Eighteenth Postgraduate Student Conference on MSc Dissertations 2011-12

Department of Civil & Structural Engineering, University of Sheffield, 2012

Seismic Strengthening of RC structures

Pramod Neupane
MSc(Eng) Structural Engineering

ABSTRACT There is a global need for seismic strengthening in many existing structures. This is
because most structures do not meet the seismic demand according to latest seismic design codes.
Various strengthening techniques have been used to rehabilitate structures for a long period of time,
however they have major shortcomings like high cost, they are time consuming and most of them
require specialised application techniques. This paper focusses on seismic strengthening of reinforced
concrete structures by Post-tensioned metal strap (PTMS) technique. By strengthening beam-column
joints and critical elements using this technique, a considerable gain in strength and ductility of the
whole structure can be achieved. To evaluate seismic response and overall structural capacity, the
results from a bare and strengthened frame tested under the BANDIT project were analysed using
DRAIN-3DX analysis. The analytical models were analysed under varying peak ground accelerations
of 0.05g to 0.35g. From several analytical results and their correlation with experimental results from
past researches, strengthening by PTMS technique was found to be efficient in increasing overall
structural strength, stiffness and ductility of the RC structure.

techniques used nowadays showed that there is a

1 INTRODUCTION
need for an efficient, low cost and easy
Structures in the past were designed according to strengthening technique to rehabilitate existing
the codes available then, which considered no or structures.
very low lateral load resistance capacity. As a
result, gravity load designed structures undergo
2 POST-TENSIONED METAL STRAP
large deformations in the event of an earthquake
(PTMS) TECHNIQUE
due to limited ductility (Ghobarah, 2000). From
the past experiences and damages caused by To mitigate the effects of structural deficiencies
earthquakes, it is noted that existing reinforced and to strengthen them to withstand seismic
concrete (RC) structures at high seismic zones loads, researchers at the University of Sheffield
need rehabilitation and seismic upgrade is a invented post-tensioned metal strap (PTMS)
necessity (Bracci et. al, 1997). technique. In this technique, metal straps used in
packaging industries are used to confine critical
The brittle failure mode of beam-column joint
beam-column joints. The tightened metal straps
leads to the reduction of overall ductility in the
prevent premature crushing of concrete and also
structure. From past earthquakes, common
increase overall strength and ductility of the
failure modes observed in RC structures were
structure by the contribution of its individual
due to inadequate anchorage length, short lap
ductility. Metal straps are clamped into position
splices, strong beam-weak column mechanism
by using strapping equipment which apply stress
and buckling of flexural reinforcements. These
in the straps and simultaneously clip them into
defects in RC structures could lead to
position by special clips. These metal straps have
progressive collapse during seismic events;
high tensile strength and can be strengthened up
however these failures can be prevented by
to 30% of their yield stress. To ensure the
implementing local strengthening techniques.
utilisation of the confining strap, an effective
Several strengthening techniques are used in lateral stress is applied on the column preceding
practice like epoxy repair, removal and to any additional loading, delivering active
replacement, concrete jacketing, steel jacketing confinement (Frangou, 1996). The low cost of
and application of fibre reinforced polymers. straps and easy operation techniques make this
However, most of them are time consuming and strengthening technique practical and it is
costly and may interrupt the buildings use. These preferred over other prevailing strengthening
inadequacies in the prevailing strengthening techniques.
2.1 Confinement Model
The confinement model chosen for this research
is generalised as follows for the determination of
confined concrete strength.
fcc = f (fco, fle) (1)
Where, fcc and fco are strengths of confined and
plain concrete respectively. fle is the confinement
induced effective lateral pressure at the point of
peak stress of confined concrete.

