seismic shear loading at flexural capacity in cantilever wall structures

278
SEISMIC SHEAR LOADING AT FLEXURAL CAPACITY
IN C A N T I L E V E R WALL S T R U C T U R E S
R VV.G. B l a k e l e y * , R.C. C o o n e y * * , L.M. M e g g e t * * *
ABSTRACT
An investigation is described into the effect of various combinations
of the normal modes of vibration of cantilever shear wall structures on the
maximum shears at flexural capacity.
It is shown that the base shear can be
much higher than would be derived by assuming a normal code lateral load
distribution of sufficient magnitude to cause flexural yielding.
The
results of elastic normal mode response spectrum analyses of a 10-storey
building considering several structural variables are presented in terms
of envelope values of the ratio of maximum base shear at flexural
capacity to that assuming a code lateral load distribution.
The same effect
is investigated with a series of step-by-step numerical integration dynamic
analyses of cantilever wall structures responding inelastically to a range
of earthquakes. On the basis of the results suggestions are made for the
shear design of cantilever walls.
1.
INTRODUCTION
The capacity design philosophy as
applied to frame or wall structures has the
objective of achieving a desirable sequence
in the failure mechanism of the structure
under severe earthquake motions. It entails
selection and suitable detailing of the
mechanisms best capable of ductility and
energy dissipation and provision in the
remaining mechanisms of sufficient reserve
strength to ensure that inelastic behaviour
is predominantly confined to the chosen
energy dissipating mechanism.
In terms of
ductile cantilever walls this approach
requires provision of sufficient shear
strength to confine energy dissipation to
ductile flexural yielding.
The New Zealand Standard Code of
Practice for General Structural Design and
Design Loadings
allows that where the
lateral force resisting system comprises
cantilever shear walls with height to
width ratios exceeding 2, width exceeding
1.5m and acceptable openings they may be
designed with a structural type factor, S,
of 1.2 or 1.0, depending on whether there
is one or more walls, provided design and
detailing is adequate to ensure that energy
is dissipated by ductile flexural yielding.
The design lateral forces may be determined
either by an equivalent static force
analysis or by a spectral modal analysis.
The latter is based on a design spectrum
comprising the curve for the basic seismic
coefficient multiplied by a scaling factor,
so chosen that the computed base shear is
no less than 0.9 times that derived from
an equivalent static force analysis. At
*
Design Engineer, Office of the Chief
Designing Engineer (Civil), Head Office,
M.W.D.
**
Senior Engineer, Office of the Chief
Structural Engineer, Head Office, M.W.D.
*** Senior Engineer, Office of the Chief
Structural Engineer, Head Office, M . W . D .
any other level the shear derived by the
spectral modal analysis is to be taken as
no less than 80% of the values computed by
the equivalent static forces method.
The
calculated shears are to be used to derive
the horizontal forces and overturning
moments.
It has been commonly assumed that the
design shear for a cantilever wall can be
determined by, first, calculating the
flexural capacity of the wall at the base
making allowance for sources of overstrength
in accordance with capacity design principles
and, second, deriving the shear at that
moment capacity assuming a distribution of
horizontal seismic forces as specified by
the code
, comprising an inverted
triangle with the addition of a concentrated
load of 0.1 times the base shear applied at
the top storey when the height to depth
ratio of the wall is equal to or greater
than 3. This lateral force diagram represents
predominantly first mode inertial forces
with the additive effect of second mode
accelerations at the top, but designers
should be aware that this does not necessarily
indicate either the critical shear forces
at capacity moment in the wall or the p o s s ibility of plastic hinges forming in the
wall above the base. There may be many
occasions in the response of a cantilever
wall structure to a major earthquake where
the higher mode forces are predominant over
those of the first mode and the combination
of forces is still sufficient to cause
flexural yielding at the base.
In such circumstances the centre of
lateral inertial loading is either lower
or higher than that predicted by the code
load distribution, depending on the direction
of combination of the applied loads due to
the various modes. Correspondingly, the
shears may be much higher at either the base
or the top of the wall than the design
values. The investigation reported in this
paper studied initially the upper limits of
B U L L E T I N OF THE NEW Z E A L A N D N A T I O N A L SOCIETY FOR EARTHQUAKE ENGINEERING, VOL.8, N 0 . 4 . DECEMBER 1975
279
likely shear at capacity moment in the wall
from an elastic normal mode response spectrum
approach.
The results were then compared
with the computed shears for three ductile
cantilever shear wall structures using a
step-by-step numerical integration computer
programme to compute the inelastic response
to five major earthquakes.
2.
NORMAL MODE RESPONSE SPECTRUM
strength are: reinforcing steel yield
strengths greater than the minimum
specified, Bauschinger effects in the
reinforcing steel,under cyclic loading,
increased contribution of the vertical
web steel to the flexural capacity with
increased imposed curvatures, concrete
compressive strengths greater than the
minimum specified, capacity reduction
factor used in the design, and actual
gravity load on the wall greater than
the design gravity load, particularly
where the critical design case is a
gravity load of 0.9D. Also, overstrength may arise because of minimum
vertical steel requirements throughout
the wall cross-section or where walls
are oversized).
