EARTHQUAKE ENGINEERING AND STRUCTURAL DYNAMICS, VOL.

25, 753-773 (1996)

EFFECT OF INFILLS O N THE GLOBAL BEHAVIOUR O F R/C

FRAMES: ENERGY CONSIDERATIONS FROM PSEUDODYNAMIC
TESTS

PAOLO NEGRO* AND GUlDO VERZELETTI'

ELSA Laboratory, Joint Research Centre of the European Commission, TP 480. I21020 Ispra (VA), Italy

SUMMARY
A series of pseudo-dynamic tests were conducted on a full-scale four-storey reinforced concrete building designed
according to Eurocodes 2 and 8. The building was 10 m long, 10 m wide, and 12.5 m high. It was designed as a ductility
class 'High' structure, for typical live loads and for a peak ground acceleration of 0.3 g and medium soil conditions.
A first test was conducted on the bare frame. The project was carried out within the framework of the European
Association of Structural Mechanics Laboratories (EASML), and was designed to assess the adequacy of the damage
indicators to be used in the calibration of Eurocode 8. The pseudodynamic test was conducted by using an artificially
generated earth-quake derived from a real earthquake (Friuli, 1976), with nominal acceleration 50 per cent larger than the
value adopted in design. The structure performed as expected. The pattern of the measured rotations was that of
a weak-beam, strongcolumn mechanism. The fundamental frequency of the structure after the test was found to be half
of the initial value, but the damage was limited and uniformly distributed.
A second experimental programme was conducted as part of the work of the Network Prenormative Research in
support of Eurocode 8, to study the influence of masonry infill panels on the global behaviour of the frame. Two
pseudodynamic tests were conducted, with different infill patterns. A test was performed by infilling the two external
frames with hollow brick masonry in all four storeys (uniform infill distribution). The test was then repeated on the
structure without inflls at the first storey, to create a soft-storey effect.The input signal was the same as in the tests on the
bare frame. The purpose of the tests was to study the effects of the different layouts of infills, as well as to calibrate the
computer models for the infills to be used in parametric analyses.
In this paper the test results are presented and the performances of the structure with different infill configurations are
compared. The globlal behaviour of the structure is compared with the predictions which could have been made with
simplified approaches. In particular, single degree of freedom energy concepts are used to verify if the differences in the
global behaviour could have been predicted. The differences in the single degree of freedom energy demands with respect
to the bare frame maly be used as a means of accounting for the presence of irregular distributions of non-structural infills
in the simplified design of the frame.

KEY WORDS: reinforcted concrete; infilled frames; regularity; experimental methods; energy

INTRODUCTION
Modern seismic codes neglect, o r take very limited account of, the effects of non-structural masonry panels.
In fact, the masonry panels strongly affect the behaviour of the main structure. In general, the presence of
non-structural masonry panels has a beneficial effect, because they significantly increase the global strength
of the structure. Om the other hand, they also increase the initial stiffness, so that the inertial forces may be
increased to a 1arg;eextent. The beneficial effect due to the increase of strength may o r may not counterbal-
ance the potentially negative effect due t o the global stiffening of the structure.
Computer models are available to conduct parametric studies on the effects of the infill panels. Generally,
phenomenologicall global models are used.' These models are of the equivalent diagonal strut type. They are

*Research engineer
'Senior research engineer, Laboratory Head

C C C 0098-8847/96/080753-2 1 Received 26 June I995

0 1996 by John 'Wiley & Sons, Ltd. Revised 26 February 1996
754 P. NEGRO AND G. VERZELETTI

simple and robust, but the calibration of the global properties is rather difficult. The experimental work
conducted to-date2 (generally on simple one-storey one-bay infilled subassemblages) does not provide data
for the calibration of the global models, since the basic properties of the material - such as the strength of
bricks and mortar, and of small wall specimens - are typically not reported.’ To fill this gap, a more refined
model is being developed at I ~ p r aThis . ~ includes 2-D smeared-crack elements for masonry and concrete, and
either a full-bond or unilateral frictionless condition at the infill-frame interface. By means of a monotonic
analysis, it is possible to calibrate the parameters required for the global models, starting from the basic
properties of the materials.
Monotonic analyses of the test structure have shown that the results are extremely sensitive to the modelling
assumptions. This confirms the need for the test to be conducted on the infilled frame, as well as the need for
continuing the refinement of the computer models, to include effects such as friction at the interface.
An important problem concerning the effects of infills is their distribution. Irregular arrangement in plan
and elevation may cause a significant concentration of damage in the frames, due to torsional effects or to the
formation of soft-storey mechanisms. After the first test on the fully infilled structure, a second test was
performed without infills in the first storey, to create a soft-storey effect with a local concentration of ductility
demands. A companion testing activity was devoted to studying the importance of the out-of-plane response
in the global b e h a v i o ~ r It
. ~ is believed that this experimental activity will allow the available computer
models to be validated and calibrated, so that extensive parametric analyses can be carried out. The results of
this study will assist in enabling the effects of the infills to be taken into account more realistically, something
which could lead to an improvement in the codes.

