## Services on Demand

## Journal

## Article

## Indicators

- Cited by SciELO
- Access statistics

## Related links

- Cited by Google
- Similars in SciELO
- Similars in Google

## Share

## Revista IBRACON de Estruturas e Materiais

##
*On-line version* ISSN 1983-4195

### Rev. IBRACON Estrut. Mater. vol.7 no.3 São Paulo May/June 2014

#### http://dx.doi.org/10.1590/S1983-41952014000300006

**Punching shear in reinforced concrete flat slabs with hole adjacent to the column and moment transfer**

**D. C. Oliveira ^{I}; R. B. Gomes^{I}; G. S. Melo^{II}**

^{I}School of Civil Engineering, Federal University of Goias, Goiânia, Brazil; E-mail: diordrum@gmail.com; rbggomes@gmail.com

^{II}Department of Civil and Environmental Engineering, University of Brasilia, Brasilia, Brazil. E-mail: guilherm@unb.br

**ABSTRACT**

The structural behavior and the ultimate punching shear resistance of internal reinforced concrete flat slab-column connections, with one hole adjacent to the column, with or without flexural moment transfer of the slab to the column was investigated. Main variables were: the existence whether or not hole, flexural reinforcement layout and ratio, the direction and sense of the moment transferred and the eccentricity of the load (M (moment transferred to column) / V (shear)) ratio at the connection - 0,50 m or 0,25 m. Seven internal slab-column joining were tested and ultimate loads, cracking, deflections, concrete and reinforcement strains were analyzed. The existence of hole adjacent to the smaller column dimension, the hole dimension, flexural reinforcement rate and placing, the variation of relation Mu/Vu in function of the load, and, than, of eccentricity of the load, influenced the slabs behavior and rupture load. Test results were compared with the estimations from CEB-FIP/MC1990 [7], EC2/2004 [12], ACI-318:2011 [1] and NBR 6118:2007 [5]. ACI [1] and EC2 [12] presented most conservative estimates, although have presented some non conservative estimates. Brazilian NBR [5], even though being partly based in EC2 [12], presented smaller conservative estimates and more non conservative estimates. A modification on all codes is proposed for taking in account the moment caused by the eccentricity at the critical perimeter for slabs with holes.

**Keywords:** structures, flat slabs, reinforced concrete, punching shear, hole.

**1. Introduction**

Flat slabs according to NBR 6118:2007 [4] are flat horizontal laminar structures, directly supported on columns. It is an alternative structural system to the conventional one, in which slabs are supported on beams. The absence of beams may present some advantages, such as saving formwork, reduced height, increased number of stories for buildings with quota limitation and greater flexibility for architectural arrangement. The adoption of flat slabs also enables greater slab panels speed execution, which makes them more economical, comparing to conventional system with beams. This increased speed can be obtained by constructive simplicity, facilitating implementation means, reducing the cutting due to the beams absence; in the cut, fold and placement of reinforcements, and concreting.

A disadvantage of flat slabs is the possibility of a punching failure for a lower load that would be the fracture load by bending. The rupture by punching occurs suddenly, with little or no warning, almost without ductility, which may lead to a progressive collapse (propagation of a failure that was originally held in a small part of the structure, in which the resulting damage is disproportionately larger than the original).

*1.1 Justification*

Despite this structural system (flat slabs) being widely used, the shear punching study is not yet theoretically a fully defined subject. International codes and national standards address this subject (shear punching), based on empirical studies. The criteria used for sizing are not the same and specific calculations given are often asked by scholars in the area. Experimental researches are crucial for understanding the many issues surrounding the use of this type of system.

Several experimental studies have been conducted abroad and in Brazil, evaluating the shear punching resistance of reinforced concrete flat slabs and addressing many aspects, such as concentrated loads, existence of shear reinforcement, border columns, existence of openings and moment transfer, quoting HANDON and HANSON [16], REGAN [20], TAKEYA [25], CARVALHO [8], GOMES and REGAN [13 and 14], PINTO [19], CORDOVIL [10], GOMES and ANDRADE [15], SANTOS [22], DIAS [11], VARGAS [27] , OLIVEIRA [17], ANDRADE [3], COELHO [9], REGAN [21], ANDRADE [2], TRAUTWEIN [26], OLIVEIRA [18], BORGES [6], SOUZA [23] and SOUZA [24]. Among these studies, we highlight the "Shear punching in Reinforced Concrete Flat Plates with Openings Adjacent to Column and Moment Transfer", SOUZA [24], in which this study was based on. This research contributes to the solution of the problem of flat slabs shear punching with openings, and flexural moment transfer of the slab to the column.