ANALYSES
A normal mode response spectrum analysis
was made of a standard 10-storey, 30.5m
tower building using the results of
Skinner
to determine extreme values of
base shear relative to shear at capacity
moment derived from the code lateral load
distribution.
2.1 Assumptions,
2.3 Procedure Followed
a.
The procedure followed for this study
is illustrated in Figs. 1 and 2 for a
particular case of a 10-storey tower building
with first mode period of vibration of 0.8
seconds, viscous damping ratio of 5%, and
cantilever wall capacity base moment of 1.4
times the design base moment, and is
itemised in general as follows:
The masses are concentrated at the
floors and have a uniform value of
4.54 x 1 0 k g m .
The storey bending stiffness ratio falls
linearly from 3500 MN/m between the base
and floor one to 1750 MN/m between
floors nine and ten.
(This variation
is felt to be consistent with variation
of wall thickness and reinforcing steel
content up a building).
The ratio of storey bending stiffness
to shear stiffness is 0.5 for all storeys.
(The results were found to be not unduly
sensitive to this parameter).
The ratios of frequency of the first
three modes of vibration is 1:3.7:7.1
in accordance with values given by
Skinner £2) f
building with the
properties given above.
Design of the building is in accordance
with the seismic design co-efficients
of the New Zealand loadings code
:
basic seismic co-efficient, C, for a
building in Zone A founded on rigid or
intermediate subsoils, Class III buildings (I = 1 . 0 ) , lateral load resistance
provided by two or more approximately
symmetrically arranged ductile cantilever
shear walls (S = 1 . 0 ) , reinforced
concrete material (M = 1 . 0 ) , and no
unusual risk (R = 1 . 0 ) . Design loads
are determined from an equivalent static
force analysis.
Response spectra '^) derived by smoothing curves of a number of earthquakes
scaled to have the same spectrum
intensity as El Centro 1940 (N-S) are
applicable to buildings in Zone A.
The foundation capacity exceeds the
wall base flexural capacity.
5
b.
c.
d.
o r
e.
f.
g.
a
2.2 Variables Considered
a.
b.
c.
Six buildings with periods in the range
0.2 seconds to 1.2 seconds inclusive.
Three values of equivalent viscous
damping, 2%, 5%, 10%.
(Generally, a
ductile cantilever wall is considered
to have an equivalent viscous damping
co-efficient of about 5%. However, a
value of 2% is feasible before widespread cracking, and damping may be as
high as 10% after non-structural damage
has occurred or where there is significant foundation compliance).
Three values of total moment overstrength at the base, M , of 1.2, 1.4
and 1.6 times the design moment, M^.
(Factors which influence this over-
a.
b.
c.
Determine the design moment at the base
of the cantilever wall from the code
equivalent static force analysis.
Apply the appropriate assumed overstrength factor to derive the capacity
moment and compute the base shear for
that moment from the code lateral load
distribution.
Calculate the second and third mode
periods and scale the earthquake response
factors, R , from the response spectra
curves for the first three modes of
vibration and the appropriate damping
ratio.
Calculate the moment at the base of the
wall for the summation of the maximum
second and third mode response forces
acting in a direction such that the
shear forces are additive at the base.
Compute the proportion of first mode
response required to make up the
difference in base moment between the
capacity moment and that due to (d)
above.
Add the applied loads due to the maximum
second and third mode responses and the
proportionate first mode response and
hence derive the maximum base shear at
capacity moment for the combination of
these modes.
Calculate the ratio of shears from (f)
and (b) above.
m
d.
e.
f.
g.
Note that although the results have
been derived for a standard 10-storey tower
building they may be applied to tower
buildings of other height with the same
distribution of mass and stiffness and the
appropriate period of vibration.
This
applies because the shape of the curves for
normal mode response forces used (2) will
be general and the results are express in
terms of ratios of shear forces.
2.4 Results
2.4.1 Maximum Base Shears
Q
The example illustrated in Figs. 1 and
2 shows that the summation of 32% of the
280
maximum first mode response, 100% second
mode and 100% third mode gives an applied
load diagram sufficient to reach flexural
capacity at the base of the wall but with a
centre of applied inertial load at 0.34 of
the height of the wall.
If the code lateral
load distribution had been assumed with
sufficient force to reach flexural capacity
at the base, the centre of applied inertial
would have been 0.73 of the height of the
building.
The ratio, V , of maximum base
shear at capacity moment, assuming a critical
combination of the first three modes of
vibration, to that assuming the lateral load
distribution specified by the code is then
2.13. This same ratio for all 54 cases
studied is listed in Table 1, showing a
range from 1.39 to 4.02.
It may be seen
that generally this ratio increases with:
increasing periods of vibration, decreasing
overstrength ratios and decreasing damping
ratios. All three trends lead to a reduced
proportion of first mode maximum response
required to allow the summation of the three
modal base moments to reach flexural capacity.
