Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 95 (2014) 112 – 121

2nd International Conference on Sustainable Civil Engineering Structures and Construction

Materials 2014 (SCESCM 2014)

Behaviour of precast concrete beam-to-column connection with U-

and L-bent bar anchorages placed outside the column panel –
Analytical study
Dicky Imam Wahjudi*, Priyo Suprobo, Hidajat Sugihardjo, Tavio
Sepuluh Nopember Institute of Technology, Keputih-Sukolilo, Surabaya 60111, Indonesia

Abstract

In an effort to obtain a more comprehensive description on the behaviour of a beam-to-column precast reinforced concrete
connection, a model to simulate the hysteretic behaviour of the cyclic response of the connection will be presented in this paper.
A mathematical model employed, which uses finite-element approach and takes into account the pinching effect that often occurs
in a concrete structure, are employed. Numerical implementation and adaptation of the model obtained in a spring element will
be done with the computer code SeismoStruct. A series of numerical data and calibrations are input to produce an adequate
response history. While not exactly the same, the model gives result showing good conformity with the experimental data.
© 2014
Crown The Authors.
Copyright © 2014Published
Published by byElsevier
ElsevierLtd.Ltd.
This is an open access article under the CC BY-NC-ND license
Peer-review under responsibility of organizing committee of the 2nd International Conference on Sustainable Civil Engineering
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Peer-review
Structures under responsibility of
and Construction organizing
Materials committee of the 2nd International Conference on Sustainable Civil Engineering
2014.
Structures and Construction Materials 2014
Keywords: hysteretic behavior; cyclic response; finite-element; pinching effect; response history.

1. Introduction

In a building with a moment-resisting frame (MRF) system, the connection between beams and columns has
become a key strength. From the analysis by engineering mechanics, it is shown that good connections allow the
transfer and distribution of internal forces between building components to take place properly. Accordingly,
establishing criteria for beam-to-column connections (BCC) having good performance is an undeniable necessity.

* Corresponding author. Telephone: +62-31-5925382; E-mail address: dicky_iw@ce.its.ac.id; dickyimamwt73@gmail.com

1877-7058 Crown Copyright © 2014 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Peer-review under responsibility of organizing committee of the 2nd International Conference on Sustainable Civil Engineering
Structures and Construction Materials 2014
doi:10.1016/j.proeng.2014.12.170
 Dicky Imam Wahjudi et al. / Procedia Engineering 95 (2014) 112 – 121 113

For precast reinforced concrete MRF, some BCCs are prepared by wet manner which result in a monolithic
connections. Most BCCs are designed and prepared by emulating it to an equivalent monolithic BCC. The use of
such method is favoured by many engineers especially when coping with MRF structures situated in high-seismicity
regions. From a monolithic MRF structure, a BCC can be analyzed by modelling it separately as a free sub-
assembly. An illustration is shown in Fig. 1.

Exterior Connection

Interior Connection

Fig. 1. BCC sub-assemblies in an MRF.

G
P P
Plastifications Plastic hinge as
are formed at Link-beam with
the beam
- ends zero length

lp

(a) (b)

Fig. 2. A BCC subjected to lateral load P.

(a) Plastification occurred at both beam ends
(b) Plastification zones idealized as plastic hinges

(a) (b)
Fig. 3. Load-displacement curve of a BCC.
(a) Fat loop
(b) Pinched loop

In the context of designing an earthquake-resistant MRF, the study on the characteristics of BCC generally
emphasizes the investigation on the load-displacement relationship due to reversal loadings. As is shown in Fig. 2, a
specimen of BCC sub-assembly is applied with sets of cyclic lateral loading P, and the resulting displacement G will
be observed. High-intensity loading will cause forming of plastification in the structure, as indicated by appearance
of cracks in concrete and occurrence of irrecoverable elongation of reinforcing steel. Pairs of data set results in a
hysteretic response graph, as is shown in Fig. 3. Fig. 3-(a) shows a common load-displacement curve of a reinforced
114 Dicky Imam Wahjudi et al. / Procedia Engineering 95 (2014) 112 – 121

concrete BCC. The curve paths of back and forth motions to form successive open loops are generated by
mechanical properties of constituent materials, which reach post-yield conditions both in tension and compression
alternately. Generally, the curve is ‘fat’ in reflecting the stress-strain relationship of materials. However, when the
loads shifted from ‘flexure-dominant’ to shear, and cracks developed in concrete resulting in the slip of steel, the
curve is rather ‘pinched’ in the middle part, as is shown in Fig. 3-(b).
In this paper, it will be presented a numerical method for reproducing the hysteretic response curve with pinching
of a reinforced concrete interior BCC subjected to cyclic loading. Then, numerical implementation and adaptation of
the model within the computer code is determined. Finally, the model is calibrated using a series of experimental
test results and typical parameters are summarized.

