Construction and Building Materials 189 (2018) 1076–1092

Contents lists available at ScienceDirect

Construction and Building Materials

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

Mechanical properties of high strength steel strand at low temperatures:

Tests and analysis
Jian Xie, Xueqi Zhao, Jia-Bao Yan ⇑
School of Civil Engineering, Tianjin University, Tianjin 300350, China
Key Laboratory of Coast Civil Structure Safety of Ministry of Education, Tianjin University, Ministry of Education, Tianjin 300350, China

h i g h l i g h t s g r a p h i c a l a b s t r a c t

Low temperature of 20 to 160 °C

increases strength but reduces
ductility of steel wires.

Proposed design equations predicts
well strengths of wires at low
temperatures.

Theoretical models predict well
nonlinear strength behaviours of steel
strands.

FE models simulate well stress-strain
behaviours of strands at low
temperatures.

Proposed empirical equations offer
means to determine strengths of
strands at low temp.

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

Article history: This manuscript studied mechanical properties of steel strands for prestressed concrete (PC) structures at
Received 8 March 2018 different low temperatures ranging from 20 °C to 160 °C. 21 tensile tests were performed to obtain the
Received in revised form 6 September 2018 stress-strain curves of steel wires in strand at different low temperatures. Empirical prediction formulae
Accepted 10 September 2018
were developed to incorporate the influences of low temperatures on mechanical properties of steel
Available online 20 September 2018
wires. Based on the stress-strain curves of single wire, theoretical and numerical models were developed
to predict the mechanical properties of multi-layer steel strands. The accuracies of these theoretical and
Keywords:
numerical models were validated by the test results. Based on the test and analysis results, empirical
Low temperatures
Steel strands
models were developed to predict the mechanical properties of multi-layer steel strands at different
Prestress concrete structures low temperatures including the elastic modulus, yield and ultimate strengths. This offers useful means
Concrete structures to calculate the mechanical properties of steel strands at low temperatures since their properties varied
Steel wires with their geometry and layout of steel wires. Finally, recommended prediction procedures are given to
Mechanical properties determine the mechanical properties of steel strands at different low temperatures.
Elastic-plastic behaviours Ó 2018 Elsevier Ltd. All rights reserved.
Finite element method

1. Introduction structures and bridges [1] as shown in Fig. 1. Due to their excellent
mechanical properties, steel strands were also used in engineering
High strength steel strands are widely used in civil engineering constructions in harsh environments with low temperatures, e.g.,
constructions, e.g., prestressed concrete (PC) structures, long-span infrastructures in cold regions, the Arctic onshore and offshore
platforms, liquefied natural gas (LNG) containers. In northern
⇑ Corresponding author. China and Tibet, the recorded lowest temperature was 53.4 °C
E-mail address: ceeyanj@163.com (J.-B. Yan). [2]. The lowest temperature in the Arctic could drop to about

https://doi.org/10.1016/j.conbuildmat.2018.09.053
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092 1077

Nomenclature

A0 cross sectional area of core wire fy, fu yield and ultimate strength of steel wire
Ai cross sectional area of helical wires of layer i f ys , f us yield and ultimate strength of steel stand
E, E1 elastic and plastic modulus of steel wire f ya ,f ua yield and ultimate strength of steel wire at ambient
Es elastic modulus of steel strand temperature
Fi total axial force of helical wires of layer i L length of core wire
F total axial force of steel strand mi number of helical wires of layer i
Hi twisting moment in tangential direction of helical wires r0 radius of core wire
of layer i ri radius of helical wires of layer i
IE factor for elastic modulus si length of helical wires length on the centreline of layer i
If u factor for ultimate strength ai lay angle of helical wires of layer i
If y factor for yield strength e0 axial strain of core wire
G0i bending moment in binormal direction of helical wires ei axial strain of helical wires of layer i
of layer i ey , eu , eF yielding, ultimate and fracture strain of steel wire
N number of steel wires in a strand w cross-sectional area of steel wire
N0i force in binormal direction of helical wires of layer i ui polar angle of helical wires of layer i
Pi pitch length helical wires of layer i j0i curvature in the binormal direction of helical wires of
Ri helix radius of steel strand layer i
T different temperature level ji curvature in the normal direction of helical wires of
T0 ambient temperature layer i
Ti orce in tangential direction of helical wires of layer i ci torsional strain of helical wires of layer i
Xi line load per unit length in normal direction of helical vi twists of helical wires of layer i
wires of layer i m Poisson’s ratio of steel wire

70 °C [3,4]. In the scenario of leakage of LNG, the external con-
crete structure of LNG containers may suffer low temperature of
about 165 °C [5–6]. Since these infrastructures suffer low tem-
peratures produced by these harsh environments, the mechanical
properties of steel strands used in these structures at low temper-
atures need to be carefully considered for the evaluation on their
structural performances.
There is extensive reported research on mechanical properties
of steel materials at low temperatures. Elices et al. [7] reported ten-
sile tests on hot rolled steel reinforcements at 20 °C, 80 °C and
180 °C. It showed that as the temperature decreased, both yield
and ultimate strengths of steel reinforcements increased but their
ductility was slightly affected. Lahlou et al. [8] studied the mechan-
ical properties of mild steel at ambient temperature and low tem-
perature 195 °C. The test results showed that strengths and
elastic modulus increased but ductility significantly decreased at
low temperatures. Yan et al. [9] carried out tensile tests on mild
and high strength steel plates within temperature ranges of

Prestressed concrete
beam Prestressed concrete
bridge
Strand

Helical wire

Helical wire
Strand

Core wire
Fig. 1. The application of a multi-layer strand.
1078 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092
+30 °C to 80 °C. They found that fracture strain, yield and ulti-
mate strengths were all increased at low temperatures. Filiatrault
et al. [10] reported the mechanical properties of steel reinforce-
ments at temperatures varying from +20 °C to 40 °C. They
observed that strengths of steel reinforcements increased at low
temperatures. Planas et al. [11] reported the tensile tests on steel
strands at 20 °C and 195 °C and test data showed that strengths
and elongations of steel strands were both increased at 195 °C.
These above experimental results show that most of previous stud-
ies focused on mild and high strength steel plates. There is still lim-
ited information on the high strength steel strands at low
temperatures. Thus, as the key components in the prestressed con-
crete structures, it is necessary to investigate the mechanical prop-
erties of steel strands at low temperatures that will provide useful
information on the analysis and design of infrastructures especially
the prestressed concrete structures built in cold regions.
Multi-layer steel strands typically consist of a straight core wire
surrounded by several layers of helical wires in a symmetrical way
as shown in Fig. 1. Different from the reinforcement or steel plate,
mechanical properties of steel strands are more complex due to
their geometric patterns and deforming mechanisms. Several geo-
metric and analytical models have been developed to study the
mechanical properties of strands based on Love’s curved beam the-
ory [12]. Machida and Durelli [13] reported the stiffness matrix of
steel strands that included the influences of bending and torsion
stiffness of wires. Knapp [14] proposed prediction equations and
investigated the deformation of core wire. Costello et al. [15,16]
presented linearized equilibrium equations that considered the
influence of Poisson’s ratio, curvature and lay angle variations.
Kumar et al. [17] extended Costello’s theory and offered a closed-
form solution for axial-torsional stiffness matrix of strands. These
developed models were based on the assumption of linear-
elastic-constitutive law of the material, and the plastic deforma-
tion in wires were neglected. Several researchers used finite ele-
ment methods to study the global response of strands that
considered plastic deformation of wires. Jiang et al. [18,19] devel-
oped a finite element model to predict the global behaviours of
strands under axial tension. Yu et al. [20] proposed a finite element
model to study behaviours of seven-wire strands under axial ten-
sion that incorporate the effect of friction on the stiffness of
strands. Abdullah et al. [21] proposed a full-scale finite element
model to observe the breakage response of seven-wire strands that
considered contact and frictions among the steel wires. Judge et al.
[22] presented a 3D elastic-plastic finite element model for multi-
layer spiral strands to simulate their global responses and local
deformations. Recently, an analytical model was developed by Foti
et al. [23] to predict the elastic–plastic behaviour of seven-wire
strands under axial torsion. It can be concluded from these previ-
ous studies that most of previous analytical models focused on
the elastic behaviours of the strands, and there are still few pro-
posed analytical models that could predict their plastic behaviours.
Due to the limited computational efficiency, some of the FE models
with complex contacts were not suitable for large-scale structural
analyses. Moreover, few FE models considered the stress redistri-
bution and contact among different wires in steel strands. Thus,
it is necessary to develop accurate theoretical and numerical mod-
els to predict global responses of multi-layer strands that consid-
ered both geometric and material nonlinearities as well as
contact algorithms among different steel wires in the strands.
Multi-wire strands are made of several twisted wires. Stress-
strain curves of strands are quite different from steel wires due
to complex geometric patterns. Under tension, the nominal
stress-strain behaviour of the multi-wire strand behaves differ-
ently from that of single wire due to stress redistribution and inter-
actions among the wires in strand. And the tensile stress-strain
behaviours of steel strands vary with the layout of the wires in
them. Moreover, tensile tests on steel strands are more difficult
to perform compared with that on steel wires especially at various
low temperatures. Thus, this manuscript aimed to develop predic-
tion methods on the mechanical properties of multi-layer strands
at different low temperatures based on tensile tests of steel wires
at ambient temperature.
This manuscript reported a study on mechanical properties of
steel strands at different low temperatures by experimental, theo-
retical and numerical methods. To achieve this objective, tensile
tests were firstly carried out on steel wires at different tempera-
tures ranging from +20 °C to 160 °C to obtain their mechanical
properties at low temperatures. Based on the test results of steel
wires, empirical prediction formulae were then developed to pre-
dict the yield and ultimate strengths of steel wires at low temper-
atures. In order to predict the mechanical properties of multi-layer
steel strands, accurate theoretical and numerical models were pro-
posed. Based on theoretical and numerical analysis results, regres-
sion analyses were finally carried out to develop empirical models
to predict the tensile stress-strain behaviours of the multi-layer
strands. Thus, with the given mechanical properties of steel wires
at ambient temperatures, the tensile behaviours of the complex
multi-layer steel strands at ambient and low temperatures can
be determined by these developed models.
2. Tests on high strength steel wires at low temperatures
2.1. Specimens
Twenty-one high strength steel wire coupons in total were pre-
pared for tensile tests at seven different temperature levels of
+20 °C, 40 °C, 70 °C, 100 °C, 120 °C, 140 °C, 160 °C. Three
identical steel wires were prepared for each temperature level. All
the high strength steel wires measured 5.2 mm in diameter and
250 mm in length, and they were cut from central core wires in
seven-wire strands. The chemical compositions of these high
strength steel wires are listed in Table 1. The two ends of each steel
wire were enlarged to facilitate the anchorage of the tested speci-
mens to the testing frame. Fig. 2 shows a representative steel wire
coupon for tensile tests.
2.2. Test setup and loading procedures
A 100-ton loading machine working with a cooling chamber
was used for the tensile tests on steel wires at low temperatures
as shown in Fig. 2. Designed target temperatures for high strength
steel wires included +20 °C, 40 °C, 70 °C, 100 °C, 120 °C,
140 °C and 160 °C. To simulate the low temperature, a cooling
chamber with an automatic control system was used. The temper-
ature in the cooling chamber can be lower to 190 °C. Cooling rate,
stabilization duration and target temperature can be pre-set in
integrated system running on a controlling PC. During tensile tests,
liquid nitrogen was sprayed into the cooling chamber through an
electromagnetic valve to low down the temperatures of testing
specimens to the target temperatures. Several thermocouples were
placed at different locations in the chamber to monitor the envi-
ronmental temperatures. Meanwhile, thermocouples worked
together with electromagnetic valve to maintain the target tem-
perature by controlling the injection velocity of liquid nitrogen.
Insulation measures were taken at the conjunction between the
testing machine and cooling chamber to reduce heat loses.
Two ends of steel wires were installed on the testing machine
through two clamps. The bottom clamp was fixed to the solid bot-
tom frame whilst the top clamp moved upwards that transferred
the tensile force to the steel wires. Three thermocouples were
directly attached to the specimens at different locations to monitor
 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092 1079

