Bulletin of Earthquake Engineering (2006) 4:1–21 © Springer 2006

DOI 10.1007/s10518-005-5407-9

The Seismic Shear Demand in Ductile Cantilever

Wall Systems and the EC8 Provisions

A. RUTENBERG∗ and E. NSIERI

Department of Civil and Environmental Engineering, Technion-Israel Institute of Technology,
Haifa, 32000, Israel
∗
Corresponding author. E-mail: avrut@tx.technion.ac.il

Received 15 March 2005; accepted 25 september 2005

Abstract. The seismic shear provisions of EC8 for ductile reinforced concrete walls, serv-
ing as the lateral load resisting system in multistorey building structures are re-examined.
Two aspects are considered: (1) single walls, or a system comprising a number of equal-
length walls, (2) a resisting system comprising walls of different lengths. It is demonstrated,
in light of recent parametric studies, that the EC8 provisions for walls in the medium- and
high-ductility classes (DC-M and DC-H, respectively) are both in need of revision. Possible
revisions of requirements and a design procedure for a wall system are presented.
Key words: ductile walls, EC8, higher modes effects, seismic shear

1. Introduction
In EC2 (CEN, 1992) as well as in other reinforced concrete codes it is tac-
itly assumed that the shear demand at the base of a flexural wall and its
distribution over the building height are obtained from the lateral force dis-
tribution leading to the design base moment. Whereas this is acceptable for
wind loading, it is not conservative for seismic design due to the effects of
higher vibration modes, both in the linear and post-elastic states. The EC8
provisions (CEN, 2004) take account of this known phenomenon, but as
is shown subsequently, they are in need of revision because they do not
predict the response faithfully for the medium- and high-ductility classes
(DC-M and DC-H respectively). As is well known, it is important to pre-
vent premature shear failure; hence a realistic estimate of the shear demand
is necessary. This need also applies to the shear demand distribution along
the building height. Moreover, the code specifies (Section 5.4.2.4) that shear
forces be redistributed among the resisting walls as the bending moments.
This shear strength allocation is shown to underestimate the shear demand
on the shorter walls, which are usually the more flexible ones.
First, the seismic shear demand on an isolated wall, or on a lateral force
resisting system comprising several walls all having the same length, as
obtained from parametric studies on two time-history suites, is compared
2 A. RUTENBERG AND E. NSIERI

with the EC8 provisions for the two ductility classes. This is followed by
a pushover procedure to estimate the shear acting at the base of each wall
in a system comprising walls of different lengths. The results of this proce-
dure are compared with nonlinear time history analyses, demonstrating the
adequacy of the pushover results.

2. Effect of higher vibration modes

The distribution of the lateral forces over the height of the structure spec-
ified by building codes for the equivalent lateral force procedure is usually
an inverted triangle, and in some codes (e.g., IBC, 2000) it becomes increas-
ingly concave with increasing fundamental period. These loading shapes are
quite adequate for predicting the moment distribution, but usually are far
off the mark for the peak shear forces in flexural wall structures. This fact
was not generally recognized until the mid-1970s. The work of Blakeley
et al. (1975) on isolated walls demonstrated that the higher modes of vibra-
tion not only increase the elastic shear demand relative to that commen-
surate with the moment demand, but they amplify the base shear more
appreciably after the formation of a plastic hinge at the wall base. Their
work brought about a change in the provisions of the New Zealand seis-
mic code (New Zealand Standards Association, 1982, 1995), and in turn in
several other seismic codes and design practice (Paulay and Priestley, 1992).
It is interesting to note that code-writing bodies in the USA have not been
quick to follow, but the Commentary to the 1999 edition of the Recom-
mended Lateral Force Requirements of the Structural Engineers Association
of California (SEAOC, 1999) already recommends adopting the amplifica-
tion factor ωv advocated by the New Zealand code. This factor is made
dependent on the number of storeys n, as follows:

Va = ωv Vd (1)
⎧ n
⎪
⎨ 0.9 + , n≤6
ωv = 10
⎪ n
⎩ 1.3 + ≤ 1.8, n>6
30
in which Va is the amplified value of the base shear Vd as evaluated by
means of the equivalent lateral force procedure. Note that Eq. 1 depends
indirectly on the fundamental period T1 through its dependence on n.
Studies by Keintzel and his co-workers (e.g., Keintzel, 1990) and later
ones by Ghosh (1992) and Seneviratna and Krawinkler (1994) have shown
that the expected amplification is much larger than that resulting from
Eq. 1, and that the extent of amplification is not only a function of the
SEISMIC SHEAR DEMAND 3

