Anda di halaman 1dari 13



The Institution of Engineers,


Malaysia


Universiti
Teknologi MARA


Universiti Malaya

12th International Conference on Concrete Engineering and Technology


12 14 August 2014

Seismic Performance Evaluation of Reinforced Concrete


Buildings by Displacement Principles
Nelson Lam
Reader in Civil Engineering
Infrastructure Engineering
The University of Melbourne

John Wilson
Executive Dean
Faculty of Engineering and Industrial Sciences
Swinburne University of Technology

Abstract
The post-yield displacement capacity of a structure subject to the transient actions of an
earthquake can be used to trade-off its ultimate strength requirements meaning that the
strength demand on the structure can be reduced should the capacity to displace be increased.
Thus, the ability of a structure to deform is as important as its ability to resist forces and
moments in ultimate conditions. However, contemporary design practices for modelling
deflection of a reinforced concrete structure are mostly aimed at ensuring fulfilment of
serviceability requirements. Force-displacement capacity models for lightly reinforced RC
columns and structural walls are introduced in this paper in a hand calculation format which
is suitable for use in design practices. The determination of the force-displacement capacity
relationships forms part of a performance based seismic evaluation of the structure. The
presented calculation techniques have been verified by comparisons with experimental results
as reported in the literature.
Keywords: Seismic performance, displacement-based design, reinforced concrete

1.

Introduction

In designing reinforced concrete members for dead loads, live loads and wind loads bending
moment values of members at their critical cross sections are checked against the respective
design moment of resistance. Similar checks for shear and axial actions have to be
undertaken. This force based approach of design does not require the post-elastic deflection
of the member to be calculated. Contemporary design practices for modelling deflection of a
reinforced concrete structure are mostly aimed at ensuring fulfilment of serviceability
requirements in pre-yield conditions.
In reality, when a structure is subject to the transient actions of earthquakes, impact and blasts
its post-yield displacement capacity can be used to trade-off its ultimate strength
requirements meaning that the strength demand on the structure can be reduced should the
capacity to displace be increased. Thus, the ability of a structure to deform is as important as
its ability to resist forces and moments in ultimate conditions. However, in a conventional
force-based design procedure, practising engineers need not be involved in modelling post


19

elastic displacement given that the trading-off phenomenon as described can be taken into
account by what is known as the behaviour factor (or structural response factor) which is
dependent on prescriptive design and detailing requirements (Eurocode 8, 2004).
The behaviour factor approach which relies on empirical information, and experience, to
justify its recommendations is simple to apply but has the shortcomings of lack of
transparencies. There are plenty of scope for making improvements to conventional design
practices which are mostly restricted to comparing applied loads with strength capacities.
Whilst empirical data on limited ductile structures (and their performance in a low seismicity
conditions) is scarce a more direct, and transparent, approach to design is warranted.
There has been ever growing references to the use of displacement principles for modelling
seismic actions in structures including RC building structures. Seismic design checks
incorporating displacement as the guiding principles can be made reliable and simple to
comprehend, and apply, provided that the displacement demand and displacement capacity
behaviour of the structure is known. Importantly, the trading-off phenomenon as described is
modelled effectively by this approach. The writing of the text book on displacement based
method of seismic design by Priestley Calvi & Kowalsky (2007) represents a milestone of
achievement in the development of this approach in the global context.
With new modeling techniques that have been developed by the authors and co-workers for
over a decade it is now possible to obtain predictions of how much horizontal drift a building
is expected to experience for given projected earthquake scenarios and subsoil conditions.
(eg. Lam & Wilson, 2004; Wilson & Lam, 2003 & 2006; Lumantarna et al., 2010, 2012 &
2013). Reliable, and unbiased, predictions of the seismically induced displacement demand of
the structure can now be made even for countries where little strong motion data has ever
been collected for conventional empirical modelling. Meanwhile, research has also been
undertaken to develop reliable correlations of the state of damage sustained by a reinforced
concrete structure with the amount of drift it is experiencing (eg. Wilson et al., in press;
Wibowo et al., 2013 & 2014a & b). The risks of collapse of the building in an earthquake can
also be ascertained should the amount of drift to cause the structure to lose its gravity load
carrying capacity be known. Refer also Wilson & Lam (2008) for a state-of-the-art review on
this topic from the perspectives of low and moderate seismicity regions.
This paper aims to provide a succinct summary of calculation techniques that are suitable for
use in a design office for estimating the displacement of a reinforced concrete column, or
structural wall, at the limit state of yielding, loss of horizontal load carrying capacity and
vertical load carrying capacity (i.e. limit state of collapse). These techniques were derived
from materials presented in international peer review journal articles of Wilson et al. (2014)
and Wibowo et al. (2013 & 2014a & b). Rigorous evaluation of the techniques based on
comparison with results from laboratory experimentations have been presented in the cited
academic articles and are not reproduced in the paper. The techniques presented have been
verified for the following range of parameters for structural walls:
-