2. The numerical method

The overall behaviour of a reinforced concrete MRF depends greatly on the behavior of its BCCs. Previous
studies have revealed that reinforced concrete BCC has many advantageous mechanical properties, such as strength,
ductility and energy dissipation capacity. The behavior of usual types of reinforced concrete BCC under cyclic
loading represents all the effects of a seismic event, which is characterized by hysteretic loops with progressive
stiffness degradation and strength deterioration that will soon lead to the failure of the structure.
Predicting the behaviour of reinforced concrete BCC is a quite complex concern, because it combines several
symptoms such as material non-linearity (plasticity, strain-hardening, concrete cracking), non-linearity of contact
and slip, geometrical non-linearity (local instability), residual stress conditions, and complicated geometrical
configurations. Under cyclic loading, this behavior is further complicated by successive static loading and
unloading. For static monotonic situations, it is nowadays possible to accurately predict the moment-rotation
response of a fairly wide-range of configurations by applying a common sectional analysis procedure [3, 7].
However, this is not the case for the cyclic situation. The same sectional analysis procedure has been applied to
simulate the cyclic hysteretic response behavior of reinforced concrete exterior BCC with rectangular or T–shaped
beam section. The model succeeded in featuring several aspects of the response, except the pinching effect [2].
At the present time, a computer program is available capable of calculating the response of the structure
subjected to cyclic loading. One program that will be mentioned in this paper, SeismoStruct, can be obtained by
downloading the installer from its website. Then, a registration will be needed after the installation is completed.
The program, which uses the finite element approach, can calculate the response of the structure by adopting several
models [14]. The author has made use of the program to calculate the response of several BCCs, for which
experimental results have been reported by previous researchers, i.e. Atalay et al, 1975 [9], and Ma et al, 1976 [15].
By comparing analytical results with experimental ones, it has been confirmed the accuracy of the program. The
model used was multilinear [4] and Richard-Abbott [5].

3. The Richard-Abbott model

The Richard-Abbott model was formerly developed in 1975, and was initially used to simulate the static
monotonic response of connections [12]. With successive improvements made through later studies by various
researchers, it finally became a convenient model to be applied to cyclic situations. A more extensive description of
this model has already been presented [13, 11], and to refresh memory, it will be outlined again below.
In this model, any generic point of a moment–curvature chart (M, I) of the loading branch, with or without
pinching, belongs to the Eq. (1). To describe the pinching effect of the orbit, two limit curves are introduced, that is
named as upper and lower bounds.

 Kot  K ht  u I
M 1
 K ht u I ..................................................................................................... (1)
ª  K  K  uI nt
º nt

«1  ot ht
»
«¬ M ot »¼

where the driving parameters are as defined in Fig. 4 and are as follows:
 Dicky Imam Wahjudi et al. / Procedia Engineering 95 (2014) 112 – 121 115

Kot = Kop + (Ko – Kop) ൈ t .................................................................................................................... (1-a)

Mot = Mop + (Mo – Mop) ൈ t .................................................................................................................. (1-b)
Kht = Khp + (Kh – Khp) ൈ t .................................................................................................................... (1-c)
nt = np + (n – np) ൈ t ............................................................................................................................ (1-d)

M Dh arctan K h
M
Upper-bound
Curve
Lower-bound
Curve

M
 K0t  K ht  u I  K ht u I
1
­°  K  K  u I nt
½° nt
M0 ®1 
0t ht
¾ M
 K0t  K ht  u I  K ht u I
M 0t
¯° ¿°
1

°­  K 0t  K ht  u I ½°
nt nt

D0 arctan K 0 ®1  ¾
M
¯° 0t °¿
I I
(a) (b)

Fig. 4. Definition of the loading branch of the Richard-Abbot model.