Figure 2-2: Schematic stress-strain curve of

active confined concrete (Moghaddam et al.,
2010)
The critical, yield and ultimate points in the
curve are defined by stress-strain equations as
Figure 2-1: Relation between lateral pressure and given in table below.
the stress of confining element at peak
(Moghaddam et al., 2010) Table 2-1: Stress-strain formulation
The effective lateral pressure of confined Point Stress (Mpa) Strain
concrete is computed by considering the Critical fcr = 0.85 fcc
equilibrium of shape of the column and spacing
of the transverse reinforcements. From the Yield
fiigure 2-1, applied lateral pressure is obtained
as: Ultimate ( - ) 1.3

(2) Following the confinement modelling, the stress-

Where, As and fs are the sectional area and stress strain curve was obtained for confined concrete
of the confining element at peak, D is the section. The confined curve shows enhanced
diameter of the confined core and S is centres stress-strain curve to that of unconfined concrete.
spacing between two adjacent transverse
reinforcement.

The volumetric ratio, is defined as

(3) ; so the above equation can be re-arranged:

(4)

The geometrical effectiveness coefficient, Ke

reflects the effectiveness of the longitudinal and
transverse reinforcement. This coefficient is
calculated by the ratio between effectively Figure 2-3: Confinement modelling for the
confined areas to the gross sectional area of the BANDIT frame
column. Ke is computed using (Saatcioglu and
Razvi, 1992) approach as:
3 EXPERIMENTAL TESTS
√( ) ( ) ( ) (5)
The analyses were performed on a full scale RC
frame experimented under the BANDIT project.
Where, bc is width of concrete core; s is spacing The frame had structural deficiencies especially
between ties; wi is distance between two laterally in the beam-column joint with deficient
supported longitudinal bars. anchorage length in the beam bottom bar and
Thus, the confinement induced effective lateral deficient lap length in the base of column’s
pressure at peak stress point of confined concrete longitudinal reinforcement bars. Structural
is calculated as; deficiencies were introduced in the frame to
portray the substandard construction practices.
fle = Ke .f1 (6) Structural deficiencies lowers overall structural
Figure 2-2 shows the comparison between performance. Figure 3-1 shows the BANDIT
confinement level gained by confined concrete frame.
section to that of unconfined concrete.
Figure 3-1: The BANDIT frame (Garcia et al.,
Figure 3-2: Analytical model
2012)
3.2.2 Fibre cross section modelling
3.1 Geometry of the frame
Every component in Element 15 was modelled
Reinforced concrete frame tested in the BANDIT
as fibres and segments. Columns were divided
project was a one-bay two-storey moment
into three segments (10%, 80% and 10%)
resisting frame regular in plan 4.26 x 4.26 m
longitudinally, while beams were treated as
with constant storey height of 3.30 m. Columns
single element. Each cross section of the element
had cross section of 260 x 260 mm with eight
was divided into sub sections of steel and
numbers of 14 mm bars in the first floor and four
concrete fibres. Each fibre cross section was
14 mm bars in the second floor along the
assigned by specifying the centre coordinates for
perimeter of the columns with 30 mm concrete
each fibre on the local axes of the element and
cover. 6 mm diameter bars were used as shear
area of each fibre.
reinforcement at 200 mm centres for columns.
Beams were of different cross sections in X and 3.2.3 Frame Damping
Y axes. In X-axis, the cross section of the beam
Drain-3DX accounts for the effect of linear
was 260 x 400 mm and 260 x 300 mm in the Y-
damping through Rayleigh damping method
axis. The flexural reinforcement used in beams
which states that the damping is proportional to
was 4 nos.-14 mm diameter bars at top and 4nos-
14 mm bars at bottom. For beam’s transverse the mass and stiffness of the structure.
reinforcement, 8 mm diameter reinforcement c = α1m+ α0k (7)
bars were used, spaced at 300 mm centres. To Where, m is the mass, α0 is the mass damping
simulate the live loads in the frame, additional coefficient, k is the stiffness and α1 is the element
masses of 13.5 tonnes and 11.0 tonnes were stiffness coefficient. The damping ratio of the
added in the first and second floor respectively. mode' n' is calculated as:
3.2 Analytical frame modelling ξi n = (8)
A 2D frame was created in the finite element
software DRAIN-3DX to model the frame. The Here, ξi n is the damping ratio for mode n and
geometry of the frame was divided into is the natural angular frequency for mode n.
elemental groups containing beam elements, ω is given by ω = 2пf; where f is the modal
column elements and beam-column joint frequency.
elements altogether summing to eight groups. A The coefficients α0 and α1 can be calculated from
sample analytical model is shown in figure 3-2. the corresponding damping ratios ξi and ξj of the
ith and jth mode. Therefore, for these modes the
3.2.1 Element modelling matrix form of equation becomes:
Element type 15 from the element library of
Drain-3DX was used to model various [ ] { } { } (9)
elementary responses for cross section of a
beam, column and beam-column joint. This
element type considers the deformable parts of Solving this equation gives the values of α0 and
the element which are sub-divided into number α1 . Since, first and second modal frequencies
of segments. The response are monitored at the govern the response, only two damping values
centre cross section in each segment. The cross are considered.
section properties are supposed to be uniform (10)
within each segment but can vary for different
segments (Powell and Campbell, 1994). (11)
The damping ratio for the second mode should
be lower than that of first mode, so they were
taken as 3% for the first mode and 2% for the
second mode, based on engineering judgement.

