Anda di halaman 1dari 10

ACI STRUCTURAL JOURNAL TECHNICAL PAPER

Title no. 108-S18

Impact of Seismic Input on Strain/Displacement Response


of Reinforced Concrete Members and Frames
by Aidcer L. Vidot-Vega and Mervyn Kowalsky

This is a study on the effect of displacement histories on the in the first phase consist of symmetrical and asymmetrical
relationship between material strains and deformation parameters reversed cyclic as well as seismic acceleration time histories.
such as displacement and drift for reinforced concrete (RC) moment The second phase consists of the analyses of RC moment
frame structures. The study is divided into two phases: 1) a phase frames under seismic acceleration time histories.
that considers RC column members subjected to symmetrical and
asymmetrical reversed cyclic displacement and seismic time
histories. Columns with 2 and 4% longitudinal steel ratios and 5 to RESEARCH SIGNIFICANCE
20% axial load ratios are analyzed; and 2) a phase that considers At the most fundamental level, current design practice is
moment frames subjected to seismic acceleration time histories. A based on monotonic moment-curvature analysis at the
total of five moment frame models are analyzed. The results are section level and monotonic pushover analysis at the
compared to monotonic analyses and simplified design expressions member or system level. Seismic structural testing over the
that were previously developed by the authors. The analysis results last several decades has shown that monotonic section analysis
support the use of monotonic section analyses to define the coupled with the plastic hinge method for member deformations
envelope of seismic response because the relationship between (or pushover analysis) can be used to accurately predict the
strains and displacements of seismically well-designed members force-deformation response of a structural system subjected
appears to be minimally affected by loading history.
to cyclic loading. What is still not understood, however, is
how the relationship between strain and displacement
Keywords: cyclic loads; load history; reinforced concrete frames; seismic
accelerations; strains. changes in RC structures under the influence of variable
displacement histories. As has been shown by Moyer and
INTRODUCTION Kowalsky (2003), four nominally identical columns subjected
Several researchers have investigated the effect of to four different displacement histories failed at four different
displacement histories on beam and column members. lateral displacement levels. In the four columns, however, the
Tomazevic et al. (1996) determined that the ultimate total tensile strain levels in the reinforcement were similar.
displacement is dependent on the displacement history This highlights the importance of strain as a critical
applied to the structure. In addition, many researchers have performance parameter to accurately define performance limit
investigated the concept of damage indexes to quantify states and the importance of understanding the relationship
damage based on the measure of absorbed energy by the between strain and displacement and how it is impacted by
structure (El-Bahy et al. 1999a,b; Tsuno and Park 2004; displacement history. The results of this research will clarify
Ingham et al. 2001, 2002). Whereas damage indexes are the impacts of different displacement and seismic acceleration
useful as an indication of absorbed energy, the drawback, histories in limits states based on material strains and will lend
from the perspective of design, is that they require knowledge confidence toward the use of monotonic section analyses for
of the load history to predict failure. Performance-based the seismic design of RC moment frames.
seismic design (PBSD) procedures, such as direct displace-
ment-based seismic design, control damage through the ANALYSIS OF SDOF COLUMN MEMBERS
control of displacements, which is based on structural and Column models
nonstructural damage criteria. In the case of nonstructural Three different columns were analyzed with two different
damage criteria, drift is a useful parameter to define target reinforced concrete square sections (Table 1). The two
displacements. In the case of structural damage criteria, sections of the three columns have 2% (s1) and 4% (s2)
however, material strain is the more useful parameter. In longitudinal steel ratio distributed around the section perimeter,
both cases, damage indexes are not generally used to define respectively. The dimensions of the sections are shown in
target displacements. Given the importance of material strain Table 1. The spacing and diameter of the transverse steel
limits in defining structural performance limit states in used for both sections is 3 and 3/8 in. (76 and 9.52 mm),
procedures such as displacement-based seismic design, it is respectively. The average transverse volumetric steel ratio
essential to have adequate relationships between material was kept constant (1.0%) for both sections. It has been
strains and displacement whose accuracy is investigated shown (Vidot-Vega 2008) that this variable does not affect
under the influence of various displacement histories. the relationship between strain and displacement. Note that
The goal of this paper is to summarize a numerical study this does not mean that a different transverse steel ratio will
on the impact of seismic input on the strain-versus-displacement
relationship in RC column members and moment frames. ACI Structural Journal, V. 108, No. 2, March-April 2011.
MS No. S-2008-309.R2 received January 12, 2010, and reviewed under Institute
This is accomplished in two phases. The first phase includes publication policies. Copyright 2011, American Concrete Institute. All rights reserved,
the analyses of column members subjected to different including the making of copies unless permission is obtained from the copyright proprietors.
Pertinent discussion including authors closure, if any, will be published in the January-
displacement histories. The displacement histories considered February 2012 ACI Structural Journal if the discussion is received by September 1, 2011.

