Ramos, A.P.; Lúcio, V.J.G. ; Regan, P.E.

; Repair and strengthening methods of flat slabs for punching;

International Workshop on Punching Shear Capacity of 1 RC flat Slabs; Trita-BKN. Bulletin, Volume 57,
Royal Institute of Technology em 2000.

Repair and Strengthening Methods of Flat Slabs for Punching

Mr. António M. P. Ramos, M.Sc., Ph.D. Student

ampr@mail.fct.unl.pt

Prof. Válter J.G. Lúcio, Ph.D.

lucio@civil.ist.pt

Prof. Paul E. Regan, Ph.D., Prof. Emeritus of University of

Westminster

Instituto Superior Técnico

Avenida Rovisco Pais
Lisboa, Portugal

ABSTRACT

The present work reports experimental research carried out on

repair and strengthening methods of flat slabs for punching. The
repair and strengthening methods studied are: strengthening using
transversal prestress, repair by substitution of the damaged
concrete; and strengthening using steel beams as a column head,
connected to the column and to the slab with epoxy resin and
mechanical expansion anchors. Four experimental test slabs
(AR1 to AR4) were produced and tested: two with transversal
prestress, one only repaired and one strengthened with a steel
column head. The execution process, its efficiency and design,
are discussed.

Key words: punching, flat slabs, repair, strengthening

1. INTRODUCTION

In flat slabs the load and moments transferred between the slab and the column cause high
stresses in the slab, near the column. These stresses produce cracking and may lead to failure.
The failure is associated with the formation of a pyramidal plug of concrete which punches
through the slab. The punching failure results from the superposition of shear and flexural
stresses in the slab, near the column. It is a local and brittle failure mechanism.

Nowadays, flats slabs are a common solution for buildings because they are economical, easy
and fast to build. The need to study suitable strengthening and/or repairing methods is
associated to the increased use of this kind of slab.
 2

There are many reasons to repair and/or strengthen flats slabs (e.g. construction or design errors,
poor quality or inadequate materials, overloading and accidents). The repair and/or
strengthening method (or methods) to be used in any particular situation depends on technical
and economical factors, and may be a complex task.

The present work reports the experimental research carried out to study one repair and two
strengthening methods for flat slabs under punching.

2. MATERIALS PROPERTIES

For accessing the concrete strength used to make the specimens, compression tests on cubes of
15x15x15cm3 (fcm,cube) were carried out. The results are showed in Table 1. The reinforcement
steel tensile yielding strength is also presented.

Table 1 – Materials properties

Concrete Steel (fsy – MPa)
Slab
age (days) fcm,cube (MPa) T6 T10 Bolts Beams
PR1 11 39.2 704 487 412 -
PR2 7 38.6 704 487 412 -
10 46.0
old
PR3 316 56.4 515 481 - -
new 15 63.9
old 700 40.9
PR4 515 481 - 377
new 51 68.3

3. STRENGTHENING USING TRANSVERSAL PRESTRESS

3.1. General

This method of strengthening consists of drilling holes through the slab, near the column, and
inserting steel bolts which are prestressed against the slab surfaces. This technique was also
described and tested, with success, by Ghali et al [1].

Two tests were carried out at the University of Westminster. The specimens were
2000x2000 mm2 and 100 mm thick. They modelled the area near a column of an interior slab
panel up to the zero moment lines. The punching load was applied by a hydraulic jack to a
200x200 mm2 area on the bottom of the slab.

The test slabs were loaded up to approximately 70 % of the predicted failure load (146 kN). A
large amount of flexural cracking was present at that stage. The test slabs were then unloaded
and strengthened with eight transversal steel bolts and the bolts were prestressed (5.0 kN each in
slab PR1 and 15.0 kN in slab PR2). The steel bolts used in these tests were cut from a 12.7 mm
threaded bar. The middle sections of the bolts were machined to a uniform diameter of 9.5 mm.
 3

The position of the strengthening steel bolts can be seen in Figure 2. Finally the test slabs were
loaded up to failure.