4 ANALYSIS RESULTS:
Analyses were performed both on bare and Figure 4-2: At 0.15g PGA
PTMS strengthened frame to study the variance
in structural response before and after Correlation between analytical and experimental
strengthening. Pushover analysis and time results at higher PGA level signifies that
history analyses were performed to evaluate the deficient structures are more vulnerable in higher
seismic response of the test frames. level of seismicity. From the time history
analyses, DRAIN-3DX has been able to predict
4.1 The BANDIT frame (bare) the structural responses of the deficient frame
The BANDIT frame had structural deficiencies under varying PGA levels.
including deficient anchorage length and splice 4.1.2 Modal frequencies
length. The structural response of a deficient
frame at different peak ground accelerations First and second modal frequencies from
(PGA) and effects of deficiencies in the overall analysis were compared to the dynamic response
response was studied. To study the effects of of experimented BANDIT bare frame. Table 4-1
deficiencies in the structure, the frame was first below shows the comparison of modal
analysed without any deficiency and analysis frequencies.
was continued followed by step by step Table 4-1 : Modal frequencies computation
introduction of deficiencies in the structure.
First mode f1 (Hz) First mode f1 (Hz)
Condition Δf1 Δf1
4.1.1 Nonlinear Time History Analysis Experiment Analysis Experiment Analysis
Initial 2.09 2.09 - 5.57 5.57 -
The nonlinear time history analysis was After PGA = 0.05g 1.68 1.7 -19% 4.64 4.5 -19%
performed considering the shaking table test data After PGA = 0.10g 1.46 1.44 -31% 4.04 3.87 -31%
obtained at different PGA levels of 0.05g, 0.10g, After PGA = 0.15g 1.29 1.17 -44% 3.69 3.31 -41%
0.15g during the BANDIT experiment. The
The analysis followed the frequency loss as
deficient frame showed high level of
obtained from the experimental result. 44% loss
deformations and could not withstand the
of first modal frequency was observed after the
seismic load of more than 0.15g PGA, so the
third test of PGA level 0.15g and 41% loss was
experiment was stopped after this excitation
observed in second modal frequency. Deficient
level.
frame showed significant loss of stiffness in the
For the analytical studies, the displacement bare frame after this PGA level.
response of top and first floor nodes were plotted
4.1.3 Static Pushover Analysis
against each level of seismic excitations and
these analytical results were compared with the Static pushover analysis was performed
experimental data obtained from full scale test. following the triangular distribution loading
To compare the time history analysis results, pattern (2:1; top floor : first floor). Capacity
some analytical comparison of top floor curves were obtained for all three cases for bare
displacement time history with experimental data frame (non-deficit bare frame, frame with
at specific PGA levels are shown: anchorage deficiency, frame with anchorage and
splice length deficiencies)

