Journal of Structural Engineering No.

34-3
Vol. 34, No.1, April-May 2007 pp. 25–32

A new concept for consideration of slab effects on building seismic

performance

C.Umarani* and Gregory A.MacRae**

Slabs exist in the majority of buildings worldwide but they are seldom modelled explicitly in analysis for design. The
slabs result in increased column demands, which may lead to failure. Since floor slabs affect the seismic response,
proper models must be developed if realistic modelling of the seismic behaviour of frames is to be undertaken. This
paper describes the development of a simple model for explicit evaluation of the slab effect on moment-resisting struc-
tural systems which considers the slab contribution to the beam overstrength at different storey drifts and the slab
opening displacements which can be related to damage. The model captures important aspects of the behaviour of a
reinforced concrete joint with a floor slab well.

One of the fundamental concepts underlying the capacity
design philosophy is that failure mechanisms in a structure
can be controlled by appropriately selecting the member
strengths. Our concrete codes recommend that in the case of
multi-storey structures, the strong column - weak beam de-
sign philosophy may be used to achieve the required level of
ductility during a major earthquake. According to this, mem-
ber strengths are selected so that strength of the structure is
limited by beam flexural strength. For capacity design to be
successful the designer should be able to assess the member
strengths under different combination of loads.
The inelastic deformation of beams in RC frames causes
column centerline to move apart. The columns may provide
restraint to elongation, and compressive force arises from
this restraint which enhances the strength of the member and
alters the behaviour of the overall structure. Elongation of
longitudinal beams containing cast-in-place slabs can cause
membrane forces to occur within slabs also enhancing the
strength of these beams. If the slab effect on the flexural
strength of beam is not considered, it is possible that the
beams will be stronger than columns and the strong column
- weak beam design will not be realized.
The contribution of a floor slab, acting as a compression
flange of a cast-in-place reinforced concrete beam, has been
recognized for a long time. Appropriate recommendations
with respect to a dependable effective width of such flanges
are incorporated in building codes throughout the world1-3 .
However similar recommendations for determining the beam
flexural strength and stiffness are not available to assess the
beam overstrength moment input into the column.
A full-scale test on a seven-storey reinforced concrete
structure tested at the Building Research Institute in Tsuku-
ba, Japan4 had base shear strength approximately 70% more
than anticipated. Much of this increased shear strength was
considered to be due to floor slab reinforcement acting as
additional tensile reinforcement to the beam. Cheung et al5
 Assistant Professor, Structural Engineering Division, Department of Civil Engineering, Anna University, Chennai - 600 025, India  Associate Professor,
Department of Civil Engineering, University of Canterbury, Christchurch, New Zealand
(Discussion on this article must reach the editors before July 31, 2007)
tested three full-scale interior and exterior beam-column joint
subassemblages with floor slabs. They reported that a sig-
nificant proportion of the slab bars in tension contributed to
the negative moment flexural strength of the beams. Tests on
subassemblies with slabs spanning between more than one
column at the University of Auckland6 , and at the Universi-
ty of Canterbury7 , both exhibited an increase in strength of
about 80%, which could be directly attributed to the effect of
the slab alone.
The contribution of slabs acting as tension flanges is gen-
erally less well understood in design practice and therefore,
in the seismic design of most building frames throughout the
world, where columns may have a tension flange on one side,
and a compression flange on the other, the effect of the floor
slab is not explicitly considered. Since the presence of the
slab can significantly affect the frame seismic behaviour, in-
creasing the potential for collapse, and since there is a pos-
sibility of slab damage,proper models of frames considering
floor slabs should be developed to consider these effects.
This paper describes the development of such a mod-
el based on the two-dimensional “Including Beam Growth”
(IBG) model by Kim8-11 . This model was used primarily for
precast beams with unbonded prestress cables, which are be-
coming popular in the United States, where little change in
behaviour may occur with repeated cycles to the same dis-
placement. The same model can be modified for steel beams,
and it can also represent reinforced concrete beams, but with
less accuracy. In beams, elongation occurs due to the follow-
ing three reasons12,13 :
(i) Geometric effects in which the flexural strains on the
tension side of the beam are greater than on the com-
pression side. This results in elongation at the beam
mid-depth, and it will be referred to as gap-opening
throughout the remainder of this paper,
(ii) Dislocation of aggregate particles restraining closure
of the cracks, and