This proportion varied from 49% to 10%
between the minimum and maximum shear ratios
in Table 1. As the proportion of first
mode forces at flexural capacity reduces,
so the influence exerted by the higher modes
on the shear forces increases.
r
Although the shear ratios of Table 1
have been calculated for cantilever walls
designed for the seismic co-efficients of
Zone A and analysed using response spectra
for earthquakes of the size of El Centro
1940 (N-S), the values are also applicable
to walls designed for the reduced co-efficients
of Zones B and C if the response spectra for
the design earthquake are considered to
reduce in the same proportion.
That is,
the values also apply to cantilever walls
designed for Zone B and analysed with
response spectra comprising acceleration
ordinates of 0.833 times those of the El
Centro-type curves, or for walls designed
for Zone C and analysed with response
spectra comprising acceleration ordinates
of 0.667 times those of the El Centro-type
spectra.
If an earthquake stronger than
El Centro was considered for Zone A the
ratios of Table 1 would increase because
the higher modes of vibration could exert
a proportionately greater influence on the
response.
The values of Table 1 represent an
envelope of maximum possible base shears
from a normal mode response spectrum
approach considering an earthquake of the
size of El Centro 1940 N - S , and assuming
that the contribution of modes higher than
the third is not likely to be significant.
From elastic considerations of ranges of
contribution of the three modes a distribution of the probability of occurrence of
base shear ratios up to the envelope values
could readily be determined.
However, account
must be taken of the limits to applied modal
forces imposed by plastic hinging, and it
is felt that the variability of base shear
forces is best investigated by means of
inelastic response computer analyses.
2.4.2. Plastic Hinge Formation at
Heights
Intermediate
Consideration of modal force combinations
can also indicate the possibility of plastic
hinge formation at levels above the base.
The moment diagram in Fig. 1 (c) shows
that the critical mode combination for shear
of 32% mode 1, 100% mode 2 and 100% mode 3,
where shears and moments are additive at
the base, is not critical for moments above
the base. However, if mode 2 acted alone a
plastic hinge could theoretically form in
the vicinity of the seventh floor for the
reverse direction of bending to that
considered above. Further, if the moments
due to mode 1 were additive to those of
mode 2 at higher floor levels, as shown
for the example of 40% mode 1 and 100%
mode 2 on the right hand side of Fig. 1 ( c ) ,
a plastic hinge could form at any point
within several stories above the base.
Wherever flexural yielding does occur it
will act as a fuse inhibiting further
increases in loading.
3.
INELASTIC TIME HISTORY SEISMIC
ANALYSES
A series of inelastic seismic analyses
were made of the time history of response
of three cantilever shear wall structures
of varying height subjected to five
different earthquake acceleration records
to determine shear and flexural hinging
characteristics.
3.1 Buildings Analysed
3.1.1 6-Storey Structure
An actual building designed in 1973
by the Ministry of Works and Development
for construction in Zone A was chosen as
an example of a 6-storey tower structure
relying on two cantilever walls for lateral
load resistance in one of the principal
axes of the building.
Lateral load resistance is provided in the other direction by
two coupled shear walls. The floor area
per storey is 8 6 9 m and the interstorey
height is 3.35m.
2
The cantilever walls are rectangular
with a member depth of 9.404m and thickness
of 4 57mm between ground and second floor,
356mm between second and fourth floor and
254mm between fourth floor and roof.
Reinforcing steel at the base comprises a total
of 72-31.8mm diameter bars concentrated
towards each end of the wall plus 34-15.9mm
diameter bars within the central region, with
a total steel percentage of 1.5%.
Design
was to a base shear co-efficient of 0.20g
with a design base moment of 59.5 MNm in
each wall. The calculated first mode period
of vibration was 0.45 seconds.
3.1.2 15-Storey
Structure
The example chosen of a 15-storey tower
structure was designed by the MWD in 1971 for
construction in Zone A. The floor area is
1 0 3 2 m per storey and the interstorey
heights are 4.27m between ground and first
floor, 3.97m between first and second
floors and 3.36m from second floor to roof.
Lateral load resistance is provided by two
coupled shear walls in one principal direction and two cantilever shear walls in the
other. Each cantilever wall has two flanges
of width 7.78m joined by a web at one end
with an overall depth of 12.19m.
The wall
thickness varies from 711mm between ground
and first floor, 559mm between second and
fifth floor, 457mm between sixth and ninth
2
281
floors and 3 56mm between the ninth floor and
the roof.
Reinforcement at the base is
provided by a 254mm by 254mm by 133kgm/m
universal column at each end of each flange
with 46-25. 4mm diameter bars in the flange,
and by 64-25.4mm diameter bars spread
uniformly through the web. The total steel
percentage at the base is 0.78%.
Design
was to a base shear co-efficient of 0.16g
and the design moment at the base of each
wall was 344 MNm.
The calculated first
mode period of vibration was 0.81 seconds.
The San Fernando earthquake of February
9 1971 was of Richter magnitude 6 . 6 with a
focal depth of 13km and slipping on a thrust
fault at 45° to the ground surface.
The
Pacoima Dam record was taken on a rock spine
in the epicentral region.
It has a maximum
acceleration of 1.2g and a duration of 10
seconds.
