You are on page 1of 11

Construction and Building Materials 145 (2017) 292302

Contents lists available at ScienceDirect

Construction and Building Materials


journal homepage: www.elsevier.com/locate/conbuildmat

Stress-strain response of A706 grade 80 reinforcing steel


David Overby, Mervyn Kowalsky , Rudolf Seracino
Department of Civil, Construction, and Environmental Engineering, North Carolina State University, United States

h i g h l i g h t s

 A706 grade 80 use in seismic design is currently restricted.


 788 tensile tests of A706-80 rebar were conducted.
 Values of five major stress-strain parameters were statistically identified.
 No observable susceptibility to strain aging.

a r t i c l e i n f o a b s t r a c t

Article history: Before ASTM A706 grade 80 rebar may be specified in the seismic design of structures, its mechanical
Received 27 July 2016 properties must be well understood and calibrated on a statistical basis. Based on the results of 788 ten-
Received in revised form 18 March 2017 sile tests of A706 grade 80 rebar encompassing bar sizes No. 4No. 18 (approx. metric No. 13No. 57), five
Accepted 26 March 2017
major stress-strain parameters are statistically evaluated and used to develop recommendations for an
Available online 11 April 2017
expected monotonic stress-strain curve. An existing material model is shown to accurately capture the
shape of the monotonic stress-strain curve. Additional tests are used to evaluate the strain-aging perfor-
Keywords:
mance of the steel.
Seismic
Reinforced concrete
2017 Elsevier Ltd. All rights reserved.
Reinforcing steel
Grade 80

1. Introduction inelastic strain capacity and strain hardening such that plasticity
is spread over a sufficiently long length. Furthermore, strength
The basic principles of seismic design follow the capacity design properties should be tightly controlled to ensure efficiency in
philosophy as outlined by Paulay and Priestley [26] that consists of design by limiting the overstrength factor for the design of capacity
three steps: (1) Locations of inelastic action are chosen; (2) The protected members and actions.
chosen locations are detailed to sustain the deformation demands In regions where high seismicity requires large quantities of
expected during the design basis earthquake; and (3) All other ele- longitudinal and transverse reinforcing steel to ensure adequate
ments of the system are protected against inelastic action. In the ductility, congestion at joints is a major problem. The use of high
case of seismic design of reinforced concrete bridges, locations of strength reinforcing steel in these cases offers a potential solution
inelastic action occur in the columns, while all other actions in to this problem; however, one of the concerns associated with the
the column (i.e. shear), and all other elements in the bridge (i.e., use of high strength rebar in seismic design is the general trend
footing, cap-beams, joints, superstructure) are protected against that as the strength of the steel increases, its maximum elongation
failure. This role is switched in the case of reinforced concrete capacity reduces, a trend which could undermine its potential ben-
frames such that the columns are designed to remain elastic while efits. As such, numerical test data must be available to validate its
the beams dissipate energy though plastic hinge formation. In all use.
cases, it is the reinforcing steel that acts as the critical link between As of the writing of this paper, ASTM A615 and ASTM A706 rep-
a ductile response and a brittle failure. As a result, for seismic resent the most typical reinforcing steel designations in the Unites
applications, it is essential that reinforcing steel exhibit sufficient States. A615 reinforcement exhibits more loosely controlled mate-
rial properties [7] making it an undesirable choice in the context of
seismic design. However, A706 specifications contain minimum
Corresponding author at: Department of Civil, Construction, and Environmental
and maximum yield strength limits, as well as larger requirements
Engineering, North Carolina State University, Mann Hall, Stinson Drive, Raleigh, NC
27695-7908, United States. for elongations while also meeting chemical composition require-
E-mail address: kowalsky@ncsu.edu (M. Kowalsky). ments which enable the steel to be weldable [8]. Because of these

http://dx.doi.org/10.1016/j.conbuildmat.2017.03.200
0950-0618/ 2017 Elsevier Ltd. All rights reserved.
D. Overby et al. / Construction and Building Materials 145 (2017) 292302 293

properties, A706 steel is routinely specified in high seismic regions, three No. 7 A706 grade 80 bars for the purpose of calibrating
especially for members expected to form plastic hinges. numerical models, with the data available on the NEES website
Prior to December 2009, the only grade of reinforcing steel [20]. Data from the Rautenberg study was also shown in a later
available in the A706 specification was grade 60. Since that time, report by Wiss et al. [34].
ASTM has included requirements for 80 ksi (550 MPa) steel The second dataset first appeared in a report by Trejo et al. [33]
(A706 grade 80) in the A706 specification. The grade designation as part of a study on the seismic performance of 24-inch (610 mm)
denotes the minimum allowable yield strength of the steel. diameter circular reinforced concrete bridge columns constructed
It should be noted that grade requirements are specified in with A706 grade 80 reinforcement. A total of six of these half-
terms of a minimum allowable value. As a consequence, actual scale columns were constructed and tested using either No. 5 or
reinforcing steel strengths are typically higher than their specified No. 6 longitudinal reinforcement, No. 3 transverse reinforcement,
values since producing mills must maintain an average strength and either A706 grade 60 or A706 grade 80 steel. Three tensile tests
that is above the minimum. The resulting material overstrength were conducted for each bar size and the results presented in their
must be accounted for in seismic design practice such that the pos- paper. This study proved to be the most informative for compar-
sible failure of capacity protected members due to higher than ison purposes at it included both stress and strain data. The same
anticipated moment demands is avoided. In the absence of actual results appeared in a later paper by the same authors [9].
data, conservative estimates of material overstrength may be used,
however, this may lead to inefficient design that not only compli-
cates construction, but increases cost. 1.2. Code limitations