JOURNAL OF STRUCTURAL ENGINEERING VOL. 34, NO.1, APRIL-MAY 2007 25

 (iii) Diagonal compression forces in the web, associated
with the action of shear reinforcement, have a hori-
zontal component which is balanced by additional ten-
sion forces in the flexural tension reinforcement. This
results in the flexural compression reinforcement not
yielding when the tension reinforcement yields during
inelastic rotation.
The latter two effects become significant during severe re-
versed cyclic loading and they are not considered in the mod-
el used in this study.
EFFECT OF BEAM GROWTH ON FRAME
BEHAVIOUR
A typical joint test configuration of a moment resisting (MR)
frame system is shown in Fig. 1.
FIG.1 TYPICAL JOINT TEST CONFIGURATION
(a) MOMENT RESISTING FRAME-ELEVATION
(b) TYPICAL JOINT TEST CONFIGURATION
For a steel beam column system, in which the beam
yields, and the beam section is symmetric, the neutral axis
is in the middle of the section, and the top of the beam sec-
tion yields in tension, while the bottom yields in compres-
sion. The distance between the points of inflection, L, does
not change as shown in Fig. 2a.
FIG.2 TYPICAL BEHAVIOUR OF SYSTEMS WITH DIFFERENT
MATERIALS
(a) STEEL
(b) REINFORCED CONCRETE/GAP SYSTEMS
For a reinforced concrete system, such as that shown in Fig.
2b, the neutral axis in the beam at the column face, is closer
to the compression side than to the tension side of the section.
Therefore, at the centre of the beam, a gap (or crack) opens
between the beam end and the column face. The gap opening
increases with storey drift and it causes the distance between
the beam points of inflection to increase. This phenomenon,
26 JOURNAL OF STRUCTURAL ENGINEERING VOL. 34, NO.1, APRIL-MAY 2007
which is commonly known as “beam growth”, was first de-
scribed by Fenwick and Irvine12 and it occurs not only in
reinforced concrete beams, but on all systems with gap open-
ing. These include precast concrete members which may or
may not be prestressed14 , steel members which are designed
to rock about one corner15 , and other similar members.
In the subassembly above, beam growth does not seem to
cause any adverse behaviour. However, in a frame, such as
that shown in Fig. 1a, the total elongation in beam length
over the width of the frame increases both with the num-
ber of storeys and the storey drift. A joint model developed
by Kim, Stanton and MacRae8-11 to consider gap opening in
reinforced concrete or precast-concrete systems is shown in
Fig. 3.
FIG.3 BEAM-COLUMN JOINT MODEL BY KIM8, 11
Kim et al.10,11 have shown that the demands on a frame
may be very different from that found from conventional
analysis as shown in Fig. 4.
FIG.4 FRAME SWAY INCLUDING BEAM GROWTH MODEL10, 11
Here it may be seen that the beam growth effect push-
es the external columns apart, thereby imposing greater de-
mands on the columns causing a greater tendency for column
inelastic action. Since, the columns are also pushing on the
beams, the moment capacity of reinforced concrete beams
is increased because of the extra compression, also increas-
ing the likelihood for column yielding. It should be noted that
while more column yielding may be expected, the probability
of a soft-storey mechanism is not necessarily increased. Kim
et al10,11 have also shown that the increase in demand due to
beam growth is not insignificant and that some of the larger
column moment demands doubled (even more) in the anal-
yses undertaken. In addition to the effects described above,
beam growth may also cause precast flooring units to become
unseated and to fall off16 . The effect of this beam growth is
not included in the majority of analyses carried out for struc-
tural design worldwide.
SLAB EFFECTS ON FRAME BEHAVIOUR
Most structures include slabs, and their presence can have
a major effect on the frame behaviour for both gap-opening
FIG.5 SLAB EFFECTS ON FRAME BEHAVIOUR
