Engineering Structures 32 (2010) 3629–3638

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Shore overloads during shoring removal

M. Azkune, I. Puente ∗ , A. Santilli
Department of Mechanical Engineering, Institute of Civil Engineering, Tecnun (University of Navarra), Manuel de Lardizabal 13, 20018 San Sebastian, Spain

article info abstract

Article history: The considerable overloads originating during the shore removal process can affect the structural safety of
Received 2 September 2009 a multistory concrete building under construction. With an incorrect shoring removal reversal of stresses
Received in revised form may occur which can cause concrete cracking and excessive deflections. Moreover, excessive overloads
24 May 2010
will damage the falsework, producing important economical losses due to its elevated cost.
Accepted 9 August 2010
Available online 9 September 2010
A measured program has been conducted during the shore stripping at different building levels.
Results show that in general the Refined Method is adequate in a conservative form for the shore removal
Keywords:
procedure.
Shore removal Then, the influence of different parameters such as concrete strength or steel reinforcement in
Multistory buildings shore removal overloads was studied. The use of the actual concrete strength and modeling the steel
Flat concrete slabs reinforcement produces a more accurate theoretical result. Nevertheless, these considerations do not
Overloads bring major changes (less than 4%).
Falsework Finally, shore overloads originating in five different shore removal procedures in a typical structure
Concrete construction have been compared, establishing some criteria for a safe shore stripping sequence. For example the best
Shoring stripping procedure found consists of removing shores by rows.
Reshoring
© 2010 Elsevier Ltd. All rights reserved.

1. Introduction interconnected slabs. The accidents that occurred in Cocoa Beach,

Florida in 1981 reported by Lew et al. [1] and in Fairfax County,
For economic reasons, there is ever more pressure to construct Virginia in 1973 reported by Carino et al. [2], where a large number
multistory concrete buildings at a faster pace. Thus, the strength of workers lost their lives, are well known examples.
of a slab usually is not enough to support an upper floor when It is recognized that a detailed knowledge of shore-slab inter-
it is cast. Consequently, self-weight of a newly poured slab is action may well have prevented some of these accidents. Nev-
distributed between lower partially hardened slabs interconnected ertheless, due to factors such as the variable nature of concrete
by shores and/or reshores. properties, determining load values and distribution between
Moreover, an excessive number of shored floors are not advis- shores and slabs is complex.
able. To keep the structure shored to the foundation would increase Several research studies have been conducted to accurately de-
construction costs considerably, due to the fact that the excessive termine the strength properties of a concrete building during con-
number of shores and falsework needed would be too high. This struction. Concrete has been studied by authors such as Price [3],
procedure could also result in exceeding ultimate shore loads on Klieger [4], Gardner and Poon [5] or Carino et al. [6], who have an-
the lower floors. alyzed the influence of factors such as curing temperature in con-
The temporary supporting structure is removed when the slab crete strength evolution.
The first known research paper on the subject of load estimation
is sufficiently resistant which allows the liberated shores to be used
was published in 1952 when Nielsen [7] presented a method which
for the construction of upper floors. This reduces the number of
was too complex to be used in practice. In 1963 Grundy and
shores needed and allows work by other trades to proceed on the
Kabaila [8] developed the Simplified Method. This pioneering work
lower floors.
was based on the following assumptions, which make the method
Structural safety, however, is also of paramount importance.
clear and easy to apply:
Load carrying capacities of shores and slabs must not be exceeded.
A significant percentage of structural collapses during construction 1. Relative to the bending stiffness of slabs, the axial stiffness of
is caused by excessive loads on both the shoring system and shores and reshores is assumed to be infinite.
2. Despite the fact that concrete properties vary with age, all slabs
are assumed to possess equal flexural stiffness.
3. The lowest level of shores or reshores is assumed to be sup-
∗ Corresponding author. Tel.: +34 943219877; fax: +34 943311442. ported on a completely rigid foundation.
E-mail address: [email protected] (I. Puente).

