Anda di halaman 1dari 15

Engineering Structures 40 (2012) 383397

Contents lists available at SciVerse ScienceDirect

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

Post-punching behaviour of at slabs strengthened with a new technique using post-tensioning


Duarte M.V. Faria , Vlter J.G. Lcio, A. Pinho Ramos
Department of Civil Engineering, Faculdade de Cincias e Tecnologia, Universidade NOVA de Lisboa, 2829-516 Caparica, Portugal

a r t i c l e

i n f o

a b s t r a c t
This work presents an experimental study concerning the post-punching behaviour of at slabs strengthened with a new technique based on post-tensioning with anchorages by bonding using an epoxy adhesive. This strengthening technique proved efcient with respect to ultimate and serviceability states. Five slab specimens were tested in the post-punching range and it was found that the post-punching resistance was on average 78% of the punching resistance. This paper reports the development of strand forces and slab displacements from the beginning of the tests, including the bond stresses developed at several stages of the loading process. It was observed that top reinforcement bars were capable of transmitting post-punching loads to the prestressing strands. Taking this into account and based on the load bath envisaged from the column to the slab, expressions for the vertical load capacities corresponding to the parts of the load path are presented and compared with the experimental results, showing their ability to predict both ultimate loads and modes of failure. Compared with other strengthening techniques, the one proposed here not only upgrades ultimate and serviceability behaviour but also adds post-punching resistance, which is a great advantage in the event of progressive collapse, since it may avoid the collapse of an entire structure, thus reducing the risk of material and human losses. 2012 Elsevier Ltd. All rights reserved.

Article history: Received 6 November 2011 Revised 9 February 2012 Accepted 1 March 2012

Keywords: Post-punching Post-tensioning Strengthening Flat slabs

1. Introduction Behaviour after punching, also known as post-punching behaviour, must not be forgotten when a at slab has to be strengthened. After a punching failure in a slab without or with small quantities of bottom reinforcement crossing the column, a sudden loss of support can occur, causing a load increase and eccentricity in the adjacent slab supports. This may lead to a progressive collapse of the entire building as slabs may fall one over another. Examples of entire buildings are those of the Harbour Bay Condominium (USA, 1981) [1], Sampoong Department Store, Seoul (South Korea, 1995) [2] and a parking garage in Gretzenbach (Switzerland, 2004) [3]. The Harbour Bay Condominium building collapse was caused by a punching shear failure that triggered a progressive collapse, during its construction. The collapse was associated with design and construction errors, namely, inadequate thickness of the slab (design error), inadequate effective depth of the reinforcing steel and improper formwork removal (construction error) [1]. The Sampoong Department Store collapse was due to punching, associated with construction faults (lower concrete compressive strength,
Corresponding author. Address: Universidade NOVA de Lisboa, Faculdade de Cincias e Tecnologia, Departamento de Engenharia Civil, Campus da Caparica, 2829-516 Caparica, Portugal. Tel.: +351 962821685; fax: +351 212 948 398. E-mail addresses: duamvf@gmail.com (D.M.V. Faria), vlucio@fct.unl.pt (V.J.G. Lcio), ampr@fct.unl.pt (A. Pinho Ramos).
0141-0296/$ - see front matter 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.engstruct.2012.03.014

reduced effective depth in the negative moment areas, reduction of the cross-section of the columns supporting the fth oor and roof), poor construction quality control, and change of use of the fth oor with increased loading [2]. Regarding the collapse of a parking garage in Gretzenbach, it was triggered by a re, when a punching failure occurred, starting a progressive collapse. It was concluded that the applied load was considerably larger than the design load and also that the slab effective depth over the column was reduced due to the misplacing of the reinforcing bars, resulting in a decreased punching resistance [3]. Examples of partial collapses are the ones that happened in Pipers Row Car Park, Wolverhampton (United Kingdom, 1997) [4] and in an apartment block, Sesimbra (Portugal, 2003, Fig. 1) [5]. In both cases, collapse was initiated by punching failure. In Pipers Row car park the partial collapse was due mainly to poor maintenance, in the areas above/ near the columns where punching failure occurred, spreading horizontally in a limited area of 15 15 m of one storey [4]. The partial collapse of an apartment block in Sesimbra was due to design and construction errors, namely the thickness of the slabs (design error), the non assurance of a correct alignment between column sections along different levels and the use of materials that were poor or different from those specied (construction errors). It is important to mention that in Fig. 1 it is possible to observe that a big part of the building is shored since its total collapse was imminent, but was avoided due to the emergency measures that were taken immediately after the collapse of part of the building.

384

D.M.V. Faria et al. / Engineering Structures 40 (2012) 383397

Nomenclature

a
b

cc cs
h

ql w

si,nal si,initial si,PP smax


Ap As Asp

d dc dint dres

fccm fcm fpk

inclination adjacent to the column, of rebars crossing the column, when the maximum post-punching load is reached is the inclination of a strand in the length between its exit from a deviator and its entry into the concrete partial factor for concrete partial factor for steel angle of inclination of the portion of the strand embedded in the slab reinforcement ratio inclination of top bars crossing the column at their start of their anchorages when the maximum post-punching load is reached bond stress corresponding to Pi,nal bond stress corresponding to Pi,initial bond stress corresponding to Pi,PP average value for maximum bond stress cross section area of a prestress strand cross section area of a bottom reinforcement that passes inside the column main reinforcement is the cross section of a rebar that crosses the top of the column and is crossed by a deviator, or is positioned between the deviator in a unidirectional arrangement average effective depth of top rebars is the strand nominal diameter is the concrete depth above the strand at the point where it enters in the concrete is the distance between the centroid of the exural reinforcement ratio and the centroid of the integrity reinforcement mean value of concrete cube compressive strength mean value of concrete cylinder compressive strength characteristic tensile strength of prestressing steel

ft fy fyd h l2 lav ldev ln Pi,nal Pi,initial Pi,PP Pstrand qd VEd,PP Vexp VPP VPP,exp VPP,rb VPP,sp