Table 1
Chemical composition of the steel wire.

C (%) Si (%) Mn (%) P (%) S (%) Cr (%) Ni (%) Cu (%) V (%)

0.801 0.227 0.75 0.009 0.006 0.248 0.014 0.012 0.003

100ton MTS
Cooling chamber Loading system

Frame
Tensile force

Detail A Liquid Nitrogen

250mm
Steel wire
Coupon

PT100
Extensometer Thermolcouples

Data acquisition
system

(a) Detail A (b) Test Setup

Fig. 2. Tensile test setup of steel wires at low temperatures.

the temperatures and achieve a balanced cooling effect. Once the 900
target temperature in the cooling chamber was achieved, it should fy
be maintained for 15 min and temperature variations should be
controlled within 1 °C to ensure that all tests were conducted
Stress (MPa)

600
under steady state condition. After that, direct tensile tests were
conducted. The loading rate was determined according to ASTM:
A370-13 [24] for low temperature tensile tests. Two strain gauges
were attached to the middle region of high strength steel wires to 300 E
record their strains. An extensometer was also used to record the
1
plastic strains after the strain gauges spoiled. The reaction forces,
displacement, temperature, strain gauges and extensometer were 0
recorded through a data acquisition system during the tests. 0 0.2 0.4 0.6 0.8
Strain (%)
2.3. Test results
Fig. 3. Determination of the elastic modulus and yield strength.
Test results presented herein included tensile stress-strain
curves at different low temperature levels, yield strength, ultimate LF  L
strength, elastic modulus, yielding strain, fracture strain, and eF ¼ ð1cÞ
L
reduction in cross-sectional area. Since the stress-strain curves of
where L denotes the original length of the measured region by
high strength steel wires have no transparent yielding plateaus,
extensometer; Ly denotes the length of the measured region at the
0.2% offset method was used to determine the yield strength and
yield; Lu denotes the length of the measured region at the ultimate;
elastic modulus according to ASTM: A370-13 [24] as shown in
LF denotes the length of the measured region at the fracture.
Fig. 3.
The indexes of yielding strain ey , ultimate strain eu , fracture A  AF
w¼ ð2Þ
strain eF and reduction in the cross-sectional area w, were used A
to describe the plasticity of steel that are defined as the following: where A denotes the area of the cross section before test; AF denotes
Ly  L the area of the cross section at the fracture point.
ey ¼ ð1aÞ
L
2.3.1. Stress-strain curves
Lu  L Fig. 4 shows the stress-strain curves of representative tests at
eu ¼ ð1bÞ
L seven low temperatures of +20 °C, 40 °C, 70 °C, 100 °C,
1080 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092

2500 Fig. 5(a) also shows that the E exhibits low correlation ratio of
0.33 with the low temperatures that implies the effect of T on E
2000 is quite limited and can be neglected.

1500 2.3.3. Yield strength and ultimate strength

MPa

20 -40 Yield strength fy and ultimate strength fu of the high strength

1000
-70 -100 steel wires can be determined from the corresponding stress–
Stress

strain curves by the methods as specified in ASTM: A370-13 [24].

-120
500
-160
0 strengths at various low temperatures are listed in Table 2. Fig. 5
0 1 2 3
Strain %
Fig. 4. Stress-strain curves of steel wires at different low temperatures.
120 °C, 140 °C and 160 °C. For high strength steel wires, the
stress increases linearly until it reaches yield strength. It can be
observed that there are no yielding plateaus in the stress-strain
curves. Beyond yielding point, the decreasing stiffness of speci-
mens resulted in the decrease of the slope of stress-strain curves.
After achieving ultimate strength, stress-strain curves show obvi-
ous decrease in stress accompanied with necking taking place in
the middle region of the specimen. Finally, fractures occurred to
all the tested specimens. Fig. 4 also shows that both yield and ulti-
mate strengths are increased as the temperature decreases.
2.3.2. Elastic modulus
The elastic modulus of high strength steel wires is determined
from the initial elastic portion of the stress-strain curves. Table 2
lists the elastic modulus of high strength steel wires at different
low temperatures. Fig. 5 (a) shows the influence of the low temper-
ature T on the elastic modulus E. It shows that the elastic modulus
of high strength steel wires increases slightly with the decreasing
temperature. As the temperature decreases from +20 °C to
160 °C, the average E value is only increased by 5%. Moreover,
-140
‘‘0.2% strain offset” method is used to determine the yield strength
as shown in Fig. 3. The determined values of yield and ultimate
4 5 (b) shows the effects of low temperature T on yield strength fy
and ultimate strength fu. They show that as the temperature
decreases from +20 °C to 160 °C, both yield and ultimate
strengths increase with the decreasing temperature, and the
increasing rate of yield strength is smaller than that of ultimate
strength. As the temperature decreases from +20 °C to 40 °C,
70 °C, 100 °C, 120 °C, 140 °C and 160 °C, the yield strength
is averagely increased by 6%, 12%, 13%, 14%, 16% and 18%, respec-
tively; and the ultimate strength is averagely increased by 6%,
12%, 14%, 16%, 18% and 20%, respectively. The correlation coeffi-
cients of yield and ultimate strengths with temperature are 0.97
and 0.96, respectively, which are much higher than that of the elas-
tic modulus. This implies both fy and fu of steel wires are more sig-
nificantly influenced by and they are more sensitive to low
temperatures. Thus, the influences of the low temperatures on fy
and fu should be carefully considered.
2.3.4. Yielding strain and ultimate strain
The measured yielding strain ey and ultimate strain eu are listed
in Table 2. Fig. 5 (c) plots the influences of the temperature T on
yielding strain ey and ultimate strain eu . These table and figure
show that both ey and eu increase as the temperature decreases
from +20 °C to 100 °C. As T decreases from +20 °C to 40 °C,
70 °C, 100 °C, the values of ey (oreu ) are averagely increased by
6% (3%), 12% (7%), and 15% (15%), respectively. At the temperature

Table 2
Details and tensile test results of the steel wires.

