Proceedings of the Institution of

Civil Engineers
Structures & Buildings 146
May 2001 Issue 2
Pages 195^202

Paper 12256
Received 07/01/2000
Accepted 21/11/2000

Keywords:
Andrew W. Beeby
concrete structures/concrete
Professor of Structural
technology & manufacture/slabs &
Design, School of Civil
plates
Engineering, University of
Leeds

Criteria for the loading of slabs during construction

A. W. Beeby
Two inequalities are derived to establish the loads that terms will be used in describing the site procedures and
may be imposed on slabs at an early age. These loads may equipment.
arise from the slab self-weight immediately after striking
the formwork or, slightly later in the construction cycle, . Formwork describes the shuttering to a slab plus the structure
when the slab is supporting construction loads from the necessary to support it from the slab below.
slabs above. In deriving the inequalities, it was concluded . Backprops are installed below the slab supporting the form-
that the criteria defining acceptable loading were related work to distribute some of the loads applied by the formwork
not to safety but to the possible impairment of service- and slab being cast to lower slabs. Backprops may be
ability performance. Results obtained during the con- required at more than one level.
struction of a seven-storey flat slab structure at BRE . Reprops are installed to support freshly struck slabs where
Cardington as part of the European Concrete Building these are considered to be insuf®ciently mature to support
Project were used to verify the proposals. their own weight and any construction loads.
. Props are adjustable vertical supports which may be used as
1. INTRODUCTION part of the formwork or as backprops or as reprops.
Two interconnected factors relating to the early age strength of
slabs have a signi®cant effect on the speed of construction of The structure itself was a seven-storey ¯at slab structure of
multi-storey reinforced concrete buildings. These are. 3 6 4 square bays with no downstand beams or column heads.
The centre to centre spacing of the columns was 7´5 m in both
. The time at which formwork can be struck; and directions and the slab had a thickness of 250 mm throughout.
. The requirements for backpropping to ensure that the slabs Fig. 1 shows the ¯oor plan, which was the same for all ¯oors.
supporting a freshly cast slab are not overloaded. The roof was constructed using an innovative form of
permanent formwork, which led to differences in the construc-
The construction of the seven-storey ¯at slab structure as part tion loads at this level. The clear storey height was 3´5 m
of the European Concrete Building Project at the Building throughout except for a rather greater height between ground
Research Establishment's Large Building Test Facility at and ®rst ¯oors. Further details of the structure, its design and
Cardington provided a unique opportunity to research these construction have been published elsewhere and will not be
issues on full scale structures under something approaching repeated here.1 A fuller description of this project is given in
laboratory conditions of control. A research programme to Beeby (2000).2
investigate early striking and backpropping was envisaged as
part of the Project from the earliest stages of planning, though 2. CRITERIA FOR LOADING SLABS AT EARLY AGES
the main thrust of the research was into the construction Since early striking of the formwork was to be a feature of the
process rather than the behaviour of the structure. The School project, it was necessary to establish some rationale for this
of Engineering at the University of Leeds was awarded a prior to the start of construction and then use the data gathered
contract by the Building Research Establishment (BRE) to during construction to validate the approach.
provide the technical input into this element of the project. In
effect, the University of Leeds worked with BRE and the The fullest treatment available to date of the problem of
Steering Committee to establish the test programme; BRE formwork striking times is given in CIRIA Report 136.3 The
installed the necessary instrumentation, carried out the actual argument is developed in this report that the critical factor
measurements and transmitted the results to Leeds. Leeds governing striking times is the strength of the members. It is
carried out the theoretical studies, processed the experimental postulated that the strength of a member is proportional to the
data and reported the results. This paper describes the work concrete strength and this assumption is used to develop rules
carried out by the University of Leeds on the de®nition of for striking times. This assumption of a linear relation between
criteria for the early loading of the ¯oors. concrete strength and member strength is certainly not true for
slab structures, which tend to have relatively low reinforcement
Before describing the details of the project, it may be helpful to percentages. In this circumstance, the ¯exural strength is
de®ne some terms. For the purposes of this paper, the following dominantly a function of the strength and quantity of the