178 ACI Structural Journal/March-April 2011


Aidcer L. Vidot-Vega is an Assistant Professor in the Department of Engineering
Science and Materials at the University of Puerto Rico at Mayagez, Mayagez,
Puerto Rico, where she received her BS and MS. She received her PhD from North
Carolina State University, Raleigh, NC, in 2008.

ACI member Mervyn Kowalsky is a Professor in the Department of Civil, Construction,


and Environmental Engineering at North Carolina State University. He is a member of
ACI Committees 213, Lightweight Aggregate and Concrete; 341, Earthquake-Resistant
Concrete Bridges; and 374, Performance-Based Seismic Design of Concrete Buildings;
and Joint ACI-ASCE-TMS Committee 530, Masonry Standards Joint Committee.

Table 1Properties of RC column models


(distributed reinforcement)
Model Column 1 Column 2 Column 3 Fig. 1Basic symmetrical reverse cyclic loading history.
Lcol, ft (mm) 8 (2438.4) 12 (3657.6) 15 (4572)
Lcol /H 3.75 4.81 7.03
18 x 26 20 x 30 18 x 26
Section, in. (mm)
(457 x 660) (508 x 762) (457 x 660)
long, dbl (s1),
0.02, 1.00 (25.4) 0.02, 1.13 (28.5) 0.02, 1.00 (25.4)
in. (mm)
long, dbl (s2),
0.04, 1.13 (28.5) 0.040,1.43 (35) 0.04, 1.13 (28.5)
in. (mm)

not affect the moment-curvature response. However, it does


imply that for the same level of strain, the drift will be
minimally affected by different transverse steel ratios.
Section 1 (s1) was analyzed with 5 and 10% axial load
ratio (ALR) (defined in Eq. (1)), while Section 2 (s2) was
analyzed with 5 and 20% axial load ratio. The yield strength Fig. 2Basic asymmetrical reverse cyclic loading history.
of the steel was set to 65 ksi (450 MPa) and the
compressive strength (28 days) was set to 4 ksi (28 MPa). validate the results from OpenSees using the above described
These values were chosen to be typical for normalweight hysteretic models. Both the global response (force-displace-
concrete and ASTM A706/A706M steel. ment) and the local response (base curvatures) obtained from
the finite element model were in close agreement with the
Column analysis models experimental results from Moyer and Kowalsky (2003).
The columns were modeled using a lumped plasticity
approach. Within OpenSees, this was accomplished by using ALR = P/fc Agross (1)
the beam with hinges element (Scott and Fenves 2006).
This element consists of plastic hinges that are modeled with
fiber elements, whereas regions outside the plastic hinge are kL + 0.022f y d bl Fu
modeled using traditional frame element properties. The L p = max ; k = 0.2 ----- 1 0.08 (2)
0.044f y d bl fy
confined and unconfined concrete in the fiber sections of the
element were modeled using the Kent and Park (1971)
concrete model with degraded linear unloading/reloading Displacement histories
stiffness according to the work of Karsan and Jirsa (1969). Four displacement histories were used in the column analyses:
This model (Concrete01 in OpenSees) assumes no tensile 1) monotonic, 2) symmetric reversed cyclic, 3) asymmetric
strength for concrete. The concrete 28-day compressive reversed cyclic, and 4) seismic. In the case of the monotonic
strength fpc, concrete strain at maximum strength epsc0, history, the columns were analyzed up to a drift ratio of 8%.
concrete crushing strain epsU, and concrete strength fpcu are The word monotonic is used to define a displacement
the required input for this concrete model. These parameters history that is one cycle to failure; it applies equally well to
were calculated as proposed by Mander at al. (1988). The section, member, or system analysis. For the symmetric
concrete strength remains constant beyond the crushing reversed cyclic history, the following drift ratio increments
point. The steel fibers were modeled using a reinforcing steel were used: 0.002, 0.005, 0.01, 0.02, 0.025, 0.03, 0.04, 0.05,
model developed by Moehle and Kunnath (2006). The 0.06, 0.07, and 0.08. The drift increments were applied at
plastic hinge lengths Lp were calculated using Eq. (2) one, three, five, and seven cycles, thus resulting in four
(Priestley 1998; Priestley et al. 2007). In Eq. (2), dbl is the distinct symmetric reversed cyclic histories. The one-cycle
longitudinal bar diameter, Fu is the ultimate steel strength, version of this displacement history is shown in Fig. 1.
and L is the length from the critical section to the point of For the asymmetric displacement history, the same drift
contraflexure in the member in SI units. Linear elastic ratio increments and the same number of cycles (one, three,
properties are used for the rest of the element. In addition to five, and seven) were employed with the exception that in the
the plastic hinge length, the required inputs for this element reversed direction, each cycle terminated at the yield drift ratio
are the modulus of elasticity E, moment of inertia I, and of the columns (0.0054, 0.006, and 0.010 for Columns 1, 2 and
cross-sectional area A for the elastic portion. Experimental 3, respectively). The one-cycle version of this displacement
results from testing of four identical RC circular columns history is shown in Fig. 2. The displacement control location for
performed by Moyer and Kowalsky (2003) were used to all the analyses was at the top of the column.

