4th International Conference on Earthquake Engineering

Taipei, Taiwan
October 12-13, 2006

Paper No. 103

NONLINEAR CYCLIC AND EARTHQUAKE RESPONSE ANALYSIS OF

REINFORCED CONCRETE STRUCTURES

H.Z. Deng 1 , D. T. Lau 2 , H.A. Khoo 3 , and K.C. Tsai 4

ABSTRACT

This paper presents a simplified method based on energy formulation for nonlinear analysis of
reinforced concrete frame structures up to ultimate failure, which has been implemented in Open
System for Earthquake Engineering Simulation framework (OpenSees). In the simplified method, a
reinforced concrete member is modeled by an equivalent member of homogeneous nonlinear material
with a derived stress-strain relationship, which satisfies the requirement that the equivalent member
has the same moment-curvature behaviour as the original member. One advantage of the simplified
method is its simplicity which can be easily implemented in most structural analysis computer
program with nonlinear modeling capacities. The developed model can accurately predict the
nonlinear hysteretic behaviour of reinforced concrete structures with frame members of arbitrary
shapes and reinforcing details under severe earthquake excitations. Numerical examples of single
bridge column structures of regular reinforced concrete members or double-skinned concrete filled
tube members and a 2-story 2-bay reinforced concrete frame are analyzed using the simplified
method under monotonic, cyclic and earthquake loadings to demonstrate the validity and accuracy of
the simplified method. The effects of confinement, steel hardening, stiffness degradation and
softening, pinching, and strength deterioration are simulated in the developed method. A correlation
study has been carried out to compare the computer simulation results by the developed method with
the experimental measurements of a full scale 3-story 3-bay reinforced concrete steel frame tested at
the National Center for Research on Earthquake Engineering (NCREE) Taiwan, as a part of the joint
research between Carleton University and NCREE. Results from the proposed method agree well
with experiment test results and predictions from finite element and fiber models.

Keywords: nonlinear analysis, composite structures, reinforced concrete, finite elements, stiffness
degradation, cyclic behavior, earthquake loading, dynamics

INTRODUCTION

Many existing older structures are inadequate in their earthquake resistance and hence require
retrofitting or rehabilitation to conform to the current seismic design requirements. Seismic retrofitting
of existing structures is a complex problem. To properly design the retrofit scheme and to minimize
the cost, it is important to determine accurately the actual resistant capacity and behavior of structures
up to failure. The difficulty of accurately predicting the inelastic and ultimate behavior of reinforced
concrete structures arises from the many factors of material nonlinear behavior, such as cracking and

1
Former Graduate Student, Department of Civil and Environmental Engineering, Carleton University, Ottawa, Canada.
Email: [email protected]. Telephone: 613-520-2600 ext. 7470
2
Professor, Department of Civil and Environmental Engineering, Carleton University, Ottawa, Canada. Email:
[email protected]. Telephone: 613-520-2600 ext. 7473
3
Assistant Professor, Department of Civil and Environmental Engineering, Carleton University, Ottawa, Canada
4
Director, National Center for Research on Earthquake Engineering, Taipei, Taiwan
crushing of concrete, and yielding and strain hardening of reinforcing steel. The interaction between
concrete and steel under high stress state and large deformation, such as anchorage slippage of the
reinforcing steel and the confinement of core concrete, are also important phenomena that needed to
be considered in an accurate modeling of reinforced concrete structures beyond yielding. When
analyzing the behavior of reinforced concrete structures subjected to severe cyclic loading, the
modeling of the interaction of the concrete and steel can greatly affect the nonlinear behavior of the
structures (Park et al. 1972), as in the hysteretic behavior including stiffness deterioration due to
cracking of concrete (Takeda et al. 1971), the pinching behavior due to shear and bond deterioration
(Popov et al. 1972), the strength deterioration and softening due to accumulative severe large
deformations and damage on the reinforcing steel and concrete (Park et al. 1982).

