Apendice C PDF
Apendice C PDF
Apendice C PDF
Abstract. Steel frames with post-tensioned bolted connections are a viable option in high
seismicity areas due to the fact that brittle failure is prevented and also because of their selfcentering connection capacity. The elements formed by steel angles have a dominant influence
on the performance of the post-tensioned semi-rigid connections; therefore, it is necessary to
know their capabilities in terms of stiffness, strength, ductility and energy dissipation. In this
paper a set of 15 experiments with angles of 152x152x10 mm and different gage values were
developed. For this aim, monotonic and cyclic tests with increasing ductility demands were
performed. The results obtained are used to propose equations to calculate the initial and postyielding stiffness. An alternative definition of ductility is proposed, and fatigue curves and
hysteretic energy (EH) capacity are estimated. Additionally, it is proposed a function that
reproduces with good accuracy the hysteretic cycles obtained experimentally. The results were
extended for the application to post-tensioned top and seat connections, and a continuous
function which defines the Moment-Rotation hysteretic curve is suggested. The proposed
equations fit well also with the experimental results found in the literature. It is concluded that
1) the capacity EH decreases as the ductility demand increases; 2) the total hysteretic energy
capacity in the steel angles is constant for different gages, and 3) the number of cycles of load to
failure decreases as the ductility demand increases.
Keyword: steel frames, self-centering, hysteretic cycle, hysteretic energy, fatigue curve,
ductility
1. Introduction
Due to the need of preventing brittle failure in welded connections of moment resisting frames
(MRF), after the 1994 Northridge Earthquake, various researchers suggested replacing welded
connections for semi-rigid connections (SRC) because they are more effective to prevent this
type of damage. In fact, analyses of steel frames with semi-rigid connections (SRF) under
earthquake ground motions have shown that the maximum values of the base shear and drifts
are reduced compared with their counterparts with rigid connections; moreover, they dissipate
more energy (Nader and Astaneh 1991, Leon and Shin 1995, Reyes-Salazar 2000). In order to
increase the rigidity of the connections and to reduce the residual drift, Ricles et al. (2001)
proposed post-tensioning the SRF with high strength strands symmetrically positioned parallel
to the axis of the beams and anchored at the ends of the frames. It was observed that the
performance of this new type of structure is superior compared with the traditional MRF (Ricles
et al. 2002, Christopoulos and Filiatrault 2002b, Garlock et al.2009, Chou and Chen 2010a,
2011, Lpez-Barraza, et al. 2013).
The post-tensioned steel frames (PTSF) are a viable alternative to replace frames with welded
connections in high seismic areas. There are two types of common devices used to dissipate
energy in PTSF which depends on the relative rotation of the beam-column joint (r). The first
*Corresponding author. Ph.D. Student, E-mail: alopezb@uas.edu.mx
a
Professor, E-mail: sruizg@iingen.unam.mx
b
Professor, E-mail: reyes@uas.edu.mx
c
Professor, E-mail: eden@uas.edu.mx
consists of elements that dissipate energy by plastic deformation (hysteretic dissipators), which
can be angles, plates o bars bolted to the connection (Ricles et al. 2001, Christopoulos et
al.2002a, Garlock et al.2005, 2007, Chou and Chen 2010b). The other devices are usually plates
placed on the flange or the web of the beam and connected to the column, which dissipate
energy by friction (Rojas et al. 2005, Wolski et al. 2009, Guo et al.2011). There are numerous
advantages of PTSF, such as: preventing brittle fracture of the connection, field welding is not
required, the connection is made with conventional materials and skills, and the damage is
concentrated in the dissipating elements, maintaining essentially elastic the beams and columns
after the action of intense earthquakes. The connection has an initial stiffness similar to that of a
typical welded connection, and it is self-centering without residual deformation. In some studies
of this type of structural system it has been found that drifts and residual drifts are small,
avoiding higher costs of repairing or permanent loss of functionality of the building after a
strong earthquake (McCormick et al. 2008, Lpez-Barraza et al. 2013). If required, it is
economical and easy to replace the damaged angles, shortening the time of interference in the
use of the building.
The rigorous analysis of the PTSF is more complex than that of frames with rigid or pinned
connections. It is difficult to capture the behavior of the bolted connections due to their large
number of components that have a large influence on the performance. These components are:
angles or bolted plates; the presence or absence of reinforcing plates; the size, strength, torque
and position (gage) of the screws. Additionally, it is necessary to determine the influence and
contribution of post-tensioned elements.