TESTS PERFORMED O N THE BARE FRAME

Design of the specimen

The general layout of the full-scale reinforced concrete test structure (Figure 1) was decided during the
EASML co-operative r e ~ e a r c h It
. ~ was a four-storey, high-ductility framed structure. Dimensions in plan
were 10 m x 10 m, measured from the column axis. Inter-storey heights were 3-0 m, except for the ground
storey (3.5 m). The structure was symmetric in one direction (that of testing), with two equal spans of 5.0 m,
while in the other direction it was slightly irregular due to the different span lengths (6.0 and 4.0 m).
All columns had a square cross-section with a 400 mm side, except for the interior column which was
450 mm x 450 mm. All beams had a rectangular cross-section, with a total height of 450 mm and a width of
300 mm. A solid slab with a thickness of 150 mm was adopted for all storeys.
The materials used for the specimens were normal-weight concrete C25/30 as specified by Eurocode 2,6
and B500 Tempcore (heat-treated steel with characteristic yielding strength 500 MPa), rebars and welded
meshes. The adoption of this kind of steel was decided on because the Tempcore process is gaining the
market in Europe. The B500 Tempcore steel does not fulfil the requirement for high-ductility structures given
either by the 1988 edition of EC8’ - used for the design of the test structure - or the version recently issued
as prenormative standard (ENV).’ The minimum characteristic value of the deformation at maximum stress
is fixed as 9 per cent in the ENV - this was prescribed in terms of strain at failure in the 1988 version
-whereas the actual value is 8.5 per cent. The lower bound of the mean value of the tensile strength to yield
strength ratio is fixed as 1.20 in the ENV - this was 1.30 in the 1988 version - whereas the actual value is
1.16. The results of the test were expected to provide information about the adequacy of this material for use
in earthquake resistant con~truction.~
The preliminary design was carried out in accordance with the prescriptions of Eurocodes 2 and 8,
assuming typical loads (additional dead load 2.0 kN/m2, to represent floor finishing and partitions, and live
load 2.0 kN/mZ),and high seismicity. The seismicity was defined by the elastic spectrum defined in EC8, with
peak ground acceleration of 0.3 g and medium soil conditions (soil type ‘B’ in EC8).
EC8 allows for the choice of three different ductility classes (Low, Medium and High), corresponding to
different values of the force reduction factor (q), and of different design methods according to the structural
regularity. The structure satisfied the regularity requirements both for vertical and horizontal configuration,
 EFFECT OF INFILLS ON THE GLOBAL BEHAVIOUR O F R/C FRAMES 755

II I/
II I I It

Direction of
Testing

,--I
Figure 1. Layout of the specimen (dimensions in metres)

so it was designed as a frame of high regularity in ductility class High, with the behaviour factor q = 5. The
design was performed using two independent planar models for each orthogonal direction, and the torsional
effects were taken into account by the simplified method prescribed by EC8.
The preliminary evaluation of the fundamental period yielded a value of T = 033 sec, which corresponds
for the soil type 13 to the flat region of the design spectrum. This yielded a total horizontal design force of
529 kN. The computed drifts, conventionally calculated according to the 1988 version of EC8, proved
smaller than the limit fixed for non-structural elements ‘fixed in such a way as not to interfere with structural
deformations’, but greater than the limit for non-structural elements ‘of brittle materials attached to the
structure’.
756 P. NEGRO AND G . VERZELETTI

All members were detailed in accordance to the rules of the 1988 version of EC8. As far as practical
detailing is concerned, it was decided to place the splicing of the bars at the columns of the ground floor in
correspondence to their mid-heights.This solution is not adopted in practice in most cases, but it is
considered by far the most rational and it allowed a simpler interpretation of the results. Details of the
reinforcement layout of the frame are reported elsewhere." The structure was loaded by means of two
double-acting, 500 kN hydraulic actuators per storey, plus one extra actuator placed orthogonally at the
third floor to prevent unwanted sway motion. Details about fixtures and control instrumentation are given in
Reference 10.

T h e high-level pseudodynamic test

An artificial accelerogram, derived from the 1976 Friuli earthquake, was adopted. The signal was
generated to fit the EC8 elastic response spectrum with a peak acceleration of 0 3 g . The reference signal is
shown in Figure 2, and the corresponding 5 per cent damping elastic response spectrum is given in Figure 3,
together with the spectrum specified by EC8 for 0.3 g and medium soil conditions. The test was performed

Ground acceleration [g]

0.4 , 1

0.3

0.2

0.1

0.0

-0.1

-0.2

-0.3
.OO 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00

Figure 2. Reference Friuli-like generated accelerogram

Spectral acceleration [g]