0141-0296/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engstruct.2010.08.007
3630 M. Azkune et al. / Engineering Structures 32 (2010) 3629–3638

This Simplified Method is the most widely used method in the casting of the floor. In both cases loads were measured on
the construction industry, and is also suggested in the shoring/ shores located underneath the slab which was being poured. The
reshoring guide published by ACI Committee 347, 2R-05 [9]. registered peak loads were compared with the construction live
Other theoretical models have been proposed in recent years. loads proposed by the ACI Committee 347 [29] and the European
In general, these models include modifications that try to improve UNE-EN 12812 [30]. It was concluded that both standards are
the results achieved by the Simplified Method, as well as to verify adequate and safe.
the accuracy of this method. In 1985 Liu et al. [10] developed the Rosowsky et al. [25] have collected experimental data during
Refined Method. This finite element based computational method the shore stripping process. The measurements consisted of
proposes a more realistic model of the actual structure. However, registering the redistribution of loads due to the removal of
this method is cumbersome for practical use. Liu et al. [11] several shores located on the same floor. They concluded that
compared on-site measurements with results obtained by both shore removal originates considerable overloads on the remaining
the Simplified and the Refined Methods, and concluded that the shores.
Refined one predicts more accurate values. Nevertheless, they During the shore stripping process of a slab, the load of a
pointed out that the Simplified Method can be used if the results removed shore is redistributed between the concrete structure
for the maximum shore load and slab moments are corrected by a and the remaining shores. This redistribution may lead to abrupt
modification coefficient that varies from 1.05 to 1.10. overloads that can cause damage to the shores, which would imply
Other models have been developed by Gardner [12], Stivaros important economical losses.
and Halvorsen [13], Chen and Mossallam [14], Mossallam and In the present work, on-site measurements have been carried
Chen [15], El-Shahhat and Chen [16], Duan and Chen [17], Fang out during several shore stripping processes of slabs. The objective
et al. [18] and Miranda de Almeida et al. [19]. of this work is to determine the magnitude of the shore overloads
But the significant amount of theoretical research published
during the shore removal process and to determine the factors that
contrasts with the absence of detailed experimental data related
affect the load’s redistribution. Several shore removal procedures
to the subject area. The first on-site measurements consisted of
are investigated and their effect on the shore overloads is
measuring shore loads and comparing experimental data with the
evaluated. The results of this study will help contractors develop
Simplified Method. For example, Agarwal and Gardner [20], Lasisi
cost effective and safe shore removal procedures.
and Ng [21] and Moragues et al. [22,23] carried out these types
of measurements during the construction of high-rise concrete
buildings. Data collected in Agarwal and Gardner [20], Lasisi and 2. Field measurements
Ng [21] referred only to shores located at intermediate levels
of the building, and the authors concluded that the Simplified Field measurements were concentrated on shore load vari-
Method acceptably predicts maximum shore and slab loads during ations during the shore removal. Special attention was paid to
construction. On the other hand, Moragues et al. [22,23] measured shores supported on the ground because experimental works such
shore loads starting from the lowest level. They pointed out that as Moragues et al. [22] and Puente et al. [24] have shown that max-
the Simplified Method overestimates the maximum shore and slab imum shore loads take place at the bottom floor. Therefore, shore
loads in 77.5% and 36.4% respectively. loads were measured during the shore removal at the two lowest
A recent extensive experimental work was carried out by levels of the studied building.
Puente et al. [24]. One hundred and two shores, distributed be-
tween three floors, were instrumented with strain gages. The 2.1. Construction site description
authors compared theoretical results proposed by different meth-
ods with field measurements. It was concluded that the Refined The measurements were conducted during the construction
Method proposed by Liu et al. [10] is the most accurate theoret- of the Playa Gaztetape Building in Getaria, a city located on the
ical method, and that Duan and Chen’s [17] Improved Simplified
Basque coast in the north of Spain. The structure is a seven-story
Method is a quite accurate method (with deviations lower than 15%
flat slab type residential apartment building with four similar
between theoretical and experimental values) which does not re-
levels underground parking. Floor to ceiling height is 2.65 m for the
quire structural analysis software.
garage levels, and 2.90 m for the residential floors. The total area is
These experimental data consisted of measuring shore loads af-
1800 m2 per floor. Each parking floor was poured in 6 sections of
ter a construction step has been determined. Therefore, measure-
250 m2 plus a ramp zone. The residential floors were poured in 8
ments do not show the evolution of loads during an operation or
sections of similar area.
during the time period between two consecutive operations.
The thickness of the slab is 25 cm for all levels with a design
Several measurements have been conducted to study the
concrete strength of 25 MPa. The shores used were adjustable
load redistribution between two consecutive constructional steps.
Rosowsky et al. [25] and Fang et al. [26] obtained continuous steel shores with an allowable shore load of 18.5 kN. The planned
registers of shore loads during the curing process of concrete slabs. shoring scheme consisted of three levels of shores with no
In both research projects, it was concluded that load variations reshores, with a construction cycle of one floor per eleven days.
during this period are mainly related to the continuous increase The instrumented shores were arranged in two different sec-
in slab and beam stiffness. tions of the parking floors in both cases, these shores were placed in
Azkune et al. [27] also registered continuous measurements of a five-column module, where concrete columns are 40×40 cm. The
shore loads during the curing process of the slab. Nevertheless, plain views of the two measurement sections are shown in Figs. 1
with respect to the previous authors, they pointed out that the and 2. The arrangement of 34 instrumented shores is also included
redistribution of loads in this phase, at least in time periods in each figure. Shores on the far left and right sides of Row 1 were
no longer than a week, are mainly determined by ambient not instrumented since the supporting shoring system at these ar-
temperature variations. Therefore, shore loads fluctuate according eas consisted of more than one shore root.
to the temperature changes registered on-site. Finally, Azkune
et al. [27] proposed modified models which predict adequately the 2.2. Measurement system set up
load variations between consecutive steps.
With respect to measurements of dynamic loads, Rosowsky Thirty four shores with strain gages were placed on the selected
et al. [25] and Azkune and Puente [28] measured shore loads during floor. Four strain gages connected in a full Wheatstone bridge
 M. Azkune et al. / Engineering Structures 32 (2010) 3629–3638 3631