*1.2 Literature 2eview*

**1.2.1 Regan [20]**

REGAN [20] evaluated the effect of openings positioned adjacently to columns in the shear punching resistance of flat slabs, and tried to minimize the loss of this resistance by using shear reinforcement. Eight square slabs of 2000 mm length and 160 mm thickness were tested. They were supported in the center of a column of 250 mm x 150 mm. A load was applied in eight loading points, two at each end of the slab. From this research the author concluded that the shear reinforcement placed beside the openings can be highly effective to restore lost strength due to the opening.

**1.2.2 Gomes e Andrade [15]**

At Furnas Centrais Elétricas SA in Aparecida de Goiânia,Goiás, GOMES and ANDRADE [15] researched the influence of shear reinforcement "stud" type on shear punching resistance of flat slabs with openings near the column region, which simulated the passage of pipes through slabs. According to the authors, the results showed that the openings reduced the shear punching resistance of a smooth reinforced concrete slab. However, the use of shear reinforcement with openings allowed to recover the loss, even in regions where the concrete was minimal.

**1.2.3 Souza [24]**

SOUZA [24] jointly investigated the effect of the use of adjacent openings and the application of bending moment. 19 squared flat slabs of 2400 mm length and thickness of 150 mm were tested. They were monolithically connected to a column with 850 mm (300 mm up to 400 mm down) tall with a rectangular cross section 200 mm x 500 mm. The slabs were loaded at the edges, from the top down. Between the slabs, the main features that differ from each other are: quantity, placement and dimensions of the openings, the rate and distribution of flexural reinforcement, the shear reinforcement and the load eccentricity (ratio M (moment transferred to the column) / V (shear)) in slab-column connection. The test system used is shown in Figure 1. The author concluded that resistance losses were greater for the slabs with an adjacent opening to the smallest column side, and the adjacent opening dimension greater than the dimension of the column (slabs L2, L3 and L4), which had as variables the rate and the flexural reinforcement position. It is also observed that for slabs with applied moment, the worst situations regarding the loss of resistance with adjacent column opening and moment transfer occurred when the moment is towards the opening region, which is more brittle and has no concrete to resist compressions in the plate bottom layer, which are increased due to the moment.

**1.2.4 Borges [6]**

BORGES [6] analyzed experimentally twenty flat square slabs of reinforced concrete, with 3000 mm length and 200 mm of thickness, aiming to investigate the behavior of slabs with rectangular columns, with some relations between the sides of the column, openings and reinforced shear. The characteristics and failure loads of slabs tested are shown in Table 1. In Figure 2 the models of the tested slabs by BORGES [6] are presented. The author concluded that the shear reinforcement used, consisting of "studs", positioned to engage slabs flexural reinforcement, showed adequate performance, leading to a failure surface formation to the outer region with shear steel, and that the resistance of slabs with openings and shear reinforcement can reach and even exceed the strength of similar slabs without openings. The shear reinforcement use in slabs with openings allowed an up to 86 % increase compared to the similar slab with openings and without such steel, and allowed the slab to reach shear punching resistance at least equal to the similar slab without opening. It was also reported that the use of additional flexural reinforcement bars in the region around the openings caused no increase to slabs shear punching resistance, although it has led to vertical displacements similar to the reference slab without openings. And the hooks used as anchorage of the main flexural reinforcement, which was intercepted by the openings, did not alter the failure loads of the slabs with openings.

**1.2.5 Souza [23]**

SOUZA [23] investigated the effect of the use of adjacent or distant column openings in eight flat reinforced concrete flat slabs of 1800 mm x 1800 mm x 130 mm. The characteristics and the slabs failure loads are shown in Table 2. Two openings were made in each slab with varying dimensions, and were located with respect to the square column with 150 mm length, as shown in Figure 3. Shear reinforcement was not used in any slabs. The failure loads were inversely proportional to the openings dimensions and their distances related to the column. The author found that openings in flat slabs of any size located near column significantly reduce shear punching resistance. We also found that the studied openings at a distance of 4 times effective depth (4d) related to column surface do not influence the load and failure mode for shear punching. Finally, it is shown that more studies should be conducted to a clearer conclusion about the influence of distant openings in the column shear punching resistance.