1.10
1.00
0.90
0.80
0.70
0.60
0.50
0.40
0.30
0.20
0.10
.OO .20 .40 .60 .80 1.00 1.20 1.40 1.60 1.80 2.00

Figure 3. Reference accelerogram elastic spectrum and EC8 elastic spectrum

 EFFECT O F INFILLS O N THE GLOBAL BEHAVIOUR OF RIC FRAMES 757

using the reference: signal multiplied by an intensity factor of 1.5. The corresponding EC8-equivalent peak
ground acceleration was then 1.5 x 0.3 = 0.45 g, a value which was thought to be representative of the
maximum seismic actions for which the frame had been designed, accounting for possible sources of
overstrength such as the overstrength from torsion which was considered in the design but not in the test.
This intensity levell is particularly meaningful in defining the damage resulting from the design-level seismic
actions. A quantitative assessment of the damage suffered by the structure will be possible only when
experimental data about the ultimate strength and ductility of the frame becomes available. To this aim,
a final test to failure will be performed after the other tests scheduled for this specimen. After that stage it will
be possible to draw definite conclusions about the performance of the structure. Post-test calculations must
also be performed to understand fully the behaviour of the structure. Data already available do, however,
allow some important conclusions to be stated.
The structure performed very well. During the test, cracks opened (and closed) in the critical regions
of the beams of the first three storeys and of most of the columns. Only the cracks at the beam-to-column
interfaces remained permanently open. Neither spalling of the cover nor local instabilities of reinforce-
ment was observed. Besides the cracks at the beam-to-column interface, which were apparent in the first
three storeys and represented evidence of local yielding in the rebars, the specimen remained apparently
undamaged.
A direct stiffness measurement was performed before and after the pseudodynamic test. With the stiffness
matrices, the vibraition eigenfrequencies were computed. The resulting fundamental frequency (0.82 Hz) was
less than half the size of that of the virgin structure (1.78 Hz). The resulting mode shapes, however, were close
to those of the virgin structure. The fact that the mode shapes did not significantly change, even though the
changes in the stiffness matrix of the specimen were significant, provide evidence that the structure was
uniformly damaged. The analysis of the maximum rotations measured at the potentially critical locations
(Figure 4) confirms the more or less uniform damage pattern (with the exception of the top storey), even
though no plastic hinges appeared to form in the beams of the external frame at the intersection with the
central column of the second storey.
Measurement of the rotations of the beams at different distances from the joint highlighted the important
role played by the slippage of the rebars in the internal joints. This effect was probably made more severe by
the adoption of the Tempcore steel. It resulted in significantly pinched forcedisplacement loops, and was
found to be responsible for most of the differences with respect to the preliminary calculation.' ' 3

Details on the lests performed on the bare frame can be found in Reference 10.

ek Internal frame External frame

Figure 4. Test on the bare frame: maximum measured rotations (max 14.9 mRad)
758 P. NEGRO AND G . VERZELETTI

TESTS ON THE INFILLED FRAME

Due to the relatively low damage suffered by the structure as a result of the bare-frame tests, no repair actions
were taken. The materials to be used for the construction of the infill panels were selected as representative of
typical light non-structural masonry. To facilitate the construction, blocks commonly available in Italy were
adopted. The blocks had dimensions of 245 x 112 x 190 (h)mm, with vertical holes taking 42 per cent of the
gross section. The average compressive strength of the blocks was found to be, respectively, 13.3 MPa in the
direction parallel to the holes (vertical), and 3.3 MPa in the direction orthogonal to the holes. The mortar
was aiso selected as typical, to reach a compressive strength of 5 MPa (cement, lime and sand in the
proportion 1 : 1 : 5).
To provide the data necessary to calibrate the global computer models, tests on small specimens were
performed at the University of Pavia to measure the resulting properties in the strong and weak directions, as
well as in the diagonal d i r e ~ t i 0 n . For
I ~ the strong (vertical) direction, a mean strength of 7.3 MPa and a mean
modulus of elasticity of 8210 MPa were obtained (values referring to the gross area). The compressive tests in
the direction orthogonal to the holes yielded a mean strength of 2.4 MPa and a modulus of elasticity of
2515 MPa. A mean value of the tensile strength of 0.28 MPa and a shear modulus of 1240 MPa were derived
from the results of the diagonal compression tests.

Unijormly infilled frame

A pseudodynamic test was conducted on the four-storey R/C frame with infills at all storeys (Figure 5) to
appraise its seismic response as compared to the bare-frame behaviour. No special provisions were made in
the construction of the panels, so that complete bond between the panels and the frame could be assumed at
the beginning of the test. No plaster was placed on the surfaces of the panels.
The loading apparatus was the same as in the test on the bare frame. Before the test, the direct
measurement of the stiffness matrix was carried out by displacing each of the storeys in turn by a small
quantity, while holding the others still. From the stiffness matrix, the corresponding elastic vibration
frequencies were computed. The fundamental frequency was found to be 3.34 Hz (for the bare frame this was
1-78Hz before the test, and 0.82 Hz after the high-level test). Due to the fragile behaviour of the mortar

Figure 5 . Uniformly infilled frame

 EFFECT OF INFILLS ON T H E GLOBAL BEHAVIOUR OF- RIC FRAMES 759

Figure 6. Soft-storey infilled frame

interface, no low-level tests were performed. A high-level test, with the same base motion as for the test on the
bare frame (0.45 g nominal base acceleration) was carried out.
The test led to the complete destruction of the masonry panels at the first and second storeys:The panels at
the third storey suffered extensive damage, and the ones at the upper storey remained almost intact.
Redundant measurements were taken at each panel up to failure, to assist in the calibration of the global
computer models to be used in the interpretation calculations. The global measured quantities will be
presented in the following sections and compared with those from the other tests.