Fig. 1. Plain view of the instrumented shores arrangement in section B.

Fig. 2. Plain view of the instrumented shores arrangement in section E.

Fig. 3. Shore instrumentation and protection.

configuration were used: two active gages were placed on opposite and two passive gages were placed to compensate for temperature.
arms to eliminate the influence of the bending strain of the shore, A picture of some instrumented shores is shown in Fig. 3(a).
3632 M. Azkune et al. / Engineering Structures 32 (2010) 3629–3638

Fig. 4. Removal process of instrumented shores.

All gages were protected against mechanical shocks and humid-

ity. A picture of some shores protection is shown in Fig. 3(b).
A four-path connector was also attached to the bridge installed
at each shore. This connector allowed an easy monitoring of shore
loads by a portable data logger. A picture of some instrumented
shores in place at the construction site is shown in Fig. 4.
Before placement on the construction site, the ratio between
the transmitted load and microstrain for each instrumented shore
was measured in a laboratory. Since steel shores were used, it
was assumed that a linear relation existed between load and
microstrain. During the calibration procedure, 8 points of the load
versus microstrain curve were measured. The mean square error
between actual values and the linear approximation was lower
than 0.2% in all cases. Thermal compensation was also evaluated
for each shore, and shores with a compensation error higher than
30 N/°C were discarded.
Finally, the precision of the measurement system was evaluated
on site using load cells placed beneath four instrumented shores.
The difference between results provided by load cells and strain
gages was less than 8% in all cases.

3. On-site measurement results

Fig. 5. Three different measured shore removal procedures.

Experimental data was collected during the shore removal at

three different slabs: the two lowest floors of one section (B) and load in most critical cases. Maximum and average values of mea-
the lowest floor of another section (E). The formwork was removed sured overloads are included in Table 1.
at 3 days after pouring. Shores were measured during the stripping From Table 1 it can be seen that important overloads have
operations. Different shore removal procedures were carried out in appeared in all cases. The most critical cases are 1 and 3. In fact, the
each case. Studied cases are listed below: maximum loads during the construction of every building occurred
when shores were supported on the ground, before the removal.
• Case 1: Shore removal at level −4, section B. Consequently, attention must be paid to the most loaded shores
• Case 2: Shore removal at level −3, section B. of a floor, especially in the case of the lowest level. In these
• Case 3: Shore removal at level −4, section E. situations, a 3 kN overload may cause some permanent damage to
Loads were measured after shores were partially stripped. In the shores. Therefore, the shore stripping process must be properly
Case 1, loads were measured for the remaining shores after the performed in order to avoid excessive overloads at the critical
removal of the four central shores of each row. This situation has shores.
been illustrated in Fig. 4. The opposite procedure was carried out
in Case 2, measuring the loads of the three central shores of each 4. Theoretical model for the shore removal operation
row after the removal of the outer shores. Finally, loads on shores
located at the central row were measured after the removal of two It was observed in the previous section that some shores can be
other rows. The procedures carried out in each case are shown subjected to considerable overloads during the shore removal pro-
schematically in Fig. 5. cedure. Therefore, it would be advantageous to obtain an optimum
Considerable overloads were observed on some shores during stripping sequence to minimize these overloads. In this respect, the
its stripping process. In the three cases studied, maximum over- ACI Committee 347 [29] Standard suggests shore/reshore stripping
loads up to 3 kN were measured; approximately 15% of the total methods that will not damage the concrete structure.
 M. Azkune et al. / Engineering Structures 32 (2010) 3629–3638 3633

Fig. 6. Shore load variations during shore removal for Case 1.