fundamental period T1 , as is implied by the dependence of ωv on the num-

ber of storeys, but also on the expected ductility demand, which in seismic
codes is the main contributor to the force reduction, or behaviour factor q
(R in North America). However, it is not presented in a form that is easy
to apply.
The European approach to this problem has changed somewhat through-
out the years, as is manifested in the successive drafts of EC8. The pres-
ent version EN 1998-1 (CEN, 2004) has different amplification formulas
depending on the design ductility class (DC). No amplification is spec-
ified for wall systems of low ductility (DC-L). For uncoupled wall sys-
tems with medium ductility (DC-M), i.e., q ≤ 3.0 a constant 50% increase
in shear demand is required. For systems designed for high ductility, i.e.,
q ≤ 4.0–6.0 (depending on redundancy) Keintzel’s (1990) formula (Eq. 5.25
therein, also Penelis and Kappos, 1997) is specified, namely:

 2  2
γRd MRd Se (TC )
ε=q · + 0.1 · ≤q (2)
q MEd Se (T1 )

in which: ε = ωv is the magnification factor, γRd the overstrength factor

(usually 1.2), MRd the design flexural strength at the wall base, MEd the
design bending moment, TC the upper limit period of the spectral accelera-
tion plateau in the design spectrum, and Se (T ) is the ordinate of the elastic
response spectrum. While the fundamental period T1 and q are directly rep-
resented, higher vibration modes affect the response through the squared
spectral ratio in the second term within the square root. The first term
ensures that the shear capacity is commensurate with the flexural strength.
Note that a different amplification factor is specified in EC8 for large
lightly reinforced walls. These walls are not expected to dissipate energy
through plastic hinging at their base, hence are outside the scope of the
present paper.
A parametric study was recently carried out at the Technion by Nsieri
(2004), using two suites consisting of 20 records each. These ensembles
have an exceedance probability of 10% in 50 years, or a return period
of 475 years (Somerville, 1997), and were developed for the SAC project
that was initiated in response to the damage to steel structures caused by
the January 1994 Northridge earthquake. The two-dimensional version of
the computer code RUAUMOKO (Carr, 2000) was used in the analysis
assuming elastic–plastic response and 5% tangent stiffness Rayleigh damp-
ing in the 1st and 5th modes. The appendix shows the analysis model and
the parameters used. Note that for a given time history q was defined as
the ratio of the minimum strength of the structure for elastic response to
the strength actually allotted to it.
4 A. RUTENBERG AND E. NSIERI

Figure 1. Mean results of ωv∗ vs. T and q from parametric study and EC8 (2004).


Figure 1 shows the variation of the mean amplification factor ωv∗ = Va Vd
with T for several values of q for the Los Angeles suite. Superposed on the
parametric results are those of Eq. 2 as well as the EC8 50% amplification
for the lower q values (DC-M). Note that Vd is taken here as triangularly
distributed at floor levels over the height of the building and causing flexural
yielding My at the base, namely:
My
Vd =
(3)
2
3
H 1 + 2n
1

where H is the building height and n is the number of storeys. It is seen

that Va is increasing with T and q. Comparing the EC8 amplification with
the parametric results it is evident that Keintzel’s formula – applicable to
DC-H – overestimates the base shear demand amplification for a wide
range of natural periods, with some underestimates at high q levels. For
DC-M walls the flat 50% amplification practically always underestimates
the predicted demand. The results for the second, i.e., the near fault (NF),
suite of records (Somerville, 1997) were found to be very similar to the LA
ones, with a mean lower by circa 8% (not shown).
It is therefore evident that with respect to base shear demand the EC8
provisions are in need of revision, particularly for the lower q values. Since
it appears that the amplification is linear in T and q the following formula
is proposed for isolated walls and structural systems comprising of flexural
walls all having practically equal lengths:
SEISMIC SHEAR DEMAND 5

Figure 2. Mean results from parametric study and proposed formula.

Va = [0.75 + 0.22(T + q + T q)]Vd (4)

Figure 2 compares the parametric results with the proposed formula. It

can be seen that this simple formula overestimates the shear demand for
low T and large q values, but otherwise follows the mean value curves
rather faithfully.
It may be interesting to note that Filiatrault et al. (1994), who studied
the problem from the Canadian code perspective, suggested that base shear
amplification be accounted for in the code by means of reduced q factors.
The need to restrict this formula to an isolated wall or to a number of
walls all having equal lengths will become apparent subsequently.

