in roof slabs (ACI 318
the member is of constant value. For some very shallow Section R18.2.3, 1989). If the member has a high ratio of
members, due to the requirements of minimum bar live to dead load, then the span/depth ratio must be
cover, compression reinforcement could be at or near the proportionally reduced in order to give satisfactory
neutral axis and be almost totally ineffective in reducing deflection performance. A prestressing force sufficient to
longterm creep deflection. produce satisfactory deflection response should always be
Shrinkage warping occurs where the centroids of the provided, regardless of whether the member is uncracked
steel reinforcement and the concrete do not coincide and at service load or it is designed as partially prestressed
the shrinkage of concrete, combined with the dimensional with tolerable flexural crack width levels which are con
stability of steel reinforcement, warps the member in a trolled by additional mild steel reinforcement.
fashion similar to a piece of bimetal subject to temper 5.2.6 Revision of structure geometryCommon solu
ature variations. Compression reinforcement reduces tions to reduce deflections include increasing the number
shrinkage warping because it brings the centroid of the of columns in order to reduce the length of the spans,
tension and compression reinforcement closer to the con adding cross members to create twoway systems, and in
crete neutral axis. While compression reinforcement re creasing the size of columns to provide more end re
duces shrinkage and warping of all flexural members, it straint to flexural members.
is especially effective for Tbeams where the neutral axis 5.2.7 Revision of deflection Limit criteriaIf deflection
is close to the compression face and far from the tension of a member is excessive, the deflection limits may be
DEFLECTION IN CONCRETE STRUCTURES 435R69
reexamined to determine if they are unnecessarily re slabs were not all built level or at the specified grade or
strictive. If experience or analysis indicates that those the method and timing of form stripping was not uni
limits (see Chapters 2 and 3) can be relaxed, then other formly applied. Also, construction loads may not have
action might not be required. Many building codes do been applied uniformly.
not set absolute limits on deflection. An engineer might 5.3.4 Delay of the first loadingThis allows the con
determine that the building occupancy, or construction crete to gain more strength before loading or helps to
conditions, such as a sloping roof, do not require the reach its design strength. Both the modulus of elasticity
normal deflection limits. EC and the modulus of rupture& will be increased. An
increase in E, increases the flexural stiffness. An increase
5.3Construction techniques in the modulus of rupture value, f, reduces the amount
5.3.1 Concrete curing to allow gain in strength of cracking or even allows the member to remain un
Deflection response is determined by concrete strength cracked with an increase in flexural stiffness EI as noted
at first loading, not by final concrete strength. If the in the next section.
construction schedule makes early loading of the con 5.3.5 Delay in installation of deflectionsensitive
crete likely or desirable, then measures to ensure high elements or equipmentSuch delay in equipment installa
strength at first loading or construction loading can be tion will have no effect on immediate or total deflection,
effective. For example, if at the design compressive except as previously noted in 5.3.1. But incremental de
strength fc of 4000 psi, the member would beuncracked flection will be reduced, namely the deflection occurring
as designed, but it is loaded when concrete strength is from the time a deflectionsensitive component is in
2500 psi, it could be excessively cracked due to a lower stalled until it is removed or the deflection reaches its
modulus of rupture at the time of loading. Even though final value. For example, if the additional longterm
its final loadcarrying capacity was satisfactory, the deflection is 1.00 in., and installation of partitions is
cracked member could still deflect several times more delayed for 3 months, the incremental deflection will be
than a similar uncracked one. Furthermore, the modulus approximately 0.50 in. or about onehalf as much as the
of elasticity of a 4000 psi concrete is higher than that of total deflection.
2500 psi concrete (see Section 5.4 of this report for the 5.3.6 Location of deflectionsensitive equipment to
effects of material selection on these parameters). avoid deflection problemsEquipment such as printing
5.3.2 Concrete curing to reduce shrinkage and creep presses, scientific equipment and the like must remain
Immediate deflection will not be greatly affected by level and should be located at midspan where the
concrete curing but additional longterm deflection will change in slope is very small with the increase in
be reduced. Assuming the longterm component of de deflection. On the other hand, because the amplitude of
flection is evenly divided between shrinkage and creep, if vibration is highest at midspan, vibrationsensitive
shrinkage is reduced 20 percent by good curing, the addi equipment may be best located near the supports.
tional longterm deflection due to shrinkage will be re 5.3.7 Provision of architectural details to accommodate
duced by 10 percent. The effect will be most pronounced expected deflectionPartitions that abut columns, as an
on members subject to high shrinkage such as those with example, may show the effect of deflection by separating
a high surface/volume ratio (smaller members), those horizontally from the column near the top even though
with thin flanges, and structures in arid atmospheres or the partition is not cracked or otherwise damaged. Archi
members which are restrained. The effect of good curing tectural details should accommodate such movements.
on creep is similar to its effect on shrinkage. Likewise, windows, walls, partitions, and other non
5.3.3 Control of shoring and reshoring procedures structural elements supported by or located under de
Many studies indicate that the shoring load on floors of flecting concrete members should be provided with slip
multistory buildings can be up to twice the dead weight joints in order to accommodate the expected deflections
of the concrete slab itself. Because the design super or differential deflections between concrete members
imposed load is frequently less than the concrete self above and below the nonstructural elements.
weight, the slab may be seriously overstressed and 5.3.8 Building camber into floor slabsBuilt in camber
cracked due to shoring loads instead of remaining un has no effect on the computed deflection of a slab. How
cracked as assumed by calculations based on design ever, cambering is effective for installation of partitions
loads. Thus, the flexural stiffness could be reduced to as and equipment, if the objective is to have a level floor
little as one third of the value calculated assuming design slab after deflection takes place. For best results, de
loads only. Furthermore, the shoring loads may be im flection must be carefully calculated using the appro
posed on the floor slabs before the concrete has reached priate modulus of concrete E, value and the correct
its design strength (see discussion in Section 5.3.1). moment of inertia I. Overestimating the deflection value
Construction of formwork and shoring should ensure can lead the designer to specify unreasonable overcam
that a sag or negative camber is not built into the slab. bering. Hence, the pattern and value of cambering at
Experience indicates that frequently the apparent deflec several locations has to be specified and the results
tion varies widely between slabs of identical design and monitored during construction. Procedures have to be
construction. Some reasons for this may be that such revised as necessary for slabs which are to be constructed
435R.70 ACI COMMITTEE REPORT
Reinhold Co., New York, Editor, M. Fintel, 1985, pp. 53 Branson, D.E., and Cristianson, M.L., TimeDepen
78. dent Concrete Properties Related to DesignStrength
Branson, D.E., and Trost, H., Unified Procedures for and Elastic Properties, Creep and Shrinkage, ACI
Predicting the Deflection and Centroidal Axis Location Special Publication, SP27, 1971, pp. 257277.
of Partially Cracked NonPrestressed Members, ACI Branson, D.E., and Kripanarayanan, KM., Loss of
JOURNAL, Proceedings, V. 79, No. 2, Mar.Apr. 1982, pp. Prestress, Camber and Deflection of NonComposite and
119130. Composite Prestressed Concrete Structures, PCI Journal,
Branson, D.E., Deformation of Concrete Structures, V. 16, Sept.Oct. 1971, pp. 2252.