Structures & Buildings 146 Issue 2 Criteria for slab loading Beeby 195

Downloaded by [ City, University of London] on [21/06/19]. Copyright © ICE Publishing, all rights reserved.
 that, since cube strength is
7·5 7·5 7·5 7·5 roughly proportional to
cylinder strength, tensile
D
1 1 strength is proportional to
(cube strength)0´6.
2 3 2 3 2
7·5
The modulus of elasticity, Ec,
also varies as a function of
4 4
C the concrete compressive
strength and design codes
give a variety of functions.
5 6 5 6 5 7·5 For example, Part 2 of
BS 8110 gives the relationship
for normal weight concrete
B 4 4
as:

2 3 2 3 2 1 Ec = 20 + 0´2fcu
7·5

1 1
where fcu is the the charac-
A
teristic cube strength and Ec
is the modulus of elasticity of
1 2 3 4 5
concrete.

Fig. 1. Floor plan of flat slab structure at Cardington; numbered points indicate positions where Eurocode 2 (converted from
measured deflections were averaged cylinder strength to cube
strength) gives the very dif-
ferent relationship:5

¯exural reinforcement and is only marginally affected by the

2 Ec = 8´8(10 + fcu)(1/3)
concrete strength. For example, halving of the concrete
strength from 37 N/mm2 to 18´5 N/mm2 in a slab with 0´5% of
Study of the results presented in Jaccoud et al.6 suggest that the
¯exural reinforcement will only lead to a reduction in the
trend of Ec is adequately re¯ected by the assumption that it
¯exural strength of 7%.
varies proportionally to (compressive strength)0´6, though this
does not give the `best ®t' to the data (see Fig. 2b).
Shear strength, which might be a critical factor in ¯at slabs, is
taken in BS 8110 to be proportional to the cube root of the 3. CONTROL OF DEFLECTIONS
concrete strength so, again, the in¯uence of concrete strength is If it is assumed that a slab will be dominantly uncracked then
relatively small.4 It therefore seems that the ultimate strength of its de¯ection will be given by
the slabs is not likely to be the critical issue de®ning the
striking time. It seems far more probable that serviceability
3 a = kwL4/EcI
considerations will be critical. The signi®cant serviceability
issues are crack widths and, more importantly, de¯ections.
where a is the de¯ection, k is a constant depending on the
These two phenomena are interrelated since the de¯ection
boundary conditions and type of loading, L is the span, I is the
depends to a marked degree on the extent of cracking.
second moment of area of the section and w is the loading per
unit area on the slab. Hence, for a given slab with a given
What follows is an attempt to develop a more rational approach arrangement of loading, the de¯ection may be taken to be
to deciding on the ability of slabs to support construction loads proportional to w/f 0´6
c , where fc is the cube strength of the
at early ages without impairing the service performance of the concrete at the time considered. In the design of the slab, it is
®nished structure. assumed that the concrete strength is equal to the speci®ed
characteristic strength, fcu and the load will be the speci®ed
In the ®rst instance, it will be helpful to consider the service load, wser. The de¯ection implicit in the design will thus
relationships between the tensile strength and elastic modulus be proportional to wser/f 0´6
cu and the de¯ection will not exceed
of concrete and the compressive strength. The tensile strength that implicit in the design if
of the concrete at any age may be related approximately to the
cube strength at that age. The relationship between cube 4 (w/wser)( fcu/fc)0´6  1
strength and tensile strength varies from source to source. For
example, Eurocode 25 assumes that tensile strength is propor- The factor (w/wser)( fcu/fc)0´6 may conveniently be called the
tional to (cylinder strength)2/3 whereas BS 8110 assumes that it Cracking Factor, Fcr giving
is proportional to (cube strength)1/2. Recent extensive studies at
EPFL Lausanne suggest (cylinder strength)0´6 (see Fig. 2a).6 This Fcr  1
a
seems a reasonable compromise and so it will be assumed here