Soft-storey frame
The masonry panels were demolished and replaced with new ones, leaving the first storey bare (Figure 6).
This was expected to lead to a soft-storey mechanism, and corresponds to the case of ‘drastic reduction if
infills in one or more storeys’ in Eurocode 8. For such cases, the code requires a local increase of the forces to
be used in design, as well as an increase in the portion of the columns of the ground floor to be detailed as for
critical regions. u p to the entire length of the column. However, none of these requirements was considered in
design.
The fundamental frequency from the stiffness measurements proved to be 1.6 Hz. The high-level test was
repeated as for the uniformly infilled frame.
As expected, the test resulted in a concentration of drift at the ground floor, and some damage was suffered
by the panels of the second storey only. As for the previous test, local measurements were taken at some of the
masonry panels. In addition, the rotations at the ends of the columns and beams of the ground and first floors
were monitored. The global results will be discussed in the next section.

DISCUSSION OF THE RESULTS

General considerations
The comparison of the results obtained in the tests with different infill configurations (Table I) is
particularly meaningful, because most of the design codes neglect the changes due to the presence of the
in fills.
760 P. NEGRO AND G. VERZELETTI

Table I. Summary of measured quantities from the three tests

~ ~~

Spectral Max top Max Max Max beam Max column

Frequency acceleration displacement base shear drift rotation rotation
Structure (Hz) (9) (mm) (VOIW (%I (mrad) (mrad)
~~ ~ ~

Bare frame 1.78 1.11 210 0.40 24 13.9 14.9

Uniformly infilled 3.30 1.16 80 062 1.1 - -
Soft storey 166 1.19 180 0.47 3.5 15.0 32.6

The measured storey displacements are shown in Figure 7. The maximum top displacement obtained for
the bare frame (about 210 mm) proved to be comparable with the one obtained for the soft-storey structure,
even though the spatial distribution of the storey drifts was obviously quite different. In spite of the almost
complete failure of the infill panels, the maximum top displacement experienced by the uniformly infilled
frame was less than two-fifths the size.
The time histories of the base shear obtained for the three tests are depicted in Figure 8. The maximum
base shear for the soft-storey structure is only slightly larger than that of the bare frame (1.45 MN). The
maximum base shear obtained for the uniformly infilled frame was almost 50 per cent larger than that of the
bare frame. It is worth mentioning that the spectral accelerations for the three structures were almost equal.
The storey-level hysteretic loops (Figure 9) provide some insight into the behaviour of the structure. The
storey shear vs. inter-storey drift diagrams for the bare frame exhibit stable dissipative loops, with amplitude
progressively decreasing from the first to the top level. Even though the loops take a pinched shape after the
first large-amplitude cycle - the main reason for this effect having been identified as the loss of bonding in
the internal joints - no strength deterioration occurs. For the case of the uniformly infilled frame, the
amplitude of the cycles also decreases with the storey level. However, severe strength and stiffness deteriora-
tion -typical of infill behaviour -can be noticed. In the case of the soft-storey frame, the energy dissipation
is largely limited to the ground floor, and the onset of pinching in the shape of the loops after the first
amplitude cycle is as evident as in the loops of the bare frame.
The inter-storey drift profiles are given in Figure 10. The values for the uniformly-infilled frame are less
than half of those of the bare frame. In the bare frame the largest value was found at the second storey instead
of the first storey, while the opposite applies to the tests on the infilled frame. For the soft-storey structure,
most of the interstorey drift took place at the bottom storey, where a value larger than 3.5 per cent was
achieved.

Damage sufSered in the various infll configurations

The testing activity on the bare frame was started to assess the applicability of various damage indicators.
In particular, the aim was to compare the effectiveness of the proposed mechanistic damage indices with
widely used damage indicators (such as maximum ductility, inter-storey drift, global stiffness degradation).
The assessment of the mechanistic damage indices will become possible only after the final cyclic test to
failure predicted for the bare frame, after the necessary repair. This will enable the ultimate displacement
capacities, as well as the sensitivity to cyclic damage, to be estimated, so that the equivalent monotonic
ultimate displacement can be determined.
Local rotation or curvature ductilities cannot be derived from the test data, since no internal force
measurements were made. Global storey ductilities can be derived from the global storey hysteretic loops,
but the comparison is of little significance. In the case of the uniformly infilled frame, the non-linearity
is substantially due to the damage in the masonry panels. In the case of the soft-storey infilled frame, the
global ductility demand at the bottom storey is much larger than in the bare frame. However, the two
quantities are not directly comparable as damage indicators, since that of the bare frame is associated

Figure 7. Measured storey displacements

EFFECT OF INFILLS ON THE GLOBAL BEHAVIOUR O F R/C FRAMES 76 1

X1.E2 Disp [mm]

2.00

-2.50 1 I
.00 .80 1.60 2.40 3.20 4.00 4.80 5.60 6.40 7.20 8.00
Bare frame
XI.E2 Disp lmml