McGraw Hill Book Co., Advanced Book Program, New Branson, D.E., Compression Steel Effect on Long
York, 1977, 546 pp. Time Deflections, ACI JOURNAL, Proceedings, V. 68,
435R72 ACI COMMITTEE REPORT
No. 8, Aug. 1971, pp. 555559. Volume Changes of High Strength Concretes with Super
Branson, D.E., Instantaneous and TimeDependent plasticizer, Journal, Japan Prestressed Concrete
Deflections of Simple and Continous Reinforced Con Engineering Association, V. 20, 1978, pp. 2633.
crete Beams, HPR Publication No.7, Part 1, Alabama Nawy, E.G., Reinforced Concrete  A Fundamental
Highway Department, U.S. Bureau of Public Roads, Aug. Approach, Second Edition, Prentice Hall, 1990, 738 pp.
1963, pp. l78. Nawy, E.G., Structural Elements: Strength, Service
Carrasquillo, R.L., Nilson, A.H., and Slate, F.O., ability and Ductility, Handbook of Structural Concrete,
Properties of High Strengh Concrete Subjected to Short McGraw Hill, New York, 1983, pp. 12.112.86.
Term Loads, ACI JOURNAL, Proceedings, V. 78, No. 3, Nawy, E.G., and Balaguru, P.N., High Strength Con
MayJune 1981, pp. 171178. crete, Handbook of Structural Concrete, McGraw Hill,
CEB Commission IV, Deformations, Portland New York, 1983, pp. 5.15.33.
Cement Association, Foreign Literature Study 547, 1968. Nawy, E.G., and Neuworth, G., Fiber Glass Rein
ComiteInternational du Beton (CEB)  Federation forced Concrete Slabs and Beams, Structural Division
Internationale de la Pricontrainte (FIP), Model Code for Journal, ASCE, New York, pp. 421440.
Concrete Structures, 1990, P.O. Box 88, CH1015, Ngab, AS., Nilson, A.H., and Slate, F.O., Shrinkage
Lausanne. and Creep of High Strength Concrete, ACI JOURNAL,
Ghali, A., Deflection Prediction in TwoWay Floors, Proceedings, V. 78, No. 4, JulyAugust 1981, pp. 255261.
ACI Structural Journal, V. 86, No. 5, Sept. Oct. 1989, pp. Nilson, A.H., Design Implications of Current
551562. Research on High Strength Concrete, American Con
Ghali, A., and Favre, R., Concrete Structures: Stresses crete Institute, SP87, 1985, pp. 85117.
and Deformations, Chapman and Hall, London and New Nilson, A.H., Hover, K.C., and Paulson, K.A.,
York, 1986, 352 pp. Immediate and LongTerm Deflection of High Strength
Grossman, J.S., Simplified Computations for Effec Concrete Beams, Report 893, Cornell University,
tive Moment of Inertia (Ie) and Minimum Thickness to Department of Structural Engineering, May 1989,230 pp.
Avoid Deflection Computations, ACI JOURNAL, Pro Park, R., and Paulay, T., Reinforced Concrete
ceedings, V. 78, No. 6, Nov.Dec., 1981, pp. 423439. Structures, John Wiley & Sons, Inc., New York, 1975,769
Also, Author Closure, ACI JOURNAL, Proceedings, V. 79, PP.
No. 5, Sept.Oct., 1982, pp. 414419. Pfeifer, D.W.; Magura, D.D.; Russell, H.G.; and
Grossman, J.S., Reinforced Concrete Design, Corley, W.G., Time Dependent Deformations in a 70
Building Structural Design Handbook, R.N. White and Story Structure, SP37, American Concrete Institute,
C.G. Salmon, editors, John Wiley & Sons, New York, 1971, pp. 159185.
1987, Ch. 22, pp 699786. Portland Cement Association, Notes on ACI 31883
Hsu, C.T., A Simple Nonlinear Analysis of Con Building Code Requrements for Reinforced Concrete
tinuous Reinforced Concrete Beams, Journal of Engi with Design Applications, Skokie, Illinois, 1984.
neering and Applied Science, V. 2, No. 4, 1983, pp. 267 Russell, H.G., and Corley, W.G., TimeDependent
276. Behavior of Columns in Water Tower Place, SP55,
Hsu, C.T., and Mirza, M.S., A Study of Post American Concrete Institue, 1978, pp. 347373.
Yielding Deflection in Simply Supported Reinforced Saucier, K.L., Tynes, W.O., and Smith, E.F., High
Concrete Beams, ACI SP43, Deflection of Concrete CompressiveStrength ConcreteReport 3, Summary
Structures, American Concrete Institute, 1974, pp. 333 Report, Miscellaneous Paper No. 6520, U.S. Army
355, Closure, ACI JOURNAL, April 1975, pp. 179. Engineer Waterways Experiment Station, Vicksburg,
Luebkaman, C.H., Nilson, A.H., and Slate, F.O., September 1965, 87 pp.
Sustained Load Deflection of High Strength Concrete
Swamy, R.N., and Anand, K.L., Shrinkage and Creep
Beams, Research Report No. 852, Dept. of Structural
of High Strength Concrete, Civil Engineering and Public
Engineering, Cornell University, February 1985.
Works Review, V. 68, No. 807, Oct. 1973, pp. 859865,
Martin, I., Effects of Environmental Conditions on
867868.
Thermal Variations and Shrinkage of Concrete Structures
in the United States, SP27, 1971, pp. 279300. Teychenne, D.C., Parrot, L.J., and Pomeroy, C.D.,
Martinez, S., Nilson, A.H., and Slate, F.O., Spirally The Estimation of the Elastic Modulus of Concrete for
Reinforced High Strength Concrete Columns, Research the Design of Structures, Current paper No. CP23/78,
Report No. 8210, Department of Structural Engineering, Building Research Establishment, Garston, Watford,
Cornell University, Ithaca, August 1982. 1978, 11 pp.
Meyers, B.L., and Thomas, E.W., Chapter 11, Elasti Yu, W.W., and Winter, G., Instantaneous and Long
city, Shrinkage, Creep, and Thermal Movement of Con Term Deflections of Reinforced Concrete Beams Under
crete, Handbook of Structural Concrete, MaGraw Hill, Working Loads, ACI JOURNAL, Proceedings, V. 57, No.
New York, Editors, Kong, Evans, Cohen, and Roll, 1983, 1, 1960, pp. 2950.
pp. 11.111.33. Wang, C.K., and Salmon, C.G., Reinforced Concrete
Nagataki, S., and Yonekura, A., Studies of the Design 5th Edition, Harper Collins, 1992, pp. 1030.
II
DEFLECTION IN CONCRETE STRUCTURES 435R73
Unified Classical and Matrix Approach, 3rd Edition, Publishing Company, Amsterdam, 1970, 622 pp.
Chapman & Hall, New York, 1989, pp. 608409. Prestressed Concrete Institute, PCI Design Handbook,
Herbert, T.J., Computer Analysis of Deflections and Chicago, Illinois, Fourth Edition, 1993.
Stresses in Stage Constructed Concrete Bridges, PCI PCI Committee on Prestress Losses, Recommenda
Journal, V. 35, No. 3, MayJune 1990, pp. 5263. tions for Estimating Prestress Losses, Journal of the
Libby, J.R., Modem Prestressed Concrete, 3rd Edition, Prestressed Concrete Institute, V. 20, No. 4, JulyAugust
Van Nostrand Reinhold Company, New York, 1984, 635 1975.
p. Rangan, B.V., Serviceability Design in Current
Martin, D.L., A Rational Method for Estimating Australian Code, SP133, American Concrete Institute,
Camber and Deflection of Precast Prestressed Members, Detroit 1993, pp. 93110.