196 Structures & Buildings 146 Issue 2 Criteria for slab loading Beeby

Downloaded by [ City, University of London] on [21/06/19]. Copyright © ICE Publishing, all rights reserved.
 10
Tensile strength = 0·38fc0·6
9

Splitting tensile strength: N/mm2 8

0
(a)
70

Modulus of elasticity = 2·7fc0·6 kN/mm2

60

50
Modulus of elasticity: kN/mm2

40

30

20

10

0
0 20 40 60 80 100 120 140 160 180 200

Compressive cylinder strength: N/mm2

(b)

Fig. 2. Prediction of concrete properties from compressive strength: (a) relationships between tensile strength and concrete
compressive strength; (b) relationship between modulus of elasticity and compressive strength (data from Jaccoud et al.6)

In a fully cracked section, an increase in Ec leads to a close to the cracking load and increasing towards the fully
compensating decrease in (a) and so the effect of concrete cracked de¯ection as the load increases above the cracking
strength is much less and the de¯ection relative to that under load. This is re¯ected in the equation given in Eurocode 2 for
the service load is more closely related to (w/wser). If (w/wser) is the calculation of de¯ection. This is
de®ned as Fw, then, for a fully cracked slab, the re¯ection will
not exceed that implicit in the design provided that 5 1/r = (1 7 z)(1/r)1 + z(1/r)2

b Fw  1´0 where

In reality, once cracking has occurred, the actual de¯ection will 6 z = (1 7 b1b2(ssr/ss)2)
lie somewhere between that of an uncracked slab and a fully
cracked slab, being close to the uncracked de¯ection at loads 1/r is the curvature at a section allowing for the effects of

Structures & Buildings 146 Issue 2 Criteria for slab loading Beeby 197

Downloaded by [ City, University of London] on [21/06/19]. Copyright © ICE Publishing, all rights reserved.
 tension stiffening; (1/r)1 is the curvature calculated on the basis 4. CONTROL OF CRACKING
of an uncracked section; (1/r)2 is the curvature calculated on The moment ®eld within the structure is proportional to w (i.e.
the basis of a cracked section; b1b2 are coef®cients depending the moment at all points in the structure would be halved if the
on the bond characteristics of the bar and whether the loading loading were halved) and the cracking moment will be
is long or short term; ssr is the stress in the reinforcement proportional to the tensile strength of the concrete. Thus, the
calculated for a fully cracked section under the cracking parameter (w/fc)0´6 will be proportional to the ratio of the
moment; and ss is the stress in the reinforcement under the moment to the cracking moment at any point and can be used
load considered calculated on the basis of a fully cracked to de®ne the extent of development the cracking. What is
section. meant here by the `extent of cracking' is the proportion of the
structure where the moment has exceeded the cracking moment
The de¯ection may be found from the curvature from the and hence ¯exural cracking will be present. In the case of a ¯at
relationship slab structure, this is likely to be the size of the area
surrounding the columns within which the cracking moment
7 De¯ection, a = $$(1/r)dx will be exceeded. There will be a substantial reduction in
stiffness in cracked zones relative to uncracked zones. Clearly,
It is found that this approach does not lend itself to the the greater is the factor (w/fc)0´6, the larger will be the cracked
de®nition of a simple factor for judging whether the de¯ection zones and the lower will be the stiffness of the structure.
will be less than that implicit in the design though it is
reasonable to assume that the appropriate factor for judging The extent of cracking implicit under service conditions in the
this would lie between Fcr and Fw. A reasonable proposition for design is de®ned by (wser/f 0´6
cu ) where wser is the design service
such a factor might be load and fcu is the speci®ed 28-day characteristic cube strength.
In any circumstance where (w/fc)0´6 exceeds (wser/f 0´6
cu ) or Fcr > 1,

c Feff  1 the cracking will be more extensive than implicit in the design.