Table 1
Measured maximum and average overloads during shore removal.
Case Row Maximum overload (kN) Average overload (kN) Maximum absolute load (kN) Average absolute load (kN)

1 1 2 3 2.94 2.40 2.00 1.87 1.48 1.05 21.78 21.86 20.59 16.45 16.09 12.42
2 1 2 3 3.14 2.08 2.13 3.04 1.43 1.45 9.50 3.68 5.37 9.09 2.89 3.67
3 2 2.66 1.86 18.67 14.83

Nevertheless, existing codes or standards do not propose adequate for the selection of an optimum shore removal sequence.
a theoretical method to model the shore removal operation. Even though it is not accurate in all cases, the model does,
An accurate model will provide necessary data concerning the nevertheless, predict adequately the trend of the redistribution of
overloads that shores are subjected to, which will help determine the loads transmitted by the removed shores.
the optimum shore removal procedure.
A 3D model is necessary for a correct modeling of the shore 4.1. Influence of slab stiffness
stripping operation. With a 2D model load variations caused by
the removal of shores located out of the modeled plane cannot In this section, the influence of slab stiffness value on load
be determined. Considering the work of Puente et al. [24], the 3D redistribution originated during shore removal is studied. The
refined method developed by Liu et al. [10] was selected for the elastic modulus Ec is the key parameter in concrete slab stiffness
comparison between theoretical and experimental results. and deflection. In this analysis, the elastic modulus was obtained
In Figs. 6–8, theoretical overloads are compared with on-site from the theoretical concrete strength development over time.
measurements. Each graph illustrates a row of shores. Each shore All slabs were poured with 25 MPa concrete (28-day design
stripping has been modeled adding two forces, one downwards for strength). The calculation of the elastic modulus was based on the
the upper slab and one upwards for the lower slab. Each one was expressions provided by the CEB-FIP 1990 [31] code.
of the same magnitude as the load in the removed shore. Different analyses were carried out in order to study the
In general, the refined method presented greater maximum influence of concrete strength over the shore stripping overloads.
shore loads than the measured results. Therefore, the refined Therefore, Case 3 was recalculated with concrete characteristic
method is considered as conservative. Depending on which case strengths of 20 and 30 MPa.
was analyzed the accuracy of the theoretical values has changed. It can be observed in Fig. 9 that the concrete strength does not
Nevertheless, the refined method correctly models the trends of have any significant influence on the shore overloads. Maximum
measured shore overloads as can be seen in Case 3, where one can shore load variations of 3.2% have been obtained with a 5 MPa
observe that in both the theoretical and experimental results the concrete characteristic variation. It can be concluded that stiffer
shore position has considerable influence. slabs lead to lower shore overloads, as the elastic modulus of
Therefore, the order in which the shore removal is carried concrete increases with its strength. Therefore, a stiffer slab will
out can affect considerably the magnitude of the overloads. In support a greater portion of the load that the removed shore was
consequence, it could be concluded that the refined method is transmitting, causing lower overloads on the remaining shores.
3634 M. Azkune et al. / Engineering Structures 32 (2010) 3629–3638

Fig. 7. Shore load variations during shore removal for Case 2.

Fig. 8. Shore load variations during shore removal for Case 3.

Fig. 9. Influence of concrete compressive strength variations on shore removal
Therefore, a variation of 5 MPa in concrete strength does not overloads.
significantly affect the theoretical overloads. Furthermore, it is as-
sumed that strength values calculated with the CEB-FIP 1990 [31] the concreting of each level and the shoring removal of the last
code, are close to actual values. level. A 25 MPa concrete characteristic strength was assumed
Another factor that affects the slab’s flexural stiffness is steel for this analysis. It can be observed that the results of the new
reinforcement. The elastic modulus of steel is much greater than analysis matched the on-site measurements better. Nevertheless,
that of concrete, so a steel reinforced slab is slightly stiffer than the introduction of steel reinforcement in the model produces
those previously considered. The analysis was repeated taking small variations, lower than 4%, on the theoretical overloads.
into account the presence of the steel reinforcement in the slab. Moreover, the model which neglects the steel reinforcement is also
An equivalent slab thickness was defined from the steel amount on the safe side.
and the ratio between the elastic modulus of steel and concrete. In conclusion, the introduction into the model of actual con-
Therefore, in this new analysis each slab has a different thickness crete strength values and steel reinforcement helps to improve
depending on its age and steel reinforcement. its accuracy. Variations registered in results, however, are not ex-
The new analysis was carried out for Case 3. The theoretical cessive. Consequently, acceptable values are obtained using the
overloads obtained are represented in Fig. 10. Table 2 shows the characteristic strength of the concrete and neglecting the steel re-
values of Ec and the slab thickness taken into account in this inforcement. Furthermore, these assumptions are both on the safe
study, considering as ‘‘age’’ the difference in the time between side.
 M. Azkune et al. / Engineering Structures 32 (2010) 3629–3638 3635

Fig. 10. Influence of steel reinforcement on shore removal overloads.