VPP,st

ultimate strength of reinforcement yield strength of reinforcement design yield strength of the reinforcement slab depth is the centre-to-centre span in the direction of the catenary. is the available bonded length stands for the length of the cantilever portion of the deviator is the clear span in the direction being considered, measured face-to-face of supports prestress force at punching failure per strand prestress force after prestress transmission per strand prestress force at post-punching maximum load per strand is the strand force factored uniformly distributed load but not twice the slab service dead load design value of the applied punching load in the postpunching stage experimental punching load post-punching maximum load experimental post-punching maximum load is the post-punching maximum load provided by top reinforcing bars is the post-punching maximum load provided by the extensions of bars beyond deviators for post-punching maximum load is the post-punching maximum load provided by the unbonded length of prestress strand to post-punching resistance

strengthening system consists of introducing steel strands in previously drilled holes in the concrete slabs that after tensioning are bonded to the concrete using an epoxy adhesive agent. The system has already been previously described in detail [6,7] (see Fig. 2). 2. Background considerations 2.1. Post-punching research Not many studies have been developed regarding post-punching behaviour, and so literature on post-punching is limited. Regan [8] tested slabs with 100 mm thickness with an effective depth of 79 mm, a column width of 200 mm and a reinforcement ratio of 0.8%. Due to the presence of three 8 mm integrity reinforcing bars in each direction, the post-punching maximum load reached about 71% of the punching load, and so Regan [8] concluded that integrity reinforcement is important to the post-punching behaviour. He observed that slabs without integrity bars reached a post-punching strength of 25% of punching strength. Therefore, the remaining 75% should be transferred through the integrity reinforcement and Regan [8], based on limited experimental results, proposed Rasmussens [9] expression to estimate it, Eq. (1):

Fig. 1. Building collapse in Sesimbra, Portugal (2003).

The performance and the post-punching behaviour of slab-column connections may be improved using special detailing with ordinary or prestressing steel. Examples of this procedure are the use of bottom bars, passing through the column and well anchored into the surrounding slab, or the use of prestress tendons that pass through the columns, positioned above the column but dropping down in the slab as it approaches midspan. This work main purpose is to show that the proposed strengthening technique is able to increase post-punching resistance. This

V PP 1:3

/2 s

q fy fcm

where VPP is the post-punching maximum load, /s is the bar diameter, fy is the yield strength of reinforcement and fcm is the mean value of concrete cylinder compressive strength.

D.M.V. Faria et al. / Engineering Structures 40 (2012) 383397


DEVIATOR SLAB FINISHING

385

PRESTRESS STEEL STRAND REINFORCED CONCRETE SLAB

COLUMN

Fig. 2. System application view.

Melo and Regan [10], performed an experimental investigation regarding the improvement of post-punching resistance of slabcolumn connections by the provision of bottom bars passing through the column and anchored in the slab. The experimental work showed that this solution can be highly effective in increasing the post-punching resistance of a slab-column connection. It was shown [10] that the nal resistance provided by the bottom bars appears to be governed either by destruction of the concrete in the zone where they are anchored in the slab (Eq. (2)) or by fracture of the bars themselves (Eq. (3)). In addition it was reported that the angle of inclination of the reinforcing bars at failure in the column vicinity varied from 22 to 26 [10].

V PP 6 0:33 V PP 6 0:5

p X d2 int fcm p 2 As f y

2 3

supports and l2 is the centre-to-centre span in the direction of the catenary. Tests by Ritz et al. [13] and Pralong et al. [14] have shown that high post-punching resistance is also possible in prestressed slabs where inclined tendons cross the columns. Ramos and Lcio [15] showed that the use of inclined prestressed strands passing through the columns can be highly effective in increasing the post-punching resistance of a at slab-column connection and also its ductility. In fact, in the specimens where the tendons cross the column, the maximum post-punching load was on average 25% higher than the experimental punching load. But in the specimens where not all the tendons crossed the column, the maximum postpunching load was on average around 40% lower than the original experimental punching load. The CEB-FIP Model Code 1990 (MC90) [16] recommends that Eq. (7) should be veried, to reduce the risk of a progressive collapse in the event of a local failure at a slab-column connection:

where As is the cross section area of a bottom rebar that passes inside the column main reinforcement and dint is the concrete depth above the bar at the point where it enters in the concrete. Eq. (2) Vpp per bar is reduced when bars are closely spaced [10]. Georgopoulos [11], based on an experimental analysis, considering the dowel action as the determinant mechanism, used Rasmussens Eq. (1) and proposed Eq. (4) to estimate the angle of inclination a of the reinforcing bars at failure and apply it in Eq. (5):

V Ed;PP 6

As fyd

s fcm sin a 1:5 fy V PP 6 sin a X As fy

where VEd,PP is the design value of the applied punching force, fyd is the design yield strength of the reinforcement. Although the MC90 [16] does not deal with prestressed slabs, it mentions that the subject should be dealt with in a special publication. Subsequent recommendations published by FIP [17] extend this rule to prestressed at slabs. To decrease the risk of a progressive collapse in the event of a local punching failure in a prestressed at slab, FIP [17] calls for reinforcement complying with Eq. (8):

V Ed;PP 6

 X fpk As fyd Ap

cs

The use of Rasmussens Eq. (1) is questionable, since it relates to the capacity of a dowel bar projecting from the vertical face of a large concrete block. The type of failure it treats is one in which the concrete under the dowel plasties over a length from the surface and the bar fails in bending near the end of the plastied length. Also, the proposed Eq. (1) doesnt take into account axial forces in the bar and the fact that the dowel is loaded at a distance from the concrete face, and so it seems physically unrealistic when compared with the performed tests. Mitchell and Cook [12] studied the possibility of the development of membrane action in various types of concrete slabs after a local failure. They concluded that tensile membrane development by well anchored reinforcing bars is capable of suspending damaged portions of the structure from the columns. They proposed Eq. (6) to estimate the integrity reinforcement.

where Ap is the cross section area of a prestressing strand that pass inside the columns main reinforcement, fpk the characteristic tensile strength of prestressing steel and cs the partial factor for prestressing steel. Neither MC90 [16] nor the FIP recommendations [17] give expressions for calculating the post-punching resistance of a slab-column connection, but their expressions are intended to reduce the risk of progressive collapse. The MC2010 [18] to be published in 2012 proposes Eq. (9) to estimate the contribution of a single bar of integrity reinforcement to post-punching strength.