**1.2.6 Rules and specifications**

The recommendations for calculation of flat slabs are presented in table 3, with and without openings, of the American Concrete Code (ACI/318-2011 [1]), of the European Code (EC2/2004 [12]) and of the Euro-International du Béton Committee (CEB-FIP, 1990 [7]) and the Brazilian Rules (NBR 6118:2007 [5].

Table 4 presents the control perimeters and their locations for flat slabs with openings, where a cut is made in the perimeter control length for slabs without openings, from radial lines, from the center of the column towards the openings. For the calculation of stresses was considered the area corresponding to the perimeter control multiplied by the effective depth of the slab (d). Only CEB-FIP/MC 1990 [7] does not state in their perimeter control prescriptions, when there are openings, to be considered.

**2. Materials and experimental program**

The experimental program was consisted of testing, until seven slabs were ruptured (L1 to L7) aiming to experimentally investigate the behavior of slab-column connections in inner regions of flat slabs with one or two adjacent column openings, and with or without flexural moment transfer of the slab to the column. Two geometric patterns were performed. They were told apart by openings, with dimensions according to Figure 4. The main variables involved in this experimental research were: 1) the presence of openings; 2) the rate and distribution of flexural reinforcement; and 3) load, with different flexural moment transfers from the slab to the column.

Model 1, slabs L1 and L2, are with slabs without opening. Model 2 comprises a square opening with 400 mm length side adjacent to the smallest side of the column, L2 to L7, monolithically connected to a pre-stressed column, with a 200 x 500 mm rectangular. The models represent a discrete flat slab model, simulating a negative moment of an internal column. Thus, the load application points of tested slabs simulate slabs inflection points, full-scale, suggesting slabs with spans ranging from 8 to 10 meters. The discrete model represents the column region and slab to be analyzed, suitable for shear punching isolated analysis, not covering the efforts of membranes in a slab panel. Regarding the load, the same was given in Table 5. The load was defined based on SOUZA [24] study, and was designed to simulate the shear punching next to the column region, both for slabs with or without loading symmetry, transferring the flexural moment from the slab to the column (discontinuity of slab flexural moment) which is a problem frequently found in projects engineering in flat slabs (when there are different loadings or different spans on either side of the column) - and also to study the applied moment intensity effects. In order to encourage the transfer of the flexural moment in a certain direction, the applied load intensity was greater on one side of the sample. On slab L4, for example, a load was applied designated as "2P", which is the same as saying that the loading on this region was twice more than on others, except to the opposite side, which had no loading (charging designated as "0") . However, the total vertical load (four corners sum of applied load) was the same in all slabs ("4P"). The load was applied from above through leaked hydraulic actuators. It was placed on metal beams resting on the actuators, and on distribution steel beams of 100 mm x 200 mm x 20 mm, fixed on the slab plates according to Figure 1. Figure 5 shows the load supported by each distribution plate for slabs L1, without the moment transfer, and L6 with moment transfer. Figure 6 shows photographs of the test system (slab L2). These actuators were anchored in four metal ties with 29 mm of diameter. The ties passed through the slab and steel beams through holes previously made, and were anchored in the reaction slab. The reaction to loading occurred in the central column, monolithically connected to the slab. The column was supported in the bottom on a cubic block of reinforced concrete with edges of 600 mm. The block was inserted to facilitate the displacement under the slab and it had the function of transmitting the test part reaction to the slab. Due to the moment transfer, column pre-stressing was used in order to prevent rotation and also to simulate the loading of the column. The pre-stressing was done by a leaked hydraulic actuator with 1500 kN capacity, supported by a metal pre-stressing box (SAC 1045 steel). This actuator was anchored by a tie with 44 mm of diameter, passing through the metal pre-stressing box, the column and the block (hole protected by steel tubes on the column and block).

Negative flexural reinforcements were composed by orthogonal meshes with 12.5 mm of diameter bars and positioned near the upper edge of the slabs (15 mm covering). Figures 7 and 8 show flexural reinforcement used in the models 1 and 2, respectively.