PCI Journal, V. 22, No. 1, JanuaryFebruary, 1977, pp. Shaikh, A.F. and Branson, D.E., NonTensioned
100108. Steel in Prestressed Concrete Beams, PCI Journal, V. 15,
Mirza, M.S., and Sabnis, G.M., Deflections of One No. 1, Feb. 1970, pp. 1436.
Way Slabs and Beams, Proceedings of Symposium, ACI Sirosh, S.N. and Ghali, N.,Reinforced Concrete
Canadian Capital Chapter, Montreal, Canada, Oct. 1971, BeamColumns and Beams on Elastic Foundation,
pp. 5387. Journal of Structural Engineering, ASCE, V. 115, No.3,
Fling, R.S., Practical Considerations in Computing March 1989, pp. 666682.
Deflection of Reinforced Concrete, SP133, American Tadros, M.K., Expedient Service Load Analysis of
Concrete Institute, Detroit, 1993, pp. 6992. Cracked Prestressed Concrete Sections, PCI Journal, V.
Naaman, A.E., Prestressed Concrete Analysis and 27, No. 6, Nov.Dec. 1982, pp. 86101.
DesignFundamentals, First edition, McGraw Hill Book Tadros, M.K., Ghali, A., and Dilger, W.H., Effect of
Co., New York, N.Y., 1982, 670 pp. NonPrestressed Steel on Prestressed Loss and Deflec
Naaman, A.E., Time Dependent Deflection of Pre tion, PCI Journal, Vol. 22, No. 2, March 1977, pp. 50.
stressed Beams by Pressure Line Method, PCI Journal, Tadros, M.K., Expedient Service Load Analysis of
V. 28, No. 2, April 1983, pp. 98119. Cracked Prestressed Concrete Sections, PCI Journal, V.
Naaman, A.E., Partially Prestressed Concrete: 27, Nov.Dec. 1982, pp. 86111. Also, see Discussion by
Review and Recommendations, Journal of the Prestressed Nilson, Branson, Shaikh, et al., PCI Journal, V. 28, No.
Concrete Institute 30 (1985), pp. 3071. 6, Nov.Dec. 1983.
Nawy, E.G., and Potyondy, J.G., Deflection Behavior Tadros, M.K., and Sulieman, H., Discussion of
of Spirally Confined Pretensioned Prestressed Concrete Unified Procedures for Predicting Deflections, by
Flanged Beams, PCI Journal, V. 16, June 1971, pp. 44  Branson and Trost, PCI Journal, V. 28, No. 6, Nov.Dec.,
59. 1983, pp. 131136.
Nawy, E.G., and Potyondy, J.G., Flexural Cracking Tadros, M.K., Ghali, A., and Meyer, A.W., Prestress
Behavior of Pretensioned Prestressed Concrete I and T Loss and Deflection of Precast Concrete Members, PCI
Beams, ACI JOURNAL , Proceedings, V. 68,1971, pp. 335 Journal, V. 30, No. 1, JanuaryFebruary 1985.
360. Tadros, M.K., and Ghah, A., Deflection of Cracked
Nawy, E.G. and Potyondy, J.G., authors closure and Prestressed Concrete Members SP86, American
discussion by D. E. Branson of Flexural Cracking Be Concrete Institute, Detroit, pp. 137166.
havior of Pretensioned Prestressed Concrete I and T Trost, H., The Calculation of Deflections of Rein
Beams, ACI JOURNAL, Proceedings, V. 68, No. 5, May forced Concrete MembersA Rational Approach, SP
1971, discussion and closure, ACI JOURNAL , Proceedings 76, American Concrete Institute, Detroit, Michigan, 1982,
V. 68, No. 11, Nov. 1971, pp. 873877. pp. 89108.
Nawy, E.G., and Huang, P.T., Crack and Deflection
Control of Pretensioned Prestressed Beams, Journal of Chapter 4
the Prestressed Concrete Institute, V. 22 (1971), pp. 3047. Prediction of Creep, Shrinkage, and Temperature
Nawy, E.G., and Chiang, J.Y., Serviceability Be Effects in Concrete Structures,ACI 209R82 (Revised
havior of PostTensioned Beams, Journal of the Pre 1986), American Concrete Institute, Detroit, 1982.
stressed Concrete Institute, V. 25, 1980, pp. 7495. Control of Cracking in Concrete Structures, ACI
Nawy, E.G., Prestressed ConcreteA Fundamental 224R80 (Revised 1984), American Concrete Institute,
Approach, Prentice Hall, Englewood Cliffs, N.J., 1989, Detroit, 1980.
738 pp. Building Code Requirements for Reinforced
Nawy, E.G. and Balaguru, P.N.,High Strength Con Concrete (AC1 31889)(Revised 1992) and Commen
crete, Handbook of Structural Concrete, McGraw Hill, taryACI 318R89 (Revised 1992), American Concrete
New York, 1983, pp. 5.15.33. Institute, Detroit, 1992, 345 pp.
Nilson, A.H., Design of Prestressed Concrete, New Recommended Practice for Concrete Formwork,
York: Wiley, 1987, pp 592. ACI 34778, American Concrete Institute, Detroit, 1978.
Neville, A. M., and Dilger, W., Creep of Concrete: Allowable Deflections, ACI 435.3R68 (Reapproved
Plain, Reinforced, and Prestressed, NorthHolland 1989), American Concrete Institute, Detroit, 1968,12 pp.
DEFLECTION IN CONCRETE STRUCTURES 435R75
Variability of Deflections of Simply Supported Graham, C.J., and Scanlon, A., Long Time Multi
Reinforced Concrete Beams, AC1 435.4R72 (Reap pliers for Estimating TwoWay Slab Deflections, ACI
proved 1989), American Concrete Institute, Detroit, J OURNAL, Proceedings, V. 83, No. 5, 1986, pp. 899908.
1972, 7 pp. Grossman, J.S., Simplified Computations for Effec
Deflection of TwoWay Reinforced Concrete Floor tive Moment of Inertia le and Minimum Thickness to
Systems: StateoftheArt Report, ACI 435.6R74 Avoid Deflection Computations, ACI JOURNAL, Pro
(Reapproved 1989), American Concrete Institute, ceedings, V. 78, No. 6, Nov.Dec. 1981, pp. 423439.
Detroit, 1974, 24 pp. Grossman, J.S., Chapter 22, Building Structural Design
Observed Deflection of Reinforced Concrete Slab Handbook, R.N. White and C.G. Salmon editors, John
Systems, and Causes of Large Deflections, ACI 435, 8R Wiley and Sons, 1987.
85 (Reapproved 1991), American Concrete Institute, Grundy, P., and Kabila, A., Construction Loads on
Detroit, 1985. Slabs with Shored Formwork in Multistory Buildings,
Flexural Strength of Concrete Using Simple Beam with ACI JOURNAL , Proceedings, V. 60, No. 12, December
Center Point Loading, ASTM C 29379, American 1963, pp. 17291738.