2.1. Shear force distribution over the building height

The distribution of shear forces on isolated structural walls over the build-
ing height has been investigated parametrically as function of T and q,
for the SAC LA10/50 ensemble consisting of 20 records (Somerville, 1997),
again assuming elastic–plastic moment–curvature relationship, plastic hinge
formation only at the base and uniform flexural stiffness. Figure 3 shows
the variation of the mean dynamic storey shear demands normalized by
the dynamic base shear for wall structures having 5, 10, 20 and 30 sto-
reys, with fundamental periods of 0.3, 0.7, 1.5, and 2.0 s, respectively, and
for q = 1, 2, 4, 6. Since the shear demand over the height does not follow
that resulting from the triangular load distribution, a design envelope of
6 A. RUTENBERG AND E. NSIERI

Figure 3. Mean dynamic storey shear demands normalized by the dynamic base shear for
wall structures: (a) 5-storey, (b) 10-storey, (c) 20-storey, (d) 30-storey.

Figure 4. Design envelope of the shear forces over the building height.

the shear forces over the building height is proposed as function of the fun-
damental natural period T , as shown in Figures 3 and 4, ξ is given by

ξ = 1.0 − 0.3 T ≥ 0.5 (5)

Each one of the expressions 0.1 H and ξ · H in Figure 4, should be taken

as an integer number of storeys. Note that the proposed shear distribution
formula is fitted to q = 1, i.e., it is conservative for larger q values. A mod-
ified form of Eq. 5 can easily be developed to cover larger q values, but
it is believed that the resulting savings would be marginal. Note also that
SEISMIC SHEAR DEMAND 7

Figure 4 resembles Figure 5.4 in EC8 (CEN, 2004), but therein it is con-
fined to walls in dual systems.
Based on a limited parameter study on 10 storey walls for q = 4 with
stepwise stiffness taper to 1/3 of the base value (Nsieri, 2004) it was con-
cluded that the effect on base shear of stiffness taper due to variations in
gravity axial compression and partial curtailment of longitudinal reinforce-
ment is not significant (less than 5%).
The distribution of shear demand over the height for walls of different
lengths acting in unison, which is different, is considered subsequently.

2.2. Effects of plastic hinges developing at higher levels

The effects of plastification above the base on the dynamic base shear
amplification factor and storey shear envelopes have been investigated for
the SAC LA10/50 ensemble consisting of 20 records (Somerville, 1997),
assuming uniform flexural strength and stiffness over the height of the wall.
The mean storey moment demands normalized by the flexural strength My
for structures having 15 and 20 storeys, with fundamental period of 1.0
and 1.5 s respectively, for q = 4, 6 are shown in Figure 5. It can be seen
that yielding may extend over a significant portion of the wall, mainly for
longer periods and high values of q (low values of flexural strength) which
can be attributed to higher mode effects.

Figure 5. Mean dynamic storey moment demands normalized by the flexural strength for
wall structures: (a) and (b)15-storey, (c) and (d) 20-storey.
8 A. RUTENBERG AND E. NSIERI

Figure 6. Comparison of mean results from parametric study.

Figure 6 shows the effect of flexural yielding developing at higher floor

levels on the mean base shear amplification for wall structures having 5,
10, 15, 20 and 30 storeys, with fundamental period of 0.5, 0.7, 1.0, 1.5,
and 2.0 s, respectively, and for q = 2, 3, 4, 5, 6. It can be seen that for wall
structures with long periods and high values of q (low values of flex-
ural strength), the base shear amplification decreases if flexural yielding is
allowed over the height of the structure. However, for low values of q this
reduction is either insignificant or not present at all.
A comparison of storey shear envelopes obtained from the analysis
without and with flexural yielding over the height is shown in Figure 7 for
wall structures with 15 and 20 storeys, having fundamental periods of 1.0
and 1.5 s, and for q = 4, 6. The effect of flexural yielding at higher levels
is reflected in a small decrease in the shear amplification along the height.
Hence it may be concluded that Eq. 5 is also applicable to cases where
plastification develops at higher elevations.

3. Shear force distribution in systems comprising several unequal walls

Consider first a one-storey structure consisting of several fixed base can-
tilever walls, each having a different length, that are all connected by a
roof diaphragm, rigid in its own plane but otherwise flexible. The system
is loaded horizontally at roof level. Within the elastic range the shear force
distribution among the walls is proportional to their stiffnesses. However,
SEISMIC SHEAR DEMAND 9

Figure 7. Mean dynamic storey shear demands normalized by the dynamic base shear for
wall structures: (a) and (b) 15-storey, (c) and (d) 20-storey.