ACI Structural Journal/March-April 2011 179


Fig. 3Steel tension strain versus drift ratio for symmetrical cyclic loadings (Column 1).

Table 2Earthquakes used in analysis obtained at the cross section at the bottom of the column
(point of maximum moment) at the extreme compression
No. Earthquake Location Year Station PGA, g Mw Near-field
fiber, and at the location of extreme tension bar, respectively,
California, with the neutral axis also measured from the extreme fiber.
1 Loma Prieta 1989 Gilroy 0.41 6.93
USA
2 Tabas Iran 1978 Dayhook 0.33 7.35
Impact of controlled displacement history
3 Erzican Turkey 1992 95 Erzican 0.52 6.69 x
on strain-displacement relationship
4 Northridge California, 1994 Rinaldi 0.83 6.69 x This section presents a comparison of the results from the
USA
static analysis using the symmetrical and asymmetrical
5 Kobe Japan 1995 KJMA 0.82 6.90
displacement histories (referred to as controlled displace-
6 Imperial California, 1940 El Centro 0.31 6.95 ment histories) and the monotonic analyses. Figures 3 to 6
Valley USA
present the steel tension and concrete compression strains as
Imperial California,
7
Valley USA
1940 Meloland 0.31 6.53 x a function of the drift ratio for the symmetrical and asymmet-
8 Adak Alaska 1971 Adak 0.18 6.50 rical cyclic histories, respectively. The drift ratio is defined
as the top displacement divided by the column length. These
figures show the results for Section 1 (s1) with 5 and 10%
Lastly, the columns were subjected to eight seismic ALR and Section 2 (s2) with 5 and 20% ALR for Column 1.
acceleration time histories (Table 2) at the base that were Similar results were obtained with other columns (Vidot-
selected in order to have a range of different characteristics Vega 2008).The first number in the legend indicates the
(frequency content, duration, and number of cycles). number of cycles at a given drift level (1, 3, 5, or 7). The drift
Earthquakes with near- and far-field characteristics were () ratios can be related to displacement ductility (/y)
considered as part of the analysis. Table 2 shows the name and values. For example, a drift of 8% represents a displacement
year of the earthquake; location; recording station; the ductility value of 13.
moment magnitude Mw , which is a measure of the total energy Comparing the symmetric displacement history results in
released by an earthquake (Hanks and Kanamori 1979); and Fig. 3 and 4, the solid lines represent the strain-versus-drift
the peak ground acceleration (PGA) in g units. ratio response for monotonic analysis, whereas the symbols
The ground motions were normalized to 0.1g to begin represent the response for the cyclic histories. Note that the
each analysis with the same level of PGA. Then, the steel tension strains are greater for the cyclic histories when
accelerations were scaled in increments of 0.1 g up to 1.5 g. compared to the monotonic analysis. The concrete compression
The elastic damping in all the dynamic simulations was strains are lower than the monotonic analysis. It is also worth
represented by 2.0% tangent-stiffness proportional noting that the magnitude of the increase in the steel tension
damping (Priestley et al. 2007). In this study, the choice of strains is the same as the reduction in the concrete compression
damping model should not have any effect on the strains. In the case of asymmetric histories (Fig. 5 to 6),
relationship between strain and displacement. the difference between monotonic and cyclic strains is
For the monotonic and cyclic displacement histories, significantly reduced. Note also that as the axial load
concrete compression and steel tension strains were obtained increases in the section, the monotonic and cyclic strains
at each drift increment. In the case of the seismic acceleration become very similar for a given drift ratio.
histories, steel tension and concrete compressive strains were The difference between cyclic and monotonic steel and
obtained at the maximum drift for each dynamic analysis. In all concrete strains (cyclic-monotonic) was quantified and is
cases, concrete compression and steel tension strains were shown in Fig. 7 and 8 for symmetrical and asymmetrical