(a) SLAB IN FRAME STRUCTURE
(b) STEEL STRUCTURE WITH STRONG SLAB
(c) STEEL STRUCTURE WITH WEAK SLAB17
(d) STEEL STRUCTURE WITH SLAB-COLUMN GAP
(e) RC/PC WITH STRONG SLAB
(f) RC/PC WITH WEAK SLAB
JOURNAL OF STRUCTURAL ENGINEERING VOL. 34, NO.1, APRIL-MAY 2007
and non-gap-opening structures. The slab is shown on a beam
in a two bay frame Fig. 5a. For example, for steel beams re-
sisting lateral frame displacements, the slab will be in com-
pression beside one column and in tension beside the column
at the far end of the beam. Where the slab is in compres-
sion, the neutral axis will probably be within the slab, and the
whole steel section may yield in tension. The strain on the ex-
treme tension side of the section may be more than twice that
expected if no slab is considered for the same hinge rotation.
Also, the moment input into the column may be significantly
larger than if there is no slab. For loading in the reverse di-
rection, the behaviour depends on the tension capacity of the
slab. If it is strong in tension, then the steel beam may well
yield completely in compression (Fig. 5b).
Again, the strains will be large and there will be a greater
probability of local buckling than if the slab is not present.
If the slab is weak in tension, then the steel beam will yield
in compression at the bottom and tension at the top (Fig. 5c).
While there will be no beam elongation at this end, the shift
of the neutral axis at the other end of the beam will cause
beam growth, and induce extra compression into the beam
which may affect the beam performance. To minimize slab
interaction with the frame, the slab may be separated so that
it does not touch the column face during the inelastic actions
which affect the frame (Fig. 5d).
For beams with gap opening expected at their ends, such
as RC beams or post-tensioned ungrouted beams, the pres-
ence of the slab restrains the gap opening. In fact, if the
slab is strong enough, gap opening due to hogging moments
may not occur and the moment input into the column may be
much greater than that expected in traditional analysis (Fig.
5e). These moments may induce a soft-storey mechanism. If
the slab is not so strong then it will crack and cause struc-
tural damage (Fig. 5f). In this respect, many of the “dam-
age free” solutions which have recently been developed14,15 ,
while they have no damage when no slab is considered, are
likely to cause severe damage when a slab exists.
COMPUTATIONAL MODELLING
The main objective of the finite element computer modelling
approach is that the behaviour and interaction of the decisive
elements should be captured accurately on a local and global
level. This should include the local deformations in the con-
tact zone of the rocking elements, the global column demand
increase due to beam / slab overstrength and beam elonga-
tion effects at the joint opening, increase of the tensile force
in the post-tensioned tendons (if any) and subsequently the
increase of compression force and change of the moment in
the beams.
The IBG model (Including Beam Growth model) devel-
oped in this study shown in Fig. 6 is based on Kim’s IBG
model8-11 . In order to reproduce the behaviour of the hybrid
frame/ RC frame and investigate some of its features, a mod-
el was developed using a two dimensional analysis program,
RUAUMOKO-2D18 . A model of a large structure can be con-
structed by assembling these, or compatible exterior joints.
Since the joint model is to be assembled for a large structure,
the efficiency of the model is critical, because it would save
computational time and storage space. Several modelling as-
sumptions are made, while keeping the integrity of the fun-
damental joint behaviour.
In Fig. 6, the nodes associated with the column are num-
bered c1-c11, while the beam nodes are designated as b1-b8.
The beam on the left, for example, spans between two nodes,
27