3.1.3
3.3 Analysis
20-Storey
Structure
A 20-storey cantilever wall structure
was designed for the purposes of this study,
with seismic co-efficient for Zone A and
importance factor for Class II structures
for consistency with the previous two
structures. The building is considered to
have a floor area of 1 4 8 8 m per storey and
interstorey heights of 4.58m between ground
and first floor and 3.36m for the remaining
storeys. Lateral resistance is provided by
two ductile cantilever walls comprising
flanges
of 9.15m length and a 15.25m overall
depth web at the end of the flanges.
The
wall thicknesses are 812mm from ground to
third floor, 711mm from third to sixth floor,
559mm from sixth to tenth floor, 4 57mm from
tenth to fifteenth floor, and 356mm from
fifteenth floor to roof.
The reinforcement
at the base comprises 88-31.8mm diameter
bars in each flange and 132-31.8mm diameter
bars spread evenly through the web with a
total steel percentage of 0.94%.
Design
was to a base shear co-efficient of 0.13g
and the design moment at the base of each
wall was 905MNm.
The calculated first mode
period of vibration was 1.2 seconds.
2
3.2 Earthquake Acceleration
Records
3.2.1 Taft, 1952, N21E
The Taft, California earthquake record
taken in July 1952 during ground motions
from the Richter magnitude 7.7 Kern County
earthquake.
The maximum acceleration is
0.16g and the period of strong shaking is
approximately 15 seconds.
3.2.2 El Centro, 1940, N-S
This record was made on alluvium 8 km
from the fault line of a Richter magnitude
6.4 earthquake.
The maximum acceleration
is 0.33g and the period of strong shaking
is approximately 12 seconds.
3.2.3
El
This is a simulated earthquake record
(3) designed to represent- the strong shaking
in a Richter magnitude 7 or greater earthquake . The maximum acceleration is 0.37g
and it has the same spectrum intensity as
the El Centro, 1940 record.
The period of
strong shaking is about 20 seconds.
3.2.4
A2
(3)
A simulated earthquake record
intended to model the shaking in the vicinity
of the fault in a Richter magnitude 8 or
greater earthquake, this has a spectrum
intensity half again as strong as that for
the El Centro, 1940 shock.
It has a maximum
acceleration of 0.44g with a period of strong
shaking of approximately 40 seconds.
3.2.5 Pacoima D a m
#
1971, S16E
Procedure
The structures were analysed with
DRAIN-2D(4), a general purpose computer
programme for the dynamic response analysis
of planar structures deforming inelastically
under earthquake excitation, using the
MWD's IBM 370/168 computer.
The structure
is idealised as a planar assemblage of
discrete elements. Analysis is by the
Direct Stiffness Method, with the normal
displacements as unknowns. Each node
possesses up to three displacement degrees
of freedom.
The structure mass is assumed
to be lumped at the nodes and static loads
may be applied prior to the dynamic loads.
A variety of structural elements is available but for this study beam-column elements
were used. These yield through the formation
of concentrated plastic hinges at the ends
of the element and permit flexural shearing
deformations to be considered.
A bilinear
moment-curvature hysteresis loop was assumed
for this study.
The dynamic response is
determined by step-by-step numerical integration, with a constant acceleration
assumption within any step. Results were
tested for sensitivity to the computational
time step and a time interval of 1/128 second
was chosen as being satisfactory.
The
viscous damping matrix at any time may be
based on the mass matrix and the current
tangent stiffness matrix or the original
elastic stiffness matrix. Where plastic
hinges form at the ends of beam-column
elements during a time step there may be
a moment unbalance on the node, which can
be corrected in subsequent time steps.
It
was found that where the stiffness-dependent
part of the damping matrix was based on the
original elastic stiffness, rather than on
the current tangent stiffness, numerical
stability was maintained and there were
only very minor differences in moment
across the nodes. The cantilever wall
structures were assumed to have 5% equivalent
viscous damping in the first mode.
The sensitivity of the response results
to the modelling of the walls was investigated.
Each wall was represented by a single column
of elements with an integer number of elements
per interstorey height.
The configuration
finally chosen had 9 elements in the first
interstorey height and decreasing numbers
of elements in succeeding storeys, with a
total of 18 elements in the 6-storey
structure, 28 elements in the 15-storey
structure and 40 elements in the 20-storey
structure.
The close spacing of elements
at the base of the walls best simulated
performance during plastic hinging in view
of the analysis procedure of concentrating
plastic rotations at the ends of each element.
As a basis for modelling the momentcurvature characteristics of the walls, each
282
section was analysed using the section design
capability of the ICES computer programme
system.
The probable strengths of the
walls were assessed by making the following
assumptions: steel yield stress of 1.25
times the minimum specified to account for
yield overstrength and Bauschinger effects
under cyclic load, gravity load on the wall
due to dead load plus 1.3 times the reduced
live load on the contributary part of the
structure, and a capacity reduction factor
of unity. A number of points on the momentcurvature curve were plotted and a bilinear
curve fitted to these points. For each wall
the curve of best fit was chosen as one with
first yield at 1.1 times design moment and
a stiffness after yield of 2% of the initial
elastic stiffness. The proximity of the
design moment to the first yield moment,
despite inclusion of probable sources of
overstrength in the latter figure, arises
because the calculated ultimate strength at
design moment for an extreme concrete fibre
compression strain of 0.003 includes a
major portion of the web steel at yield
strength, whereas this steel will be at a
lower stress at first yield of the extreme
bars in the tension flange. Thus, but for
over strength first yield in the wall could
occur at a moment up to 25% less than the
design moment.