Aspect ratio (a) : 0.5 a 4.0


Axial load ratio (n) : 0.02 n 0.50
Longitudinal reinforcement ratio (v): 0.2% v 2.0%
Transverse reinforcement ratio (h): 0.0% h 1.0%


20

Schematic diagrams showing the monotonic force-displacement relationships for RC


columns and walls are shown in Figures 1a & 1b respectively. A force-displacement capacity
diagram representing structural walls and columns in support of a building can be
transformed and overlaid on an Acceleration-Displacement Response Spectrum Diagram
representing the projected seismic actions forming part of the seismic performance evaluation
process. Detailed descriptions of the methodology can be found in Wilson & Lam (2006).
Lightly reinforced moderate and slender walls (a>1)

Fu

Fu

Flf

Flf

Fy

Fy

Lightly reinforced squat walls (a < 1)

Faf
FCR

FCR

CR y

lf

af

CR y

(a) Columns

Figure 1

2.

peak

(a) Walls

Force-displacement models of RC columns and walls

Cracking (Fcr, cr)

The lateral moment of resistance to cracking MCR (and the associated lateral cracking strength
FCR) and displacement of a RC column, or a RC wall, can be calculated by employing
classical beam theory assuming linear elastic uncracked behaviour of the concrete in order
that gross-sectional properties may be assumed for all cross-sections. If provisions in the
Australian standard (AS3600) are to be adopted the design flexural tensile strength of
concrete may be taken as 0.6(f c) where f c is the characteristic cylinder strength of
concrete. Recommendations by another valid code of practice may be adopted as appropriate
given that the modelling principle is generic.
With lightly reinforced structural walls it is important to check the value of MCR against the
ultimate moment of resistance Mu of the critical cross section. Should the value of MCR > Mu
a single crack is likely to form at the base of the wall resulting in a potential concentration of
plastic deformation leading to non-ductile behaviour (Wibowo et al., 2013).
The curvature of the critical cross section at the state of cracking (cr) can be estimated
simply by dividing MCR by the flexural rigidity, EIgross, of the cross section where Igross is the
gross second moment of area. The amount of displacement, cr, of a singly loaded cantilever
model of length h is accordingly equal to cr h2/3 for a cantilever model and cr h2/6 for a
column with both ends fixed. The corresponding drift ratio, cr /h , is typically in the order of
0.10 % (Wilson et al., in press).


21

3.

First Yield (Fy, y)