Figure 8. Force-displacement relationships for four one-storey walls having equal dis-
placement (Paulay and Restrepo, 1998).

following yielding of the longer wall, the additional load will be carried by
the shorter ones, as can be seen in Figure 8. It is, therefore, evident that
design for shear on the basis of relative stiffness underestimates the shear
10 A. RUTENBERG AND E. NSIERI

Figure 9. Two-storey wall system: (a) properties and loading, (b) floor forces and
deflected shapes (Rutenberg, 2004).

forces on the shorter wall, since the shear force capacity should be made
proportional to the flexural strength. This point is known, and, as noted,
has been addressed in the technical literature as well as in EC8 (1st term
in Eq. 2), and other seismic codes.
Consider now the two-storey system supported by the two walls shown
in Figure 9(a). Wall 1 is fully fixed at its base, whereas Wall 2, which
is longer and hence stiffer, is hinged. The two horizontal pin-ended rigid
members model the floor slabs connecting the two walls. For simplicity
assume that the external horizontal force acts only at roof level. Note that
the hinge at the base of Wall 2 is in fact a plastic one, and has formed at a
horizontal force level H . An additional force H in the same direction will
act only on Wall 1 provided that Wall 2 follows its deflected shape without
resistance. However, this requires that another hinge forms in Wall 2 at first
floor level, which usually is not the case. To fix ideas, let Wall 2 be very
much stiffer than Wall 1. Therefore, under the force increment H , Wall 2
will enforce on Wall 1 its straight-line deflected shape at every floor level.
The resulting forces on the system and the deflected shapes are shown in
Figure 9(b), assuming simple flexural behaviour (i.e., ignoring shear defor-
mation), rigid floors and no foundation rotation. It is seen that the addi-
tional shear force on Wall 1 V1 (0) = 2.5H = 1.25M(0) h, (M(0) is
the base moment increment due to H and h is the storey height), and
the shear on Wall 2 is reduced by 1.5H . In fact, for a rigid Wall 2 and a
large number of equal storeys,
 it can easily be shown (e.g., Karman and
√ 
∼
Biot, 1940) that V1 (0) = 3 − 3 M(0)/ h, which is quite close to the
two storeys value, and in this case is independent of the vertical distribu-
tion of the incremental horizontal loading.
SEISMIC SHEAR DEMAND 11

In an elastic-perfectly-plastic system all this takes place without a

change in the plastic hinge moment because the sense of curvature remains
the same. Real systems are, of course, much more complicated than that:
they are unlikely to be fully fixed at the foundation level; strain penetration
is likely to be important, rocking may even take place, shear deformations
may become important with falling aspect ratio, and floors are not infi-
nitely rigid in their own planes. Also, the assumption of concentrated plas-
ticity is a very crude approximation for walls of some depth, as the hinge
lengths will increase with increasing H , and the transition of the force–
displacement curve to the plateau is not abrupt, as assumed in this model.
Furthermore, the elastic–plastic model is a crude description of the hys-
teretic response of reinforced concrete walls. Therefore, in many instances
this effect could be less pronounced than in this case, also because Wall
2 is not necessarily very much stiffer than Wall 1. Yet, this simple model
does show that the shear demand on the shorter wall may be much larger
than that evaluated on the basis of the usual assumptions. In other words,
it may not be correct to assume, as is often done in practice, that shear
force demand is proportional to bending moment demand, even in cases in
which the effect of higher modes of vibration is considered.
An important aspect of the moment and shear transfer to the still uny-
ielded walls is the large force acting mainly on the first floor diaphragm
N , which, for a stiff wall with a plastic hinge at its base and large num-
√ 
ber of storeys, can reach the value of 6 2 − 3 M(0) / h. Note that this
force depends on the base moment, rather than on the base shear. This
floor force can manifest itself either as an axial force when the lines of
action of the walls coincide or as shear when the walls are parallel.
For increasing statically applied lateral load it is possible to follow the
post first-yield shear force distribution among a number of walls – as plas-
tic hinges form at the base of each one – either by means of successive
applications of Eq. 6, successive static linear analyses, or pushover analysis.
These equations take the following form:
    
V if (0) = V (0) + α M (0) Ih If h I if I (6a)
    
V ih (0) = V (0) − α M (0) h I ih I (6b)

In which Vif (0) and Vih (0) are the respective shear forces and Iif
and Iihare the respective moments of inertia of the ith fixed and hinged
walls, I and Ih are the moment
 of inertia sums of all walls and the
√ 
hinged walls respectively, and α = 3 − 3 for walls having uniform flex-
ural stiffness over the height. In multistorey buildings the second term in
Eq. 6 is usually much larger than the first one, so that in fact Vi (0), as
12 A. RUTENBERG AND E. NSIERI

Figure 10. Cyclic pushover load shape as function of “Time”.