Table 2
Elastic modulus and equivalent slab thickness of each level in Case 3.
Fig. 12. Theoretical loads before removing the lowest level of shores.
Level Age (days) Ec (GPa) he (cm)

−4 29 32.079 25.55 41 MPa and its modulus of elasticity is 35 GPa. The elasticity mod-
−3 18 31.035 25.57
ulus of wooden shores is 7.75 GPa, and the compression strength of
−2 6 27.688 25.65
wood is 5.6 MPa. For the present study, it is assumed that there are
two levels of shores and one level of reshores, with a construction
rate of one floor per week.

5.2. Refined structure model

The Refined Method was selected for the comparison between

different shore removal procedures. In the previous section, it was
demonstrated that the Refined Method adequately predicts the
shore overloads originated during this operation. Different shore
stripping procedures were analyzed for the lowest level, the most
critical with respect to the falsework.
In the model, the elastic deformation of columns was neglected.
That is, vertices A, B, C and D in Fig. 11 of the studied slab were
considered as fixed supports. It is assumed that slab edges are free,
i.e. the effects of slab continuity are ignored.
Theoretically obtained shore load distribution is shown in
Fig. 12, before the beginning of shore removal on the lowest level.
Fig. 11. Plan view of the analyzed module (shore numbered). In this situation, the two poured slabs are shored to the foundation.
From this figure, it can be seen that, on the lowest level, the
5. Optimum shore removal sequence most loaded shores are located in the central zone of the module.
Therefore, special attention must be paid to the removal of these
From on-site measurements it has been observed that shore shores.
overloads are highly influenced by the shore’s position in the
module under consideration. Therefore, greater or lower shore 5.3. Shore removal procedures
overloads will result depending on the order in which shores are
Different, commonly employed, on-site shore stripping se-
removed.
quences have been applied throughout the studied structure. Over-
The objective, in this section, is to establish general criteria
loads on all shores were calculated in each case.
for an optimum shore removal sequence. The optimum procedure
should lead to greater overloads at less loaded shores, with no • Procedure 1: a typical zig-zag shore removal. It consists of
significant overloads on the shores that bear a greater load. That is, removing the shores in one direction from one of the shore rows
the objective in a stripping process in which the maximum shore located at one slab edge, and then continuing with the adjacent
row in the opposite direction and so on until the opposite row is
load is not increased.
reached. This operation is carried out in two possible directions:
procedure1A: stripping according to larger shore distance rows
5.1. Description of the studied structure
(X direction in Fig. 11) as shown in Fig. 13(a), and procedure 1B:
stripping according to shorter shore distance rows (Y direction
Different shore stripping procedures have been applied over the in Fig. 10) as shown in Fig. 13(b).
typical structure studied by Liu et al. [10]. The plain view of the • Procedure 2: shore removal in spiral. The shore located at the
analyzed structure module is shown in Fig. 11. Main dimensions of module center is removed first, and the operation is completed
the four columns module are also included. removing shores in a spiral. The shore removal order is shown
All floors are similar. Story height is 2.80 m, with 18 cm slab in Fig. 13(c).
thickness. The cross-section of the shores is 50 × 100 mm. Regard- • Procedure 3: stripping alternating rows. The most loaded row
ing material parameters: 28-day cylinder strength of concrete is of shores, the central row in direction X , is removed first.
3636 M. Azkune et al. / Engineering Structures 32 (2010) 3629–3638

Fig. 13. Shore removal order by the different procedures.