The positive flexural reinforcements used in model 1 was composed of an orthogonal mesh with 10 bars of 6.3 mm of diameter in two directions, spaced in each 24 cm, as shown in Figure 9. For model 2 samples, the bars coinciding with the opening position were cut, so that it was inserted, without replacement bars, as shown in Figure 9.

The steel used was CA-50 and CA-60 types. In order to obtain these material mechanical properties, samples were tested as NBR 6152 (1992) [4]. Table 6 shows the characteristics (mechanical properties) of steels tested.

The concrete used was a self-compacting concrete (SCC), settled to reach 30 MPa (characteristic compression strength). Table 7 shows the proportion of the materials used to produce one concrete cubic meter.

The slabs were shaped in steel formworks. A metal tube with 830 mm of length and 75 mm of diameter was embedded in each formwork, in the column center, for subsequent passage of the central tie; and four rectangular metal tubes with external measures 60 mm x 100 mm for later lateral passage of the ties. For slabs with openings shaping, molds with Expanded Polystyrene (Styrofoam) were made. Figure 10 shows photographs of metallic formwork and reinforcement before L7 slab concreting and L1 and L2 moulded slabs. After concreting, water was placed on the slabs and they were covered with plastic sheets for seven (7) days. Water was replenished twice a day for the first three days and once on remaining days.

**3. Results**

*3.1 Vertical displacement*

In Figures 11 and 12 are shown the graphs of *vertical displacement x position relative to the center of the slab*, in each axis, for slabs L1 (without openings and without applied moment) and L3 (with opening and moment applied in the direction parallel to the largest side of the column). The behavior in both directions was symmetrical in slab L1. In slab L3, there was rotation in *WE* direction where displacements in each side were opposed. However, in *W* side there was a displacement down, and up in *E* side. The *W* maximum displacements were on average 2.1 times higher on the maximum displacement in *E* side. This was due to the applied load, which was "2P" intensity in *W* side, while the opposite side (*E*) has not received a load. In the *NS* direction, the maximum displacements in *N* side were slightly higher, 0.07 mm on the load of 25 kN and 1.01 mm on the load of 200 kN.

*3.2 Load and failure mode*

All slabs were cracked by puncturing. Table 8 shows the main characteristics of the slabs and their failure loads.

The L2 slab (V_{u} = 266 kN, M_{u} = 116.8 kN.m), without opening and moment applied parallel to the longest side of the column, had a 38 % loss compared to L1 slab (reference slab). This failure load reduction was due to the flexural moment transfer from the slab to the column in that sample.

The moment influence on decreasing shear resistance could also be observed comparing L4 and L5 slabs. Both of them had parallel moment applied to the longest side of the column towards the 400 mm x 400 mm opening. The difference between the models was the applied moment intensity, which was higher on L4 slab. As a result, the latter showed a of 68 % loss compared to the reference slab, while on L5 slab the loss was lower, 50 %.

Comparing the slabs with openings and moment transfer (L3 to L7) with L2 slab, without opening and moment transfer, it was observed that reducing the perimeter of the slab-column link (due to adjacent opening) does not result in load loss if the flexural moment is not applied towards the opening. The slabs L2 (V_{u} = 266 kN, M_{u} = 116.8 kN.m) and L3 (V_{u} = 250 kN, M_{u} = 113.7 kN.m), despite the last one having opening, slabs showed very close failure loads. It is also interesting to observe that when the moment intensity applied on the slab with opening was reduced, L6 slab (V_{u} = 305 kN, M_{u} = 65.8 kN.m), resulted in a failure load even greater than L2 slab without opening. This indicates that the transfer of slab flexural moment to the column is more damaging to the shear resistance, than the existence of adjacent openings to the column.

*3.3 Cracking and surface failure*

In slabs L1 and L2, the radial cracks initiated at the column. In slabs L3, L6 and L7, such cracks initiated at the same time, in the column and the openings corners. In slabs L4 and L5, the radial cracks initiated at the corners of the openings. These cracks widened toward the slabs edges. Subsequently or at the same loading, circumferential cracks appeared. The slabs cracking view is shown in Figure 13.