1.50

1.00

.50

.oo
-.so
storey-4

i
storey-3
-1.50
+ storey-2
-2.00 storey-1
Time [see]
-2.50 I I
.W .80 1.60 2.40 3.20 4.00 4.80 5.60 6.40 7.20 8.00
Uniformly infilled frame
XI.E2 Disp [mml

2.00

1.50

1.00

.50

.Mi

-.50
storey-4
-1.00
O storey-3

-lSO t W 1 + storey2
-2.00 - - storey-]
Time [sec]

.OO .80 1.60 2.40 3.20 4.00 4.80 5.60 6.40 7.20 8.00
Soft-storey infilled frame
762 P. NEGRO AND G. VERZELETTI

Base SheariUNl
1.50
1

IP " 1
Time [secl
.I w
.00 .80 1.60 2.40 3.20 4.00 4.80 5.60 6.40 7.20 8.00
Bare frame
Base Shear[MNJ
2.50

2.00
1.50

1.00
.50
.00
-.50
-1.00

-1.50
-2.00
~. I
~

.OO .80 1.60 2.40 3.20 4.00 4.80 5.60 6.40 7.20 8.00
Uniformly infilled frame
Ease Shear[MN]
2.00, # 1

.00 .80 1.60 2.40 3.20 4.00 4.80 5.60 6.40 7.20 8.00
Soft-storey infilled frame
Figure 8. Time histories of base shear

with a strong-column, weak-beam mechanism, while that of the soft-storey infilled frame is associated
with a bottom-storey local sidesway mechanism, as can be seen from the maximum measured rotations
(Figure 11). Due to the different dissipative mechanisms activated in the tests, the global ductilities are not
related to the member-level ductilities in the same fashion.
 rn
1
1
rn
2
4i

1.50 1.50

1.20 - 120 -
m
.90 - .90 - m
z
.60 - .60 - z
.30 - .30 -
.00 - .MI -
-.30 - -.30 -
-.&I - -.60 -
-.90 - -.90 -
-1.20 - -1.20 -
Drip3 [mm] Drip4 [mml
-1.50 -1.50

Figure 9. Storey-level hysteretic loops: (A) Bare frame

 Shear1 [MN]
2.so , I 2.50 , Shear2 [MN] 1
2.00
IS0
1.00
.so
.00
-.so
-1.00
-IS0
-2.00 -2.00 1 Drift2 Imml
I - .
1
-7 Tn -2.50 I
-40.00 -32.00 -24.00 -16.00 -8.00 .W 8.00 16.00 24.00 32.00 40.00 40.00 -32.00 -24.00 -16.00 -8.00 .00 8.00 16.00 24.00 32.00 40.00

Shear3 [MN] Shear4 IMN]

2.50 I I P
2.00 <
t n
- J P
1.50 1.so N
m
1.00 - r
m
4
.so - A
.00 -
-SO - -.so
-1.00 - -1.00
-1.50
-1.50 -
-2.h
Dr@3 [ m m l
s/
-2.00
1
1 Drift4
- Imml
- . 1
-2.50 -2.50 I
40.00 -32.00 -24.00 -16.00 -8.00 .00 8.00 16.00 24.00 32.00 40.00 40.00 -32.00 -24.00 -16.00 -8.00 .OO 8.00 16.00 24.00 32.00 40.00

Figure 9. (B) uniformly infilled frame

EFFECT OF INFILLS ON THE GLOBAL BEHAVIOUR OF R/C FRAMES 765

1 l l I l I I I l

I I I I I I I I I L I I I I I I I I
i

4
E'
I.
$-
766 P. NEGRO AND G. VERZELETTI

Storey

soft-storey

' unifonninf

* bareframe

.OO SO 1.00 1.50 2.00 2.50 3.00 3.50 4.00

Figure 10. Maximum inter-storey drift profiles

Internal frame External frame

Figure 11. Test on the soft-storey infilled frame: maximum measured rotations (max 32.6 mRad)

It is difficult in this case to relate the maximum inter-storey drifts to the damage suffered by the structure.
The maximum inter-storey drifts measured during the tests (Figure 10) are 1.13 per cent for the uniformly
infilled frame, 2.4 per cent for the bare frame and 3-55 per cent for the soft-storey infilled frame. The largest
values, however, were found at the bottom storey for the infilled frame, and at the second storey for the bare
frame. It is interesting to note that the ratio of the maximum rotations measured at the member ends of the
 EFFECT OF INFILLS ON THE GLOBAL BEHAVIOUR OF RIC FRAMES 767

bottom storey for the tests on the infilled frame and on the soft-storey infilled frame is not comparable with
the ratio of the corresponding maximum drifts. This is further evidence for the different dissipative
mechanisms activated during the tests.
Stiffness degradation can be regarded as a damage indicator. The global stiffness matrices were evaluated
from the on-line mleasurements by using the procedure developed by Igarashi et a l l 4 This procedure is based
on a second-order updating of the initial stiffness matrix, similar to the BFGSI5 stiffness updating method
used in the solution of non-linear equations with quasi-Newton strategies.” The initial values of the stiffness
matrix were measured via direct stiffness measurement before the test, and the update procedure was applied
continuously by using the displacements and restoring forces measured during the test. The resulting time
histories of the vibration frequencies are given in Figure 12. The progressive deterioration of stiffness is
noticeable. Moreover, the final values of the frequencies are in agreement with the frequencies computed
from the stiffness matrix directly measured after the test. However, the effects of the failure of the infill panels
are not apparent.
Energy can also be used as an indicator of the expected damage.16 This will be addressed in the following
section.