Next, adjacent rows are removed, and so on until rows located stripping procedures are included in Table 3. In this table shore
between columns are reached. Shores located in the same row represents the shore numbered according with Fig. 11.
are removed in order, starting from a slab edge to the opposite It can be concluded that different shore overloads were ob-
one. The shore removal sequence is shown in Fig. 13(d). tained depending on the stripping procedure. The most critical sit-
• Procedure 4: similar to Procedure 3, but also alternating the uation was observed in Procedure 1A, since an absolute load of
removal of the shores located within a row. First, the central 15.35 kN was reached at shore 20, after a 5.28 kN overload. This is
shore of the row should be removed, then both adjacent shores, perhaps one of the most common on-site procedures, where work-
and so on until the removal of the shores located at both edges ers start removing shores from a slab edge and continue by rows
of the row. ACI Committee 347 [29] also recommends beginning until the opposite slab edge is reached.
with the removal of shores placed in the middle of a bay, since Applying the same criterion for the other direction, maximum
a more adequate slab load distribution is obtained. The shore absolute shore load is reduced by 9% (13.97 kN), with a 3.90 kN
removal order is shown in Fig. 13(e). overload (26% reduction). Therefore, overloads are considerably
• Procedure 5: removal of the most loaded shore. Apparently, reduced by taking the precaution of stripping by rows in the
this is the optimum procedure. It consists of removing the most direction where the distance between columns is less.
loaded shore at every moment. This procedure could not be Spiral shore removal did not reduce maximum overloads. Late
used on-site, since the worker cannot know which is the most removal of shores located at the central row edges (shores 2
loaded shore at a given time. This analysis, however, will allow and 20) led to considerable overloads on both shores. Maximum
the evaluation of the differences with other more common overload originated on shore 2 was 5.16 kN, with a maximum
shore removal procedures described previously. Shore removal absolute shore load of 15.23 kN.
order is shown in Fig. 13(f). Procedure 3 involved a significant improvement with respect
to previous procedures. Since the most loaded row was removed
A new analysis was carried out for each procedure when a first, maximum overloads were generated on the less loaded
shore was removed, calculating the overloads originated on the shores. Consequently, maximum absolute shore loads decreased.
remaining shores at each stage. Maximum shore overload was 4.05 kN, causing a maximum
absolute shore load of 12.97 kN at shore 1. Therefore, with respect
5.4. Analysis results to the initial case 1A, maximum overload value was reduced by 23%
and maximum absolute load by 16%.
The maximum overloads (max ol) and the maximum absolute Procedure 4 did not lead to significant improvements when
loads (max absl) that were transmitted by shores during the compared to Procedure 3. Albeit the overloads on shores placed
 M. Azkune et al. / Engineering Structures 32 (2010) 3629–3638 3637

Table 3
Maximum shore overloads and maximum absolute loads during shore removal (kN).
Shore Before removal Procedure 1A Proc 1B Proc 2 Proc 3 Proc 4 Proc 5
Max ol Max absl Max ol Max absl Max ol Max absl Max ol Max absl Max ol Max absl Max ol Max absl

1 8.92 – 8.92 1.26 10.18 4.10 13.01 4.05 12.97 3.09 12.01 3.14 12.05
2 10.07 0.49 10.56 1.43 11.50 5.16 15.23 1.92 11.99 1.80 11.87 1.92 11.99
3 8.92 0.76 9.68 4.10 13.01 2.95 11.87 1.71 10.62 3.98 12.90 4.05 12.97
4 7.18 0.48 7.66 – 7.18 1.03 8.21 1.02 8.20 1.46 8.64 1.46 8.64
5 9.04 2.21 11.26 1.04 10.08 1.32 10.37 3.59 12.63 2.61 11.65 2.74 11.79
6 10.23 2.63 12.86 2.82 13.05 2.04 12.27 1.27 11.51 1.04 11.27 1.27 11.51
7 9.04 1.50 10.54 3.63 12.67 1.17 10.21 2.66 11.70 3.43 12.48 3.59 12.63
8 7.18 0.19 7.37 1.03 8.21 1.11 8.29 1.47 8.65 1.46 8.64 1.46 8.64
9 7.54 0.62 8.15 0.24 7.77 1.57 9.10 1.57 9.10 1.29 8.83 1.29 8.83
10 9.17 2.08 11.24 1.18 10.34 1.98 11.14 3.67 12.83 1.85 11.02 1.85 11.02
11 10.34 3.35 13.69 3.21 13.55 – 10.34 1.28 11.62 – 10.34 – 10.34
12 9.17 2.61 11.77 3.70 12.87 1.23 10.40 2.82 11.99 2.60 11.76 2.60 11.76
13 7.54 1.00 8.54 1.57 9.10 1.30 8.84 1.57 9.10 1.29 8.83 1.29 8.83
14 7.18 1.11 8.29 0.36 7.55 1.47 8.65 1.46 8.64 1.32 8.51 1.32 8.51
15 9.04 2.79 11.84 1.01 10.05 2.08 11.12 3.48 12.52 2.35 11.40 2.35 11.40
16 10.23 3.52 13.75 3.19 13.42 0.69 10.92 1.06 11.29 0.69 10.92 0.69 10.92
17 9.04 2.12 11.16 3.52 12.56 0.80 9.85 2.74 11.79 3.16 12.20 3.16 12.20
18 7.18 0.79 7.98 1.47 8.65 0.90 8.09 1.03 8.21 1.32 8.51 1.32 8.51
19 8.92 2.98 11.90 0.20 9.11 2.44 11.36 2.48 11.40 3.14 12.05 2.95 11.87
20 10.07 5.28 15.35 3.90 13.97 4.58 14.64 – 10.06 1.92 11.99 1.59 11.65
21 8.92 4.38 13.30 2.52 11.44 3.77 12.69 3.14 12.05 4.05 12.97 3.83 12.75