Item T (°C) E (GPa) E (GPa) fy (MPa) fy (MPa) fu (MPa) fu (MPa) ey (%) ey (%) eu (%) eu (%) eF (%) eF (%) w (%) w (%)
A-1 20 211.7 206.5 1799.4 1808.5 1937.1 1947.2 1.03 1.03 3.41 3.54 3.78 3.83 0.20 0.19
A-2 20 209.0 1817.9 1961.8 1.03 3.64 3.76 0.19
A-3 20 198.8 1808.2 1942.6 1.04 3.58 3.94 0.18
B-1 40 212.3 206.3 1931.3 1917.9 2100.3 2069.9 1.08 1.10 3.57 3.63 3.79 3.87 0.20 0.20
B-2 40 202.4 1903.3 2080.5 1.13 3.80 4.02 0.21
B-3 40 204.1 1919.2 2028.8 1.08 3.53 3.79 0.18
C-1 70 208.6 205.0 2035.0 2017.1 2200.1 2186.1 1.15 1.15 3.73 3.78 3.92 3.97 0.17 0.17
C-2 70 199.8 2003.4 2168.0 1.14 3.75 3.92 0.19
C-3 70 206.5 2012.9 2190.2 1.17 3.85 4.08 0.15
D-1 100 205.7 208.1 2043.8 2036.9 2218.0 2221.3 1.21 1.19 4.03 4.07 4.28 4.36 0.21 0.20
D-2 100 207.9 2039.0 2223.0 1.18 4.00 4.25 0.19
D-3 100 210.6 2027.7 2222.9 1.17 4.19 4.56 0.20
E-1 120 213.5 210.1 2079.2 2063.9 2280.2 2258.2 1.17 1.18 3.87 3.90 4.16 4.27 0.17 0.16
E-2 120 212.7 2064.3 2267.7 1.17 3.93 4.46 0.19
E-3 120 204.1 2048.3 2226.7 1.19 3.91 4.20 0.11
F-1 140 213.7 217.4 2093.4 2096.0 2284.3 2299.5 1.20 1.18 4.17 4.02 4.41 4.33 0.19 0.17
F-2 140 217.5 2089.9 2297.0 1.18 3.82 4.18 0.17
F-3 140 221.2 2104.6 2317.2 1.16 4.07 4.38 0.15
G-1 160 215.9 215.6 2127.7 2133.1 2349.4 2327.4 1.18 1.17 3.98 3.83 4.30 4.19 0.15 0.13
G-2 160 209.7 2139.9 2315.9 1.19 3.66 4.22 0.16
G-3 160 221.1 2131.7 2316.8 1.15 3.83 4.06 0.09

Number in three identical specimens

A-1 E, E, fy, fy, fu, fu denote elastic modulus, average elastic modulus, yield strength, average yield strength, ultimate strength,
Temperature level: 1~7 denote 20, -40, -70,
-100, -120, -140, -160ćˈrespectively.
average ultimate strength of steel wires, respectively. ey, ey, eu, eu, eF, eF denote yield strain, average yield strain, ultimate strain, average ultimate strain, facture strain, average
facture strain of steel wires, respectively. w and w denote reductions in cross-sectional area and average reductions in cross-sectional area, respectively.
 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092 1081

300 fy fu
2500

y = -0.061x + 204.54 y = -2.150x + 1999.7

250 R² = 0.33 2300

Strength (MPa)
R² = 0.96
E (GPa)

2100
200
1900
y = -1.775x + 1855.8
R² = 0.97
150 1700
-200 -160 -120 -80 -40 0 40 -200 -160 -120 -80 -40 0 40
T( ) T( )
(a) Effect on E (b) Effect on fy and fu
u F y 0.21

4.00
0.18
3.25 1.20
Strain (%)

(%)
1.10
2.50
1.00 0.15
-200-160-120 -80 -40 0 40
1.75

1.00 0.12
-200 -160 -120 -80 -40 0 40 -200 -160 -120 -80 -40 0 40
T( ) T( )
(c) Effect on y, u and F (d) Effect on

Fig. 5. Effect of low temperature on different mechanical properties of steel wires.

of 100 °C, both ey and eu of high strength steel wires are 1.15
times of those values at ambient temperature. However, once the
temperature is below 100 °C, i.e., as T decreases from 100 °C
to 160 °C, both ey and eu slightly decreased. As T decreases from
100 °C to 120 °C, 140 °C, 160 °C, ey (oreu ) are averagely
reduced by 0.95% (4.14%), 0.57% (1.21%), and 1.30% (6.05%), respec-
tively. Thus, it can be concluded that as the temperature is below
100 °C, the influences of the low temperature on ey and eu are
marginal and even can be neglected.
Fig. 6 shows the typical failure modes of the tested specimen
and fracture surfaces of the cross section. It can be observed that
obvious necking occurred to the specimens after the tests at tem-
peratures of +20 °C to 100 °C. However, as temperature T is below
100 °C, failure mode of the tested specimens changes from duc-
tile mode to brittle mode.
2.4. Proposal stress-strain curves for the single wire at low
temperatures

2.3.5. Fracture strain and reductions in cross-sectional area

The fracture strain eF and reductions in cross-sectional area w
are used to describe the plasticity of steel at low temperatures.
Fig. 5 (c) and (d) depicts the influences of low temperature T on
eF and w, respectively. Table 2 lists the test results of eF and w at
various low temperatures. As T decreases from 20 °C to 100 °C,
eF increases with the decrease of temperature. As T decreases from
+20 °C to 40 °C, 70 °C, 100 °C, the average eF value is increased
by 1%, 4%, and 14%, respectively. However, as T is below 100 °C,eF
exhibits slight changes. As T decreases from 100 °C to 120 °C,
140 °C, 160 °C, the eF is averagely reduced by 2.03%, 0.81%,
and 3.90%, respectively. Meanwhile, w ratio almost remains at
the same level as the temperature decreases from +20 °C to
100 °C, which implies the effect of temperature T on w are quite
limited within this temperature interval. As T decreases from
100 °C to 120 °C, 140 °C, 160 °C, w is averagely reduced by
22%, 16%, and 33%, respectively.
The effects of low temperatures on yield strength fy and ulti-
mate strength fu of the single wire are shown in Fig. 5 and Table 2.
Based on the reported test data in Table 2, the mathematical rela-
tionships between the mechanical properties of steel wires and the
low temperature T are developed, along with regression analyses.
Thus, mathematical relationships between the yield (or ultimate)
strength and temperatures can be given as follows:
f y ¼ f ya eað1=T1=T 0 Þ ð3Þ
f u ¼ f ua ebð1=T1=T 0 Þ ð4Þ
where f ya and f ua express the yield strength and ultimate strength of
steel wires at ambient temperature, respectively; f y and f u express
the yield strength and ultimate strength of steel wires at tempera-
ture T, respectively, in K; T0 is the ambient temperature; a and b are
1082 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092

70 °C, 100 °C, 120 °C, 140 °C and 160 °C were used for
the determinations on A, B, a, b and k at different low temperatures
as listed in Table 3. Hence, the equation of regression analysis for
predicting the yield and ultimate strengths are recommended with
the given yield and ultimate strengths at ambient temperature as
follows:

f y ¼ f ya e50:17ð1=T1=T 0 Þ ð10Þ

f u ¼ f ua e54:31ð1=T1=T 0 Þ ð11Þ

where T0 and T denote the ambient temperature and low tempera-

ture, in K, and 114 K  T  293 K.
According to Eqs. (10) and (11), the simplified equations can be
obtained to predict the yield and ultimate strengths of steel wires
as follows:
(a) Fracture modes of the steel wires
f y ¼ f ya e0:0010ðT 0 TÞ ð12Þ
Ductile fracture Brittle fracture