180 ACI Structural Journal/March-April 2011


Fig. 4Concrete compression strain versus drift ratio for symmetrical cyclic loading
(Column 1).

Fig. 5Steel tension strain versus drift ratio for asymmetrical cyclic loadings (Column 1).

cyclic histories, respectively, for the sections under The total steel tension strain that the reinforcing bars are
consideration. In general, the difference between cyclic subjected to is a combination of the flexural strain and
and monotonic steel and concrete strains increases as the growth strain (Moyer and Kowalsky 2003). The flexural
number of cycles and drift ratio are increased and the strain is obtained directly from monotonic analyses and flexural
axial load ratio is decreased. Strain differences as high as strength theory, whereas the growth strain is a function of the
0.012 and 0.0038 were obtained for symmetric and displacement history and material stress-strain relationships.
asymmetrical histories, respectively. Figure 9 shows the growth strain as a function of the drift
The differences between monotonic and cyclic results can ratio obtained for Column Model 1 with symmetrical cyclic
be partially explained by the calculation of the growth steel histories with one and seven cycles per drift level. As
strain gr , which is defined as the strain in the cross section expected, gr increases as the drift ratio increases and is
when the column is at zero displacement. The strain is somewhat larger for the column subjected to seven cycles of
uniform when the column is subjected to zero lateral loading at each drift level. The growth strain decreases as the
displacement; however, that strain is not necessarily zero. axial load in the section increases, which is expected, and
The growth strain can be thought of as the vertical elongation also explains why the strain differences between monotonic
of the column due to cyclic loading, and it has also been and cyclic analyses are greatly reduced as the axial load ratio
termed the residual strain at zero column displacement. increases. It is important to note, however, that in all cases,

ACI Structural Journal/March-April 2011 181


Fig. 6Concrete compression strain versus drift ratio for asymmetrical cyclic loading
(Column 1).

Fig. 7Strain difference for column models with symmetrical


cyclic loadings.

Fig. 9Growth strain for Column Model 1 for Section 1


(s1) and section 2 (s2) using symmetrical cyclic loadings
with one and seven cycles per drift.

Fig. 8Strain difference for column models with asymmetrical


cyclic loadings.
the growth strain is shown to be much larger than the difference
between the monotonic and cyclic strains, implying that a
direct addition of growth strain and flexural tension strain
may not be appropriate. Consider Fig. 10, which shows the
total steel strain as a function of the drift ratio for Column 1,
Section 1, subjected to one cycle of loading at each drift level Fig. 10Growth strain illustration for symmetrical cyclic
as well as monotonic loading. There are four curves shown loading (Column 1 - s1).

182 ACI Structural Journal/March-April 2011


Fig. 11Steel tension strain versus drift ratio for seismic time histories (columns).

