See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/344403309

Mechanical properties of steel reinforcing bars for concrete structures in central

Colombia

Article  in  Journal of Building Engineering · January 2021

DOI: 10.1016/j.jobe.2020.101858

CITATION READS
1 987

3 authors:

Julian Carrillo Harold Esteban Lozano Guzman

Nueva Granada Military University Nueva Granada Military University
199 PUBLICATIONS   815 CITATIONS    1 PUBLICATION   1 CITATION   

SEE PROFILE SEE PROFILE

Carlos A Arteta
Universidad del Norte (Colombia)
48 PUBLICATIONS   182 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Ground Motion Prediction Equations for Colombia View project

Ground Motion Models for Northern South America View project

All content following this page was uploaded by Julian Carrillo on 08 October 2020.

The user has requested enhancement of the downloaded file.

Share your article!
We are pleased to let you know that the final version of
your article Mechanical properties of
steel reinforcing bars for
concrete structures in central
Colombia is now available online, containing full
bibliographic details.

To help you access and share this work, we have

created a Share Link – a personalized URL providing 50
days' free access to your article. Anyone clicking on
this link before November 27, 2020 will be taken directly
to the final version of your article on ScienceDirect,
which they are welcome to read or download. No sign up, registration or fees are
required.

Your personalized Share Link:

https://authors.elsevier.com/a/1btDu8MyS8yKT2

We encourage you to use this Share Link to download a copy of the article for your
own archive. The URL is also a quick and easy way to share your work with
colleagues, co-authors and friends. And you are welcome to add the Share Link to
your homepage or social media profiles, such as Facebook and Twitter. 
Mechanical properties of steel reinforcing bars for concrete structures in central
Colombia
Julian Carrilloa,d,1, Harold Lozanob, Carlos Artetac,d
a Full professor, Department of Civil Engineering, Universidad Militar Nueva Granada, UMNG, Bogotá, Colombia
b Research Assistant, Universidad Militar Nueva Granada, UMNG, Bogotá, Colombia
c
Assistant professor, Department of Civil Engineering, Universidad del Norte, Barranquilla, Colombia
d Colombian Earthquake Engineering Research Network, CEER, www.ceer.co

Abstract

Reinforcing steel imported into Colombia from various countries, which does not meet the requirements of
local building codes, hinders the compliance with the life safety limit state intended by the local building code.
The results of a statistical analysis of the tensile mechanical properties that characterize the stress-strain curves
of steel bars marketed in Bogotá are presented and discussed. The experimental program was planned to
provide statistics of the main mechanical properties of grade 60 steel bars with bar diameter ranging from 9.5
mm to 25.4 mm subjected to monotonic tension loading. The study comprises 85 monotonic tensile tests and
85 bending tests, including five specimens per diameter for each of three manufacturing companies evaluated.
The measured results are compared with those reported by different studies conducted previously in Colombia,
Mexico, Canada and the US. Characteristic graphical stress-strain curves are proposed for each diameter
studied. These curves, along with the statistics reported for each bar diameter can be used for future analytical
studies on the structural performance and reliability of reinforced concrete structures in central Colombia.

Keywords: deformed bars, tensile testing, stress-strain curve, strain hardening, concrete structures.

1 Introduction

Actual mechanical properties of the reinforcing steel allow evaluating the expected variability of reinforced
concrete (RC) performance, and thus reduce the uncertainty in the seismic response assessment of structural
systems. Currently, there are no recent studies evaluating the mechanical properties of reinforcing steel
describing the constitutive stress-strain relationship of bars available in central Colombia. The mechanical
properties of reinforced concrete materials must comply with current regulations for earthquake-resistant
constructions. Requirements prescribed by the Colombian Code for Earthquake-resistant Construction, NSR-
10 [2] for RC structures are primarily those prescribed by version 2008 of ACI-318 [3]. In 2018, Toro [1]
reported illegal steel imported into Colombia from various countries, which does not meet the requirements of
local building codes and materials standards. This contraband material causes not only marketing reduction
for local steel producers but also, as indicated by Muñoz [4], an imminent risk to the structural safety of
concrete structures in Colombia, which is critical because more than 85% of the population in Colombia lives
in intermediate and high seismicity areas. A main problem of steel coming irregularly to Colombia from other
countries (e.g. China, Turkey or India) is that such steel may not comply with the required production
standards, exacerbating the collapse risk for earthquake-resistant construction, hence hindering the compliance
with life safety limit state intended by NSR-10.
Previous studies for characterizing the mechanical properties of reinforcing steel have been conducted
in different cities of Colombia such as Bogotá, Soacha, and Ocaña, as well as in other countries such as Mexico,
the US, and Canada. In Colombia, Cabrera and Koller [5] reported on the measured mechanical properties of
steel bars marketed in Bogotá and Soacha. These authors tested not only certified national steel bars marketed
in central Colombia, but also imported steel bars without quality certificate based on Colombian standards.
González et al. [6] conducted a study on the tensile behavior of steel bars marketed in Colombia from 7
different companies. Arévalo and Arévalo [7] compared the measured mechanical properties of bars
commercialized in Ocaña with those reported by each manufacturer. Mander and Matamoros [8] studied the
mechanical properties of steel bars marketed in the US, considering rebar of various strength grades. Tavio et
al. [9] studied the ratio between the maximum strength and the yield strength of bars marketed in India.

1
Corresponding author. Tel.: +57 1 6500000 Ext. 1705; fax: + 57 1 6370557. Cra. 11 #101-80, E-mail:
[email protected]
1
Rodríguez and Botero [10] conducted a study on the tensile behavior of steel bars commercialized in Mexico
City. These authors reported the compliance probability of yield strength of the rebar.
Recently, Carrillo et al. [11] evaluated the yield strength, maximum tensile strength, and ductility
capacity of electro-welded wire meshes marketed by four manufacturers in Bogotá. The test campaign of the
study comprised wires with diameter of 4 mm, 5 mm, and 6 mm, spaced at 150 mm. Results reported by these
authors demonstrated that welded-wire meshes available in Bogota do not meet the ductility requirements of
reinforcement for concrete structures subjected to seismic demands. Although the wire mechanical properties
cannot be compared with those of deformed bars, results of previous studies on steel wires available in Bogotá
motivate new studies for characterizing the mechanical properties of all types of reinforcing steel marketed
there. For instance, updated research evaluating the mechanical properties of reinforcing steel and describing
the constitutive stress-strain relationship of bars available in central Colombia is missing.
This article presents and discusses the results of a statistical analysis of the tensile mechanical properties
that characterize the stress-strain curve of steel bars marketed in Bogotá. It is worth recognizing that the range
of response observed in the estimation of an empirical stress-strain curve for steel is due to the inherent
variability of the material properties, the production process of the bars, and can also be attributed to measuring
errors, among other unknown causes. Hence, the range of observed parameters describing the constitutive
curves is associated to aleatory variability and epistemic uncertainty which are captured combined in all
measurements presented herein. The main objective of the study is providing statistics regarding the main
mechanical properties of grade 60 (G60) steel bars (fy = 60 ksi = 420 MPa) with bar diameter (db) ranging from
9.5 mm (3/8”) to 25.4 mm (1”) subjected to monotonic tension loading for application in the code-based
structural engineering practice. The mechanical parameters reported herein are (see Fig. 1) the modulus of
elasticity (Es), yield strength (fy), maximum (tensile) strength (fsu), yield strain (εy), strain hardening modulus
that characterizes the initial slope of the strain-hardening branch (Esh), strain at the onset of strain hardening
(sh), strain at maximum strength fsu (εsu), the maximum-to-yield strength ratio (k = fsu/fy), and bending capacity.
Stress and strain values at fracture are not evaluated in this study (fu and εu). The study comprises 85 monotonic
tensile tests and 85 bending tests, including five specimens per diameter for each of three manufacturing
companies evaluated. The tensile tests allowed obtaining the main parameters defining the stress-strain curves
of the bars, from which the corresponding statistical analysis is performed. The experimental campaign also
included physical and mechanical properties such as actual unit mass and bending capacity. Characteristic
graphical stress-strain curves are proposed for each diameter studied. These curves, along with the statistics
reported for each bar diameter can be used for future analytical studies on the structural performance and
reliability of RC structures in central Colombia.