The aim of the present study is to develop a simple modeling and flexible analytical procedure for the
evaluation of nonlinear behavior of reinforced concrete structural members up to the ultimate failure
that can be easily implemented in most structural analysis computer programs with nonlinear
modeling capabilities. Two formulations of the simplified method by direct equilibrium and strain
energy approaches can be applied. In the simplified method, the nonlinear behavior of a complex
reinforced concrete member predicted by an equivalent model member with a derived material stress-
strain relationship gives the same moment-curvature response as the original member. Only the strain
energy approach is presented in this paper. The direct equilibrium approach can be found in Deng et al.
(2003). The nonlinear behavior of concrete cracking, the behavior of concrete softening after peak
stress are considered, and the effect of confinement on the concrete strength contribution to the overall
behavior of the structures are taken into account. The elasto-plastic behavior of reinforcing steel with
strain hardening, and the interaction between concrete and reinforcing steel materials are modeled.
The equivalent model is implemented in the Open System for Earthquake Engineering Simulation
framework (OpenSees) (Mazzoni et al. 2005) to analyze the nonlinear behavior of reinforced concrete
structures. Under cyclic loading, the equivalent model can account for the hysteretic behavior of
reinforced concrete structures with consideration of stiffness degradation, pinching effects, strength
deterioration and softening. An empirical relationship has been developed to account for the influence
of axial load in a reinforced concrete member on the unloading stiffness of its equivalent member. The
proposed simplified method is validated by comparing the simulation results to the experimental
results of a double-skinned concrete filled tube member (Tsai et al. 2002), simulation results of a 2-
story 2-bay reinforced concrete frame (Mazzoni et al. 2005), and experimental results of a 3-story 3-
bay reinforced concrete-steel (RCS) frame tested at the National Center for Research on Earthquake
Engineering (NCREE) Taiwan (Chen et al. 2003 and Cordova et al. 2003). The reinforced concrete
frames are subjected to monotonic, cyclic and earthquake loadings. In the paper, the proposed
simplified model results are compared with experimental data, and numerical results by the fiber layer
model, or numerical results by the finite element model. The proposed simplified model results have
also been previously validated by Deng et al. (2003) for a cantilever reinforced concrete beam.

SIMPLIFIED METHOD: ENERGY BASED FORMULATION

Figs.1 and 2 illustrate the basic steps in the derivation of the specified stress-strain relationship of the
nonlinear equivalent rectangular member material from the original member section. For a rectangular
reinforced concrete member illustrated in Fig.1, the equivalent model has the same cross-sectional
area as the original reinforced concrete member section and the equivalent width beq is equal to the
sectional area divided by the height of the section. There are two approaches in the formulation of the
simplified method, namely, direct formulation and energy based formulation. In the direct formulation
as presented in Deng et al. (2003), selected states in the behavior of the reinforced concrete members
are analyzed to establish points on the specified stress-strain relationship of the nonlinear
homogeneous material model of the equivalent member directly by strain compatibility and
equilibrium of internal resultant forces. However, in the preliminary study by Deng et al. (2003), it
was found that the derived stress-strain relationship for the equivalent model determined from the
direct formulation in some cases can become unstable after the elastic limit is exceeded due to
difficulty in convergence. An alternative approach by the energy method is thus formulated to derive
the equivalent stress-strain relationship. Procedures similar to that for the direct formulation are
applied on selected states in the behavior of the reinforced concrete member. However, instead of
equilibrium of internal resultant forces, strain energy is used in the derivation of the specified stress-
strain relationship of the nonlinear equivalent member material. The strain energy based approach can
eliminate the instability in the derived stress-strain relationship and give a better prediction of the
reinforced concrete member behavior by the equivalent simplified model. In the energy based
formulation, the derived stress-strain relationship is determined by setting the strain energy and the
curvature of the equivalent section to equal that for the reinforced concrete member at selected states
shown in Fig.2. For structural member under flexure, the strain energy is determined by

M
U = ∫ ∫φ EI
L
d φ dx = ∫ ∫ε σ d ε dV
V
(1)

where E is the elastic modulus and I is the moment of inertia of the member. Linear relationship is
assumed for the stress-strain behavior between two consecutive modeling states. The strain energy of
an equivalent member of solid cross-sectional area can be expressed by strain energy density in terms
of equivalent stress and strain as