The failure surfaces, slabs L1 to L7, started on the slab upper surface (tensioned surface) and extended towards the column-slab junction, the slab underside (compressed surface), leading to a variety of inclinations, which originated the shear punching "cone". In slabs with opening, it was possible to see the failure surface formation through itself. The surface slope in relation to the underside of the slab L1 ranged between 31° and 53°. In L2 slab, failure surfaces occurred with slopes ranged from 27° to 85°. In L3 slab, this variation was from 26° to 32°. In slabs L4 to L7, variations were, respectively, from 22° to 67°, 24° to 41°, 17° to 36° and 2° to 41°. Figures 14 and 15 show the surfaces failure configuration for slabs L1 to L4.

*3.4 Comparison between experimental and theoretical results*

Rules for estimated loads allow the concrete strength, for the test date, as equal to concrete strength characteristics (fck ≅ fc). Because it is a scan of experimental results, no safety factor was adopted.

The results estimated by some rules ended up being lower to those experimentally found. It is worth noting that in this study, the slabs were subjected only to the efforts of shear punching and moment transfer, and the codes are committed to predict many other conditions not included in the tests, and that may happen in a real situation, such as possible asymmetric loads concentrated near the column, horizontal forces, cracking of concrete at early ages, adding arrows due to charges maintenance for a long period, unfavorable load construction conditions due to re-shoring, heavy equipment, among others. Figure 16 shows graphically the comparison between "experimental loading" / "estimated loading" (L1 slab) and "experimental stress" / "estimated stress" (slabs L2 to L7) relations. The "estimated load" and "estimated stress" were obtained with the studied standards usage.

The CEB-FIP/MC1990 [7] does not provide specifications about openings use in flat slabs. The failure load was calculated only for the L1 and L2 slabs.

The NBR 6118:2007 [5], as well as other standards showed results against safety for L1 slab (reference slab), and the ACI-318: 2011 [1] was the closest one to the experimental result. For L2 slab, without openings, estimated standards were similar to those experimentally obtained. The NBR 6118:2007 [5] had "experimental stress" / "estimated stress" relation equal to 0.97. The ACI-318: 2011 [1] was the most conservative standard: it showed "experimental stress" higher than "estimated stress" in 25 %. For L3 and L6 slabs, both with moment applied in the opposite direction to the opening, all standards presented estimates for safety, highlighting the ACI-318: 2011 [1] and EC2/2004 [12], which showed the most conservative results, related to "experimental stress" / "estimated stress" ranging between 1.26 and 1.54. For NBR 6118: 2007 [5] this ratio ranged between 1.10 and 1.28. In slabs L4 and L5, both with moment applied towards the opening, most of the rules showed estimate against safety, including NBR 6118:2007 [5], with ratio "experimental stress" / "estimated stress " ranging from 0.66 and 0.68. The only exception was ACI-318: 2011 [1], which showed results favoring safety for L4 slab. In slab L7, only ACI-318: 2011 [1] showed results favoring safety (ratio "experimental tension" / "estimated stress" equal to 1.17), although the estimated EC2/2004 [12] was fairly close to the experimental result (ratio "experimental stress" / "estimated stress" equal to 0.97). For the same slab, the NBR 6118: 2007 [5] showed a correlation of 0.84 against the security.

*3.5 Comparison between experimental results and SOUZA's results [24]*

**3.5.1 Vertical displacement**

The slabs vertical displacements behavior pattern in this study was similar to SOUZA's Slabs [24], verifying: 1) openings that led to an increase of vertical displacements around them; 2) in the direction that the moment was applied, there was a spin and the most loaded side displaced towards the load application, while the opposite edge displaced opposite the stressed load direction; 3) the direction to which the moment was not applied, displacements showed by all slabs were similar to those obtained in the reference slab L1, without openings and without moment; 4) in slabs with openings and with parallel moment applied to the longest side of the column, openings did not result in large differences of displacements on the side with opening, in relation to the slab without opening with moment applied in the same direction; 5) slab with opening and parallel moment applied to the column shortest side is the one that shows greater displacements at the most loaded edge. The probable cause would be slab-column connection inertia, being this parallel one lower than the column shortest side.