Society for Testing and Materials. Jensen, V.P., Solutions for Certain Rectangular Slabs
AguinagaZapata, M., and Bazant, Z.P., Creep Continuous Over Flexible Supports, University of Illinois
Deflections in Slab Buildings and Forces in Shores Engineering Experimental Station Bulletin No. 303, 1938.
During Construction, ACI JOURNAL , Proceedings, V. 83, Jokinen, E.P., and Scanlon, A., Field Measured
No. 5, 1986, pp. 719726. TwoWay Slab Deflections, Proceedings, 1985 Annual
ASCE Task Committee on Finite Element Analysis Conference, CSCE, Saskatoon, Canada, May 1985.
of Reinforced Concrete Structures, Finite Element Kripanarayanan, K.M., and Branson, D.E., Short
Analysis of Reinforced Concrete, American Society of Time Deflections of Flat Plates, Flat Slabs, and TwoWay
Civil Engineers, New York, 1982, 553 pp. Slabs, ACI JO U R N AL, Proceedings, V. 73, No. 12,
Branson, D.E.,Instantaneous and TimeDependent December 1976, pp. 686690.
Deflections of Simple and Continuous Reinforced Lin, T. Y., LoadBalancing Method for Design and
Concrete Beams, HPR Publication 7, Part 1, Alabama Analysis of Prestressed Concrete Structures, ACI
Highway Department, Bureau of Public of Roads, Aug., JOURNAL , Procedings, V. 60, 1963, pp. 719742.
1963, pp. l78. Liu, X., Chen, W.F., and Bowman, M.D., Construc
Branson, D.E.,Deformation of Concrete Structures, tion Load Analysis for Concrete Structures, Journal of
McGrawHill Book Company, New York, 1977, 546 pp. Structural Engineering, V. 111, No. 5,1985, pp. 10191036.
Cook, R.D., Concepts and Applications of Finite Meyers, B.L. and Thomas, E.W., Chapter 11, Elasti
Element Analysis, John Wiley and Sons, 1974. city, Shrinkage, Creep and Thermal Movement of Con
Gallagher, R. H., Finite Element Analysis: Funda crete, Handbook of Structural Concrete, McGraw Hill,
mentals, PrenticeHall, Inc., New Jersey, 1975. New York, 1983, pp. 11.111.33.
Gardner, N.J., Shoring, Reshoring and Safety, Nawy, E.G., Reinforced ConcreteA Fundamental
Concrete International, V. 7, No. 4, 1985, pp. 2834. Approach, 2nd edition, PrenticeHall, 1990, 738 pp.
Ghali, A., Deflection Prediction in TwoWay Floors, Nawy, E.G., Prestressed ConcreteA Fundamental
ACI Structural Journal, V. 86, NO. 5, 1989, pp. 551562. Approach, Prentice Hall, 1989, 739 pp.
Ghali, A., Deflection of Prestressed Concrete Two Nawy, E.G., and Neuewrth, G., Fiber Glass Rein
Way Floor Systems, ACI Structural Journal, V. 87, No. 1, forced Concrete Slabs and Beams, Structural Division,
1989, pp. 6065. ASCE Journal, 1977, pp. 421440.
Ghali, A., and Favre, R., Concrete Structures: Nilson, A.H., and Walters, D.B. Jr., Deflection of
Stresses and Deformations, Chapman and Hall, London TwoWay Floor Systems by the Equivalent Frame
and New York, 1986, 350 pp. Method, ACI JOURNAL, Proceedings, V. 72, No. 5, May
Ghosh, S.K,Deflections of TwoWay Reinforced 1975, pp. 210218.
Concrete Slab Systems, Proceedings, International Design of PostTensioned Slabs, PostTensioning
Conference on Forming Economical Concrete Buildings, Institute, Glenview Illinois, 1976, 52 pp.
Portland Cement Association, Skokie, 1982, pp. Ramsay, R.J., Mina. S.A., and Macgregor, J.G.,
29.129.21. Monte Carlo Study of Short Time Deflections of Rein
Gilbert, R.I.,Deflection Control of Slabs Using forced Concrete Beams, ACI JOURNAL, Proceedings, V.
Allowable Span to Depth Ratios, ACI JOURNAL , 76, No. 8, August 1979, pp. 897918.
Proceedings, V. 82, No. 1, Jan.Feb. 1985, pp. 6777. Rangan, B.V., Prediction of LongTerm Deflections
Gilbert, R.I.,Determination of Slab Thickness in of Flat Plates and Slabs, ACI JOURNAL, Proceedings, V.
Suspended PostTensioned Floor Systems, V. 86, No. 5, 73, No. 4, April 1976, pp. 223226.
Sep.Oct. 1989, pp. 602607. Rangan, B.V., Control of Beam Deflections by
Graham, C.J., and Scanlon, A., Deflection of Rein Allowable Span to Depth Ratios, ACI JOURNAL, Pro
forced Concrete Slabs under Construction Loading, ceedings, V. 79, No. 5, Sept.Oct. 1982, pp. 372377.
SP86, American Concrete Institute, Detroit, 1986. Sbarounis, J.A., Multistory Flat Plate Buildings 
435R76 ACI COMMITTEE REPORT
Construction Loads and Immediate Deflections, Con Vanderbilt, M.D., Sozen, M.A. and Siess, C.P., De
crete International, V. 6., No. 2, February 1984, pp. 7077. flections of MultiplePanel Reinforced Concrete Floor
Sbarounis, J.A., Multistory Flat Plate Buildings: Slabs, Proceedings, ASCE, V. 91, No. ST4 Part 1, August
Measured and Computed OneYear Deflections, Con 1965, pp. 77101.
crete International, V. 6, No. 8, August 1984, pp. 3135. Zienkiewicz, O.C., The Finite Element Method,
Scanlon, A., and Murray D. W., Practical Calculation McGrawHill Book Co., 1977.
of TwoWay Slab Deflections, Concrete International,
November 1982, pp. 4350. Chapter 5
Simmonds, S.H., Deflection Considerations in the Building Code Requirements for Reinforced
Design of Slabs, ASCE National Meeting on Transporta Concrete (ACI 31889)(Revised 1992) and Commen
tion Engineering, Boston, Mass., July 1970, 9 pp. taryACI318R89 (Revised 1992), American Concrete
Tam, K.S.S., and Scanlon, A., Deflection of Two Institute, Detroit, 1992, 345 pp.
Way Slabs Subjected to Restrained Volume Change and Alsayed, S.H., Flexural Deflection of Reinforced
Transverse Loads, ACI JOURNAL, Proceedings, V. 83, Fibrous Concrete Beams, ACI Structural Journal, V. 90,
No. 5, 1986, pp. 737744. No. 1, Jan.Feb. 1993, pp. 7276.
Timoshenko, S. and WoinowskyKrieger, S., Theory of Grossman, Jacob S., Simplified Computation for
Plates and Shells, McGrawHill Book Co., New York, Effective Moment of Inertia, Ie and Minimum Thickness
1959. to Avoid Deflection Computations, ACI JOURNAL , Pro
Thompson, D.P., and Scanlon, A., Minimum Thick ceedings, V. 78, Nov.Dec. 1981, pp. 423439.
ness Requirements for Control of TwoWay Slab Deflec
tions, ACI Structural Journal, V. 85, No. 1, Jan.Feb. * This report was submitted to letter ballot of the committee and approved in
1986, pp. 1222. accordance with Institute balloting procedures.