(a). Without pinching (b). With pinching

The parameter t, ranging in the interval [0, 1], defines the transition law from the lower to upper bound curve. It
must reproduce, as closely as possible, the shape of experimental curves, and is given by this expression:
t1
ª I Ilim t1 º
t « » ............................................................................................................................. ....... (2)
¬« I Ilim   1 ¼»
t1

where t1, t2 and Ilim are three experimentally calibrated parameters. The parameter Ilim is related to the maximum
experienced deformation in the direction of the loading branch to be described, and is evaluated to the following
relationship:

Ilim C  Io  Imax  ................................................................................................................................ (3)

where |Io| is the absolute value of the deformation corresponding to the starting point of the current excursion, Imax
is the maximum absolute value of the deformation experienced in the previous loading history in the direction of
loading branch to be described, and C is a calibration parameter. The unloading branch is assumed to be linear with
a slope equal to the initial stiffness Ko up to the interception with the straight line obtained by drawing a parallel to
the hardening line going through the origin. This allows the Bauschinger effect to be considered.

M M
C< 1 Kh
My+
K0
C= 1 My+ K0
C> 1
I My-
Iy
I
_I_ Imax

(a) (b)

Fig. 5. Definition of the Richard-Abbot model parameters.

(a). C parameter (b). Unloading branch

The cyclic action in the inelastic range produces an accumulation of plastic deformation, until ductility of the
system is locally exhausted and failure occurs due to fracture. In most cases, the repetition of loading is
accompanied by ‘response weakening’ as a consequence of strength deterioration and stiffness degradation. This can
be taken into account, both for strength and stiffness, using these expressions:
116 Dicky Imam Wahjudi et al. / Procedia Engineering 95 (2014) 112 – 121

§ Eh ·
M o,Re d M o u ¨1  iM u ¸ ........................................................................................................... (4-a)
¨ M y u I u ,o ¸
© ¹
§ Eh ·
K o,Re d K o u ¨1  iK u ¸ .............................................................................................................. (4-b)
¨ K o u I u ,o ¸¹
©
where I u ,o is the corresponding ultimate value in the case of one single excursion from the origin (monotonic
loading), Eh is the hysteretic energy accumulated in all previous experienced excursions, My represents conventional
yield resistance, Ko is the initial stiffness, and coefficients iM and iK are empirical parameters related to the damage
rate in strength and stiffness.
Hardening due to cyclic plastic deformation is considered to be isotropic. Furthermore, experimental results of
constant deformation amplitude tests did not show the occurrence of strength deterioration. This implies that cyclic
hardening increases in a few early cycles and then becomes stable.

M o,inc M o . . . . . . . . . . . . . . . if Imax ൑ Iy ....................................................................................... (5-a)

§ Imax  I y · if I > I ......................................................................................... (5-b)
M o,inc M o u ¨1  H h u ¸ max y
¨ I y ¸¹
©
where Mo and Mo,inc are the initial and increased value of strength, Imax is maximum deformation reached in the
loading history, Iy is the yield deformation, Hh is an empirical coefficient defining the level of isotropic hardening.
For more detailed information and further discussion on the subject, readers are advised to examine the indicated
references.