where Feff is the effective de¯ection factor The width of cracks depends on the geometry of the section
considered and the average strain at the tension face. For a
8 Feff = (1 7 z)Fcr+ zFw
fully cracked section, this strain is roughly proportional to the
applied loading and hence the crack widths in a fully cracked
The expression for z may be simpli®ed to a function of the slab will not exceed those implicit in the design if Fw  1.
load. A brief parameter study suggests that the best result is
obtained if the following relationship is used As with de¯ections, however, the crack width will actually be
somewhat smaller than that calculated assuming a fully cracked
section but the degree of this reduction can only be estimated if
10 z = 1 7 (wcr/w)
the load causing cracking is known. If this can be calculated,
then the crack widths would be satisfactory provided that
where wcr is the load at which the slab cracks.
zFw  1. Under normal construction conditions, however, this
would be dif®cult to assess and it is more reasonable to make
It is likely to be awkward to calculate the cracking load for the
the safe assumption that satisfaction of inequality (a) will
slab and it will be safe, and much more convenient to assume
ensure that the extent of cracking is not greater than implicit in
that the de¯ection will not exceed that implicit in the design by
the design and satisfaction of inequality (b) will ensure that the
satisfying the two inequalities:
crack width is not greater than is implicit in the design.

a Fcr  1´0
Ensuring that performance (both from the point of view of
cracking and de¯ections) under construction loads does not
b Fw  1´0 affect future performance thus now requires that two conditions
are met:
The factors considered above do not address the possibility that
early loading may result in excessive de¯ections due to creep.
a Fcr  1´0
Unfortunately, no such simple approach seems possible for
considering creep deformation due to early loading as has been
developed for short-term effects. A major problem is that no b Fw  1´0
clear statement can be made about the assumptions on time of
loading implicit in design. A short parameter study, using the From the earlier discussion, it will be remembered that
creep model given in the CEB Model Code7 and considering application of the two inequalities above is slightly conserva-
various scenarios for striking and loading the slabs, suggests tive and that cracking and de¯ections will not exceed those
that early loading does not have a very signi®cant effect on the implicit in the design provided that
®nal total de¯ection under normal circumstances. CIRIA Report
136 arrives at the same conclusion.3 If this is accepted then the c Feff  1
inequalities (a) and (b) remain valid for long-term effects also
and thus remain a reasonable basis for decisions on formwork Some limited exceedance of Fw may therefore be allowed
striking, repropping and backpropping to ensure adequate without damaging future performance. It is doubtful, however,
serviceability performance. if designers would consider signi®cant exceedance of the

198 Structures & Buildings 146 Issue 2 Criteria for slab loading Beeby

Downloaded by [ City, University of London] on [21/06/19]. Copyright © ICE Publishing, all rights reserved.
 design service load during construction as reasonable and so supporting the formwork and backprops were calculated using
major violations of (b) should be avoided. This is particularly the method ®rst proposed by Grundy and Kabaila.8 This method
important if signi®cant amounts of redistribution have been makes the following assumptions.
carried out in the design.
(a) The slabs behave elastically and are assumed to have equal
It is proposed that inequalities (a) and (b), as well as being used stiffness.
to establish the striking time for the formwork, may also be (b) The props supporting the slabs and formwork may be
used to establish the loading that may be imposed on ¯oors at regarded as being in®nitely rigid.
any time during the construction process. The studies of the (c) Shrinkage and creep of the concrete may be ignored.
Cardington structure show that inequality (a) governs the (d ) The foundation may be regarded as rigid.
striking time and will permit the striking of formwork (e) The loads applied to a slab by formwork and backprops are
considerably earlier that do the current rules. Inequality (b) assumed to be uniformly distributed.
tends to be critical for the slab supporting the formwork at the
time of concreting the next ¯oor above. The design service load for the slabs at Cardington was
10´5 kN/m2.
There is an issue of uncertainty that needs to be addressed
when considering the practical use of inequality (a). This is in BRE placed load cells under the majority of the props
the assessment of the strength of the concrete used to assess Fcr. supporting the formwork and all the backprops within the bays
In the work presented here, the concrete strength has been considered. As far as possible, BRE staff recorded the following
determined by use of temperature matched cured cubes and this information from the time the backprops were ®xed to the time
was considered to give a very reliable estimate of the actual the formwork was struck.
concrete strength. Other means of assessing strength could be
used and these might be considerably less reliable. It would be . The ambient temperature and relative humidity.
necessary in such cases to introduce some factor into the . The loads in the instrumented props supporting the formwork
equation to allow for this uncertainty. This issue is currently and the backprops.
under consideration by CONSTRUCT and recommendations will . The temperatures in the concrete of the slab being cast,
be given in their `Guide for Flat Slab Formwork and Falsework' measured by thermocouples.
when this document is ®nalised. It is not considered further . The deformations of all slabs within the instrumented areas
here. measured by transducers.