PTSF connections are basically SRC with post-tensioned (PT) elements which can be
idealized as two springs in parallel, one nonlinear and one linear, representing the SRC and the
PT elements, respectively.
The presence of the SRC has been studied mainly in two ways; the first considers the
connection as a single piece, and describes its behavior using the moment- rotation curve (M-r)
(Richard and Abbott 1975, Yang and Jeon 2009). The parameters of the corresponding
equations are usually obtained from experimental results. In the second option, the parts of the
connection are modeled with finite elements using fiber elements, assigning to each fiber a
force-displacement relationship (Shen and Astaneh 2000, Ricles et al. 2001). The first option
has the advantage that when implemented it in a frame analysis program the number of elements
required is small. However, the disadvantage is that no information is provided about the
performance of the parts of the connection.
With the purpose of determining the performance and evaluating the influence of the elements
that compose the semi-rigid post-tensioned connection, experiments have been developed
finding that the behavior is nonlinear beginning at the start of the deformation (Ricles et al.
2001, 2002, Garlock et al. 2005, Chou and Chen 2010b). The typical M-r curve is as a shaped
flag, which characterizes the self-centering capacity as well as the energy dissipation. In the
studies mentioned above the proposed hysteretic models have been fitted with straight segments
(bi-linear or multi-linear). It is noticed that in the case of connections with bolted angles, it is
very important to know the performance of the angles due to its influence on the behavior of the
connection. With the aim of determining their initial and post-yielding stiffness, strength,
ductility, low cycle fatigue and hysteretic energy dissipation capacity, experiments have been
conducted with isolated angles of 8, 10, 13, 16 and 19 mm thick (Shen and Astaneh 1999,
Garlock et al. 2003) by changing the gage, size and strength of screws and the type of load. It
has been found that the ultimate strength is 3 times the yield strength, that the hysteresis loops
are stable, that degradation of strength and stiffness do not occur, and that the ductility capacity
ranges between 8 and 10.
The present study aims to improve the understanding of the performance of connections with
bolted top and seat angles, post-tensioned with high strength strands. In order to observe the
performance of the angles used in the connections, the authors performed a series of 15
experiments with angles of 152x152x10 mm for different gage values, subjected to cyclic
loading, increasing the ductility demands. From the results, equations to calculate initial and
post-yield stiffness are proposed. A new definition of ductility is suggested, fatigue and
hysteretic energy capacity curves are built as function of ductility demand. Additionally, it is
proposed a continuous function that accurately reproduces the hysteretic cycles obtained
experimentally. The results obtained for isolated angles, were extended for the application to the
post-tensioned top and seat connection and a continuous function which defines the M-r curve
of the hysteretic cycles is proposed. The proposed equation satisfactorily fits the experimental
results found in the literature.
2. Experimental program
2.1 Test specimen and instrumentation
The general configuration of the test specimens is shown in Fig. 1(a); to achieve symmetry each
specimen consists of two angles placed back to back. An A490 steel screw with diameter of
25.4 mm is used to attach the angles to the loading machine. The mechanical properties,
obtained from the average of two coupons, are: yield stress (Fy) = 382 MPa, ultimate strength
(Fu) = 552 MPa, modulus of elasticity (E) = 213 GPa and yield strain (y) = 0.002. A load of
magnitude 2V is applied with an actuator as shown in Fig. 1 a, by symmetry each angle support
a load of magnitude V. The displacement of the heel angle with respect to the column flange ()
is measured using a LVDT of 25 mm (see Fig. 1(c)). To measure displacements at various
points of interest, a camera KRYPTON K600 was used, Fig. 1(a) shows the location of ten LED
sensors. Based on their relative displacements the following can be measured:
L1 and L3, L3 and L5 measure the axial deformation of the angle wings.
L1 and L2, L5 and L6, measure the relative angle sliding with respect to the beam and
column flanges respectively.
L9 and L10 measure the displacement of the actuator piston (check the displacements of
the internal meter of the MTS machine).