20 -40 -70 -100 -120 -140 -160 f u ¼ f ua e0:0011ðT 0 TÞ ð13Þ

where T0 and T denote, respectively the ambient temperature and

low temperature, in °C, and 160 °C  T  20 °C.
Table 4 lists the predicted yield and ultimate strengths at differ-
ent low temperatures by Eqs. (10)–(13). These predicted values are
compared with those test results as shown in Fig. 7 (a) and (b). It
A-2 B-2 C-3 D-1 E-2 F-1 G-3 can be observed that the predictions by Eqs. (10)–(13) show high
(b) Fracture surfaces of the steel wires correlations with the experimental values and the correlation coef-
ficients R2 are all larger than 0.9. It can be also found that the cor-
Fig. 6. Fracture of the steel wires at different low temperatures. relation coefficients R2 for predictions by Eqs. (12) and (13) are
both 0.98 that are larger than those by Eqs. (10) and (11). This
implies Eqs. (12) and (13) offers more accurate predictions for yield
the sensitive coefficients of steel wires for yield strength and ulti-
and ultimate strengths of steel wires at different low temperatures.
mate strength, respectively, in 1/K.
Fig. 8 (a) and (b) compares the predictions by Eqs. (10)–(13)
The relationships between coefficient a and b can be obtained as
with the test data of 21 specimens. For Fig. 8 (a), most of the pre-
follows:
dictions fall within the scope 10% error of the test results and the
lgðf 0 =f u Þ average test-to-prediction are both 0.98. The coefficient of varia-
b¼a ð5Þ tion (COV) listed in Table 6 are 0.04 and 0.05, respectively. Fig. 8
lgðf 0 =f y Þ
(b) shows that the predictions fall within the scope 5% error of
lgðf 0 =f u Þ test results. Furthermore, the average test-to-prediction for yield
k¼ ð6Þ and ultimate strengths are about 1.0 and coefficients of variation
lgðf 0 =f y Þ
(COV) are only 0.01.
where k denotes the constant value related to the mechanical prop- According to these figures and data in Table 4, it can be con-
erties of steel wires at different low temperatures. The fracture cluded that all of the empirical formulae Eqs. (10)–(13) can offer
strength of the steel wires f0 is independent with the temperature satisfactory predictions on the yield and ultimate strengths of steel
that can be obtained by tensile tests. wires at different low temperatures. In particular, Eqs. (12) and
In Eqs. (3) and (4), the temperature is in Kelvin degree. In order (13) can offer more accurate predictions. Thus, once the yield
to facilitate to engineering applications, the unit of temperature and ultimate strengths of steel wires at ambient temperature are
can be converted from Kelvin degrees to Celsius degrees. The sim- given, Eqs. (10)–(13) can be used to determine the yield and ulti-
plified equations can be modified as the following: mate strengths of steel wires at different low temperatures within
20 °C to 160 °C. In addition, according to the test results, the
f y ¼ f ya eAðT 0 TÞ ð7Þ effects of low temperatures on elastic modulus and plastic modu-
lus is quite limited and can be neglected. Thus, stress-strain curves
f u ¼ f ua eBðT 0 TÞ ð8Þ for steel wires at low temperatures can be determined by the given
modulus and strengths obtained by Eqs. (10)–(13).
where A and B are the sensitive coefficients of steel wires for yield
strength and ultimate strength, respectively, in 1/.°C
Similar to a and b, the relationships between coefficient A and B Table 3
can be obtained according to Eqs. (5) and (6) and it is written as Different coefficients obtained from the regression analysis on test results.
follows: T (°C) a (1/K) b (1/K) A (1/K) B (1/K) k
B ¼ kA ð9Þ -40 67.44 70.13 0.0010 0.0010 1.040
-70 72.83 77.22 0.0012 0.0013 1.060
where the value of k in Eq. (9) can be calculated by Eq. (6). -100 50.76 56.22 0.0010 0.0011 1.107
With the test data in Table 2, different coefficients, A, B, a, b and -120 42.79 47.99 0.0009 0.0011 1.122
k are calculated by Eqs. (3)–(9). Due to the regression analysis and -140 36.38 41.01 0.0009 0.0010 1.127
-160 30.79 33.27 0.0009 0.0010 1.080
experimental studies focused on the mechanical properties of steel
Mean 50.17 54.31 0.0010 0.0011 1.090
wires at low temperatures, the experimental results of 40 °C,
 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092 1083

Table 4
Predicted yield and ultimate strengths at different low temperatures by the developed design equations.

T (°C) fy (MPa) fu (MPa) fy1 (MPa) fu1 (MPa) fy/fy1 fu/fu1 fy2 (MPa) fu2 (MPa) fy/fy2 fu/fu2
20 1799.4 1937.1 1808.5 1947.2 0.99 0.99 1808.5 1947.2 0.99 0.99
20 1817.9 1961.8 1808.5 1947.2 1.01 1.01 1808.5 1947.2 1.01 1.01
20 1808.2 1942.6 1808.5 1947.2 1.00 1.00 1808.5 1947.2 1.00 1.00
40 1931.3 2100.3 1889.3 2041.5 1.02 1.03 1920.3 2080.1 1.01 1.01
40 1903.3 2080.5 1889.3 2041.5 1.01 1.02 1920.3 2080.1 0.99 1.00
40 1919.2 2028.8 1889.3 2041.5 1.02 0.99 1920.3 2080.1 1.00 0.98
70 2035.0 2200.1 1949.7 2112.3 1.04 1.04 1978.8 2149.8 1.03 1.02
70 2003.4 2168.0 1949.7 2112.3 1.03 1.03 1978.8 2149.8 1.01 1.01
70 2012.9 2190.2 1949.7 2112.3 1.03 1.04 1978.8 2149.8 1.02 1.02
100 2043.8 2218.0 2034.0 2211.4 1.00 1.00 2039.1 2222.0 1.00 1.00
100 2039.0 2223.0 2034.0 2211.4 1.00 1.01 2039.1 2222.0 1.00 1.00
100 2027.7 2222.9 2034.0 2211.4 1.00 1.01 2039.1 2222.0 0.99 1.00
120 2079.2 2280.2 2111.5 2302.7 0.98 0.99 2080.3 2271.4 1.00 1.00
120 2064.3 2267.7 2111.5 2302.7 0.98 0.98 2080.3 2271.4 0.99 1.00
120 2048.3 2226.7 2111.5 2302.7 0.97 0.97 2080.3 2271.4 0.98 0.98
140 2093.4 2284.3 2216.5 2426.8 0.94 0.94 2122.3 2321.9 0.99 0.98
140 2089.9 2297.0 2216.5 2426.8 0.94 0.95 2122.3 2321.9 0.98 0.99
140 2104.6 2317.2 2216.5 2426.8 0.95 0.95 2122.3 2321.9 0.99 1.00
160 2127.7 2349.4 2366.5 2605.2 0.90 0.90 2165.2 2373.6 0.98 0.99
160 2139.9 2315.9 2366.5 2605.2 0.90 0.89 2165.2 2373.6 0.99 0.98
160 2131.7 2316.8 2366.5 2605.2 0.90 0.89 2165.2 2373.6 0.98 0.98
Mean 0.98 0.98 1.00 1.00
Cov 0.04 0.05 0.01 0.01

fy1, fy2 denote the yield strengths predicted by Eqs. (10) and (12), respectively. fu1, fu2 denote the ultimate strengths predicted by Eqs. (11) and (13), respectively.

Fig. 7. Comparisons of the scatters of predictions by the design equations with the test results for different steel wires.

Fig. 8. Comparisons of predictions by the design equations with test results for 21 steel wires.
1084 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092

0 1 0 1 0 1
3. Theoretical models on predicting stress-strain behaviours of sinui sinai cosui sinui cosai
B C B C B C
steel strands t i ¼ @ cosui sinai A; ni ¼ @ sinui A; bi ¼ @ cosui cosai A
cosai 0 sinai
Multi-layer strands were usually used in the prestressed con-
ð17Þ
crete structures. Different from single wire, multi-layer strands
were made through twisting several steel wires together as an where ti, ni and bi are tangent, normal and binormal unit vectors,
integrity. These geometric and processing differences would pro- respectively.
duce the differences of nominal stress-strain curves of the steel The relationship between the local coordinate system and the
strands from those of single straight wire. Compared with tensile global coordinate system is expressed as follows:
tests on steel strands, tensile tests on wires are much easier to per- 0 1 0 10 1
form. This section makes efforts to develop the theoretical models ni cosui sinui 0 eX
B C B CB C
to predict the stress-strain curves of multi-layer steel strands @ bi A ¼ @ sinui cosai cosui cosai sinai A@ eY A ð18Þ
based on the tensile stress-strain curves of steel wires. ti sinui cosai cosui cosai cosai eZ

3.1. Geometric model and elastic-plastic constitutive law of steel The curvature in the normal direction, binormal direction and
strands twists of helical line in Frenet-Serret frame before loading can be
defined as follows:
The multi-layer strand usually consists of a straight core wire 2
sin ai sinai cosai
with radius r0, surrounded by several layers of helical wires in a ji ¼ 0; j0i ¼ ; vi ¼ ð19Þ
Ri Ri
symmetrical way. The helix radius Ri and pitch length Pi in unload-
ing stage can be calculated as follows: The length of helical wires and core wire the centerline are
defined as si and l. After axial deformation, the small change of heli-
X
i1
Ri ¼ r 0 þ 2rj þ r i ð14Þ cal wires, core wire, lay angle, polar angle and helix radius are
j¼1 defined as dsi , dl, dai , dui and dRi . The axial strain of core wire
and helical wires can be expressed as e0 ¼ dl=l and ei ¼ dsi =si ,
2pRi respectively. The polar angle is expressed as follows:
Pi ¼ ð15Þ ui ¼ sinai si =Ri .
tanai
According to [15], the helical wires strain ei and torsional strain
where r0 is the radius of core wire, and ri is the radius of helical ci of helical wires can be obtained as follows:
wires of layer i. ai denotes lay angle of helical wires of layer i as
shown in Fig. 9. ei ¼ e0  tanai dai ð20Þ
In unloading stage, the centerline of the helical wires are given
P
by the following form: dui ei r 0 e0 þ i1
j¼1 r j ej þ r i ei
ci ¼ Ri ¼  dai þ m ð21Þ
ui s tanai Ri tanai
rðui Þ ¼ Ri cosui eX þ Ri sinui eY þ Ri eZ ð16Þ
tanai
where m is the Poisson’s ratio of steel wire.
where a right-handed Cartesian frame {eX , eY , eZ } is used to describe The change of helix radius can be influenced by Poisson’s ratio
the geometric structure of the steel strand, and eZ coincides with and the deformation between helical wires and core wire. Due to
the strand centerline as shown in Fig. 9. ui denotes the polar angle the effect of interwire deformation on small-diameter steel strands
of helix. This angle can be measured around the eZ relative to eX . is quite limited, for simplifying the calculation, interwire deforma-
A local coordinate system attached to the centerline of the heli- tion is neglected, and thus, only the influence of Poisson’s ratio is
cal wire is used to describe the local deformation as shown in considered in this work. The change of helix radius can be calcu-
Fig. 10 (a). Eq. (17) can be obtained as follows: lated as follows:

Fig. 9. A multi-layer strand with cross section.

 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092 1085

s
b
s

s s

(a) (b) (c )
Fig. 10. (a) Centerline of a helical wire. (b) Deformation through a wire acting on a helical wire [15]. (c) Force and moment resultants cross section [24].