**3.5.2 Failure load**

In this study slab L6 (Vu = 305 kN, Mu = 65.8 kN.m, fc = 45.6 MPa, d = 124 mm, ρ = 1.19 %), with a 400 mm x 400 mm opening adjacent to the smallest column side and situated on the stressed edge, with parallel moment applied toward to the longest side of the column, had a failure load very close to SOUZA's L12 slab [24] (Vu = 319 kN, Mu = 74.4 kN.m, fc = 37.8 MPa, d = 123 mm, ρ = 1.48 %), with a 200 mm x 200 mm opening adjacent to the lowest side of the column, placed on the stressed edge, with a moment applied towards the longest parallel direction column side. This means that a 200 mm x 200 mm opening adjacent to the smallest side of the column and compressed by the applied moment can be as damaging to the failure load as a 400 mm x 400 mm opening adjacent to the lowest column side and tensioned by the moment applied to the variables used in this study. It is noteworthy that all SOUZA's slabs [24] have identical geometric characteristics to the slabs of this study, varying their dimension (200 mm x 200 mm, 200 mm x 300 mm or 400 mm x 400 mm) and openings location (parallel to the longest or shortest column side).

For slabs with failure parallel moment applied to the shortest column side, the use of two 300 mm x 200 mm openings adjacent to the longest side of the column is less damaging to the failure load than a 400 mm x 400 mm opening adjacent the lowest column side. In this study, it was observed a 20% load loss in slab L7 (Vu = 260 kN, Mu = 44.5 kN.m, fc = 46.8 MPa, d = 121 mm, ρ = 1.24 %) in relation to SOUZA's L18 slab [24] (Vu = 322 kN, Mu = 53.1 kN.m, fc = 37.3 MPa, d = 126 mm, ρ = 1.05 %), with two 300 mm x 200 mm openings adjacent to the column longest side.

For slabs with failure moment applied parallel to the longest column side, the partial slab-column connection loss in the region on the corners caused by two openings of 200 mm x 200 mm adjacent to the smallest side of the column, is less damaging to the failure load than a single 400 mm x 400 mm opening adjacent to the column lowest side and located in the compressed edge. Regarding SOUZA's L10 slabs [24] (Vu = 189 kN, Mu = 83.0 kN.m, fc = 34.2 MPa, d = 123 mm, ρ = 1.24 %), with two openings of 200 mm x 200 mm adjacent to the lowest side of the column, the L4 slab in this study (Vu = 137 kN, Mu = 59.0 kN.m, fc = 44.6 MPa, d = 123 mm, ρ = 1.20 %), with 400 mm x 400 mm opening adjacent to the smallest side of the column and located in the compressed edge, showed a 28 % lower failure load. Table 11 presents the characteristics of mentioned slabs.

**3.5.3 Concrete and flexural reinforcement deformations**

It can be observed that such slabs in this study as in SOUZA's slabs [24] that the reinforcement flow was reached on several points, mainly in the columns region, as expected. The discontinuous flexural reinforcement bars (which ended in the opening) were little solicited, indicating that these reinforcements are not effective against bending and therefore punching. Concerning the concrete, the largest deformation (compression) in L1 reference slab, with no opening and with no applied moment, were observed near the column center, perpendicularly directed to its longest side. For slabs with applied flexural moment, the instrumented points on the load side, parallel to the moment direction, had the highest deformations. Some strain gauges situated in the least loaded sides showed that there was tension, and the most loaded sides showed compression, indicating these slabs rotation.

**4. Conclusions**

Regarding the failure load, it was confirmed through bibliography, which refers to shear punching resistance reduction, against the moment transfer from the slab to the column (L2 slab without opening, with applied moment, showed failure load 38% lower than L1 reference slab, without opening and without applied moment). In slabs with openings, the moment transfer to the column led to a decrease in resistance between 28% and 68%, relative to the L1 slab reference. However, the worst situations regarding opening slabs resistance loss, in slabs adjacent the column, and with moment transfer, occur when the moment is towards the opening region, which is more fragile and has the same volume of concrete to resist compressions in the bottom layer of the slab. When the moment is not applied towards the opening region, the failure load is very close, or even higher than a slab without opening (with moment transfer), as comparison between the slab L2 (V_{u} = 266 kN, M_{u} = 116.8 kN.m) with L3 slabs (V_{u} = 250 kN, M_{u} = 113.7 kN.m), with opening and moment applied in opposite direction to the opening region, and L6 (V_{u} = 305 kN, M_{u} = 65.8 kN.m), identical to the previous slab. However, the moment was applied with less intensity. Thus, based on test samples, flexural moment transfer from the slab to the column is more damaging to the shear strength, than opening adjacent to the column.