V Ed;PP 6

As fyd ft =fy k sin ault 6

p X 0:5 fck

cc

dres bint

As

2 q d ln l2 /d fy

where qd is the factored uniformly distributed load but not less than twice the slab dead load, /d is a shear reduction factor, ln is the clear span in the direction being considered, measured face-to-face of

where ault depends on the type and ductility class of the reinforcing bars, k stands for the characteristic value, ft is the ultimate strength of reinforcement, fy is the yield strength of reinforcement, dres is the distance between the centroid of the exural reinforcement ratio and the centroid of the integrity reinforcement and bint is the control perimeter activated by the integrity reinforcement after punching [18]. For example, for ductility class C, ault = 25 and 1.15 6 (ft/fy)k 6 1.35. It is interesting to observe that Eq. (9) has many resemblances with Eqs. (2) and (3), but take into account more factors such as the ductility class of the reinforcing bars.

386

D.M.V. Faria et al. / Engineering Structures 40 (2012) 383397 Table 1 Geometric and material properties.

At this point, is important to keep in mind that while both, the MC90 [16] and the FIP recommendations [17], use the design value of the vertical reaction at the slab-column connection under an ultimate load combination, MC2010 [18] uses an accidental load combination. As the vertical reaction at the slab-column connection under an ultimate load combination is approximately twice that when using an accidental load combination, it is possible to conclude that both expressions result almost in the same amount of steel reinforcing bars. 3. Experimental work 3.1. Specimens and tests The experimental work consisted of testing ve reduced-scale at slabs that had previously been tested to failure by punching. These slabs measured 2300 2300 mm2: slabs DF2 and DF3 were 100 mm thick, and slabs DF5 to DF7 were 120 mm thick. Slab DF7 was strengthened bidirectionally while the others were strengthened unidirectionally. The punching load was applied by a hydraulic jack positioned under the centre of the slab, via a 200 200 50 mm steel plate. Eight points on the top of the slab were connected to the strong oor of the laboratory using prestress

Specimen depth. Average effective depth. c Average reinforcement ratio.


b

strands and spreader beams. Fig. 3 gives a plan and section of the system. These specimens simulated the area around an internal extending to the zero moment lines. The bottom reinforcement of the slab consisted of 6 mm rebars every 200 mm, in both orthogonal directions. In slabs DF2 to DF3 the top reinforcement consisted of 10 mm rebars every 60 mm and in slabs DF5 to DF7 it was 10

2300

200

2000

1200

Fig. 3. Plan and section AA (dimensions in mm).

2300

700

200

D.M.V. Faria et al. / Engineering Structures 40 (2012) 383397

387

(a)

Anchorage Load Cell Steel Beam

Deviator Top End Strand

11
Load Cell Bottom End

(b)

(c)

DF2/DF3

DF5 to DF7

DF2/DF3 and DF5/DF6

DF7

Fig. 4. Details of test specimens (a) detailed prestress prole, (b) deviator geometry and (c) top steel reinforcing bars above the column (dimensions in mm).

mm rebars every 75 mm, in both orthogonal directions. The reinforcing bars were straight and reached to the edges of the slabs. The slab thickness (h), average effective depth (d) and top exural reinforcement ratios (ql) are presented in Table 1. In the strengthened slabs, except DF7, the prestress was in the direction of the rebars at smaller effective depth. Strands 3 and 4 in Fig. 3 are only used in slab DF7. Regarding the prestress prole, the strands were positioned 50 mm from the column faces and a radius

of 2500 mm was adopted for the deviators and, therefore, for the strands. The slope of the strands in the embedded portion was about 1/5 (11.3), as shown in Fig. 4, where a longitudinal section of the prestress strand is shown. In Fig. 4 also shows the top reinforcement and the positioning of the deviators relative to the column. In this gure ldev stands for the length of the cantilever portion of the deviator, which was designed to be 175 mm, but due to construction errors was around 170 mm for slabs DF2 and

388

D.M.V. Faria et al. / Engineering Structures 40 (2012) 383397

6 8 1 2 3 9 7 4 5
400

400

posed lengths of the strands between the deviators and the slabs. During the prestress operation it was possible to correlate the force measured in the load cells with the strains measured in the strain gauges. After the release of the provisional anchorages and transfer of the prestress forces to the concrete, the forces in the strands were computed using the relation obtained in the rst stage. 3.3. Materials properties Compression tests of 150 mm (fccm) cubes were carried out on the same day as the test of the corresponding slab. The results are given in Table 1. Along with cylinder strengths (fcm) taken as 0.80 of cube strengths. The yield strengths (fy) and ultimate strengths (ft) of the bar reinforcement are also given in Table 1. The prestress strands used were 15.2 mm diameter with a 139.5 mm2 cross section, having a maximum force of 273 kN, a 0.1% proof force of 246 kN and a modulus of elasticity of 197.4 GPa. The epoxy adhesive was also tested for its exural and compressive resistance. Briey the results from exural and compressive testing showed that the average tensile strength was 49.1 MPa and the average yield compressive strength was 108.8 MPa. 4. Discussion of the results 4.1. Load/evolution of strand forces As explained, each strand had a load cell placed at one end and a pair of strain gauges glued to it, and so it was possible to compute the load in each strand during tests. It was found that the force/ strain measurement relation was approximately linear within the test range, making it easy to correlate the strain measurement with the strand load during the tests [7]. This correlation was used when determining the development of strand forces in the post-punching testing stage. An example of a force/strain measurement relation may be seen in Fig. 6. Each strand has its own relation. Fig. 7 shows the strand loads versus slab loads from the start of the transmission operation. After this operation, the slabs were loaded until punching failure and the evolutions of strand forces were approximately linear. After the punching failures there sudden losses of strand forces occurred, followed by increases when the loads were increased again in the post-punching stage. From Fig. 7 we can see that each pair of strands exhibited a similar behaviour, both in relation to the load/strand force development and for the maximum attained strand force. In some cases, near the maximum post-punching load, one of the strands started to lose force while its pair gained force at about the same rate, which may be due to the differences in the damage sustained at either side of the load (specimens DF5 and DF6 Fig. 7), causing a rotation of the deviators, which was more pronounced to one side of the column (Fig. 8) due to the load eccentricity. This reduced the strand eccentricity and consequently decreased the strand force on that side. The forces decreased equally in both strands of specimen DF3, meaning that in this case the load was applied centrally and the force reduction is due to deviator rotation. The values obtained for the strands 3 and 4 of specimen DF7 are due to problems that occurred during the injection procedure. These problems arose because the hole in the bottom of the slab was not correctly sealed and so, under the pressure of the injection, cracks formed in the sealing material that allowed the bonding agent to leak, which prevented the adequate involvement of the strands and the bonding agent. It can also be noted that the strand forces mostly exceeded the value they had at the time of punching failure, meaning that the contribution of bonded prestressing steel strands is signicant at the post-punching stage.