N , the first floor diaphragm force, depends on the base moment rather
than on the base shear.
It is more difficult to predict the peak shear forces in the walls during
an earthquake by means of the above approach. The main difficulty lies in
the fact that at the lateral load level for which all the walls have yielded
the shear forces acting on them, except on the most flexible one, are not
at their peak values – as can be inferred from the simple two-storey exam-
ple presented above. On the other hand, pushover analysis (up to yielding
of all walls) with load reversal, or cyclic pushover, as shown in Figure 10,
can estimate the sought peak values. In fact such an analysis is not really
needed if it is noted that full load reversal is commensurate with doubling
the base shear at yield Viy on each wall. Therefore, an estimate of the peak
shear Vi,max on wall i can be obtained from:
   
Vi,max = 2 Viy  − Vity  (7)

in which Vity is the base shear of the wall when all walls have yielded.
This type of analysis requires that the base shear on the system be
known. One can use, of course, the ωv values in seismic codes (e.g., EC8
Eq. 5.25, Eq. 2 herein), but, as already shown, these are not sufficiently
accurate; hence the use of Eq. 4 is recommended. Note, however, that the
applicability of Eqn. 4 is predicated on the assumption that the total base
shear acting on a group of walls is equal to that on an isolated wall whose
stiffness and strength are equal to the respective sums of these properties
of the individual walls. This assumption was checked for the two 20-record
suites used for the analyses, and was found to be fully justified.
Note, however, that the pushover analysis proposed herein leads to the
formation of plastic hinges at base level only. Time history analyses have
shown that in the more rigid walls plastic hinges may, and often do, prop-
agate to higher levels. This further shear redistribution cannot be captured
by the proposed procedure, yet, as will be shown subsequently, the devia-
tions due to plastic hinge spreading are not excessive.
SEISMIC SHEAR DEMAND 13

Figure 11. Four walls lateral load resisting system for 10-storey building. On the left
typical lateral inertia-force distribution at maximum shear.

Table I. Properties of the four Walls for models 1, 2 and 3

 
Wall lw (m) EI (kNm2 ) I I My (kNm) My My

Model 1: Stiffness Ratios 1 : 2:3:4

1 3.50 6.90 × 106 0.1 7876 0.1391
2 4.41 13.8 × 106 0.2 12504 0.2209
3 5.05 20.7 × 106 0.3 16382 0.2894
4 5.56 27.6 × 106 0.4 19846 0.3504

69.0 × 106 1.0 56608 1.0000
Model 2: Stiffness Ratios 1 : 2:4:8
1 3.50 6.90 × 106 0.0667 7876 0.1097
2 4.41 13.8 × 106 0.1333 12504 0.1742
3 5.56 26.7 × 106 0.2667 19911 0.2773
4 7.00 55.2 × 106 0.5333 31504 0.4388

102.5 × 106 1.0000 71795 1.0000
Model 3: Stiffness Ratios 1 : 3 : 9 : 27
1 2.50 2.52 × 106 0.025 4018 0.0609
2 3.61 7.56 × 106 0.075 8356 0.1268
3 5.21 22.68 × 106 0.225 17385 0.2637
4 7.50 68.04 × 106 0.675 36165 0.5486

100.8 × 106 1.000 65924 1.0000

The influence of shear deformation in the walls and of the in-plane flex-
ibility of the floor diaphragms has also been considered. As expected, both
reduce to some extent the redistribution of shear in the walls. Some results
are presented in the following section.
14 A. RUTENBERG AND E. NSIERI

3.1. Numerical examples

The following examples compare the results of the static cyclic pushover
analysis with those of time history analyses for three symmetric structural
models shown schematically in Figure 11. The four walls can be consid-
ered as representing one half of the lateral load resisting system in a sym-
metric 10-storey building. These models differ in the stiffness and strength
ratios among their respective walls. Model 1 represents a system with
moderate variation of properties among its walls (stiffness ratios: 1:2:3:4),
Model 3 has large variations (1:3:9:27), while Model 2 has intermediate
ones (1:2:4:8). The relevant structural properties of the walls are given in
Table I. A simplified typical lateral force distribution over the height of the
building at peak base shear is shown on the left hand side of Figure 11. It
is assumed that the floor slabs are rigid in their own plane and completely
flexible out of plane. The walls are modelled as elastic-perfectly plastic with
uniform flexural stiffness over the height, and the mass of the structure was
adjusted to give a fundamental vibration period T = 1.0 s. Rayleigh damp-
ing of 5% in the 1st and 5th modes using the tangent stiffness matrix was
assumed.
The static approximation of the peak shear forces at the base of the
walls was carried out by cyclic pushover analysis for q = 4 up to apeak
base shear Vmax = Va (Eq. 4), which was applied at the level heq = My Vmax
above the base (My = yield moment of the combined 4 walls). Hand calcu-

Figure 12. Cyclic pushover results for Walls 1 and 4: base shear vs. time (Model 3).
SEISMIC SHEAR DEMAND 15

Figure 13. Cyclic pushover results for Walls 1 and 4: base shear vs. roof displacement
(Model 3).