For both RC columns and structural walls the moment of resistance of the critical cross
section at the state of initial yield, My (and the associated lateral initial yield strength Fy) may
be calculated using classical free body diagram analysis of the cross section which ensures
strain compatibility based on the assumption of plane sections remaining plane. The
calculations can be simplified further by assuming zero tensile strength of the concrete which
means ignoring contributions by tension stiffening which can be significant with lightly
reinforced and lightly axially loaded sections. This type of calculations can be implemented
by what is known as fibre element analysis which is operable in a user-friendly purposely
developed computational environment open to public access (eg. RESPONSE 2000;
CUMBIA 2007) or on EXCEL spreadsheets (Lam et al., 2011). For members subject to
significant axial pre-compression errors arising from the assumption of zero tensile strength
of concrete tends to be minor (Wibowo et al., 2014b). Furthermore, the effects of tension
stiffening is not so important in conditions of cyclic stress reversals which is expected in the
response of a structure to severe earthquake ground shaking (Wibowo et al., 2014b citing
recommendations by Park & Paulay, 1975).
An alternative simplified method of estimating the value of My circumventing the need of
fibre element analysis is to first undertake traditional (code based) calculations for
determining the ultimate flexural strength (Mu) of the cross section and then take My to be
equal to 0.8 times Mu (Wilson et al., in press).
Section curvature at first yield, y, can be found simply by dividing the calculated value of My
by the effective flexural rigidity, EIeff , of the cross section where Ieff is the effective second
moment of area and is not to be confused with the gross second moment of area, Igross. The
amount of displacement, y, is accordingly equal to y h2/3 for a cantilever model and y h2/6
for column with both ends fixed. According to review of the literature by Wilson et al. (in
press) the following recommendations have been made for rectangular cross-sections to take
into account the important influence of the axial pre-compression of the cross section :
(a) FEMA356 (2000)
Ieff = 0.7 Ig for axial load ratio n 0.5
Ieff = 0.5 Ig for axial load ratio n 0.3
Ieff = 0.5 Ig + (n 0.3) Ig for axial load ratio 0.3 < n < 0.5

(1a)
(1b)
(1c)

(b) Paulay & Priestley (1992)


100

+ n I gross
I eff =
f

(2)

The accuracies of results obtained from equations (1) and (2) have been evaluated by the
authors based on comparison with results from fibre element analyses. The comparative
analyses were based on a RC column with square cross section, longitudinal reinforcement
ratio, v, of 1% and axial pre-compression ratio ranging from 0 to 0.6 (Figure 2a). The
sensitivity of the moment-curvature relationships to changes in the value of axial precompression is evident from results of the simulations (Figure 2b). A bi-linear model was
then calibrated to match with each of the curves that were simulated from the fibre element
analyses. The effective stiffness value (EIeff) of the column was then taken as the slope of the



22

calibrated bi-linear model. It is shown in the comparative analyses that estimates made by
equation (2) were highly consistent with those from fibre element analyses (Figure 2c).
Estimates of EIeff based on the use of equations (1a) (1c), hence estimates for y and y,
have been found to be conservative meaning that estimates for the value of y could be
lower than the actual values. Estimates for y is expected to be particularly conservative for
shear dominated structural walls because shear deformation has been ignored in the
calculations. On the other hand, these same equations can underestimate the value of EIeff for
lightly reinforced structural walls (v << 1%) which are subject to a low level of axial precompression. It is noted that the listed equations can only be applied to RC columns and
structural walls of rectangular cross sections. Fibre element analyses will need to be
employed for modelling other types of cross-sections (Lam et al., 2011).
60

420

420

60

N24 bars
Grade 40 concrete
500 MPa Deformed Bars
(fy= 500 MPa fsu= 600 MPa)

(a) Cross-section

(b)

simulated moment-curvature relationships

(c) Ratio of EIeff/EIgross


Figure 2 Effective Flexural Rigidity of a Cracked RC Section

4.

Peak Resistance (Fu, u or peak)

For both RC columns and structural walls the peak moment of resistance, Mu, (and the
associated lateral peak strength, Fu) can be calculated using standard ultimate limit state
design procedures. For many rectangular, and circular, cross sections design charts are
available for determination of the value of Mu. The amount of ultimate displacement, u, of a
singly loaded cantilever model of length h is calculated as the sum of the yield displacement,



23

y, and plastic displacement, pl, which can be estimated using equation (3) the basis of which
can be illustrated using the schematic diagram of Figure 3. The value of u can be obtained
using equation (4) in the final step of the calculation.
Ku.d

0.004

u
(a) Curvature at onset of spalling

D/2

Lp
0.0025

u - y

y
(b) Yield Curvature

(c) Plastic Hinge Rotation

Figure 3 Curvature and Rotation of Plastic Hinge at Peak Conditions


pl = p Z

where

p = (u y )L p

(3a)

u =

(3b)

cu

or cu
K u .d K u ." w

(3c )

cu is the strain of concrete at the onset of spalling and may be taken as 0.004
K u d or K u " w is the depth of the compression zone
Eqn (3c) can be simplified into u =
d

0.024 0.024
or
based on taking k u = 1
6
d
"w

(3d )

is the effective depth of concrete; " w is the length of the structural wall

y =

2 y
D

or

2 y

(3e)