In the modelled momentcurvature relationship the strain hardening
branch of the curves after yield allows for
increasing contribution of the web steel
with increasing curvatures. The initial
elastic stiffness was chosen to allow for
flexural cracking by an assumption of a
moment of inertia of 0.5 times that of the
gross section at the base, the proportion
increasing uniformly with height to unity
over approximately the top third of the
structure.
The output of dynamic response analysis
programmes is, of course, subj ect to the
limits of accuracy with which the structural
properties and performance can be modelled
in computer analysis.
In this study the
reliability of results is enhanced since
the primary interest is in ratios of shears
under different modal combinations rather
than absolute values.
3.4
Results
An example is shown in Fig. 3 of the
time history of response of the 15-storey
structure to 20 seconds of the Bl simulated
earthquake.
The curves plotted are displacement at the top of the wall against time
and base shear against time, and points have
been plotted for the ratio of base moment to
base shear. The plot of the ratio of base
moment to base shear indicates the variability
of the modal combinations.
These only give
rise to significant shear forces at or
approaching flexural yield in the structure
and values of this parameter which were
recorded while part of the wall was yielding
are indicated.
At small force levels a
number of points were outside the range of
the plot.
This occurred when the modal
combination was such that the base shear
approached zero or was in the opposite
sense to the base moment.
The results for all analyses are
summarised in Table 2. They are presented
in terms of: the maximum displacement
ductility factor at the top of the building.
u; the ratio of maximum base moment to '
code design base moment, M / M ; ; the ratio
of maximum base shear to code design base
shear, V
/ V a ; the minimum ratio of base
moment to base shear while the moment
curvature loop for the base element is on
the strain hardening branch, (Mj /Vj )min,
expressed as a proportion of the wall
height, H; and the maximum ratio of base
shear during strain hardening to base shear
at that base moment assuming a code lateral
load distribution - equivalent to V
of
Section 2. The values of V
for the
structures responding to the El Centro and
Bl earthquakes are in the range 1.3 to 1 . 6
and may be compared with the envelope
values for structures of similar period,
damping and overstrength given in Table 1.
Maximum flexural overstrength for the
inelastically analysed buildings are given
by the ratio of M
/ M a , although the overstrength when V
was recorded may be
rather less depending on the position on
the strain hardening branch of the momentcurvature curve. Allowance should be made
for an extra source of strength in all
three walls, relative to those of Table 1,
since they were designed on co-efficients
for public buildings rather than those for
Class III structures. As expected the
ratio V
increases for the stronger earthquakes , A2 and Pacoima Dam records.
m a x
c
m a x
=)
:>
r
r
m a x
r
m
a
x
r
The response characteristics of the
15-storey wall during a portion of the
time history of Fig. 3 is shown in Fig. 4.
The curves plotted are at increments in
the same yield excursion.
The applied
load diagram illustrates that during the
time interval the second mode contribution
has reversed from being additive to the
first mode near the base and subtractive
near the top, to being additive near the
top and subtractive near the base.
In
consequence the first increment gives
critical conditions for shear near the base
and the second increment gives critical
conditions for shear near the top.
Clearly,
a capacity design shear envelope must allow
for modal shear increases at the top as
well as at the base of the wall. The base
shear in this half cycle was at a value
greater than would be assessed from capacity
shear design, based on the code load
distribution, for a period of approximately
0.1 seconds.
The ratio of base moment to base
shear, M^/V^, represents a convenient
measure for assessing the variability of
the modal combinations at flexural capacity.
Frequency curves are plotted in Fig. 5 for
each of the structures using all values of
t / b recorded while the structures were in
the strain hardening range for all earthquake analyses. The frequency curves show
a skew distribution with the mode less
than the mean, that is positive skewness.
Wherever the value of M^/V^ is less than
that for an assumed code load distribution,
approximately 0.72H, the shears near the
base will be higher than that assumed from
this distribution at that moment.
Similarly,
where the values of M^/V^ are greater than
that applicable to the code load distribution
the shears near the top of the wall will be
greater.
M
V
As predicted in Section 2.4.2, plastic
hinges formed in a number of storeys above
283
the base in the taller structures. For the
analyses using the El Centro, Bl and A2
records hinging occurred at some stage in the
response in all elements up to maximum heights
of 8 storeys (plus a hinge at the eleventhstorey) for the 20-storey structure, 5
storeys for the 15-storey structure and 1
storey for the 6-storey structure.
Extent
of hinging was to almost twice these heights
under the Pacoima Dam record.