Fig. 12Concrete compression strain versus drift ratio for seismic time histories (columns).

in Fig. 10. The solid curve represents the steel tension strain differences between the monotonic and cyclic strains are smaller
as a function of drift ratio for the monotonic analysis. As than the growth strain (strain at zero column displacement).
expected, at zero drift, the steel tension strain at the critical
section is zero. The remaining three curves show the relationship Impact of seismic time histories on
between steel tension strains and drift ratio for the response strain-displacement relationship
that occurred to a drift ratio of 0.08, 0.09, and 0.10. In each In addition to the controlled cyclic displacement histories,
case, at zero drift ratio, the steel tension strain in the critical seismic time histories were also considered as shown in
section is equal to the growth strain. As the column lateral Table 2. Fifteen analyses were performed with each earth-
displacement increases, note that the difference between the quake for a total of 120 analyses per combination of column
steel tension strain versus drift for the cyclic loading and the section and axial load. Figures 11 and 12 present steel
monotonic steel tension strain versus drift is not constant. tension (at extreme tension bar) and concrete compression
This implies that the growth strain, which is measured at zero strains (at extreme compression fiber) as a function of drift
displacement, is not constant and actually reduces as the ratio, respectively, for the eight earthquakes considered in
lateral displacement (drift) increases. Of course, this is this study for Columns 1 and 3. The results are compared
contradictory to the assumption that the growth strain can be against monotonic analyses (solid line). It can be noted that
added to the flexural tension strain, as suggested by Moyer the strains from the seismic time histories are very similar to
and Kowalsky (2003). This observation explains why the the strains from the monotonic analysis among all earth-

ACI Structural Journal/March-April 2011 183


a consequence, the results shown in this paper should not be
extrapolated to frames that do not satisfy this failure mechanism
(for example, older RC frames with deficient detailing
resulting in soft story mechanisms or extensive column
hinging due to dynamic amplification of moments in taller
buildings). For the monotonic pushover analyses, the lateral
load was applied by distributing the loads in the stories in
proportion to the product of the mass and displacement as is
done in the direct displacement-based design procedure. The
forces are distributed linearly for buildings up to four stories
and parabolic for buildings taller than four stories.
Fig. 13Strain differences versus drift ratio for Section 1
with 10% axial load (seismic time histories). Analysis techniques and displacement histories
Analyses of the RC moment frames were conducted under
Table 3Properties of RC frame models the influence of seismic acceleration time histories (applied
at the base). The same ground motions (Table 2) were
Model Case 1 Case 2 Case 3 Case 4 Case 5
considered as those for the member analyses previously
# story, discussed. Similarly, the material models are the same as
4, 2 8, 2 8, 2 12, 2 16, 3
bays
previously described for the member analyses. The bond slip
LB/HB 10 8 9 8 12
(yield penetration) was not directly modeled in the members.
Lcol, The plastic hinge length (Eq. (2)), however, does account for
12 (3657) 12 (3657) 12 (3657) 12 (3657) 12 (3657)
ft (mm)
this indirectly through the length of strain penetration, thus
Lcol /H 6.1 5.6 5.6 5.2 4 addressing the impact of bar slip on the strain-versus-
Beams, 16 x 20 18 x 24 18 x 24 18 x 26 26 x 33 displacement response (Priestley 1998; Priestley et al. 2007).
in. (mm) (406 x 508) (457 x 610) (457 x 610) (457 x 660) (660 x 838)
Again, the elastic damping in all the dynamic simulations
Columns, 24 x 24 26 x 26 26 x 26 28 x 28 36 x 36 was represented by 2.0% tangent-stiffness proportional
in. (mm) (610 x 610) (660 x 660) (660 x 660) (711 x 711) (914 x 914)
damping (Priestley et al. 2007). Steel tension and concrete
long
(beams)
0.017 0.020 0.020 0.023 0.030 compression strains at maximum drifts were obtained at each
analysis. The strains were obtained at the critical section in
long
(columns)
0.027 0.040 0.041 0.054 0.058 the beam plastic hinges because the desired mechanism of
inelastic deformation for these frames involves the formation of
the plastic hinges at the ends of the beams.