DEFLECTION IN CONCRETE STRUCTURES 435R77
APPENDIX BDETAILS OF THE SECTION considered in the analysis of displacements (see Section B7).
CURVATURE METHOD FOR CALCULATING A linear stressstrain relationship is also assumed for the
DEFLECTIONS* reinforcement. The method in this appendix can be applied
This appendix presents a general method for calculating not only for steel reinforcement but also for other reinforce
displacements (translations and rotations) in prestressed and ment materials, such as fiberreinforced polymers (Hall and
nonprestressed reinforced concrete plane frames and beams. Ghali 1997; ISIS Canada 2001)
The method is based on analysis of strain distribution at indi
Cracking changes the distribution of internal forces in stati
vidual sections to determine axial strain and curvature under
cally indeterminate structures. For example, sections that
prescribed loading conditions.
crack over the supports of continuous beams result in reduced
negative moments at these sections and increased positive
B1Introduction moments at midspans. This redistribution of bending
The body of this report discusses control of deflection by
moments can be important in deflection calculations.
different means, including the choice of thickness of
The procedure presented in this appendix is intended to
members, the selection of the material, and the construction
compute deformations in service; it does not track the
techniques. It also discusses simplified calculation methods.
behavior as the load approaches ultimate. Approximations
The method in this appendix represents a more comprehensive
are involved for nonlinear effects when the concrete stress
method for the calculation of deformations in plane frames.
exceeds onehalf its compressive strength and when
Starting with the calculation of axial strain and curvature at
unbonded posttensioning is used. Strain compatibility of
individual sections, the variation of these parameters over
unbonded tendons and the adjacent concrete is not satisfied
the length of members is determined and used to calculate
(Ariyawardena and Ghali 2002).
the translations or the rotations at any section. Thus, the
B1.1 Notation
comprehensive analysis requires more calculations than the
simplified methods; it also requires more given (or assumed) A = crosssectional area
data. It accounts for the timedependent effects of creep and B = first moment of crosssectional area
shrinkage of concrete and relaxation of prestressing steel, Ct = ratio of creep strain to the initial strain (occurring
using timedependent material parameters that need to be immediately after stress application)
included in the given data. The comprehensive analysis is dA = elemental area
recommended when the deflection is critical and accuracy is E = modulus of elasticity
necessary. In this case, the analysis gives at any section of a fct = tensile strength of concrete
plane frame strain distribution that can be used to calculate I = second moment of crosssectional area
displacement components, composed of translations in two Ie = effective second moment of crosssectional area
orthogonal directions and a rotation. Thus, the analysis can l = member length
give vertical deflections as well as horizontal drifts. M = bending moment
The general method in this appendix determines the N = normal force
displacements (translations and rotations) in prestressed and n = ratio of modulus of elasticity of reinforcement to
nonprestressed reinforced concrete plane frames. The modulus of elasticity of concrete
general method is based on an analysis of strain distribution P = absolute value of prestressing force
at a section considering the effects of a normal force and a t = time
moment caused by applied loads, prestressing, creep and
y = distance measured downward from a horizontal xaxis
shrinkage of concrete, and relaxation of prestressing steel.
= coefficient to account for the effects of bond quality
The calculated axial strain and the curvature at various
of reinforcement, cyclic loading, and sustained
sections of the frame can then be used to calculate displace
loading on tension stiffening
ment by virtual work or other classical techniques. The
sectional analysis can accommodate the effects of creep and = d/dy = slope of stress diagram
shrinkage of concrete and relaxation of the prestressing steel. = deflection or camber
A sensitivity analysis of the uncertainties can be performed = strain
to determine the effect of varying the timedependent param O = strain at reference point O
eters on the calculated displacements. SH = unrestrained shrinkage of concrete
The sectional analysis is intended for service conditions. A = dimensionless coefficient used to calculate curva
linear stressstrain relationship can be assumed for the tures due to creep and shrinkage in reinforced non
concrete under service conditions, provided that the concrete prestressed sections (See Section B5.3)
stress does not exceed about half the compressive strength. = stress
At a crack location, the concrete section in tension is O = stress at reference point O
ignored. At an uncracked location, reinforcement bonded to = aging coefficient
the concrete causes increased stiffness, which should be r = reduction multiplier of the intrinsic relaxation
= curvature (= d/dy = slope of strain diagram)
*
p = intrinsic relaxation
435R Appendix B became effective January 10, 2003. For a list of Committee
membership at the time of the creation of this section, please see the end of the Appendix. pr= reduced relaxation
435R78 ACI COMMITTEE REPORT
Subscripts the initial deflection to give the deflection due to creep and
1 and 2 = uncracked state and cracked state, with concrete shrinkage combined.
in tension ignored, respectively The analysis method can determine the upper and the lower
c = concrete bounds of probable deflections by varying the input parame
cr = cracking ters and repetition of the calculation. In statically indetermi
g = gross concrete cross section nate structures, the timedependent effects can develop
m = mean value interpolated between States 1 and 2 changes in the stress resultants; computation of these changes
ns = nonprestressing steel is discussed in Ghali, Favre, and Elbadry (2002).
ps = prestressing steel
B3Crosssectional analysis outline
B2Background Crosssectional analysis is applicable for initial and
Calculation of deflections of prestressed or nonprestressed timedependent strains and stresses in prestressed and
reinforced concrete members is complicated by several nonprestressed reinforced concrete structures. Cracking
factors, including shrinkage and creep of concrete, relaxation may occur depending on the amount of prestressing
of prestressing reinforcement, and cracking. The analysis provided. The analysis assumes that structures are
presented in this appendix can be used to determine the initial composed of members, for which plane cross sections
and the timedependent stresses, strains, and displacement at remain plane after deformation. The strain distribution over
service loads for prestressed or nonprestressed reinforced the cross section, or a part of the cross section in the case of
concrete members for which the internal forces are known. composite cross section, can be defined by two parameters:
The procedure accounts for cracking; equilibrium and the normal strain O at an arbitrary reference point O and the
compatibility requirements of basic mechanics are satisfied. curvature , that is, the gradient of the strain over the depth
The accuracy of deformation calculations for reinforced of the section. The parameters O and can be used to calculate
concrete structures depends upon the thoroughness of the displacement (translations and rotations) by established struc
analysis method and the accuracy of the parameters used as tural analysis methods, such as virtual work or moment area.
given data. These parameters include the moduli of elasticity
of concrete, nonprestressing and prestressing reinforce B4Material properties
ments, the creep coefficient of concrete, the unrestrained This section defines the material parameters required as
shrinkage of concrete, relaxation of the prestressing rein input in the calculation of displacement. The parameters are:
forcement, and, when cracking occurs, the tensile strength of Ec = the modulus of elasticity of concrete;
the concrete. It is impossible to eliminate the error caused by Ct = the creep coefficient;
the uncertainty of the input parameters because they depend = the aging coefficient;
upon variables, such as the properties of the concrete, the SH = the free shrinkage of concrete; and
ambient temperature, and the relative humidity. An analysis p = the intrinsic relaxation of the prestressing steel.
satisfying the two basic requirements of mechanics, equilib Guidance on the values of Ec, Ct, SH, , and p to be
rium and compatibility, can reduce the error in the calculations. used in design are given in ACI 209R, CEBFIP (1990), and
This appendix presents such a method for structural concrete Magura, Sozen, and Siess (1964). Refer to Section B10 for
frames. A study of the sensitivity of the calculated displace example values of these parameters.
