Sunteți pe pagina 1din 8

The 23rd Biennial Conference of the Concrete Institute of Australia

Moment Resisting Connections with Post-Installed Reinforcement


Jakob Kunz1
1
Hilti Corp., FL-9494 Schaan, Principality of Liechtenstein

Synopsis: As a result of the increasing need for rehabilitation and upgrade of existing structures, post-
installed reinforcing bars in concrete are rapidly gaining in importance. For such applications, the Hilti HIT
injection system with the first approval in Germany has proved to be reliable and the anchorage is
designed on the basis of Eurocode 2. Only straight ends of bars are possible for installation reasons for
post-installed bars. Careful modelling of the load path in the concrete is therefore required in moment
resisting connections such as frame nodes, where hooks or bends are used for cast-in bars.
For such a frame node a design concept with strut-and-tie model is presented. A series of 18 tests on full
scale specimens with cast-in hooked and straight bars and post-installed connection bars has been
performed. The anchorage depth, the concrete strength and the reinforcement ratio have also been
varied. The test results are evaluated in this paper with respect to the strut-and-tie model. The analysis
shows, that the proposed design concept gives very accurate results for reasonable anchorage lengths,
but that the model overpredicts the loading capacity for very short anchorage lengths. In this case a
design based on the shear strength of the concrete is proposed.

Keywords: post-installed reinforcement, structural design, strut-and-tie modelling, frame node, testing

Applications
Concrete construction joints and the design of their reinforcement are
subjects which arise every day in the course of planning and work
execution. If the connecting reinforcement in an existing building member
does not meet requirements, it can be subsequently bonded in to be
rigidly connected using the Hilti HIT injection system (fig. 1). This is the
well-proven method of subsequently anchoring reinforcement which is
suitable for use on construction sites and has obtained the first DIBt
approval (1). The necessary design work can be carried out in accordance
with the reinforced-concrete standard Eurocode 2 (2). The method has
proven to be very reliable and economical. Rebars with lengths of up to
two meters can be anchored. Rebars subsequently bonded in using the
Hilti HIT injection system can be regarded as those normally cast in
concrete. In cases where the tensile forces have to be redirected, the
cast-in connecting bars are bent in the pertaining direction. Bonded-in
reinforcement, however, can only be installed with straight bar ends in
drilled holes (fig. 2). A suitable modelling of the flow of forces in the
existing building
member should be Fig.1: Injection of adhesive to
carried out so that the bond reinforcement
ultimate and
serviceability limit states can be verified. Among the
most frequent applications are, for example, the rigid
connection of retaining walls or piers (fig. 1) to existing
base plates or of cantilever slabs or corbels to existing
Fig. 2: Cast in place bonded in reinforced-concrete walls or columns.

Strut-and-Tie Modelling
Strut-and-tie models are used for design work in the ultimate limit state (ULS) of continuous and
discontinuous regions of reinforced-concrete structures, for example as per ACI 318-05 (3). The tensile
splitting forces must be given attention in the case of the concrete compression members and, if they
cannot be taken up by the concrete, reinforcement must be installed to do so. In the case of statically
indeterminate structures, the sectional forces may be calculated in accordance with the plastic theory if
the ductility requirements have been met.

ISBN 0 909375 78 X 479 Concrete 07


Strut-and-Tie Model of Frame Node
It is common practice to use hooked bars for
frame corners with opening moment6. For N1
setting reasons, post-installed bars can only M1
be applied with straight ends. Therefore,
V1
careful modeling of the node is required. Fig.
3 shows strut-and-tie models for standard
frame nodes and for situations with straight
reinforcement as described by Kupfer et
al.(4). The forces N, V, and M in Figures 2
and 3 result from the structural analysis
outside the node region. The loading of the
wall in Figures 2 and 3 results in a tensile M3
force in the reinforcement on the left hand V2
side and in a compression force on the right
C z3
hand side. Initial tests and computer N3
simulations led to the consideration that the N2
V3
straight bar has a tendency to push a M2
concrete cone against the interface with the
wall. Thus the compressive stress is at Fig. 2: Strut-and-Tie Model for Cast-in-Place Hooked
interface is not concentrated on the outside of Bars
the wall, but distributed over a larger part of
the interface, which leads to a reduced lever
arm in the wall section. A reduction factor of N1
0.85 is recommended by Kupfer et al.(4). M1
Therefore, a first approximation of the lever
arm in the wall section is z1R = 0.85·z1 where V1
z1 is the standard lever arm: z1 = 0.9·d, and d
as the static height from the compression d
surface of the wall to the center of the vertical
tension bars [300 mm – concrete cover (30
mm) – db/2]. While the equilibrium inside of z1R
frame nodes with cast-in hooked bars can be
modeled with the compression strut C M3
V2 FS0
continuing from the vertical compression
C2t C0 zo
force and anchored in the bend at the level of lb z3
the lower reinforcement (Fig. 2), straight bars S
N2 C2c T N3
are anchored by bond stresses at a level V
M2 3
above the lower reinforcement (Fig. 2b). The
strut-and-tie model requires that the angle T
between the inclined compression strut C0 Fig. 3: Strut-and-Tie Model for Post-Installed Straight
and the horizontal direction is 30º to 60º (refer bars
to Fig. 3).
As bending cracks are expected to occur along the bar from the top of the base concrete, the anchorage
zone is developing from the lower end of the bar and its length Ɛb, shown in Figure 3, is that required to
develop the steel yield strength using the bond capacity:

f y ˜ db
"b (1)
4 ˜ f bd

fy = Design yield strength of the vertical bar


db = Diameter of the vertical bar
fbd = Design bond strength of cast-in bar to concrete or of the adhesive mortar

The strut-and-tie model (Fig. 3) assumes that the compression strut C0 is anchored at the center of the
anchorage zone. For a drilled hole of depth tb and a concrete cover to the center of the upper
reinforcement of the base element of cs, the lever arm inside the node z0 is:

480
"b
z0 tb   cs (2)
2
The lever arm inside the node z0 is smaller than the lever arm of the slab z3. The tension in the upper slab
reinforcement in the node region, FS0, is higher than the tension calculated for the slab with z3; the tensile
resistance of the upper reinforcement must therefore be checked separately. The shallower the
embedment of the post-installed vertical bar is, the more the moment resistance of the slab in the node
region is reduced as compared to a node with hooked bar.
On the left hand side of the anchorage zone shown in Fig. 2b, the compression force is continuing through
the struts C2t and C2c to the tension and compression zones of the slab and the equilibrium of the
horizontal forces is given. The vertical components of C2t and C2c are taken up by tensile stresses in the
concrete. Normally there is no vertical reinforcement in the slab to take up the tension force S. The loads
and thermal solicitations of a slab do not lead to horizontal cracking; therefore it is possible to attribute the
tension force S to the tensile capacity of the concrete. On the safe side, the maximum splitting stress has
been taken as that caused by a concentrated load C0 on the center of the anchorage zone. It has been
shown that the occurring maximum splitting stress maxҏ sp can be calculated as (4):

max V sp
§ V  V3 ˜ z1 · ˜ §1  z0 · ˜ §1  " b · ˜ § 2.42 · d f
¨ M1  2 ¸ ¨ ¸ ¨ ¸ ¨ ¸ (3)
2 z ¹ © 2˜ z ¹ © b˜ z2 ¹
ct
© ¹ ©
M1, V2, and V3 are external forces on the node (refer to Fig. 2b), z1 is the inner lever arm of wall section
outside node region (z1 = 0.9·d), and b is the width of the wall section. If the calculated maximum splitting
stress is smaller than the tensile strength of the concrete fct, then the base plate can take up the splitting
forces without any additional shear reinforcement.

Test Programme
At the American University of Beirut, eighteen specimens were tested in five sets (8) (table 1). Each set
contained three to four specimens. The variables considered in the study included: bar size, nominal
concrete compressive strength, installment history (pre-installed or post-installed), anchorage type of pre-
installed bars (hooked or straight), depth of embedment, and the Hilti product used in the post-installing
procedure. To evaluate the effect of the two different Hilti products on the bond behavior of post-installed
reinforcing bars relative to pre-installed hooked or straight bars in each set, the bar size, the nominal
concrete compressive strength, and the embedment depth were kept constant.

Design of the Specimens


The 18 test specimens are identified in table 1. The 70-cm deep test specimen consisted of two identical
30x75 cm vertical elements anchored in a 170x35 cm base. The specimen simulated the rigid connection
of two walls to a slab. The 70-cm depth allowed the placement of 7 bars on each face of the two vertical
elements and the horizontal base element. Transverse reinforcement was placed in all elements. In all 18
specimens, the longitudinal reinforcement of the
20 30 base consisted of two layers of seven 20-mm
70 30 20
bars each. The clear cover to the longitudinal
bars in all elements was 3 cm. The test set-up
allowed the application of two equal compressive
2(7I12/14) loads below the tip of both identical vertical
75 elements subjecting them to combined flexure
and shear, and consequently subjecting the
reinforcement on the inner face of the elements
2(7I20)
to tensile force. The inner face reinforcing bars
were either pre-installed in the base with hooked
hef
anchorage or pre-installed in the base with
35 straight anchorage or post-installed in the base
after the base was cast using Hilti HIT injection
system.
The nominal concrete compressive strength in
170
Set 1 was 20 MPa. The vertical reinforcement on
Fig. 4: test specimen dimensions either face of the two vertical elements consisted

481
of seven 12-mm bars. The variable was the anchorage method of the inner face tensile bars of the vertical
elements: pre-installed with hooks, pre-installed straight, post-installed using Hilti HY 150, or post-installed
using Hilti RE 500 and the anchorage depth.
To check the effect of short straight anchorage on the relative behavior of pre-cast and post-installed bars,
the three specimens of Set 3 were tested. The nominal concrete compressive strength was 20 MPa and
the bar size of the vertical elements was 12 mm. A short anchorage of 15 cm was provided. The three
specimens of Set 5 were identical to those of Set 3 with the exception of using 14-mm bars in the vertical
elements instead of 12-mm bars.