400

400

400

400

Fig. 5. Displacement transducers.

Model DF2
100

Strands Force (kN)

80 60 40 20 0 0 500 1000

Strand 2 y = 0,037x R2 = 0,9982

Strand 1 y = 0,0316x R2 = 0,9984

Strand 1 Linear (Strand 1) 1500 2000

Strand 2 Linear (Strand 2) 2500 3000

Strain Measurement (x10-3)


Fig. 6. Force/strand strain measurements for specimen DF2.

DF3, 150 mm for slabs DF5 and DF6 and 140 mm for slab DF7. Also, the exposed length of the strands is greater than the one presented in Fig. 4 due to the fact that an initial portion of the drilled hole was damaged during the drilling process [7]. This is taken into account when computing post-punching resistance in Section 5.2. The loading stages were: before the prestress operation and while the bonding agent was curing, the vertical load was kept constant at about 40% of the punching failure load of non-strengthened similar slabs. Only after the release of the provisional equipment and full transfer of the prestress to the concrete, was the vertical load increased. When punching failure occurred a sudden force drop was registered. The specimens were then loaded again to study their post-punching behaviour, until no further load increment could be sustained.

3.2. Monitoring Four load cells were used (Fig. 3) to measure the vertical load applied to the specimens, one in each spreader beam. Up to nine displacement transducers were used to measure the vertical displacement of the top surface of the slabs during the tests (Fig. 5). Displacement transducers 8 and 9 were only used for specimen DF7. Displacement transducers were supported by steel beams outside the test models. The forces installed in each prestress strand were measured with a load cell during the prestressing operation, while using the temporary strut [7] and with one pair of strain gauges for each strand during the loading of the slab, that were glued on the ex-

400

400

D.M.V. Faria et al. / Engineering Structures 40 (2012) 383397

389

Model DF2
300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0 0
punching

Model DF6
300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0 0
punching Strand 2 Strand 1 post-punching

Strand 1

Strand 2
Strand 2 Strand 1

Load (kN)

Load (kN)

post-punching

Strand 1 20

Strand 2 40 60 80 100 120

20

40

60

80

100

120

Strands Force after Transmission (kN)

Strands Force after Transmission (kN)

Model DF3
300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0 0 Strand 1 Strand 2
punching Strand 1 Strand 2
post-punching

Model DF7
340 320 300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0 0
punching

Load (kN)

Load (kN)

Strand 2 Strand 1

post-punching

Strand 1 20

Strand 2 40 60 80 100

20

40

60

80

100

Strand Force after Transmission (kN)

Strands Load after Transmission (kN)

Model DF5
300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0 0
punching Strand 1 Strand 2

Model DF7
340 320 300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0 0 Strand 3 Strand 4
Strand 3

Load (kN)

post-punching

Load (kN)

Strand 4

Strand 1

Strand 2

20

40

60

80

100

120

20

40

60

80

100

Strands Force after Transmission (kN)

Strands Load after Transmission (kN)

Fig. 7. Strand forces during tests with load for strengthened specimens DF2, DF3, DF5, DF6 and DF7.

Table 2 presents the results for strand loads and the installed bond stresses after transmission and at punching failure, using the uniform bond stress model [19] and it is possible to conclude that bond stresses dont reach the maximum value obtained in pull-out tests of 12.0 MPa [19]. Bond stresses at post-punching maximum load are treated in Section 5.2.6. 4.2. Load/displacement results This section presents the results obtained from the analysis of the slabs vertical displacements, taken as averages of pairs of values measured by displacement transducers, symmetrically opposed in relation to the centre of the slabs. Fig. 9 shows displacements for slabs DF2, DF3, DF5, DF6 and DF7 measured at the several displacement transducers, and the three different stages mentioned above (prestress transmission, punching failure and post-punching stage) can be seen in all of them, with each stage being indicated in the graphs. In the graphs, for example, D1 and D5 represent the average of the values registered by displacement transducers 1 and 5, relative to D3. Its important

to mention that displacements measurements may not be accurate, since these displacement transducers were positioned onto the top surfaces of the slabs and the measurements may be affected by spalling at the tops of the slab. The punching failures caused sudden decreases of load together with large displacements. When the loads were increased again the stiffness were much lower than during the test up to punching. The post-punching behaviour was similar in most tests with stiffness decreasing progressively until the load/displacement lines became horizontal. Tests were generally terminated when top bars broke near the column. Regarding specimen DF3 it should be noted that after punching failure there was an almost complete discharge and that during the new loading this specimen showed an approximately linear behaviour until it reached a load of 110 kN. This load coincides roughly with the loads sustained by the other slabs immediately after punching, except for specimen DF7 where this value is about 125 kN (Fig. 9). Beyond that point the stiffness of the specimens fell progressively until it reached the above-mentioned horizontal baseline. Tests were terminated when top reinforcing bars rupture occurred, just after they reached a stage of

390

D.M.V. Faria et al. / Engineering Structures 40 (2012) 383397

orthogonal to the deviators, the other rebars were locked by them, which also contributed to the post-punching resistance. Model DF3 was practically totally unloaded due to a loading problem in the load actuator, that didnt work as in the other models. In Fig. 9, is possible to observe that Model DF3 presented a lower stiffness during post-punching loading, when compared to other models. This smaller stiffness may be associated with the fact that this model presented the lowest bonded lengths (see Table 3), and so pull-out behaviour also presented a smaller stiffness than other models. This is shown in Fig. 7, where is possible to observe that the strand loading growing rate was relatively smaller than in other slabs, leading to lower strand forces at maximum postpunching load. Results from the different four load cells on the reaction beams, used in each model, are not presented separately since these measured approximately the same load values, that is, total load was equally distributed among them. At the undersides of the slabs, openings formed following the outlines of the loading plates. Fig. 10 shows that the sloping sides of the inverted pyramids of concrete above the plates crumbled leaving approximately vertically faced blocks supporting the top bars. A clear rotation of the deviators over the slabs can also be seen. Figs. 8, 10 and 11 present pictures of the several specimens after the post-punching stage. In Figs. 10 and 11, it is possible to observe that concrete was cracked around the area where the strands enter in the slab, and this is related to the observations by Melo and Regan [10], who proposed Eq. (2), to quantify a load per bar (without R) corresponding to failure resulting from this type of damage and to the upper limit of Eq. (9) presented by MC2010 [18], corresponding to a concrete breakout above the strand when it enters in the concrete slab. 5. Post-punching loads
Fig. 8. Rotations of deviators in specimen DF3.