ments to the values of the input parameters is presented in
B4.1 CreepThe lines AB and BC in Fig. B4.1(a) repre
Ghali, Favre, and Elbadry (2002); a website for this book
sent the variation of a stress increment c, introduced on
provides three computer programs to assist in the calculations.
concrete at time t0, and sustained to time t. The corre
For statically determinant structures, the equilibrium sponding variation of strain is represented by curve EFG in
requires that the stress resultants at any cross section are not Fig. B4.1(b). The total corresponding strain, instantaneous
to be changed by the timedependent effects of creep, plus creep, at time t can be expressed by:
shrinkage, and relaxation, and that the stress resultants be
equal to the known normal force N and the bending moment
M. Compatibility requires that the strains in bonded rein c
c ( t ) =  [ 1 + C t ( t, t 0 ) ] (B41)
forcement and concrete are equal. For prestressing steel, this Ec ( t0 )
requirement applies to the change in strain occurring after
transfer of the prestress force to the concrete. Most of this where Ec(t0) is the modulus of elasticity of concrete at time
appendix deals with the analysis of stress and strain distribu t0. Ct(t,t0) is the creep coefficient, which is equal to the creep
tions in a cross section of a member. After the strain distri strain in the period t0 to t divided by the instantaneous strain.
bution has been determined, calculation of the deflected When the stress increment c is introduced gradually over
shape of the structure represents a geometry problem that has the period t0 to t, represented by Curve ADC in Fig. B4.1(a),
been covered in many textbooks (Ghali and Neville 1997); the total strain at time t is (refer to Curve EHI in Fig. B4.1(b)):
also refer to Section B8. The procedure of analysis presented
in this appendix can be used to evaluate the validity of
c
simplified analyses that do not adhere to the requirements of c ( t ) =  [ 1 + C t ( t, t 0 ) ] (B42)
equilibrium and compatibility, such as use of a multiplier to Ec ( t0 )
DEFLECTION IN CONCRETE STRUCTURES 435R79
c pr = r p (B45)
c ( t ) = 
 (B43)
E c ( t, t 0 )
where r is a function of the initial steel stress and the total
prestress change (Ghali et al. 2002); a value of r = 0.8 is a
where is the aging coefficient, and Ec (t,t0) is the age reasonable value to use in practice.
adjusted modulus of elasticity of concrete.
Examples of stresses that are introduced gradually include B5Sectional analysis
timedependent effects of shrinkage, prestress loss, and The procedure in Section B5.2 satisfies equilibrium and
support settlement. The aging coefficient is mainly a function strain compatibility at all reinforcement layers and applies to
of t0 and t (Ghali, Favre, and Elbadry 2002); a value of 0.8 can prestressing and nonprestressing reinforcement in cracked or
be used in most cases because varies between 0.7 and 0.9. uncracked sections. In a section posttensioned by a
Tabulated values of are given in ACI 209R. The aging coef unbonded internal or external tendon, the change in strain in
ficient is used as a multiplier to the creep coefficient, Ct(t, t0), the tendon due to loads applied after prestressing or due to
as shown in Eq. (B42). The aging coefficient was introduced creep, shrinkage, or relaxation is not compatible with the
by Trost (1967) and Baant (1975) to calculate the total strain, change in strain in the adjacent concrete. The procedure
including creep, due to a gradually introduced stress incre presented in Section B7.2 ignores this incompatibility; thus,
ment, as represented by Curve ADC in Fig. B4.1(a). it implies an approximation when applied to a structure
B4.2 ShrinkageThe strain that results from unrestrained having unbonded internal or external prestressed tendons
shrinkage of concrete between time t0 and t, SH(t,t0), (Ariyawardena and Ghali 2002).
depends mainly upon the size and the shape of the member, B5.1 Review of basic equationsThe following is the
the properties of concrete, the ambient air humidity, and the derivation of equations for the distribution of stress and
ages of concrete at t0 and t (ACI 209R). The same parameters strain over a homogeneous elastic cross section subjected
also influence the value of the creep coefficient Ct (t,t0). For to a normal force N at an arbitrary reference point O
the analysis of deformations in a specified period t0 to t, the combined with a moment M about a horizontal axis through
given data include the value of the hypothetical shrinkage O (Fig. B5.1). Assuming that a plane cross section remains
that would occur in the period if the concrete were free to plane, the stress and strain variations over the depth are
shrink. This value is: expressed by equations of straight lines:
where the time ts is the start of drying shrinkage. When the = O + y (B51b)
analysis is for ultimate shrinkage and creep, t = .
B4.3 Relaxation of prestressing steelThe loss of tensile where y is a distance of any fiber, measured from the reference
stress in a tendon that is elongated and then maintained at a horizontal axis through O, and and are strain and stress
constant length and temperature is referred to as the intrinsic at any fiber. The strain parameters O and define the strain
relaxation and is denoted p. The value p depends on the distribution; O is the strain value at O; = d/dy = curvature.
435R80 ACI COMMITTEE REPORT
M
= E (B52b) =  (B58b)
EI
where E is the modulus of elasticity of the homogeneous
material. Choosing O at a fixed point (for example, the top fiber)
rather than the centroid has the advantage that the location of
The stress resultants are:
the centroid does not have to be determined for the
uncracked and cracked stages. The equations in this section
N = dA (B53a) will be used for the structural concrete section, with A, B, and
I being properties of the transformed section, which is the
area of the concrete plus the area of the reinforcement multi
M = y dA (B53b) plied by the modular ratio (Ens or Eps for reinforcement
divided by Ec or Ec for concrete); where Ens and Eps are the
Substitution of Eq. (B51a) and (B51b) into (B53a) and moduli of elasticity of nonprestressing and prestressing rein
(B53b), respectively, gives the stress resultants forcement, respectively; and Ec and Ec are modulus and age
adjusted modulus of concrete, respectively. The transformed
section has modulus of elasticity equal to that of concrete.
N = AO + B (B54a)
When cracking occurs, only the area of concrete in compres
sion is included in the transformed section.
M = BO + I (B54b)
B5.2 Instantaneous and timedependent stress and
strainThis section describes a procedure to analyze imme
where A, B, and I are the crosssectional area and its first and diate increments of strains (t0) and stresses (t0) in a
second moments about a horizontal axis through O, respec prestressed or nonprestressed section and the changes in
tively. Thus, these values in a period t0 to t. The times t0 and t represent
instances in multistage construction when a load or
prestressing is applied, or when the support conditions are
A = dA; B = ydA; I = y2dA. modified. The term immediate means initial before occur
rence of creep, shrinkage, and relaxation.
When N and M are known, the corresponding strain and The given data are as follows (Fig. B5.2):
stress parameters can be determined by solution of Eq. (B54a) Gross concrete dimensions of the section, areas Aps and
and (B54b) and by using Eq. (52a) and (52b): Ans (with subscripts ps and ns referring to prestressing
and nonprestressing reinforcement, respectively);
IN BM
O = 
2
 (B55a) Coordinates y defining the location of each area;
AI B Magnitudes of normal force N and bending moment M
introduced at t0;
BN + AM Elasticity moduli Ec(t0), Eps , and Ens;
=  (B55b)
AI B
2 Creep and aging coefficients of concrete Ct(t,t0 ) and
(t,t0), respectively; and
Unrestrained shrinkage of concrete SH(t,t0), and
IN BM reduced relaxation pr.