Figure 1. Sketch of a stress-strain curve of deformed bar sample subjected to monotonic tension loading.

2 Previous Studies

2.1 Mexico
In 1995, Rodríguez and Botero [10] performed a statistical analysis of the main mechanical properties defining
the monotonic tensile stress-strain curve of steel bars marketed in Mexico. These authors reported the tests of
100 specimens having 600 mm in length with bar diameters (db) ranging between 7.9 mm (5/16”) and 31.8
mm (1¼”). Rodríguez and Botero [10] reported difficulties in determining the yield strength of the small
diameter bars (7.9 mm, 9.5 mm and 12.7 mm) because the yield plateau was not clearly defined in the stress-
strain curves. From the statistical analysis, these authors reported that the Normal distribution fit well the data
collected for each mechanical property. Rodríguez and Botero [10] also found that there is a 8.7% probability
that steel bars do not comply with the minimum yield strength (fy > 420 MPa, 60 ksi) in Mexico City.
2
 In 1984, Villanueva and Meli [12] reported a non-compliance probability of 21% for the nominal yield
strength of rebars in Mexico. Rodríguez and Botero [10] argued that quality control of reinforcing steel
production in Mexico must be enhanced to achieve the acceptance limits prescribed by European codes where
the acceptance criterion of fy is defined in terms of the 5th percentile. For tensile maximum strength (fsu),
Rodríguez and Botero [10] found that the mean values are similar to those reported by Villanueva and Meli
[12]. Because a yield plateau in the stress-strain curves of the small diameter bars was not clearly defined, the
mean and coefficient of variation (CV) of measured values of εy and εsh were not consistent and therefore
Rodríguez and Botero [10] assumed εy = εsh for these small diameter bars.

2.2 United States and Canada

In 2004, Bournonville et al. [13] studied the mechanical properties of steel bars produced in the United States
and Canada with diameters ranging between 9.5 mm (3/8”) and 25.4 mm (1”). The experimental campaign
comprised tests per ASTM A706 [14] standard of 100 samples from each of 29 steel manufacturers. The
mechanical properties analyzed were fy, fsu, and elongation. The results were reported in terms of manufacturer,
bar diameter, and steel grade. The summary statistics reported by Bournonville et al. [13] include the mean,
median, standard deviation (SD), CV, minimum and maximum values, asymmetry and kurtosis. Bournonville
et al. [13] found that less than 0.1% of the specimens did not comply the minimum values for fy and fsu. These
authors also noted that 1.2% of the measured values did not fulfill the minimum requirement for the elongation
parameters.
In 2017, Overby et al. [15] conducted a statistical analysis on the behavior of G80 steel bars (fy = 80 ksi
= 550 MPa) in the United States, using tensile tests of 788 specimens with diameter varying between 9.5 mm
(3/8”) and 57.3 mm (2 ¼ ”). Parameters evaluated in the study were Es, fy, fsu, y, sh and su (see Fig. 1), to
obtain representative values for defining analytical stress-strain curve for G80 steel bars. Overby et al.
observed that the stress-strain curves for G80 steel bars exhibited branches similar to those observed for G60
steel bars. These authors identified that 7% of the specimens did not comply with the minimum value of fy,
which is 550 MPa for G80 steel. The greatest bar diameters such as N11 (34.9 mm), N14 (43.0 mm), and N18
(57.3 mm) are those bar diameters failing in complying with this minimum value. The mean value of Es value
reported by Overby et al. [15] is 192,500 MPa (27,9 ksi), which is below the classical 200 GPa reported in
building codes.
In 2019, Mander and Matamoros [8] developed a numerical model to simulate the behavior of steel bars
that are currently used in the US buildings. Different grades of steel bars prescribed by US codes were included
in this study. For G60 steel bars, mean and CV for fy, fsu, Esh, εsh and εu and were 483 MPa and 5%, 786 MPa
and 3%, 6895 MPa and 21%, 0.013 and 39%, and 0.13 and 18%, respectively, and the mean value of the
elongation capacity was 16%. Mander and Matamoros [8] found the mean value of fsu/fy ratio for G60 steel
bars is 1.36 for steel bars in the United States.

2.3 Colombia
In 2005, González et al. [6] assessed experimentally the main mechanical properties of reinforcing steel bars
available in Colombia, such as fy, fsu, and fu, and the corresponding strain values (εy, εsu, εu). The specimens
were subjected to monotonic tensile loading to characterize the stress-strain curves. The experimental program
comprised 477 specimens from 7 different companies with diameter ranging between 9.5 mm (3/8”) and 25.4
mm (1”). Similarly to Rodríguez and Botero [10] in Mexico, the statistical analysis performed by González et
al. [6] confirmed that the measured data fit well the normal distribution. All the specimens tested by González
et al. [6] complied with the yield strength limits (fy-min = 420 MPa and fy-max = 540 MPa) prescribed by ASTM
A706 [14]. The elongation results found by González et al. [6] in Colombia are similar to those reported by
Rodríguez and Botero [16] in Mexico with the elongation of bars surpassing the 14% required by ASTM A706.
González et al. [6] proposed stress-strain curves associated with the 5th and 95th percentiles for each diameter
analyzed. These authors concluded that the quality control in Colombia to produce national steel in 2005 was
adequate.
In 2017, Cabrera and Koller [5] conducted a study on the tensile behavior of steel bars marketed in
Colombia, mainly in Bogotá and Soacha. These authors found not only certified steel from three national
companies but also found steel imported from Turkey and China. The experimental campaign of the study
included 40 specimens with nominal diameter of 1.27 mm (½”). Reported mechanical properties were nominal
mass, fy, fsu, and elongation. Cabrera and Koller found that 35% of all samples tested in monotonic tension
(30% imported and 5% national), did not meet the yield strength range (i.e between 420 MPa and 540 MPa)
required per ASTM A706. These authors also found that steel bars imported from China and Turkey can be
categorized as medium and low strength because mean values of fy were 350 MPa and 240 MPa, respectively.
3
Cabrera and Koller [5] observed that only 57.5% of the tested specimens meet the mass requirement per ASTM
A706, and 77.5% of tested specimens comply with the fsu requirement per ASTM A706 (i.e., fsu > 550 MPa).
In addition, Cabrera and Koller [5] found that all the samples comply with fsu/fy ratio, which must be at least
1.25 according to ASTM A706, and that the mean value of fsu/fy is 1.35 for N4 bars. Based on results of
numerical models, Cabrera and Koller argued that the reduction of yield strength from 420 MPa to 298 MPa
(minimum value observed in the study) may increase by 66%, 71% and 41% the areas of rebars required in
concrete beams, columns and footings, respectively.
In 2018, Arévalo and Arévalo [7] conducted a study on the tensile behavior of steel bars marketed in
Ocaña, Colombia. The experimental program reported by Arévalo and Arévalo [7] included 15 specimens
from three different steel companies. This author reported physical parameters of bars, the stress-strain curves
and the main mechanical properties such as fy, fsu, y and elongation. Arévalo and Arévalo [7] compared the
measured values with those reported by each company, which are associated to requirements per ASTM A706.
Arévalo and Arévalo [7] verified that the specimens studied meet the maximum tensile strength and elongation
values prescribed by ASTM A706. However, the curves reported by Arévalo and Arévalo [7] reflect that the
yield strength does not meet the minimum value established by ASTM A706. According to the curves reported
by Arévalo and Arévalo [7], the yield strength values of some companies are roughly 10% smaller than 420
MPa. The elongation of the specimens remained above 15%, hence the parameter is compliant with ASTM
A706.