5. OUTLINE OF EXPERIMENTAL WORK AND RESULTS These data were recorded electronically and sent by e-mail to
In the planning of the project, it had been decided to project teams.
concentrate the instrumentation in two panels: the panels
between Grid lines 2, 3, A and C (see Fig. 1). These were chosen 6. PRELIMINARY DISCUSSION OF RESULTS
so as to cover an edge panel and an interior panel, which were Table 1 summarises the loads on the various ¯oors at critical
as `typical' as possible and which were not complicated by stages in the construction together with the concrete strengths
awkward boundary conditions. Some information was also at each stage and the calculated values of the control
obtained for the corner panel, AB-12 but this paper will parameters Fcr and Fw based on the calculated and the
exclusively be concerned with Panels AB-23 and BC-23. measured loads. The design service load, wser, has been taken as
10´5 kN/m2 and the speci®ed characteristic cube strength was
The procedure followed in the striking of the formwork was 37 N/mm2. In the table, the construction stages relate to a fuller
that inequality (a) was used, together with the design construc- programme in the main report but are maintained here for
tion loads on the ¯oor, to establish a concrete strength at which convenience of reference. The construction step relevant to
the formwork could be struck. Once this strength had been each stage is given in Table 2.
achieved, the formwork for the whole ¯oor was generally
lowered away from the slab over a period of a few hours. The most immediately noticeable difference between the
Backprops were inserted as early as practicable before casting calculated loads and measured loads on the slabs is that the
of the next slab. measured forces are always signi®cantly less than the calcu-
lated values. Only in one case (load on slab 3 at Stage 13) do
There are many possible procedures that might be used for the calculated and measured loads correspond. The overestima-
installing backprops but the procedure generally adopted at tion of the loads is not necessarily unexpected or unreasonable;
Cardington was the following. When backprops were installed, the calculated loads are design loads that have been selected to
they were placed in position but generally not tightened cover the worst likely situation during construction. It is
signi®cantly so that the backprops were only loaded when suspected that the construction loads were well below the
additional loads are applied to the system through the props design values of 0´75 kN/m2 on existing slabs and 1´5 kN/m2 on
supporting the formwork. Generally, as there was no reprop- slabs being concreted and the measurements simply re¯ect this.
ping, the freshly struck slab thus supported its own weight and Whether there is scope for any reduction of the assumed
any construction loads up until the insertion of the backprops. construction imposed loads is not an issue which will be
This meant that the most signi®cant load carried by the pursued further here.
backprops resulted from the casting of the next ¯oor above.
Comment must be made on the very high calculated loads on
The `design' loadings for the backprops and the ¯oors Floor 6 at Stage 26. This was the result of the backprops being

Structures & Buildings 146 Issue 2 Criteria for slab loading Beeby 199

Downloaded by [ City, University of London] on [21/06/19]. Copyright © ICE Publishing, all rights reserved.
 Stage Floor Design floor Measured Temperature Values based on design loads Values based on measured
load floor load matched cube loads
(kN/m2) (kN/m2) strength
(N/mm2) Fcr Fw Fcr Fw