O = 
2
(B56a)
E ( AI B ) The values of N and M, respectively, represent the total
normal force at the reference point O, and the bending
BN + AM moment introduced at t0 including any statically indeterminate
=  (B56b) effects of loads and prestressing. It is assumed that just
2
E ( AI B )
before introducing N and M, the section is subjected to linearly
varying strain and stress defined by the values at O and the
When O is chosen at the centroid, B = 0, and Eq. (B55a), derivatives with respect to y.
(B55b), (B56a), and (B56b) take the more familiar forms As an example, consider the cross section at the midspan
of the posttensioned, simple beam subjected at time t0 to a
N prestressing force P, and the member selfweight per unit
O =  (B57a)
A length q in Fig. B5.3(a) illustrates the meaning of the
DEFLECTION IN CONCRETE STRUCTURES 435R81
Step 2Determine parameters for the hypothetical strain Determine the stress resultants that cancel the effect of the
change that would occur, in the period t0 to t, if creep and prestressing steel relaxation
shrinkage were unrestrained, using:
Nrelax = Apspr (B515a)
O free(t,t0) = SH (t,t0) + CtO(t,t0) (B511a)
Mrelax = Aps yps pr (B515b)
free = Ct(t,t0)(t0) (B511b)
The sum of the two normal forces and the two bending
If the initial stress is not zero, creep due to stresses intro moments calculated in this step (Eq. (B514) and (B515))
duced earlier than t0 should be added, and the creep coefficients represent fictitious forces N and M, which would prevent
should be included in the input. ACI 209R, Ghali, Favre, and deformations due to creep, shrinkage, and relaxation.
Elbadry (2002), and CEBFIP (1990) give guidance on creep Eliminate the fictitious restraint by application of N and
coefficient values. M on the ageadjusted transformed section (whose properties
are A, B, and I). The ageadjusted transformed section is
Step 3Calculate the two parameters defining the distri
composed of Ac , nns Ans and nps Aps; where nns or nps is the
bution of a hypothetical stress gradually introduced between
modulus of elasticity of reinforcement divided by Ec(t,t0).
t0 and t to prevent the strain change calculated in Step 2,
Calculate the changes in strain and stress parameters (changes
using Eq. (B52a) and (B52b).
in O, , O, ) by substitution of N, M, Ec(t,t0), A, B,
and I in Eq. (B55a), (B55b), (B56a), and (B56b).
Orestraint = Ec O free (B512a) Update the initial stress parameters by adding the stress
parameters determined in Steps 1, 3, and 4. Update the initial
restraint = Ec free (B512b) strain parameters by adding the strain parameters determined
in Steps 1 and 4.
where Ec is the ageadjusted modulus of elasticity of concrete. B5.3 Commentary on the general procedure and on a
special case (nonprestressed sections subjected to M without
Ec(t,t0) = Ec(t0)/[1 + Ct(t,t0)] (B513) N)The general procedure gives the two parameters O,
defining the strain distribution in an uncracked prestressed or
nonprestressed section subjected to N and M. When cracking
Step 4Calculate A, B, and I for concrete alone, without occurs, an additional step is required; refer to Section B6.
reinforcement, and substitute these values together with Calculation of displacement in a plane frame from the strain
O restraint and restraint in Eq. (B54a) and (B54b) to obtain parameters is a classical geometry problem that can be
the resultants of stress required to prevent creep and solved more easily by virtual work or other methods. Trans
shrinkage deformations verse deflection in beams can be determined from the curva
ture only (without the need for O). The equations
Ncreep, shrinkage = Aconcrete O restraint + (B514a) presented below, which are derived from the general proce
dure in Section B5.2, give the immediate and the longterm
Bconcrete restraint curvatures at nonprestressed reinforced concrete sections
subjected to bending moment M.
Consider a nonprestressed reinforced concrete cross
Mcreep, shrinkage = BconcreteO restraint + (B514b)
section (Fig. B5.4) subjected at time t0 to a bending moment
M, without normal force. The general procedure of analysis
Iconcrete restraint in Section B5.2 reduces to Eq. (B111) to (B113) for initial
DEFLECTION IN CONCRETE STRUCTURES 435R83
curvature (t0 ) and the curvature changes ()creep and neutral axis at t0 (the centroidal axis of the trans
()shrinkage due to creep and shrinkage occurring between formed section at t0); and
t0 and t. The longterm curvature at time t is the sum of the yc = the ycoordinate of the centroid of Ac , measured
curvatures determined by the three equations downward from the centroid of the ageadjusted
transformed section.
I In most cases, the tension steel area As near the bottom
(t0 ) = g c (B516)
I fiber is larger than the compression steel area As and the
values of y and yc are positive and negative, respectively
()creep = (t0 )Ctcreep (B517) (Fig. B5.4(a) and (b)).
The symbol Ac means the active area of concrete at time t0.
Thus, when the previous equations are used for a cracked
cs

()shrinkage =  (B518) section, Ac is the area of concrete within the compression zone
d shrinkage at time t0. Derivation of the equations in this Appendix from
the general procedure in Section B5.2 and graphs for the coef
where c is the instantaneous curvature at an uncracked ficients creep and shrinkage are available (Ghali, Favre, and
homogenous concrete section (without reinforcement) Elbadry 2002).
intermediate values between the strain parameters in B7.3 Other tension stiffening approachesThe complete
uncracked and the cracked states. stressstrain curve for concrete in tension has been used by
B7.1 Bransons effective moment of inertiaFor a many researchers to model the tensionstiffening effect
member subjected to bending moment without a normal (Gilbert and Warner 1978; Polak and Vecchio 1993; Scanlon
force, only the curvature m is required to calculate trans and Murray 1974). This approach can be incorporated into
verse deflection. For a nonprestressed reinforced concrete the section curvature procedure by appropriate modification
section subjected to bending moment M without normal of the transformedsection properties. Introduction of the
force, Bransons equations (Branson 1977) for mean curva nonlinear stressstrain relationship would involve an itera
ture are: tive solution. Gilbert and Warner (1998) have modeled the
effect by increasing the effective stiffness of the reinforcing
M
m =  (B71a) steel, while Kaufmann and Marti (1998) have introduced a
EI e tensionchord model based on assumed bond stress distribu
tion between cracks.
M cr 4 M cr 4
I e = 
 I g + 1 
 I (B71b)
M M 2 B8Deflection and change in length of a frame member
The following equation can be used to calculate the trans
where verse deflection center at midlength of a straight member of
Ie = the effective moment of inertia; a frame
Ig = the second moment of gross concrete area about its cen
troidal axis, ignoring the presence of reinforcement; l
2
I2 = the second moment of area of the transformed center =  ( m1 + 10 m2 + m3 ) (B81)
96
cracked section about its centroidal axis;
Mcr = cracking moment; and
M = the moment on the section. where
B7.2 CEBFIP approachFor any prestressed or nonpre l = the member length; and
stressed reinforced section, the mean strain parameters Om m1, m2 , and m3 = mean curvatures at one end, at the cen
and m can be determined by interpolation between ter, and at the other end, respectively.
uncracked and cracked states Similarly, the rotation at member ends end and the
change in member length l can be determined by:
Om = (1 )O1 O2 (B72a)
l
m = (1 )1 + 2 (B72b) end1 =  ( m1 + 2 m2 ) and (B82)
6
where
O1 and 1 l
= strain parameters calculated using the end2 =  ( 2 m2 + m3 )
uncracked transformed section; 6
O2 and 2 = strain parameters calculated using the
cracked transformed section l
l =  ( Om1 + 4 Om2 + Om3 ) (B83)
= an empirical interpolation coefficient, 6
expressed by CEBFIP Model Code MC
90 (1990) as follows: where Om1, Om2, and Om3 are mean normal strains at one
end, at the center, and at the other end, respectively.