3 Experimental program

The test program presented herein comprises 85 monotonic tensile tests and 85 bending tests of deformed steel
bars and was conducted in the Materials and Structures Laboratory of the Universidad Militar Nueva Granada
in Bogotá, Colombia. The experimental campaign includes reinforcing bars with a diameter (db) ranging
between 9.5 mm (3/8”) and 25.4 mm (1”). Steel bars were obtained from 3 different manufacturing companies
that commercialize reinforcing bars locally in Bogotá. These companies are identified herein as companies A,
B, and C. For tensile and bending tests, 10 specimens for each diameter of each company (5 specimens for
each type of test) were considered. The nomenclature assigned to ach specimens is X-NY-Z, where the first
letter X symbolizes the company (A, B or C), the second letter Y indicates de nominal diameter of the bar in
one-eight of an inch as custom in Colombia (e.g. N3 to N8 are bars with db of 3/8” to 8/8=1”), and the third
letter Z is the ID of the specimen.

3.1 Tensile tests

The ASTM A370 [17] standard contains the requirements and procedures for performing tensile and bending
tests of steel bars used in Colombia. When the mechanical requirements established in the standards are not
met, tests must be performed per ASTM A370 to determine the chemical composition of the steel because it
directly affects the mechanical parameters of the bars. The mass and length of each of the 85 specimens was
initially obtained to estimate the actual mass before tensile testing of the bars. In Colombia, the reinforcing
steel quality control is based on the ASTM A706 [14] as follows: the minimum and maximum values of yield
strength (fy) are 420 MPa and 540 MPa, respectively; the minimum value of tensile strength is (fsu) is 550 MPa,
the minimum elongation associated to a gauge length of 200 mm is 12% for bar diameters equal or greater
than 22.1 mm (7/8”), otherwise it is 14%.
The length of each test specimen for the tensile test was 600 mm. The tests were performed in a Controls
MC-66 universal testing machine with tensile load capacity of 1,000 kN. ASTM A370 specifies that loading
rate shall vary between 69 MPa/min and 700 MPa/min. The specimens in this study were tested using a mean
loading rate value of 385 MPa/min. Gauge marks were embed on each bar for defining an initial gauge length
of 200 mm for computing elongation after tensile testing per ASTM A370. The elongation capacity of each
bar specimen was assessed by comparing the change in length between the gauge marks with the initial 200
mm. Post-yield TML FLA-5-11-5L strain-gauges were adhered to the specimens for estimating the strain
during tensile testing. Load and strain data were recorded with a sampling frequency of 10 Hz. The
instrumentation of this study allowed obtaining stress-strain curves and identifying the mechanical parameter
of interest: modulus of elasticity (Es), yield strength (fy), maximum (tensile) strength (fsu), yield strain (εy),
strain hardening modulus (Esh), strain at the onset of strain hardening (sh), and strain at maximum strength fsu
(εsu) (see Fig. 1).

4
3.2 Bending tests
The bending test is prescribed in ASTM A370 as a method for assessing the ductility capacity of the material.
The 85 bending tests were performed using a bar bending testing machine, which ensures free displacements
at the ends of the bar. For the test, the bar specimen is bent 180° around a pin. The pin diameter is a function
of the diameter of each specimen as follows: 3 times de bar diameter (3d) for diameters ranging between N3
to N5, and 4d for bar diameters ranging between N6 to N8. The tests were performed on 600 mm length
specimen to ensure free bending. For compliance, the specimen must not have any cracks in the outer bending
zone after testing.

4 Results and discussion

This section summarizes and discusses the results obtained from the physical characterization, as well as the
tensile and bending tests. The results are contrasted with results of studies conducted in Colombia and other
countries. As recommended by Rustom [18], three levels of variation of the data are considered in this study
to characterize the measured results: high homogeneity for results with CV lower than 5%, moderate
homogeneity for CV around 20%, and low homogeneity for CV higher than 50%.

4.1 Mass, elongation and bending

Table 1 shows the mean values of the measured results of mass and elongation of each diameter studied.
Results in the table include bars from all the companies. Table 1 also shows the values of these parameters
prescribed by ASTM A706 [14], and whether the results comply or not with the standard. It is noted that only
specimens N3 comply with the prescribed nominal mass, although a marginal difference smaller than 3% is
observed between measured and nominal mass for the others, which exhibit less unit weight. Table 1 shows
that the mean value of elongation of N3, N4 and N5 bars do not meet the minimum elongation requirement
per ASTM A706. This contrasts elongation values greater than 15% for bars in Ocaña, Colombia [7]. Cabrera
and Koller [5] also reported 15% of elongation for N4 bars in Bogota and Soacha. Bar elongations reported by
Arévalo and Arévalo, Cabrera and Koller are similar to those found by Mander and Matamoros [8], who
reported elongations of 16% for bars in the US. Bournonville et al. [13] also reported 14% of elongation for
bars with diameters ranging between N3 and N8 in the US and Canada. In terms of 180° bending test, there
was no evidence of cracking in the outside-radius of bent for any specimen. Hence, all companies and
diameters comply with the requirement per ASTM A706.

Table 1. Measured results of mass and elongation percentage

Bar Mass, kg/m Elongation, %
diameter Measured Nominal Compliance Measured Prescribed Compliance
N3 0.56 0.560 Yes 13.7 14 No
N4 0.98 0.994 No 12.1 14 No
N5 1.52 1.552 No 13.8 14 No
N6 2.20 2.235 No 14.3 14 Yes
N7 2.95 3.042 No 13.3 12 Yes
N8 3.91 3.973 No 13.9 12 Yes

4.2 Mechanical properties from the stress-strain curves

Figs. 2 to 7 show the stress-strain curves measured for N3 to N8 bars. For comparison, the same scale was
used in the vertical and horizontal axes of the figures. The stress is obtained as the ratio between the measured
load and the nominal bar area of each diameter evaluated. Strains are directly recorded by the strain-gauges
adhered to the surface of the bars. Most of the stress-strain curves show an apparent strain hardening branch
after yielding; nevertheless, a salient feature of curves sets A-N4 and B-N5 is the lack of, or reduced strain
range of the yield plateau, which may be associated to values of chemical composition of the base material
higher than those prescribed by ASTM A706. A well-defined yield plateau is apparent in the stress-strain
curves of all N8 bars in Fig. 7. Lastly, it is worth noting that all stress-strain curves of company C exhibited a
well-defined yield plateau.

5
 Figure 2. Stress-strain curves of N3 bars of the companies: (a) A, (b) C.

Figure 3. Stress-strain curves of N4 bars of the companies: (a) A, (b) B, (c) C.

Figure 4. Stress-strain curves of N5 bars of the companies: (a) A, (b) B, (c) C.

Figure 5. Stress-strain curves of N6 bars of the companies: (a) A, (b) B, (c) C.

Figure 6. Stress-strain curves of N7 bars of the companies: (a) B, (b) C.