Figure 1 – Test geometry (PR1 and PR2)

Figure 2 – Position of the strengthening steel bolts

The bottom and top reinforcement consisted of 6 mm rebars every 200 mm and 10 mm rebars
every 75 mm, respectively. The average effective depths were 80 mm.

The strains in the steel bolts (in the machined area), the vertical displacements at six points and
the total applied load at each load stage, were measured during the tests.

3.2. Test results

In slab PR1, the force distribution between the eight bolts was not uniform. During loading the
bolts at the middles of the edges were more highly stressed. At a total applied load of 230 kN,
the force in the least stressed bolt was only 37 % of the force in the most stressed bolt. This is
because the corner bolts were at a greater distance from the column edges than the others. Three
bolts yielded at failure.
 4

Test Slab PR1

35

Bolt 1 Bolt 2
30
Bolt 3 Bolt 4

Force in the bolt (kN)

25
Bolt 5 Bolt 6

20 Bolt 7 Bolt 8

15

10

0
60 80 100 120 140 160 180 200 220 240
Applied Force (kN)

Figure 3 – Force evolution in the bolts (Test Slab PR1)

In slab PR2 the corner bolts were placed closer to the column. For this reason the force
distribution between the eight bolts was more uniform than in the previous model. In test PR2
no bolts yielded at failure.

Test Slab PR2

35
Bolt 1 Bolt 2
30
Bolt 3 Bolt 4
Force in the bolt (kN)

25
Bolt 5 Bolt 6

20 Bolt 7 Bolt 8

15

10

0
60 80 100 120 140 160 180 200 220 240 260
Applied Force (kN)

Figure 4 – Force evolution in the bolts (Test Slab PR2)

In both tests the force in the bolts stayed more or less constant in the first load stages. When the
applied load reached the predicted ultimate load of the slab without strengthening, the loads in
the bolts increased quickly.

In slab PR1 the failure load was 240 kN and the failure surface was internal to the perimeter
defined by the strengthening bolts. The slope of the failure surface with the horizontal plane was
about 40°.

In slab PR2 the failure load was 250 kN, the failure surface was external to the perimeter
defined by the strengthening bolts and its slope with the horizontal plane was about 28°.

The comparison between the experimental and the predicted failure loads is shown in Table 2.
The measured small eccentricities of the applied load were considered in Pu,eff , which was
estimated following the recommendations of the codes, considered. In the quantification of the
punching resistance the average material strengths, without partial coefficients, were used.
 5

In slab PR1 the increase of strength, as compared to a slab without bolts, was limited by the
resistance of the unreinforced failure surface between the load and the bolts. The resistance
increased due to a steeper failure surface than would be normal in punching.

For slab PR2 the failure surface inside the bolts would have had a steeper inclination due to the
better positioning of the corner bolts and, at the same time, the perimeter outside the bolts was
smaller than in PR1. For those reasons the punching failure occurred externally to the perimeter
defined by the bolts. The failure was influenced by the imminent bending failure of the slab
(PRm,b = 266.5 kN).

Table 2 – Comparison between experimental and predicted failure loads(slabs PR1 e PR2)

Pu,exp (kN) Pu,eff(kN) (2) PRm(kN) (3)

Slab (1) Code Pu,eff/PRm

BS 8110 [2]] 252 211 1.19

EC2 [3]] 276 200 1.38
PR1 240
ACI 318 [4]] 256 170 1.51
MC90 [5]] 252 231 1.09
BS 8110 [2]] 259 210 1.23
EC2 [3]] 288 198 1.45
PR2 250
ACI 318 [4]] 263 169 1.55
MC90 [5]] 259 230 1.13
(1) Pu,exp - experimental failure load; (2) Pu,eff - effective punching load; (3) PRm - theoretical
punching strength without bolts

4. REPAIR BY SUBSTITUTION OF THE DAMAGED CONCRETE

4.1. General

Test PR3 was made at the Universidade Técnica de Lisboa. The specimen was 1800x1800 mm2
and 100 mm thick. Otherwise it was identical to PR1 and PR2.