5.1. Comparison of post-punching and punching loads Maximum post-punching loads are presented in Table 4 and compared with punching failure loads. A ratio between the post-punching maximum load and the punching load (VPP,exp/Vexp) greater than unity was never registered, and the average value for that ratio was about 0.78. It should be noted that Ramos [20] and Ramos and Lcio [15] obtained an average value for the relationship VPP,exp/Vexp of 1.25 in specimens in which the prestress strands were positioned above the column. In these cases there were four strands with 12.7 mm nominal diameter in both orthogonal directions.

constant loading and increasing deformation. In model DF3, although no rupture of bars occurred, tests were stopped in order to avoid any damage to monitoring equipment. In the post-punching stage the specimens exhibited large deformations and the rebars could be seen to be detached, especially in the vicinity of the column (Fig. 10). It is important to note that the top reinforcement was held by the deviators and this is the reason, in some cases, for the local bending and the rupture of the bars. As the reinforcement bars with higher effective depth were positioned
Table 2 Strand loads and bond stresses for different test stages. Specimen DF2 DF3 DF5 DF6 DF7 Strand S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S3 S4 Pi,initial 48.0 40.5 50.5 52.0 58.3 59.4 53.9 74.3 55.3 51.5 46.4 33.0
a

(kN)

si,initial b (MPa)
3.60 2.93 5.20 4.65 4.58 4.05 4.76 4.11 4.53 4.33 3.67 2.53

Pi,nal 84.3 76.7 78.1 78.6 76.1 97.6 79.0 96.7 67.2 62.7 33.3 33.0

(kN)

si,nal d (MPa)
6.32 5.54 8.04 7.03 5.97 6.66 6.98 5.34 5.72 5.46 2.40 2.30

Pi,PPe (kN) 102.5 92.83 52.29 52.77 64.93 110.93 99.78 109.42 96.02 91.55 32.68 32.97

a b c d e

Prestress load after transmission in each strand. Bond stress corresponding to Pi,initial. Prestress load at punching failure in each strand. Bond stress corresponding to Pi,nal. Prestress load at post-punching maximum load in each strand.

D.M.V. Faria et al. / Engineering Structures 40 (2012) 383397

391

Model DF2
300 275 250 225 200 175 150 125 100 75 50 25 0
0,0 5,0
punching post-punching

Model DF5
D6 and D7

D2 and D4

Load (kN)

D1 and D5

D1 and D5 D6 and D7 D2 and D4

300 275 250 225 200 175 150 125 100 75 50 25 0


0,0 5,0 10,0

punching

Load (kN)

post-punching

D1 and D5 D6 and D7

D2 and D4

D1 and D5 D6 and D7 D2 and D4

10,0

15,0

20,0

25,0

30,0

35,0

40,0

45,0

15,0

20,0

25,0

30,0

35,0

40,0

45,0

Displacement (mm)

Displacement (mm)

Model DF3
300 275 250 225 200 175 150 125 100 75 50 25 0
0,0
punching

Model DF6
300 275 250 225 200 175 150 125 100 75 50 25 0
0,0 5,0
punching

D1 and D5
post-punching

Load (kN)

post-punching

D1 and D5

Load (kN)

D6 and D7

D2 and D4 D6 and D7

D2 and D4

D1 and D5 D6 and D7 D2 and D4

D1 and D5 D6 and D7 D2 and D4

5,0

10,0

15,0

20,0

25,0

30,0

35,0

40,0

45,0

10,0

15,0

20,0

25,0

30,0

35,0

40,0

45,0

Displacement (mm)

Displacement (mm)

Model DF7
325 300 275 250 225 200 175 150 125 100 75 50 25 0
-5,0 0,0
punching

D8 and D9 D2 and D4

D1 and D5 D6 and D7
post-punching

Load (kN)

D1 and D5 D6 and D7 5,0 10,0 15,0 20,0 25,0 30,0

D2 and D4 D8 and D9 35,0 40,0 45,0

Displacement (mm)
Fig. 9. Displacement evolution with load.

5.2. Development of a proposal for estimating post-punching resistance In this section the development of a design proposal is presented based on the load path envisaged from the column to the slab. This load path may be divided in the following parts: 1. Transfer of vertical force from the concrete to the top bars over the column, which involves bearing stresses that tend to crush and split the concrete. 2. Transfer of force by the top bars to the deviators with the extensions of the bars beyond the deviators anchored in the concrete at a reduced inclination. 3. Transfer of force along the deviators, which requires exural resistance by the deviators. 4. Transfer of force by the strands between the ends of the deviators and the starts of the strand anchorages in the concrete. 5. Anchorage of the strands in the concrete.

Expressions for the participation of each Part in the post-punching resistance are described next. Before describing the expressions, it must be said that as referred before, tests were terminated after they reached a stage of constant loading and increasing deformation. At this point, except in DF3, several rebars ruptured and so this was the main reason why slabs were not able to develop any more resistance. Nevertheless, it may be observed in Figs. 10 and 11 that the concrete at the corner of the columns was spalled away below the bars and that the concrete was cracked around the area where the strands entered in the slab. Also, from those gures, it is possible to observe that extensions of the bars beyond the deviators spalled the concrete cover over a certain length, and were inclined at a reduced angle w between the deviators and their anchorages in the outer concrete, thus carrying loads proportioned to sin w. Fig. 12 presents a model of the deected shape of the slabs, showing the parts of the load path described above. As the bottom reinforcement did not cross the column, it could not participate in the post-punching resistance.