6
 Figure 7. Stress-strain curves of N8 bars of the companies: (a) A, (b) B, (c) C.

Postprocessing of the stress-strain curves allows estimating the characteristic mechanical parameters for
statistical and analytical analysis. The mechanical parameters reported herein are: Es, fy, fsu, εy, Esh, sh, εsu, and
k (fsu/fy) (see Fig. 1). Fig. 8 shows box and whisker plots summarizing the dispersion of four of these parameters
(fy, fsu, k and εsu). The data is organized by company and bar diameter, also including aggregate results for both.
For the fy data (Fig. 8a), reference lines at 420 MPa and 540 MPa set the minimum and maximum fy limits per
ASTM A706. For fsu (Fig. 8b), a reference line at 550 MPa sets the minimum fsu limit per ASTM A706. For
k=fsu/fy (Fig. 8c), a reference line at 1.25 sets the minimum k limit per ACI-318 [3].
It is noted in Fig. 8a that there are data points in test sets A-N3, B-N5 and C-N7 below the minimum
yield strength required. Of these three cases, that of company B is critical because the median strength is also
below the minimum requirement. Nevertheless, although these data points are lower than the aggregate median
for N5, all aggregate medians comply with the minimum requirement. Note at the bottom of the figure that, in
the aggregate plots for bar diameter (except for the N3 data), the inter-quartile ranges of the box plots lie close
to 420 MPa, without much dispersion (CV~3% on average). This, along with the fact that no data point
surpasses the maximum limit of fy, indicate that the production process of the companies focuses on complying
with the minimum strength requirement, and do not allow room for conservatism.
The maximum strength data is shown in Fig. 8b. All data points are above the minimum required tensile
strength limit of 550 MPa. According to Rustom [18], the CV values of fsu in this study are indicative of high
homogeneity. Except for the data in A-N3, data of companies A and B exhibit dispersion smaller than that of
company C. It is also noted that the maximum strength for company B is uniform across all bar diameters.
Although maximum strength for Company C meet the minimum strength requirement, it shows an apparent
decreasing trend with increasing bar diameter. Company C also shows larger dispersion for the two smaller
bar diameters, as company A does for the N3 set.
The maximum-to-yield strength ratio data is shown in Fig. 8c. It is noted that all data points comply
with the minimum k value of 1.25, being the data for company C closest to the minimum limit. The compliance
with minimum k value per ACI-318 [3] is an important feature of the steel employed for seismic applications
as it promotes the spread of plasticity over a longer portion of the plastic hinge region of members subjected
to axial load and moment.
The strain at maximum stress data is presented in Fig. 8d. This strain level marks the onset of bar
necking, upon which the real tensile stresses cannot be recovered from the engineering stress-strain curve in a
reliable manner anymore. Although the su data shows high dispersion due to the larger inherent variability
associated to strain measurements as compared to force-based measurement such as strength (e.g., see ASTM
A706 results in Bournonville [13], and Mander and Matamoros [8]), data of su measured in this study is lower
than that reported in previous studies. It is noted that all data points are above 8% strain which is deemed here
as a minimum value to consider the bar material as ductile. A reference line at su = 10% is introduced for
comparison purposes as the median of data across company and bar diameter hovers around it. In fact, the
median of the aggregate data including companies and bar diameter is 10%.
The aggregate data, including every specimen and company, comprises at least 70 data points per
parameter from which empirical cumulative distribution functions (eCDF) are estimated and normality tests
are performed. The goodness-of-fit tests by Lilliefors [19] and Anderson and Darling [20] are implemented
along with inspection of the quantile-quantile (qq) plots. An advantage of these two statistical tests with respect
to the commonly employed Kolmogorov-Smirnov test [21] is that the population mean and standard deviation
do not need to be known, as is the case of the data set presented herein. The null hypothesis of these tests is
that the “sample distribution is Normal.” The latter is statistically rejected when the p-value is lower than the
chosen significance level, typically 5%, as selected here. Fig. 9 shows two examples of the histograms and qq-
plots of the observed data of the parameters in which one rejects and the other fails to the reject the null
7
hypothesis (i.e., the data is Normally distributed). In Fig. 9, pL indicates the p-value for the Lilliefors test [19],
while pAD is that of the Anderson-Darling test [20]. It is noted that, for the aggregate dataset, fy does not follow
a normal distribution (i.e. the fy data has p-values larger than 0.05), and that Es does follow a Normal
distribution. Both claims are confirmed by inspection of the qq-plots, in which the tails of the fy-data deviates
from the normality assumption. Table 2 summarizes the results obtained for the Lilliefors and Anderson-
Darling tests for each parameter of the stress-strain curve considered here, including elongation. It is noted
that three parameters statistically rejected both goodness-of-fit tests, for example, fy, sh and Esh, although, for
the latter, the p-values are not too small. Others confirm normality with both tests, for example Es and
elongation, while the rest have mixed results. For these cases, the authors inspected the qq-plots and made the
final decision in the last column of Table 2. In summary, except for fy, sh, all parameter samples comply with
the normality assumption.

(a) 550 540 (b)

650

fsu (MPa)
fy (MPa)

500
A

A
450 600
420
400 550 550
550 540
650

fsu (MPa)
fy (MPa)

500
B

B
450 600
420
400 550 550
550 540
650

fsu (MPa)
fy (MPa)

500
C

450 600
420
400 550 550
550 540
650
Aggregate

Aggregate

fsu (MPa)
fy (MPa)

500

450 600
420
400 550 550
Agg.

Agg.
N3

N4

N5

N6

N7

N8

N3

N4

N5

N6

N7

N8
1.7
(c) (d) 0.12
1.6
k = fsu / fy

1.5
εsu

0.10
A

0.10
1.4
1.3 1.25
0.08
1.7
0.12
1.6
k = fsu / fy

1.5
εsu

0.10
B

0.10
1.4
1.3 1.25
0.08
1.7
0.12
1.6
k = fsu / fy

1.5
εsu

0.10
C

0.10
1.4
1.3 1.25
0.08
1.7
0.12
1.6
k = fsu / fy
Aggregate

Aggregate

1.5
εsu

0.10
0.10
1.4
1.3 1.25
0.08
Agg.

Agg.
N3

N4

N5

N6

N7

N8

N3

N4

N5

N6

N7

N8

Figure 8. Dispersion of recorded data: (a) fy, (b) fsu, (c) k=fsu/fy, (c) εsu.

8
 Table 2. p-values for test of normality
Normality
Lilliefors test Anderson-Darling test
Parameter assumption
pL Pass pAD Pass Pass
fy 0.001 No ~0.000 No No
fsu 0.047 No 0.051 Yes Yes
k 0.074 Yes 0.022 No Yes
y 0.224 Yes 0.043 No Yes
sh ~0.000 No ~0.000 No No
su 0.014 No 0.142 Yes Yes
Elongation 0.508 Yes 0.521 Yes Yes
Es 0.590 Yes 0.647 Yes Yes
Esh 0.035 No 0.016 No Yes

Figure 9. Histograms and qq-plots: (a, b) yield strength, (c, d) modulus of elasticity.

4.3 Summary statistics of the stress-strain curve parameters

Table 3 summarizes the average (X) and the coefficient of variation (CV) of the parameters estimated from the
stress-strain curve. The data is organized by company name and bar diameter for a subsequent comparison
with other studies. From the aggregate results by diameter and company, it is noted that the parameters
resulting with high homogeneity (i.e. with average CV < 5%) [18] is fsu, followed by fy, k and Es, with CV <
7%. The rest of the parameters may be classified to have moderate homogeneity with the following CVs: su
(~9%), elongation (~14%), y (~23%), Esh (~20%), and sh (~43%). This last statistic confirms the annotation
regarding the variance in the observed yield plateau extension, which also affects how Esh is estimated. The
data in Table 3 shows that mean value of fy for all the specimens of the different companies were greater than
420 MPa and lower than 540 MPa, as indicated in ASTM A706 [14]. The variations for fy obtained in this
study for the different companies of all diameters were lower than 8.3% as shown in Table 3, so they are
considered values with moderate homogeneity as proposed by Rustom [18].
The aggregate data of each bar diameter and each company are key parameters for code-based design.
The eCDF of the aggregated data allows estimating the probability that a given parameter takes on a value less
than or equal to certain test level; for example, F(z) = P(fy ≤ z) indicates the probability that fy is less than or
equal to test value z. Fig. 10 shows the eCDF along with the theoretical Normal CDF estimated with the mean
(X), and the standard deviation (s) of the aggregated data contained in Table 4. The second (P2), the median
(P50) and the ninety-eighth percentile (P98) are marked as horizontal lines for reference, and are also
summarized in Table 4. Table 4 also includes the standard error (SE) for confidence interval estimations. The
9
P2 value can be used by the local community of civil engineers as a safe lower bound limit for analysis and
design purposes. The mean and standard deviation can be used to recover the theoretical normal density
functions when applicable, for example for future reliability studies on local structures. Note that the theoretical
normal distribution does not fit well the fy and sh empirical data as concluded before in Table 2.