σ ( eq ) j + σ ( eq ) j −1 (2)
Δu ( eq ) j = {ε ( eq ) j − ε ( eq ) j −1}
2

where Δu(eq)j is the change of strain energy density from the (j-1)th to the jth state, σ(eq)j is the equivalent
stress at the jth state, ε(eq)j is the strain at the jth state, u(eq)j-1 and ε(eq)j-1 are the stress and strain at the (j-
1)th state, respectively. After the strain energy density Δu(eq)j is obtained from two consecutive states of
the reinforced concrete member, the equivalent stress σ(eq)j in Eq. (2) can be solved.

The axial load carried by a structural member not only alters the stress distribution in the member
cross-section, but also has an influence on the degrading behavior in its stiffness and strength under
cyclic loading. Roufaiel and Meyer (1987) have shown that the axial force increases the yield moment
of a member section. Numerical analyses conducted in this study also show that the axial load has
significant influence on the initial stiffness of the equivalent member under cyclic loading. In the
formulation here, the stress induced by the axial load is added to the equivalent stress as

P
σ (eq) j + (3)
Aeq

where σ(eq)j is the equivalent stress computed by Eq. (2), P is the axial load, Aeq is the cross-sectional
area of the equivalent member.

COMPUTER IMPLEMENTATION

In the present study, computer modules based on the computer-aided toolbox Mathcad have been
developed for the implementation of the simplified equivalent analysis model and procedures to
generate the required stress-strain relationship of the nonlinear equivalent member material. The
developed computer modules can be used in conjunction with other computer structural analysis
programs with user specified nonlinear material modeling capacities for detailed analysis nonlinear
reinforced concrete structural members of arbitrary shape. Modules of the simplified method for
monotonic and cyclic loading are implemented in the OpenSees. Many hysteretic models have been
developed over the years to model the hysteretic behavior of reinforced concrete structures under
cyclic loading. However, few models can represent all the important characteristics. The hysteretic
material model in the OpenSees is adopted and modified for analyzing the behavior of structural
members subjected to cyclic loading using the simplified method and is able to incorporate the
important phenomena that contribute to prediction of the nonlinear behavior of reinforced concrete
structures, such as stiffness degradation, pinching effect, strength deterioration and softening. The
derived equivalent material model properties are implemented in the hysteretic material model in
OpenSees, which has the hysteretic relationship as shown in Fig.3. The basis of the equivalent model
in modeling the cyclic response behavior of a reinforced concrete member is the equivalent stress-
strain relationship derived by the simplified method. Point 1 denotes the cracking state of the concrete,
point 2 denotes initial yielding of the reinforcing steel, and point 3 is the ultimate strength of the
reinforced concrete member. During cyclic loading, the unloading and reloading branches of the
hysteretic model are established by the stiffness degradation parameter β, the pinching effect modeling
parameter c1 and c2, and strength deterioration and softening parameter D. modeling of these
phenomena are discussed in following sections. The unloading path from point 3 to point 4 has an
unloading stiffness Ku. When reloading in the negative direction, the load path targets the pinching
point 5 according to the pinching parameters c1 and c2, which range from 0 to 1. On continuing
loading beyond point 5 in the opposite direction, the load path aims toward the yielding point 6. On
unloading from the maximum at point 7, the response follows with the same unloading stiffness Ku
toward point 8. During reloading in the positive direction, the load path targets pinching point 9, after
which it points towards a reduced maximum response at point 10 because of strength deterioration and
softening. The parameters controlling the hysteretic behavior of the model are presented as follow:

Stiffness Degradation

In this study, the unloading stiffness is defined by a modified degrading stiffness hysteretic model,
formulated by Saiidi and Sozen (1979)

