See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/340248495

Seismic assessment of a R/C building with setbacks using nonlinear static and
dynamic analysis procedures

Conference Paper · October 1998

CITATIONS READS

6 26

2 authors, including:

Andreas Kappos
Khalifa University
224 PUBLICATIONS   4,502 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Database for the seismic fragility assessment of as-built and retrofitted bridges using machine learning techniques View project

Upgrade of existing bridges View project

All content following this page was uploaded by Andreas Kappos on 28 March 2020.

The user has requested enhancement of the downloaded file.

 1
Seismic assessment of a R/C building with setbacks using nonlinear static and dynamic analysis
procedures

A.J. Kappos and S.G. Scott

Department of Civil Engineering, Imperial College, London, UK

ABSTRACT: The paper addresses multistorey reinforced concrete (R/C) frames with setbacks in the upper
part. Available experimental and field evidence is reviewed, followed by a case-study involving the design
and assessment of a ten-storey frame with two large setbacks (equal to about 40% the length of the storeys
below) at its top. The structure is designed to the provisions of Eurocode 8 (EC8) as a ductility class ‘M’
building irregular in elevation for a PGA of 0.25g, using multimodal analysis, as required by the code. It is
then assessed using both push-over analysis and dynamic time-history analysis for selected input motions
scaled to various intensities. It is found that the seismic performance of the structure is satisfactory, even for
motions twice as strong as the design earthquake.

1 INTRODUCTION increases in ductility demands in the region of the

setbacks.
The issue of irregularity is currently one of the most
controversial ones in seismic design. While it has
long been recognised that past building failures
during earthquakes had often as the main cause 2 EXPERIMENTAL AND FIELD EVIDENCE
irregular configurations in plan and/or in elevation,
it is not yet clear whether buildings designed and Observed damage in buildings after strong
detailed to the provisions of modern seismic codes, earthquakes indicates an inferior performance of
but not satisfying the currently accepted ‘regularity buildings with setbacks (Penelis and Kappos 1997).
criteria’, can still perform adequately (i.e. in a way In a detailed study of damage statistics after the
essentially similar to that of regular structures) 1978 Thessaloniki (Volvi) earthquake, Penelis et al.
during a strong earthquake. (1989) report that whereas the percentage of
The present paper addresses multistorey R/C damaged buildings was 16.6% in the case of regular
frames with setbacks, i.e. a reduction of the length elevations (no setbacks), it increased to 25.3% for
of the building along its height (irregularity in buildings with one setback, 28.4% for buildings
elevation). The paper focuses on buildings with with two setbacks, and 30.4% for buildings with
setbacks in the upper floors (not on tower-like three or more setbacks. No specific information is
structures where one large setback occurs in the available regarding the distribution of damage in the
lowest part of the building). The former setback buildings, in particular whether the area
architectural form is quite common in cases where a around the setbacks was the most heavily damaged.
relatively narrow road separates two multi-storey Some specific evidence is available from the
buildings, to permit adequate sunlight exposure of Alkyonides (Greece) 1981 earthquake, during which
the lower floors. a 6-storey R/C building in Athens with a 66%
Early analytical studies (Aranda 1984, Costa et setback at the top storey suffered failure of all
al. 1989) involving buildings with setbacks have columns of the setback, whereas the rest of the
raised some concern regarding the behaviour of this structure was almost intact (Spyropoulos, 1982).
class of structures, as they have indicated significant The first author has observed at least one similar
A.J.Kappos & S.G.Scott 2

case in the town of Loutraki, during the same of irregular structures is still open to discussion. It
earthquake. has to be pointed out that the foregoing tests
involved application of unidirectional earthquake
input, hence they did not address the problem of
3.0m
torsion in setback buildings.
In the following, a case-study is presented,
involving the design and assessment of a ten-storey
frame with two large setbacks (equal to about 40%
the length of the storeys below) at its top. The
structure is designed to the provisions of Eurocode 8
(EC8) for ductility class ‘M’ buildings irregular in
elevation, for a PGA of 0.25g. It is then assessed
using both push-over analysis and dynamic time-
history analysis for selected input motions scaled to
various intensities.