Table 3. Summary of statistics for mechanical properties of steel tested

Statistical
Company Es Esh
Diameter Paramete fy (MPa) y sh fsu (MPa) su k=fsu/fy
* (MPa) (MPa)
r
X 450.0 0.0024 201727 0.0112 6623 644.9 0.1034 1.44
A
CV, % 7.7 8.2 3.5 10.0 10.2 4.1 0.9 5.1
X 508.9 0.0030 214427 0.0150 4279 671.5 0.1149 1.32
N3 C
CV, % 1.0 7.4 3.6 6.7 4.6 2.2 4.0 1.9
X 476.2 0.0026 207372 0.0129 5582 656.7 0.1111 1.38
Aggregate
CV, % 8.3 15.1 4.6 17.4 23.8 3.8 6.3 5.9
X 439.4 0.0030 215028 0.0103 5591 632.7 0.1014 1.44
A
CV, % 3.2 3.6 4.8 5.8 10.2 0.3 10.4 3.2
X 440.5 0.0023 195180 0.0163 6019 647.6 0.1044 1.47
B
CV, % 1.7 10.3 8.1 9.9 4.7 0.8 9.3 1.9
N4
X 446.0 0.0025 189913 0.0298 5293 618.3 0.1042 1.39
C
CV, % 1.9 6.4 5.8 7.1 8.2 1.8 5.3 2.1
X 441.9 0.0026 200040 0.0188 5656 632.9 0.1034 1.43
Aggregate
CV, % 2.3 13.3 8.1 45.7 8.9 2.2 8.0 3.4
X 436.5 0.0037 203994 0.0158 7752 641.5 0.0961 1.47
A
CV, % 1.3 9.3 4.5 6.7 25.2 0.4 10.0 1.3
X 402.2 0.0020 202804 0.0052 4684 656.8 0.1030 1.64
B
CV, % 4.6 2.1 4.0 7.4 5.3 0.1 9.4 4.5
N5
X 447.5 0.0034 206960 0.0263 5318 609.7 0.0975 1.36
C
CV, % 3.4 3.0 6.0 10.7 20.7 2.1 2.4 1.5
X 428.7 0.0030 204586 0.0158 5918 636.0 0.0989 1.49
Aggregate
CV, % 5.6 26.3 4.6 57.6 30.8 3.4 8.1 8.3
X 445.7 0.0033 205091 0.0104 6637 685.5 0.1063 1.54
A
CV, % 0.9 6.3 4.8 2.8 3.9 0.3 4.8 0.8
X 461.8 0.0029 187781 0.0132 5215 671.7 0.0921 1.45
B
CV, % 1.2 5.9 5.6 11.6 8.2 0.3 6.6 1.3
N6
X 445.5 0.0034 208220 0.0274 5399 604.3 0.1072 1.36
C
CV, % 2.6 10.7 5.4 9.0 10.9 1.2 12.8 1.7
X 451.0 0.0032 200364 0.0170 5874 653.8 0.1019 1.45
Aggregate
CV, % 2.4 10.5 6.7 46.2 16.1 5.7 10.9 5.4
X 440.0 0.0024 208514 0.0135 6221 643.7 0.0867 1.46
B
CV, % 1.0 6.2 6.2 8.7 6.6 0.8 28.1 1.2
X 422.6 0.0022 214559 0.0246 4238 569.4 0.0949 1.35
N7 C
CV, % 3.0 3.1 2.8 13.4 19.5 0.8 6.5 3.5
X 431.3 0.0023 211537 0.0191 5229 606.6 0.0908 1.41
Aggregate
CV, % 3.0 7.1 4.8 33.0 23.2 6.5 19.0 4.9
X 447.1 0.0034 199348 0.0123 5692 653.9 0.1003 1.46
A
CV, % 1.0 7.3 4.2 12.1 7.6 0.3 10.5 0.8
X 432.3 0.0025 214708 0.0143 6311 639.4 0.1103 1.48
B
CV, % 1.5 10.2 7.2 13.6 12.2 1.0 8.6 1.7
N8
X 440.4 0.0045 221290 0.0196 5078 602.7 0.1049 1.37
C
CV, % 1.2 4.7 3.3 9.9 12.4 2.0 5.0 1.8
X 440.0 0.0035 211782 0.0154 5694 632.0 0.0985 1.44
Aggregate
CV, % 1.8 23.6 6.3 22.6 13.2 3.6 28.0 3.6
N3 to N8 A-B-C X 443.0 0.0029 205485 0.0167 5664 635.8 0.1026 1.44
aggregate aggregate CV, % 5.2 23.0 6.4 42.5 19.7 4.9 9.3 5.9
* Companies A and B do not fabricate N7 and N3 bars, respectively.

Results in Table 4 shows that P2 for fy is below the minimum limit allowed by the standards, and that
for fsu is above it. The lack of conservatism involved in the former result must be considered by designers, code
developers, and peer reviewers when conceiving RC structures in central Colombia. A yield strength value of
the rebar lower than anticipated directly reduces the yield strength of otherwise compliant structures. This may
10
be critical for RC members with lower reinforcing steel areas or steel areas controlled by minimum
requirements, and large gross cross-section inertia whose cracking moment may surpass the yield capacity,
resulting in an unexpected brittle response at ultimate-loading level demand.

Table 4. Summary statistics of the aggregated data

Parameter X s SE P2 P50 P98
fy (MPa) 443 23 2.6 391 442 508
fsu (MPa) 636 31 3.6 567 642 685
k 1.44 0.085 0.010 1.30 1.45 1.68
y 0.0029 0.0007 0.0001 0.0020 0.0029 0.0045
sh 0.0167 0.0071 0.0008 0.0052 0.0145 0.0304
su 0.1026 0.0096 0.0011 0.0849 0.1000 0.1235
Elongation 0.14 0.020 0.002 0.10 0.14 0.18
Es (MPa) 205485 13239 1489 176710 205230 228037
Esh (MPa) 5664 1118 126 3592 5766 8214

P98
1.0 1.0 1.0
eCDF
0.8 0.8 0.8

0.6 P50 Normal fit 0.6 0.6

Fn (x )
CDF

0.4 0.4 0.4

0.2 0.2 0.2

P2
0.0 0.0 0.0
400 440 480 520 560 580 600 620 640 660 680 1.3 1.4 1.5 1.6 1.7
fyfy(MPa)
[MPa] ffsu [MPa]
su (MPa)
fy/fsu
k

1.0 1.0 1.0

0.8 0.8 0.8

0.6 0.6 0.6

Fn (x )
CDF

0.4 0.4 0.4

0.2 0.2 0.2

0.0 0.0 0.0

0.0020 0.0030 0.0040 0.08 0.09 0.10 0.11 0.12 0.005 0.015 0.025
eyy esu
su
esh
sh

1.0 1.0 1.0

0.8 0.8 0.8

0.6 0.6 0.6

Fn (x )
CDF

0.4 0.4 0.4

0.2 0.2 0.2

0.0 0.0 0.0

0.10 0.12 0.14 0.16 0.18 180000 200000 220000 4000 6000 8000 10000
Elongation
Elongation EEs [MPa]
s (MPa) EEsh(MPa)
[MPa]
sh

Figure 10. Empirical cumulative frequency versus theoretical normal distribution function for different parameters.

Recognizing that a limited number of datapoints prevents a robust estimation of X and CV, bootstrapping
methods [22, 23] allow estimating the distribution of the mean and the standard deviation of the parameters
11
under evaluation. Figs. 11a and 11b show examples of the distribution of X for fy organized by company name
and bar diameter. This type of analysis is repeated for the CV and other variables in consideration. Confidence
intervals (CI) plots are a simple way to find out if the difference between two means is statistically significant.
For example, if the CI overlap, the difference is not significant. The 95% CI of X are presented for fy in Figs.
11c and 11d. Solid vertical lines reference the minimum a maximum limits of the required yield strength in
Colombia per ASTM A706. The vertical dashed line is the mean value of the aggregated data shown in Table
4. Tables 5 and 6 contain the 95% CI for X organized by company name and bar diameter, respectively.
The CI range for company C is greater than the aggregate mean value, while that of company B is lower
than it. The mean of company A coincides with that of the complete dataset. Although the dispersion of the
mean is larger for company C as compared to companies B and A, the overlap of all CI’s indicates that the
difference in the mean yield strength of companies A, B, and C is not be statistically significant. A similar
claim can be made for the fsu data for companies A and B, but not for company C whose CI is below the
aggregate mean, and does not overlap with those of the former (see Table 5). As for the bar diameter data in
Fig. 11d, there is not a clear trend with diameter of the CI width, nor a trend with bias with respect to the mean
of the aggregate. It is noted that production of N3 bars generates the largest dispersion of the yield strength
data, also exhibiting the highest mean value. The reduced conservatism in the mean yield strength data of N5
bars is worth considering as it is the second with largest CI width, with the smallest mean yield strength, less
than the aggregate mean, but specially because of the lower bound of the mean confidence interval falling
below the minimum yield strength limit allowed for Colombia. Hence, there is a chance that the yield strength
capacity of N5 bars produced in Bogotá is limited. The difference between the mean yield strength of the N3
set and that of the N4, N5, N6, and N8 sets is statistically significant, as is that of the N6 set with respect to
the N5, N7 and N8 sets. These same conclusions do not hold for the fsu data, whose mean CI’s overlap for
almost all bar diameter pairs, except for the N7 whose CI width is the largest, below the aggregate mean, with
a significant mean strength difference when compared to the N3 and the N6 sets. Regarding the strain at
maximum capacity, the mean su CI’s overlap among companies and bar diameter; hence their difference is
not statistically significant. Furthermore, except for the N3 set, the lower and upper CI limits among the bar
diameter groups are within 0.6% strain. For the sake of briefness, similar analyses regarding other parameters
are left to the reader, following results contained in Tables 5 and 6.

