Effects of Confined Concrete Models On Simulating RC

13th World Conference on Earthquake Engineering
Vancouver, B.C., Canada

August 1-6, 2004
Paper No. 1498
EFFECTS OF CONFINED CONCRETE MODELS ON SIMULATING RC

COLUMNS UNDER LOW-CYCLIC LOADING
Zongming HUANG 1 Wenfeng ZHOU 2 Tao CHEN1 Juan CHEN1
SUMMARY
Material models are the base of computer simulation analysis of structure. Many confined concrete models
have been put forward until now; however, there are few comparative studies on effect of confined
concrete models on structural member analysis. In this paper, four models are selected as the
representatives and applied to nonlinear dynamic analysis program for space frames which is based on the
beam-column element using the finite element flexibility method and the fiber model. Park’s test on
columns that are combined axial and bending under cyclic loading are taken as the calibrations,
comparisons to material and structural member level are drawn and the analytical results of various models
are discussed. From this study it is concluded that despite the significant difference in the stress-stain
curve of each confinement model the difference in structural member analysis reduces obviously but still
remarkable. The simulation using the Mander model shows better agreement with the experimental results.
INTRODUCTION
The accuracy of nonlinear analysis of RC structure depends strongly on material models. Various
analytical models available in the literature for confined concrete are mostly limited to material level and
comparative study on applicability of the different concrete model to RC member anaysis are rarely made;
however, engineer community would pay more attention to the accuracy of the model in simulation at
structural or member level. An ideal confined concrete model should be the following: 1. good precision in
simulation at member or structural level; 2. to reflect the dominant aspects of confinement, fewer
parameters, and explicit in physics.
In this paper, four models are selected as the representatives and applied to nonlinear dynamic analysis
program for space frames which is based on the beam-column element using the finite element flexibility
method and the fiber model. On the basis of experimental results by Park [2], simulation and comparison
are made on combined axial and bending columns under cyclic loading at material level. The effect of
different hysteretic rules and skeleton curves on simulation results are investigated at material and member
level respectively.
1
.College of Civil Engineering, Chongqing University, Chongqing, China.
2
.Quality-monitoring Station in Guangzhou, Guangzhou, China.
This work was sponsored by National Natural Science Foundation of China under Grant No.59978055
RIENFORCED CONCRETE COLUMN SPECIMEN BY PARK
In Reference 2, the effect of some factors such as different tie configuration, ratio of axial compress force
to strength, volumetric ratio of transverse reinforcement and tie yield strength and concrete strength, etc
had been taken into account on combined axial and bending columns under cyclic loading. The ratios of
longitudinal reinforcement of four members are all 0.0179 and their yield strengths are 375Mpa. Fig.3
shows the scheme of the experiment and other more details can be referred to Reference 2.
Tab. 1 Test data shown in Reference 2
ratio of axial
fc Volumetric ratio
axial load compress section fyh tie spacing
(Mpa of transverse
specimen (KN) force type (MPa) (mm)
) reinforcement
to strength
1 23.1 1815 0.260 A 0.015 297 80
2 41.4 2680 0.214 A 0.023 316 75
3 21.4 2719 0.420 B 0.020 297 75
4 23.5 4265 0.600 B 0.035 294 72
FIBER-MODEL BASED BESM-COLUMN ELEMENT
The experimental results of column specimens by Park were studied by applying a nonlinear analytical
method to a RC section, i.e., fiber model. As is well known, the RC section would be divided into many
fibers paralleled to the longitudinal axis, adopting Bernoulli-Euler’s assumption, the section’s moment
curvature relationship was set up based on the uniaxial stress-strain relationship of the concrete and the
longitudinal bar. The influence of shear and bond slip between the concrete and steel was neglected. In the
concrete model, the difference in the compressive strength and descending branch between the core
concrete and the cover concrete were considered. The beam-column element is cast in the framework of
the finite element flexibility based beam theory by Spacone et al [9]. Finally, the Menegotto and Pinto law
modified by Filippou et al [10] to include isotropic strain hardening is used for the reinforcing steel.
Fig. 1 Details of test specimen Fig. 2 Comparison of stress-stain relationships

calculated from different models
CALCULATION BASED ON EACH MODEL FOR SPECIMEN 1 AT MATERIAL LEVEL
In the light of the significant difference in various confinement models at material level, only the
computing results of four models for Specimen 1 are presented in this paper (Tab. 2 and Fig. 2). The
results show that:
(1) Because each model is put forward based on its own experiment, the behavior of concrete material
scatters, and experimental technique and measurement are not unified, each model is quite different at
material level. For Specimen 1, the difference of fcc amounts to 39.95/23.5=1.7; the difference of ε cc
reaches to 0.0075/0.0022=3.41.
(2) Dispersion of descending branch is greater than that of ascending branch. The Mander model shows
good ductility of the material, but the Gill model and the Zhang model show relatively poor ductility of
the material.
COMPARISON OF CONFINED CONCRETE AT MEMBER LEVEL
Effect of hysteretic rules