in the center of the module were reduced, maximum loads were 3. Overloads generated during shore removal are influenced by
similar to those calculated in Procedure 3. Therefore, the removal the stiffness of slabs and shores. Hence, use of actual con-
of shores alternating the position within the row did not reduce crete strength and modeling of steel reinforcement produces
maximum loads when compared with Procedure 3. more accurate theoretical results. Nevertheless, relatively low
Finally, maximum loads did not decrease when Procedure 5 was strength variations and steel reinforcement consideration do
checked. In fact, the procedure of removing the most loaded shore not lead to important variations of theoretical values (less than
is almost the same as Procedure 4. 4%). Therefore, modeling the slab with its project characteristic
Therefore, from a practical point of view, Procedure 3 gets strength and without its steel reinforcement leads to less accu-
optimum results. Hence, the shore removal process should be rate but acceptable results.
carried out by rows: the most loaded row of shores should be
4. The overload supported by each shore is influenced by its rela-
removed first, then the adjacent rows and finally the less loaded
tive position within the module. Therefore, different overloads
rows. The use of this procedure will be especially important when
are produced depending on the shores which are removed. Con-
maximum shore loads are present. During the construction of
any multistory concrete building, removal of the lowest level of sequently, the stripping order affects the maximum shore loads
shores, which are supported on ground is the most critical, since considerably.
maximum shore loads occur at this stage. For a typical 4 column 5. The best stripping procedure found consists of removing shores
module supported by shores with a similar tributary area, the by rows. First, the most loaded row must be removed, then the
most loaded row is located in the center according to the direction adjacent rows and so on until the removal of less loaded rows.
corresponding with the longest bay (X direction in the example For a typical symmetric 4 column module, the most loaded row,
studied here). in the most critical situation, is the central one in the longest bay
direction.
6. Conclusions 6. Although an alternating removal of the shores located within a
row does not produce a reduction on shore overloads, it could
The overloads originating during the shore removal process be beneficial for the concrete slab. By stripping the shores in the
can cause shore failure in some cases. Damage to the falsework middle first, the slab will be loaded as designed. First, the central
can produce important economic losses due to its elevated cost. shore of the row should be removed, then both adjacent shores
Therefore, on-site measurements were conducted during the and so on until the removal of shores located at both edges of
shore stripping at different levels. Based on these experimental the row.
measurements, the adequacy of the Refined Method for the shore
removal operation modeling was evaluated. Finally, the shore
overloads originated in different shore removal procedures were Acknowledgements
studied. From this study the following conclusions can be stated:
This research was sponsored by the Basque Government
1. During the shore removal process, shores are subjected to
(Departamento de Educación, Universidades e Investigación and
considerable load increments. Overloads of up to 3 kN (15% of
the total load, approximately) have been reached during on- Departamento de Industria, Comercio y Turismo, Project number
site measurements, which represent approximately 10% of the UE2005-1) and the Spanish Government (Ministerio de Fomento,
shore capacity. Programa Nacional de Construcción, Project Ref. 80003/A04). Ulma
2. The Refined Method models adequately the shore removal Construcción also sponsored this study and supplied part of the
procedure. The Refined 3D Method predicts correctly trends material needed during the measurements. The authors also wish
within the redistribution of loads. The theoretical model to acknowledge the collaboration of Construcciones Imaz in the
proposes overload values greater than real ones. Therefore, it collection of experimental data. The opinions expressed in this
is a model that proposes values which lean on the safe side, as work are those of the writers and do not necessarily reflect the
shown in Figs. 6–8. points of view of the sponsors.
3638 M. Azkune et al. / Engineering Structures 32 (2010) 3629–3638