392

D.M.V. Faria et al. / Engineering Structures 40 (2012) 383397

Although there was no column reinforcement in the present specimens, the bearing stresses did not cause the failures, and it seems unlikely that they would be critical in normal circumstances. 5.2.2. Load-path Part 2 In the tested specimens and after punching, it was observed that the post-punching load was being supported by the top reinforcing bars as they were locked by the strands deviators. So, even though prestress strands do not cross the column top they can promote an increased post-punching load via the locking of the top rebars that cross the column top. This includes top bars that do not cross under the deviators, since these bars were positioned under the top bars that were locked by the deviators. In light of the experimental results, the following expression to predict the maximum Part 2 post-punching resistance, may based only on the top reinforcement bars that cross the column top according to Eq. (10):

V PP;rb

sin a Asp f y

10

where VPP,rb is the post-punching maximum load provided by top reinforcing bars, a is the top reinforcing bar slope with the horizontal when the post-punching maximum load is reached, Asp is the cross section of each rebar that crosses the column top and fy is the yield strength of the rebars. Each rebar that is soundly anchored in both sides of the column must be considered twice. Average values for a were recorded for each specimen (Table 3). Values proposed for a are 29 and 16 for the unidirectional case, for bars perpendicular and parallel to the deviators, respectively, and of 24 for the bidirectional one. These values were obtained from the average of the values registered in tests, and are represented in Fig. 12. Regarding the resistance provided by the bars continuing at an angle w beyond the deviators can be estimated from Eq. (11).

V PP;sp
Fig. 10. Longitudinal rebars detachment and concrete cone crumbling (specimen DF2 after post-punching stage).

sin w Asp f y

11

5.2.1. Load path Part 1 This part corresponds to the existence of bearing stresses that tend to crush and split the concrete, which was observed in all tests (Figs. 10 and 11). According to Melo and Regan [10], failure of the concrete in columns is unlikely to be a problem as the bars of the slab pass inside the cage of the column reinforcement.
Table 3 Reinforcement bar slopes, damage and bonded lengths. Specimen Reinf. bar a ()
a b

where VPP,sp is the post-punching maximum load provided by the extensions of the bars beyond the deviators and Asp is the cross section of one bar crossing the column including bars between unidirectional deviators. Measured values of w are given in table 3 and it is proposed that w should be taken as 6o and 7o for bars respectively perpendicular and parallel to uni-directional deviators and as 9o were there are deviators in both directions. These values were obtained from

Reinf. bar extensions w ()


a b

Strand

Angle h ()

Damage at the top (mm)

Initial bonded length (mm)

DF2 DF3 DF5 DF6 DF7

35 37 30 14 25

15 20 13 16 23

7 6 6 6 9

8 7 7 7 9

S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S3 S4

10.0 10.0 10.3 10.8 12.0 12.3 13.3 12.3 12.0 12.1 11.4 11.5

200 175 200 205 160 150 180 90 215 220 215 218

279 290 203 234 267 307 237 379 246 241 291 301

Bars with higher effective depth. Bars with lower effective depth. Measured between the face of damaged concrete and the axis of the steel strand in its initial position, that is, against the bottom face of the hole (Fig. 12). d Bonded length before any concrete breakout.
b c

D.M.V. Faria et al. / Engineering Structures 40 (2012) 383397

393

Fig. 11. Views of specimens DF5, DF6 and DF7 after post-punching stage.

the average of the values registered in tests, and are represented in Fig. 12. The results and the observation of Fig. 10 for Model DF7 show that concrete cover was not so damaged by splitting as in other slabs. Also, Fig. 9 for Model DF7 shows that during postpunching the registered behaviour was somewhat stiffer than in other models, showing that orthogonally positioned deviators provide a better locking of top reinforcing bars, thus making the behaviour stiffer. Angles a and w are dependent on the distance between the outline of the column and the deviators, so the angles appointed above are only valid for deviators positioned at distances from the column similar to the ones tested. 5.2.3. Load path Part 3 This part refers to the transfer of force along the deviators. Forces acting on a deviator are presented in Fig. 12. The deviator must be designed so that exural and shear resistances are sufcient for the loads in the gure. The resistance of the deviators used in the tests was not governing, and so in the following comparisons with the tests results, this part will be neglected. At the

end of the tests, deviators were subjected to a visual inspection and no deformation was observed. 5.2.4. Load path Part 4 Concerning the transfer of force by the strands between the ends of the deviators and the starts of the strand anchorages in the concrete, Eq. (12) must be respected:

V PP;st

sin b Pstrand

12

where VPP,st is the post-punching maximum load provided by the unbonded length of a prestress strand, b is the angle of inclination of the strand with the horizontal (Fig. 12) and Pstrand 6 Apfp is the strand force. In Figs. 10 and 11, it is possible to observe that the inclinations b of the strands between the deviators and the concrete were greater than the angle of inclination of the portion of the strand embedded in the slab, which was of about h % 11. 5.2.5. Load path Part 5 Regarding Part 5, the anchorage of the strands in the concrete is taken into account limiting Pstrand to smaxpdclav, where

394

D.M.V. Faria et al. / Engineering Structures 40 (2012) 383397

Fig. 11 (continued)

Table 4 Comparison of punching and post-punching maximum loads. Specimen Vexp a (kN) VPP,exp b (kN) VPP,exp/Vexp
a b

DF2 272.9 230.6 0.84

DF3 254.6 190.1 0.75

DF5 295.0 194.0 0.66

DF6 292.7 230.0 0.79

DF7 319.5 284.0 0.89

Punching failure load. Post-punching maximum load.

smax = 12.0 MPa [19], dc is the strand nominal diameter and lav is
the available bonded length, the determination of lav is described in the next section. 5.2.6. Design procedure The design procedure to determine the maximum post-punching resistance must be the following: reinforcing bars and strand forces are related to each other, since Eq. (10) must be equal to the sum of Eq. (11) with Eq. (12); so, in a rst step, it must be as-

sumed that reinforcing bars are yielded in order to nd Pstrand using the above methodology. If Pstrand > Aspfp or Pstrand > smaxpdclav then the yielding of the strand or bond capacity, respectively, govern design and thus the post-punching resistance. If Pstrand is less than both Aspfp and smaxpdclav, the top reinforcing bars govern the resistance. Deviators must be designed according to Section 5.2.3, taking in account Fig. 12d and Pstrand computed before. Another issue, is the fact that the differences of the angles of inclination b and h, originate a localized force (Pstrandsin(bh)), that induces concrete breakout. The destruction of concrete will occur until the concrete thickness (dint) above the strand is enough to prevent further concrete breakout (Fig. 12). This concrete thickness (dint) may be computed using Eq. (2), and after nding it, lav may be computed. In order to perform all the calculations above it is necessary to know the angle b and (bh), and it may be found from experimental results, by the application of the referred methodology, assuming that top reinforcing bars crossing the column are yielded, since in tests reinforcing bar ruptures were observed, except in model DF3. The results from this analysis are presented