In order to discuss to effect of different hysteretic rules, a comparison is drawn in different confined
concrete models by adding hysteretic rules to them. Simulative calculations of Specimen 1 based on the
Mander model after adding different hysteretic rules are only presented here. These rules are used in this
paper:
(1) Rule 1: Because the calculated stiffness of the unloading curve is larger than Ec when ε U approaches
to ε cc, borrowing ideas from the Blakeley model, the unloading curve less-than ε cc is modified to
straight line and the rate of slope is Ec. The specification of ε re and assumptions of concrete in tension
are retained in the Mander model (Fig. 3).
(2) Rule 2:Holding the specification of ε pl and ε rein the Mander model, the unloading curve are modified
to two-fold line when ε U is greater than ε cc (Fig. 4), and
ε Q = 0.5 × (ε pl + εU ) , σ Q = 0.4 × σ U
(3) Rule 3: Holding the specification of ε pl and ε re in the Mander model and adding Blakeley-hysteretic
rules (Fig. 5). More details can be referred to Reference 8.
(4) Rule 4: New hysteretic rules is presented in this paper: the unloading curve is modified to two-fold line
when ε U is greater than ε cc(Fig. 6), and
ε pl = 0.5εU , ε Q = 0.75εU , σ Q = 0.4σ U
Fig. 3 Rule 1 Fig. 4 Rule 2
Fig. 5 Rule 3 Fig. 6 Rule 4

The following can be found in Fig. 7:
(1) When using the rule 3 to simulate, strength drops suddenly in the unloading curve of Blakeley
hysteretic rules, which causes capacity and ductility of the member to reduce and the specimens have
been destroyed at u= 6.
(2) When the deformation is small, the simulative results of three other hysteretic rules almost coincide
with each other, and they are yet consistent when the deformation is relatively greater.
(3) Hysteretic rules proposed in this paper could simulate well hysteresis characteristic of specimens,
which is easy to use and code. To avoid the effect of hysteretic rules, the hysteretic rules of two-fold
line presented in this paper are applied to each confined model in the following.
Fig. 7 The comparison among different hysteretic rules
Effect of skeleton curves

Test results of specimens at different displacement ductility u is listed in Tab. 2, which is corresponding to
computing results of loading at u=2, 4 and 6. Fig.8, 10, 12, 14, 15 and 16 show simulative results of
different models on four specimens. For the sake of figure clarity and because of no failure of specimen 1
and 2 in simulation , only one hysteresis loop is shown in Fig.8 and 10 at each ductility ratio respectively.
Fig.9, 11, 13 and 17 are experimental curves of four specimens.
Tab. 2 The force of each specimen calculated
by each model according to u=2,4 and 6 ('*'= near failure)
Mande
Displacemen Zhang Sheikh Gill
Test result r
Specimen t Model Model Model
(KN) Model
(mm) (KN) (KN) (KN)
(KN)
1-1 10.5 1267 1134 1061 1060 1086
1-2 22.5 1311 1185 1038 1050 1061
1-3 33.72 1317 1204 980* 1022 943*
2-1 8.477 1402 1434 1425 1346 1461
2-2 17.02 1410 1360 1334 1348 1380
2-3 25.65 1417 1398 1335 1374 1388
3-1 6.98 1205 1182 1071 1092 1169
3-2 14.0 1175 1218 998* 1046 1067
3-3 20.93 1166 1208 938* 1002 559*
4-1 5 1225 1219 1053 1058 1248
4-2 0 1387 1301 999 991 1187
4-3 15.5 1416 1341 1135* 965 failure
It can be observed that:
(1) The difference of simulative results at member level for each model are less than that at material level,
and the max is 1341/965=1.39.
(2) When ductility is small (u=2), the difference of computing results of capacity of each model is small;
with increasing of ductility, the difference becomes large gradually.
(3) When the ratio of axial compress force to strength is small (Specimen 1 and 2), the variance of
calculative results among these models is small; with increasing of the ratio, the variance grows.
(4) Simulative results of the Mander model and the Sheikh model are comparatively satisfactory, and those
of the Zhang model and the Gill model have been destroyed at high ductility. The Mander model is
the best in simulative results.
Fig. 8 The simulated comparison of Specimen 1 Fig. 9 Test result of Specimen 1

between the Mander model and the Skeikh model
Fig. 14 The simulated comparison of Fig. 15 The simulated comparison of
Specimen 3 between Specimen 3 between
the Zhang model and the Gill model the Mander model and the Skeikh model
Fig. 16 The simulated comparison of Fig. 17 Test result of Specimen 4