quakes. The trends are maintained for all columns and Strain-drift relationships for moment frames
different axial load cases beyond those shown in the paper. The strain-versus-drift behavior observed in the dynamic
The difference between strains (concrete and steel) from the analysis was compared to expressions previously developed
seismic time histories and the monotonic analysis was as by the authors, which were based on monotonic analysis.
large as 0.0035 for a specific earthquake (EQ4), however, These expressions, shown in Eq. (3) and (4) for steel tension
they are generally much smaller than 0.0035 (Fig. 13). Note and concrete compression strains, respectively, were developed
that only one section is shown in Fig. 13, and it was selected as a design tool. While a detailed discussion of the development
because it had the largest difference between the seismic and of these equations is outside the scope of this paper, a brief
monotonic strains. It is worth noting that in all cases, discussion is warranted. The interested is referred to Vidot-
concrete compression strain limits are more affected by Vega (2008) for further details. These equations were
changes in the strain and drift relationship because the same obtained through the use of moment curvature analysis of
shift in strain is a larger proportion of the total strain for several rectangular sections with distributed longitudinal
concrete compression than for steel tension. reinforcement and monotonic analysis of RC frames with a
beam-sway mechanism. From the section analysis, rela-
ANALYSIS OF MOMENT FRAMES tionships between curvature and material strain (steel and
Moment frame models concrete) were obtained. Then, curvatures were related to
A total of five moment frame models were developed. The drift in the RC frames by the use of plastic hinge lengths and
beam and column elements were modeled using the beam data from the monotonic analysis. The purpose of these
with hinges element (Scott and Fenves 2006). Several beam equations is to obtain the interstory drift required to reach
aspect ratios (LB/HB) were considered. The models consisted specific levels of concrete compression or steel tension
of four-story frames with two bays (LB/HB = 10), eight-story strains in the beams of RC moment frames to be used in
frames with two bays (LB/HB = 8 and 9), 12-story frames performance-based design methods such as direct displacement-
with two bays (LB/HB = 8), and 16-story frames with three based design.
bays (LB/HB = 12). The properties of the sections and In these equations, HB is the depth of the beams, Lp is the
elements of these models are shown in Table 3. Rectangular beam plastic hinge length (Eq. (2)); y is the yield strain of
beam and square column sections were used in all the models the steel; LB is the beam length; s is the steel strain; c is the
with distributed longitudinal reinforcement around the concrete strain; Ky is the yield curvature (Priestley 1998),
section perimeter. Constant column lengths (Lcol) were used which is given by Eq. (5); and s is the mechanical reinforce-
at all building floors. These frames were designed according ment ratio (Eq. (6)), which is a function of the longitudinal
to the beam-sway mechanism, where plastic hinges form at steel ratio long , the yield steel strength fy , and concrete
the ends of the beams and at the first-floor column base. As compression strength fc .

184 ACI Structural Journal/March-April 2011


Fig. 14Steel tension versus drift ratio for seismic time histories for frame models.

Steel tension strain tension and concrete compression) obtained at each maximum
drift from the analyses using the eight earthquakes (EQ1 to
L L EQ8) scaled at different PGAs. These results are shown for the
j = 0.5 y ------B- + ( 1.75 s s K y ) ------p- 0.70 s
0.15 0.080
(3) story that has the highest interstory drift (critical story).
H B H B
Note that, in general, there does not appear to be a signif-
icant difference on the relationship between strains and drift
Concrete compression strain ratio for monotonic analysis when compared to the seismic
analysis. The trends are maintained in all frame models
j = 0.5 y ------B + ( ( 1.6 s + 4 ) c + 0.004 K y ) ------p 0.85 c
L L 0.065
(4) beyond those described in this paper. It is observed that the
H B H B drift equations predict the drift for a given strain limit quite
well, especially for the steel strain limits. For concrete strain
where limits, the agreement is better for drifts less than 6%. Drift
levels of 6% correspond to beam displacement ductilities
Ky = yHB = 2.1y (5) that varied from 6.4 to 9.5, depending on the dimensions of
the beams in each model. The strain differences do not
exceed 0.0035 in any case. The highest differences were
f obtained for the four-story frame model while the smallest
s = long ----y- (6)
f c differences were noticed for the 16-story frame case. The
differences varied for each ground motion considered. Also,
the characteristics of the structure influenced the response to
Impact of seismic acceleration time histories
Whereas the impact of seismic time histories on the strain- a given ground motion. Several researchers have found
versus-displacement relationship was small for single similar results in which the story drifts are strongly dependent
columns, it is perhaps more important to study the impact on on the ground motion and structure characteristics (Gupta
moment frame structures. If the trend observed for single and Krawinkler 2000; Krawinkler et al. 2003; Ruiz-Garcia
columns continues for moment frame structures, it lends and Miranda 2006). The relationship between strains and
confidence to current practice where strain-based drift limits displacement, however, is minimally affected by the
states are based on monotonic analysis. This section presents different ground motions and the trends are maintained in all
the results obtained from the frame analyses using seismic building cases considered in this study. This is certainly a
acceleration time histories. Also, as previously discussed, a useful outcome because it builds more confidence in the use
comparison was performed using the drift equations developed of the strain limit states based on monotonic section analysis,
by Vidot-Vega (2008). as well as the use of monotonic analysis to describe the
Figure 14 shows the steel tension strain as a function of the envelope of the response of a structure subjected to cyclic
drift ratio for four of the frame models. Figure 15 presents and seismic ground motions.
the concrete compression strains as a function of the drift
ratio for four of the frame models. Also shown in these CONCLUSIONS
figures is a comparison with the monotonic results (solid This study presented a numerical study on the effects of
line) and the values obtained with the proposed drift equations displacement history on the strain-displacement relationship
(dash-dot line). The points correspond to the strains (steel of RC column members and moment frames. Column