However, the
rotations imposed on the hinges at higher
levels were not large. The limiting height
of wall below which 90 per cent of the total
cumulative plastic rotation was suffered was
computed and gave average values as follows:
0.25 times the wall depth for the six-storey
structure, 0.5 times the wall depth for the
15-storey structure and 1.0 times the wall
depth for the 20-storey structure.
curvatures in a cantilever wall.
These
curvatures are unlikely to be reached in
most earthquakes. Therefore, since the
moment-curvature characteristics of
cantilever wall sections customarily have
an upward sloping post-yield branch the
calculated flexural capacity will usually
be rather higher than that which will be
achieved.
In lieu of a more accurate
assessment of likely maximum moment the
multiplying factors could be reduced by up
to 1 0 % . The buildings analysed to derive
the above factors were designed as public
buildings and would have design strengths
approximately 1.3 times those of similar
Class III buildings or 0.8 times those of
similar Class I buildings. The effect of
such changes in strength varies with period
of vibration of the structure and proportionate increases in the shear ratio V due to
such strength increases may be approximately
derived from Table 1. Appropriate values
of the multiplying factor, so derived for
all three classes of structure and
including the reduction discussed above,
are presented in the next section.
These
values are all based on a "design" earthquake for Zone A of the size of El Centro,
but they would have to be increased if an
earthquake of the size of A2 was considered.
As discussed in Section 2.4.1 these same
ratios would apply to buildings designed
for the co-efficients of New Zealand seismic
zones B and C.
r
4.
DESIGN
APPLICATIONS
The analyses of Sections 2 and 3 have
demonstrated the possibility of the modes
combining in such a way as to give shear
forces in cantilever walls considerably
greater than would be predicted from a codespecified distribution of applied loads
sufficient to cause flexural yielding.
It
has also been shown that flexural yielding
may extend for several storeys above the
base.
Observations of the structural
performance of some cantilever shear wall
structures during earthquakes have shown
extensive diagonal tension cracking often
over several storeys, two recent examples
being the Indian Hills Medical Centre during
the San Fernando 1971 earthquake ( ) and
the Hotel Trueba during the Veracruz 1973
eartyquake
Performance of other walls
may have been helped by conservative shear
design.
5
a
In Fig. 6 the envelope curves are plotted
for maximum shears recorded at all levels
during response of the three structures to
the El Centro and Bl earthquake records.
It may be seen that most values are enclosed
by curves derived by multiplying the shear
loads computed from a code load distribution
of applied loads at maximum flexural overstrength (that is maximum moment recorded
during the response since moment increases
with curvature during strain hardening) by
factors which took the following values:
1.4 for the 6-storey building, 1.6 for
the 15-storey building and 1.7 for the 20storey building.
Such curves could form the
basis of a design method in which, first,
the flexural capacity at the base of a wall
is computed, then, the magnitude of the
applied loads (assuming a code load
distribution) to achieve that capacity at
the base is calculated, and, finally, the
shears corresponding to the applied loads
are multiplied by factors such as those
above to give a shear design envelope.
It
is recognised that the shear design loads
derived from this approach can represent a
large shear steel requirement, particularly
at the base, where the contribution of the
concrete to shear resistance cannot be
relied on in design.
Some reduction of these
factors could be justified on the following
grounds.
Calculation of flexural capacity
assuming a concrete strain at the extreme
compression fibre of 0.003, as is common
practice for reinforced concrete sections,
gives a moment corresponding to very high
It may be argued that the code
^
allows energy dissipation in shear in walls
designed with a structural type factor, S,
of 1.6. Multiplication of the shear forces
by the factors suggested above and allowing
for overstrength can represent a more
severe shear design requirement for walls
dissipating energy in flexure than for
those dissipating energy in shear.
However,
the consequences of shear failure may well
be more serious in a ductile cantilever wall
than in a squat shear wall where shear deformations may be more readily controlled.
Also, the code
allows that the foundation
system need not be designed to resist
forces and moments greater than those
resulting from a lateral force corresponding
to S x M = 2, where M is the material factor.
Thus the foundation could act as a fuse before
the shear capacity derived from the suggested
factors is reached, but this could not always
be relied on. With all aspects considered
the design method suggested above is considered to be justified.
This approach encourages the designer to refine his flexural
design to minimise overstrength and is
preferable to a factored load approach
using S = 2, as allowed by the Code
,
since this may be unsafe in shear in many
cases.
Although the inelastic dynamic analyses
showed that plastic hinging can extend up to
half the height of a tall cantilever wall,
the major inelastic rotations were concentrated near the base.
In such circumstances
under cyclic loading in the member the
contribution of the concrete to the shear
resistance deteriorates and should not be
considered in design.
The 90 per cent limits
of total cumulative plastic rotations quoted
in the previous section may be taken as a
guide to the height of wall for which
concrete contribution to the shear strength
should be neglected.
In an actual wall the
284
plasticity may be spread further by
inclined flexure-shear cracks.
Above
this level it appears justifable to include
the concrete in shear design, but consideration should be given to provision of extra
ties for confinement of the concrete in
compression or for prevention of buckling
of the longitudinal steel wherever plastic
hinges may form.