2.2 Ductility definition
Ductility () is usually defined as the ratio of max.to y. There are several alternatives to
define the yield displacement at the angles. y may be regarded as the displacement when the
first fiber yield, or that produced when the angle mechanism occurs. Such deformation states
correspond to the beginning and the end of the transition zone respectively in the V- graph,
bounded by Points 1 and 2 shown in Fig. 2. Shen and Astaneh (1999) assume that y
corresponds to the plastic mechanism formation from which ductility values from 8 to 10 are
reported. Since for the first load levels, the stresses in the angles are due to bending and shear;
yielding of the material gradually occurs with different intensity in each section, therefore, it is
not observed a sudden loss of rigidity, but it occurs in a wide range of displacements already
defined as the transition zone. Considering that the capacity to dissipate EH of the angles is an
important parameter, a value of y obtained from the intersection of two straight lines as shown
in Fig. 2 is suggested in this study, in such a way that the enclosed areas by these lines and the
V- monotonic are the same. Table 2 shows the values of y obtained according to this
definition.
2.3 Test sequence
Monotonic and cyclic tests for a total of 15 specimens were developed, with gage (g1) of 80,
90, 100 and 108 mm. For each g1, a test was carried out monotonically from which the yield
displacement (y), the yield strength (Vy), the maximum displacement (max) and the maximum
force (Vmax) were determined, as defined below.
After obtaining y, cyclic tests were performed for each gage with maximum ductility
demands of 3, 6, 12 and 18. Table 1 shows the pattern of cyclic loading controlled by
displacements, which is similar to that proposed by SAC (Structural Engineers Association of
California, Applied Technology Council, California Universities for Research in Earthquake
Engineering) (SAC/BD 1997) The amplitudes in each cycle are modified in proportion to y,
until the max corresponding to the ductility demand imposed on each test is reached. For the
ductility of 3, 6, 12 and 18 the loading and reloading process is maintained until failure occurs.
The loading speed in all cases was 1.5 mm/s.
Table 2 shows the results of some specimens tested. The parameter used in the specimen
names are self-explaining, for example L152-10-g90-D6 means that a sided angle of 152 mm,
10 mm thick, 90 mm gage and a ductility demand of 6, were considered. The letter M in the
Table 2 corresponds to a monotonic load test. Fig. 1(b) shows the gage (g1) and the effective
gage (g2), the first is defined as the distance from angle heel to the screw center, and the second
as the distance from the fillet side angle to the screw head, which is precisely the point where
plastic hinges are formed. g2 largely defines the stiffness and the ductility and depends on the
thickness of the angle, the size of the screws, and g1.
3. Experimental results
3.1 Failure patterns
Fatigue failure occurred in the angles in all cases. Because of the large size and strength of the
bolts, they remained perfectly elastic. A mechanism with three plastic hinges was formed; the
first, parallel to the face of the column screws, and the others, in the angle fillets (see Fig. 1(c)).
Fractures occurred in the fillet angle on the side of the column, except for the Specimens 5 and
10; in these cases the screws are loosened, losing clamping with the column, allowing greater
freedom of rotation and the fracture at the fillet on the side of the beam. In the monotonic tests,
corresponding to the first flow, is almost unnoticeable. When the mechanism is formed,
values were observed to be 2.3, 3.3, 4.5 and 5.5 mm in the specimens, for gages of 80, 90, 100
and 108 mm, respectively. From the displacements recorded by the LEDs, it is observed that the
bolts remain elastic, axial strains in the wings of the angles are negligible compared to . There
was no sliding of the angles relative to the beam. In the Specimens 5 to 10 small displacements
relative to the column flange were observed since, as already mentioned, due to the number of
cycles the bolts lost some tightening. In spite of this, no significant influence was observed in
the hysteresis loops.
3.2 Stiffness
The initial stiffness (Kei) corresponds to the elastic behavior, it was determined by averaging
the rigidity obtained in the first cycles of loading in the range of very small deformations before
the transition. All of these cycles start at 0.05 mm and finalize between 0.2 and 0.6 mm,
according to the gage increment. The post-yielding stiffness (Kp) is calculated after the
transition zone, this remains practically constant up to the maximum displacement of the cycle.
Table 2 shows the values of Kei and Kp. The reason of Kp to Kei (rp) varies from 0.09 to 0.15. Fig.