X
i1
dRi ¼ ðmr 0 e0 þ 2mrj ej þ mr i ei Þ ð22Þ
j¼1

The change of curvature in the normal direction, binormal

direction and twists of helical line after axial deformation can be
defined as follows:
dji ¼ 0 ð23aÞ

2 2
sin ai sin ai
dj0i ¼  ð23bÞ
Ri þ dRi Ri

sinai cosai sinai cosai

dvi ¼  ð23cÞ
Ri þ dRi Ri
 Fig. 11. Bi-linear curve of steel wire.
where ai denotes the lay angle after deformation and can be

expressed as ai ¼ ai þ dai . rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
An elastic-plastic constitutive law is introduced in this theoret- 3  
f ¼ sij  C epij sij  C epij  rs  0 ð29Þ
ical models. Suppose that any point on the cross section of helical 2
wires is Sðr s ; hs Þ, and its axial strain es and tangential strain cs as
shown in Fig. 10 (b) can be determined as follows: where constant C is obtained by a simple tensile test and equals to
2/3Ep in Eq. (29). The symbols sij denotes the stress deviator,
es ¼ ei  dj0i rs coshs ð24Þ sij ¼ rij  rm dij . The symbols epij denotes the plastic strain and
defined as follows:
cs ¼ rs dvi ð25Þ
0 1
Accordingly, the axial strain es and torsional strain cs can be
eps 0  12 cps sinhs
decomposed into purely elastic strainee , ce and plastic strainep , cp , epij ¼ B
@ 0 meps 1
2
cps coshs C
A ð30Þ
p p
i.e., es ¼ ee þ ep , cs ¼ ce þ cp . In the elastic region, the normal stres-
1
2
cs sinhs  12 c s coshs meps
ses r is obtained by adopting the generalized Hooke’s law as In the plastic region, the incremental plastic strain can be writ-
follows: ten as:
E
r ¼ ee E; s ¼ c ¼ Gce ; ð26Þ depij ¼ deij  deeij ð31Þ
2ð1 þ mÞ e
The incremental strain resulting from a stress increment that
where E denotes the elastic modulus of high strength steel wires.
considered the classic Prandtl-Reuss associated flow rules is intro-
In the plastic region, the normal stress r can be calculated as
duced as follows:
follows:
r ¼ rs þ Ep ep ð27Þ @f
depij ¼ dk ð32Þ
@ rij
EE1
Ep ¼ ð28Þ @f
E  E1 @ rij
deij
dk ¼ 2r2 ð33Þ
where E1 and rs denotes the plastic modulus and the first-yielding s
ð3G þ Ep Þ
9G
stress of the high strength steel wires are shown in Fig. 11.
Subsequent yield criterion f is defined by the von-Mises yield where the symbol k is a proportional coefficient that satisfies flow
criterion, along with a kinematic hardening law as follows: rules and consistency conditions.
1086 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092

Newton-Raphson method was used for calculations of incre-
mental steps of elastic-plastic constitutive law [23,25].
Based on Love’s curved beam equations, the effect of Poisson’s
ratio and lay angle variations are considered, and the effect of
interwire friction and shear deformability is ignored. As shown in
Fig. 10 (c), the following equilibrium equations are established:
N0i ðvi þ dvi Þ þ T i ðj0i þ dj0i Þ þ X i ¼ 0
stresses and tangential stresses of helical wires can be calculated
through the elastic-plastic constitutive law defined by Eqs. (24)–
(33) that is proposed by using the von-Mises yield criterion, along
with a kinematic hardening law. Based on Love’s curved beam
equations and Costello’s model [16], the equilibrium equations
Eqs. (34) are established, and the resultant axial force can be calcu-
lated by Eqs. (35)–(37). Thus, the stress-strain behaviours of steel
ð34aÞ
strands can be obtained by Eqs. (14)–(38).

G0i ðvi þ dvi Þ þ Hi ðj0i þ dj0i Þ  N 0i ¼ 0 ð34bÞ

3.2. Model validations
where T i and Hi respectively express the force and the twisting
moment in tangential direction of helical wires, N0i and G0i respec- To verify the accuracy of the elastic-plastic mechanical model,
tively express the force and the bending moment in binormal direc- the theoretical results were compared with experimental data
tion of helical wires and X i is line load per unit length in normal reported by Utting and Jones [26]. The geometric and material
direction. details of the tested steel strands are listed in Table 5 [18,26].
The twisting momentHi and bending moment G0i in helical wires The proposed theoretical model in this manuscript is also com-
are derived through integration over the cross sectional areaAi : pared with Costello’s theory that is based on linearly elastic consti-
Z tutive law. The presented stress-strain curves of seven-wire and
Hi ¼ r s sdAi ð35aÞ nineteen-wire steel strands by the developed theoretical models
Ai
in this manuscript resemble well with curves proposed by Costello
Z [15] in the linear portion, and these models also can give accurate
G0i ¼  rs coshs rs dAi ð35bÞ predictions in the plastic portion, as depicted in Fig. 12. For seven-
Ai
wire strand, the curve by the proposed model agrees well with the
The force T i in tangential direction of helical wires can be experimental curve [26] in both linear and plastic portions. The dif-
obtained as follows: ferences of elastic modulus between the theoretical results and
Z test results are only about 1%. And in plastic range, the axial force
Ti ¼ rs dAi ð35cÞ is slightly larger than test results with a small difference of less
Ai
than 3%. For nineteen-wire strand, it can be noted that the theoret-
The total axial force of each layer of helical wires can be calcu- ical results are slightly higher than the test results of Utting and
lated as follows: Jones [26]. That is due to the fact that contact deformation is
ignored in the proposed model.
F i ¼ mi T i cosðai þ dai Þ þ mi N0i sinðai þ dai Þ ð36Þ
where mi represents the number of helical wires in each layer.
4. Numerical model of stress-strain behaviours of steel strands
The resultant axial force acting on the steel strand is the combi-
nation of the force on each wire of steel strand as follows:
4.1. Finite element model
Z X
i
F¼ r0 dA0 þ Fi ð37Þ 4.1.1. General
A0 j¼1
A three-dimensional finite element model was developed using
where A0 is the cross sectional area of core wire. general commercial software ABAQUS to simulate the tensile beha-
In summary, the value of axial strain of steel strand equal to viours of steel strands at different low temperatures. ABAQUS/
axial strain of core wire and the axial stress of steel strand is Standard type of implicit solver was used for the analysis on steel
obtained as follows: strands.

F
rstrand ¼ Pi
A0 þ j¼1 mi Ai
estrand ¼ e0
The above equations define the elastic-plastic strain-stress
behaviours of steel strands. The geometric model of steel strands
is defined by Eqs. (14)–(19). In the solution process, the axial strain
of steel strand is assigned, which equals to the axial strain of core
wire e0 . The axial strain ei and torsional strain ci of helical wires
defined by Eqs. (20)–(21) can be derived from the geometric rela-
tionship between core wire and helical wires. Within a local coor-
dinate system, the normal stresses of core wire, the normal
ð38aÞ 4.1.2. Elements and material properties
All the components were modelled by three-dimensional eight-
node continuum elements with reduced integration point and
ð38bÞ
hourglass control (C3D8R) in ABAQUS element library. This ele-
ment consists of three translation degrees of freedom at each node
and one integration point.
Convergence studies were also performed to find an appropriate
mesh density that considered both computing accuracy and effi-
ciency. Finally, the mesh sizes for the radical elements are 1/10
of the wire diameter and the longitudinal elements are 1/80 of
helical pitch, respectively. Fig. 13 shows the FE model with final-
ized mesh size.

Table 5
The geometric and material data.

Type r0 (mm) r1 (mm) r2 (mm) a1 (°) a2 (°) E (GPa) E1 (GPa) fy (MPa) fu (MPa) m
1*7 1.97 1.865 – 11.8 – 188 24.6 1540 1800 0.3
1*19 1.83 1.665 1.665 14.6 14.4 188 24.6 1540 1800 0.3

r0, r1, r2 denote the radius of core wire, first layer helical wires and second layer helical wires, respectively. a1 and a2 denote the lay angle of first layer helical wires and second
layer helical wires, respectively. E, E1, m, fy and fu denote the elastic modulus, plastic modulus, Poisson’s ratio, the yield and ultimate strengths, respectively.
 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092 1087

Table 6
Predicted elastic modulus, yield and ultimate strengths at different low temperatures by theoretical and numerical models.