2 1 7.00 6.00 24 0.86 0.67 0.74 0.57

5 1 8.25 N/A 52 0.64 0.79 N/A N/A
6 2 7.00 6.00 24 0.86 0.67 0.74 0.57
9 1 10.50 8.20 65 0.71 1.00 0.56 0.78
2 11.75 10.32 47 0.97 1.12 0.85 0.98
10 3 7.50 6.00 23 0.95 0.71 0.76 0.57
13 1 9.33 6.90 67 0.62 0.89 0.46 0.66
2 9.33 7.35 50 0.74 0.89 0.58 0.70
3 10.58 10.57 45 0.90 1.01 0.90 1.01
14 4 7.00 6.00 28 0.79 0.67 0.68 0.57
17 3 10.50 8.50 51 0.82 1.00 0.67 0.81
4 11.75 9.90 58 0.85 1.12 0.72 0.94
18 5 7.00 6.00 27 0.79 0.67 0.69 0.57
21 4 10.50 8.20 59 0.76 1.00 0.59 0.78
5 11.75 10.32 42 1.04 1.12 0.91 0.98
22 6 7.50 6.50 14 1.28 0.71 1.10 0.62
25 5 9.70 7.80 48 0.79 0.92 0.64 0.74
6 12.69 9.50 38 1.19 1.21 0.89 0.90
26 6 15.25 10.10 38 1.43 1.45 0.95 0.96
7 0Í00 1.20 25 0Í00 0.00 0.14 0.11
27 7 7.00 6.00 34 0.70 0.67 0.60 0.57

N/AöNot available

Table 1. Loading parameters for slabs

Stage Construction activity Application of the formula in CIRIA Report 136 is not
2 Formwork to first floor struck straightforward as the design calculations for the ¯oors are not
5 Second floor concreted available and the CIRIA report emphasises that the analysis
6 Formwork to second floor struck method used to assess the moment must be the same as was
9 Third floor concreted used in the design. A rough estimate, based on the measured
10 Formwork to third floor struck loads and using the coef®cients in BS 8110 for analysis,
13 Fourth floor concreted
14 Formwork to fourth floor struck suggests that the CIRIA approach would indicate that the slabs
17 Fifth floor concreted were overstressed by a factor of between 1´4 and 2´2 on striking
18 Formwork to fifth floor struck the formwork.
21 Sixth floor concreted
22 Formwork to sixth floor struck The condition when the slab is supporting the prop forces from
25 Roof concreted
26 Backprops between fifth and sixth floors struck concreting the ¯oor above, which the proposed method
27 Formwork to roof struck suggests will be the most critical condition, is shown to be less
critical by the CIRIA method with overstresses only up to about
Table 2. Construction stages in Table 1 40%. Clearly the proposed method permits much earlier form-
work striking and rather higher loads at later stages in the
construction.

accidentally removed before the formwork was struck. For this It was also found that the loads carried by the backprops were
case, the calculation has assumed that the whole of the weight considerably smaller than calculated using the Grundy and
of the freshly cast roof slab would be supported by the 6th Kabaila method. This results from ignoring the stiffness of the
¯oor. In fact, it is clear that, after removal of the backprops, the props and the different locations of the props supporting the
roof actually supported a signi®cant proportion of its self- formwork and the backprops. The development of improved
weight, even though the formwork had not been stripped. methods of calculating the forces in backprops will be the
subject of a further paper.
It will be seen from Table 1 that, ignoring the accidental stage
26, while calculation based on the design loads suggests that 7. DEFLECTIONS AFTER CONSTRUCTION
Fcr may exceed 1´0 by up to 18% and Fw by up to 12%, during The de¯ections of all ¯oors before and after striking the
the actual construction Fw was never exceeded and Fcr only formwork, and at several later occasions, were measured by
once by 10%. This was on striking the formwork to Floor 6 precise levelling. It is useful to look at the resulting data and
where the concrete strength was lower than for the other ¯oors attempt to establish whether the level of loading during
and the formwork for the roof was already being assembled construction had any discernible effects on the longer-term
above. performance of the ¯oors. For simplicity, only the results