3 shows the results of Specimens 9 and 12, which correspond to a monotonic and a cyclic test
with the same gage, respectively, and the same ductility demand. It can be observed that:
The monotonic test curve is the envelope of the hysteretic cycles.
The origin of the reloading curve changes when increasing the cycle amplitude,
however, it always cross the end point of the previous cycle. It indicates a hysteretic cycle
shift which is produced by the Bauschinger effect.
There is no appreciable loss of initial stiffness.
There is an increment of Kp for the cycles with large displacements due to strain and
geometry hardening (also found by Shen and Astaneh 1999, and Garlock et al. 2003).
CN 3149 1.63
(1)
(2)
The hysteretic energy dissipated per unit length of the angles (EHb) can be calculated as
follows:
E Hb
EH
b
(3)
where b is the length of the angle. By taking b=152 mm and applying Eq. (3) to the results of
this study, Eq. (4) is obtained, which is valid to calculate the hysteretic energy capacity per unit
length, for angles of 10 mm thickness and any gage amplitude.
(4)
displacement in the elastic range. Fig. 6 shows the model used; without considering axial
deformation it has two degrees of freedom in the node B: a linear displacement in y direction ()
and the angular displacement (). The node A is fixed and represents the action of the column
screw, the node C is restricted in x direction and to rotation; but it is free in the y direction.
The rigidity EI is constant for the portions AB and BC of the idealized angle. The lengths gc
and gv are parallel to the column and beam axis, respectively. The force V is applied to the end
C of the member BC. The details of the analysis procedure are shown in (Lpez-Barraza 2013),
where is found that the displacements and can be calculated as
gc
1
g g v
V
3EI 4 g c
1
g v
3
c
(5)
gc
1
V
2 EI 4 g c
1
g
v
(6)
where, E is the elasticity modulus and I is the moment of inertia of the longitudinal section of
the angle. The moments at the ends of the bars are considered positive anticlockwise and are
given by Eqs. (7), (8) and (9). It must be noted that the moment MAB has the largest magnitude
followed by MBA, so that first plastic hinge will be formed at point A, and then, two plastic
hinges will be simultaneously developed at point B, one at each side of the angle heel, as
observed in the experiments.
M AB
2g
Vg v 1 c
gv
gv
4
gc
M BC M BA
M CB
(7)
2Vg v
gv
4
gc
(8)
Vg v
gv
4
gc
(9)
Eq. (5) can be expressed as V=Ki which relates the force and displacement of the angle,
where the initial stiffness K i is given by:
4gc
1
g v
3EI
Ki 3
gc gc
1
gv
(10)
To calculate Ki from Eq. (10), gc and gv lengths are measured from the face of the screw to the
outside of the angle, it leads to the best approximation of the experimentally measured stiffness
(Kexp). Kexp is calculated as the average of Kei values for each group with the same gage by using
the experimental results shown in Table 2. In general Eq. (10) overestimates the initial stiffness
by about 8.5%. In Fig. 7 the ratio (CK) of Kexp and Ki is plotted against gc/t. By a regression
analysis the Eq. (11) is obtained, from which one can calculate the correction factor CK.
Multiplying Ki per CK the corrected initial stiffness (Ki*) is obtained, which is closer to the
experimental values, and the average error is reduced to 2%. Table 3 shows the values of CK
calculated with Eq. (11) together with Ki and Ki*, the last column shows the error of Ki* with
respect to Kexp. Eq. (10) and Eq. (11) constitute a simple option to calculate the initial stiffness
of such angles. There is a linear variation of the initial stiffness with respect to gc/t, which is
shown in Fig. 8 where the values of Kexp vs. gc/t are plotted.
CK 0.063(
gc 2
g
) 0.882( c ) 2.036
t
t
(11)
V Va
k (1 rp )( a )
k (1 r )( )
p
a
1
Vy
1
N
rp k ( a )
(12)
Where, V = force, =displacement, rp = ratio of the initial stiffness and rigidity of postyielding, Vy = yield force, N = parameter that defines the curvature of the transition zone, =
path of the unloading and reloading cycles, Va and a are the force and displacement where the
cycle starts loading or unloading, and k represents the stiffness at initial loading, which can be
kei or ki.