As for design rules, they must be safe and even conservative, especially about failure to slabs punching, fragile and no warning ruptures, comparison of experimental results in this study with rules estimates showed that the rules requirements are not meeting the desired security. The results were not satisfactory, some even against the security, especially when the moment is applied to the parallel direction of the column largest dimension towards the opening region. As shown in Table 9, ACI-318: 2011 [1] showed that the standard was more conservative, with arithmetic average of relations τu / τr1 equal to 1.21, and EC2/2004 [12] was the closest to experimental results, with arithmetic average of relations νu / νR,c equal to 1.06, for slabs with applied moment. Regarding L1 reference slab, with no opening and no applied moment, all rules mentioned showed estimates against the safety, according to Table 10. The code that showed closest experimental result was the ACI-318: 2011 [1] with respect V_{u} / V_{Calc} equal to 0.92.

**5. Thanks**

The authors thank CNPq and CAPES for the financial support. They also thank the enterprises Carlos Campos Consultoria Limitada and Realmix Concreto Limitada for having collaborated with materials testing and characterization.

**6. Bibliographic References**

[01] ACI COMMITTEE 318. Building Code Requirements for Reinforced Concrete and Commentary - ACI 318/2011. Farmington Hills, Michigan, American Concrete Institute, 2011. [ Links ]

[02] ANDRADE, J. L. S. de. Estudo Experimental da Inclinação de Estribos Abertos em Lajes Cogumelo de Concreto Armado. Dissertação de Mestrado, Faculdade de Tecnologia, Departamento de Engenharia Civil e Ambiental, Universidade de Brasília 2000, 142 p. [ Links ]

[03] ANDRADE, M. A. S. de. Punção em lajes cogumelo - Estudo do Posicionamento da Armadura de Cisalhamento em Relação à Armadura de Flexão. Dissertação de Mestrado, Escola de Engenharia Civil, Universidade Federal de Goiás, 1999, 156 p. [ Links ]

[04] ASSOCIAÇÃO BRASILEIRA DE NORMAS TÉCNICAS. NBR 6152:1992 - Materiais Metálicos - Determinação das Propriedades de Tração. Rio de Janeiro, 1992. [ Links ]

[05] ASSOCIAÇÃO BRASILEIRA DE NORMAS TÉCNICAS. NBR 6118:2003 - Projeto de Estruturas de Concreto - Procedimento. Rio de Janeiro, 2007. [ Links ]

[06] BORGES, Liana de Lucca Jardim; MELO, Guilherme Sales Soares de Azevêdo; GOMES, R. B.; REGAN, P. E. Punching shear of reinforced concrete flat plates with openings. ACI Structural Journal, v. 110, p. 547-556, 2013. [ Links ]

[07] CEB-FIP (1990). CEB-FIP Model Code 1990: Final Draft. Bulletin D'Information, Committe Euro-International du Beton, Lausanne, July. 1991. [ Links ]

[08] CARVALHO, E. M. L. Puncionamento de Lajes Protendidas. Dissertação de Mestrado, COPPE-UFRJ, Rio de Janeiro, 1982. [ Links ]

[09] COELHO, A. E. G. Puncionamento em Lajes Cogumelo de Concreto Armado com Resistência de 30 MPa e Armadura de Cisalhamento Vertical e Inclinada. Dissertação de Mestrado, Faculdade de Tecnologia, Departamento de Engenharia Civil e Ambiental, Universidade de Brasília, 1999, 133 p. [ Links ]

[10] CORDOVIL, F.A.B. Punção em Placas de Concreto Armado. 1995, 393p. Tese de Doutorado. Departamento de Engenharia de Estruturas e Fundações, Escola Politécnica da Universidade de São Paulo, 1995. [ Links ]

[11] DIAS, D. P. Reforço ao Puncionamento em Lajes-cogumelo. Dissertação de Mestrado, COPPE/UFRJ, Rio de Janeiro, 1997. [ Links ]

[12] EUROCODE 2. Design of concrete structures - Part 1: General Rules and Rules for Buildings. European Prestandard ENV 2004-1-1:2004. European Committee for Standardization, Brussels, 2004. [ Links ]

[13] GOMES, R. B.; REGAN, P. E. Punching strength of slabs reinforced for shear with offcuts of rolled steel I section. Magazine of Concrete Research, London, United Kingdom, v. 51, n.2, p. 121-129, 1999. [ Links ]