D.M.V. Faria et al. / Engineering Structures 40 (2012) 383397

395

(a)

(b)

(c)

(d)

(e)

Fig. 12. Different parts of the load path and geometrical parameters, (a) section of the slab passing thru the longitudinal axis of the deviator, (b) section thru the axis of the slab in a direction parallel to the deviators, (c) section thru the axis of the slab in a direction perpendicular to the deviators, (d) Part 3 acting forces and (e) deviation force due to slab deformation inducing concrete breakout.

in Table 5. In model DF6, four bars instead of the three that would be expected (accordingly to Fig. 4) were mobilized since the concrete cone around the column didnt crumble completely as the rest of the models. From the observation of Table 5 it is possible to conclude that the angle (bh) ranges between 0.7 and 4.4 and no relation between (bh) and any other tests variable is
Table 5 Determination of b and dint in order to avoid concrete breakout for tested models. Spec. DF2 DF3 DF5 DF6 DF7 Eq. (10) (kN) 213.5 242.0 184.5 175.6 207
a

found. So, it is proposed to perform the design calculations using two values for (bh), one of 0.5 and other of 4.5, thus obtaining two values for post-punching resistance, adopting the lower of them. It may be also concluded that in some cases concrete breakout occurred since dint > measured dint. These results match well with the observed behaviour, as in those models, some cracks

Eq. (11) (kN) 73.7 84.2c


b

Str S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S3 S4

(bh) 0.7 0.7 4.4 3.9 0.6 0.3 2.3 3.3 1.9 1.8 2.5 2.4

Pi,PP sin (bh) (kN) 1.3 1.2 4.0 3.6 0.7 0.6 4.0 6.3 3.2 2.9 1.4 1.4

dint (mm) (Eq. (2)) 22 21 39 37 17 15 41 51 36 35 24 24

Measured dint (mm) 24 20* 25* 28* 23 21 31* 9* 34* 35 33 33

lav (mm) 279 284 126 186 267 307 192 178 235 241 291 301

si,PP (MPa) e
7.7 6.9 8.7 5.9 5.1 7.6 10.9 12.9 8.6 8.0 2.4 2.3

10.7 14.7 12.6 15.6 15.7

63.2b 73.7c 54.2b 63.2c 54.2b 63.2c 81.1b 81.1


c

In this slab, 4 reinforcing bars in each direction were accounted instead of 3. Bars with lower effective depth. Bars with  higher effective depth.  d b arcsin Eq: 10Eq: 11 . Pi;PP e Bond stress corresponding to Pi,PP taking into account lav. Correspond to models where dint > measured dint and so, lav < initial bonded length (Table 3).
b c

396

D.M.V. Faria et al. / Engineering Structures 40 (2012) 383397

Table 6 Comparison of the design proposal with the experimental results. Spec. Eq. (10) (kN) Eq. (11) (kN) Std Pstrand (kN) (bh) = 0.5 DF2 DF3 DF5 DF6 DF7 197.1 197.1 197.1 197.1 210.8 136.9 136.9 117.4 117.4 162.2 S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S3 S4 82.6 82.6 80.3 76.8 92.1 90.0 83.6 90.0 56.2 55.8 59.0 58.5 (bh) = 4.5 60.1 60.1 58.9 57.0 70.2 69.0 65.2 69.0 42.8 42.6 44.4 44.1 dint (mm) (Eq. (2)) (bh) = 0.5 16 16 16 16 18 18 17 18* 14 14 15 14 (bh) = 4.5 42* 42* 42* 41* 48* 48* 46* 48* 37* 37* 38* 38* lav (mm) (bh) = 0.5 279 290 203 234 267 307 237 334 246 241 291 301 (bh) = 4.5 177 163 109 162 145 183 170 196 231 233 261 276

s (MPa)a
(bh) = 0.5 6.2 6.0 8.3 6.9 7.2 6.1 7.4 5.6 4.8 4.8 4.2 4.1 (bh) = 4.5 7.1 7.7 11.3 7.4 10.2 7.9 8.0 7.4 3.9 3.8 3.6 3.3

a *

Bond stress corresponding to Pstrand taking into account lav. Correspond to models where dint > measured dint and so, lav < initial bonded length.

Table 7 Comparison of experimental post-punching maximum loads and its predictions. Specimen (Table 5)Eq. (10)/VPP,exp (Table 6)Eq. (10)/VPP,exp DF2 0.93 0.85 DF3 1.26 1.04 DF5 0.95 1.02 DF6 0.76 0.86 DF7 0.73 0.74 Average 0.93 0.90 COV 0.23 0.14

appeared around the start of the strands bonded length, as may be observed in Figs. 10 and 11. In these cases the available bonded length lav is computed using the calculated dint, allowing the correct determination of bond stresses, that in few cases attains values close to the limit of 12.0 MPa [19], meaning that rupture of steel bars occurred simultaneously with the maximum bond capacity of the bonded strands (Table 5). In Table 5 it is possible to observe that only in specimen DF6 bond stresses si,pp reached values close to 12.0 MPa, which is consistent with the less stiff bond behaviour registered during testing and presented in Fig. 7. After the description of the load path and of the design procedure, it is clear that the way this strengthening technique works is different from conventional post-tensioning. In conventional post-tensioning only strands passing through the columns can be effective in increasing the post-punching resistance of a at slabcolumn connection. Prestress strands that dont pass through the columns are not attached to the column in any way, since they pass within reinforcement layers, and so they are ineffective regarding post-punching resistance. Contrarily, if prestress strands pass through the columns, then there are some similarities with the strengthening technique, namely with Load Path Parts 1 and 4 and care must be also taken regarding concrete breakout that arises from deviation forces due to the slab deformation. The strengthening technique is limited by the capacity of top reinforcement bars that cross the column, and also by the bond in the strands anchorages, which does not exist in conventional posttensioning. 5.3. Comparison of the design proposal with the experimental results In this section the design procedure described in Section 5.2.6 is applied and results are compared with the experimental results. It is assumed: (bh) = 0.5 and 4.5; a of 29 and 16 for the unidirectional case, for bars perpendicular and parallel to the deviators, respectively, and of 24 for the bidirectional one; w of 6 and 7 for the unidirectional strengthening case, for bars perpendicular and parallel to the deviators, respectively and of 9 for the bidirectional one, as mentioned in Section 5.2. The damaged lengths and h