For additional explanation some hysteretic cycles with different ductility demands, were
separated from Fig. 3, which are shown with continuous line in Fig. 9; the plot of Eq. (12) is
shown with dashed line. The values of the parameters used in the equation are: k = 35.9 kN/mm
(calculated using Eq. (10) and (11)), rp=0.09, Vy = kDy=61.0 kN, N=1.5 and =2.0. Fig. 9 shows
good approximation between the hysteretic cycles obtained experimentally and those obtained
from Eq. (12).
As earlier mentioned, for identical specimens, monotonic curves constitute the envelope of the
hysteretic cycles. The monotonic curve can be adjusted with Eq. (12) by slightly changing the
values of the parameters: for the first loading (virgin curve) =1.0, a=0 and Va=0. For
unloading, =2.5, a and Va are the maximum values achieved in the loading cycle. k, rp, N and
Vy are unchanged. Fig. 10 shows with continuous line the monotonic curve obtained
experimentally for the Specimen 9 (part of Fig. 3); the dashed line is the plot of Eq. (12).
(13)
where Ts is the resultant of tendon force and F is the force induced by the interaction of the
floor system with the beam. By making moment summation about the center of rotation,
without considering F, it is obtained:
M= Tsd2 +Vad1+MaT+MaC
(14)
Eq. (14) shows that angles and tendons work as springs in parallel, the Tsd2 term represents
the contribution of the post-tensioned tendon to the moment of the connection, the remaining
terms are the contributions of angles. In summary, it can be stated that the flexural strength of
the connection is the superposition of the contribution of the tendons and that of the angles.
6.2 Contribution of post-tensioned tendons
To obtain the M-r curve of the post-tensioned tendons, it is necessary to transform the axial
stiffness (ks) of the tendons into a rotational equivalent stiffness (ks), which can be done as
follows:
The tension in the tendons (Ts) is the sum of the initial tension imposed (T0) and the additional
tension caused by the connection gap (T). Thus,
Ts T0 T
By expressing T in terms of the connection gap (tension elongation)
(15)
Ts T0 k s s
(16)
where s is the elongation of the tendons due to the gap. From Fig. 11(c), d2 is the level arm of
the resultant axial force in the beam with respect to center of rotation. Then, the following
relationship can be established:
s 2d 2 r
(17)
The factor 2 in Eq. (17) is to consider the two connections which influence the deformation of
the tendons in the interior of a bay. Substituting Eq. (17) into Eq. (16) it gives:
Ts T0 2k s d 2 r
(18)
Eq. (18) is the relationship between the strength of the tendons and the rotation of the
connection. Multiplying both sides of Eq. (18) per d2, it is obtained:
M s T0 d 2 2k s d 22 r
(19)
by considering
and
k s 2k s d 22
(20)
M d T0 d 2
(21)
M s M d k s r
(22)
ks in Eq. (22) is the contribution of the tendons to the rotational stiffness of the connection
and Md is known as the decompression moment. The later is the result of the initial tension so it
is a constant, moreover, it represents the value of the moment just when the connection starts
opening.
6.3 Contribution of the angles
The contribution of the angles to the connection resistant moment can be calculated as
follows:
M V Va d1
(23)
In the elastic range, Va=Ki , where Ki is the initial stiffness and of the gap in the
connection. It can established that =d1 r, then Va=Ki d1 r. Substituting Eq. (23) it is obtained:
M V K i d12 r
(24)
From Eq. (24) it is observed that the contribution of the angles to the rotational stiffness of the
connection is:
kV K i d12
(25)
M Ma
kV (1 rp )( a r )
k (1 r )( )
p
a
r
1 V
My
1
N
(rp kV k s )( a r )
(26)
Ma and a are the moment and rotation at the beginning of each cycle of loading or unloading,
respectively, for the initial loading cycle Ma=Md and a =0.0, for the unloading phase Ma and a
takes the maximum preceding values, My=Vy d1. rp, N and have the same meaning as defined
before.
To illustrate the use of Eq. (26), a connection was designed, similar angles to that of Specimen
12, were considered. The beam section is a W18x46 with plates 25 mm thick placed in the
flanges as reinforcement. Four tendons of 150 mm2, resistance of 279 kN and 8 m in length are
used. The values of T0, d2, d1, Md, ks, kv, rp and My, are 433.8, kN, 242 mm, 519 mm, 105 kNm, 1757.8 kN-m/rad, 10235 kN-m/rad, 0.06, and 23.2 kN-m, respectively. Substituting these
values in Eq. (25) with Ma=Md=105 kN.m, a=0 and =1 for the loading cycle, and Ma=222
kN.m, a=0.033 rad and =2 for the unloading cycle the graph shown in Fig. 12 is obtained,
which resembles the typical form of flag found in previous experimental studies (Ricles et al.