Specimen 4 between
the Zhang model and the Gill model
CONCLUSION
1. A significant difference exists in stress-strain relationships of different confined concrete models. In

member analysis, although the difference has been greatly reduced, but still is remarkable. Thus, the
confined concrete model should be carefully selected in structural analysis.
2. Skeleton curves and hysteretic rules play an important role in numerical simulation at structural
member.
3. The primary factor affecting simulative accuracy of confined concrete at member level is ductility of
descending branch in the material model, which is just the maximal dispute among the current models.
The main factors affecting the mechanical behavior of confined concrete should be further investigated.
4. The results in this paper show that the Mander model obtained the better agreement with the test results.
REFERENCES
1. Mander J. B, Priestley M. J. N and Park. R. "Theoretical stress-strain model for confined concrete". J
Strut Div ASCE, 1988; 114(8): 1804-1826.
2. Park R. Priestley. M. J. and Gill W. D. "Ductility of square-confined concrete columns". J Struct Div
ASCE,1982; 108(4): 929-950.
3. Park.R. Kent D.C. and Sampson R.A. "Reinforced concrete members with cyclic loading". J Stuct Div
ASCE,1972; 98(7): 1341-1359.
4. Sheikh S.A. "A comparative study of confinement models". ACI Struct J,1982; 79(4):296-306.
5. Sheikh S. A and Uzumeri S. M. "Analytical model for concrete confinement in tied columns". J Struct
Div ASCE,1982;2703-2722.
6. Thompson K. J and Park R. "Moment-curvature behavior of cyclically loading structural concrete
members". Proceedings of the Institution of Civil Engineers (London). Part 1 - Design & Construction.
1980;69(2): p317-341.
7. Yankelevsky D. Z, Reinhardt H. W. "Model for cyclic compressive behavior of concrete". J Struct
Engng ASCE,1987;113(2);228-240.
8. Zhang Xiuqin, Guo Zhenhai, Wang Chuanzhi. "Stress-stain curvilinear equation of confined concrete
under cyclic loading". J Building Structure,1982;9:16-20.
9. Spacone E., Filippou F.C., and Taucer F.F. "Fiber beam-column model for non-linear
analysis of R/C frames: Part I. Formulation". Earthquake Engineering and Structural
Dynamics, 1996; 25: 711-725.
10. Filippou F.C., Popov E.P., and Bertero V.V. "Modeling of reinforced concrete joints under
cyclic excitations". J Struct Engng ASCE;109(11):2666-2684.
NOTATION
The following symbols are used in this paper

fc = compressive strength of confined concrete;
fyh= tie yield strength;
ε U = the strain corresponding to the unloading point;
ε cc= the strain corresponding to the maximum concrete stress;
ε pi= the strain corresponding to the unloading point whose stress is zero;
ε re= the strain corresponding to the intersection between reloading curve and skeleton curve;
ε Q= the strain corresponding to Point 'Q';
Ec = the tangent modulus of elasticity of the concrete;
σ U= the stress corresponding to the unloading point;
σ Q= the stress corresponding to Point 'Q'.

Effects of Confined Concrete Models On Simulating RC

Uploaded by

Copyright:

Available Formats

Effects of Confined Concrete Models On Simulating RC

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Effects of Confined Concrete Models On Simulating RC

Uploaded by

Copyright:

Available Formats

13th World Conference on Earthquake Engineering

Vancouver, B.C., Canada

EFFECTS OF CONFINED CONCRETE MODELS ON SIMULATING RC

Zongming HUANG 1 Wenfeng ZHOU 2 Tao CHEN1 Juan CHEN1

FIBER-MODEL BASED BESM-COLUMN ELEMENT

Fig. 1 Details of test specimen Fig. 2 Comparison of stress-stain relationships

CALCULATION BASED ON EACH MODEL FOR SPECIMEN 1 AT MATERIAL LEVEL

COMPARISON OF CONFINED CONCRETE AT MEMBER LEVEL

Effect of hysteretic rules

Fig. 3 Rule 1 Fig. 4 Rule 2

Fig. 5 Rule 3 Fig. 6 Rule 4

Fig. 7 The comparison among different hysteretic rules

Effect of skeleton curves

Fig. 8 The simulated comparison of Specimen 1 Fig. 9 Test result of Specimen 1

Fig. 10 The simulated comparison of Specimen 2 Fig. 11 Test result of Specimen 2

Fig. 12 The simulated comparison of Specimen 3 Fig. 13 Test result of Specimen 3

Fig. 16 The simulated comparison of Fig. 17 Test result of Specimen 4

1. A significant difference exists in stress-strain relationships of different confined concrete models. In

The following symbols are used in this paper

You might also like