4. Computational implementation

The aforementioned computational procedure is now available as a computer code, namely SeismoStruct [14]. To
employ the program, first the structure modelling is performed. Data input includes the geometry of structures,
sectional dimensions, material properties and loading schemes. When plastification is intended to occur in the
structure, information about plastic hinges need also be clearly defined. The numerical implementation in this study
will be conducted by utilizing this software.
There are five BCC specimens will be presented to analyze their response behavior numerically, and then to
compare the results with experimental observations. The specimens consist of one monolithic BCC, two precast
BCC with U-bent bar anchorages placed outside the column panel, and two precast BCC with L-bent bar
anchorages. The experimental results have been reported in other paper from the same conference event (this
proceeding), and the resulting hysteretic curves are presented in Figs. 6-(a) to 6-(e) of the report [6]. Illustrations of
the five specimens are delivered in Figs. 6-(a) to 6-(e) of this paper.
Method of finite element was employed in this computational job, and hence structural modelling should be set
first. Using SeismoStruct, the structure is modelled into segments of linear element, and the model for BCC
specimen is shown in Fig. 7-(a). Meshing rules for beam and column sections are presented in Figs. 7-(b) and 7-(c).
To describe the cyclic behavior of the constituent materials, mathematical models are fixed for concrete [8] and
reinforcing steel [10].
As the specimens were loaded until experiencing post-yield conditions, plastic hinges are certainly formed in the
connections. Therefore, data for plastic hinges, which in such cases occurred at beam fixed-ends, must be input.
Each hinge, which is recognized as Link-beam, needs 30 parameters to be defined, i.e. 15 for the ascending branch
(subscript a) and 15 for the descending branch (subscript d). They are: Ka and Kd – initial stiffness for the upper-
bound curves, Ma and Md – strength for the upper-bound curves, Kpa and Kpd – post-yield stiffness for the upper-
bound curves, Na and Nd – shape parameter for the upper-bound curves, Kap and Kdp – initial stiffness for the lower-
bound curves, Map and Mdp – strength for the lower-bound curves, Kpap and Kpdp – post-yield stiffness for the lower-
bound curves, Nap and Ndp – shape parameter for the lower-bound curves, t1a and t1d – empirical parameter related to
pinching, t2a and t2d – empirical parameter related to pinching, Ca and Cd – empirical parameter related to pinching,
 Dicky Imam Wahjudi et al. / Procedia Engineering 95 (2014) 112 – 121 117

iKa and iKd – empirical coefficient related to stiffness degradation, iMa and iMd – empirical coefficient related to
strength deterioration, Ha and Hd – empirical coefficient defining the level of isotropic hardening, and – Emax-a and
Emax-d – maximum value of deformation (i.e. rotation) reached in the loading history. Summary of values of Link-
beam parameters used in the analysis is presented in Table 1.

1600 mm 450 1600 mm 1600 mm 450 1600 mm

Stirrup : I10 - 100 mm

Stirrup : I10 - 100 mm

II 1285 mm III 1285 mm

I II I

450 mm 450 mm

400 400 400 400

400 400

Stirrup : I10 - 100 mm

Stirrup : I10 - 100 mm

Stirrup : I10 - 100 mm Stirrup : I10 - 100 mm Stirrup : I10 - 100 mm Stirrup : I10 - 100
1285 mm
1285 mm

(a) (b)

1600 mm 450 1600 mm 1600 mm 450 1600 mm

Stirrup : I10 - 100 mm

Stirrup : I10 - 100 mm

III I
1285 mm III 1285 mm
3 D16

II I II 3 D16

450 mm 450 mm

400 400 400 400 400 400

Stirrup : I10 - 100 mm

Stirrup : I10 - 100 mm

Stirrup : I8 - 80 mm Stirrup : I10 - 100 mm
Stirrup : I10 - 100 Stirrup : I10- 100
1285 mm 1285 mm

(c) (d)

1600 mm 450 1600 mm

I
Stirrup : I10 - 100 mm

III
1285 mm
3 D16
3 D16

II 3 D16

450 mm

400 400
Stirrup : I10 - 100 mm

Stirrup : I10 - 100 mm Stirrup : I10- 100 mm

1285 mm

(e)

Fig. 6. BCC specimens investigated in this research

(a) Specimen 1 – monolithic
(b) Specimen 2 – with U-bent bar anchorage
(c) Specimen 3 – with L-bent bar anchorage
(d) Specimen 4 – with U-bent bar anchorage
(e) Specimen 5 – with L-bent bar anchorage
118 Dicky Imam Wahjudi et al. / Procedia Engineering 95 (2014) 112 – 121

(b) (c)

(a)

Fig. 7. Structural modelling applied to the BCC specimen.

(a). The specimen modelled as linear elements (b). Meshing for beam section (c). Meshing for column section

Table 1. Link-beam parameter used in the analysis.