Figure 4-1: At 0.10g PGA

Slight variations in the analytical and
experimental results were observed at this PGA
level, however overall structural response was Figure 4-3: Pushover curves for BANDIT frame
predicted by the analysis.
The graph shows pushover curves for bare frame
with and without considering deficiencies. With
the introduction of deficiencies, the frame
showed overall reduction of base shear capacity to predict the structural response of strengthened
along with stiffness. The deficient frame showed frame under varying peak ground accelerations.
base shear capacity of 51.16 kN at the yield point
4.2.2 Pushover Analysis
which progressed to collapse with the formation
of soft storey mechanism in the first floor at 73.5 A pushover analysis was performed to compute
kN. The capacity curve showed the dominance the capacity of the PTMS strengthened frame.
of deficient anchorage length over deficient The base shear capacity vs. displacement curve
splice length in the frame. Overall ductility showed the yield point at 68.30 kN, however the
factor was reduced by 38% in the bare frame due collapse point was not observed in the
to anchorage and splice deficiency in the frame. strengthened frame at considered triangular
loading condition. The reason for this is the
4.2 PTMS strengthened BANDIT frame
limitation of DRAIN-3DX which could not show
High strength metal straps of dimension 0.8 mm deformations in the frame until 400 mm control
x 25 mm were used to confine the beam-column displacement considered for the analysis.
joints and base of the column. Beam-column
4.2.3 Modal frequencies
joints were strengthened by metal straps
confining in the transverse direction, followed by The modal frequencies for the strengthened
stretching straps in longitudinal direction to frame were computed and compared with the
provide orthogonal confinement grid. experimental results obtained from full scale test.
4.2.1 Time History Analysis Table 4-2: Modal frequencies
The modal analyses were conducted for Condition
First mode f1 (Hz)
Δf1
Second mode f2 (Hz)
Δf2
Experiment Analysis Experiment Analysis
increasing level of peak ground accelerations Initial 1.65 1.65 NA 5.01 5.01 NA
(PGA) from 0.05g to 0.35g and compared with After PGA = 0.05g 1.56 1.53 -7.3% 4.66 4.62 -7.8%
After PGA = 0.10g 1.5 1.49 -9.7% 4.59 4.57 -8.8%
experimental results. Top floor displacement After PGA = 0.15g 1.47 1.47 -10.9% 4.33 4.32 -13.8%
records for strengthened frame at specific PGA After PGA = 0.25g 1.28 1.26 -23.6% 3.9 3.8 -24.2%
After PGA = 0.35g 0.99 0.97 -41.2% 3.7 3.6 -28.1%
levels are shown:
Obtained analytical result showed 41% loss of
first modal frequency after the analytical test of
0.35g PGA. The second modal frequency
showed 28% reduction in initial frequency.

5 COMPARISON OF RESULTS
5.1 Comparison of Static Pushover Analyses
Figure 4-4: At 0.10g PGA
Static pushover analysis was performed in the
test frame before and after the application of
PTMS and compared to study the variance in
structural performance. The result showed an
improved performance of the frame after
strengthening as shown in figure 5-1.

Figure 4-5: At 0.15g PGA

PTMS strengthened frame could withstand
seismic excitation of 0.35g PGA without any
significant deformation.