FIG.6 IBG MODEL
b5 and b7, placed at the centroid of the beam section. Node
b5 near the joint is connected to b1 on the top and b3 on the
bottom of the beam. The vertical distance, c, between b1 and
b5, is the confined beam depth which is the beam depth ex-
cluding the clear cover to shear reinforcement. The nodes b1,
b5 and b3 and b2, b6 and b4 are connected by stiff members.
This enforces rigid body movement of the nodes relative to
each other. The column nodes c1 through c4 are located at
the column faces. The joint is assumed to be rigid and no
shear deformation is considered and this is done by slaving
all the degrees of freedom of nodes c1 to c8 with the center
node c7. The columns are lined up vertically to node c7 and
can be linked to the nodes of the joint above and below. The
column end nodes, c10 and c11 are restrained against later-
al displacement. The unbonded prestressing steel (if any) is
linked at the far end of each beam, between nodes b7 and b8.
The mild steel bars at the interfaces are located at the appro-
priate positions. The beam-column interface is represented
by the multispring element with ten springs. Extra nodes are
provided on either side of the interface which are not shown
in the figure. These nodes are slaved to move with the column
face of the beam end.
Various elements used in the models are as follows:
1. Giberson one component beam model is used to mod-
el the beams as well as column, where only elastic de-
formations are expected to occur. The beams and col-
umn are assumed to be linearly elastic and no strength
degradation is assumed. The columns have rigid end
blocks. The depth of the blocks is assumed to be 80
percent of the adjoining beam depth to take into ac-
count of softening of beam-column joint due to cracks.
The elastic modulus of the concrete is obtained from
its measured compressive strength, fck (in N/mm2 ), as
5000vfck (in N/mm2 ). Half the gross moment of iner-
tia for the beams and the column are used to take in-
to account flexural cracking. The stiffnesses of these
members are assumed constant throughout.
2. Truss elements that are prestressed before lateral load
is applied to the subassemblage are used to model the
unbonded post tensioned steel in prestressed members
(not used in RC members); The stress- strain curve for
post tensioned steel is assumed to be bilinear with a
change in slope occurring at the yield point. In all but
28 JOURNAL OF STRUCTURAL ENGINEERING VOL. 34, NO.1, APRIL-MAY 2007
exceptional cases post tensioned steel is expected not
to yield.
3. Multi-spring gap elements with ten springs are used to
model the gap opening / closing at the beam-column
interface for concrete beams. The springs carry on-
ly compression and the tensile strength of the con-
crete is neglected. Kim8-11 assumed that all inelastic
deformation occurred in the grout itself between the
beam and column when used with precast concrete el-
ements. Such an assumption is obviously very sensi-
tive to the thickness of grout used, and it makes no
sense when the grout thickness is zero (such as in the
case of RC members). In this study, a more reasonable
approach for estimating the confined concrete failure
length in unbonded post-tensioned beams developed
by El-Sheikh et al19 was used. Properties for the in-
terface concrete are given based on the compressive
strength equations of CEB-FIP model code20 for con-
fined concrete
4. Truss elements are used to model the steel in the beam
at the interface. The bars are designed to yield cyclical-
ly and in doing so, dissipate energy. The stress-strain
relationship is idealized as bilinear with a strain hard-
ening ratio of 0.02. The length of the bars, Ls (in-
cluding additional debonded length) over which the
bars elongate and contract, is much longer than the
interface grout thickness in the model, Lmodel. Ma-
terial properties are thus scaled to mimic the actu-
al strength and stiffness of the bars, viz., As,model =
As x(Lmodel =Ls ); fsy,model = fsy x(Ls =Lmodel ).
THE CONCEPT
The concept proposed to model the behaviour of floor
slabs is new, simple and appropriate for use in two-
dimensional analysis of reinforced concrete frames with
traditional hysteretic beam-column line elements. A typ-
ical interior beam-column-slab subassemblage is shown
in Fig. 7. The subassemblage is shown with boundary
conditions typical of those applied to beam-column sub-
assemblage test specimens. The pinned supports are locat-
ed at midheight of the columns and midspan of the beams,
which are the approximate locations of the inflection points
in a frame under lateral load. The portion of slab modelled
by strut element (hereafter called slab element) in the beam-
column-slab subassemblage is shown in Fig. 8.
The model which contains one extra “slab element” to
consider slab effects, is shown in Fig. 9. This model can be
used with either gap opening or non gap opening beam el-
ements. The strut models the portion of the slab outside the
column lines as shown in Fig. 8. While the strut considers
the effect of the slab, it does have some limitations. Bending
effects of the slab are ignored (although they can be consid-
ered if the strut is not pin-ended). Also, any interaction of the
slab with the column sides is ignored as the slab strut is not
connected to the column itself - it only connects either side
of the slab. While the strut itself is a simple element, it needs
to be calibrated in order to accurately represent the slab be-
haviour. Calibration of the slab element has been carried out
based on experimental results5 using the computer program,
Ruaumoko. All reinforcement in the slab, as well as a good
cracking model for the concrete are considered to make the
analysis as realistic as possible using an approach similar to
that by Fenwick et al21 .