Table 5. 95% confidence interval of the mean organized by company

Parameter Company A Company B Compay C
fy (MPa) 438-450 427-443 441-460
fsu (MPa) 643-660 647-657 600-622
k 1.45-1.49 1.47-1.53 1.35-1.37
y 0.0029-0.0033 0.0023-0.0025 0.0029-0.0034
sh 0.0112-0.0129 0.0109-0.014 0.0221-0.026
su 0.0889-0.1012 0.0991-0.1069 0.1001-0.1068
Elongation 0.12-0.13 0.14-0.15 0.13-0.15
Es (MPa) 201315-208903 195920-207635 204098-213668
Esh (MPa) 5906-6840 5396-5990 4668-5244

Table 6. 95% confidence interval of the mean organized by bar diameter

Parameter N3 N4 N5 N6 N7 N8
fy (MPa) 450-499 437-447 417-440 446-456 423-438 436-444
fsu (MPa) 640-670 626-640 625-646 635-671 584-629 620-643
k 1.34-1.44 1.41-1.46 1.43-1.55 1.41-1.49 1.36-1.44 1.41-1.46
y 0.0024-0.0029 0.0024-0.0028 0.0026-0.0034 0.003-0.0034 0.0022-0.0024 0.0031-0.0039
sh 0.0115-0.0143 0.0147-0.0232 0.0114-0.0202 0.0134-0.021 0.0154-0.0228 0.0137-0.0172
su 0.0748-0.1104 0.0993-0.1075 0.095-0.1029 0.0967-0.1074 0.095-0.1053 0.097-0.1091
Elongation 0.13-0.15 0.11-0.13 0.13-0.14 0.14-0.15 0.12-0.15 0.13-0.15
Es (MPa) 201485-213330 192205-207859 200142-209368 193717-206877 205319-217112 204939-218492
Esh (MPa) 4597-5878 5375-5877 5105-6885 5378-6130 4501-5922 5329-6092

12
Figure 11. Bootstrapping distribution of the mean and coefficient of variation of fy organized by company (A and C) and
by bar diameter (B and D).

Confidence interval for the coefficient of variation are summarized in Table 7. Following Rustom [18]
homogeneity classification, CV CI’s with high homogeneity are underlined if the lower and upper bound are
less than 5%, those with high heterogeneity are highlighted with bold fonts if at least the upper limit is greater
than 50%, while the rest are categorized as having moderate heterogeneity. It is noted that the fy variance is
small for all companies and the production of N4, N6, N7, and N8 bars. Results for N3 and N5 bars are higher,
although lower than 11% and 7%, respectively. This trend is consistent with the mean results CI width
described before for these two sets. The fsu variance is also small for steel produced by companies A and B,
and for N4, N5, and N8 bar sets. The ultimate-to-yield strength (k) data is highly homogeneous for rebar
produced by companies A and C, and bars in the N4, and N8 sets. As for the highly heterogenous data, it only
comprises sh values for N4, N5 and N6 bars. It is also noted that the homogeneity in fy, fsu and k parameters
for N4 and N8 bars production in Bogotá is high.

Table 7. 95% confidence interval of the CV (%).

Company Company Company
Parameter N3 N4 N5 N6 N7 N8
A B C
fy (MPa) 0.22-0.23 0.23-0.23 0.22-0.23 3.2-10.5 1.7-2.6 3.1-6.9 1.5-3 1.1-3.9 1.2-2.3
fsu (MPa) 2.3-4.1 1.4-2.3 3.4-6.2 1.5-5.4 1.3-2.8 2.1-4.3 3.7-6.4 4.7-6.9 2.4-4.3
k 2.49-4.4 2.41-6.6 1.81-3 2.8-7.7 2.4-4.2 5.3-9.8 3.9-6.3 2.6-6 2.4-4.5
y 11.5-20 11.1-16.7 19.3-28.7 8.6-18.8 9.1-16.1 17.2-31.7 7.1-12.8 3.1-9.2 17.3-29.1
sh 13.7-21.8 20.3-43.5 16.5-26.9 10.2-22.2 34.1-53 37.6-75.3 35.3-51.1 21.5-39.8 16.1-28.1
su 9.7-21.9 7.2-11.8 6.4-10.9 13.4-38.2 5.8-9.5 4.2-10.9 6.6-13.6 2.3-12.5 4.9-15.9
Elongation 9.3-15.5 7.8-12 13.3-19.9 3.6-14 9.9-19.1 5.8-9.9 7.1-14.9 11.5-21.1 10.4-19.6
Es (MPa) 3.4-5.9 5.6-9.1 4.6-7.8 2.5-5.8 5.4-9.8 3-5.8 4.4-8.4 2-6.3 4.7-7.9
Esh (MPa) 9.1-26.1 9.6-17 12.4-19.3 12.1-23.6 5.6-11.3 15.1-38.9 9.9-15.4 12.6-29 7.1-18.5

13
4.4 Comparison of results with other authors and countries
Table 8 summarizes the results by bar diameter of studies carried out in Colombia and other countries. The
results by Bournonville et al. [13] only include the ASTM A706-G60 steel. Fig. 12 compares the results for fy,
fsu and su of this study and those in Table 8 for the US and Mexico, organized by bar diameter. For the US
data, except for the N3 data, the results show an apparent consistency (i.e. less variation) of the mean variation
among the different bar diameters for fy and fsu. The mean yield strength and maximum strength of the
Colombian tests are the smallest among all countries. The US data exhibits the largest average yield strength
and the Mexico data does so for the maximum tensile strength. The CV of the fy and fsu data of the US hover
around 5%, which is indicative of its high homogeneity and consistency among the bar diameters. For
Colombia, CVs are smaller than 6% in most cases, except for fy of the N3 set, and fsu for the N7 set. The Mexico
data for fsu exhibits the highest homogeneity, while the results for fy are comparable to those of Colombia,
specifically, the yield strength of N3 bars in both countries does not seem well controlled. Regarding the strain
at maximum strength data, the average strain capacity of the Mexican bars is 25% larger than that of the
Colombian set, and comparable to those of the US. However, the Colombia set shows more consistency in
strain capacity among bar diameters. As expected for this type of measurement, the strain data exhibits, in
general, moderate homogeneity, which, for Colombia, reduces with increasing bar diameter.