ACI Structural Journal/March-April 2011 185


Fig. 15Concrete compression strains versus drift ratio for seismic time histories for
frame models.

models with 2 and 4% of longitudinal steel ratio were occur. This will ultimately require experimental studies with
analyzed with several axial load levels. The columns were careful measurement of strains well into the nonlinear range.
subjected to monotonic, symmetrical, and asymmetrical 7. It is important to note that the aforementioned
cyclic displacement histories and seismic acceleration time conclusions do not imply that damage is only a function of
histories. Also, a total of five frame building models were the maximum displacements. The effect of the inelastic
analyzed under the influence of seismic time histories. From excursions and duration of them cannot be excluded in the
the column and frame analyses, the following conclusions assessment of damage of RC structures. This study has
were drawn: addressed the impact of loading history on the relationship
1. The relationship between strain and displacement is between strain and displacement. Currently, performance
affected most by symmetrical displacement histories. limit states used for design are based on monotonic analyses
Asymmetrical displacement histories and seismic time because the envelope of a cyclic response is well defined by the
histories had less impact on this relationship. This is likely monotonic curve. This in turn leads to the assumption that
due to larger accumulation of residual strain that occurs the strain-versus-deformation relationship is unique and
under symmetric reversals of loading. This observation is independent of load history. Of interest here was the accuracy
consistent with the experimental results of Moyer and of that assumption.
Kowalsky (2003), which consisted of three symmetric and
one asymmetric cyclic tests (no seismic displacement histories NOTATION
were considered in their experimental work). Agross = gross section area
2. The differences between strains obtained using monotonic ALR = axial load ratio
and cyclic or seismic histories were small in magnitude. Small Aslong = longitudinal steel area
differences can be more significant for concrete strain limits. dbl = longitudinal bar diameter
E = concrete modulus of elasticity
3. The relationship between strain and displacement is less
Fu = ultimate steel strength
affected as the axial load in the column increases. fc = compressive concrete strength
4. The simple addition of the growth steel strain to the fy = yield strength of the longitudinal steel
flexural tension strain is not correct, as the difference H = column depth
between cyclic and monotonic strains reduces with increased HB = beam depth
lateral deformation. Ky = yield dimensionless curvature
5. Similar trends were obtained from the moment frame L = member length
analyses. Small differences in strains for a given displacement LB = beam length
were noted depending on the earthquake and characteristics Lcol = column length
of the structure under consideration. Lp = plastic hinge length
P = axial load
6. It can be concluded that monotonic section analyses and c and s = concrete compression and steel tension strains
the strain-based limit states derived from them can be used y = steel yield strain
with reasonable confidence for seismic design of concrete = displacement ductility
moment frames, as seismic time histories generally had a j = interstory drift
minimal impact on the relationship between strain and y = frame yield drift
displacement. Further research is still needed to quantify the long = longitudinal reinforcement ratio
effect of load history on the strain at which key limit states s = mechanical reinforcing ratio