The slab was loaded until punching failure occurred at 180 kN (1st phase). After that, the
damaged concrete (old concrete) was removed and substituted by a repair concrete (new
concrete). The concrete surface was sound and roughened to promote mechanical adhesion
between the two concretes. All the dirt, oil, grease, and other bond-inhibiting materials were
removed. Prior to pouring the repair concrete, surface was brought to a saturated surface-dry
condition. Finally the slab was loaded again until another punching failure occurred at 183 kN.

During the tests, the deformations of three points on the slab surface and the applied load were
measured. In the first test, the strains in the top reinforcement were also measured.
 6

4.2. Test results

In the second phase the inclination of the failure surface with the horizontal was smaller than in
the first phase and the punching failure occurred through the old concrete and not through the
bonding surface of the two concretes. It should be noted that no bonding agent was used.

The experimental failure load and the theoretical resistance, according to four codes, are
presented in Table 3. Effective punching loads were estimated considering the applied load
eccentricities. The BS 8110 upper limit of 40 MPa for the useful concrete strength was
neglected.

From the results presented in Table 3 it can be seen that the ratios Pu,eff/PRm are higher in phase
one than in phase two. The ratio Pu,eff/PRm of 2nd phase is approximately 0.84 of the
corresponding value in 1st phase. This value measures the effectiveness of the repair method and
can be called the “monolithic coefficient”.

Table 3 – Comparison between experimental and predicted failure loads (Slab PR3)

Phase Pu,exp (kN) Code Pu,eff(kN) PRm(kN) Pu,eff/PRm

BS 8110 [2]] 211 222 0.95
EC2 [3]] 207 230 0.90
1st phase 180
ACI 318 [4]] 218 185 1.18
MC90 [5]] 211 243 0.87
BS 8110 [2]] 193 238 0.81
EC2 [3]] 211 273 0.77
2nd phase 183
ACI 318 [4]] 195 204 0.96
MC90 [5]] 193 261 0.74

5. REPAIR AND STRENGTHENING WITH STEEL BEAMS

5.1. General

Test PR4 was made at the Universidade Técnica de Lisboa. The specimen was 3300x3300 mm2
and 100 mm thick with a column cross section of 200x200 mm2 (Figure 5). The specimen was
intended to model an interior panel of a slab up to the lines of maximum span moment. The
bottom and top reinforcement used is shown in Figure 6. The average effective depth was
80 mm.

The slab was previously taken to failure by eccentric punching. All the damaged concrete was
removed and substituted by a repair concrete, as in the test previously described.

The slab was strengthened by the addition of steel beams acting as a column head. The steel
beams were glued to the column and to the slab bottom surfaces with epoxy resin and connected
to the column side faces with steel expansion anchors.
 7

To apply positive moments to the edges of the specimen, and to give zero relative rotations of
the opposite slab edges, steel beams were used (Figure 7). The load was applied at 16 points of
the slab top surface by means of a steel structure (Figure 5).

Figure 5 – Test geometry (Slab PR4)

Bottom Reinforcement Top reinforcement

Figure 6 – Slab reinforcement (Slab PR4)

The slab was finally loaded until failure. Eight displacement transducers were used to measure
the vertical deflections of the slab surface, and another two measured the horizontal relative
displacements between the slab surface and the steel beams. Eight rosettes, with three strain
gauges each, were glued to the steel beams at 100 mm from the support, to analyse the evolution
of the stresses in the column head. The applied load was measured with four load cells.
 8

Figure 7 – Steel structure for the boundary conditions at mid span

5.2. Test results

The yielding of the top reinforcement began at a total applied load of 250 kN.