Table 8. Results of previous studies in Colombia and other countries

Statistical fy Es Esh fsu Elon.
Author Country Diameter y sh su fsu/fy
parameter (MPa) (Mpa) (Mpa) (Mpa) %
X 471.1 - - 0.0076 - 752.5 0.14 - -
N3
CV, % 8.7 - - 37.5 - 3.0 24.8 - -
X 437.0 - - 0.0082 - 719.6 0.15 - -
N4
CV, % 7.9 - - 30.6 - 4.8 8.4 - -
Rodríguez
X 472.2 - - 0.0098 - 746.8 0.11 - -
and Botero Mexico N5
CV, % 1.4 - - 9.2 - 2.3 6.7 - -
[10]
X 456.6 - 0.0088 - 732.1 0.12 - -
N6
CV, % 3.7 - - 28.7 - 1.5 12.4 - -
X 447.1 - - 0.0072 - 741.1 0.11 - -
N8
CV, % 4.8 - - 34.7 - 2.3 8.2 - -
X 510.4 - - - - 685.6 - - 14.8
N3
CV, % 4.2 - - - - 4.4 - - 0.9
X 471.9 - - - - 648.2 - - 15.4
N4
CV, % 5.4 - - - - 5.7 - - 1.4
X 472.9 - - - - 647.9 - - 15.5
Bournonville N5
CV, % 5.4 - - - - 5.4 - - 1.6
et al. US
X 477.7 - - - - 655.9 - - 15.2
[13] N6
CV, % 4.7 - - - - 5.0 - - 1.9
X 480.0 - - - - 664.1 - - 15.2
N7
CV, % 4.8 - - - - 4.6 - - 2.3
X 475.0 - - - - 659.6 - - 15.6
N8
CV, % 4.8 - - - - 4.6 - - 2.0
Tavio et al.
India NE* X 494.6 - - - - 636.6 1.29 -
[9]
Mander and X 482.7 - - 0.013 6895 655.0 0.13 1.36 16.0
Matamoros US NE*
CV, % 5.0 - - 39.0 21.0 5.0 18.0 4.0 13.7
[8 ]
N3 X 470.3 0.0024 215600 0.013 - 659.7 0.11 - -
Gonzales et
Colombia  -
al. [6] CV, % 5.4 5.4 34.8 - 4.6 10.6 - -
N8
Cabrera and
Colombia N4 X 471.0 - - - - 633.0 - 1.35 15.0
Koller [5]
Arévalo and
Colombia NE* X - - - - - 558.9 - 17.3
Arévalo [7]
* The bar diameter is not reported.

González et al. [6] reported that fsu of steel bars in Colombia with diameter ranging between N3 and N8
was 1.25 times greater than fy. The mean value of fsu/fy reported by Cabrera and Koller [5] was 1.35 for N4
bars in Colombia, that reported by Mander and Matamoros [8] was 1.36 for steel bars in the US, and that
reported by Tavio et al. [9] was 1.29 for steel bars in India. When comparing with the other studies analyzed,

14
the highest mean value of fsu/fy is found in the present study, that is, mean k = 1.44. Regarding Es, most building
codes assume Es = 200,000 MPa. However, the modulus of elasticity reported by Rodríguez and Botero [10]
was 215,600 MPa for Mexico City. Table 3 shows that, the mean value of Es for the aggregate data reported
herein is 205,485 MPa, exceeding the value assumed by NSR-10 Colombian code [2] and ACI-318 [3].
However, mean results for test sets B-N4, C-N4, B-N6, and A-N8 are below the reference modulus between
2.4 and 6.1%. In terms of yield strain, the mean value of εy reported by González et al. [6] was 0.0024 for bars
in Colombia, while the values measured herein vary between 0.0022 and 0.0045, with an aggregate CV
describing the data as moderately homogeneous. This result steam from the CV found in the present study for
the N5 and N8 bars (26.3% and 23.6%, respectively).
Higher values of εsh are related to the extended of a well-defined yield plateau. The minimum and
maximum values of εsh found herein, aggregated by diameter, are 0.013 and 0.019, respectively, while the
mean value of the aggregated data is 0.017. Table 3 shows that the CV of εsh is the highest among all variables
reported herein, and is associated to the disparity observed in the stress-strain curves in Figs. 2 to 7, where
some bar diameters have a well-defined plateau while for others there is a lack of it. This indicates that the
chemical composition of the bars differs from one company to another and among bar diameters. Rodríguez
and Botero [10] observed that the yield plateau of small-diameter bars (N3, N4 and N5) in Mexico was not
well-defined; hence, they assumed that εy = εsh for these cases. For other bars, the εsh reported by the authors
varies between 0.0072 and 0.0098. Mander and Matamoros [8] reported a mean value of εsh = 0.013 for steel
bars in the US, which is equal to the mean value reported by González et al. [6] for bars in Colombia (see
Table 8).
In terms of Esh which characterizes the initial slope of the strain-hardening branch (see Fig. 1), the values
obtained in the present study varied between 4500 MPa and 8500 MPa, with mean value equal to 5664 MPa
and CV of 13.2%. The CV of Esh in the present study is indicative of moderate homogeneity. Mander and
Matamoros [8] found that the CV and mean value of Esh for bars in the US is 21% and 6895 MPa, respectively,
being the mean value 21.7% higher than that found in this study.
(a) Country / db (b) Country / db
Colombia Mexico US Colombia Mexico US
550 540 10

500
Ave. 5
5

Ave.
450 Ave.

420
0
N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE

(c) Country / db (d) Country / db

Colombia Mexico US Colombia Mexico US
Ave. 10

8
700
Ave. 6
5
Ave. 4
600
2
550
0
N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE

(e) Country / db (f) Country / db

Colombia Mexico US Colombia Mexico US
30
0.15

Ave. Ave.
20
20
Ave.
0.10
10
5

0.05 0
N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE N3 N4 N5 N6 N7 N8 NE

Author This study S2018 CK2017 RB1995 B2004 MM2019

Figure 12. Comparison of mean and coefficient of variation for fy, fsu and su organized by bar diameter and country.

15
4.5 Characteristics stress-strain curves
Characteristic values of mechanical parameters for computing monotonic tensile stress-strain curves of bars in
Bogotá are proposed in this section. Rodríguez and Botero [10] argue that the 5th percentile (P5) is commonly
used as an acceptance criterion for fy and is specified in different European codes. Rodríguez and Botero [10]
also state that the maximum value of fy (e.g. 540 MPa per ASTM A706) must be taken into account in
earthquake-resistant engineering, because of the additional flexural overstrength in reinforced concrete
elements. Unexpected flexural over-strength may exacerbate additional forces on members whose design did
not anticipate it. For example, shear failures may be triggered in beams over-strengthened in flexure when
subjected to seismic shaking that generate the development of probable moments at the ends.
The mechanical parameters used to construct analytical stress-strain curves based on the results of this
study are associated to the 5th (P5) and 95th (P95) percentiles. The P95 level is proposed as an acceptance
parameter of the maximum yield strength defined in ASTM A706 [14]. Table 9 summarizes the P5 and P95
for the mechanical properties aggregated by bar diameter in Bogotá. The elastic zone (i.e., for 0 ≤ εs ≤ εy) is
defined by Hooke's law in Eq. (1), which indicates that the tensile strength is proportional to the strain through
the modulus of elasticity in Table 3. In the yield plateau (y < s ≤ sh), the strength remains constant, equal to
fy as shown in in Eq. (1). The model proposed by Mander [24] in Eq. (2) is used to determine the strain
hardening branch of the tensile stress-strain curve. Mander proposes two methods to calculate the value of M.
The first method uses Eq. (3), which depends on the basic mechanical parameters described through the article,
while the second method, shown in Eq. (4), adds two transitional parameters fsh1 and εsh1. It is noted that the
value of M is fundamental for computing the stress-strain curve since it characterizes the shape of the strain
hardening branch.

𝐸  𝑖𝑓 0 𝜀 𝜀
𝑓 (1)
𝑓 𝑖𝑓 𝜀 𝜀𝑠 𝜀

 
𝜎 𝑓 𝑓 𝑓 (2)
 

 
𝑀 𝐸 (3)

𝑀   (4)
 

Table 9. Values associated with percentiles P5 and P95 for different mechanical parameters
Nominal fy fsu
Percentile εy εsh εsu M
Diameter (MPa) (MPa)
P5 412.1 0.0022 0.010 617.5 0.059 1.87
N3
P95 513.3 0.0032 0.016 681.5 0.119 2.63
P5 428.0 0.0021 0.010 607.1 0.093 1.95
N4
P95 454.1 0.0031 0.031 652.0 0.116 3.42
P5 387.0 0.0020 0.005 601.8 0.089 1.66
N5
P95 458.9 0.0039 0.029 657.2 0.113 3.81
P5 435.8 0.0027 0.010 599.2 0.087 1.84
N6
P95 466.6 0.0037 0.029 686.5 0.121 3.44
P5 410.9 0.0021 0.012 565.2 0.066 1.72
N7
P95 444.1 0.0026 0.028 648.4 0.113 2.95
P5 428.3 0.0023 0.011 595.3 0.082 2.17
N8
P95 450.4 0.0046 0.021 654.8 0.117 3.18

Rodríguez and Botero [4], Mander [24] and Gonzales et al. [6] explain that parameters fsh1 and εsh1 describe
intermediate experimental points in the strain hardening branch. A well-defined and practical criterion is not
proposed for obtaining these values. The first method is preferred in this study to determine M by means of

16
Eq. (3). This value is reported in Table 9 at the P5 and P95 levels. The characteristic analytical stress-strain
curves of steel bars marketed in central Colombia are shown in Fig. 13, including P5 and P95 levels, showing
good agreement with the experimental data.

