External Prestressing Concrete Columns With Fibrous Composite Belts
External Prestressing Concrete Columns With Fibrous Composite Belts
External Prestressing Concrete Columns With Fibrous Composite Belts
1631
FRPRCS-7 1633
occurs before the FRP sheet is fully utilized. Thereby, it becomes all too natural to think
of prestressing the FRP wraps before applying it as a confining device, hoping that a
more efficient use of expensive composites can be achieved. By doing so, the
confinement is no longer limited to passive mechanism but includes the active
confinement, as well.
On the other hand, reported experimental results show that similar levels of lateral
pressure for similar types of concrete do not yield similar performance. Therefore, it has
been suggested that concrete be referred to as a restraint-sensitive material, rather than a
pressure-sensitive material. This approach pronounces the lateral stiffness of the
confining device as a very important factor in enhancing the seismic performance of
concrete columns. In fact, the confining pressure is dependent on the dilation
characteristics of concrete, which in turn are affected by the lateral stiffness of the
confining device.
Based on the above-mentioned two paragraphs, a question may be raised, as follows:
what would be the difference between seismic performances of the two concretes which
are confined by fibrous materials with two different levels of initial lateral pressure (say
zero, and also, some prestressing value) while keeping the lateral stiffness of the two
cases the same as each other? In other words, what is the effect of active confinement
provided by prestressing of FRP in comparison with that of passive confinement? This
paper is trying to shed more light on this point through an experimental program.
The idea of seismic retrofitting concrete columns by external prestressing has also
been applied to high-strength steel by several researchers 1, 2, 3. Some researchers adopted
lateral pre-tensioning of FRP using resin injections under pressure or using an expansive
agent to apply pressure on the jacket 4, 5. The method proposed by the authors adopts a
different technique, since it uses a simple prefabricated steel device to manually prestress
the FRP. Moreover, the technique is applied to the FRP belts instead of continuous
sheets.
RESEARCH SIGNIFICANCE
This paper presents results of an experimental study on passive and active
(prestressed) composite systems for improved seismic performance of concrete columns.
Considering that the prestressed passive confinement system is relatively new, there is a
need to conduct research for addressing the related design concerns of external
prestressing concrete columns. Among these are, the effects of different initial confining
pressure, yet the same lateral stiffness of the confining device, on deformation ductility of
retrofitted columns. This issue has been investigated by conducting cyclic loading test on
five square columns which were retrofitted by carbon and aramid/epoxy belts.
FRPRCS-7 1635
which crossbar can be put. It should be explained that, herein, crossbar refers to a
metallic device with a threaded hole at its both ends. When the two-ply belt is wound
around the column, its both ends can be clamped together by putting a couple of
crossbars into the end eye-hooks, and then, passing bolt through the threaded hole of the
crossbar. Then, prestressing can be given to the belts by manually screw driving the bolts
of crossbars. The implementation of the prescribed prestressing method is simple, yet
effective in arriving at high prestressing levels, and can be applied on site without using
heavy machinery.
Before wrapping the column with FRP belts, the corners of square section were
rounded up to 2.5 cm to avoid stress concentration, and also, to generate corners hoop
stresses. After rounding, the corner concrete surface was well prepared by grinding and
primer so as to provide smooth surface to facilitate prestressing of the belts. It should be
explained that although the larger corner radius is more desirable for non-circular
sections to increase the area of effectively confined concrete, the amount of concrete
cover brings a limit to the amount of corner radius. Considering our specimens were
scaled down by a geometrical scale factor of 2.4 and regarding the existing cover
concrete, the 2.5-cm corner radius was a maximum value in practice. The prestressing
was gradually increased by fastening the bolts of prestressing device, and continued up to
either about a sixth of tensile strength of fibers for carbon/epoxy belts or about a third of
fiber strength for aramid/epoxy belts. The strain in the belts was monitored by strain
gauges pasted on the surface of the belts. For the level of prestressing which was used in
this study, the relaxation during several days after applying the initial prestressing was
negligible. The intervals of belts were selected as 7.5 cm, which was enough to
accommodate the prestressing device (i.e. couplers) between two successive belts. The
position of couplers was changed for every other belt, as indicated in Figure 2. For each
specimen five belts were used, and were distributed within the bottom part of the column
with about height of 37.5 cm, which was equal to 1.5 times the dimension of cross
section.