The overall lack of experimental data on A706 grade 80 rebar in


1.1. Summary of prior research the literature is reflected in the hesitancy of design codes to allow
its use in regions expected to form plastic hinges. In some cases,
Likely owing to the relatively recent addition of grade 80 rebar the use of A706 grade 80 reinforcement is directly restricted while
to the ASTM A706 specification, there is an overall lack of experi- in others it is passively restricted by setting upper limits on yield
mental data available in the published literature. This is com- strength that are below 80 ksi (550 MPa). A brief summary of the
pounded by the limited production of A706 grade 80 reinforcing guidelines (or lack thereof) for use of A706 grade 80 steel in design
steel as mills are hesitant to produce steel that is currently in lim- codes is presented below. More detailed summaries may be found
ited use, and users are reluctant to specify the material given the elsewhere as well as in several of the reports already mentioned
limited test data available. As of the writing of this paper, five [14,34,21,33].
reports [29,34,21,33,9] were found to either directly reference or ACI 318-14 Section 20.2.2 limits deformed reinforcement used
include material test results associated with A706 grade 80 rebar. in special seismic systems to be of grade 60 or lower because of
As several of the reports were written regarding the same data, the insufficient data to confirm applicability of existing code provi-
available A706 grade 80 stress-strain data is limited to two data- sions for structures using the higher grade [A706 grade 80] [4].
sets consisting of twelve tensile tests and the accompanying However, the commentary to Section 18.2.6 makes provision for
stress-strain curves. The available experimental data is further lim- higher grades where sufficient test data is available to support
ited in that only a few bar sizes have been considered and that their use: Section 18.2.1.7 permits alternative material such as
strains were generally not provided to accompany the included ASTM A706 Grade 80 if results of tests and analytical studies are
yield and tensile strength data. A short summary of the past presented in support of its use [4].
research is included below, however, the interested reader is Section 3.2.1 of the Caltrans SDC 1.7 states that the capacity of
referred to Overby et al. [24] and Overby [25] for a more thorough concrete components to resist all seismic demands except shear,
discussion of past research. Table 1 presents the results from the shall be based on most probable (expected) material properties
twelve tests. to provide a more realistic estimate for design strength [11].
Rautenberg et al. [29] presented the findings of a study on the While the code offers recommendations for the expected stress-
applicability of high strength reinforcement in reinforced concrete strain properties of ASTM A706 grade 60 rebar, the use of ASTM
columns resisting lateral earthquake loads. Their research, which A706 grade 80 reinforcing steel is not directly addressed.
was based on testing conducted as part of Rautenbergs PhD disser- Based on research by Shahrooz et al. [32], the AASHTO LRFD
tation at Purdue in 2011 [28], aimed to evaluate the 60 ksi Bridge Design Specification [2] permits the use of reinforcing steel
(414 MPa) limit imposed by the American Concrete Institute with specified minimum yield strength of up to 100 ksi (690 MPa)
(ACI) on the yield strength of rebar used in regions expected to for all elements and connections in Seismic Zone 1 where permit-
form plastic hinges [3]. They evaluated 8 columns constructed ted by specific articles. Section C5.4.3.3 states that Reinforcing
from either ASTM A706 grade 60, A706 grade 80, or A1035 grade steels with a minimum specified yield strength between 75.0 and
120 longitudinal reinforcement. They conducted tensile tests on 100 ksi may be used in seismic applications, with the Owners

Table 1
Summary of literature test data (1 ksi = 6.9 MPa).

Reference Bar No. of Yield Point (0.2% Yield Point (0.0035 Onset of Strain Tensile Strength Ultimate Strain Elong.% in 8
Size Tests offset) EUL) Harding inch
Stress, Strain, Stress, Strain, Stress, Strain, Stress, Strain, Stress, Strain,
ksi in/in ksi in/in ksi in/in ksi in/in ksi in/in
Rautenberg et al. 7 1 83 119 11.7
[29] 7 1 83 117 15.6
7 1 84 118 14.8
Trejo et al. [33] 3 3 85.6 0.0055 73.3 0.0035 N.A. N.A. 120.5 0.0947 85.2 0.1378 13
4 3 86.2 0.0051 85.4 0.0035 85.9 0.0084 114.3 0.1066 86.8 0.1555 14
5 3 86.1 0.0048 84.3 0.0035 85.5 0.0098 114.0 0.1225 93.9 0.1893 15
294 D. Overby et al. / Construction and Building Materials 145 (2017) 292302