N T (°C) D (mm) a1(°) a2(°) fys,a(MPa) fus,a (MPa) Es,a(GPa) fys,e (MPa) fus,e(MPa) Es,e(GPa)
3 20 6.2 5 – 1801.7 1939.4 197.3 1766.5 1890.7 194.0
3 40 6.2 5 – 1913.0 2072.1 197.3 1873.6 2019.7 194.0
3 70 6.2 5 – 1971.3 2141.3 197.3 1928.2 2087.4 194.0
3 100 6.2 5 – 2031.3 2213.5 197.3 1984.3 2157.5 194.0
3 120 6.2 5 – 2072.4 2263.4 197.3 2032.3 2205.4 194.0
3 140 6.2 5 – 2114.2 2313.6 197.3 2071.6 2254.5 194.0
3 160 6.2 5 – 2156.9 2365.1 197.3 2111.6 2304.6 194.0
3 20 6.2 10 – 1781.5 1926.1 189.3 1748.0 1870.0 188.8
3 40 6.2 10 – 1891.6 2059.4 189.3 1858.2 1997.6 188.8
3 70 6.2 10 – 1949.2 2128.9 189.3 1912.2 2064.6 188.8
3 100 6.2 10 – 2008.6 2200.3 189.3 1967.8 2133.9 188.8
3 120 6.2 10 – 2049.2 2250.7 189.3 2005.8 2181.3 188.8
3 140 6.2 10 – 2090.6 2300.6 189.3 2044.5 2229.8 188.8
3 160 6.2 10 – 2132.8 2352.5 189.3 2084.0 2279.3 188.8
3 20 6.2 15 – 1749.2 1887.3 176.9 1700.0 1838.4 176.7
3 40 6.2 15 – 1857.4 2020.1 176.9 1776.8 1963.8 176.7
3 70 6.2 15 – 1913.9 2088.6 176.9 1866.4 2029.6 176.7
3 100 6.2 15 – 1972.2 2161.7 176.9 1913.4 2097.6 176.7
3 120 6.2 15 – 2012.1 2210.6 176.9 1942.5 2144.1 176.7
3 140 6.2 15 – 2052.7 2261.4 176.9 2017.8 2191.6 176.7
3 160 6.2 15 – 2094.2 2313.3 176.9 2056.4 2240.0 176.7
7 20 11.4 5 – 1802.8 1920.7 197.7 1752.8 1851.5 193.1
7 40 11.4 5 – 1914.3 2051.0 197.7 1834.5 1971.0 193.1
7 70 11.4 5 – 1972.6 2119.6 197.7 1874.5 2031.7 193.1
7 100 11.4 5 – 2032.7 2190.5 197.7 1931.0 2096.7 193.1
7 120 11.4 5 – 2073.7 2239.3 197.7 1951.0 2140.4 193.1
7 140 11.4 5 – 2115.6 2289.0 197.7 1982.7 2184.4 193.1
7 160 11.4 5 – 2158.3 2339.5 197.7 2010.6 2229.8 193.1
7 20 11.4 10 – 1785.8 1897.0 191.0 1723.5 1809.1 184.1
7 40 11.4 10 – 1896.3 2026.1 191.0 1822.5 1926.7 184.1
7 70 11.4 10 – 1954.0 2093.7 191.0 1859.3 1987.9 184.1
7 100 11.4 10 – 2013.5 2163.6 191.0 1896.1 2050.4 184.1
7 120 11.4 10 – 2054.2 2211.7 191.0 1951.4 2093.9 184.1
7 140 11.4 10 – 2095.7 2260.7 191.0 1988.2 2137.7 184.1
7 160 11.4 10 – 2138.0 2310.4 191.0 2025.0 2182.3 184.1
7 20 11.4 15 – 1758.0 1859.4 179.4 1607.9 1723.2 171.0
7 40 11.4 15 – 1866.8 1985.5 179.4 1709.7 1854.2 171.0
7 70 11.4 15 – 1923.6 2051.5 179.4 1789.0 1914.5 171.0
7 100 11.4 15 – 1982.2 2119.8 179.4 1827.0 1977.1 171.0
7 120 11.4 15 – 2022.2 2166.8 179.4 1851.3 2019.0 171.0
7 140 11.4 15 – 2063.1 2214.7 179.4 1874.1 2060.1 171.0
7 160 11.4 15 – 2104.8 2263.8 179.4 1895.5 2104.6 171.0
19 20 13 10 5 1796.8 1933.5 194.9 1737.5 1812.8 186.8
19 40 13 10 5 1907.9 2063.1 194.9 1838.8 1929.6 186.8
19 70 13 10 5 1966.0 2130.7 194.9 1888.6 1990.9 186.8
19 100 13 10 5 2025.9 2202.5 194.9 1953.6 2054.1 186.8
19 120 13 10 5 2066.8 2251.6 194.9 1989.5 2096.7 186.8
19 140 13 10 5 2108.5 2300.8 194.9 2038.0 2140.4 186.8
19 160 13 10 5 2151.1 2352.2 194.9 2077.4 2182.0 186.8
19 20 13 10 10 1782.7 1912.8 189.9 1703.2 1795.1 177.8
19 40 13 10 10 1892.9 2043.5 189.9 1822.3 1911.7 177.8
19 70 13 10 10 1950.6 2111.0 189.9 1874.5 1972.6 177.8
19 100 13 10 10 2009.9 2181.4 189.9 1928.1 2035.5 177.8
19 120 13 10 10 2050.6 2230.6 189.9 1964.0 2078.1 177.8
19 140 13 10 10 2092.0 2279.3 189.9 2000.0 2121.7 177.8
19 160 13 10 10 2134.2 2330.1 189.9 2035.9 2166.2 177.8
19 20 13 10 15 1761.4 1884.7 181.8 1681.7 1763.5 166.6
19 40 13 10 15 1870.3 2012.5 181.8 1774.2 1878.2 166.6
19 70 13 10 15 1927.2 2078.8 181.8 1821.2 1938.5 166.6
19 100 13 10 15 1985.9 2148.0 181.8 1886.3 2000.1 166.6
19 120 13 10 15 2026.1 2195.6 181.8 1919.5 2042.1 166.6
19 140 13 10 15 2067.0 2244.1 181.8 1953.0 2085.3 166.6
19 160 13 10 15 2108.7 2293.9 181.8 1986.3 2129.1 166.6
19 20 13 15 10 1774.2 1898.6 185.9 1667.5 1769.1 172.6
19 40 13 15 10 1883.9 2025.8 185.9 1748.2 1884.0 172.6
19 70 13 15 10 1941.2 2090.3 185.9 1787.0 1943.1 172.6
19 100 13 15 10 2000.3 2162.5 185.9 1824.4 2001.5 172.6
19 120 13 15 10 2040.8 2207.6 185.9 1898.2 2038.9 172.6
19 140 13 15 10 2082.0 2259.5 185.9 1926.4 2071.0 172.6
19 160 13 15 10 2124.0 2308.0 185.9 1950.0 2114.4 172.6

