Professional Documents
Culture Documents
Zhao 2007
Zhao 2007
⎛ d b ( mm ) f y ( MPa ) ⎞1⁄α
s y ( mm ) = 2.54 ⎜ ------------------- -------------------------- ( 2α + 1 )⎟ + 0.34 (3)
⎝ 8437 f c ′ ( MPa ) ⎠
EXAMPLES OF APPLICATION
To demonstrate the applicability of the zero-length
section element with the proposed material models and
the corresponding improvements to the analysis results, cyclic
responses of two concrete columns and a bridge T-joint
system were simulated using a fiber-based finite element
program and the results were compared with the experimental
Fig. 8—Influence of parameter Rc on cyclic bar stress versus data. For all examples, the existing Kent-Scott-Park concrete
slip relation. model with zero tension and the Giuffré-Menegotto-Pinto
steel model with isotropic strain hardening elements were
where σ* is the bar stress ratio, s* is the slip ratio, suy is used, respectively, to model the concrete and steel fibers.
the stress limit ratio, and sy′ is the slip recovered The Giuffré-Menegotto-Pinto steel model with isotropic
elastically as determined by the return stress divided strain hardening does not include any ratcheting effects. The
by the initial slope (K), as illustrated in Fig. 7. Kent-Scott-Park concrete model with zero tension assumes
• In Eq. (4), coefficient Rc, with typical values in the zero tension capacity, thus the tension stiffening is ignored.
range of 0.5 to 1.0, defines the shape of the reloading For all analytical simulations with the strain penetration
curve. Depending on the anchorage detail and the effects, the model parameters were determined as follows:
corresponding mechanism, it is possible for a bar with the yield slips were calculated as per Eq. (3) using the
sufficient anchorage length to exhibit pinching hysteretic reported material properties; the ultimate bar strengths were
behavior in the bar stress versus slip response, especially taken as 1.5fy as per Priestley et al.;22 the ultimate slips were
when it is anchored into a joint. The coefficient Rc will approximated to 35sy; the b factors were taken as 0.5; and the
permit the pinching characteristic to be accounted for in Rc factors were taken as 1.0 for the columns and 0.7 for the
the analytical simulation of the flexural member. The T-joint system. The reason for using two different Rc factors
lower end value of Rc will represent significant pinching was that the longitudinal bars in the cantilever columns had
behavior while a value of 1.0 will produce no pinching ample anchorage length and 90-degree hooks at the end,
effect, as demonstrated in Fig. 8. A comprehensive test whereas the column bars were terminated into the T-joint
program is required to establish a procedure to determine with straight ends and an anchorage length of 22db. The
the value of Rc. In the absence of test data, the Rc values suitable Rc values were determined by comparing the cyclic
chosen for the examples may be used in fiber-based analysis results with the measured force-displacement
analysis of similar structural problems. responses of the test units.
CONCLUSIONS
Well-designed flexural concrete members experience
rotations at the fixed end(s) due to bond slip that occurs as a
Fig. 11—Comparison of experimental and analytical results result of strain penetrating along fully-anchored longitudinal
for T-joint specimen IC1:35 (a) force versus displacement; bars into the adjoining concrete members. Focusing on
and (b) moment versus column end curvature. column and wall longitudinal bars anchored in footings and
bridge joints, an efficient method is proposed in this paper to
to-column joints. The concrete strengths on the day of testing model the bond slip rotation using a zero-length section
were reported to be 31 MPa for the column and 40 MPa for element that can be employed in nonlinear fiber-based analysis
the cap beam and joint. Under constant axial load of 400 kN, of concrete structures. A constitutive model that expresses
the column was subjected to cyclic lateral loading at a height the bar stress versus loaded-end slip response was developed
of 1829 mm above the column-to-cap beam interface. The for the steel fibers of the zero-length section element using
yield lateral displacement for the T-joint system was suitable experimental data reported in the literature. The
reported to be 17 mm with the corresponding lateral adequacy of the proposed monotonic response for the steel
resistance of 250 kN. The test joint experienced strength fibers was illustrated by comparing the theoretical and
deterioration at lateral displacement of 103 mm due to measured bar stress versus loaded-end slip responses of two
formation of large joint cracks and subsequent joint damage. pull-out tests conducted on fully anchored bars in concrete.
The simulation model included six fiber-based beam- Because of the lack of cyclic test data in the literature, the
column elements for the cap beam and four beam-column hysteretic rules for the bar stress versus loaded-end slip
elements for the column. An additional fiber-based beam- response were established using the available test data and
column element with the elastic column section properties observed responses of concrete members under cyclic loading.
modeled the joint. The zero-length section element (with the Advantages of the proposed method to improve fiber-
model parameters of sy = 0.51 mm, fy = 448 MPa, b = 0.5, based analysis of concrete structures was demonstrated by
and Rc = 0.7) was located between this elastic element and simulating cyclic response of two concrete cantilever
the adjoining column element. columns and a bridge tee-joint system. Simulated responses
Figure 11(a) compares the measured force-displacement were compared with the observed responses at both global
hysteresis response of the test unit with the analytical results and local levels. The analyses that used the proposed method
obtained with and without the strain penetration effects. The to model the strain penetration effects satisfactorily captured
analysis, which included the strain penetration effects, the deflections, force versus displacement hysteresis
produced force-displacement response that closely matched responses, strains in the longitudinal reinforcing bar, and
with the measured response in both loading directions. The section curvature of the test units. When the strain penetration
joint shear failure experienced by the test unit towards the effects were ignored, the force resistance at a given lateral
end of testing was not accounted for in the analytical model, displacement was overestimated, along with portraying
and hence the analysis slightly overestimated the force larger hysteresis loops. Most importantly, the local response
resistance at the maximum displacement. On the other hand, parameters such as the steel strain and section curvature,
the analysis that did not include the strain penetration effects which indicate the extent of structural damage, were
overestimated both the lateral load resistance and the grossly overestimated.
unloading-reloading stiffness. Based on these observations, it was concluded that: 1) the
The advantages of incorporating the strain penetration strain penetration effects should not be ignored in the analysis of
effects in the analysis is more pronounced in Fig. 11(b), in concrete members; and 2) the zero-length section element
which the column moment versus curvature histories at the incorporating the proposed constitutive model for the steel
beam-to-column intersection are compared. The analysis fibers can be used in nonlinear fiber-based analysis to accurately
that ignored the strain penetration effects overestimated the capture the strain penetration effects and thus the global and
column end curvature by approximately 90% towards the local responses of concrete flexural members. The proposed
end of the test, indicating that the bar slip due to strain method is versatile in that it can be used for modeling
penetration greatly affects the local response measures that concrete flexural members without limiting cross-sectional
are indicative of damage to the plastic hinge region. A shapes or direction of the lateral load. In addition, the
significant improvement to the moment-curvature response proposed constitutive model for the bar stress versus slip
prediction was obtained when the analysis included the response can be employed to capture the strain penetration
strain penetration effects. The predicted moment-curvature effects in models of concrete structures developed using
hysteretic loops, however, are noticeably broad along the other types of elements.
reloading path prior to intersecting the curvature axis. This
discrepancy is expected to be diminished when the values of ACKNOWLEDGMENTS
The study reported in this paper is from a PreNEESR project supported by the
the model parameters, especially su, b, and Rc, are refined. National Science Foundation (NSF) under Grant No. CMS-0324559. The
As previously discussed, an experimental investigation authors gratefully acknowledge the support of S. McCabe and D. Foutch,
designed to quantify the bar stress versus slip response as a who have served as program directors for this grant. The authors also