δ m −β (4)
Ku = Ko ⋅ ( )
δy

where Ku is the unloading stiffness of the equivalent member, Ko is the initial stiffness, δm is the
maximum deformation attained in the direction of the loading during the previous loading cycles, δy is
the deformation of the equivalent member at yielding, β is a parameter that determines the degraded
unloading stiffness. In the study by Roufaiel and Meyer (1987), it was shown that the axial force
increases the yield moment of a section but reduces the ultimate failure curvature considerably.
Furthermore, for a damaged reinforced concrete column member, analysis results by Kwan et al.
(2001) showed that the elastic unloading behavior of steel dominates the unloading behavior of
deteriorated member under cyclic loading. In this study, analyses of the example structures indicate
that the axial load has significant influence on the unloading stiffness of the equivalent member
subject to cyclic loading. A proposed equation for the unloading stiffness parameter β of the
equivalent member is expressed as

P ρf y
b 0 + b1 + b2 ' (5)
Po fc
β =
P ρf y
1 + b3 + b4 '
Po fc

where P is the axial load, Po is the maximum axial load resistance, ρ is the reinforcement ratio, fc’ is
the compressive strength of concrete, fy is the yield strength of the reinforcing steel, b0, b1, b2, b3, and
b4 are the constants, which are calibrated by the steel concrete fiber model to be 0.030, 14.02, 0.151,
15.54 and 7.075. This empirical expression for β is used in the determination of the unloading stiffness
for the equivalent model of the structure members subject to cyclic and earthquake loading in this
study.

Pinching Effect

For columns or beams with low shear span to depth ratio, D’Ambrisi and Fillipou (1999) have found
that the pinching effect is an important phenomenon in modeling the hysteretic behavior of the
reinforced concrete structures. Laboratory tests conducted on reinforced concrete specimens by Popov
et al. (1972) have found that there is a strong correlation between the degree of pinching and the
magnitude of shear at the section, and that pinching effect reduces the load resistance of the member
during reloading. D’Ambrisi and Filippou (1999) have proposed a pinching model that defines a target
pinching point in the reloading process based on the maximum loading point in the reloading cycle.
The pinching point (δp, Fp) 9 in Fig3 is determined by the pinching parameters c1 and c2 as

δ p = ( c 1 − 0 . 5 ) δ o , max (6)
F p = c 2 Fm (7)

where δo,max is the sum of the absolute values of two unloading deformations from the previous load
cycle, and Fm is the maximum strength attained in previous load cycle in the reloading direction as
shown in Fig.3. In this study, the values of pinching parameters of c1 and c2 are set based on the
experimental observation and results.

Strength Deterioration and Softening

Experiments have shown that the rate of strength deterioration and softening under cyclic loading
increase with axial force and lateral deformation (Park et al. 1982). Park and Ang (2003)
recommended that the strength degradation and softening be controlled by a hysteretic damage index,
which is a function of ductility and absorbed hysteretic energy that gives

δ D = δ m (1 + D ) (8)
FD = Fm (1 − D) (9)

where δD and FD are the current maximum deformation and strength due to strength deterioration and
softening respectively as shown in Fig 3, D is the current value of the hysteretic damage index for the
deformation and strength as shown in Fig.3. In this study, the value of damage index D is set based on
the experimental observation and results.

NUMERICAL EXAMPLES

To demonstrate the validity of the equivalent model and analysis procedure, numerical examples of
reinforced concrete structures are analyzed using the simplified method. The equivalent model is
implemented into OpenSees for analysis of the nonlinear behaviour of reinforced concrete structures.
Simulations of the static monotonic and cyclic loading behavior of the numerical examples are carried
out using the steel concrete fiber model and the equivalent model. For both models, the structural
members are modeled in OpenSees by the flexibility-based nonlinear beam-column fiber element,
which is based on the non-iterative force formulation and the spread of plasticity along the element.
This allows the columns and beams to be modeled by a single element without discretization error in
computation (Neuechofer and Fillipou 1997). In the simulation with the steel concrete fiber model, the
material models are steel02 and concrete01 are used. An equivalent member of homogeneous material
properties is derived in the simplified method for each structural member. The numerical examples
and the comparison of the analysis results are presented below.