Table II. Comparison of peak shear values (kN)

4
V1 V2 V3 V4 Base shear i=1 Vi

Model 1: Stiffness ratios 1 : 2 : 3 : 4

Mean 1651 1995 2416 3387 7015 9449 #
Pushover analysis 1655 2044 2536 3371 7090 9606
Stiffness proportional 702 1403 2105 2805 7015 7015
Strength proportional 976 1550 2030 2459 7015 7015
Model 2: Stiffness ratios 1 : 2 : 4 : 8
Mean 1884 2437 3408 6572 9242 14301 #
Pushover analysis 1901 2567 3676 6132 9330 14276
Stiffness proportional 615 1230 2460 4937 9242 9242
Strength proportional 1014 1611 2555 4062 9242 9242
Model 3: Stiffness ratios 1 : 3 : 9 : 27
Mean 1187 2077 3274 7595 8234 14133 #
Pushover analysis 1208 2118 3560 7009 8256 13895
Stiffness proportional 206 618 1853 5557 8234 8234
Strength proportional 501 1044 2171 4518 8234 8234

# non-simultaneous.

lation for Model 2, using successive applications of Eqs. 6 followed by Eq.

7, is illustrated in Rutenberg (2004).
The dynamic analyses were carried out using RUAUMOKO (Carr, 2000)
with the same ensemble of the SAC LA10/50 20 accelerograms.
16 A. RUTENBERG AND E. NSIERI

Table III. Effect of plastification spreading above base: comparison with push-
over results

V1 V2 V3 V4 Base Shear

Model 1: Stiffness Ratios 1 : 2 : 3 : 4

Mean 1859 2349 2561 3192 6925
Pushover analysis 1655 2044 2536 3371 7090
Model 2: Stiffness Ratios 1 : 2 : 4 : 8
Mean 1884 2437 3408 6572 9242
Pushover analysis 1901 2567 3676 6132 9330
Model 3: Stiffness Ratios 1 : 3 : 9 : 27
Mean 1231 2373 4503 6527 8011
Pushover analysis 1208 2118 3560 7009 8256

In both analyses it was first assumed that plastic hinges could form only
at the wall bases. Then this restriction was relaxed. Figure 12 shows the
variations of the total base shear and the shear forces on Walls 1 and
4 with “time”, and Figure 13 shows these variations with roof displace-
ment. Note the small residual displacements and shear forces. A compar-
ison of the pushover analyses with the mean time-history results is given
in Table II. This table also presents the shear forces that the structure
would be designed for using routine procedures. Comparing the base shear
demands on the three models it is evident that for Model 1, with its very
moderate stiffness ratios (1:2:3:4), the routine procedures are already very
off the mark for the most flexible wall, and with increasing stiffness ratios
they become increasingly unconservative for all the walls. It is also seen
that the proposed approach yields very reasonable estimates of the seismic
shear force demand on the walls.
Note that when plastic hinges are permitted to develop at higher lev-
els the results may be less satisfactory, but are still quite adequate for
design purposes for systems with moderate variation of properties among
the walls. Table III compares the peak shear forces for this case with the
cyclic pushover results reported in Table II. This comparison, rather than
with the mean dynamic shear forces in walls with plastic hinging confined
to the base, is made in order to show the expected error to be expected
when shear design is based on pushover analysis. It can be seen that the
maximum underestimate (21%) is in Wall 3 of Model 3, which has the larg-
est wall stiffness spread (1:3:9:27). Errors in the other two models are much
smaller.
The distribution of the shear demand over the height of each wall was
also studied. It was found that the envelope proposed for an isolated wall
SEISMIC SHEAR DEMAND 17
Table IV. Effect of shear deformation on base shear (kN); factored El Centro
record


4
V1 V2 V3 V4 Base shear Vi
i=1

Model 3: Stiffness Ratios 1 : 3 : 9 : 27

Neglecting shear 1187 2077 3274 7595 8234 14133
deformation
Pushover analysis 1208 2118 3560 7009 8256 13895
Shear deformation 1120 1842 2950 6213 8315 12125
(uncracked member)
Pushover analysis 1112 1866 3040 5524 8256 11542
Shear deformation 1014 1663 2453 5110 8820 10240
(cracked member)
Pushover analysis 1007 1660 2394 4279 8256 9340

Figure 14. Base shear vs. In-plane floor stiffness for 2 mass distributions: factored El Cento
record.