Parameter to describe BCC Specimen 1 BCC Specimen 2 BCC Specimen 3 BCC Specimen 4 BCC Specimen 5

Ka (kN.m/rad) 53755.00 71094.00 77517.00 71094.00 77517.00

Ma (kN.m) 173.40 299.90 203.70 299.90 206.60
Kpa (kN.m/rad) 524.70 260.44 307.58 260.44 307.58
Na 0.60 0.60 0.60 0.60 0.60
Kap (kN.m/rad) 53755.00 71094.00 77517.00 71094.00 77517.00
Map (kN.m) 78.00 134.96 70.25 134.96 80.64
Kpap (kN.m/rad) 524.70 260.44 307.58 260.44 307.58
Nap 0.60 0.60 0.60 0.60 0.60
t1a 38.00 42.00 40.00 38.50 39.50
t2a 0.50 0.50 0.51 0.50 0.50
Ca 1. 1. 1. 1. 1.
iKa 0.000006 0.000004 0.000005 0.000006 0.000006
iMa 0.028 0.03 0.028 0.03 0.031
Ha 1.0E-07 1.0E-07 1.0E-07 1.0E-07 1.0E-07
Emax-a (rad) 0.1 0.1 0.1 0.1 0.1
Kd (kN.m/rad) 53755.00 64286.00 46105.00 64286.00 46105.00
Md (kN.m) 173.40 152.10 162.40 152.10 164.70
Kpd (kN.m/rad) 524.70 305.87 199.38 305.87 199.38
Nd 0.60 0.60 0.60 0.60 0.60
Kdp (kN.m/rad) 53755.00 64286.00 46105.00 64286.00 46105.00
Mdp (kN.m) 78.00 68.45 52.47 68.45 64.20
Kpdp (kN.m/rad) 524.70 305.87 199.38 305.87 199.38
Ndp 0.60 0.60 0.60 0.60 0.60
t1d 38.00 42.00 40.00 38.50 39.50
t2d 0.50 0.50 0.51 0.50 0.50
Cd 1. 1. 1. 1. 1.
iKd 0.000006 0.000004 0.000005 0.000006 0.000006
iMd 0.028 0.03 0.028 0.03 0.031
Hd 1.0E-07 1.0E-07 1.0E-07 1.0E-07 1.0E-07
Emax-d (rad) 0.1 0.1 0.1 0.1 0.1
 Dicky Imam Wahjudi et al. / Procedia Engineering 95 (2014) 112 – 121 119

5. Discussion of the result

nalyses bring about load–displacement relationship curves. These analytical data, superimposed with
experimental ones for each specimen, are presented in Figs. 8-(a) to 8-(e). This will provide ease and clarity to the
reader in comparing both data.

(a) (b)

(c) (d)

(e)

Fig. 8. Analytical–Experimental Comparison of the five BCC specimens

1. BCC Specimen 1 – monolithic
2. BCC Specimen 2 – with U-bent bar anchorage
3. BCC Specimen 3 – with L-bent bar anchorage
4. BCC Specimen 4 – with U-bent bar anchorage
5. BCC Specimen 5 – with L-bent bar anchorage
120 Dicky Imam Wahjudi et al. / Procedia Engineering 95 (2014) 112 – 121

For all specimens, the numerical curves show good agreement with the experimental results. They also display
conformity in strength, especially in large displacements. The phenomena of stiffness degradation and strength
deterioration are clearly visible in both results, although the analytical results give somewhat larger values than the
experimental ones.
When compared with the experimental study, the analytical results indicate a fatter curve. This is much more
noticeable in BCC Specimens 2 and 4 (those with U-bent bar anchorages). A somewhat different situation occurred
in BCC Specimens 3 and 5 (L-bent bar anchorages), where the hysteretic curves generated analytically are almost as
pinched as those generated experimentally. This outcome is likely due to the application of higher content of the
reinforcing steel in the cast-in-place concrete area in BCC Specimens 2 and 4 has led to a calculated stiffness that is
greater than the actual value. The result will be seen in higher values of load, particularly in the reloading region
(zones around zero displacement).

6. Conclusions

A numerical calculation procedure has been carried out to predict the hysteretic response behaviour of precast
reinforced concrete BCC specimens subjected to cyclic loading. The calculations were performed using the
computer code SeismoStruct which comprises of a finite-element based structural analysis. To take into account the
pinching effect of the resulting curve, the Richard-Abbot model was chosen in order to give a better description on
the characteristics of moment-rotation relationship in the plastic hinge regions. The performed calculations have
resulted in curves of response history showing similar characteristics with the experimental results. From the
comparison, it was also observed the quantity of stiffness with its degradation and strength with its deterioration
were almost similar. Particular circumstances have occurred in the BCC specimens with U-bent bar anchorages, in
which analytical results displayed a slightly fatter shape of curve. Nevertheless, in general, analytical results have
shown good conformity with the experimental ones.