"w

y is the strain of steel at the onset of yielding; and is equal to 0.0025 for 500 MPa steel
D is the gross dimension of the column in the direction of loading
"
L p is the plastic hinge length which can be taken as D or w as a first cut estimate
2
2
f

L p = 0.2 u 1h + 0.1" w + 0.022 f y d b for structural walls


f

as per recommendations by Priestley, Calvi & Kowalsky (2007)


f u is ultimate strength of reinforcement

(3f)

d b is the diameter of the main longitudinal reinforcement


Z

is the distance from the point of loading and the centre point of the plastic hinge

u = y + pl

5.

( 4)

Ultimate Conditions (Flf, lf or m)

RC columns and lightly reinforced squat walls (a 1) can be susceptible to rapid


deterioration in their lateral resistance when displaced beyond the peak limit because of
degradation in shear strength. The loss of lateral resistance due to shear degradation is



24

illustrated in the schematic diagram of Figures 4a and 4b. The shear degradation model
presented in these figures can be represented by the algebraic expression of equation (5a).
Flf
Vtot

0.8Fu
= 1 k ( ) where = lf
Vtot
y

(5a)

0.3e5.7 n
based on curve fitting of experimental data (Wibowo et al.,2014b)
9a
n is axial load ratio
= 1 for a 3
= 2 for a 2
k=

(5b)

(5c)
(5d)

Algebraic manipulation of equation (5a) results in equation (6) which provides estimates for
the value of lf based on limited (20%) degradation in the lateral resistance of the member.

lf =

0.8 Fu
1 + k

k
Vtot

(6)

For moderate or slender RC walls the post peak lateral resistance is more robust. Thus, lateral
resistance Fu can be sustained up to the limit of m. Calculation for the value of m based on
the use of equation (3) is similar to the calculation for u except that equation (3c) which
defines the limiting value of curvature ductility is modified as follows:
su
0.06 0.072
to represent the limit of bar buckling
=
=
(1 K u )" w 5 " w " w
6
ccu
u =
to represent the limit of crushing of confined concrete
Ku" w

u =

ccu = 0.004 + 1.4 h sm

(6a)

(6b)

f yh

( 6c )

f 'c

where
h is the transverse reinforcement ratio
sm is the strain - hardening strain of the horizontal confinement steel
f yh is the yield strength of the horizontal confinement steel

Vtot

degraded shear strength

Initial shear strength

Vtot

Fu
Flf = 0.8 Fu

Limit of loss
of lateral
resistance

Fy

k
Force-displacement curve controlled by flexure

lf

(a) Limit of loss of lateral resistance

(b) Shear degradation model

Figure 4 Loss of lateral resistance due to shear degradation





25

6.

Collapse Conditions (Faf , af)

The displacement limit of axial failure of RC columns has been found to be most sensitive to
the axial load ratio (n) as revealed by results of parametric studies (Wibowo et al., 2014b).
Transverse reinforcement ratio (h) is influential too as shown in Figure 5. The drift limit
predictions are based on 50% degradation in the lateral resistance of the member (ie. from Fu
to Faf = 0.5Fu). In regions of low and moderate seismicity RC columns and walls are typically
unconfined. Thus, the lowest curve in Figure 5 is to be adopted for determining the value of
af for this class of columns. Longitudinal reinforcement ratio (v) is also an important
parameter but influence of this parameter is minor in conditions of high pre-compression.
Conservative estimates of the value of af (in terms of percentage drift) are summarised in
Table 1 for a range of axial load ratios based on recommendations made in Wibowo et al.
(2014b).

Figure 5 Axial load failure drift limit


Excerpt from Wibowo et al. (2014b)

Table 1 Conservative estimates of axial load failure drift limit


Axial load ratio
0.1
0.2
0.3
0.4

7.

Axial load failure drift limit (%)


3.5
1.9
1.1
0.9

Worked Example

The methodology for constructing a force-displacement capacity model is illustrated with the
worked example of a column which is 4 m in length, has both ends fixed, and is subject to an
axial force of 1000 kN. A typical cross-section of the column is shown in Figure 2a. Concrete
grade is 40.