200 Structures & Buildings 146 Issue 2 Criteria for slab loading Beeby

Downloaded by [ City, University of London] on [21/06/19]. Copyright © ICE Publishing, all rights reserved.
 obtained on 23 November
1998 are considered here. 15 20
This is about eight months Location 1 Location 2
after completion of the struc-
ture. To get a more consistent

Deflection: mm
10 15
picture, the de¯ections of all
nominally similar locations
on a particular ¯oor have 5 10
been averaged. The locations
considered are marked as 1 to
5 on Fig. 1. 0 0
(a) (b)
The de¯ections are all calcu- 25 15
lated relative to the columns Location 3 Location 4
and have been corrected for 20
Deflection: mm

the measured de¯ections at 10

the column locations. The 15

design of each ¯oor was
5
different and details of the 10
design procedures used are
given in Furness et al.1 The 0 0
critical factor in the design, (c) (d)
as far as this exercise is 15 15
concerned, is that the weight Location 5 Location 6
of reinforcement supplied
varied from 16´7 tons in the 10 10
Deflection: mm

3rd ¯oor to 25´52 tons in the

6th ¯oor. Because of the
slender nature of the ¯oors, it
5 5
was found necessary in the
basic designs to increase the
quantities of reinforcement
above those required for ulti- 0 0
0·6 0·8 1·0 1·2 0·6 0·8 1·0 1·2
mate to meet the de¯ection
Cracking factor, Fcr Cracking factor, Fcr
requirements. This was not (e) (f)
done for the 3rd ¯oor and
hence the low weight of steel
Fig. 3. The deflection at various locations at an age of approximately 260 days plotted against the
in this slab. cracking factor, Fcr

Floors 1 and 2 had roughly

the minimum reinforcement
necessary to meet the de¯ection requirements (17´7 and Location Increase (%)
17´9 tons respectively) while the reinforcement in Floors 4 to 6
was rationalised in various ways, leading to higher steel 1 90
2 109
weights. The precise effect of these changes in reinforcement 3 104
weight is dif®cult to calculate but it is reasonable to suppose 4 112
that the de¯ection would decrease as the weight of reinforce- 5 76
ment increased if the concrete had cracked to any signi®cant 6 115
degree.
Table 3. Percentage increase in deflection predicted by
To allow for this, the de¯ections have been normalised by regression analysis between values of Fcr of 0Í6 and 1Í1
multiplying them by the weight of reinforcement in the slab
divided by 19´6, the average weight of reinforcement used.
These normalised average de¯ections are plotted in Fig. 3(a)±(f) Overall, this suggests that the maximum value of the parameter
against the maximum value of Fcr to which each ¯oor was Fcr reached during the construction process does relate directly
subjected (taken from Table 1). Lines were ®tted to the data for to the de¯ection at later ages. Hence excessive values of this
each location by linear regression analysis and this shows a during construction will lead to larger de¯ections in the long
signi®cant increase in de¯ection with increase in Fcr in all term. De¯ections have been obtained for Locations 3 and 6
cases. To gain a clearer idea of the magnitude of this increase, (respectively mid span of edge and interior panels) at an age of
the percentage increase in de¯ection predicted by the regres- 500 days. At this time the slabs were all carrying an imposed
sion lines for a change in Fcr from 0´6 to 1´1 has been load of 3´27 kN/m2. This gave a total load on the slabs of
calculated and is tabulated in Table 3. 9´27 kN/m2, which is close to the design service loading of

Structures & Buildings 146 Issue 2 Criteria for slab loading Beeby 201

Downloaded by [ City, University of London] on [21/06/19]. Copyright © ICE Publishing, all rights reserved.
 10´5 kN/m2. Nevertheless, the
value of Fcr reached is only 35 20

around 0´7 because of the Location 3 Location 6

concrete strength being sig- 30

ni®cantly above the speci®ed 15

Deflection: mm
25
characteristic value at this
time. This is below the values
of Fcr reached in all slabs 20
10
during construction.
15