References

[1] ACI Committee 318, Building Code Requirements for Structural Concrete (ACI 318-08), Farmington Hills, Michigan, 2008.
[2] D.I. Wahjudi, Mathematical Model on Moment-Rotation Behavior of Reinforced Concrete Connections due to Cyclic Loading, Journal
IPTEK, ISSN. 0853-4098, Vol. 3, No. 2, Nov. 1992.
[3] D.I. Wahjudi, Study on Moment-Curvature Behavior on the Fulfilment to the Capacity and Ductility Requirements of Reinforced Concrete
Moment Resisting Frame, Research Report, Lemlit ITS, Surabaya, 1994, (in Indonesian).
[4] D.I. Wahjudi, P. Suprobo, H. Sugihardjo and Tavio, Multilinear Model to Simulate Hysteretic Response Behavior of Reinforced Concrete
Beam-to-Column Connection due to Cyclic Loading, Proc. 3rd National Seminar on Aplikasi Teknologi Prasarana Wilayah (ATPW),
Surabaya, 2012, (in Indonesian).
[5] D.I. Wahjudi, P. Suprobo, H. Sugihardjo and Tavio, Numerical Implementation in Modelling Hysteretic Cyclic Response Behavior with
Pinching of A Reinforced Concrete Beam-to-Column Connection, Proc. 8th Asia–Pacific Structural Engineering & Construction
Conference and 1st International Conference on Civil Engineering Research, Surabaya, 2012.
[6] D.I. Wahjudi, P. Suprobo, H. Sugihardjo and Tavio, Behavior of Precast Concrete Beam-to-Column Connection with U- and L-Bent Bar
Anchorages Placed outside The Column Panel – Experimental Study, Proc. 2nd International Conference on Sustainable Civil Engineering
Structures and Construction Materials, Yogyakarta, 2014.
[7] Imbsen & Associates, Inc., XTRACT – Cross Section Analysis Program for Structural Engineers, Downloadable software with registration
from URL: http://www.imbsen.com, 2002.
[8] J.B. Mander, M.J.N. Priestley and R. Park, Theoretical Stress-Strain Model for Confined Concrete, Journal of Structural Engineering,
ASCE., Vol. 114, No. 8, 1988.
[9] M.B. Atalay and J. Penzien, The Seismic Behavior of Critical Regions of Reinforced Concrete Components as Influenced by Moment,
Shear and Axial Force, Report No.: EERC 75-19, UCB, California, 1975.
[10] M. Menegotto and P.E. Pinto, Method of Analysis for Critically Loaded RC Plane Frames, IABSE Preliminary Report for Symposium on
Resistance and Ultimate Deformability of Well-Defined Repeated Loads, Lisbon, 1973.
[11] P. Nogueiro, L. Simoes da Silva, R. Bento and R. Simoes, Numerical Implementation and Calibration of A Hysteretic Model with
Pinching for The Cyclic Response of Steel Joints, International Journal of Advanced Steel Construction, 2007.
[12] R.M. Richard and B.J. Abbott, Versatile Elasto-Plastic Stress-Strain Formula, Journal of the Engineering Mechanics Division, ASCE.,
1975.
 Dicky Imam Wahjudi et al. / Procedia Engineering 95 (2014) 112 – 121 121

[13] R. Simoes, L. Simoes da Silva and P. Cruz, Cyclic Behavior of End-plate Beam-to-Column Composite Joints, International Journal of
Steel and Composite Structures, 2001.
[14] SeismoSoft, Inc., SeismoStruct – Computer Program for Static and Dynamic Nonlinear Analysis of Framed Structures, Downloadable
software with registration from URL: http://www.seismosoft.com, 2012.
[15] S.Y.M. Ma, V.V. Bertero and E.P. Popov, Experimental and Analytical Studies on The Hysteretic Behavior of Reinforced Concrete
Rectangular and T-Beams, Report No.: EERC 76-2, UCB, California, 1976.