26

Basic properties
1.5
E c = (24 ) 0.043 40 = 32 GPa
Gross sectional properties :
Ag = 0.42 2 m 2 ; I g =

0.42 0.42 3
= 0.0026 m 4
12

E c I g = 83200 kNm 2
n=

1000
P
=
= 0.14
2
Ag f 'c 0.42 40 10 3

f 't = 0.6 40 = 3.8 MPa


1000
P
=
= 5670 kN/m 2 or 5.7 MPa
Ag 0.42 2

At state of cracking

f 't + P = 0.42 3800 + 1000 = 117 kNm

Ag
6
0.42 2

M
117
= 0.0014
cr = cr =
E c I g 83200

M cr =

cr =

BD 2
6

cr h 2
6

Fcr =

117
= 58.5 kN
2

0.0014 4 2
= 0.0037 m or 4 mm approx.
6

At state of first yield


M u = 310 kNm from design charts for rectangular sections
Take M y 0.8M u = 250 kNm
My

250
= 125 kN
2
100
E c I eff 100

=
+ n =
+ 0.14 = 0.34 E c I eff = 0.34 83200 = 28290 kNm 2
500
E c I g f y

My
y h 2 0.0088 4 2
250
=
= 0.0088 y =
=
= 0.023 m or 23 mm approx.
y =
6
6
E c I eff 28290
Fy =

At peak

M u = 310 kNm from design charts for rectangular sections Fu =

0.024 0.024
=
= 0.067
p = (u y ) L p
u =
d
0.36
0.42
= 0.012 rad
p = (0.067 0.0088)
2
pl 0.023 + 3.6 0.012 = 0.066 m or 66 mm

u = y + pl

pl 2 p L p = 2 p 2

M u 310
=
= 155 kN
2
2

0.42
= 3.6 p
2

At ultimate (allowing for degradation of shear strength)


0.8M u = 250 kNm

0.8 Fu =

0.8M u 250
=
= 125 kN
2
2

0.8 Fu

1 + k
Vtot
k
e 5.7n
2
e 5.70.14
where k = 0.3
; n = 0.14 and a =
= 4.8 k = 0.3
= 0.16
9-a
0.42
9 - 4.8
Shear strength of section Vtot is estimated to be 300 kN assuming stirrup bar dia. of 12 mm and bar spacing of 300 mm

lf =

lf =

125
0.023
= 0.1 m or 100 mm approx.
1 + 1 0.16
300
0.16


27

At Collapse (loss of axial resistance)


0.5M u = 155 kNm
For n = 0.14

af =

af
h

100 = 3.5

0.5Fu =

0.5M u 155
=
= 77 kN
2
2

3.5 1.9
0.04 = 2.9 based on interpolating values listed in Table 1
0.1

2. 9
4 = 0.116 m or 116 mm
100

Results of calculations for the force-displacement pairs of values are summarized in Table 2.
The capacity curve is accordingly shown in Figure 6.
Table 2 Force-Displacement values of limit states of column in worked example
Limit state
Forces (kN)
Displacement (mm)
Cracking
58.5
4
First Yield
125
23
Peak
155
66
Ultimate
125
100
(loss of 20% of strength)
Collapse
77
116
(loss of 50% of strength)

Figure 6 Capacity model of column in worked example

8.

Conclusions

Capacity models for lightly reinforced RC columns and walls have been presented, and
illustrated by example, in a hand calculation format which is suitable for use in design
practices. Estimates for the values of the force-displacement pairs: Fcr - cr, Fy - y, Fu - u or
peak, Flf - lf and Faf - af required for the construction of the capacity models have been
covered. The determination of the force-displacement capacity relationships forms part of a
performance based seismic evaluation of the structure. The aim of this summary paper is to
illustrate calculation techniques which have been verified by comparisons with experimental
results as reported in the literature. A review of the academic literature in this topic and
details of the verifications of the presented relationships can be found in the cited references
and are not reproduced herein.



28

9.

Acknowledgements

The authors acknowledge the financial assistance provided by the Australian Research
Council with grants DP1096753 and DP0772088. The proof reading of the script and
checking of the calculations in section 7 by Dr Elisa Lumantarna is also gratefully
acknowledged.

10.