The failure load was 375 kN. A flexural failure occurred instead of a brittle punching failure. In
this respect the repair and strengthening operation was successful.

The axial stresses and the shear forces in the steel beams were nearly proportional to the applied
load.

Table 4 - Comparison between experimental and predicted failure loads (Slabl PR4)
Exterior Control
Bending Interior Control Perimeter
Perimeter
Pu,exp (kN)
Slab PRm,b (kN) Pu,eff (kN) PRm,int (kN) Pu,eff (kN) PRm,ext (kN)
(1) Code
(2) (3) (4) (5) (6)

BS 8110 [2]] 385 515 (195+320) 378 491

EC2 [3]] 431 512 (192+320) 431 533
PR4 375 380
ACI 318 [4]] 393 494 (174+320) 379 582
MC90 [5]] 386 533 (213+320) 379 530
(1) Pu,exp - experimental failure load; (2) PRm,b - theoretical bending resistance; (3,5)Pu,eff - effective punching
load; (4) PRm,int - theoretical punching resistance obtained by adding the slab strength to the shear strength of the
steel column head; (6) PRm,ext - theoretical punching resistance considering a control perimeter exterior to the steel
column head

A comparison between the experimental and predicted failure loads is shown in Table 4. It can
be seen that the experimental failure load (375 kN - flexural failure) is quite similar to the
theoretical flexural resistance of the slab. Slipping was observed between the slab bottom
surface and the steel beams, so the flexural resistance was calculated by adding the slab and the
beam flexural resistance. The theoretical punching resistance is higher than the flexural
resistance, and for that reason a flexural failure occurred.
 9

6. CONCLUSIONS

Two methods of strengthening and one of repair, for flat slabs with punching problems, were
described and tests were performed to analyse their efficiency.

The repair method analysed consists in replacing the concrete damaged through punching failure
by a repair concrete. This technique showed good results. With this method a punching
resistance similar to that of the undamaged slab may be obtained if the surface of the old
concrete is well prepared, a suitable repair concrete is used and a good curing is performed. The
“monolithic coefficient” was approximately 0.84.

One repair and strengthening method was tested. This technique was applied to a slab, which
was damaged by punching. The damaged concrete was replaced by a repair concrete and the
model was strengthened with a column head of steel beams connected to the column and slab by
epoxy resin and steel anchors. This strengthening method shows to be able to develop a good
punching resistance, increasing also the bending resistance.

A strengthening method useful for cases where the punching failure has not occurred was also
studied. This method consists on using prestressed steel bolts through the slab thickness near the
column. Two slabs were tested and the achieved punching resistance was higher than those
predicted for the concrete slabs alone. With this method it seams that, if the prestress is not
enough to force a punching failure surface outside the perimeter of the bolts, the punching
resistance of the system may be obtained by adding the resistance of the concrete slab to the
prestressing force applied to the bolts.

REFERENCES

[1] Ghali, A.; Sargious, M. A.; Huizer, A.: “Vertical Prestressing of Flat Plates Around
Columns”, Shear in Reinforced Concrete, ACI, Special Publication SP 42, 1974, Detroit, Vol. 2,
pp. 905-920.
[2] British Standards Institution: “BS8110 - Structural use of concrete”, London, 1985.
[3] Norma Portuguesa NP ENV 1992-1-1: ”Eurocódigo 2: Projecto de Estruturas de Betão -
Parte 1.1: Regras Gerais e Regras para Edifícios” (”Eurocode 2 : Design of Concrete
Structures - Part 1.1 : General Rules and Rules for Buildings”), April, 1998.
[4] American Concrete Institute: “Building code requirements for reinforced concrete”, ACI
Committee 318, Detroit, 1995.
[5] Comité Euro-International du Béton: “CEB-FIP - Model Code 1990”, Bulletin
d´information nº213/214, May, 1993.