2002, Garlock et al. 2005).
6.5 Comparison with experimental results
For the purpose of validating the accuracy of Eq. (26) it is plotted and compared with
experimental results of post-tensioned beam-column assemblies (semi-rigid post-tensioned
connection) developed by Garlock et al. (2005). The specimens considered are: Spec. 16s-45,
Spec. 20s-18 and Spec. 36s-30, tested under cyclic loading for different levels of drift. The
specimens correspond to an interior connection.
The sections of the beam and the column are W36x150 and W14x398, respectively. Angles
203x203x19 mm of 406 mm long, and plates of 25x356 mm with variable length are placed as
reinforcement in the flanges, both of steel Fy=345 MPa. The tendons are of 140 mm2, steel A416 and their number is variable. In the specimens notation "16s-18", 16s refers to sixteen
strands and 18 to the initial tension of each strand in kips.
Based on the geometry and material information reported, the required parameters in Eq. (26)
are calculated. Fig. 13 reported by Garlock et al. (2005) shows the experimental results, the
graphics obtained with Eq. (26) for specimens Spec. 16s-18, Spec. 20s-18 and Spec. 30s-36 are
also presented in the Figure. This indicates a very good approximation between the equation
plot and the experimental results that represent different combinations of relative contribution of
angles and tendons, to the strength and stiffness of post-tensioned semi-rigid connections. The
implementation of this model in a computer program for the dynamic analysis of post-tensioned
frames with semi-rigid connections is simple and efficient since it is sufficient to consider the
connection as an additional element of the structure while retaining the parameters that
influence the strength, stiffness and energy dissipation capacity. An alternative very useful and
powerful tool for calculating the parameters associated to the contribution of the angles in the
proposed hysteretic model is the use PRCONN computer program developed by Richard and his
collaborators (Richard ,1993).
7. Conclusions
To improve the knowledge of the performance of steel angles used in bolted connections
under cyclic loading, a set of experiments with isolates steel angles with different gages (g2) was
developed in order to know the rigidity, capacity to dissipate hysteretic energy (EH), and number
of load cycles to failure for different ductility demands (). Mathematical models (Eq. 12) were
built to represent the connections and the force-displacement hysteretic cycles (V-). The results
of the isolates angles were extended to post-tensioned semi-rigid connection and a continuous
function (Eq. 26) that accurately models the moment-rotation (M-r) hysteretic cycles was
proposed. Some important observations are:
For a given , EH does not vary with g2.
For a given g2, EH decreases as increases.
In all cases, when decreases EH increases.
There is a linear variation between EH and .
There is a large variation while obtaining Ki in the experimental results, but its influence
is small when compared with the large inelastic deformations, therefore, it does not
significantly affect the final EH capacity.
The post-yielding stiffness in the angles varies from 0.09 to 0.15 of Ki.
Hysteretic cycles are stable and no significant degradation in strength and stiffness
occurs.
Eq. (12) models with good accuracy the hysteretic cycles (V-) of angles.
Eq. (26) reproduces with good accuracy the hysteretic cycles (M-r) of post-tensioned
semi-rigid connections, so it is a good option to implement it as a hysteretic model in
computer programs of nonlinear dynamic analysis.
Acknowledgments
The support given by PROMEP of the Secretara de Educacin Pblica de Mxico, the
Universidad Nacional Autnoma de Mxico under Project PAPIIT-IN107011-3 and the
Universidad Autnoma de Sinaloa is greatly appreciated.
References
Chou C-C, Chen JH. (2010). Column restraint in post-tensioned self-centering moment frames.
Earthquake Engineering and Structural Dynamics, 39( 7), 751774.
Chou C-C, Chen JH. (2010). Tests and analyses of a full-scale post-tensioned RCS frame subassembly.
Journal Constructional Steel Research, 66(11), 13541365.
Chou C-C, Chen JH. (2011). Analytical model validation and influence of column bases for seismic
responses of steel post-tensioned self-centering MRF system. Engineering Structures, 33, 26282643.