approval, only as permitted in the AASHTO Guide Specifications for bars from three different heats for each of the 10 sizes. Within each heat and bar
size, three specimens were cut from each of three individual 20-foot (6 m) lengths
LRFD Seismic Bridge Design [2]. This implies that A706 grade 80
of bar. The total possible number of tests amounted to 810, but only 59 of the antic-
reinforcing steel is permissible, subject to specific constraints. ipated 81 No. 18 bar tests were performed due to an incompatibility of the test
Section 8.4.1 of the AASHTO Guide Spec. for LRFD Seismic setup with the transverse ribs on the bars from one of the mills. A summary of
Bridge Design [1] states that ASTM A 706 Grade 80 reinforcing the tensile testing matrix is provided in Table 2. All tests were conducted with
steel may be used in capacity-protected members as specified in the bars in the as-rolled condition. Further details related to materials and testing
may be found in Overby et al. [24].
Article 8.5 but shall not be used in members where plastic hinging
is expected. It is further stated in the accompanying commentary
2.2. Equipment used
that this allowance was made due to the strength control and elon-
gation characteristics of A706 grade 80 reinforcement, and that it As part of the experimental program, two distinct test setups were employed.
has not been permitted in plastic hinge regions due, in part, to a For smaller reinforcing bars (No. 4 through No. 10), a 200-kip (890 kN) MTS univer-
lack of stress-strain data. Only ASTM A615 grade 60 (in seismic sal testing machine was employed. However, for the No. 11, 14, and 18 bars, the
design categories B and C, with the owners approval) or A706 expected strength of the specimens exceeded the capacity of the MTS system. As
a consequence, a custom-built testing system was developed (Fig. 1) for the testing
grade 60 reinforcing steel is allowed in members expected to form
of these larger bars.
a plastic hinge. The MTS machine was operated at a displacement rate of 1 in/min (25.4 mm/
The Washington Department of Transportation Bridge Design min) in order to satisfy the testing speed requirements of ASTM specification
Manual [35] Section 5.1.2 permits the unrestricted use of A706 A370 Standard Test Methods and Definitions for Mechanical Testing of Steel Prod-
ucts which specifies an upper and lower pre-yield and post-yield testing speed as a
grade 80 reinforcement in regions having Seismic Design Category
function of the free length of the bar. All rebar samples were 30 inches (760 mm)
(SDC) A, but limits its use to only capacity protected members for long, which corresponded to a free length of 20 inches (500 mm).
SDCs B, C, and D. The design of the large bar testing system (see Fig. 1) consisted of three 200-kip
Section 1.5.5.1.17 of the Oregon Department of Transportation (890 kN) double-acting hydraulic jacks operated by an electric hydraulic pump,
Bridge Design and Drafting Manual is specifically devoted to the providing a total capacity of 600 kips (2670 kN). Reaction plates for the jacks were
5-inch (127 mm) thick and hexagonal in shape and contained a 3-inch (76 mm)
use of ASTM A706 grade 80 reinforcement [23]. The manual states
diameter hole at their centroids. The entire system was operated in an upright man-
that A706 grade 80 reinforcement may not be used in members ner such that the bars were tensioned vertically. The test specimens were anchored
designed for plastic seismic performance such as bridge columns at the top and bottom of the setup using a wedge-chuck system (Figs. 2 and 3) in
due to limited experimental testing. which the chucks reacted on the opposite faces of the plates as the jacks. A single
specimen length of 48 inches (1220 mm) was used for all No. 11 through No. 18
The Alaska DOT currently uses A706 grade 60 rebar for the
tests in order to accommodate the geometry of the test setup. An advantage of
design of members expected to form a plastic hinge; however, the three-jack two-plate system is that it leaves the entire middle length of the test
A706 grade 80 has been specified for capacity protected members specimen exposed to accommodate the Optotrak markers used to determine
in accordance with the AASHTO specifications (Elmer Marx, strains. A 200-kip (890 kN) load cell was used to record the forces and was placed
AKDOT, personal communication, April 1, 2016). on a fourth jack near the test setup but still connected to the same hydraulic lines.
This configuration of the load cell away from the bar served as a way to indirectly
Examination of these code provisions leads to the conclusion
measure bar forces in excess of 200 kips (890 kN) using a single 200-kip (890 kN)
that many of these provisions recognize the lack of test data, and load cell. The force readings from the load cell were adjusted to account for losses
as a consequence place limitations on the use of the material. occurring across the hydraulic lines using an adjustment factor obtained by per-
The implication is that pending the availability of additional data, forming a subset of tests in which a second 200-kip (890 kN) load cell was placed
directly in-line with the test specimen. A detailed explanation of the calibration
use of the material may be allowed, with some codes (ACI 318)
process is described in Overby et al. [24].
already stipulating this. An Epsilon Class B1 2-inch (50.8 mm) gage length extensometer was used to
record strains for all No. 4 through No. 10 bar tests. In addition to the extensometer,
1.3. Research objective the Optotrak Certus HD non-contact position measurement system [19] was
employed as well. It should be noted that all recommendations provided in this
paper were based on the data from the Optotrak system as it was used on all bar
Despite the limited test data on A706 grade 80 reinforcing steel
sizes and demonstrated to be consistent with the extensometer data obtained from
in the current literature, there is significant potential in its use as an the No. 4 through No. 10 bar tests. The extensometer was not used on the No. 11
alternative to A706 grade 60 steel. For example, congestion into through No. 18 bar tests as the more violent nature of the large bar fractures posed
members adjacent to column plastic hinges, such as footings and a risk of damage to the instrument.
The Optotrak system is a 3D noncontact position measurement system that
cap-beams, would be reduced given the smaller amount of steel
simultaneously tracks the location of target LED markers with an RMS accuracy
employed when compared to grade 60 steel. This could result in a of 0.1 mm and a resolution of 0.01 mm [19]. The system operates in the infrared
potential savings in material costs, transportation, and labor. As a (IR) spectrum where markers flash IR light at a predefined frequency of up to
result, a research program was initiated at NC State University to 4600 Hz. The outputs from the Optotrak are the x-y-z coordinates of each marker
quantify the non-linear behavior of A706 grade 80 reinforcing steel. relative to a pre-defined origin at each frame record. Strains are calculated by divid-
ing the change in distance between markers by the initial distance between them.
The primary objectives of the research described in this paper
Multiple gage lengths can be established on a single test specimen, thus allowing
are to (1) develop a database of tensile test results from a variety for calculation of the distribution of strain over the instrumented region (see
of bar sizes for identification of the expected mechanical proper- Fig. 4). This, combined with the Optotraks inherent ability to measure large, inelas-
ties, (2) investigate the shape of the monotonic stress-strain curve tic strains without causing damage to the system itself, affords a unique advantage
over more traditional instrumentation such as strain gages and extensometers. Fur-
including yielding and strain hardening behavior, (3) evaluate the
ther discussion on the uses and benefits of the Optotrak system can be found in the
capability of existing monotonic reinforcing steel models to simu- following resources: [13,24,25].
late the experimental findings, (4) provide numerical recommen- A spacing of 2 inches (50.8 mm) was chosen for the Optotrak markers in order to
dations for stress-strain parameters required to fully define a be consistent with the 2-inch gage (50.8 mm) length of the extensometer as well as
monotonic reinforcing steel model, and (5) investigate the strain conformance with the testing provisions of ASTM A370 (ASTM A370-15). Including
six of these 2-inch (50.8 mm) gage lengths inherently offered a way of measuring
aging susceptibility of the steel.
strains over three overlapping 8-inch (203 mm) gage lengths (Fig. 4) as strains can
be calculated between any two markers regardless of whether or not they are adja-
2. Materials and methods cent. Including three 8-inch (203 mm) gage lengths increased the likelihood that
fracture could be captured in an instrumented region of the test specimen. Addition-
2.1. Program summary ally, the ability to provide strain data in terms of an 8-inch (200 mm) gage length
offered compatibility with existing test data also in terms of an 8-inch (200 mm)
A total of 788 monotonic tensile tests on the ten major imperial bar sizes (No. 4 gage length. Ultimately, little difference existed between 2-inch (50.8 mm) and 8-
through No. 18 approx. metric No. 13 through No. 57) were conducted on A706 inch (200 mm) gage length measurements [25]. Final recommendations related to
grade 80 rebar manufactured by three different producing mills. Each mill provided strains were therefore based on the 2-inch (50.8 mm) gage length data.
D. Overby et al. / Construction and Building Materials 145 (2017) 292302 295

Table 2
Test matrix (1 ft = 305 mm).

Bar size Mills Heats 20-ft bars Specimens Tests


No. 4 3 3 3 3 81
No. 5 3 3 3 3 81
No. 6 3 3 3 3 81
No. 7 3 3 3 3 81
No. 8 3 3 3 3 81
No. 9 3 3 3 3 81
No. 10 3 3 3 3 81
No. 11 3 3 3 3 81
No. 14 3 3 3 3 81
No. 18 3 3 3 3 59
Total number of tensile tests 788

Fig. 2. Wedge-chuck system used to anchor No. 11, 14, and 18 bars (No. 18 bar
shown).

Fig. 1. Custom testing rig used for tension tests of No. 11, 14, and 18 bars (No.
18 bar shown).

3. Results and discussion

2.3. Strain age testing 3.1. Expected stress-stain behavior

Strain aging can be defined as the process by which a reinforcing bar develops Table 4 summarizes the results of the tensile testing portion of
increased strength and reduced ductility over time following inelastic deformation. the experimental program. The values in the table represent the
This behavior arises as a result of dislocation pinning at the molecular level in
combined results of all 788 tests for each parameter. Table 5 breaks
which small, interstitial atoms like carbon and nitrogen are freed through plastic
deformation of the material and allowed to migrate through the crystal structure the findings down according to the individual bar sizes. Fig. 5
until they accumulate at dislocation sites (irregularities in the molecular structure), demonstrates the location of each parameter on the stress-strain
preventing further slipping of the crystal planes [12]. The increased resistance to curve. Further clarification on how each parameter was deter-
sliding of the crystal planes results in an increase in strength (both yield and ulti-
mined and a discussion of the distribution of the data are pre-
mate) and reduction in ductility of the material. Numerous articles and papers have
been written on the strain aging properties of rebar [27,15,31,18]; however, no such sented in Sections 3.3.13.3.6. The stated recommendations for
investigation appears to exist for A706 grade 80 rebar. each parameter coincide with the mean value obtained for that
A total of 30 strain aging tests consisting of two bar sizes (No. 5 and No. 7) were parameter.
conducted for 3 separate pre-strain levels (0.0075, 0.0150, and 0.0300) and 5 aging The ASTM A706 specification sets a lower limit on the percent
periods. An additional 9 tests were conducted on just No. 7 bars, using the same
elongation at fracture as a means of ensuring the ductility require-
pre-strain levels, in order to investigate the impact of storing the bars at sub-
freezing temperatures on the rate of strain aging. The complete test matrix is pre- ments of the steel are met by the manufacturer. This process is
sented in Table 3. All of the strain age test specimens were selected from the same generally achieved by fitting the fractured ends of the bar together
mill and heat in order to better isolate the variables being considered. and re-measuring what was an 8-inch (200 mm) gage length prior
The following procedure was used for conducting the strain aging tests: (1) ten-
to testing. It should be noted though that because necking and frac-
sile test to one of the three pre-strain levels; (2) reduce force to zero and unload bar
from test setup; (3) store bar at ambient laboratory or freezing temperatures for
ture are localized phenomena, necessarily resulting in a non-
one of the five aging periods; and (4) tensile test to failure. uniform distribution of strain over the length of the bar, this
296 D. Overby et al. / Construction and Building Materials 145 (2017) 292302