Double Skinned Concrete Filled Tube (DSCFT) member

For the DSCFT member (Tsai et al. 2002) with thickness of 2mm for the external and internal steel
tubes as shown in Fig.4, a parametric study on the confinement coefficient k is carried out by
increasing k from 1.231 as determined by the modified Kent and Park (1982) confined concrete model
based on the spiral reinforced steel, to 2.5 and 3.7 because the two concentric circular thin steel tubes
provide better confinement compared to the spiral reinforcing steel. Material properties and hysteretic
parameters used in the analysis are shown in Table 1. Results in Figs.5 shows that the behavior of the
DSCFT member can be more accurately represented by the confinement coefficient k=3.7 when
compared to test results by Tsai et al. (2002) in monotonic loading. A comparison of the cyclic
analytical results obtained by equivalent model in OpenSees based on the derived stress-strain
relationship of the equivalent member material with the confinement coefficient k= 3.7 and test results
by Tsai et al. (2002) is shown in Fig.6. It is observed that the simplified model consistently predicts a
lower conservative moment capacity. The predicted unloading behavior of the DSCFT is in good
agreement with the experimental results. The pinching effect and the ultimate moment capacity
predicted by the computer simulation agree reasonably well with the test data. However, beyond the
peak moment, the actual behavior of strength and softening of the DSCFT member decays much
rapidly than the equivalent model.

2-Story 2-Bay Reinforced Concrete (2B2B RC) Frame

To calibrate the hysteretic parameters of the equivalent model implemented in OpenSees for the
analysis of nonlinear behavior of the reinforced concrete structures, a full-scaled 2-story 2-bay RC
frame (Mazzoni et al. 2005), which consists of reinforced concrete columns and beams, are analyzed
for push-over and cyclic loading simulations. Only the result of cyclic loading simulation is discussed
here. The material properties and the dimensions of the structural member sections are presented in
Tables 2. In the cyclic loading simulation, a lateral displacement control iteration scheme of a linear
time series, and the gravity loads distribution of the frame shown in Fig.7 are considered in generating
the behavior of the frame. The cyclic loading pattern is shown in Fig.8.

In the cyclic loading simulation, the empirical expression β for the stiffness degradation, the pinching
modeling parameters c1 and c2 and the strength deterioration and softening parameter D are employed
to predict the response of the frame. As shown in Fig 9, the simulation results of the 2-story 2-bay RC
frame under reversed cyclic loading by the equivalent model agree well with results by the steel
concrete fiber model. The major characteristics in the hysteretic behavior of reinforced concrete frame
under cyclic loading are captured, such as the initial loading stiffness, the maximum deformation at
each loading cycle, the stiffness degradation at unloading, and the pinching effect in reloading..

3-Story 3-Bay Reinforced Concrete Column with Steel Beam (3S3B RCS) frame

For further validation of the accuracy and modeling capability of the equivalent model of reinforced
concrete frame simulation under earthquake loading, a correlation study is carried out to compare the
simulation results with the test data by Chen et al. (2003) at the National Center of Research on
Earthquake Engineering (NCREE) on a full-scale 3-story 3-bay RCS frame, as shown in Fig.10. The
design and construction details of the frame can found in Chen et al. (2003) and Cordova et al. (2003).
The full-scale RCS frame was tested and loaded pseudo-dynamically at NCREE to monitor the
structural response under ground motions obtained by scaling the Chi-Chi ground motion record to the
earthquake hazard level corresponding to 50%, 10%, and 2% chance of exceeding in 50 years. The
first test results of the 50% chance of exceeding in 50 years are compared with the simulation results
of the equivalent model implemented in OpenSees. The material properties and the hysteretic
parameters are given in Table 3.