Energy considerations
Absorbed energy from test results. The time histories of the total absorbed energy at each storey are
depicted in Figure 13 for the three tests. The total absorbed energy represents the sum of elastic strain energy
and hysteretic energy. However, it is apparent from the figure that the elastic energy represents a small
portion of the total absorbed energy.
From Figure 1.3 one can understand how the energy dissipation was distributed among the different
storeys. In the bare-frame structure, most of the energy dissipation took place in the first two storeys, with
more or less the same amount of dissipation. The energy dissipated at the third storey was still noticeable,
while the energy dissipated at the top storey was almost negligible. The effect of the infills in the uniformly
infilled structure was to reduce further the higher storeys’ contribution to energy dissipation, though they still
contributed significantly to the total energy dissipation. For the soft-storey infilled structure, almost all the
energy dissipation took place at the bottom storey.
The large-amplitude drops in the graphs correspond to the points where the individual masonry panels
failed. It is interesting to recall how this piece of information could not be found in the time histories of the
vibration frequencies.
In Figure 14 the total absorbed energies for the complete structure are given. Since no viscous damping
was specified during the test, the total absorbed energy is equal to the difference between the input energy and
the kinetic energy. The analysis of the input energy provides the opportunity to compare the test results with
the energy demands which could have been predicted by using simple SDOF considerations. The need to
check the validity of applying the results obtained by SDOF analysis of MDOFs by means of experiments on
whole multi-storey buildings has already been expressed.I6 If this comparison were satisfactory, energy
considerations could represent a simple way to account for the presence of the infills in design, for instance by
modifying the design forces according to the different input energy demands with respect to that of the bare
frame.
Housner’s assumption. A simple way of using SDOF calculations to predict the energy demand for an
MDOF structure is to use Housner’s as~umption.’~ According to this assumption, one can estimate the
energy demand as the maximum kinetic energy of the corresponding linear SDOF system. Since we have
already noticed that the spectral pseudo-acceleration is roughly the same for the three structures - approx-
imately 1.2 g for the high-level earthquake - one can easily compute the kinetic energy of the linear SDOF
system corresponding to each of the three structures. The effective masses were obtained from the eigenvec-
tors of the measured stiffness matrices. The energy demands obtained for the bare frame, the uniformly
infilled structure and the soft-storey infilled structure are, respectively, 169, 41 and 224 kJ. These values
compare unfavourably with the total absorbed energies at the end of the tests (Figure 14). This is further
evidence that the energy demands computed by using Housner’s assumption tend to be underestimated.
768 P. NEGRO AND G . VERZELETTI

25.00
22.50 - -
20.00 - -
17.50 - -
15.00 - -

Frequency tHzl
25.00
22.50
-

20.00

10.00 -

2.50
.OO
.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
Soft-storey infilled frame
Figure 12. Time histories of frequencies from tangent stiffness updates
EFFECT O F INFILLS ON THE GLOBAL BEHAVIOUR O F RIC FRAMES 769

X1.E2 Energy[k.I]
1.40
1
1.20

1.00

.80

.60
storey-4
.40 storey-3
storey-2
.20
storey-1
.00
.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
Bare frame
XI.E2 Energy[kJ]
1.40

.60 -
V
storey-4
.40 - 0
storey-3
+ storey-2
.20 -
x storey-1
.w
.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
Uniformly infilled frame
XI.E2 Energy[k.I]
4.00 , 1

* 3.50

3.00

2.50 1
2.00 -
1.50 - - storey-4

1.00 - - storey-3
+ storey-2
storey-1

.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00

Soft-storey infilled frame

Figure 13. Time histories of storey-level absorbed energy

770 P. NEGRO AND G. VERZELETTI

Absorbed Energy [kJ]

5.00

4.50

4.00

3.50

3.00

2.50

2.00

I .so
sof-storey

1.W
uniform inf

.so bare frame

.00
.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00

Figure 14. Time histories of total absorbed energy

Absolute us. relative energy. Uang and Bertero have shown that absolute input energy based on an elastic
perfectly plastic SDOF system can provide a good estimate of the energy demands for a multi-storey
building." Absolute energy is the energy evaluated with reference to the absolute system, i.e. including
energy terms related to ground motion, while relative energy is evaluated in the equivalent fixed-base system.
To clarify the differences between the two definitions, it is useful to recall the equation of motion for the
SDOF subjected to ground motion:
mu,+ cd +fs =0 (1)
where rn is the mass of the system, c the viscous damping coefficient,& the structural restoring force, u, the
total displacement of the mass and u the relative displacement of the mass with respect to the ground. By
introducing the notation u, = u + ug, where ug is the input ground motion, equation (1) becomes
mv+cd+&= -mu, (2)
By integrating equation (1) with respect to u, we obtain the 'absolute energy equation'
Ei = Ek 4- Ec + E, (3)
The first term on the right is the 'absolute kinetic energy'
rnd:
Ek =- (4)
2
the second term is the damping energy