are the ones presented in Table 3, measured in each specimen, since in a real case it is possible to measure both. The main conclusions that arise from the observation of the results presented in Table 6 are that assuming the average values a, w and extreme values for b h, mentioned above, the estimated post-punching resistance is governed by reinforcement bars failure, as observed in tests for all models, except in model DF3, due to reasons exposed in Section 4.2. Adopting a smaller value for (bh), implies an increase of Pstrand relative to adopting a higher value of (bh). Higher bond stresses are attained for smaller values of Pstrand, since for these values, higher deviation forces are originated at the points where the strands enter in the slab thus inducing concrete breakout, diminishing the available bonded length. Table 7 gives the ratios between the post-punching resistance predicted by Eq. (10) and experimental values, both, when applying geometry details measured in tests and the average geometry values regarding the angles a, w and extreme values for b. It is clear that for model DF6, the consideration of four reinforcement bars (from Table 5) crossing the column underestimates post-punching resistance, and this may be due to the fact that, when considering the fourth bar, its angle a was much lower than the ones of the remaining bars and the average between the different values a for all other bars masked the results, causing a decrease of the vertical component of bar forces. It may be concluded that the average relations are both lower than unity and provide conservative predictions of post-punching maximum loads.

6. Conclusions This paper describes the experimental research conducted on a new reinforced concrete slab strengthening technique and presents the results obtained regarding its post-punching behaviour. Five slab specimens were tested in the post-punching range and it was found that on average the post-punching resistance is 78% of the punching resistance, using in most cases only two prestressing strands that did not cross the column top. It was found that postpunching resistance is ensured by different physical phenomena and a design proposal is presented based on the load path envisaged from the column to the slab. This load path may be divided in the following parts: transfer of vertical force from the concrete to the top bars over the column, which involves bearing stresses that tend to crush and split the concrete; transfer of force by the top bars to the deviators with the extensions of the bars beyond the deviators anchored in a portion of the concrete; transfer of force along the deviators, which requires exural resistance by

D.M.V. Faria et al. / Engineering Structures 40 (2012) 383397

397

the deviators; transfer of force by the strands between the ends of the deviators and the starts of the strand anchorages in the concrete; and anchorage of the strands in the concrete. Expressions for the vertical load capacities corresponding to the parts of the load path are presented and compared with the experimental results, showing its adequacy to predict the observed mode of failure and the post-punching capacity. Compared with other strengthening techniques, the one proposed here, apart from upgrading ultimate and serviceability behaviour, also adds post-punching resistance. This is a great advantage for the prevention of a progressive collapse since it reduces the probability of a total collapse in the event of a local punching failure. Acknowledgements This work received support from the Fundao para a Cincia e Tecnologia Ministrio da Cincia, Tecnologia e Ensino Superior through Scholarship Number SFRH/BD/37538/2007 and Project PTDC/ECM/114492/2009. We would like to thank CONCREMAT for making the slab specimens VSL for the prestress strands and HILTI Portugal who supplied the epoxy adhesive (HIT-RE 500) and all the equipment related to adhesives and drilling. References
[1] Delatte, Norbert J. Beyond failure: forensic case studies for civil engineers. Berlin: ASCE Press; 2009. [2] Gardner NJ, Jungsuck Huh, Chung Lan. Lessons from the Sampoong Department store collapse. Cem Concr Compos 2002;24(6):5239. [3] Ruiz MF, Muttoni A, AndKunz J. Strengthening of at slabs against punching shear using post-installed shear reinforcement. ACI Struct J 2010;107(4): 43442.

[4] Wood J. Pipers row car park, Wolverhampton. Quantitative study of the causes of the partial collapse on 20th March 1997. Health and safety executive report. UK; 2008. [5] Faria D, Lcio V, Ramos A. Flat slab strengthening systems. SILE011 International Seminar on Structural Connections, Lisbon; March 2011 [in Portuguese]. [6] Faria DMV, Lcio VJG, Ramos AMP. Strengthening of reinforced concrete at slabs using post-tensioning with anchorages by bonding. Fib symposium concrete: 21st century superhero, London; June, 2009. [7] Faria DMV, Lcio VJG, Ramos AMP. Strengthening of at slabs with posttensioning using anchorages by bonding. Eng Struct 2011;33:202543. [8] Regan PE. Behaviour of reinforced concrete at slabs, CIRIA report 89, Construction Industry Research and Information Association, London; 1981. [9] Rasmussen BH. Betonindstobde Tvaerbelastede Bolts og Dornes Baereevne, Bygningsstatiske Meddelelser, 34, Copenhagen; November, 1963. p. 3955. [10] Melo GSSA, Regan PE. Post-punching resistance of connections between at slabs and interior columns. Mag Concr Res 1998;50(4):31927. [11] Georgopoulos T. Verbesserung des Bruchverhaltens punktfrmig gesttzter Stahlbeton-platten. Bautechnik 1988;65:26770. [12] Mitchell D, Cook WD. Preventing progressive collapse of slab structures. ASCE J Struct Eng 1984;110(7):151332. [13] Ritz P, Marti P, Thurlimann B. Versuche ber das biegeverhalten von vorgespannten platen ohne verbund, Institut fr Baustatik und Konstruktion, ETH, Zurich, Bericht no. 7305-1; April, 1975. p. 114. [14] Pralong J, Brandli W, Thurlimann B. Durchstanzversuche an stahlbeton-und spannbeton-platten, Institut fr Baustatik und Konstruktion, ETH, Zurich, Bericht no. 7305-3; 1979. [15] Ramos A, Lcio V. Post-punching behaviour of prestressed concrete at slabs. Mag Concr Res, Thomas Telford 2008;60(4):24551. [16] Comit Euro-International du Betn. CEB-FIP model code 1990. Thomas Telford, London; 1993. p. 437. [17] Fdration Internationale de la Prcontrainte. Recommendations for the design of post-tensioned slabs and foundations. SETO, London; May, 1998. p. 44. [18] Federation International du Beton. Model Code 2010. Voted version for publication in 2012. [19] Faria DMV, Lcio VJG, Ramos AMP. Pull-out and push-in tests of bonded steel strands. Mag Concr Res 2011;63(9):689705. [20] Ramos A. Punching in presstressed concrete at slabs. PhD thesis. Technical University of Lisbon, Lisbon; 2003 [in Portuguese].

Anda mungkin juga menyukai