FIG.7 TYPICAL BEAM-COLUMN-SLAB SUBASSEMBLAGE
FIG.8 BEAM-COLUMN-SLAB SUBASSEMBLAGE
FIG.9 IBGS MODEL
Figure 10 shows the EBG (Excluding Beam Growth)
model which is a conventional model used in many building
analyses11 . A single bilinear rotational spring between nodes
connected to the elastic beam and column elements simu-
lates inelastic rotational action at the beam-column interface.
The nodes connected are slaved together in the horizontal and
JOURNAL OF STRUCTURAL ENGINEERING VOL. 34, NO.1, APRIL-MAY 2007
vertical directions only. The elastic stiffness of the spring is
made very high to minimize elastic deformation there. Flex-
ural yielding occurs at the moment capacity from the IBG
model. A bilinear curve, or a flag-shaped hysteresis curve,
could be used for the rotational spring. The post-yield stiff-
nesses is found by trial and error to provide the same force-
displacement curve as that from the IBG subassemblage
model. Because this model does not consider beam growth,
it is useful for comparison with the IBG model in a multibay
frame.
FIG.10 EBG MODEL
CALIBRATION OF SLAB ELEMENT IN IBGS
MODEL
The slab element in the IBGS model is calibrated for a RC
subassemblage based on experimental results by Cheung et
al5 . Figure 11 shows the pattern of major cracks formed on
the top surface of slab. It may be seen that the crack pat-
tern is consistent with diagonal compression forces being
introduced over the length of the beams. The tensile yield-
ing of all longitudinal slab bars was measured when loads
were applied to the beams. Thus the whole width of the
slab of the test unit was effective as a tension flange. In this
calibration, all reinforcement in the slab was considered in
FIG.11 PATTERN OF MAJOR CRACKS ON TOP OF SLAB5
29
the development of the slab axial element using an approach
similar to that by Fenwick et al6 .
The tensile yielding of all longitudinal slab bars was mea-
sured when loads were applied to the beams. Thus the whole
width of the slab of the test unit was effective as a tension
flange. In this calibration, all reinforcement in the slab was
considered in the development of the slab axial element us-
ing an approach similar to that by Fenwick et al6 .
Calibration of the slab element was carried out in two
different ways. In both methods a single strut in each quar-
ter of the slab was used to represent the series of diagonal
forces apparent from the crack pattern. Method 1: Struts are
considered to move out from the beam at a distance of d, the
effective depth of the beam beam, from the face of the col-
umn at an angle of 45 (Fig. 12a). The length of slab element
FIG.12 COMPRESSION STRUT LOCATIONS
(a) METHOD 1, (b) METHOD 2
was assumed equal to the column width plus 2d. With this
assumption only three longitudinal bars of Cheung’s test unit
were effective in taking the tensile forces as shown. Method
2: The struts are positioned at the face of the column itself
and extend in the slab at an angle of 45 (Fig. 12b). The
length of slab axial element is the column width. In this case
30 JOURNAL OF STRUCTURAL ENGINEERING VOL. 34, NO.1, APRIL-MAY 2007
all the longitudinal bars were effective in taking the tensile
forces.
For calibration of the slab element finite element mod-
els as shown in Figs. 13a and 13b were used. These consider
the slab alone, decoupled from the frame, in order to obtain
appropriate characteristics for the slab axial element. (It is
also possible to consider the full slab element model in con-
junction with the frame but this was not done in this model).
Only the plane of the slab was considered, the longitudinal
and transverse reinforcement were connected to the points A
and B (Fig. 13) using rigid links. B was assumed to be pinned
and A could move in longitudinal direction (horizontally). A
displacement was imposed at A to the left, and the forces pro-
duced allowed a force-displacement curve, was plotted. This
force-displacement curve is identical to that required for the
slab axial element.
FIG.13 FINITE ELEMENT SLAB MODEL
(a) METHOD 1, (b) METHOD 2
COMPARISON WITH SUBASSEMBLY TESTS
Experimental results of unbonded prestressed concrete sub-
assemblage tests without a slab by Day22 and Day’s theo-
retical estimate of the behaviour are compared with those
obtained from IBG and EBG joint models (Method 2) de-
veloped in this study without strain hardening. Figure 14a
describes the force-displacement curves reported by Day. Future work involves evaluating the effect of beam
The experiment did not follow the theoretical estimate be- growth/slab effects on steel and RC/PC frame seismic re-
cause joint degradation occurred in this test unit which was sponse. A simple preprocessor is being developed so that
not considered in the model. The IBG and EBG results the slab effects can easily be incorporated into a structural
should therefore be compared with Day’s theoretical curve, analysis program, which can be used for research of different
rather than with the experimental data. It may be seen that building types. Joint deformation effects will also be consid-
the EBG and IBG models in Fig. 14b have the same backbone ered.
curve as would be expected because the EBG was calibrated
to the IBG model. The IBG/EBG unloading curves are differ-
ent because the flag-shaped hysteresis curve in RUAUMOKO
does not currently allow non-trapezoidal flag shapes for the
rotational spring in the EBG model. The backbone curve
for Day’s theoretical data compares well with the IBG/EBG
model.