The simulations results, shown in Figs.11, correlate well with the experimental data, except that after
the intensive shaking of about 27 seconds, the roof displacement of the frame from experimental data
is larger than that predicted by the simulation results. This is probably due to the initial damages
already sustained by the frame in some preliminary tests as reported Chen et al. (2003) and Cordova et
al. (2003), resulting in the immediate deterioration and softening in the frame capacity after the
intensive loading during the earlier part of the loading time history in the actual structure during the
test that is not duplicated in the computer model simulation results. In Fig.12, the correlation of the
corresponding base shear results of the test and simulation by the equivalent model show that the
equivalent model can effectively capture the major characteristics in the earthquake response of the
frame. The equivalent model slightly overestimates the base shear response of the frame. The
discrepancies between the experimental data and the simulation results are attributed to the stiffness
decay due to damages indicated by Chen et al. (2003) and Cordova et al. (2003) on the structure
components, buckling of the longitudinal reinforcement of the reinforced columns and local buckling
of the steel beam. In OpeenSees, the behavior of columns splice and the bolted splice, which may have
contributed to some dissipation of the seismic energy in the actual frame are not modeled. The
stiffness of the actual frame could have been decayed by the damage, thus the actual experimental data
give larger displacement after intensive excitation period and smaller base shear. Simulations using the
more time consuming steel concrete fiber model have also carried out for the frame. Although results
are not shown, it is found to be similar to that from the equivalent model.

CONCLUSIONS

This study presents a simplified modeling and flexible analysis approach for the nonlinear analysis of
reinforced structural members. Numerical examples are presented to verify the procedures of the
simplified analysis method and the derived equivalent model, which includes DSCFT member
subjected to a lateral load, a full-scaled 2-story 2-bay RC frame subjected to the push-over and cyclic
loading, and a full-scaled 3-story 3-bay composite RCS moment frame under earthquake excitation.
Analytical results obtained from the simplified method are in good agreement with the fiber model in
OpenSees and the experimental results. The research shows that the nonlinear response of reinforced
concrete members can be accurately predicted by the simplified method, and that the simplified
method is an efficient tool for evaluating the nonlinear behavior of reinforced concrete structural
members up to ultimate failure. The advantage of the simplified method is that it can be easily
implemented with most structural analysis programs that have the capabilities of user specified
nonlinear material models. The developed model can accurately predict the nonlinear hysteretic
responses of reinforced concrete and other reinforced concrete structures with frame members of
arbitrary shapes and reinforcing details under severe earthquake excitations.

ACKNOWLEDGEMENTS

The research supports provided by the Natural Sciences and Engineering Research Council Canada
and the National Center for Research on Earthquake Engineering (NCREE) Taiwan, and the National
Science Council Taiwan for the collaboration between Carleton University and NCREE are gratefully
acknowledged. Technical information provided by M.L. Lin, Y.S. Yuan and C.H. Chen of NCREE are
appreciated. Special thanks to Shadi Ostovari for her initial works and the technical assistance from
Kate Chang on this simplified method.