For carbon/epoxy belts used in this study, the prestressing level of about a sixth of the
ultimate strain seemed to be a maximum value which could be achieved practically on
site due to the fact that higher prestressing values might cause some local damage to the
fibers during the prestressing procedure and might facilitate the fibers rupture during the
cyclic test. It should be noted that in the proposed prestressing technique, the belts are in
direct contact with the concrete surface and also the prestressing is applied before the
belts are impregnated with the epoxy resin. The concrete surface was well prepared by
grinding and using primer before wrapping with the belts so as to provide a smooth
surface during prestressing procedure. However, the direct contact between surface
aggregates and belts, especially at the higher prestressing levels and consequently under
the resulting higher interacting pressure between concrete and fibers, may cause some
local damage to the fibers. On the other hand, for the aramid/epoxy belts (with twisted
aramid fiber bundles), the higher prestressing levels may be achieved, for instance, the
prestressing level of about a third of tensile strength could be easily applied in the present
paper. In terms of handling the belts and applying the prestressing in the way which has
been suggested in here (i.e. forming a two-ply belt from a loop, wrapping around the
FRPRCS-7 1637
distributed evenly along the bottom 37.5-cm of the column. This column was retrofitted
by use of non-prestressed carbon belts, as can be implied from /0 appeared in the last
part of the specimen name. Thereby, the only available external confinement for this
column was passive one. Comparing with the original column, R04M1-C/0 showed a
much better response in terms of increased maximum lateral strength, and more
importantly, the considerable deformation ductility, as indicated in Figure 3. The
specimen could maintain its lateral capacity up to about R= 7.0% with no drop. Although
R04M1-C/0 experienced the first diagonal crack at early stages, i.e. about R= 1.0% that
was almost the same stage as for the original column R04M1-0, diagonal cracks did not
propagate thanks to the induced external lateral pressure by carbon/epoxy belts. In other
words, the retrofitting scheme could stop the drop in the lateral capacity and provided a
desirable ductile flexural response. The cyclic test continued until one of the belts (the
second belt from the bottom) failed at about R= 9.0%. After this stage and at R= 10%,
another belt failed and longitudinal steel bars buckled because of losing their cover
concrete and lack of any confining device. Although the corners of the cross section had
been rounded up to 2.5 cm before wrapping the column with carbon/epoxy belts, the
failure of the belts was initiated at the corners, indicating the relatively small corner
radius and the concentrated lateral pressure at those critical areas of non-circular sections.
The column R04M1-C/6 was the same as R04M1-C/0 except that its carbon/epoxy
belts were prestressed up to about a sixth of the ultimate strain of the carbon fibers, that
was about 2500 microns (see the Table 2), and could be implied from /6 which
appeared in the last part of the specimen name. Because of the prestressing which was
introduced to the belts in R04M1-C/6, the external confinement in this test column was
not limited to only passive confinement but included the active confinement as well. With
reference to Figure 3, it can be concluded that the peak strength and deformation ductility
for R04M1-C/6 did not show considerable difference from those of R04M1-C/0. In other
words, an increase in the initial lateral pressure can not lead in considerable upgrading of
seismic performance unless the lateral stiffness of confining device is increased. The
above-mentioned observation is important as it helps us understand the initial confining
pressure can be appreciated if it is accompanied with some increase in lateral stiffness at
the early stages of cyclic loading.
The observed damage in terms of cracks at the final stages of cyclic loading test was
less in some parts in R04M1-C/6 comparing with the non-prestressed version, i.e. the test
column R04M1-C/0. This limited level of damage can be explained by the difference
between the actively and passively confined concretes. That is, active confinement can
limit widening of the cracks once they appear. On the other hand, very high level of
prestressing would not necessarily limit the damage level. This is because when the
cracks appear and propagates, a certain amount of pre-tensioning would be useful so as to
avoid widening the cracks; however, the pre-tensioning values higher than that may cause
the two adjacent sides of cracks to be overlapped. Overlapping the two sides of a crack
under too much prestressing values can cause some sort of additional damage. In this
capacity, as we are concerned with earthquake-damaged concrete columns, where the
cracks propagated after the quake, external prestressing can provide a quick solution in
limiting the level of further damage under aftershocks or even future earthquakes.
FRPRCS-7 1639
explained for carbon series, implied while keeping the lateral stiffness of the confining
device constant, the prestressing effects would be very limited. This conclusion can be
verified by some other observations. For instance, the past research carried out by the
authors 6, 7, 8 demonstrated that by locating steel angles between aramid/epoxy belts and
concrete surface at the corners of square section, the prestressing of belts could lead in
remarkable increase in shear strength, and more importantly, in ductility of the columns.
For the case of the past research with steel corner angles, the aramid/epoxy belts were
used, and for the reasons which mentioned earlier, higher prestressing levels could be
utilized. However, the main source of difference between the results comes from the
combination of high prestressing level with some definite increase in lateral stiffness
provided by the steel corner angles. It should be noted that for square sections, in which
the confinement effects work mainly through the corners of the cross section by arching
action, it would be good enough if the increase in lateral stiffness is provided only at the
corners of cross section. In this regard, it would be interesting to see (Figures 3 and 4)
that all four columns retrofitted by either carbon or aramid/epoxy belts, and whether the
belts prestressed or not, show a similar deformation ductility. It has been known,
however, that passive confinement can not be appreciated in terms of enhancing flexural
strength under low axial compression. The present research, through dealing with the
retrofitted columns which developed flexural response, addressed, in more details, the
effects of active confinement on lateral strength and, more importantly, on the
deformation ductility of the retrofitted columns.