N denotes the number of wires. D is the diameter of steel wires. a1, a2 are the lay angle of the first layer steel wires and second layer steel wires. Es,a, fys,a, fus,a denote the elastic
modulus, yield and ultimate strengths predicted by theoretical model. Es,e, fys,e, fus,e denote the elastic modulus, yield and ultimate strengths predicted by numerical model.
1088 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092
140
105
Axial force (kN)
70
35
0 0
0.000 0.005
Strain
(a) Seven-wire strand
Fixed end
U(1)=Free
U(2)=Fixed
U(3)=Fixed
Displacement directions:
1=radial
2=circumferential
3=longitudinal
Fig. 13. FEM for strand.
A nonlinear isotropic model that adopted von-Mises yield crite-
rion was used for steel strands. Typical bi-linear stress-strain
curves with strain hardening was adopted in this model, as shown
in Fig. 11. The values of the yield, ultimate strengths, elastic mod-
ulus and Poisson’s ratio were determined according to the test
results.
4.1.3. Boundary conditions, and loading
Boundary conditions of the strand used in the FE modelling are
shown in Fig. 13. Two major steps that include 1) coupling the two
ends of the steel strands to two reference points, and 2) defining
the boundary condition to these two reference points at both ends.
In order to simulate the boundary conditions, a local cylindrical
coordinate system was created in radial, circumferential and longi-
tudinal directions. Taking into account of the Poisson effect, the
radial contraction of wires was allowed, and the translation along
the radial direction was free at two ends. For circumferential
motion, torsional rotation was fixed at both two ends of the strand
to prevent unwinding of steel wires. Furthermore, the displace-
ment along longitudinal direction was fixed at one end, but
released at the loading end. The tensile force was applied by
displacement-control on the loading end to simulate the loading
condition of the tensile test.
4.1.4. Interactions
A surface-to-surface type of contact was used to define the
interactions among the surfaces of different wires in the steel
strand. This contact type described the two interacting surfaces
in both normal direction and tangential direction. Hard contact
and penalty friction type contact algorithms were used to define
the contacts in these two directions of the interacting surface,
respectively.
320
240
Axial force (kN)
160
Test [25] Test [25]
Analytical 80 Analytical
FEM FEM
Costello's model Costello's model
0.010 0.015 0.000 0.005 0.010 0.015 0.020
Strain
(b) Nineteen-wire strand
Fig. 12. Axial load–axial strain curves.
A finite sliding algorithm allowing large rotations and deforma-
tions, was adopted to define the contact surface interactions. Sur-
Loading end face of core wire was chosen as the master surface because of its
U(1)=Free axial stiffness was greater than the helical wires. In the contact
U(2)=Fixed interacting algorithm, the surface-to-surface contact can reduce
U(3)=Free
the risk of the large, undetected penetrations of master surface into
the slave surface and obtain a smoother contours. Furthermore,
this approach of contact also can improve stability at the corners
and edges. More accurate solution can be obtained via this
approach, but required more computational time and space. Pen-
alty contact algorithm was adopted to describe the interaction in
the tangential direction to the contact surfaces. An isotropic Cou-
lomb law was used to model the interwire frictional effect. How-
ever, preliminary studies have shown that the interwire frictional
effect can be expressed by a constant friction coefficient varying
from 0.1 to 0.2 [21]. Different finite models were established with
different friction coefficients. The calculation results showed that
the FE model with a constant friction coefficient 0.1 fitted test data
very well than 0.2.
4.2. Model validations
To verify the accuracy of the finite element (FE) model, the
results from finite element analysis (FEA) are compared with the
test results reported by Utting and Jones [26]. The geometric and
material details of the steel strands are listed in Table 5, which
are taken from [18,26]. The stress-strain curves from FEA are com-
pared with the test results in Fig. 12. It can be observed that
numerical results resemble well with those experimental curves.
For seven-wire and nineteen-wire strands, the differences of elastic
modulus between the numerical results and test results are both
only about 1%. And for seven-wire strand, the numerical results
in plastic range is slightly larger than test results with a small dif-
ference of less than 2%. Thus, it can be concluded that the devel-
oped FEM can give reasonable predictions on the stress-strain
behaviours of multi-layer steel strands.
4.3. Discussions on localized analysis
In addition to global response, the exact finite model proposed
in this manuscript can be used for localized analysis of strands
with stressing. Taking seven-wire strands for example, from the
contours of the equivalent plastic strain as shown in Fig. 14, it
can be seen that the onset of plastic deformation is confirmed by
the occurrence of equivalent plastic strain in core wire. Fig. 15(a)
and (b) shows the von-Mises and normal stress distribution in
 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092
Core wire
Helical wire
Fig. 14. Equivalent plastic strain at 0.76% elongation.
cross section at the onset of the plastic deformation in the strand.
At that time, the axial force is 89kN and the axial strain of the
strand is 0.0076. It can be observed that the plastic deformation
initiates from the contacting surfaces between the helical wires
and the core wire. The maximum stress also developed along the
contact lines. As the axial strain of the strand equals to 0.0095, full
cross section yield as shown in Fig. 15(c) and (d). From the normal
stress contours as shown in Fig. 15 (b) and (d), it can be seen that
the contacting stresses can be obtained as expected between the
surfaces of the adjacent helical wires.
5. Development of stress-strain curves of steel strands at low
temperatures
Even though stress-strain curves of the multi-layer strands at
low temperatures can be predicted by the developed theoretical
and numerical models with the given curves of steel wires at dif-
Fig. 15. (a) von-Mises stress at 0.76% elongation. (b) Normal stress at 0.76% elongation. (c) von-Mises stress at 0.95% elongation. (d) Normal stress at 0.95% elongation.
1089
ferent low temperatures, it is still complex for the structural engi-
neers. This section further simplified this procedure, and makes
efforts to developed empirical formulae to predict the mechanical
properties of the representative multi-layer strands used in the
engineering constructions at different low temperatures with the
only given test data of the steel wires at ambient temperature.
Before carrying out this study, the representative steel strands,
i.e., three-wire, seven-wire, and nineteen-wire strands were
selected. Considering these steel strands were fabricated with dif-
ferent lay angles and pitch lengths, 70 cases were considered in
this study. With the validated theoretical and numerical models,
the yield strength, ultimate strength and elastic modulus of these
different cases are determined and listed in Table 6. In addition,
the material properties of these strands adopts the test values
reported in this paper.
The mathematical relationships between mechanical beha-
viours of the single wire at ambient temperature and the global
response of steel strands at low temperatures were developed,
along with multiple regression analyses. The factors are defined
as follows:
Es
IE ¼ ð39Þ
Ea
f ys
If y ¼ ð40Þ
f ya
f us
If u ¼ ð41Þ
f ua
where IE , If y and If u are the factors for elastic modulus, yield strength
and ultimate strength, respectively. Ea , f ya and f ua are the elastic
1090 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092

modulus, yield strength and ultimate strength of steel wires at For regression models established by p predictors from total n
ambient temperature, respectively. Es , f ys and f us are the elastic predictors (p < n), the best subset is that with the smallest value
modulus, yield strength and ultimate strength of steel strand at dif- of Mallows Cp index. The correlation coefficient R2 is a value
ferent low temperatures, respectively. between 0 and 1, which describes the degree of correlation
Considering the influences of temperature T, pitch length P and between the data and the regression results. Meanwhile, the accu-
the number of wires N, the exponential models were developed in racy of the regression models can be evaluated by the standard
regression analysis. For nineteen-wire steel strands, the pitch error of the regression S. Thus, the combinations and the number
length of helical two layers of steel wires are defined as P1 and of predictors in the regression models can be determined. It is
P2. Therefore, the separate regression models for three-wire, noteworthy that regression models should be developed with the
seven-wire and nineteen-wire steel strands are established in this lowest possible number of predictors that ‘‘adequately” describe
work. The assumed regression analysis equations are given as the data.
follows: The regression analysis results are listed in Table 7. It can be
( observed that: (1) For three-wire and seven-wire steel strands,
cN d P f egT for 3 - wire and 7 - wire steel strands the regression models 4E, 3fy and 3fu are recommended consider-
IE ¼ f f
cN d P11 P22 egT for 19 - wire steel strands ing the above evaluation criteria. In model 3fy and 3fu, the regres-
ð42Þ sion models are developed with the number of wires N, pitch
length P and temperature T. However, model 4E is developed with-
( out considering temperature T. (2) For nineteen-wire steel strands,
cN d P f egT for 3 - wire and 7 - wire steel strands
If y ¼ f f
the regression models 6E, 6fy and 6fu are recommended. In model
cN d P 11 P22 egT for 19 - wire steel strands 6fy and 6fu, the regression models are developed with the pitch
ð43Þ length P1, P2 and temperature T. However, model 6E is developed
only with pitch length P1, P2. For nineteen-wire steel strands, the
(
number of wires N shows lower correlations with data which is
cN d P f egT for 3 - wire and 7 - wire steel strands
If u ¼ f f neglected in regression analyses.
cN d P11 P22 egT for 19 - wire steel strands
The recommended regression models for factors on elastic
ð44Þ modulus, yield strength and ultimate strength are given as follows:
where c, d, f1, f2 and g are constants that can be determined by
(1) For three-wire and seven-wire steel strands
regression analysis. T is the low temperature, in °C, and
160 °C  T  20 °C.
In order to facilitate multiple linear regression analysis, Eqs. Es
(42)–(44) needs a logarithmic transformation. Several regression IE ¼ ¼ 0:64N0:055 P0:092 ð45Þ
Ea
analysis methods are available for this work, e.g., stepwise regres-
sion method, best subset method, forward selection method and f ys
backward elimination method. The best subset method was used If y ¼ ¼ 0:87N0:058 P0:04 e0:001T ð46Þ
f ya
in this study to evaluate the significant predictors for IE , If y and
If u . The selection of proper subset of predictors for regression mod- f us
els must meet three evaluation criteria of Mallows Cp index, the If u ¼ ¼ 0:86N0:047 P 0:038 e0:001T ð47Þ
f us
correlation coefficient R2, and standard error of the regression S.

Table 7
The best subset regression analysis of lnIE, lnIfy, lnIfu on lnN, lnP and T.

Response Model n R2 Mallows Cp S lnN lnP1 lnP2 T

IE 1E 1 0.05 298.1 0.046 x* –
2E 1 0.66 85.6 0.028 x –
3E 2 0.66 84.4 0.028 x – x
4E 2 0.89 2.0 0.016 x x –
5E 3 0.89 4.0 0.016 x x – x
6E 2 0.953 2.0 0.010 x x
7E 3 0.953 4.0 0.010 x x x
8E 4 0.953 4.0 0.010 x x x x
Ify 1fy 1 0.76 129.0 0.031 – x
2fy 2 0.80 102.5 0.028 x – x
3fy 3 0.95 4.0 0.015 x x – x
4fy 1 0.869 292.4 0.022 x
5fy 2 0.121 2094.4 0.059 x x
6fy 3 0.990 4.0 0.006 x x x
7fy 4 0.990 4.0 0.006 x x x x
Ifu 1fu 1 0.87 253.7 0.025 – x
2fu 2 0.90 177.0 0.022 x – x
3fu 3 0.98 4.0 0.009 x x – x
4fu 1 0.962 235.6 0.012 x
5fu 2 0.034 6586.0 0.063 x x
6fu 3 0.996 4.0 0.004 x x x
7fu 4 0.996 4.0 0.004 x x x x
*
denotes considered predictors in the subset regression analysis. n is the number of considered predictors.
 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092 1091