(5)
 EFFECT OF INFILLS ON THE GLOBAL BEHAVIOUR OF R/C FRAMES

the third is the absorbed energy

s
E, = fsdu

and the term on the left is the ‘absolute input energy’

s
Ei= mij,dv,

Analogously, by integrating equation (2) with respect to u one obtains the ‘relative energy equation’
EI = EL + E, + E,
The second and third terms on the right remain unchanged with respect to the absolute energy equation. The
first term on the right represents the ‘relative kinetic energy’, i.e.

and the term on the left is the ‘relative input energy’:

It has been demonstrated that the relative input energy is practically independent of the ductility demand
and the damping, and the energy spectra for non-linear systems are very similar to those of linear system^.^
According to Uanig and Bertero the absolute input energy is also practically independent of the ductility
demand,16 except for the case of signals with extremely long duration. These facts should make the use of
energy concept a valuable tool for assessing the differences in the global behaviour of the same structure in
different infill Configurations.
The limitations (ofthe SDOF procedures in dealing with structures exhibiting soft-storey mechanisms have
already been reported.” However, the need to check the validity of applying the results obtained by SDOF
analysis to MDOFs by means of experiments on whole multi-storey buildings has also been expressed.16The
results of the tests performed on the frame with different infill configurations will be used for this purpose. If
SDOF energy conisiderations were sufficient to highlight the differences in the global behaviour of infilled
frames, the differences in the input energy demands with respect to the bare frame may represent a simple way
to account for the presence of irregular infill distributions in design.
A constant-ductility elastic perfectly plastic response spectrum was constructed by iterating on the
strength-reduction factor until the reference ductility was obtained, with a 1 per cent tolerance, and both
absolute input energy and relative input energy were computed according to the definitions given in
equations (7) and (10).This was done for a target ductility value of 5, which was the global ductility assumed
in design for the bare frame. The choice of the ductility value was not expected to yield large differences in the
input energy, due )tothe substantial independence of the input energy from the ductility claimed by different
author~.’’.’~For the same reason, zero damping was assumed in constructing the response spectrum. The
resulting input energies - absolute and relative - are given in Table 11, together with the input energies
computed for the elastic SDOF system, and they are compared with the measured input energies. The
measured input energies were computed for the four-DOF system as the sum of absorbed and - absolute or
relative - kinetic energy.
From Table 11, lit can be concluded that neither the relative nor the absolute SDOF input energy demands
reflect the measured input energy. The assumption that the input energies are independent of the ductility
level was not verified. The comparison is more satisfactory for the elastic SDOF energies. Absolute input
energies compare relatively better than relative input energies.
Even though the comparison of the SDOF analysis and the test results was not satisfactory, and further
studies are probably needed to clarify the importance of the ductility demand in the SDOF input energy, we
772 P. NEGRO AND G. VERZELETTI

Table 11. Computed values of absolute (Ei) and relative (Ef) input energy (kJ)
Effective SDOF p = 5 Elastic SDOF Measured
Frequency mass
Structure (Hz) (kgx lo3) Ei Ef Ei Ef Ei Ef

Bare frame 1.78 295 366 353 444 400 398 377
Uniformly infilled 3.30 275 164 117 269 184 303 258
Soft storey 1.66 337 401 379 552 509 536 511

can still conclude that the SDOF input energy was sufficient to highlight the differing global structural
behaviour for the different infill configurations. From the SDOF input energies one could in fact be
forewarned about the substantially different energy demands for the uniformly infilled frame and the
soft-storey infilled frame with respect to the bare frame.
Since no simple, yet accurate, methods exist to account for the distribution of non-structural infills in
moment resisting frames, a method based on the simple modification of the design forces according to the
differences in the SDOF energy demand with respect to the bare frame should be explored.

CONCLUSIONS
The global results of the pseudodynamic tests conducted on a full-scale four-storey reinforced concrete frame
designed according to Eurocodes 2 and 8 have been presented. The tests were conducted on the bare frame,
as well as on the frame with two different configurations of non-structural masonry infills.
The effects of the infills on the global behaviour of the structure have been discussed, and the response of
the structure in the three configurations has been compared with the predictions which could have been
made by means of simplified SDOF techniques. The following conclusions can be stated:
The presence of light non-structural masonry infills can change the response of the structure to a large
extent. The presence of a regular pattern of infills to a large extent prevents energy dissipation from
taking place in the frame (this was true in spite of the progressive complete failure of the panels at the
first two storeys). Irregularities in the panels result in unacceptably larger damage to the frame. In
general, the effect of the non-structural infill panels cannot be neglected in design.
The differences in the behaviour of the structure with different distributions of infiils can be captured by
means of simplified SDOF techniques based on energy considerations. The differences in the SDOF
energy demands with respect to the bare frame may be used as a means to account for the presence of
irregular distributions of non-structural infills in the simplified design of the frame.