3 DESIGN OF THE STRUCTURE

The structure studied (Fig.1) has two large setbacks

equal to 6 and 4 m at the top two storeys, which
amount, respectively, to 38% and 40% of the
dimension of the storeys below. Therefore, they far
exceed the 10% limit prescribed by the EC8 (CEN
1994) and the structure is classified as irregular in
6.0m
elevation.
6.0m 4.0m
The structure has previously been designed as a
Figure 1: Geometry of frame. regular frame (same overall geometry, no setbacks)
according to the CEB (1985) Seismic Code and to
While the foregoing evidence indicates a rather
the EC8 (Kappos 1997). The same materials (C20
consistent trend, it has to be emphasised that most of
concrete and S400 steel), the same design ground
these buildings had been designed on the basis of
acceleration (Ad=0.25g), and the same ductility
rough hand-methods of static analysis, and also the
class (medium, ‘M’) were used.
level of detailing required by the then applicable
code was quite low, and poor construction practices It is pointed out that the behaviour factor (q) used
often made it even lower. to derive the design seismic actions is slightly
different in each case; for the regular frame
Recent experimental studies involved frames
structure q=3.0 for the CEB MC design, and q=3.75
with setbacks designed and detailed to modern
for the EC8 design, whereas for the irregular frame
codes, such as the American UBC. Both small-scale
q=3.0 is specified by EC8.
structures (Wood 1992), and 1/4 scale structures
(Shahrooz and Moehle 1990) were tested on the The key difference from the case of regular
shake table for earthquake intensities up to three buildings, in addition to the 20% lower behaviour
times the design one. Contrary to previous analytical factor, was the compulsory use of multimodal
and field evidence, the shake table tests have clearly analysis. For comparison purposes, the forces
indicated that irregular reinforced concrete (R/C) resulting by applying the simplified modal or
structures can indeed perform adequately, even equivalent lateral force (ELF) procedure of EC8,
when subjected to earthquakes significantly stronger which applies to regular structures only, were also
than the one they were designed for, and it has calculated.
consequently been argued that current code The first 4 modes of vibration were considered in
provisions regarding regularity tend to be the multimodal analysis. These had periods 1.06s,
overconservative. Furthermore, the issue of the 0.41s, 0.25s and 0.19s, with contributing masses
compulsory use of multimodal analysis in the design corresponding to 78.6%, 11.1%, 3.7% and 2.5%
A.J.Kappos & S.G.Scott 3

respectively of the total mass. The cumulative mass nevertheless, the responsibility of the designer is
represented by these 4 modes is 95.9% of the total, higher, since weaknesses in modelling, and (more
which safely exceeds the EC8 requirement of 90%. often) misinterpretation of the results are more
It is worth pointing out that the corresponding common in dynamic analysis.
periods for the regular frame were 0.98s, 0.37s,
0.22s, and 0.16s, which correspond to 85-92% the
values for the setback frame.
4 METHODOLOGY FOR ASSESSMENT OF
Fig. 2 compares the modal force and the SEISMIC PERFORMANCE
equivalent lateral force distribution along the height
of the building; it is seen that multimodal analysis
gives about 20% higher force at the top of the The seismic behaviour of the EC8 designed setback
frame, but equivalent lateral forces are higher in structure was then assessed using two analytical
most other storeys. The base shear in the ELF procedures:
procedure is 2% higher than that calculated from 1. Nonlinear static (‘push-over’) analysis, using
multimodal analysis. The foregoing comparisons a proportional loading pattern determined from the
imply that had the structure been designed using the (elastic) multimodal analysis.
(simpler) ELF procedure, the design action effects 2. Nonlinear dynamic (time-history) analysis,
would have been underestimated by up to 20% in using two earthquake records scaled to selected
the upper part (note that some members are fractions of the spectrum intensity of the design
controlled by minimum reinforcement earthquake (in this case the EC8 elastic spectrum for
requirements), whereas the lower part of the frame ag=0.25g).
would have been reinforced in essentially the same
Push-over analysis is indispensable for
way as in the case of multimodal analysis.
estimating the possible overstrength of a structure
and identifying possible weaknesses in some parts
10 of it (e.g. soft storeys).
9
8
7
6
storey
5
4 MODAL
3 ELF
2
1

0 20 40 60 80
4.1 Assessment using push-over analysis
storey shear (kN)