[14] GOMES, R. B; REGAN, P. E. Punching resistance of RC flat slab with shear reinforcement. Journal of Structural Engineering (New York, N.Y.), EUA, v. JUNE, n.vol.125, p. 684-692, 1999. [ Links ]

[15] GOMES, R. B.; ANDRADE, M.A.S. de. Punching in Reinforced Concrete Flat Slabs with Holes. In: Proceedings of Developments in Computer Aided Design and Modelling for Structural Engineering. Edinburgh-UK, pp.185-193, 1995. [ Links ]

[16] HANSON, N.W.; HANSON, J.M. Shear and Moment Transfer Between Concrete Slabs and Columns. Journal. PCA Research and Development Laboratories. Vol. 10, no 1, pp 2-16, 1968. [ Links ]

[17] OLIVEIRA, D. R. C. Análise Experimental de Lajes Cogumelo de Concreto de Alta Resistência com Armadura Inclinada de Punção. Dissertação de Mestrado, Faculdade de Tecnologia, Departamento de Engenharia Civil e Ambiental, Universidade de Brasília, 1998, 137 p. [ Links ]

[18] OLIVEIRA, D. R. C. Análise Experimental de Lajes Cogumelo de Concreto Armado com Pilares Retangulares. Tese de Doutorado, Faculdade de Tecnologia, Departamento de Engenharia Civil e Ambiental, Universidade de Brasília, 2003, 214p. [ Links ]

[19] PINTO, R. C. A. Punção Assimétrica em Lajes. 1993, 145p. Tese de Mestrado, Engenharia Civil, COPPE/UFRJ, 1993. [ Links ]

[20] REGAN, P. E. Design for Punching Shear. The Structural Engineer, vol. 52, n° 6, p. 197-207, June, 1974. [ Links ]

[21] REGAN, P.E. Punching Tests of Reinforced Concrete Slabs with and without Shear Reinforcement with Openings Adjacent to Columns. 35 Marylebone Road London NW1 5LS. School of the Built Environment, University of Westminster, London, July. 1999. [ Links ]

[22] SANTOS, V. C. F. Resistência ao Puncionamento de Lajes Cogumelo de Concreto Armado. Projeto Final de Graduação, UnB, 1995, 48p. [ Links ]

[23] SOUZA, Raphael Miranda de; MELO, Guilherme Sales Soares de Azevêdo; GOMES, Ronaldo Barros. Ligações Laje-Pilar de Lajes Lisas, com Furos Adjacentes ao Pilar: Estudo Experimental e Comparações com a NBR 6118:2003. In: 53º Congresso Brasileiro do Concreto CBC2011, 2011, Florianópolis. Anais do 53º Congresso Brasileiro do Concreto CBC 2011. v. 1. p. 1-16. [ Links ]

[24] SOUZA, Raphael Miranda de; MELO, Guilherme Sales Soares de Azevêdo; GOMES, Ronaldo Barros. Estudo de ligações laje-pilar de lajes lisas, com furos adjacentes ao pilar e transferência de momento: estudo experimental e comparações com a NBR 6118:2003. In: 53º Congresso Brasileiro do Concreto CBC 2011, 2011, Florianópolis. Anais do 53º Congresso Brasileiro do Concreto CBC 2011, 2011. v. 1. p. 1-16. [ Links ]

[25] TAKEYA, T. Estudo Experimental da Ruína de Ligações Laje-pilar em Bordas de Laje Cogumelo. Dissertação de Mestrado, Esc. de Eng. de São Carlos, USP, , 1981, 241p. [ Links ]

[26] TRAUTWEIN, Leandro Mouta; BITTENCOURT, Túlio Nogueira; GOMES, R. B.; DELLABELLA, J. C.. Punching strength of flat slabs with unbraced shear reinforcement. ACI Structural Journal, v. 108, p. 197-205, 2011. [ Links ]

[27] VARGAS, E. N. Z., Punção em Lajes Cogumelo de Concreto de Alta Resistência Reforçado com Fibras de Aço. Dissertação de Mestrado, EESC/USP, São Carlos, 1997. [ Links ]

Received: 22 Jan 2014

Accepted: 18 Mar 2014

Available Online: 02 Jun 2014