ACKNOWLEDGEMENTS
The preliminary design and the tests on the bare frame were conducted in the framework of a joint research
project by the European Association of Structural Mechanics Laboratories. An invaluable contribution was
made by the members of the Reinforced Concrete Working Group of the Association, and by E.C. Carvalho,
Head of the group. The tests on the infilled frame were performed as part of the activities of the Network
Prenormative Research in support of Eurocode 8 (PRECS), funded by the European Commission under its
programme on Human Capital and Mobility. The contribution of M.N. Fardis, leader of the task group on
infilled frames, and of G.M. Calvi, co-ordinator of PREC8, was much appreciated.

REFERENCES
1. M. N. Fardis and G. M. Calvi, ‘Effects of infills on the global response of reinforced concrete frames’, Pror. 10th European c m j
earthquake eng., Vienna, 4, 2893-2598 (1984).
2. C E B Bulletin No. 220, ‘Behaviour and analysis of reinforced concrete structures under alternate action inducing inelastic response’,
Chapter 5: ‘Reinforced concrete infilled frames’ (to be published).
 EFFECT OF INFILLS O N T H E GLOBAL BEHAVIOUR O F R/C FRAMES 773

3. D. Combescure, P. Pegon and A. Anthoine, ‘Modelling of the in-plane behaviour of masonry infilled frames’, in A. S. Elnashai (ed.),
European Seismic Design Practice, Balkema, Rotterdam. 621-629, 1995.
4. P. Negro and C. A. Taylor, ‘Effect of infills on the global seismic behaviour of R/C frames: results of pseudodynamic and shaking
table tests’, 11th world conf earthquake eng., Acapulco, Mexico, 1996 (submitted).
5. E. C. Carvalho, ‘Cooperative research on the seismic response of reinforced concrete structures’, Contract no. 4504-91-10 ED ISP P,
Final Report, Lisbon, 1993.
6. Eurocode No. 2, ‘Common unified rules for concrete structures’, Report E U R 8848 E N , Commission of the European Communities,
1994.
7. Eurocode No. 8, ‘Structures in seismic regions - design’, Part 1: general and building, May 1988 edition, Report E U R 12266 E N ,
Commission of the European Communities, 1988.
8. Eurocode No. 8, ‘Design provisions for earthquake resistance of structures, Parts 1-3: general rules - Specific rules for various
materials and elements’, EN V 1988-1-3, CEN, Brussels, 1994.
9. M. Pipa and E. (3. Carvalho, ‘Reinforcing steel characteristics for earthquake resistant structures’, Proc. 10th European conf:
earthquake eng., Vienna, 4, 2887-2892 (1994).
10. P. Negro, G. Verzeletti, G. E. Magonette and A. V. Pinto, ‘Tests on a four-storey full-scale R/C frame designed according to
Eurocodes 8 and 2 preliminary report’, Report E U R 15879, European Commission, Joint Research Centre, Ispra, Italy, 1994.
11. A. V. Pinto, P. Negro, P. Pegon and A. Arede, ‘Analysis of the four-storey R/C building to be tested in the ELSA-reaction wall
facility’, Proc. 10th European conf earthquake eng., Vienna, 3, 2331-2336 (1994).
12. P. Negro, G. Verzeletti, G . Magonette and V. Renda, ‘Pseudo-dynamic testing of a four-storey full-scale reinforced concrete frame
building designed in accordance with Eurocodes 2 and 8’, Proc. 10th European con$ earthquake eng., Vienna, 3,2323-2329 (1994).
13. M. Calvi and S. Santini, ‘Preliminary tests on infill masonry’, P R E C 8 Progress Report, Pavia, Italy, 1994.
14. A. Igarashi, F. Seible and G. A. Hegemier, ‘Development of the pseudodynamic technique for testing a full scale five-storey shear
wall structure’, Proc. U S - J a p a n seminar on the development and future directions of structural testing techniques, Hawaii (1993).
15. R. Fletcher, Practical Methods of Optimization, Wiley, New York, 1987.
16. V. V. Bertero andl C.-M. Uang, ‘Issues and future directions in the use of an energy approach for seismic-resistant design of
structures’, in P. Fajfar and H. Krawinkler (eds), Nonlinear Seismic Analysis and Design qf Reinforced Concrete Buildings, Elsevier,
New York, 1992.
17. G. W. Housner, ‘Limit deisgn of structures to resist earthquake’, Proc. 1st world conf: earthquake eng., Berkeley, CA (1956).
18. C.-M. Uang and 11. V. Bertero, ‘Evaluation of seismic energy in structures’, Earthquake eny. struct. dyn. 19, 77-90 (1990).
19. T. Zahrah and W. J. Hall, ‘Earthquake energy absorption in S D O F structures’, J . struct. eng. 110, 1757-1772 (1984).
20. H. Krawinkler and A. A. Nassar, ‘Seismic design based on ductility and cumulative damage demands and capacities’, in P. Fajfar
and H. Krawinkler (eds), Nonlinear Seismic Analysis and Design of Reinforced Concrete Buildings, Elsevier, New York, 1992.