Variation of strain in carbon and aramid/epoxy belts
Figure 5 demonstrates the strain variation in one of the belts located in the bottom part
of the four retrofitted test columns. For the columns which were retrofitted by nonprestressed carbon or aramid/epoxy belts, lateral strain at low drift angles could not get a
considerable increase. This phenomenon could be predicted as passive confinement could
be mobilized only after the concrete dilation became a relatively high value. This would
be considered as a problem for passive confinement which could not avoid the initiation
of cracks (i.e. diagonal ones) at early stages of cyclic lateral loading, as explained earlier.
On the other hand, from Figure 5 it is seen that the maximum lateral strain observed in
the belts of R04M1-C/0 during cyclic loading test could not go beyond about 7000
microns. Also for R04M1-A/0, except for the asymmetric variation of strain at larger
drifts due to the sudden spalling of the concrete located between the two first belts in one
of the web sides of the column from R=9%, the strains were less than 7000 microns. It
was, however, not expected to arrive at the same value mentioned in Table 2 as for tensile
strength of fibers because in this case the belts followed the perimeter of a square section,
and also, they were in direct contact with the concrete surface. But another issue which
mainly limited the increase in the belt strain was related to dilation tendency of the
concrete. The strain of 0.7% as a limit for passive confinement provided by FRP sheets
had been already reported 9. Considering the above-mentioned observations, it became all
too natural to think of introducing a prestress to the belts so that the initial lateral strain
could get an appreciable value at the early stages of cyclic loading, and on the other hand,
a better use from the whole strength capacity of composites could be made. However, as
mentioned earlier, the increase in initial confining pressure should be combined with
FRPRCS-7 1641
REFERENCES
1. Yamakawa, T., Kamogawa, S., and Kurashige, M., Seismic Performance and Design
of RC Columns Retrofitted by PC Bar Prestressing as External Hoops, Journal of
Structural and Construction Engineering, AIJ, Japan, No. 537, 2000, pp. 107-113.
2. Saatcioglu, M., and Yalcin, C., External Prestressing Concrete Columns for Improved
Seismic Shear Resistance, Journal of Structural Engineering, ASCE, V. 129, No. 8,
2003, pp. 1057-1070.
3. Gamble, W. L., Hawkins, N. M., and Kaspar, I. I., Seismic Retrofitting Experience
and Experiments in Illinois, Proceedings of 5th National Workshop on Bridge Research
in Progress, National Center for Earthquake Engineering Research (NCEER), State
University of New York at Buffalo, Buffalo, New York, 1996, pp. 245-250.
4. Saadatmanesh, H., Wrapping with Composite Materials, Non-Metallic (FRP)
Reinforcement for Concrete Structures, 1995, 1(1), pp. 593-600.
5. Priestly, M. J. N., and Seible, F., Design and Seismic Retrofit Measures for Concrete
and Masonry Structures, Construction and Building Materials, V. 9, No. 6, 1995, pp.
365-377.
6. Yamakawa, T., Nasrollahzadeh Nesheli, K., and Satoh, H., Seismic or Emergency
Retrofit of RC Short Columns by use of Prestressed Aramid Fiber Belts as External
Hoops, Journal of Structural and Construction Engineering, AIJ, Japan, No. 550, 2001,
pp. 135-141.
7. Nasrollahzadeh Nesheli, K., Yamakawa, T., Satoh, H., and Inaba, H., Experimental
Investigation of RC Short Columns Retrofitted by Prestressed Aramid Fiber Belts as
External Transverse Reinforcement, Proceedings of Japan Concrete Institute (JCI),
Japan, V. 23, No. 1, 2001, pp. 961-966.
8. Nasrollahzadeh Nesheli, K., Yamakawa, T., and Satoh, H., Experimental Study on
Retrofitting of Shear Critical RC Columns using Pre-tensioned Aramid Fiber Belts,
Proceedings of 1st Conference on Application of FRP Composites in Construction and
Rehabilitation of Structures, edited by Nasrollahzadeh Nesheli, K., Building and Housing
Research Center (BHRC), Tehran, 2004, pp. 95-104.
9. Building Disaster Prevention Association of Japan (BDPAJ)., Design Guidelines for
Retrofit, Repair and Construction for Reinforced Concrete and Steel Reinforced Concrete
Buildings utilizing Continuous Fiber Reinforced Material, 1999. (in Japanese)
FRPRCS-7 1643
Figure 3 Measured lateral strength (V) versus drift angle (R) for carbon series.
FRPRCS-7 1645
Figure 4 Measured lateral strength (V) versus drift angle (R) for aramid series.