in particular, the transition from elastic to inelastic behavior and


the curvature of the strain hardening region. Identification of the
stress and strain values at yielding and the onset of strain harden-
ing necessarily resolves the first of these issues; however, the
Wedge shape of the strain hardening region is not as readily obtainable.
Grip While not as critical as the expected values of stress and strain
at yield and ultimate and the strain at the onset of strain harden-
ing, the shape of the strain hardening curve is important in model-
ing applications such as moment curvature analysis. As such, it
becomes important to assess the ability of existing rebar models,
formulated for a different grade of steel, to accurately describe this
region.
Fig. 6 illustrates the ability of the Raynor model [30] to capture
the shape of the strain hardening region when calibrated with the
mean values presented in Table 4. Additionally demonstrated in
the figure is the similarity in shape of the A706 grade 80 strain
hardening curve to that of an A706 grade 60 curve. It is further
demonstrated in Overby et al. [24] that the experimental stress-
strain curves align consistently with existing A706 grade 80 curves
found in the literature. These findings indicate that there is no need
to develop a new monotonic stress-strain model.

No.18 Bar 3.3. Statistical analysis by parameter

It is clearly evident from Fig. 6 that a degree of variability exists


in the test results. Cumulative distribution functions (CDFs) offer a
Fig. 3. Interface between bar and wedge grips. means to graphically illustrate this variability as well as highlight
trends or anomalies in the dataset. Quantitatively, they can be used
to identify the percentage of data in a dataset that exists at or
approach to determining ductility is subjective to the location of
below a given value. Thus the CDF curve for a dataset is always
the fracture relative to the boundaries of the gage length. In mod-
centered about its median (as opposed to mean) value. This char-
eling applications, it is the ultimate tensile strain (strain at max
acteristic is demonstrated in Figs. 712 which present the CDFs
stress) that is used as the measure of ductility of the bar as this
for each of the five measured stress-strain parameters and the
marks the point of maximum strain before the onset of necking.
modulus of elasticity. Each figure contains the CDF of the individ-
For this reason, the percent elongation at fracture was not closely
ual data points (referred to as the empirical cdf or ecdf) as well as
investigated. A limited dataset of percent elongations at fracture
a best-fit probability distribution (excluding the modulus of
has been presented in Overby [25].
elasticity).
Three experimental data CDFs are plotted for each of the strain
3.2. Shape of the stress-strain curve and application of material models parameters. This approach offers a way to illustrate the variability
between the six 2-inch (50.8 mm) gage lengths used to capture
One of the expressed objectives of the present research was to strains for each test. One of the curves is composed entirely of
assess the shape of the stress-strain curve for A706 grade 80 rebar, the minimums of the six values from each test, one is composed

Fig. 4. Optotrak marker arrangement used for tensile tests (No. 4 bar shown).
D. Overby et al. / Construction and Building Materials 145 (2017) 292302 297

Table 3
Strain aging test matrix showing number of tests by bar size and aging period.

No. of tests 48 hours 10 days 30 days 60 days 6 months


No. 5 3 3 3 3 3
No. 7 3 3 3 3 3
No. 7 (freezing) 3 3 3

Table 4
Summary of tensile testing results and design recommendations by parameter (1 ksi = 6.9 MPa).

Modulus of Yield strength (ksi) Yield Strain at onset of strain Tensile strength Ultimate tensile Tensile-to-yield
elasticity (ksi) strain hardening (ksi) strain ratio
ADM1 EUL2 0.2%
OM3
Mean 27,888 85.0 84.3 85.1 0.0033 0.0074 112.5 0.0954 1.32
St. Dev. 1601 3.03 3.20 2.93 0.0003 0.0019 3.65 0.0055 0.03
COV 5.74% 3.56% 3.79% 3.45% 9.03% 26.17% 3.24% 5.80% 2.19%
95th Percentile 30,322 89.2 88.7 89.4 0.0038 0.0109 118.9 0.1024 1.36
5th Percentile 25,111 79.0 78.5 79.4 0.0029 0.0048 106.4 0.0845 1.28
Recommendations 28,000 85.0 0.0033 0.0074 112.5 0.0954 1.32
1
Autographic Diagram Method stress associated with the top-of-the-knee of the yield plateau.
2
Extension Under Load Method stress at a strain of 0.0035.
3
0.2% Offset Method stress at the intersection of the 0.2% offset line with the stress-strain curve.

Table 5
Summary of tensile testing means and standard deviations by bar size (1 ksi = 6.9 MPa).

Bar size Modulus of Yield strength (ksi) Yield strain Onset of strain Tensile strength Ultimate tensile
elasticity (ksi) hardening (ksi) strain
Mean St. Dev. Mean St. Dev. Mean St. Dev. Mean St. Dev. Mean St. Dev. Mean St. Dev.
No. 4 28,894 1482 88.0 2.48 0.0034 0.0003 0.0072 0.0020 114.2 2.59 0.0922 0.0062
No. 5 28,542 1407 86.7 2.14 0.0032 0.0002 0.0084 0.0015 112.1 2.04 0.0945 0.0055
No. 6 27,797 1716 86.1 2.06 0.0031 0.0002 0.0085 0.0025 113.5 2.84 0.0958 0.0057
No. 7 27,902 1204 86.3 3.01 0.0033 0.0004 0.0078 0.0012 114.2 3.28 0.0971 0.0045
No. 8 28,089 981 86.0 1.38 0.0032 0.0002 0.0069 0.0017 113.9 2.50 0.0957 0.0037
No. 9 28,126 1035 85.1 1.48 0.0031 0.0003 0.0065 0.0015 113.6 3.53 0.0956 0.0051
No. 10 28,472 1172 84.2 1.55 0.0031 0.0002 0.0056 0.0012 114.6 3.44 0.0959 0.0041
No. 11 26,982 1338 83.6 2.85 0.0034 0.0003 0.0084 0.0023 110.1 2.69 0.0955 0.0056
No. 14 26,143 1330 81.7 1.63 0.0034 0.0003 0.0076 0.0014 108.7 2.43 0.0971 0.0062
No. 18 27,994 2323 80.7 2.51 0.0033 0.0003 0.0076 0.0014 107.6 3.35 0.0945 0.0073