(Eq. 5) can be quite conservative for the more flexible walls in the systems
presented here.
The effect of considering shear deformation in the walls is shown in
Table IV. As expected, shear flexibility affects the longer walls (smaller
aspect ratio) more than the shorter ones. Note that cyclic pushover predicts
quite well (with the possible exception of V4 ) the peak shear also for these
cases.
18 A. RUTENBERG AND E. NSIERI

Table V. Effect of shear deformation and in-plane floor flexibility on base shear
(kN); factored El Centro record

V1 V2 V3 V4 Base
shear

Shear deformation & in-plane floor 1149 2048 2970 7584 7856
flexibility neglected – uncracked member
Shear deformation – uncracked member 967 1531 2593 5968 6918
& in-plane floor flexibility
Shear deformation – cracked member 907 1430 2351 5289 7782
& in-plane floor flexibility

The influence of in-plane floor flexibility for two different assumptions

on the distribution of mass among the walls is presented parametrically
in Figure 14 for the El Centro (1940) NS record factored by 1.85. It is
seen that with falling in-plane floor rigidity the effect of mass distribution

becomes more pronounced. Note that floor rigidities lower than 106 kN m
are not realistic for reinforced concrete floors.
The combined effect of shear deformation and in-plane floor flexibility
is summarized in Table V, assuming  that floor diaphragm stiffness between
two adjacent walls is 7.5 × 106 kN m and floor mass distribution is propor-
tional to wall cracked flexural rigidities. Note that the designations cracked
and uncracked refer to shear cracking (Park and Paulay, 1975). It is seen
that some reduction in the shear demand is to be expected, particularly in
the stiffest wall.
It may also be interesting to compare the sequence of plastic hinge
formation under seismic excitation with the static cyclic one. Whereas it
appears that it is always the stiffest wall that yields first, further hinging
does not follow the static sequence. For example, for Model 3 under the
factored El Centro record, yielding of Wall 1 follows that of Wall 4, then
Wall 3 and finally Wall 2 (i.e., 4:1:3:2 rather than 4:3:2:1).

4. Summary and Conclusions

This paper is concerned with the seismic shear force demand on flexural
walls serving as the lateral load resisting system in multistorey buildings.
Two main factors affecting the shear demand were studied: (1) higher
modes of vibration in the linear and post yield ranges and (2) shear
force redistribution from walls with plastic hinges to those still remain-
ing elastic. It has been shown by parametric analysis, using a suite of
20 accelerograms that: (1) the shear amplification factors specified in
SEISMIC SHEAR DEMAND 19

seismic codes of the type given in the New Zealand code (Eq. 1) are not
conservative, particularly for walls designed for high ductility demands,
i.e., for large strength reduction factors; (2) the EC8 (CEN, 2004) shear
amplification formula (Eq. 2) is in need of calibration, since it does
not adequately predict the expected response for DC-H structures and is
unconservative for DC-M ones. Note that these observations are valid not
only for structures designed by the equivalent lateral force procedure, but
also for those designed using modal analysis since most of the shear ampli-
fication occurs after the plastic hinge was formed. However, when modal
analysis is performed it already provides the correct shear amplification
for elastic analysis, i.e., for q = 1, hence the amplification should reflect
only the nonlinear effects. Eq. 4 or another simple formula based on the
data leading to it can replace the one given in seismic codes. An enve-
lope (Eq. 5) for the shear demand over the wall height has also been
provided.
The study on multistorey multi-wall systems shows that non-simultaneous
yielding at the bases of walls has a significant effect on the base shear distri-
bution among them. It is demonstrated that the shear demand on the flexi-
ble walls is likely to be much larger than is commensurate with their relative
stiffness, or even with their relative flexural strength. Using the shear demand
given in Eq. 4, a simple pushover procedure is proposed to estimate the base
shear in each wall.
Whereas most of the important parameters have been considered, the
effect of the strain hardening ratio on the shear demand in systems of
unequal walls needs further study since preliminary analyses indicate that
some additional shear amplification in the shorter walls can be expected.
Shear deformation in the walls and the in-plane flexibility of the floor
diaphragms mitigate the amplification to some extent. Yet, a modification
of seismic code provisions to reflect shear force amplification with increas-
ing natural period and ductility, as well as the redistribution due to succes-
sive formation of plastic hinges is called for.
This study was confined to structures in which flexural walls form the
sole seismic-force resisting system. A study of wall-frames – more common
structures – is now under way.