REFERENCES

Chen, C.H., Lai, W.C., Cordova, P., Deierlein, G.G., Tsai, K.C. (2003). Pseudo-dynamic test of full-scale rcs
frame: Part I – Design, construction, and testing”. Proceeding of the International Workshop on Steel and
Concrete Composite Construction, Taipei, Taiwan. pp.107-118.
Cook, R.D., Malkus, D.S., and Plesha, M.E., (1989). Concepts and applications of finite elements analysis. 3rd
Edition. John Wiley & Sons, Inc. N.Y. USA.
Cordova, P., Chen, C.H., Deierlein, Lai, W.C., G.G., Tsai, K.C. (2003). Pseudo-dynamic test of full-scale rcs
frame: Part II – Analysis and design implications. Proceeding of the International Workshop on Steel and
Concrete Composite Construction, Taiwan. pp.118-131.
D’Ambrisi, A. and Filippou, F.C. (1999). Modeling of cyclic shear behavior in RC members. ASCE Journal of
Structural Division, 125(10): pp.1143-1150.
Deng, H.Z., Chang, Y.Y., Lau, D.T, Ostovari, S., and Tsai, K.C. (2003). A simplified approach for nonlinear
response analysis of composite structural members. Proceedings of the International Workshop on Steel and
Concrete Composite Construction, Taipei, Taiwan. pp.207-216
Kent, D.C. and Park, R. (1971). Flexural members with confined concrete. Proceedings of the American Society
of Civil Engineers Journal of Structural Division, 97(ST7): 1969-1990.
Kwan, W.P., and Billington, S.L. (2001). Simulations of structural concrete under cyclic load. ASCE Journal of
Structural Engineering, 127(12): 1391-1401.
Mazzoni, S., Mckenna, F., and Fenves, G.L. et al. (2005). OpenSees command Language manual. Release 1.6.
Pacific Earthquake Engineering Research Centre, University of California, Berkeley, Calif.
Neuechofer, A. and Filippou, F.C. (1997). Evaluation nonlinear frame finite element models. ASCE Journal of
the Structural Engineering, 123(7): 958-966.
Park, R., Kent, D.C., and Sampson, R.A. (1972). Reinforced concrete members with cyclic loading. ASCE
Journal of Structural Division, 98(7): 1341-1360.
Park, R., Priestley, M.J.N., and Gill, W.D. (1982). Ductility of square-confined concrete columns. ASCE Journal
of Structural Engineering, 108(4): 929-950.
Popov, E.P., Bertero, V.V. and Krawinkler, H. (1972). Cyclic behavior of three r.c. flexural members with high
shear. EERC Report 72-5, Earthquake Engineering Research Center. University of California, Berkeley,
Calif.
Roufaiel, M.S.L., and Meyer, C. 1987. Analytical modeling of hysteretic behavior of r/c frames. ASCE Journal
of the Structural Engineering, 113(3): 429-444.
Saiidi, M.,and Sozen, M.A.(1979).Simple and complex models for nonlinear seismic response of reinforced
concrete structures. Structural Research Series 465, Civil Engineering Studies, University of Illinois, Urbana,
I11.
Takeda, T., Sozen, M.A., and Nielson, N.N. 1970. Reinforced concrete response to simulated earthquakes,”
Proceeding of the ASCE Journal of the Structural Division, 96(ST12): 2257-2573.
Tsai, K.C., Lin, M.L., Lin, Y.S., and Wang, T.F. 2002. Double skinned concrete filled tubes for bridge piers.
Research Paper, National Taiwan University and Sinotech Engineering Consultant, Taipei, Taiwan.

Table 1. Material properties of DSCFT member

Concrete E c ( MPa) f c' ( MPa) f cr' ( MPa) εc Hysteretic β 0.425
25045 28 3.175 0.002 parameters c1 0.05

Steel E s (MPa) f y ( MPa) f u (MPa) εu in equivalent c2 0.05

201000 305.2 416.1 0.10 model D 0.005

Table 2 Dimension and material properties Table 3 Dimension and material properties
of the 2s2b rc frame members of the 3s3b rcs frame members
1E1,1E2,2E1,2E2 1I1,2I1 1B1,1B2,1B1,2B2 1E1,1E2,3I1,3I2 1I1,1I2,2I1,2 2E1,2E2,3E1,3E2
I2
12 #8 Bars 8 #11 Bars 12 #11 Bars 4 #11 Bars
12 #11 Bars 8 #9 Bars
75mm 75mm 75mm
75mm 75mm 75mm 650mm 650mm 650mm
610mm 686mm 610mm
175mm

650mm 175mm 650mm 650mm

610mm 610mm 610mm

Concrete Steel Concrete Steel

E c (Mp 23630 E s (Mpa) 206843 Ec (Mpa) 30018 Es (Mpa) 210000
f c' (Mpa 27.58 f y (Mpa) 413.7 '
f c (Mpa) 44.5 f y (Mpa) 527
0.002 Strain hardening Strain
εc 0.02 εc 0.002 0.02
ratio b hardening
Hysteretic parameters in equivalent model Hysteretic parameters in equivalent model
β c1 c2 D β c1 c2 D
1E1,1E2,1 0.425 0.0 0.0 0.0 1E1,1E2 0.225 0.001 0.001 0.005
2E1,2E2,2 0.275 0.0 0.0 0.0 2E1,2E2 0.325 0.001 0.001 0.003
1B1, 1B2 0.275 0.0 0.0 0.0 3E1,3E2 0.200 0.001 0.001 0.003
2B1, 2B2 0.275 0.0 0.0 0.0 1I1,1I2 0.255 0.001 0.001 0.005
2I1,2I2 0.255 0.001 0.001 0.003
3I1,3I2 0.225 0.001 0.001 0.003