Figure 13. Characteristic stress-strain curves for bars in central Colombia for different diameters: (a) N3, (b) N4, (c) N5,
(d) N6, (e) N7, (f) N8.
5 Conclusions

This article presented and discussed the results of a statistical analysis of the tensile mechanical properties that
characterize the stress-strain curve of steel bars marketed in Bogotá. On average, the bars marketed in Bogotá
comply with the standards mandatory for Colombia. The monotonic tensile tests allowed obtaining the stress-
strain curves of the bars, from which the corresponding statistical analysis of the mechanical parameters of N3
to N8 bars followed. It was observed that the Normal distribution fits suitably most of the results measured.
The statistical analysis presented in Table 3 showed that the mean value of yield strength (fy) of almost all bars
for the different companies is greater than 420 MPa and lower than 540 MPa, which are the minimum and
maximum values required by ASTM A706. Only the N5 bars of company B did not comply with the minimum
required fy with a recorded mean of 402 MPa. The highest recorded value of fy is below the maximum permitted,
with the mean of 509 MPa for the N3 bars of company C.
The average modulus of elasticity for all diameters was 205,484 MPa, although test sets B-N4, C-N4,
B-N6, and A-N8 had mean Es below the 200,000 MPa assumed by most building codes worldwide. The
elongation percentage was the parameter that partially comply with ASTM A706. The percentage elongation
values of the N3, N4 and N5 bars were lower than the values established by ASTM A706. The minimum
elongation value required in ASTM A706 for diameter N3, N4 and N5 was 14%, while the average values
reported herein were 13.7%, 12.1% and 13.8%, respectively. The minimum elongation percentage required by
ASTM A706 for the N6, N7 and N8 bars is 12%; reported values in this study exceeded this value.
The paper also proposed characteristic values of the mechanical properties needed for defining analytical
monotonic stress-strain curves of steel bars marketed in Bogotá for diameters from N3 to N8. The proposed
curves reflected good agreement between the experimental and numerical values, and may be used to study
the seismic behavior of RC structures in central Colombia.

Acknowledgments
The authors recognize the Vicerrectoría de Investigaciones at the Universidad Militar Nueva Granada
(UMNG) for financing the Grant IMP-ING-2933. The authors also would like to thank undergraduate student
Verónica Abuchar and graduate student Jefferson Piedrahita at Universidad del Norte for data and figure
processing. The authors also acknowledge to USI Library of Institute of Engineering at UNAM for providing
some of the references. The information contained in this article represents exclusively the opinion of the
authors and does not necessarily represent the opinions of the sponsor.

17
 References
[1] Toro C. (2017). The steel arrives to Colombia without technical specifications, En Obra, May. (in Spanish)
[2] NSR-10 (2010). Colombian Code for Earthquake Resistant Construction. Colombian Association for Earthquake
Engineering, AIS, Colombia. (in Spanish)
[3] ACI Committee 318 (2008). Building code requirements for structural concrete and commentary (ACI 318).
American Concrete Institute, ACI. Farmington Hills, MI.
[4] Muñoz H. (2012). Manual of Gerdau-Diaco Steel for earthquake-resistant constructions, 3rd Ed. Gerdau-Diaco,
Bogotá. (in Spanish)
[5] Cabrera A. and Koller M. (2017). Tensile behavior of structural Steel per NTC 2289 standard from different
reinforcing steel producers. Bachelor Thesis, Universidad Católica de Colombia, Bogotá, Colombia. (in Spanish)
[6] González V., Botero J., Rochel R., Vidal J. and Álvarez M. (2005). Mechanical properties of steel used in Colombia.
Ingeniería y Ciencia, 1(1):67-76. (in Spanish)
[7] Arévalo J. and Arévalo Jh. (2018). Analysis of mechanical and metallographic properties of deformed bars
commercialized by different hardware stores in Ocaña, Norte de Santander. International Conference on
Engineering, INGENIO 2018, 382-390.
[8] Mander T. and Matamoros A. (2019). Constitutive modeling and over-strength factors for reinforcing steel. ACI
Structural Journal, 116(67):219-231, doi: 10.14359/51713320.
[9] Tavio, Anggraini R., Raka P. and Agustiar (2017) Tensile strength/yield strength (TS/YS) ratios of high strength
steel (HSS) reinforcing bars, International Seminar on Metallurgy and Materials. Surabaya. Indonesia.
[10] Rodríguez M. and Botero J. (1995). Behavior of reinforcing bars subjected to monotonic and reversed-cyclic loading
including buckling. Institute of Engineering, National University of Mexico, UNAM. (in Spanish)
[11] Carrillo J., Díaz C. and Arteta C. (2019). Tensile mechanical properties of the electro-welded wire meshes available
in Bogotá, Colombia. Construction and Building Materials, 195:352-362, doi: 10.1016/j.conbuildmat.2018.11.096
[12] Villanueva J. and Meli R. (1984). Statistical study of properties of materials and structures in Mexico City. Institute
of Engineering, National University of Mexico, UNAM. (in Spanish)
[13] Bournonville M., Dahnke J. and Darwin D. (2004). Statistical analysis of the mechanical properties and weight of
reinforcing bars. Research Report, University of Kansas.
[14] ASTM A706 (2008). Standard Specification for low-alloy steel deformed and plain bars for concrete reinforcement.
American Society of Testing Materials, ASTM. West Conshohocken, PA, 6 pp.
[15] Overby D., Kowalsky M. and Seracino R. (2017). Stress-strain response of A706 grade 80 reinforcing steel.
Construction and Building Materials, 145: 292-302, doi: 10.1016/j.conbuildmat.2017.03.200.
[16] Rodríguez M. and Botero J. (1996). Aspects of seismic behavior of reinforced concrete structures considering the
mechanical properties of reinforcing bars produced in Mexico. Institute of Engineering, National University of
Mexico, UNAM. (in Spanish)
[17] ASTM A370 (2019). Standard test methods and definitions for mechanical testing of steel products. American
Society of Testing Materials, ASTM. West Conshohocken, PA, 46 pp.
[18] Rustom A. (2012). Descriptive statistics, probability and inference: a conceptual and applicated vision. 1st Ed.,
Universidad de Chile. (in Spanish)
[19] Lilliefors H. (1967). On the kolmogorov-smirnov test for normality with mean and variance unknown. American
Statistical Association, 62(318), 399-402, doi:10.1080/01621459.1967.10482916.
[20] Anderson T. and Darling D. (1952). Asymptotic theory of certain "goodness of fit" criteria based on stochastic
processes. Annals of Mathematical Statistics, 23(2), 193-212, doi: 10.1214/aoms/1177729437.
[21] Massey F. (1951). The kolmogorov-smirnov test for goodness of fit. American Statistical Association, 46(253), 68-
78. doi:10.1080/01621459.1951.10500769.
[22] Efron B. (1979). Bootstrap methods: another look at the jackknife. Annals of Statistics, 7(1), 1-26.
[23] Davison A. and Hinkley D. (1997). Bootstrap methods and their application. Cambridge University Press,
Cambridge, United Kingdom.
[24] Mander J. (1984) Seismic design of bridge piers. University of Canterbury, Canterbury, New Zealand.

18

View publication stats