Tensile Strength The Kolmogorov-Smirnov goodness-of-fit test was used to eval-


uate the distributions of the experimental data. Five probability
distributions were considered for each parameter (normal, lognor-
mal, beta, gamma, and Weibull) using a 5% significance level. Pre-
vious research has investigated the use of the normal, lognormal,
and beta distributions to characterize rebar yield and tensile
Ulmate Tensile Strain

strengths for various grades and specifications [5,17,22,10]. Such


Onset of Strain Hardening
Stress

findings are useful in the area of probabilistic seismic risk assess-


ment in which sampling techniques such as Latin Hypercube sam-
pling are used within a Monte Carlo simulation to develop seismic
fragility curves that take into consideration the expected material
Yield Strain

properties and their associated distributions. A unique aspect of


the current work is the consideration of the strain parameter dis-
tributions in addition to the strength parameter distributions for
A706 grade 80 reinforcing steel. Table 6 summarizes the probabil-
ity distributions that were considered for each parameter and
Strain assigns a rank to which distributions were found to be the most
likely fits based on a 5 percent significance level.
Fig. 5. Identification of parameters on the monotonic stress-strain curve.

3.3.1. Yield strength


entirely of the maximums of the six values from each test, and a Three methods of determining the yield strength were evalu-
third curve is obtained by taking the average of the six values from ated: the Autographic Diagram Method or top-of-the-knee, the
each test. These curves are defined as the Min ecdf, Max Extension Under Load Method, and the Offset Method. The Exten-
ecdf, and Mean ecdf respectively, in the figures below. Only sion Under Load (EUL) yield strength was taken as the value of
the average strain CDFs have been evaluated for a best-fit proba- stress corresponding to a strain of 0.0035. The Offset (OM) yield
bility distribution. strength was taken as the value of stress corresponding to the
298 D. Overby et al. / Construction and Building Materials 145 (2017) 292302

A706 grade 80 Raynor model

A706 grade 80 experimental data


A706 grade 60 curve

Fig. 6. Comparison of the Raynor material model and an A706 grade 60 curve to the 788 grade 80 curves.

Yield Strength, MPa 1.00


538 552 565 579 593 607 621 634 648 662 676 690 0.90
1.00 Mean = 27,888 ksi
ASTM (192,280 MPa)
0.90 min 0.80
allowable Mean = 85.0 ksi
0.80 (586 MPa) 0.70
0.70
0.60
Probability
Probability

0.60
0.50
0.50
0.40 E=29000
0.40 Empirical CDF
Beta CDF Min - ecdf
0.30 0.30
Mean - ecdf
0.20 0.20
ASTM Max - ecdf
0.10 max
allowable 0.10
0.00
78 80 82 84 86 88 90 92 94 96 98 100 0.00
Yield Strength, ksi 22000 26000 30000 34000 38000 42000 46000 50000
Modulus of Elascity, ksi
Fig. 7. Yield strength beta and empirical CDFs including all bar sizes.
Fig. 9. Modulus of elasticity empirical CDFs considering all bar sizes.

1.00
1.00
0.90 Mean = 0.0033
0.90 Mean = 0.0074
0.80
0.80
0.70
0.70
Probability

0.60
0.60
Probability

0.50
Min - ecdf
0.40 0.50
Max - ecdf
Min - ecdf
0.30 Mean - ecdf 0.40
Gamma Max - ecdf
0.20 0.30
Mean - ecdf
0.10 0.20 Lognormal CDF
0.00 0.10
0.0020 0.0025 0.0030 0.0035 0.0040 0.0045 0.0050
Yield Strain,in/in 0.00
0.0020 0.0040 0.0060 0.0080 0.0100 0.0120 0.0140 0.0160
Fig. 8. Yield strain gamma and empirical CDFs including all bar sizes. Onset of Strain Hardening, in/in

Fig. 10. Strain at onset of strain hardening empirical CDFs including all bar sizes.
Lognormal distribution shown for reference.
intersection of the stress-stain curve with a 0.2% offset line running
parallel to the linear elastic region of the curve. All three methods
are permitted by ASTM A370 [6]. did not exhibit well-defined yield plateaus (98 of 788 tests) were
As very little difference existed in the results obtained between therefore not included in this dataset. The resulting yield strength
the three methods, the expected yield strength was identified as CDF is presented in Fig. 7.
the value of stress corresponding to the top-of-the-knee of the From the figure, it is immediately apparent that several of the
stress-strain curve at the onset of yielding. Test specimens that tests (48 out of 690 tests) fell below the ASTM lower limit of
D. Overby et al. / Construction and Building Materials 145 (2017) 292302 299

Tensile Strength, MPa beta distribution was identified as the best fitting distribution
676 689 703 717 731 745 758 772 786 800 814 827 841 855 based on the Kolmogorov-Smirnov statistic. This observation is
1.00 consistent with the work by Bournonville et al. [10] who found
ASTM
0.90 min the beta distribution to be an acceptable fit for A706 grade 60
allowable rebar.
0.80
0.70
Mean = 112.5 ksi 3.3.2. Yield strain
Probability

0.60 (776 MPa) The expected yield strain was identified as the value of strain
0.50 corresponding to the stress at the top-of-the-knee of the stress-
0.40 Empirical CDF
strain curve at the onset of yielding. Test specimens that did not
exhibit well-defined yield plateaus (98 of 788 tests) were not
0.30 95th Percenle
included in this dataset. The yield strain CDFs of the min, max,
0.20 Lognormal CDF
and mean Optotrak strains are presented in Fig. 8.
0.10 There did not appear to be any indication that the yield strain
0.00 varied as a function of bar size. No. 14 bars had the highest average
98 100 102 104 106 108 110 112 114 116 118 120 122 124 yield strain of 0.0034 and No. 10 bars had the lowest average yield
Tensile Strength, ksi strain of 0.0031. There was a 1.4% difference between the mean
yield strains based on a 2-inch gage length and those based on
Fig. 11. Tensile strength lognormal and empirical CDFs including all bar sizes. an 8-inch (200 mm) gage length. The median average yield strain
including all bar sizes and based on a 2-inch gage length was
0.0032.
1.00
As with the yield strengths, the gamma and beta distributions
0.90
Min - ecdf were found to be acceptable fits to the yield strain data at a 5 per-
0.80 cent significance level. The gamma distribution was identified as
Mean - ecdf
the best fitting distribution based on the Kolmogorov-Smirnov
0.70 Max - ecdf
statistic.
5% of Means
Probability