Figure 5-1: Pushover curves comparison

From the comparison of pushover curves for
Figure 4-6: At 0.35g PGA bare and strengthened frame, result showed
Observations from the time history analysis significant enhancement in overall base shear
showed proper correlation of displacement time capacity after applying PTMS strengthening
histories between experimental and analytical technique. The base shear force at yielding point
results. So, DRAIN-3DX analysis has been able in the bare frame was 51.16 kN, which increased
to 68.30 kN for the strengthened frame. This The modal frequency loss at 0.15g PGA in the
showed 33.5% increment in base shear capacity bare frame was obtained to be equal to that of the
of the frame at yield point. Also, it was observed strengthened frame at PGA level of 0.35g. The
from the graph that strengthened frame gained comparison showed significant reduction of
considerable stiffness by PTMS intervention. modal frequency loss in the strengthened frame.
This shows gain in structural stiffness in the
5.2 Comparison of maximum floor
frame with the application of PTMS
displacements
strengthening technique.
From the displacement time history analyses of
bare and strengthened frame, a comparison was 6 CONCLUSIONS
made to study the displacement control induced
by strengthening technique considered. Table 5-1 The structural response of RC structures under
below shows the first and second floor maximum seismic loading were studied. These studies were
displacements obtained from analytical and mainly focussed on studying the performance of
experimental results performed on bare and the Post-Tensioned Metal Strap (PTMS)
PTMS strengthened BANDIT frame. strengthening technique. Non-linear time history
and static pushover analyses were carried out on
Table 5-1: Maximum displacements (mm) the analytical model created considering the
PTMS strengthened structural layout of experimented BANDIT
Bare frame
PGA level Floor no.
building, (mm) frame. The analytical and experimental results
Experiment Analysis Experiment Analysis showed a proper correlation. This suggested the
0.05g
2 17.5 16.1 20.5 20.76 reliability of analysis approach.
1 10.5 10.86 13.3 13.78

0.10g
2 44.7 45 52 48.72 From the comparison between analytical results
1 24.8 23.61 29.4 26.99
2 81.8 74.7 78.9 66.5
of seismic response of the bare and strengthened
0.15g
1 31.3 35.2 41.6 40.85 frame in terms of various parameters as
0.25g
2 - - 125.9 110.23 discussed in chapter 5, PTMS strengthening
1 - - 60.7 58.67
2 - - 162.3 146.57
technique showed effective gain in overall
0.35g
1 - 75.3 78.8 structural strength, stiffness and ductility. This
was achieved by an increased confinement level
The bare frame could undergo maximum
gained by the application of post-tensioned metal
displacement of 81.8 mm at the PGA level of
straps in the beam-column joints. Thus, seismic
0.15g and the experiment was stopped due to
strengthening can be achieved efficiently by
severe cracks in the second floor joint in the bare
PTMS strengthening technique.
frame. But after the application of strengthening [

technique, the frame showed a very well

response with maximum displacement of 162.30 ACKNOWLEDGEMENTS
mm in the second floor. These records were well I gratefully acknowledge professor Kypros
identified by the analysis, as can be seen in the Pilakoutas for his valuable comments and
table. The analytical result from the second floor constructive criticisms during dissertation.
displacement at PGA level 0.15g in strengthened Research scholars Reyes Garcia Lopez and
frame showed 11% reduction in maximum Yasser Helal deserve immense thanks for
displacement than that obtained in bare frame. guiding me and providing necessary data
The first floor displacement showed an average required for the dissertation.
increment of 25% displacement while the second
floor displacement was almost same between REFERENCES
bare and strengthened frame. This shows damage Ghobarah, Ahmed, 2000. Seismic assessment of
reduction in the second floor by PTMS existing RC structures. Progress in Structural
strengthening. Engineering and Materials, 2(1), pp.60–71.
5.3 Comparison of modal frequencies Frangou, M., 1996. Strengthening of concrete by
lateral confinement. The University of Sheffield.
From table 4-1 and 4-2, the first and second
modal frequencies were compared between the Garcia, R. et al., 2012. Seismic Strengthening of
frames before and after strengthening. The result Deficient RC Buildings Using Post-Tensioned
showed significant reduction of frequency loss in Metal Straps : An experimental Investigation. 15
the strengthened frame. At 0.15g PGA level, the WCEE.
bare frame showed the frequency loss of 44% in Moghaddam, Hasan, Samadi, Maysam &
the first mode, while it was reduced to 10.9% in Pilakoutas, Kypros, 2010. Compressive behavior
the strengthened frame. Similar response was of concrete actively confined by metal strips,
observed in the second modal frequency where part B: analysis. Materials and Structures,
bare frame showed 33.8%, this was reduced to 43(10), pp.1383–1396.
13.8% in the strengthened frame at PGA 0.15g.