Materials
In all five sets, the reinforcing bars of each size were from the same heat of steel and had parallel
deformation pattern and identical rib geometry. The bars met ASTM specifications and were Grade 60.
The average tested yield stresses were 557 MPa for the 12-mm bars, 555 MPa for the 14-mm bars, and
592 MPa for the 20-mm bars.
Two non air-entrained concrete mixes were used to obtain nominal concrete strength values of 15 and 20
MPa.

Test procedure
A compressive force was applied at around 16-cm from the tip of the two identical cantilever vertical
elements. The force was applied using two hydraulic rams having a capacity of 300 kN each and operated
by a single hydraulic pump. The compressive force was transferred uniformly to the faces of the vertical
concrete elements by means of two steel beams. The load, applied by the hydraulic rams, was monitored
by an electronic pressure transducer and was measured at the pump by a pressure gauge. To monitor the
load-deflection history of the specimen, four Linear Variable Differential Transducers (LVDT’s) were
mounted.

Test Results
The following table shows the relevant details of the specific test specimens as well as the measured test
data.
strength at day
length hef [cm]

elements ǻmax
ultimate load

Deflection of
Tensile bars

of testing f'c
Anchorage

Anchorage

Measured
of vertical
elements

Concrete

Pmax [kN]
Test no.

method
Set no.

vertical
[MPa]

[mm]

1 25 7 Ø 12 Cast-in, bend 19.4 169.09 29.97


2 25 7 Ø 12 Cast-in straight 28.9 139.77 17.04
1
3 25 7 Ø 12 HIT HY 150 26.1 156.44 23.87
4 25 7 Ø 12 HIT RE 500 26.1 164.22 33.39
5 29 7 Ø 12 Cast-in, bend 11.7 155.43 32.37
6 29 7 Ø 12 Cast-in straight 14.0 129.41 15.00
2 7 29 7 Ø 12 HIT HY 150 15.3 133.80 23.68
8 29 7 Ø 12 HIT RE 500 14.4 149.36 29.69
9 15 7 Ø 12 Cast-in straight 25.9 67.66 5.79
3 10 15 7 Ø 12 HIT HY 150 25.8 76.92 9.36
11 15 7 Ø 12 HIT RE 500 32.5 105.16 18.21
12 29 7 Ø 14 Cast-in, bend 17.0 237.87 26.50
13 29 7 Ø 14 Cast-in straight 17.6 176.78 18.23
4
14 29 7 Ø 14 HIT HY 150 19.6 178.60 34.17
15 29 7 Ø 14 HIT RE 500 21.6 201.21 48.83
16 15 7 Ø 14 Cast-in straight 31.8 87.53 5.25
5 17 15 7 Ø 14 HIT HY 150 32.5 104.22 22.90
18 15 7 Ø 14 HIT RE 500 19.6 95.75 29.40
Table 1: Test Results

482
Test Evaluation
Hamad et al. (9) have analyzed the test results considering the design provisions of ACI 318. The
following section is a re-evaluation of the test results with respect to the strut-and-tie model shown in
section 2.
According to section 3.2 the yield strength of the wall inner reinforcement (7 bars) Fy is 441kN for the I12
bars and 598kN for the I14 bars. For the reduced lever arm z1R Kupfer (4) recommends a reduction factor
of 0.85. With a static height of the wall section of 270mm the lever arm to be considered is 243mm. The
lever arm of the external force with respect to the slab’s longitudinal top reinforcement y is 600mm.
Assuming that the connecting reinforcement reaches its yield capacity at failure load, the external load
Vreq reqired to yield the connecting reinforcement is
y
Vreq Fy ˜ (4)
z1 R
The lever arm in the node z0 has been evaluated from the bond strengths of the adhesive mortars and of
cast in straight bars according to formula (2); for hooked bars the entire anchorage length was considered
as lever arm z0. The angle T between the inclined compressive strut D0 and the horizontal direction can
then be calculated as
z0
tgT (5)
z1R

Figure 5 shows the ratio of the maximum


1.4 measured horizontal force of the cylinder Pmax
1.2 to the force Vreq required to reach steel yield
according to the strut-and-tie model of section
1 set 1
2 (formula (4)) against the angle T according to
P max /Vreq [-]