Chou C-C, Chen J-H. (2011). Seismic design and shake table test of a steel post-tensioned self-centering
moment frame with a slab accommodating frame expansion. Earthquake Engineering and Structural
Dynamics, 40, 1241-1261.
Christopoulos, C., Filiatrault, A. and Uang, C.M. (2002a). Self-centering post-tensioned energy dissipating
(PTED) steel frames for seismic regions, University of California, Report No. SSRP-2002/06.
Christopoulos, C. and Filiatrault, A. (2002b). Seismic response of post-tensioned energy dissipating
moment resisting steel frames, Proceedings of the 12th European Conference on Earthquake
Engineering, London, UK, paper No. 61.
Garlock, M., Ricles, J., and Sause, R. (2003).Cyclic load tests and analysis of bolted top-and-seat angle
connections. J. Struct. Eng., 129(12), 16151625.
Garlock, M., Ricles, J., and Sause, R. (2005). Experimental studies on full-scale post tensioned steel
connections. ASCE Journal of Structural Engineering, 131(3), 438-448.
Garlock, M., Sause, R. and Ricles, J. (2007). Behavior and design of post-tensioned steel frames system.
ASCE Journal of Structural Engineering, 133 (3), 389.
Guo, T., Song, L. and Zhang, G. (2011). Numerical simulation of the seismic behavior of self-centering
steel beam-column connections with bottom flange friction devices, Eathquake Engineering and
Engineering Vibration, 10, 229-238. DOI: 10.107/s11803-011-0061-5.
Leon, R.T. and Shin, K.J. (1995). Performance of Semi-rigid Frames, Proceedings of Structure Congress,
p. 1020-1035, April.
Lpez-Barraza, A., Bojrquez, E., Ruiz, S.E. and Reyes-Salazar,A. (2013). Reduction of maximum and
residual drifts on post-tensioned steel frames with semi-rigid connections, Advances in Materials
Science and Engineering, 2013, ID 192484, doi:10.1155/2013/192484.
Lpez-Barraza, A. (2013) Diseo ssmico de marcos de acero con conexiones semi-rgidas, basado en
energa. Ph.D. dissertation, Universidad Nacional Autnoma de Mxico, Mxico City (in progress).
McCormick, J., Aburano, H., Ikenaga, M. and Nakashima, M. (2008). Permissible residual deformation
levels for building structures considering both safety and human elements. 14th Conference on
Earthquake Engineering. Beijing, China, 12-17 October.
Nader, M.N. and Astaneh, A. (1991). Dynamic behavior of flexible, simirigid and rigid frames, Journal of
Construction Steel Research, 18, 179-192.
Reyes-Salazar, A. (2000). Dissipation of energy in steel frames with PR connections, Structural
Engineering and Mechanics, 9(3), 241-256.
Richard, R.M., and Abbott, B.J. (1975). Versatile elastic plastic stress-strain formula, ASCE Journal of
Engineering Mechanics, 101(4), 511-515.
Richard R.M., PRCONN (1993), Moment-rotation curves for partially restrained connections, RMR
Design Group, Tucson, Arizona.
Ricles, J.M., Sause, R., Garlock, M. and Zhao, C. (2001). Post-tensioned seismic-resistant connections for
steel frames, ASCE Journal of Structural Engineering, 127(2), 113-121.
Ricles, J.M., Sause, R., Peng, S.W. and Lu, L.W. (2002). Experimental evaluation of earthquake resistant
post-tensioned steel connections, ASCE Journal of Structural Engineering, 128(7), 850-859.
Rojas, P., Ricles, J.M. and Sause, R. (2005). Seismic performance of post-tensioned steel moment
resisting frames with friction devices. ASCEJournal of Structural Engineering, 131(4), 529-540.
SAC/BD-97/02 Protocol for fabrication, inspection, testing, and documentation of beam-column
connection tests and other experimental specimens, by P. Clark, K. Frank, H. Krawinkler, and R. Shaw.
Appendix E Loading protocol for stepwise increasing cyclic test
Shen, J., and Astaneh-Asl, A. (2000). Hysteretic model of bolted-angle connections. J. Constr. Steel Res.,
54, 317343.
Shen, J., and Astaneh-Asl, A. (1999). Hysteretic behavior of bolted angle connections. J. Constr. Steel
Res., 51, 201218.