Figure 4 shows the de¯ec-

tions at these two locations at 10 5
0·6 0·8 1·0 1·2 0·6 0·8 1·0 1·2
500 days plotted against the Cracking factor, Fcr Cracking factor, Fcr
highest values of Fcr reached (a) (b)
during construction, which
are the highest values reached Fig. 4. The deflection at locations 3 and 6 at an age of approximately 500 days plotted against the
at any time during the life of cracking factor, Fcr
the ¯oors so far. It can be
seen that a clear relationship
between the maximum value
of Fcr reached and de¯ection at this time is still obtained. This 9. ACKNOWLEDGEMENTS
reinforces the view that Fcr is an excellent parameter for This paper is published by kind permission of the Building
de®ning the stiffness of slabs as a function of load history and Research Establishment. The co-operation of Dr Richard Moss
concrete strength. Furthermore, the relationship has remained of BRE is particularly acknowledged.
effective regardless of the different amounts of creep that have
occurred at different ages.
REFERENCES
The effectiveness of the other parameter, Fw, has not been able 1. FURNESS L., CHANA P. S. and MOSS R. M. Design of an in-situ
to be tested from the Cardington data since all slabs except the concrete building for research. Paper presented at IStructE
®rst ¯oor were loaded during construction to approximately the Seminar on Aspects of Cardington Full-Scale Testing, 13
same value of Fw (between 0´94 and 1´01). May 1999.
2. BEEBY A. W. Early Striking of Formwork and Forces in
8. CONCLUSIONS Backprops. Building Research Establishment Report BR
It is proposed that the early loading of slabs either due to 394, 2000.
striking of formwork or due to other construction loading 3. HARRISON T. A. Formwork Striking TimesÐcriteria, predic-
should satisfy the following two relationships: tion and methods of assessment. CIRIA Report 136.
Construction Industry Research and Information Associa-
a Fcr  1´0 tion 1995.
4. BRITISH STANDARDS INSTITUTION. Structural use of ConcreteÐ
b Fw  1´0 Part 1. Code of practice for design and construction. BSI
Milton Keynes, 1997, BS 8100: Part 1.
where Fw is the loading factor, (w/wser), Fcr is the cracking 5. BRITISH STANDARDS INSTITUTION. Eurocode 2, Design of
factor, (w/wser)( fcu/fc)0´6, w is the load per unit area of the slab concrete structures, Part 1: general rules and rules for
at the time considered, wser is the design service loading on the buildings. BSI Milton Keynes, 1992, DD ENV 1992-1-
slab, fcu is the characteristic concrete strength assumed in the 1:1992.
design and fc is the concrete strength at the time considered. 6. JACCOUD J-P., FARRA B. and LECLERCQ A. Tensile strength,
Modulus of ElasticityÐBond-tension stiffeningÐLimit
Experimental results obtained from the European Concrete state of Cracking. Report of joint CEB/FIP Working Group
Building Project at Cardington suggest that the above equations on HSC/HPC. IBAP, EPF Lausanne, March 1995.
are soundly based and provide an acceptable and economic 7. CEB BULLETIN D'INFORMATION No. 213/214: CEB/FIP Model
means of deciding upon the capacity of slabs during construc- Code 1990, CEB Lausanne, May 1993.
tion. In particular, the de¯ection of the ¯oors at Cardington 8. GRUNDY P. and KABAILA A. Construction loads on slabs with
were shown to vary linearly with the maximum value of Fcr to shored formwork in multi-storey buildings. Journal of the
which the ¯oor had been subjected at ages of up to 500 days. American Concrete Institute, 1963, December.

Please email, fax or post your discussion contributions to the secretary: email: sismeyh@ice.org.uk; fax: +44 (0)20 7799 1325; or
post to Helen Sismey, Journals Department, Institution of Civil Engineers, 1^7 Great George Street, London SW1P 3AA.

202 Structures & Buildings 146 Issue 2 Criteria for slab loading Beeby

Downloaded by [ City, University of London] on [21/06/19]. Copyright © ICE Publishing, all rights reserved.