0.8 set 2 formula (5). It clearly appears that formula (4) is


set 3
0.6 a very accurate prediction for angles T larger
set 4
than 30º to 35º, but that the formula
0.4 set 5
overpredicts the test results by about 100% for
0.2 angles T smaller than 30º.
0
Figure 6 shows specimen 8 (post-installed,
0 20 40 60
hef=290mm, T=44º) after the test. The crack
angle T [degs]
pattern corresponds to the flow of forces shown
in figure 3. Indeed, the average of Pmax/Vreq for
Fig. 5: test data compared to valued predicted by
test sets 1, 2 and 4 is 0.98 with a coefficient of
strut-and-tie model
variation of 10.7%.

Fig. 6: specimen 8, hef=290mm Fig. 7: specimen 11, hef=150mm

483
On the other hand, the failure shown in figure 7 for specimen 11 (post-installed, hef=150mm, T=16º) is a
concrete cone pulled out by the connection bars; this reminds of a shear failure of the concrete. In the
following the strength of the specimens with short embedment is calculated from the shear capacity of the
concrete. Eurocode 2 defines the shear strength of a section without shear reinforcement in sect. 4.3.2.3
by:
VRd 1 [W Rd ˜ k ˜ (1.2  40 ˜ U1 ] ˜ bw ˜ d (6)

WRd with fck = fc’-8, WRd= 0.24 MPa (specs 9, 10) and 0.30 MPa specs 11, 16, 17 and 18
k size factor, =1.6-d. d is assumed 0.11m
bw depth of specimen =700mm
U1 longitudinal reinforcement ratio. With hef=150mm, bw=700mm, 7bars I20 -> U1=0.02

With a safety factor for concrete of 1.5 and assuming that the characteristic resistance is about 70% of the
average resistance of a concrete structure, the expected ultimate shear capacity is

1.5
Vu VRd 1 ˜ (7)
0.7
Vreq Assuming that the capacity of the specimens with short embedment is
defined by the shear strength in two planes (figure 8), the cylinder load
Vreq required to reach shear capacity is, similar to formula (4):

y
Vreq 2Vu ˜ (8)
z1 R
VR,u VR,u
For test sets 3 and 5 the average of Pmax/Vreq is 0.96 with a coefficient of
Fig. 8: shear design variation of 8.8%. This shows that a shear design approach is a good
approximation in the case of short anchorage lengths.

Conclusions
The evaluation of the test series described in section 3 has shown that moment resisting frame node
connections with post-installed straight reinforcement bars can be safely designed using the proposed
strut-and-tie model. The deviations between theoretically predicted and measured load capacities are
small.
The tests also show that such connections with short anchorage lengths do not have a plastic type of
failure, but the failure is comparable to a brittle shear failure. Therefore, post-installed, moment resisting
frame node connections should be made with relatively long anchorage lengths, so that the flow of forces
can develop as assumed in the strut-and-tie model.

References

1. DIBt: Bewehrungsanschluss mit Hilti Injektionsmörtel HIT-HY 150, Allgemeine


bauaufsichtliche Zulassung Z 21.8 -1648
2. Pr EN 1992-1 part 1-1 (Eurocode 2): Planung von Stahlbeton- und Spannbetontragwerken.
3. Building Code Requirements for Structural Concrete (ACI 318-05). American Concrete
Institute, 2005
4. Kupfer, H., Münger, F., Kunz, J., Jähring, A.: Nachträglich verankerte gerade
Bewehrungsstäbe bei Rahmenknoten. Bauingenieur 1/2003. Springer Verlag.

484
5. Hilti Fastening Technology Manual B2.11. Hilti Technical Service, issue 2005
6. Borgstede, D: Diplomarbeit am Lehrstuhl für Massivbau, Prof. Zilch, TU München, 2002
7. Leonhardt, F.: Vorlesungen über Massivbau, Teil 2: Sonderfälle. Springer, Berlin 1981.
8. Jyengar K.T.S.R.: Two dimensional theories of anchorage zone stresses in post-tensioned
prestressed concrete beams. Journ. ACI. Proc. Vol. 59, Nr. 10, pp 1443 – 1466.
9. Hamad B.S et al.: Experimental Evaluation of the Bond Strength of Bonded-In Reinforcement
Using Hilti Injection Systems. ACI structural journal, V 103, No. 2, March – April 2006.

Acknowledgements
The strut-and-tie design model proposed in this paper has been developed in close cooperation with
Professor H. Kupfer from Technical University of Munich, Germany. The test series presented in the paper
has been carried out at the American University of Beirut under the direction of Professor B. S. Hamad.

485
486

S-ar putea să vă placă și