FIG.14 MODELLING OF DAY’S EXPERIMENTS22

(a) DAY’S EXPERIMENTAL RESULTS AND PREDICTION
(b) IBG AND EBG MODEL RESULTS

The beam-column-slab joint (IBGS) model was com-

pared against Cheung’s experimental results in Figs. 15(a)
- 15(c). Here the column is held in place at the top while
the beams are alternately moved up and down. These ini-
tial analyses capture the trends of the experiments, especially FIG.15 LOAD-DISPLACEMENT RESPONSE OF EAST BEAM22
the strength on the loading part of the curve. The increase (a) EXPERIMENTAL RESULT22
in strength due to the presence of the slab was 21%. Fur- (b) ANALYTICAL RESULT (SLAB EFFECT CONSIDERED)
ther modification to the model to account properly for strain (c) ANALYTICAL RESULT (SLAB EFFECT NOT
hardening and for shear deformation within the hinging zone CONSIDERED)
is likely to improve the comparison further.

JOURNAL OF STRUCTURAL ENGINEERING VOL. 34, NO.1, APRIL-MAY 2007 31

CONCLUSIONS
Most models of structures do not consider beam-growth or
slab effects explicitly. The majority of “damage free” struc-
tures will have substantial damage to the building slab. Slab
effects can increase the moment demand on the column, and
the shear demand on the beam. A beam-column-slab joint
model is proposed to consider gap opening and for evalu-
ating the slab effect in two-dimensional nonlinear analysis
of reinforced concrete frames. The model proposed above is
one of the first proposed to explicitly consider slab effects in
building analysis. Comparison of simulated and observed re-
sponse indicates that the proposed model can represent key
aspects of the beam-column-slab joint behaviour well. Fu-
ture work involves evaluating the effect of beam growth/slab
effects on steel and reinforced concrete and / or prestressed
concrete frame seismic response.
ACKNOWLEDGEMENT
The authors are grateful for the NPEEE (National Program
on Earthquake Engineering Education) of the Indian govern-
ment which has funded the study leave of the second author
at the University of Canterbury. The authors express their
thanks to Dr. Richard Fenwick and Dr. Athol Carr for their
valuable guidance.
REFERENCES
1. ——— “Code of practice for the design of concrete
structures”, NZS 3101: Part 1: 1982. Standards Asso-
ciation of New Zealand, Wellington, 1982, pp. 127.
2. ——— “Building code Requirements for reinforced
Concrete ACI Committee 318”, ACI 318–89, American
Concrete Institute, Detroit, 1983, 353 pp.
3. ——— “Indian standard code of practice for plain and
reinforced concrete”, Code of Practice Indian Standard,
IS 456 (2000) Bureau of Indian Standard, New Delhi,
36–37 pp.
4. “Earthquake effects on reinforced concrete structures -
U.S.-Japan research”, ACI SP-84, ACI, Detroit, Mich.,
1984.
5. Cheung, P., Paulay, T. and Park, R.. “A Reinforced
Concrete Beam-Column Joint of a Prototype One-way
Frame with Floor Slab Designed for Earthquake Resis-
tance”,Research Report, 87–6, University of Canterbury,
New Zealand, 1987.
6. Fenwick, R. C. and Davidson, B. “Elongation in Ductile