0.60
Weibull CDF 3.3.3. Modulus of elasticity
0.50
The modulus of elasticity was defined as the slope of the linear
0.40 Mean = 0.0954 elastic portion of the stress-strain curve. Specifically, it was defined
0.30 as the slope of the line passing between 0.2 times the ADM yield
strength and 0.8 times the ADM yield strength. Because a value
0.20
of stress and strain are necessary to define this parameter it is indi-
0.10 rectly based on Optotrak readings and therefore has three empiri-
0.00 cal CDF curves as shown in Fig. 9.
0.0600 0.0700 0.0800 0.0900 0.1000 0.1100 0.1200 0.1300 It should be noted that due to the nature in which the modulus
Ulmate Tensile Strain, in/in of elasticity was calculated, the average value cannot be backed
into by using the recommended yield strength and yield strain.
Fig. 12. Ultimate tensile strain Weibull and empirical CDFs including all bar sizes.
In other words, dividing fy recommended by ey recommended does not
equate to Ey average. Additionally, a rounded value of 28,000 ksi
(193,050 MPa) has been recommended for the modulus of elastic-
80 ksi (550 MPa). While this behavior was limited solely to the No. ity as it is anticipated that the most probable use of the data in
11, 14, and 18 bars, it was demonstrated through additional testing design or computational simulation would be to specify the
that this was not a result of the adjustment factor previously expected yield strength of the steel along with the modulus of elas-
described. Further discussion on this behavior has been presented ticity and then calculate a corresponding yield strain as part of the
in Mirza and MacGregor [17], Lim [15], and Nowak and Szerszen analysis. To illustrate, applying this approach would result in a
[22]. Another observation of the testing was that different bar sizes yield strain of 0.003. The modulus of elasticity was not evaluated
originating from the same heat could have distinctly different yield for a best-fit probability distribution.
strengths. This observation is addressed in greater detail in Overby No. 4 bars had the highest average modulus of elasticity of
et al. [24]. Table 5 summarizes the mean yield strengths for each 28,894 ksi (199,215 MPa), and No. 14 bars had the lowest average
bar size. The mean values ranged from 80.7 ksi to 88.0 ksi modulus of elasticity of 26,143 ksi (180,250 MPa); however, there
(556 MPa to 607 MPa) with the No. 18 bars having the lowest did not appear to be any indication that the modulus of elasticity
mean value and the No. 4 bars having the highest. The median varied as a function of bar size when considering the averages of
yield strength including all bar sizes was 85.2 ksi (587 MPa). the other bar sizes. The median average modulus of elasticity
The gamma and beta distributions were found to be acceptable including all bar sizes and based on the 2-inch gage length yield
fits to the yield strength data at a 5 percent significance level. The strains was 27,943 ksi (192,660 MPa).

Table 6
Probability distributions found to be acceptable fits to each parameter from the KS test at a 5% significance level by rank.

Normal Lognormal Beta Gamma Weibull


fy 1 2
ey 2 1
esh
fu 2 1 3 4
eu 1
300 D. Overby et al. / Construction and Building Materials 145 (2017) 292302

3.3.4. Onset of strain hardening consistent trend for these three sizes to have the lowest tensile
The onset of strain hardening was identified as the value of strengths a combined mean of 108.8 ksi (750 MPa) versus a com-
strain corresponding to the intersection of a horizontal line passing bined mean of 113.7 ksi (784 MPa) for the No. 4 through No. 10
through the 0.2% Offset Method yield strength and the slope of the bars. There did not appear to be any indication that the tensile
initial portion of the stain hardening curve. Test specimens that did strength varied as a function of bar size in the remaining sizes.
not exhibit well-defined yield plateaus are included in this dataset The median tensile strength including all bar sizes was 112.6 ksi
because they still exhibited strain hardening. In these cases, the (776 MPa).
onset of strain hardening generally coincided with the intersection The normal, lognormal, beta, and gamma distributions were
of the 0.2% offset line with the stress-strain curve. found to be acceptable fits to the tensile strength data at a 5 per-
The bimodal nature of the data (Fig. 10) illustrates the variabil- cent significance level. The lognormal distribution was identified
ity in yield plateau lengths from just past the yield point (around as the best fitting distribution, though followed closely by the nor-
0.4% strain) to greater than 1% strain. The combined dataset indi- mal and the beta distributions.
cates a decrease in yield plateau length as bar size increases from
No. 6 to No. 10. This trend was not reflected in any of the mills indi-
3.3.6. Ultimate tensile strain
vidually and was shown to result from a single mill having No. 6
The ultimate tensile strain was identified as the value of strain
bars with long yield plateaus and No. 10 bars with little to no yield
corresponding to the maximum value of stress recorded during the
plateaus. As a result, the No. 6 bars averaged the highest overall
test. This parameter is illustrated graphically in Fig. 12.
onset of strain hardening strains at 0.0085, and the No. 10 bars
There did not appear to be any indication that the ultimate ten-
averaged the lowest overall onset of strain hardening strains at
sile strain varied as a function of bar size. No. 7 bars had the high-
0.0056. There was a 0.2% difference between the mean onset of
est average ultimate tensile strain of 0.0971 and No. 4 bars had the
stain hardening strains based on a 2-inch (50.8) gage length and
lowest average ultimate tensile strain of 0.0922. There was a 0.06%
those based on an 8-inch (200 mm) gage length. The median aver-
difference between the mean ultimate tensile strains based on a 2-
age onset of strain hardening strain including all bar sizes and
inch (50.8) gage length and those based on an 8-inch (200 mm)
based on a 2-inch gage length was 0.0073.
gage length. The median average ultimate tensile strain including
Due to the bimodal shape, none of the considered probability
all bar sizes and based on a 2-inch gage length was 0.0955.
distributions offered acceptable fits to the data at the considered
Only the Weibull distribution was found to be an acceptable fit
significance level. The lognormal distribution offered the closest
to the ultimate tensile strain data at a 5% significance level.
fit despite failing the goodness-of-fit test and is therefore shown
in Fig. 10 for reference.
3.3.7. Analysis of variabilities
The preceding section utilized cumulative distribution func-
3.3.5. Tensile strength tions to illustrate the degree of variability in the experimental
The expected tensile strength was identified as the maximum results for each of the parameters presented as well as some break-
value of stress recorded during the test. This represents the point down of the data by bar size. While a similar approach of present-
at which strain hardening transitions to strain softening or ing and comparing CDF curves could also be used to make
necking. comparisons between the producing mills and heats for each
As indicated in Fig. 11, none of the tensile strength values fell parameter, the required number of figures becomes impractical.
below the ASTM lower limit of 100 ksi (690 MPa). The adjustment A much more concise approach is to consider the coefficient of
to the No. 11, 14, and 18 bar data previously described revealed a variation (CV) of the test results within each of these categories.

Table 7
Mill coefficients of variation and average CV across the mills.