References

CUMBIA (2007) by Montejo, L.A. and Kowalsky, M.J., Set of Codes for the Analysis of Reinforced
Concrete Members, Report No. IS-07-01, Constructed Facilities Laboratory, North Carolina State
University, Raleigh, NC.
EN 1998-1, Eurocode 8: Design of structures for earthquake resistance Part 1: General rules, seismic
actions and rules for buildings, BSI, 2004.
FEMA 356 (2000) NEHRP guidelines for the seismic rehabilitation of buildings, Federal Emergency
Management Agency, Washington DC, USA.
Lam, N.T.K., Wilson, J.L. (2004) Displacement Modelling of Intraplate Earthquakes Invited paper,
International Seismology and Earthquake Technology Journal (special issue on Performance Based
Seismic Design; Ed Nigel Priestley), 2004. Indian Institute of Technology, Vol.41(1), paper no. 439: pp.
15-52.
Lam, N.T.K., Wilson, J.L. and Lumantarna, E. (2011). Force-Deformation Behaviour Modelling of
Cracked Reinforced Concrete by EXCEL spreadsheets , Invited paper in Special Issue entitled
Computing Technologies and Concrete Structures in Computers and Concrete. 8(1): 43 57.
Lumantarna, E, Lam,.N.T.K., Wilson, J.L. and Griffith, M.C. (2010). Inelastic Displacement Demand of
Strength Degraded Structures, Journal of Earthquake Engineering. 14: 487-511.
Lumantarna, E., Lam, N.T.K. & Wilson, J.L.(2013), Displacement Controlled Behaviour of
Asymmetrical Single-Storey Building Models Journal of Earthquake Engineering. 17: 902-917.
Lumantarna, E., Wilson, J.L. & Lam, N.T.K. (2012) Bi-linear displacement response spectrum model for
engineering applications in low and moderate seismicity regions Soil Dynamics and Earthquake
Engineering, 43: 85-96.
Park, R. and Paulay, T. (1975), Reinforced Concrete Structures, John Wiley & Sons Inc.
Paulay, T. and Priestley, M.J.N. (1992), Seismic design of reinforced concrete and masonry buildings,
John Wiley & Sons, Inc.
Priestley, M.J.N., Calvi, G.M. and Kowalsky, M.J.N. (2007), Displacement-Based Seismic Design of
Structures, IUSS Press, Pavia, Italy.
RESPONSE (2000) by Bentz, E.C., Sectional Analysis of Reinforced Concrete Members, PhD thesis,
University of Toronto.
Wibowo, A., Wilson, J.L., Lam, N.T.K. and Gad, E.F. (2014b), Drift Performance of Lightly Reinforced
Concrete Columns, Engineering Structures. 59: 522-535.
Wibowo, A., Wilson, J.L., Lam, N.T.K. and Gad, E.F. (2014a), Drift Capacity of Lightly Reinforced
Concrete Columns, Australian Journal of Structural Engineering. 15(2): 131-150.
Wibowo, A., Wilson, J.L., Lam, N.T.K. and Gad, E.F.(2013), Seismic performance of lightly reinforced
structural walls for design purposes, Magazine of Concrete Research. 65: 809-828.
Wilson, J.L., Wibowo, A., Lam, N.T.K. and Gad, E.F.(in press), Drift behaviour of lightly reinforced
concrete columns and structural walls for seismic design applications, Manuscript no. S14-002R1
accepted for publication on 12 May 2014.



29

Wilson, J.L., Lam, N.T.K. (2008), Recent Developments in the Research and Practice of Earthquake
Engineering in Australia, Special Issue of Invited papers on Earthquake Engineering, Australian Journal
of Structural Engineering. 8(1): 13-28.
Wilson, J.L., Lam, N.T.K.(2006), Earthquake Design of Buildings in Australia by Velocity and
Displacement Principles, Australian Journal of Structural Engineering Transactions, Institution of
Engineers, Australia. Vol. 6(2): 103-118. Awarded Warren Medal.
Wilson, J.L., Lam, N.T.K. (2003), A recommended earthquake response spectrum model for Australia,
Australian Journal of Structural Engineering. Institution of Engineers Australia, 5(1):17-27 (2003).


30

Anda mungkin juga menyukai