In this analysis lateral forces were applied at each

Figure 2: Comparison of storey shears given by the storey of the structure, the relative size of which was
equivalent lateral force method and the modal method proportional to the storey forces given by the
(first 4 modes). multimodal (elastic) analysis; forces were gradually
increased until the structure failed. In line with
previous studies (e.g. Kappos 1991) failure was
Several investigators (e.g. Shahrooz and Moehle taken to correspond to a 3% global drift (top
1990) have suggested appropriate force distributions displacement equal to 3% the total height). The
for setback buildings, to be used for ELF-based development of plastic hinges was tracked, which
design. However, these procedures tend to be case- permitted an assessment of the capacity design
specific and probably inconvenient to adopt in a procedure (i.e. ensuring the columns are stronger
code which has to be applicable to all types of than the beams, to avoid inelastic instability of the
buildings. Hence, for the time-being, it appears that building subsequent to the formation of plastic
multimodal analysis is indeed required, to achieve a hinges).
conservative design of all parts in a setback
building. This causes little practical problems, since The analysis of the structure was carried out
modal spectrum analysis options are currently using the program DRAIN-2DX (Prakash et al
available in most commercial programs; 1993). Standard point hinge modelling was used for
A.J.Kappos & S.G.Scott 4

R/C members, with elastoplastic behaviour of each 5 DISCUSSION OF RESULTS

hinge and proper account for (yield moment) - (axial
force) interaction in columns; the latter is an 5.1 Results from push-over analysis
important feature in the case of exterior columns,
where significant axial force variation occurs due to
the varying overturning moment. Further discussion Fig. 3 shows the (base shear)-(top displacement)
of the modelling assumptions may be found in curve calculated from the push-over analysis; it is
Kappos (1991, 1997). seen that the yield strength of the structure VC is
10% higher than the shear VB at the formation of the
first hinge (in a beam) and 53% higher than the
4.2 Assessment using time-history analysis design shear VA. Given that subsequent to yielding
the base shear continues to increase (although
This type of analysis was also carried out using probably not to the extent Fig. 3 suggests, since no
DRAIN-2DX (Prakash et al 1993), using essentially strength degradation was considered in the inelastic
the same model as previously, with elastoplastic model), it is clear that the frame possesses an
hysteresis constitutive laws governing the behaviour overstrength of 70% or more.
of each hinge.
The possibility of failure in each member, as well
as in each storey of the buildings, was checked by 800
applying appropriate local as well as global failure 700
VC
criteria. 600
VB
Local failure corresponds either to the 500
Base shear
exceedance of the available plastic rotation capacity (kN) 400 VA
of a R/C member, taking the adverse effect of high 300

shear stresses into account, or to the development of 200

a shear force exceeding the corresponding capacity 100

0
of the member; the latter is checked with respect to 0 0.2 0.4 0.6 0.8 1
all possible failure modes, that is diagonal tension, Displacement (m)
diagonal compression (web crushing) and sliding
shear. As the shear failure criterion is based on Figure 3: Base-shear against top displacement from
strength rather than on deformability, it is generally pushover analysis.
more conservative than the rotational capacity The first hinges form in the beams of the
criterion. Global failure is assumed to coincide with intermediate span. The spreading of plastic hinges
storey failure; a dual criterion based on a limiting along the height of the structure can be seen in Fig.
interstorey drift of 2% and the development of a 4. At the point of effective yield (point a) several
sidesway collapse mechanism involving all vertical beam hinges have formed, whereas column hinging
members has been adopted for assessing storey is restricted to locations specified by the Code (base
failure. More details on local and global failure and top storey), with the exception of one hinge in
criteria used may be found in Kappos (1991, 1997). the interior column of the first setback. At a total
Two input motions were used for the analysis, drift of 1% (point b) a few more column hinges
the well-known El Centro (1940) N-S record, which appear but no possibility of sidesway mechanism
facilitates comparisons with other studies, and the involving columns is detected anywhere in the
Thessaloniki (1978) N30E record, which had been structure. Even at a drift of 3% (point c) (well
found to produce most of the critical response beyond the limit of 1% implied by EC8) no column
quantities in the corresponding regular frame in a mechanism forms. It is interesting to note that more
previous study (Kappos 1997), where a larger column hinges form at the second storey than in the
number of records had been used. two top storeys (where the setbacks occur). It
appears therefore, that the code procedure, although
not precluding the formation of column hinges
(despite the capacity design used), does in fact
provide an adequate protection against the formation
of unfavourable collapse mechanisms.
A.J.Kappos & S.G.Scott 5