Acknowledgements
This study was supported by the Israeli Ministry of Construction and
housing through a grant to the National Building Research Institute at the
Technion.
20 A. RUTENBERG AND E. NSIERI

Appendix
The parameter study from which Eq. 4 was derived was based on the ana-
lytical model shown in Figure A1 and with the parameters given in Table
A1. Storey height was 3.0 m in all cases.

Figure A1. Analytical model of wall structures used in the parameter study.

Table A1. Number of storeys and fundamental periods of wall

structures used in the parameter study

Fundamental period (s) Number of storeys

5 10 15 20 25

0.30 0.50 1.00 1.50 2.50

0.50 0.70 1.25 1.75 2.75
T1 0.70 1.00 1.50 2.00 3.00
1.25 1.75 2.25
1.50 2.00 2.50
SEISMIC SHEAR DEMAND 21

References
Blakeley, R.W.G., Cooney, R.C. and Megget, L.M. (1975) Seismic shear loading at flexural
capacity in cantilever wall structures. Bulletin New Zealand National Society Earthquake
Engineering 8, 278–290.
Carr, A.J. (2000) RUAUMOKO – Program for inelastic dynamic analysis. Department of
Civil Engineering, University of Canterbury, Christchurch.
Comite Europeen de Normalization (CEN) (2002) Eurocode (EC) 2: Design of concrete
structures – Part 1 General rules and rules for buildings (prEN-1992-1-1), Brussels.
CEN (2004), Eurocode (EC) 8: Design of structures for earthquake resistance – Part 1 Gen-
eral rules, seismic actions and rules for buildings (EN 1998-1), Brussels.
Filiatrault, A., D’ Aronco, D. and Tinawi, R. (1994) Seismic shear demand of ductile canti-
lever walls: a Canadian code perspective. Canadian Journal of Civil Engineering 21, 363–
376.
Ghosh, S.K. (1992) Required shear strength of earthquake-resistant reinforced concrete
shearwalls, in P. Fajfar and H. Krawinkler (eds.), Nonlinear Seismic Analysis and Design
of Reinforced Concrete Buildings, pp. 171–180, Elsevier, London.
IBC (2000) International Building Code. International Code Council, Falls Church, Virginia.
Karman, T. von and Biot, M.A. (1940) Mathematical Methods in Engineering. McGraw-Hill,
New York.
Keintzel, E. (1990) Seismic design shear forces in reinforced concrete cantilever shear wall
structures. European Journal of Earthquake Engineering 3, 7–16.
New Zealand Standards Association. (1982) and (1995) NZS 3101 Code of Practice for the
design of Concrete Structures (Parts 1 & 2). Standards Association of New Zealand,
Wellington.
Nsieri, E. (2004) The seismic nonlinear behaviour of ductile reinforced concrete cantilever
wall systems. M.Sc. thesis, Technion, Israel Institute of Technology, Haifa, Israel.
Park, R. and Paulay, T. (1975) Reinforced Concrete Structures. John Wiley & Sons, New
York.
Paulay, T. and Priestley, M.J.N. (1992) Seismic Design of Reinforced Concrete and Masonry
Buildings. John Wiley & Sons, New York.
Paulay, T. and Restrepo, J. (1998) Displacement and ductility compatibility in buildings
with mixed structural systems. Journal New Zealand Structural Engineering Society 11(1),
7–12.
Penelis, G.G. and Kappos A.J. (1997) Earthquake-Resistant Concrete Structures. E & FN
SPON (Chapman and Hall), London, U.K.
Rutenberg, A. (2004) The seismic shear of ductile cantilever wall systems in multistorey
structures. Earthquake Engineering & Structural Dynamics 33, 881–896.
SEAOC (1999) Recommended lateral force requirements and commentary. Structural Engi-
neers Association of California, Sacramento, CA.
Seneviratna, G.D.P.K. and Krawinkler, H. (1994) Strength and displacement demands for
seismic design of structural walls. Proceedings of the 5th US National Conference on
Earthquake Engineering, Chicago, Vol. 2, pp. 181–190.
Somerville, P. et al. (1997) Development of ground motion time histories for Phase 2 of the
FEMA/SAC steel project. Report SAC/BD-97/4, SAC Joint Venture, Sacramento, Calif.