b ε b eq ε (e q )(i-2 ) ε (e q )(i-1 ) ε (e q )(i) σ (e q )(i-2 ) σ (e q )(i-1 ) σ (e q )(i

a h (i-
c ΣδΜ n
Mi b
ΣδΜ n
h
d φi h
φi

(a) Reinforced concrete section (b)Equivalent member

Fig.1. Equivalent stress and strain
 b eq
σ
fs 2 4 6 4 5
fc 2 3 6
h 3 h
5 εs
1 1
εc
(a)R.C. section (b) Concrete model (c) Steel model (e)Equivalent section (f) Equivalent stress-strain
Typical states of M M
R.C. member 4
3 5
1. Cracking of concrete 2 6
2. Yielding of steel
3. Peak stress in concrete 1
4. Steel hardening
5. Residual stress in concrete
φ φ
6. Steel ultimate fracture
(d) R.C. member moment-curvature (g)Equivalent member moment curvature
Reinforced concrete member Equivalent member
Fig.2. Simplified analysis model of reinforced concrete member section

3
10
Fm 2
Fy (δ D, F D )

1100 mm
to
1 Concrete
9 Ku
8 Ko ti
0 4

δ o, max
δy δm δ
5
Di
180 mm
De
300 mm

6
7

Fig.3. Equivalent model for cyclic loading Fig.4. DSCFT member (Tsai et al. 2002)

200
140

120 150

100 100
M o m e n t (k N -m )

80 50
Moment (kN-m)

60
0
-10 -5 0 5 10
40 Experimental Results
-50
Energy Approach k=3.7
20 Energy Approach k=2.5
Energy Approach k=1.231 -100
0
0 5 10 15 20 25 30 -150 Simplified method
-2
Curvature (10 rad/m) Experimental (Tsai et al. 2002)
-200
-2
Curvature (rad*10 /m)

Fig.5. Moment-curvature of DSCFT member Fig.6. Results of cyclic loadings for DSCFT member
 P/2 P P/2
2B1 2B2
H
150

P e a k D is p la c e m e n t (m m )
100
2E1 2I1 2E2
50
P 2P P
0
H/2 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
1B1 1B2 -50

-100
1E1 1I1 1E2
-150

7000mm 7000mm

Fig.7. 2s2b rc frame Fig.8. Cyclic loading pattern for 2s2b rc frame

3B1 3B2 3B3

1500

1000 3E1 3I1 3I2 3E2

2B1 2B2 2B3

B ase S hear (kN )

500

0
-150 -100 -50 0 50 100 150 2E1 2I1 2I2 2E2
-500

-1000
Steel Concrete Fiber Model 1B1 1B2 1B3
Equivelent Model
-1500 1E1 1I1 1I2 1E2

Roof Displacement (mm)

7000mm 7000mm 7000mm

Fig.9. 2s2b rc frame cyclic loading simulation result Fig.10. 3-story 3-bay rcs frame (NCREE)

2500
Chi-chi TCU082 EW, PGA=0.276g
200 2000
Chi-chi TCU082 EW, PGA=0.276g
R o o f D is p la c e m e n t (m m )

150 1500
1000
B a s e S h e a r (k N )

100
50 500

0 0
-500
-50
-1000
-100
Equivalent model -1500
-150 Equivalent model
Experimental results -2000
-200 Experimental results
0 5 10 15 20 25 30 35 40 45 -2500
0 5 10 15 20 25 30 35 40 45
time (s) time (s)

Fig.11. 3s3b rcs frame roof displacement under Fig.12. 3s3b rcs frame base shear under Chi-
Chi-chi TCU082 EW CC50/50 excitation chi TCU082 EW CC50/50 excitation