800 10

700
.
c
9
b.
600 .
a 8

500 7
Base shear 6
(kN) 400
300 Storey5 demand
4 supply
200
100 3

0 2
0 0.2 0.4 0.6 0.8 1 1
Displacement (m)
0
0 0.01 0.02 0.03 0.04
. . . Hinge rotation (rad)
. . ... .. . ... ..
. . . .. .. . . .. .. .
. .. .. ... . .. .. ... .. Figure 5: Required and available plastic hinge rotations
. . . .. ... .. . .. .. .. in the interior columns (for Aef=0.25g).
. . . . ... ... . .. ... ... .
.. . . .. ... ... . .. ... ... .
. .. . . ... .. .. . ... .. .. .
. . . . . .... .... . . .... .... . 10
. . . . .. .... ... . .. .... ... .. 9
. . . . . . . . . 8
(a) (b) (c)
7
6
Figure 4: development of plastic hinges during pushover
Storey5 demand
analysis.
4 supply
3
5.2 Results from time-history analysis 2
1
0
When the structure is subjected to the input motions 0 0.01 0.02 0.03 0.04
Hinge rotation (rad)
normalised to the design intensity (0.25g), column
hinges generally form at predetermined locations
(base and top of the structure), the only exception Figure 6: Required and available plastic hinge rotations
being the top of a 9th storey internal column, where, in the exterior beam supports (positive bending, for
however, the plastic rotation demand is very low Aef=0.25g).
(0.610-3 rads). In fact, as shown in Fig. 5, in all
cases the required column plastic rotations were
well within the available capacities. As pointed out
10
in previous studies (Kappos 1997) this might
9
indicate an over-design of the columns with respect
8
to confinement. 7
Beam hinging did occur, as expected, but, as 6
shown in Figures 6 and 7, plastic rotation demands Storey5 demand
were again lower than the available capacities. 4 supply

During the ‘serviceability’ earthquake (0.10g) no 3

2
plastic hinges occurred, either in the columns or the
1
beams. The interstorey drifts (Fig. 8) that occurred 0
during this analysis (the largest being 0.21% in the 0 0.01 0.02 0.03 0.04
Hinge rotation (rad)
third storey) are small enough to ensure no damage
in the non-structural elements.
Figure 7 Required and available plastic hinge rotations in
the interior beam supports (positive bending, for
Aef=0.25g).
A.J.Kappos & S.G.Scott 6

3%, depending on the formation of sidesway

10
9
SERVICEABILITY EQ mechanism). Importantly, the inelastic drifts instead
8 DESIGN EQ of concentrating in the area of the setbacks,
7 COLLAPSE EQ exhibited their peak values in the lower part of the
6 frame. Note that similar trends had been observed in
Storey5
the case of a setback frame tested by Wood (1992).
4
3 Moreover ductility demands under the ‘collapse
2 earthquake’ (0.50g) were within available
1
capacities. This earthquake caused the formation of
0
0 0.2 0.4 0.6 0.8 1 1.2 14 plastic hinges in the columns, 8 of which were in
Drift ratio (%)
the permitted places. Of the remaining 6, the largest
was only 0.005 rads (so small that it can be assumed
Figure 8: Maximum interstorey drifts that occurred in the that the columns would still be almost intact) and a
serviceability, design and collapse earthquakes collapse mechanism was avoided.
(Aef=0.10g, 0.25g and 0.50g respectively).
The shear strength requirements under the 0.5g
earthquake are shown in Figures 9 and 10 for the
most critical columns; it is worth pointing out that
10 hoops alone can carry the entire dynamic shear (no
9 demand need for ‘concrete contribution’), which again is a
8
7
supply sign of over-conservatism.
6
storey 5
4
3
2 10
1 9 regular
0 8 irregular
0 100 200 300 400 500 600 7
shear (kN)
6
Storey5
Figure 9: Required and available shear capacities in the 4
external columns during the collapse earthquake 3
2
(Aef=0.50g).
1
0
0 0.1 0.2 0.3 0.4 0.5 0.6
Drift ratio (%)
10
9 demand Figure 11: Comparison of maximum interstorey drifts in
8 supply the regular and irregular frames when subjected to the
7
design earthquake (Aef=0.25g).
6
storey 5
4
3
Shown in Figure 11 is a comparison of the
2
1
maximum interstorey drifts of the setback frame
0 with those of the regular frame (Kappos 1997),
0 200 400
shear (kN)
600 800 when they are both subjected to the same record (El
Centro) scaled to the design earthquake intensity. It
Figure 10: Required and available shear capacities in the can be seen that there is no significant difference
internal columns during the collapse earthquake between the performances of the two frames, as the
(Aef=0.50g). maximum drift is about the same (0.5%) for both
structures; the regular frame is characterised by a
somewhat more uniform distribution of drift along
As shown in Fig. 8, the maximum drifts under the height. The foregoing comparisons indicate that
the ‘survival’ earthquake (0.50g) were up to slightly the design of the irregular frame was appropriate.
more than 1%, well below the failure value (2% to
A.J.Kappos & S.G.Scott 7