References [17] Duan MZ, Chen WF. Improved simplified method for slab and shore load
analysis during construction. Project report CE-STR-95-24. West Lafayette
[1] Lew HS, Carino NJ, Fattal SG. Cause of the condominium collapse in Cocoa (IN): Purdue University; 1995.
Beach, Florida. Concr Int 1982;4(8):64–73. [18] Fang DP, Zhu HY, Geng CD, Liu XL. Floor load distribution in reinforced concrete
[2] Carino NJ, Leyendecker EV, Fattal SG. A review of the Skyline Plaza collapse. buildings during construction. ACI Struct J 2001;98(2):149–56.
Concr Int 1983;5(7):35–42. [19] Miranda De Almeida JF, Silva MR, Ramalho MA. A new procedure for the
[3] Price WH. Factors influencing concrete strength. ACI J Proc 1951;47(2): analysis of construction loads in multistory reinforced concrete structures.
417–32. Struct Des Tall Spec Build 2003;12:293–315.
[4] Klieger P. Effect of mixing and curing temperature on concrete strength. ACI J [20] Agarwall RK, Gardner NJ. Form and shore requirements for multistory flat slab
Proc 1958;54(6):1063–81. type buildings. ACI J Proc 1974;71(11):559–69.
[5] Gardner NJ, Poon SM. Time and temperature effects on tensile, bond, and [21] Lasisi MY, Ng SF. Construction loads imposed on high-rise floor slabs. Concr
compressive strengths. ACI J Proc 1976;73(7):405–9. Int: Des Construct 1979;1(2):24–9.
[6] Carino NJ, Hamlin MJ, Snell LM. Properties of concrete at early ages. Concr Int [22] Moragues JJ, Catalá J, Salort V, Sirvent PL. Transmisión de Cargas entre Forjados,
1989;11:51–4. durante el Proceso Constructivo: Medidas Realizadas en Obra. Hormigón y
[7] Nielsen K. Loads on reinforced concrete floor slabs and their deformations Acero 1991;179:37–47.
during construction. Bulletin no.15, final report. Stockholm: Swedish Cement [23] Moragues JJ, Catalá J, Pellicer E. An analysis of concrete framed structures
and Concrete Research Institute. Royal Institute of Technology; 1952. during the construction process. Concr Int: Des Construct 1996;18(11):44–8.
[8] Grundy P, Kabaila A. Construction loads on slabs with shored formwork in [24] Puente I, Azkune M, Insausti A. Shore–slab interaction in multistory reinforced
multistory buildings. ACI J Proc 1963;60(12):1729–38. concrete buildings during construction: an experimental approach. Eng Struct
[9] ACI Committee 347. Guide for shoring/reshoring of concrete multistory 2007;29(5):731–41.
buildings (ACI 347.2R-05). Farmington Hills (MI): American Concrete Institute; [25] Rosowsky DV, Philbrick Jr TW, Huston DR. Observations from shore load
2005. p. 18. measurements during concrete construction. J Perform Construct Facil ASCE
[10] Liu XL, Chen WF, Bowman MD. Construction load analysis for concrete 1997;11(1):18–23.
structures. J Struct Eng ASCE 1985;111(5):1019–36. [26] Fang DP, Zhu HY, Geng CD, Liu XL. On-site measurements of load distribution
[11] Liu XL, Chen WF, Bowman MD. Construction loads on supporting floors. Concr in reinforced concrete buildings during construction. ACI Struct J 2001;98(2):
Int: Des Construct 1985;7(12):21–6. 157–63.
[12] Gardner NJ. Shoring, reshoring, and safety. Concr Int: Des Construct 1985;7(4): [27] Azkune M, Puente I, Insausti A. Effect of ambient temperature on the
28–34. redistribution of loads during construction of multi-storey concrete structures.
[13] Stivaros PC, Halvorsen GT. Shoring/reshoring operations for multistory Eng Struct 2007;29(6):933–41.
buildings. ACI Struct J 1990;87(5):589–96. [28] Azkune M, Puente I. Evolución de la distribución de cargas entre puntales y
[14] Chen WF, Mossallam K. Concrete buildings analysis for safe construction. Boca forjados durante la construcción. Hormigón y Acero 2007;245:79–91.
Raton: CRC Press; 1991. [29] ACI Committee 347. Guide to formwork for concrete. 1988.
[15] Mossallam KH, Chen WF. Determining shoring loads for reinforced concrete [30] UNE-EN 12812. Cimbras, Requisitos de comportamiento y diseño general.
construction. ACI Struct J 1991;88(3):340–50. 2008.
[16] El-Shahhat M, Chen WF. Improved analysis of shore-slab interaction. ACI [31] CEB-FIP Model Code. Comité Euro-International du Béton—Federation Inter-
Struct J 1992;89(5):528–37. nationale de la Precontrainte, Lausanne. 1990.