The future of structural design appears
to lie in a probabilistic approach when more
is known of the applied loads and the
structural performance. For such an
approach the analyses reported here could
be extended to cover more buildings and
earthquake loadings to determine an
extreme value distribution of the parameter
V , representing the critical modal combination effects on shear loadings.
r
5.
CONCLUSIONS
An investigation was made of the
effects of the higher modes of vibration on
the maximum shears in cantilever walls when
they are loaded seismically to their flexural capacity.
It has been shown that the
various modes may combine in such a way as
to reach yield moment at the base of the
wall but with an applied load diagram such
that the centre of inertial load is very
much lower than would occur with the
normal code load distribution comprising
an inverted triangle with possibly a point
load at the top of 10% of the base shear.
Alternatively, the centre of inertial load
may be higher than assumed.
Consequently
a shear design
based on such an assumed
load distribution could underestimate the
imposed shear loadings, both at the base
and higher in the wall, at flexural
capacity and lead to a premature failure.
An elastic normal mode response spectrum
analysis was made of a 10-storey building
responding to an El Centro sized earthquake . The effect of the variables damping,
overstrength, and building period was
found on the ratio of maximum base shear at
flexural capacity, assuming a critical
combination of the first three modes of
vibration, to that assuming the lateral
load distribution specified by the code.
The value of this ratio was found to vary
from 1.4 to 4.0, increasing in turn with
increasing period of vibration, decreasing
damping ratio and decreasing flexural
overstrength.
The elastic normal mode response
spectrum approach was augmented by a
series of inelastic seismic analyses of
the time-history of response of three
cantilever shear wall structures of varying
height subjected to five different earthquake
acceleration records. The maximum value of
the ratio of actual base shear at yield to
base shear at that moment assuming code
lateral load distribution was found to be
1.4 for an El Centro 1940 N-S earthquake
and 1.6 for the Bl simulated record.
Greater values were computed for stronger
earthquakes.
It was found that plastic
hinging could extend well above the base of
a wall, up to half the height of a tall
structure.
However, plastic rotations
were concentrated near the base.
It is recommended that the design of
ductile cantilever walls be based on capacity
design principles and the following
ure is suggested:
proced-
(i)
Compute design base moment from either
the equivalent static force method or spectral
modal analysis as allowed by the code
.
(ii) Design and detail flexural reinforcement.
(In the capacity design approach
there are clearly advantages if the section
design is refined to minimise overstrength.)
(iii) Compute the capacity moment of the
section allowing a realistic assessment for
at least the following sources of overstrength: actual yield strength of steel
greater than minimum, specified, increases in
steel strength under reversed cyclic loading
beyond yield inherent in the Bauschinger
effect, actual concrete strengths greater
than the minimum specified, gravity loads
on the wall greater than the minimum design
case, neglect of the capacity reduction
factor incorporated in design, and the full
yield stress contribution of all web steel.
(iv) Determine a shear force
diagram
based on the code-specified distribution of
applied loads sufficient to reach capacity
moment of the wall at the base, such moment
being calculated as in (iii) above.
(v)
Multiply the shear force diagram
determed as in (iv) above by a factor which
has no less than the following values
according to class of structure:
No. of Storeys
1
6
10
15
to 5
to 9
to 14
to 20
Class I
1.0
1.2
1.4
1.4
Class II
1.2
1.3
1.5
1.6
This will give the shear design
Class I I I
1.3
1.5
1.7
1.8
envelope.
(vi)
In design of shear reinforcement neglect
the contribution of the concrete to the
shear resistance where concentrated plastic
rotations are anticipated.
The results of
this study give some indication of the
extent of such rotations. Above this level
consideration should be given to the use of
extra ties for confinement of the concrete
in compression or for prevention of buckling
of the longitudinal steel under moderate
plastic rotations.
ACKNOWLEDGEMENTS
The permission of the Commissioner of
Works, N. C. McLeod, to publish this paper
is acknowledged.
The opinions expressed
are not necessarily those of the Ministry
of Works and Development.
Grateful acknowledgement is made of the advice and guidance
received from O. A. Glogau, Chief Structural
Engineer, and B. W. Buchanan, Senior Design
Engineer, MWD.
K. E. Williamson, Lewis and
Williamson, Auckland first brought the
authors' attention to the problem assessed
in this paper.
Special acknowledgement
is made for his continued interest and helpful comments.
REFERENCES
1.
Standards Association of New Zealand,
"Draft New Zealand Standard Code of
Practice for General Structural Design
285
and Design Loadings", DZ 4203, Part 4,
Earthquake Provisions, March 1975.
50 pp.
Skinner, R. I., "Earthquake - Generated
Forces and Movements in Tall Buildings",
Bulletin 166, New Zealand Department of
Scientific and Industrial Research, 1964.
106pp.
Jennings, P. C., Housner, G. W. and
Tsai, N. C., "Simulated Earthquake
Motions for Design Purposes", Proceedings
of the Fourth World Conference on
Earthquake Engineering, V o l . 1 Session
A l , Santiago, Chile, 1969. pp. 145-160.
Kanaan, A. E. and Powell, G. H.,
"General Purpose Computer Programe for
5.
6.