Step 3. Based on the results in Step 2, the stress-strain curves of

steel strands at low temperatures can be obtained by the theo-
(2) For nineteen-wire steel strands retical models, i.e., Eqs. (14)–(38) and FE model.
Step 4. For three-wire, seven-wire and nineteen-wire steel
Es strands, with the obtained mechanical properties of steel wires
IE ¼ ¼ 0:44P0:037
1 P 0:099
2 ð48Þ
Ea at ambient temperature (Step 1), the mechanical properties of
steel strands can be directly calculated by Eqs. (45)–(50).
f ys
If y ¼ ¼ 0:56P0:082
1 P 0:032
2 e0:001T ð49Þ
f ya 7. Conclusions

f us This manuscript firstly reported the test results on the mechan-

If u ¼ ¼ 0:72P0:032
1 P0:022
2 e0:001T
f ua
In Eqs. (45)–(50), Ea denotes elastic modulus of steel wires at
ambient temperature, in GPa, f ya and f ua denotes yield strength
and ultimate strength of steel wires at ambient temperature, in
MPa; T is the low temperature, in °C, and 160 °C T  20 °C. In
Eqs. (45)–(47), the number of wires N is equal to 3 or 7; P denotes
the pitch length of steel strands, in mm. In Eqs. (48)–(50), P1 and P2
denote the pitch lengths of two layers helical wires of steel strands,
in mm. Otherwise, it is worth noting that the pitch length P can be
defined with helix radius R and lay angle a through a geometric
relationship P ¼ tan
2pR
a.
6. Proposed prediction methods on stress-strain curves of steel
strands at low temperatures
In order to facilitate the engineering application of the steel
strands at different low temperatures, the flowchart is shown in
Fig. 16 and following procedures are recommended:
Step 1. The tensile tests are carried out to obtain stress-strain
curves of steel wires at ambient temperature.
Step 2. Eqs. (10)–(13) can be used to predict the mechanical
properties of steel wires at different low temperatures.
ð50Þ ical properties of high strength steel wires at different low temper-
atures. Empirical formulae were also proposed to predict the low-
temp stress-strain behaviours of steel wires. Followed, theoretical
and numerical models were developed to predict the stress-
strain behaviours of multi-layer steel strands at low temperatures.
Empirical prediction equations were finally developed to predict
the tensile stress-strain behaviours of the multi-layer strand with
the given mechanical properties of steel wires at ambient temper-
ature. Based on these experimental, theoretical and numerical
studies, the following conclusions can be both drawn:
(1) As the temperature decreased from +20 °C to 160 °C, yield
and ultimate strengths of steel wires were increased by about
20%; However, the elastic modulus E of steel wires exhibited
neglected marginal changes. As the temperature decreased
from +20 °C to 100 °C, the values of yield strain ey , ultimate
strain eu and failure strain eF were averagely increased by
about 15%. As the temperature was below 100 °C, the influ-
ences of the low temperature on ey , eu and eF were marginal.
The low temperature only affected w ratio as it was below
100 °C, and w ratio was averagely reduced by 33% as the
temperature decreased from 100 °C to 160 °C. The frac-
ture mode of steel strands changed from ductility to brittle
mode at the low temperature of 100 °C.

Start

Determine curves of steel wires at ambient temperature from tensile tests

Determine f y, f u of steel wires at different low

temperatures by Eq.(10)-(13)

Determine mechanical Calculation of

properties of steel strand
by theoretical method by
Eq.(14)-(28)
Determine mechanical representative steel strand
properties of steel strand
at low temperatures by Eq.
by FE method (45)-(50)

Determine curves of steel strands at different low temperatures

End

Fig. 16. The flowchart of proposed prediction methods on stress-strain curves of steel strand at low temperatures.
1092 J. Xie et al. / Construction and Building Materials 189 (2018) 1076–1092
(2) Empirical prediction models, i.e. Eqs. (10)–(13), were devel-
oped to predict the mechanical properties of steel wires at
low temperatures. Validations of the predictions against
the test results proved their accuracies.
(3) This manuscript developed theoretical models to predict the
mechanical properties of multi-layer steel strands. These
models considered the geometric and material nonlinearity
of multi-layer strands. Their accuracies were checked by test
results, and proved to be capable of predicting stress-strain
behaviours of steel strands at low temperatures.
(4) A three-dimensional finite element model was also devel-
oped to simulate the stress-strain behaviours of multi-
layer steel strands at low temperatures. It simulated the dif-
ferent geometric details and material nonlinearities.
Through extensive validations of the FE predictions against
the test results, the FE model was proved to be capable of
simulating stress-strain curves of steel strands at low
temperatures.
(5) Based on the theoretical and numerical results, regression
models were also developed to predict the mechanical prop-
erties of representative types of steel strands at low temper-
atures. Eqs. (45)–(50) can be used to predict the elastic
modulus, yield and ultimate strengths of three-wire,
seven-wire and nineteen-wire steel strands at low tempera-
tures with the given mechanical properties of steel wires at
ambient temperature. This may facilitate determination of
mechanical properties of steel strands at low temperatures
especially during the engineering design process.
So far, the study was limited to ultimate tensile behaviours of
steel strands and steel wires at low temperatures. More future
works will be needed on their dynamic behaviours since these Arc-
tic offshore structures are exposed to impacting ice loads.
Conflicts of interest statement
The authors declare that they have no conflicts of interests.
Acknowledgment
This work was financially funded by the National Natural
Science Foundation of China (Grant No. 51608358). The authors
gratefully express their gratitude for the financial supports.
References
[1] E. Stanova, G. Fedorko, S. Kmet, V. Molnar, M. Fabian, Finite element analysis of
spiral strands with different shapes subjected to axial loads, Adv. Eng. Softw.
83 (2015) 45–58.
[2] J. Xie, X. Li, H. Wu, Experimental studies on the axial-compression performance
of concrete at cryogenic temperatures, Constr. Build. Mater. 72 (2014) 380–
388.
[3] D.L. Gautier, K.J. Bird, R.R. Charpentier, et al., Assessment of undiscovered oil
and gas in the arctic, Science. 324 (5931) (2009) 1175–1179.
[4] A. Palmer, K. Croasdale, Arctic Offshore Engineering, World Scientific
Publishing, Singapore, 2013.
[5] B.B. Kanbur, L.M. Xiang, S. Dubey, F.H. Choo, F. Duan, Cold utilization systems
of LNG: a review, Renew. Sust. Energ. Rev. 79 (2017) 1171–1188.
[6] J.B. Yan, J. Xie, Experimental studies on mechanical properties of steel
reinforcements under cryogenic temperatures, Constr. Build. Mater. 151
(2016) 661–672.
[7] M. Elices, H. Corres, J. Planas, Behavior at cryogenic temperatures of steel for
concrete reinforcement, ACI Struct. J. 83 (3) (1986) 405–411.
[8] D. Lahlou, K. Amar, K. Sala, Behavior of the reinforce concrete at cryogenic
temperatures, Cryogenics. 47 (2007) 517–525.
[9] J.B. Yan, J.Y.R. Liew, M.H. Zhang, J.Y. Wang, Mechanical properties of normal
strength mild steel and high strength steel S690 in low temperature relevant
to Arctic environment, Mater. Des. 61 (2014) 150–159.
[10] A. Filiatrault, M. Holleran, Stress-strain behaviour of reinforcing steel and
concrete under seismic strain rates and low temperature, Mater. Struct. 34 (5)
(2001) 235–239.
[11] J. Planas, H. Corres, M. Elices, Behaviour at cryogenic temperatures of tendon
anchorages for prestressing concrete, Mater. Struct. 21 (1988) 278–285.
[12] A.E.H. Love, A Treatise on the Mathematical Theory of Elasticity, fourth ed.,
Dover Publications, New York, 1944.
[13] S. Machida, A.J. Durelli, Response of a strand to axial and torsional
displacements, J. Mech. Eng. Sci. 15 (1973) 241–251.
[14] R.H. Knapp, Derivation of a new stiffness matrix for helically armoured cables
considering tension and torsion, Int. J. Numer. Meth. Eng. 14 (1979) 515–520.
[15] G.A. Costello, Theory of wire rope, 2nd ed., Springer, New York, 1997.
[16] G.A. Costello, R.E. Miller, Lay effect of wire rope, J. Eng. Mech. Div. 105 (4)
(1979) 597–608.
[17] K. Kumar, J.J. Cochran, Closed-form analysis for elastic deformations of multi-
layered strand, J. Appl. Mech. 54 (1987) 898–903.
[18] W.G. Jiang, M.S. Yao, J.M. Walton, A concise finite element model for simple
straight wire rope strand, Int. J. Mech. Sci. 41 (1999) 143–161.
[19] W.G. Jiang, J.L. Henshall, J.M. Walton, A concise finite element mode for three-
layered straight wire rope strand, Int. J. Mech. Sci. 42 (2000) 63–86.
[20] Y.J. Yu, Z.H. Chen, H.B. Liu, X.D. Wang, Finite element study of behavior and
interface force conditions of seven-wire strand under axial and lateral loading,
Constr. Build. Mater. 66 (2014) 10–18.
[21] A.B.M. Abdullah, Jennifer A. Rice, H.R. Hamilton, Gary R. Consolazio, An
investigation on stressing and breakage response of a prestressing strand
using an efficient finite element model, Eng. Struct. 123 (2016) 213–224.
[22] R. Judge, Z. Yang, S.W. Jones, G. Beattie, Full 3D finite element modelling of
spiral strand cables, Constr. Build. Mater. 35 (2012) 452–459.
[23] F. Foti, A.L. Roseto, Analytical and finite element modelling of the elastic–
plastic behaviour of metallic strands under axial–torsional loads, Int. J. Mech.
Sci. 115–116 (2016) 202–214.
[24] ASTM, A370-13. Standard Test Methods and Definitions for Mechanical Testing
of Steel Products, ASTM International, West Conshohocken, United States,
2013.
[25] L. Xiang, H.Y. Wang, Y. Chen, Y.J. Guan, L.H. Dai, Elastic-plastic modeling of
metallic strands and wire ropes under axial tension and torsion loads, Int. J.
Solids. Struct. 129 (2017) 103–118.
[26] W.S. Utting, N. Jones, The response of wire rope strands to axial tensile loads-
Part I. Experimental results and theoretical predictions, Int. J. Mech. Sci. 29 (9)
(1987) 605–619.