CV fye eye esh fue esu


Mill 1 3.14% 7.74% 20.23% 2.28% 4.98%
Mill 2 3.67% 10.37% 24.67% 3.17% 6.73%
Mill 3 3.82% 7.87% 30.20% 3.74% 4.93%
Averages 3.54% 8.66% 25.03% 3.06% 5.55%

Table 8
Averages of coefficients of variation of all tests within each category (1 ft = 305 mm).

AVG CV fye eye esh fue esu


Mills 3.54% 8.66% 25.03% 3.06% 5.55%
Heats 1.72% 7.08% 14.79% 1.74% 5.18%
20-ft Bars 0.34% 4.40% 3.56% 0.25% 3.82%

Table 9
Number of standard deviations of strain age test away from benchmark value.

Tensile strength Ultimate tensile strain


No. 7 0.0075 0.0150 0.0300 No. 7 0.0075 0.0150 0.0300
2 days 0.53 0.98 0.73 2 days 1.17 2.06 1.09
10 days 0.17 0.26 0.23 10 days 0.02 0.05 1.00
30 days 0.05 0.11 0.22 30 days 5.54 1.85 4.73
60 days 0.22 0.35 0.05 60 days 0.92 0.72 0.84
180 days 1.19 0.02 0.07 180 days 0.45 2.05 1.32
D. Overby et al. / Construction and Building Materials 145 (2017) 292302 301

Table 7 illustrates this approach by first showing the coefficient of  Best-fit probability distributions were presented for both of the
variation for all of the tests related to a particular mill and then strength parameters as well as the strain parameters based on
summarizing these down to an average coefficient of variation the Kolmogorov-Smirnov goodness-of-fit test.
across the three mills. Practically, this can be seen as the average  The yield strength values were shown to be well represented by
variability within a given mill. Following this same approach, the beta distribution, while the yield strain values were shown
Table 8 presents the average coefficients of variation of 1) tests to be well represented by the gamma distribution. The strain at
within each of the mills, 2) tests within each of the heats (average the onset of strain hardening values showed the highest degree
variability within a given heat), and 3) tests within each of the 20- of variability and were not found to be well represented by any
foot (6 m) bars (average variability within a single 20-foot (6 m) of the considered distributions. The tensile strength values were
length of rebar). shown to be well represented by the lognormal distribution,
The results follow what would be expected in that there is more and lastly, the ultimate tensile strain values were shown to be
variability between tests within a heat than tests within a single well represented by the Weibull distribution.
20-foot (6 m) bar and more variability between tests within a mill  The A706 grade 80 stress-strain curve is nearly identical in
than tests within a heat. Past studies of reinforcing steel mechan- shape to the A706 grade 60 curve.
ical properties have indicated similar results [5]. The high variabil-  The shape of the strain hardening curve may be accurately char-
ity in the length of the yield plateau likely results from the fact that acterized using currently available monotonic rebar models.
this parameter is sensitive to a number of factors related to the  About 7% of the tests fell below the ASTM minimum yield
manufacturing process (grain refinement due to rolling, cooling, strength requirement of 80 ksi (550 MPa). This behavior was
etc.) as well as chemical composition [15,27]. As such, high vari- limited to the No. 11, 14, and 18 bars.
ability in this parameter is not unexpected.  The average modulus of elasticity tended to be about 1000 ksi
(6900 MPa) lower than the commonly recommended value of
3.3.8. Strain aging results 29,000 ksi (200,000 MPa).
Table 9 summarizes the results of the No. 7 bar strain age tests  The No. 11, 14, and 18 bars, overall, had lower yield and tensile
for the tensile strength and ultimate tensile strain parameters. The strength than the No. 4 through No. 10 bars. However, no trends
values in the table represent the number of standard deviations of were evident for any other parameters as a function of bar size.
the strain age test result from a benchmark value which is taken as  A706 grade 80 rebar did not show susceptibility to strain aging
the mean of the nine tensile test results from the same mill and for any of the considered pre-strain levels, even up to an aging
heat. Negative values imply that the strain age test result fell below period of 6 months, however, further studies on larger bar sizes
the benchmark value and positive values imply that it exceeded may be worthwhile.
the benchmark value.  A706 grade 80 rebar may be reliably specified for capacity pro-
As indicated in the table, the majority of the strain age test tected members which would reduce congestion in joints and
results fell within one standard deviation of the benchmark value. result in an economic benefit in terms of reduced material
The results in the table also illustrate little to no trend with either and labor costs.
aging period or pre-strain level. An indicator that observable strain  Future research will focus on behavior of plastic hinge regions
aging had occurred in the tests would be a decrease in the ultimate reinforced with A706 grade 80 rebar as spread of plasticity
tensile strain with increasing aging period and an increase in the and limit state strains may be influenced by the different mate-
tensile strength with increasing aging period. While not shown, rial response.
the same lack of a definable trend was observed in the No. 5 bar
tests and the low temperature No. 7 bar tests. As such, there was
no indication of strain aging susceptibility of A706 grade 80 rebar. Acknowledgments

The research described in this paper was supported by the Cal-


4. Conclusions
ifornia Department of Transportation under agreement #65A0525.
Their support is gratefully acknowledged. The conclusions are
An assessment of the stress-strain behavior of ASTM A706 grade
those of the authors alone and should not be construed as to imply
80 reinforcement was conducted at NC State University through
endorsement by Caltrans. The authors would like to recognize the
experimental testing and analysis. A total of 788 tensile tests of
three producing mills (Cascade, Gerdau, and Nucor) who graciously
No. 4 through No. 18 bars (approx. metric No. 13 through No.
donated reinforcing bars for this research. Additionally, we wish to
57), in the as-rolled condition, originating from three different pro-
acknowledge the help of the Concrete Reinforcing Steel Institute
ducing mills and including multiple heats of steel from each mill
(CRSI) and Bethany Hennings in coordinating with the mills to
provided a substantial dataset of test results that was otherwise
acquire the steel and for providing special access to the CRSI data-
lacking in the literature. These results were used to identify the
base of mill tensile test results. Finally, the authors would like to
expected values of key parameters necessary to define the mono-
extend their gratitude to the staff at the NC State University Con-
tonic stress-strain curve of A706 grade 80 rebar: the yield strength,
structed Facilities Laboratory, including Greg Lucier, Jerry Atkin-
yield strain, strain at onset of strain hardening, tensile strength,
son, and Jonathan McEntire, for their support and patience in
and ultimate tensile strain. The modulus of elasticity was also eval-
helping conduct the tests as well as several of the many undergrad-
uated. A subset of tests focused on the strain aging performance of
uates, including Aaron Stroud and Grayson Fulp, who helped pre-
the steel. Based on the work just described, the following conclu-
pare the bars for testing.
sions are presented:

 The current work increased the body of publicly available A706 References
grade 80 stress-strain data by over 650%. Recommendations for
[1] AASHTO, Guide Specification for LRFD Seismic Bridge Design, second ed. with
yield stress, yield strain, modulus of elasticity, strain at the 2012 Interim Revisions., American Association of State Highway and
onset of strain hardening, ultimate tensile strain, and tensile Transportation Officials, Washington, DC, 2011.
strength were provided in Table 4 that may be used as the basis [2] AASHTO, AASHTO LRFD Bridge Design Specifications, Customary U.S. Units,
seventh ed. with 2016 Interim Revisions., American Association of State
for future building code requirements. Highway and Transportation Officials, Washington, DC, 2014.
302 D. Overby et al. / Construction and Building Materials 145 (2017) 292302

[3] ACI Committee 318, Building Code Requirements for Structural Concrete [21] NIST, Use of High-Strength Reinforcement in Earthquake-Resistant Concrete
(ACI318-11) and Commentary, American Concrete Institute, Farmington Hills, Structures. NIST GCR 14-917-30 Report, prepared by the NEHRP Consultants
MI, 2011, p. 503. Joint Venture, a partnership of the Applied Technology Council and the
[4] ACI Committee 318, Building Code Requirements for Structural Concrete (ACI Consortium of Universities for Research in Earthquake Engineering, for the
318-14) and Commentary, American Concrete Institute, Farmington Hills, MI, National Institute of Standards and Technology, Gaithersburg, Maryland, 2014.
2014. [22] A.S. Nowak, M. Szerszen, Calibration of design code for buildings (ACI318):
[5] D.E. Allen, Statistical study of the mechanical properties of reinforcing bars, Part 1-statistical models for resistance, ACI Struct. J. 100 (2003) 377382.
Building Research Note No. 85, Division of Building Research, National [23] ODOT, Bridge Design and Drafting Manual rev. April 2015, Oregon
Research Council, Ottawa, Canada, 1972, p. 8. Department of Transportation, 2015.
[6] ASTM A370, Standard Test Methods and Definitions for Mechanical Testing of [24] D.T. Overby, M. Kowalsky, R. Seracino, A706 Grade 80 Reinforcement for
Steel Products, American Society for Testing and Materials, West Seismic Applications Research Report CA16-2563, 2015. http://www.dot.
Conshohocken, PA, 2015, p. 50. ca.gov/newtech/researchreports/reports/2015/CA16-2563_FinalReport.pdf.
[7] ASTM A615/A615M-16, Standard Specification for Deformed and Plain [25] D.T. Overby, Stress-Strain Behavior of ASTM A706 Grade 80 Reinforcement
Carbon-Steel Bars for Concrete Reinforcement, ASTM International, West (MS thesis), North Carolina State University, 2016. https://repository.lib.ncsu.
Conshohocken, PA, 2016, p. 8. edu/handle/1840.16/11339.
[8] ASTM A706/A706M-16, Standard Specification for Deformed and Plain Low- [26] T. Paulay, M.J.N. Priestley, Seismic Design of Reinforced Concrete and Masonry
Alloy Steel Bars for Concrete Reinforcement, ASTM International, West Buildings, Wiley, New York, 1992, p. 744.
Conshohocken, PA, 2016, p. 7. [27] L.N. Pussegoda, Strain Age Embrittlement in Reinforcing Steels (Ph.D. thesis),
[9] A.R. Barbosa, T. Link, D. Trejo, Seismic performance of high-strength steel RC Dept. of Mechanical Eng., University of Canterbury, New Zealand, 1978.
bridge columns, J. Bridge Eng. (2015) 04015044, http://dx.doi.org/10.1061/ [28] J.M. Rautenberg, Drift Capacity of Concrete Columns Reinforced with High-
(ASCE)BE.1943-5592.0000769. Strength Steel (PhD dissertation), Purdue University, West Lafayette, IN, 2011,
[10] M. Bournonville, J. Dahnke, D. Darwin, Statistical Analysis of the Mechanical p. 263.
Properties and Weight of Reinforcing Bars, Research Report SL 04-1, University [29] J.M. Rautenberg, S. Pujol, H. Tavallali, A. Lepage, Drift capacity of concrete
of Kansas, 2004. columns reinforced with high-strength steel, ACI Struct. J. 110 (2) (2013) 307
[11] CALTRANS, CALTRANS Seismic Design Criteria Version 1.7, Department of 317.
Transportation, California, 2013. [30] D.J. Raynor, D. Lehman, J. Stanton, Bond Slip response of reinforced bars
[12] A.H. Cottrell, B.A. Bilby, Dislocation theory of yielding and strain ageing of iron, grouted in ducts, ACI Struct. J. 99 (5) (2002).
Proc. Phys. Soc. Sect. A 62 (1) (1949) 4962. [31] J.I. Restrepo-Posada, L.L. Dodd, R. Park, N. Cooke, Variables affecting cyclic
[13] J.C. Goodnight, S. Fulmer, N. Brown, M. Kowalsky, J. Nau, Experimental behavior of reinforcing steel, J. Struct. Eng. 120 (11) (1994) 31783196.
applications of 3D position monitoring systems for earthquake engineering [32] B.M. Shahrooz, R.A. Miller, K.A. Harries, H.G. Russell, Design of Concrete
research, in: Proceedings of the 10th National Conference on Earthquake Structures Using High-Strength Reinforcement, NCHRP Report 679,
Engineering, Earthquake Engineering Research Institute, Anchorage, AK, 2014. Transportation Research Board, National Research Council, Washington, DC.,
[14] D.P. Gustafson, Raising the grade, Concr. Int. 32 (4) (2010) 5962. 2011.
[15] W.T. Lim, Statistical Analysis of Reinforcing Steel Properties (Masters thesis), [33] D. Trejo, A.R. Barbosa, T. Link, Seismic Performance of Circular Reinforced
University of Canterbury, Christchurch, New Zealand, 1991, p. 151. Concrete Bridge Columns Constructed with Grade 80 Reinforcement, Research
[17] S.A. Mirza, J. MacGregor, Variability of Mechanical Properties of Reinforcing Project SRS 500-610, 2014.
Bars, J. Struct. Div., ASCE 105 (1979) 921937. [34] Wiss, Janney, Elstner and Associates, Inc. (WJE), Determination of Yield
[18] A. Momtahan, R.P. Dhakal, A. Reider, Effects of strain-ageing on New Zealand Strength for Nonprestressed Steel Reinforcement, WJE Project No. 2013.4171,
reinforcing steel bars, Bull. N. Z. Soc. Earthquake Eng. 42 (2) (2009) 179186. Dec. 31, 2013.
[19] NDI, Optotrak Certus User Guide Rev. 5, Northern Digital Inc, 2011. [35] WSDOT, Bridge Design Manual LRFD. M 23-50.15, Department of
[20] NEES, NEES Project Warehouse, 2009. http://nees.org/warehouse/project/904 Transportation, Washington, 2015.
(last accessed Aug. 21, 2015).