186 ACI Structural Journal/March-April 2011


REFERENCES McKenna, F.; Fenves, G. L.; Scott, M. H.; and Jeremic, B., 2000, Open
El-Bahy, A.; Kunnath, S.; Stone, W.; and Taylor, A., 1999a, Cumulative System for Earthquake Engineering SimulationOpenSees, http://
Damage on Circular Bridge Columns: Benchmark and Low Fatigue Tests, opensees.berkeley.edu.
ACI Structural Journal, V. 96, No. 4, July-Aug., pp. 633-641. Moehle, J., and Kunnath, S., 2006, Reinforcing Steel: OpenSees Users
El-Bahy, A.; Kunnath, S.; Stone, W.; and Taylor, A., 1999b, Cumulative Manual, http://opensees.berkeley.edu.
Damage on Circular Bridge Columns: Variable Amplitude Tests, ACI Moyer, M., and Kowalsky, M., 2003, Influence of Tension Strain on
Structural Journal, V. 96, No. 5, Sept.-Oct., pp. 711-719. Buckling of Reinforcement in Concrete Columns, ACI Structural Journal,
Gupta, A., and Krawinkler, H., 2000, Estimation of Seismic Drift V. 100, No. 1, Jan.-Feb., pp. 75-85.
Demands for Frame Structures, Earthquake Engineering & Structural Priestley, M. J. N., 1998, Brief Comments on Elastic Flexibility of RC
Dynamics, V. 29, pp. 1287-1305. Frames, and Significance to Seismic Design, Bulletin of the New Zealand
National Society of Earthquake Engineering, V. 31, No. 4, pp. 246-259.
Hanks, T., and Kanamori, H., 1979, A Moment Magnitude Scale,
Priestley, M. J. N.; Calvi, G. M.; and Kowalsky, M. J., 2007, Displace-
Journal of Geophysics, V. 84, No. B5, pp. 2348-2350.
ment-Based Seismic Design of Structures, IUSS Press, Italy, 721 pp.
Ingham, J.; Liddell, D.; and Davidson, B., 2001, Influence of Loading
Ruiz-Garcia, J., and Miranda, E., 2006, Evaluation of Residual Drift
History on the Response of a RC Beam, Bulletin of the New Zealand
Demands in Regular Multi-Storey Frames for Performance-Based
National Society of Earthquake Engineering, V. 34, No. 2, pp. 107-124.
Seismic Assessment, Earthquake Engineering & Structural Dynamics,
Ingham, J.; Liddell, D.; and Davidson, B., 2002, An Assessment of V. 35, pp. 1609-1629.
Parameters Describing a Reinforced Concrete Beam, Bulletin of the New Scott, M., and Fenves, G., 2006, Plastic Hinge Integration Methods for
Zealand National Society of Earthquake Engineering, V. 34, No. 2, pp. 107-124. Force-Based Beam-Column Elements, Journal of Structural Engineering,
Karsan, I., and Jirsa, J., 1969, Behavior of Concrete under ASCE, V. 132, No. 2, Feb., pp. 244-252.
Compressive Loadings, Journal of the Structural Division, ASCE, V. 95, Tomazevic, M.; Lutman, M.; and Petkovic, L., 1996, Seismic Behavior
No. 12, pp. 2543-2563. of Masonry Walls: Modeling of Hysteric Rules, Journal of Structural
Kent, D. C., and Park, R., 1971, Flexural Members with Confined Engineering, V. 122, No. 9, Sept., pp. 1048-1054.
Concrete, Journal of Structural Engineering, ASCE, V. 97, No. 7, July, Tsuno, K., and Park, R., 2004, Prediction Method for Seismic Damage
pp. 1969-1990. on RC Bridge Columns, Structural/Earthquake Engineering, JSCE, V. 21,
Krawinkler, H.; Medina, R.; and Alavi, B., 2003, Seismic Drift and No. 2, July, pp. 97-111.
Ductility Demands and their Dependence on Ground Motions, Engineering Vidot-Vega, A. L., 2008, The Impact of Load History on Deformation
Structures, V. 25, pp. 637-653. Limit States for the Displacement-Based Seismic Design of RC Moment
Mander, J. P.; Priestley, M. J. N.; and Park, R., 1988, Theoretical Stress- Frame Buildings, PhD dissertation, Department of Civil, Construction and
Strain Model for Confined Concrete, Journal of Structural Engineering, Environmental Engineering, North Carolina State University, Raleigh,
ASCE, V. 114, No. 8, Aug., pp. 1804-1826. Raleigh, NC.

ACI Structural Journal/March-April 2011 187


View publication stats

Anda mungkin juga menyukai