Inelastic Dynamic Response of Plane
Structures", Earthquake Engineering
Research Centre, University of
California, Berkeley, Report No. EERC
73-6, April 1973. 101 pp.
Pinkham, C. W., "A Review of the Repair
of Two Concrete Buildings Damaged by
the San Fernando Earthquake", Journal
of the American Concrete Institute,
Proc. Vol. 70, No. 3, March, 1973.
pp. 237 - 241.
Irvine, H. M., "The Veracruz Earthquake
of 28 August, 1973", Bulletin
of the
New Zealand National Society for Earthquake Engineering, Vol. 7, No. 1, March
1974. pp. 2 to 13.
TABLE 1
RATIO OF MAXIMUM BASE SHEAR TO SHEAR DERIVED FROM CODE
LOAD D I S T R I B U T I O N AT FLEXURAL CAPACITY
M =1„4M.
o
a
M =1.2M,
o
a
PERIOD
T
M =1.6M.
o
a
(seconds)
X = 2%
X=5%
X=1G%
A=
2%
X=5%
X=10%
X=2%
X=5%
X=10%
0.2
1.58
1.55
1.53
1.49
1.47
1.45
1.43
1.41
1.39
0.4
1.93
1.71
1.63
1.80
1.60
1.54
1.70
1.53
1.47
0.6
2.36
2.02
1.83
2.17
1.87
1.71
2.02
1.76
1.62
0.8
2.82
2.32
2.05
2.55
2.13
1.89
2.68
1.99
1.78
1.0
3.33
2.68
2.32
3.00
2.44
2.13
2.75
2.26
1.99
1.2
4.02
3.20
2.76
3.59
2.89
2.50
3.27
2.65
2.31
286
TABLE 2
RESULTS OF I N E L A S T I C DYNAMIC ANALYSES
V
M
M
max
a
15 STOREY
CANTILEVER
WALL
20 STOREY
CANTILVER
WALL
0.97
1.2
0.6
TAFT
2.6
1.5
1.3
EL CENTRO
2.7
2.5
1.3
Bl
4.1
3.4
2.8
A2
6.6
5.7
4.3
PACOIMA
1.09
1.12
0.72
TAFT
1.22
1.18
1.14
EL CENTRO
1.26
1.26
1.16
Bl
1.36
1.28
1.22
A2
1.43
1.29
PACOIMA
1.04
1.24
0.83
TAFT
1.79
1.66
1.74
EL CENTRO
max
1.67
1.84
1.76
Bl
d
1.76
1.96
2.10
A2
2.44
3.09
3.35
PACOIMA
0.66H
M,
V
v
DAM
1.48
V
V
6 STOREY
CANTILEVER
WALL
;
b
min
V
r max
DAM
TAFT
0.53H
0.49H
0.57H
0.51H
0.45H
0.45H
Bl
0.46H
0.42H
0.38H
A2
0.33H
0.27H
0.23H
PACOIMA
-
DAM
EL CENTRO
DAM
1.12
-
1.34
1.42
1.26
EL CENTRO
1.40
1.56
1.60
Bl
1.55
1.53
1.89
A2
1.78
2.69
3.15
PACOIMA
TAFT
DAM
287
Pdssible capacity envelope
if wall overstrong near top
3 2 % mode I + 1 0 0 % mode 2
f 1 0 0 % mode 3
Design moment
(spectral modal analysis
1 0 0 % mode 2
Design moment
(equivalent static forces)
- 4 0 % model
+ 1 0 0 % mode 2
Capacity moment
( M = 14 M )
0
d
Capacity moment
M = 1-4 M )
0
0
APPLIED LOAD
(a)
(MM)
SHEAR
(b)
(MN)
-I
150
MOMENT
(MNm)
(c)
F I G U R E 1 : D E S I G N F O R C E S FOR 10 S T O R E Y CANTILEVER W A L L
(T-j = 0 8 0 S E C O N D S )
d
2 8 8
74H
0
1
2
3
4
6
5
8
7
9
10
11
TINE
12 13
I
U
(seconds)
1
,
,
r
15 16 17 18
19
20
F I G U R E 3: R E S P O N S E OF 15 S T O R E Y C A N T I L E V E R W A L L S T R U C T U R E TO B1 E A R T H Q U A K E
11 898secs
shear diagram at m a x i m u m
f l e x u r a l overstrength
for
-.ode load distribution.
11 813secs
30
20
10
0
-10
-20 - 3 0
DEFLECTION (mm)
-1
0
APPLIED
1
2
3
LOADS (MN)
U
10
SHEAR
5
(MN)
MOMENT (MNm)
F I G U R E 4: I N C R E M E N T S IN R E S P O N S E OF 15 S T O R E Y C A N T I L E V E R W A L L S T R U C T U R E T O B1 E A R T H Q U A K E
(a)
SHEAR (MN)
6 storey wall
S H E A R (MN)
(b) 15 s t o r e y w a l l
SHEAR ( M N )
(c) 20 s t o r e y wall
FIGURE 6 : M A X I M U M S H E A R ENVELOPES FOR EL C E N T R O A N D B1 E A R T H Q U A K E S