Seismic-Resistant Reinforced Concrete Frames.” Proc of
the Tom Paulay Sym, Farmington Hills, MI, American
Concrete Institute, SP 157-7, 1995, pp. 143–170.
7. Fenwick, R., Davidson, B., and Lau, D. B. N. “Inter-
action between ductile RC perimeter frames and floor
slabs containing precast units”, Proc. of NZSEE Conf.,
Wairakei, 2005.
8. Kim, J., “Behaviour of Hybrid Frames under Seismic
Loading.” PhD Thesis, Dept., of Civil Eng., University
of Washington, Washington, 2002.
32 JOURNAL OF STRUCTURAL ENGINEERING VOL. 34, NO.1, APRIL-MAY 2007
9. Kim, J., Stanton, J. F., and MacRae, G. A. “Effects of
Beam Growth on the Behaviour of Reinforced Concrete
Frames.” 7th U.S. Nat’l Conf. Ear. Engg. July 2002,
Boston.
10. Kim J., Stanton J. F., MacRae G. A., “Approximate
Methods of Accounting for Beam Growth Effects”, 13th
World Conf. on Ear. Engg., Vancouver, B.C., Canada,
August 1–6, 2004, Paper No.504.
11. Kim J., Stanton J., and MacRae G. A., “Effect of Beam
Growth on Reinforced Concrete Frames”, J. of strut.
Eng. ASCE., Vol. 130 No. 9,September 2004, pp. 1333–
1342.
12. Fenwick, R. C. and Irvine H. M., “Reinforced Concrete
Beam-Column Joints for Seismic Loading- Part II Exper-
imental Results” Bull. of the New Zealand Nat. Soc. for
Ear. Engg., Vol. 10 No. 9, December, 1977, pp. 174–185.
13. Fenwick, R., and Megget L. M. “Elongation and Load
Deflection Characteristics of Reinforced Concrete Mem-
bers Containing Plastic Hinges”, Bull. of the NZ Nat.
Soc. for Ear. Engg., Vol. 26 No. 1, March 1993, pp. 28–
41.
14. Priestley M. J. N. and MacRae G. A., “Seismic Tests
of Precast Beam-to-Column Joint Subassemblages with
Unbonded Tendons”, PCI J., January-February 1996, pp.
64–80.
15. Filiatrault A., ”Gap-Opening Systems”, 13th World
Conf. on Ear. Engg., Vancouver, B.C., Canada, August
1–6, 2004.
16. Matthews, J., Bull, D. K. and Mander, J., Investigating
the Load Paths of Floor Diaphragm Forces during Se-
vere Damaging Earthquakes, Proc. of the Comb. Conc.
Soc. and Ready Mix, Rotorua, TR24, 2001, pp. 122–131.
17. Clifton G. C.. “Semi-Rigid Joints for Moment-Resisting
Steel-Framed Seismic-Resisting Systems”, HERA Re-
port R4-134, New Zealand Heavy Engineering Research
Association, Manukau City, 2004.
18. Carr, A.J. “RUAUMOKO users manual”, University of
Canterbury: 2004. www. ruaumoko. co. nz
19. El-Sheikh, M.T., Pessiki, S., Sause, R. and Lu, L.W.,
“Moment - Rotation Behaviour of Unbonded Post-
Tensioned Precast Concrete Beam-Column Connec-
tions,” ACI Struct. J. 2000, Vol. 97 No. 1: pp. 122–131.
20. “CEB-FIP Model Code 1990”, Comite Euro-
International du Beton (CEB) and the Federation Inter-
nationale de la Precontrainte (FIP)., Design Code,
Thomas Telford Ltd.
21. Fenwick, R., Bull, D.K., Macpherson, C. and Lind-
say, R., “The Influence of Diaphragms on Strength of
Beams”, Proc of NZSEE 2006 conf, New Zealand, 2006
22. Day, S., “Cyclic load testing of Precast Hybrid Frame
Moment-Resisting Connection”s, Master’s Thesis, Dept.
of Civil and Environ. Engg., Univ. of Washington, 1999.