Methodology. Earthquake Engineering and

6 CONCLUSIONS Structural Dynamics, 20(2):167-176.
Kappos A.J. 1997. A Comparative Assessment of R/C
Structures designed to the 1995 Eurocode 8 and the
A review of field evidence regarding R/C buildings 1985 CEB Seismic Code. Intern. Jnl. Structural
with setbacks at the top has indicated that they are Design of Tall Buildings, 6(1):59-83.
more heavily damaged than structures regular in Penelis G.G. and A.J. Kappos 1997. “Earthquake-
elevation. Although no conclusive evidence is resistant Concrete Structures”. E & FN SPON
available regarding the location of possible damage (Chapman & Hall), London (592 pp).
concentration, setback floors have suffered serious Penelis, G.G., D. Sarigiannis, E. Stavrakakis and K.C.
damage on some occasions. Stylianidis 1989. A statistical evaluation of damage to
The performance of the multi-storey setback buildings in the Thessaloniki, Greece earthquake of
June, 20, 1978. Proceedings of 9th World Conf. on
frame analysed can be considered as completely
Earthq. Engng. (Tokyo-Kyoto, Japan, Aug. 1988),
satisfactory. The capacity design approach Tokyo:Maruzen, VII:187-192.
succeeded in ensuring failure in the beams and not Prakash, V., G.H. Powell and F.C. Filippou 1993.
in the columns, thus avoiding the formation of a DRAIN-2DX User Guide. SESM Section, Dept. Civ.
collapse mechanism. Engineering, UC Berkeley.
The fact that during the serviceability earthquake Shahrooz, B.M. and J.P. Moehle 1990. Seismic response
all the members remained elastic and the interstorey and design of setback buildings. Journal of Structural
drifts were small enough to avoid damage to non- Engineering, ASCE, 116(5):1423-1439.
structural elements, fulfils the EC8 requirement. Spyropoulos, P.J. 1982. Report on the Greek earthquakes
Although for the design earthquake minor of February 24-25, 1981. Concrete International
column hinging occurred in places violating the (ACI), 2:11-15.
EC8 requirements, it was so small that it could be Wood, S.L. 1992. Seismic response of R/C frames with
ignored from a strength-degradation point of view. irregular profiles. Journal of Structural Engineering,
The interstorey drifts were small enough not to ASCE, 118(2):545-566.
cause any significant damage to infill panels.
Peak interstorey drift values were essentially the
same as for a corresponding frame without setbacks,
and drift distribution indicated that damage did not
tend to concentrate in the setback area.
It appears that the EC8-prescribed procedure for
the design of setback buildings is an appropriate
one, although it can be argued that the detailing of
columns can be made less conservatively.

REFERENCES

Aranda, G.R. 1984. Ductility demands for R/C frames

irregular in elevation. Proceedings of 8th World Conf.
on Earthq. Engng. (San Francisco, Calif., July 1984),
IV:559-566.
CEN Techn. Comm. 250/SC8 1994. Eurocode 8: Design
Provisions for Earthquake Resistance of Structures -
Part 1: General Rules (ENV 1998-1-1/2/3). CEN,
Brussels.
Costa, A.G., C.S. Oliveira, and R.T. Duarte, 1989.
Influence of vertical irregularities on seismic response
of buildings. Proceedings of 9th World Conf. on
Earthq. Engng., (Tokyo-Kyoto, Japan, Aug. 1988),
Tokyo:Maruzen, V:491-496.
Kappos A.J. 1991. Analytical Prediction of the Collapse
Earthquake for R/C Buildings: Suggested

View publication stats