Weaver W. and Gere J.M. (1980). Matrix analysis of framed structures (Second Edition). Van Nostran
Reinhold Company.
Wolski, M., Ricles, J.M. and Sauce, R. (2009). Experimental study of a self-centering beam-column
connection with bottom flange friction device. ASCEJournal of Structural Engineering, 135(5), 479488.
Yang J.G and Jeon S.S. (2009). Analytical model for the initial stiffness and plastic moment capacity of an
unstiffened top and seat connection under a shear load, International Journal Steel Structures, 9(3), 195205.
2V
(a)
(b)
L2
L1
60 mm
LED
L3
g2
g1
L6
L7
(c)
L8
LVDT
L9
Plastic hinge
L10
V [Load]
Vy
2
1
0
[Displacement]
100
80
V [kN]
L5
60 mm
L4
CYCLIC
MONOTONIC
60
40
20
0
-20 0
10
15
[mm]
20
-40
-60
-80
-100
1600
CN [C y c les]
1400
1200
1000
CN = 3149-1.63
800
600
400
200
0
0
10
15
20
[Ductility]
60
g2/t=3.9
g2/t=5.0
50
g2/t=6.0
g2/t=6.9
E H [kN.m]
40
Ec. (2)
30
20
10
0
10
15
20
[Ductility]
V
C
gv
gc
100
K exp , [kN/mm]
1.0
0.9
0.8
0.7
0.6
80
60
40
20
0
g c /t
10
5.0
7.0
g c /t
V [ kN]
50
0
0
10 12 14 16 18 20
-50
[ mm ]
Experimental
-100
Equation
100
50
0
0
-50
-100
9.0
100
V [ kN ]
CK = K exp / K i
1.1
8 10 12 14 16 18 20
[ mm ]
Experimental
Equation
11.0
Shim plate
r
T
Va Ma
PT strands
d1
PM
d2
(a)
(b)
(c)
C
Reinforcing
plate
Ma
Angle
(a)
(b)
(c)
M [ kN-m ]
250
200
150
100
50
0
0
0.01
0.02
0.03
0.04
r [rad]
Fig. 12 Typical curve M-r post-tensioned semi-rigid connection
Eq. (26)
Load
step
1
2
3
4
5
6
7
8
9
30
10
0.3y
0.6y
1.0y
1.5y
2.0y
3.0y
6.0y
9.0y
12.0
y
18.0
y
Specimen
.
Name
L152-10-g80-M
3.9
y
(mm)
0.60
m
(mm)
12.7
g2/t
Vy
(kN)
65.0
Vm
(kN)
110.4
Kei
(kN/mm)
45.79
Kp
(kN/mm)
6.67
EH
(kN.m)
*
Cycles
L152-10-g80-D12
3.9
7.2
12
95.2
68.15
118
33.05
L152-10-g80-D18
3.9
11.0
18
103.7
140.40
61
23.60
L152-10-g90-M
5.0
1.00
19.8
50.0
103.2
24.66
3.10
L152-10-g90-D3
5.0
3.0
57.0
57.56
1559
51.00
L152-10-g90-D6
5.0
6.0
73.0
77.44
213
42.50
L152-10-g90-D12
5.0
12.0
12
88.6
76.61
43
24.65
L152-10-g90-D18
5.0
18.0
18
99.3
60.25
25
17.35
L152-10-g100-M
6.0
1.70
18.0
47.5
87.7
26.40
2.32
10
L152-10-g100-D3
6.0
5.1
74.8
49.43
359
48.10
11
L152-10-g100-D6
6.0
10.2
67.9
28.91
116
42.25
12
L152-10-g100-D11
6.0
18.0
11
89.0
29.11
43
29.55
13
L152-10-g108-M
6.9
1.75
22.0
40.0
78.1
14.41
2.00
14
L152-10-g108-D6
6.9
10.5
64.8
29.67
90
33.90
15
L152-10-g108-D10
6.9
18.0
10
78.1
20.75
32
29.70
Specimens
1_3
4_8
9_12
13-15
gc /t
6.25
7.31
8.36
9.20
CK
1.016
1.045
0.935
0.746
Ki*
(kN/mm)
86.86
58.23
35.87
21.91
% Error
2.4
-1.8
6.7
1.4