2
f ct
= 1  (when 1 max > fct, cracked) (B73) The deflection center and the rotation end are measured
1max
from the chord, which is the straight line joining the two ends.
Equations (B81) to (B83) assume parabolic or straight
= 0 (when 1 max < fct , uncracked) (B74) line variations of m and Om over the length of a frame
member. The method of virtual work (Ghali and Neville
where 1997) gives the translation in any direction and the rotation
fct = the tensile strength of concrete; at any section of a frame having any variation of m and Om.
1 max = calculated tensile stress at extreme fiber when Equations (B84) to (B86), which are derived by virtual
cracking is ignored; and work, can be used instead of Eq. (B81) to (B83) with any
= coefficient to account for the effects of bond variation of m and Om over the length of a frame member:
quality of reinforcement, cyclic loading, and
sustained loading on tension stiffening.
l2
In most cases, = 0.5 can be used when deformed bars are 1 l l
x
employed and cyclic or sustained loads are applied.
center = 
2 0 m x dx + 
2 l 2 m 1 l dx (B84)
DEFLECTION IN CONCRETE STRUCTURES 435R85
Fig. B10.2Example section subjected to moment: (a) cracked section at age t0 (neutral
axis through centroid of transformed section); and (b) properties of ageadjusted trans
formed section.
where ( 300 10 )
6
a1 = b/2 ()shrinkage =  (0.950) = 7.92 106 in.1
36
a2 = nAs + (n 1) As
a3 = nAs d (n 1)Asds The longtime curvature at time t:
This gives:
a1 = 6 in.
(t) = (54.25 + 20.51 + 7.92) 106 = 82.7 106 in.1
a2 = 48.75 in.2
a3 = 1475 in.3 B11References
c = 12.13 in. B11.1 Referenced standards and reportsThe following
The centroid of the transformed section at t0 coincides standards and reports listed were the latest editions at the
with the bottom edge of the compression zone. Other time this document was prepared. Because these documents
geometrical properties of the section are: are revised frequently, the reader is advised to contact the
I = 30,510 in.4 sponsoring group to refer to the latest version.
Ac = 144.4 in.2
The centroid of the ageadjusted transformed section is at: American Concrete Institute
y = 5.11 in. 209R Prediction of Creep, Shrinkage, and Tempera
I = 61,020 in.4 ture Effects in Concrete Structures
yc = 11.16 in. 318/318R Building Code Requirements for Structural
Ic = 19,670 in.4 Concrete and Commentary
435R88 ACI COMMITTEE REPORT
These publications may be obtained from: Press, London and New York, website: www.sponpress.com/
concretestructures.
American Concrete Institute Ghali, A., and Neville, A. M., 2003, Structural Analysis: A
P.O. Box 9094 Unified Classical and Matrix Approach, 5th Edition, Spon
Farmington Hills, MI 483339094 Press, London and New York, 844 pp.
Gilbert, R. I., and Warner, R. F., 1978, Tension Stiffening
B11.2 Cited references in Reinforced Concrete Slabs, Journal of the Structural
Ariyawardena, N., and Ghali, A., 2002, Prestressing with Division, V. 104, No. ST.
Unbonded or External Tendons: Analysis and Computer Hall, T. S., and Ghali, A., 1997, Prediction of the Flexural
Model, Journal of Structural Engineering, ASCE, V. 128, Behavior of Concrete Members Reinforced with GFRP
No. 12, pp. 14931501. Bars, Society for the Advancement of Material and Process
Baant, Z. P., 1975, Prediction of Concrete Creep Effects Engineering, V. 42, pp. 298310.
Using AgeAdjusted Effective Modulus Method, ACI
ISIS Canada, 2001, Reinforcing New Structures with
JOURNAL, Proceedings V. 69, No. 4, Apr., pp. 212217.
Fiber Reinforced Polymers, Design Manual No. 3.
Branson, D. E., 1977, Deformation of Concrete Structures,
Kaufmann, W., and Marti, P., 1998, Structural
McGrawHill, New York, 546 pp.
Concrete: Cracked Membrane Model, Journal of Struc
CEBFIB, 1990, Model Code for Concrete Structures,
tural Engineering, V. 124, No. 12, pp. 14671475.
CEB, Thomas Telford, London.
Ghali, A., and Azarnejad, A., 1999, Deflection Prediction Magura, D.; Sozen, M. A.; and Siess, C. P., 1964, A
of Members of Any Concrete Strength, ACI Structural Study of Stress Relaxation in Prestressing Reinforcement,
Journal, V. 96, No. 5, Sept.Oct., pp. 807816. PCI Journal, V. 9, No. 2, pp. 1357.
Ghali, A., and Elbadry, M., 1986 and 2000, Users Manual Polak, M. A., and Vecchio, F. J., 1993, Nonlinear Analysis
and Computer Program CRACK, Research Report CE851, of Reinforced Concrete Shells, Journal of Structural
Department of Civil Engineering, University of Calgary, Engineering, V. 119, No. 12, pp. 34393462.
Calgary, Alberta, Canada, Feb. Updated and renamed Scanlon, A., and Murray, D. W., 1974, Time Dependent
version: Reinforced and Prestressed Members, Computer Reinforced Concrete Slab Deflections, Journal of the
program RPM, American Concrete Institute, software code Structural Division, V. 100, No. ST9, pp. 19111924.
076AG2.CP. Trost, H., 1967, Auswirkungen des Superpositionsprinzips
Ghali, A.; Favre, R.; and Elbadry, M., 2002, Concrete auf Kriechund RelaxationsProbleme bei Beton und Spann
Structures: Stresses and Deformations, 3rd Edition, Spon beton, Betonund Stahlbetonbau, V. 62, No. 10, pp. 230238.
*
Members who prepared Appendix B.
DEFLECTION IN CONCRETE STRUCTURES 435R77
degrees Fahrenheit (F) ....................... degrees Celsius (C) ...................... t, = (tF  32)/1.8
degrees Celsius (C) ......................... degrees Fahrenheit (F) ..................... tF = 1.8t, + 32
*This selected list gives practical conversion factors of units found in concrete technology. The reference sources for information on SI units and more exact
conversion factors are ASTM E 380 and E 621. Symbols of metric units are given in parenthesis.
t E Indicates that the factor given is exact.
$ One liter (cubic decimeter) equals 0.001 m3or 1000 cm3.
0 These equations convert one temperature reading to another and include the necessary scale corrections. To convert a difference in temperature from Fahrenheit
degrees to Celsius de divide by 1.8 only. i.e., a change from 70 to 88 F represents a change of 18 F or 18/1.8 = 10 C deg.