Materials 13 03842 v2

materials
Article
Determination of the Bending Properties of Wire
Rope Used in Cable Barrier Systems
Dawid Bruski
Department of Mechanics of Materials and Structures, Faculty of Civil and Environmental Engineering,
Gdańsk University of Technology, Gabriela Narutowicza 11/12, 80-233 Gdańsk, Poland; [email protected];
Tel.: +48-58-348-6149

Received: 19 June 2020; Accepted: 27 August 2020; Published: 31 August 2020
Abstract: This paper presents research on the bending properties of 3 × 7 19-mm wire rope commonly
used in road cable barriers. A total of 19 experimental tests were conducted. In addition, two nonlinear
3D numerical models of the wire rope using beam and solid finite elements were developed. Based on
these models, four numerical simulations were carried out. The numerical results were validated
against the experimental ones and a very good agreement was obtained. The main result of the
research is the determination of the moment–curvature relationship for the wire rope considered.
The effect of prestretching on the rope performance is discussed. The numerical results are analyzed
in this paper in detail, including the behavior of the wire rope under bending and analyses of the
cross-sectional and contact stresses. Suggestions concerning the type of finite element for wire rope
modeling are also given. The results can be used, for example, in numerical simulations of crash tests
of cable barriers.
Keywords: 3 × 7 wire rope; cable barrier; moment–curvature relationship; bending properties;

modeling; numerical simulations; FEM; experimental tests
1. Introduction
Wire ropes are in widespread use as structural members. A wire rope mostly consists of strands
helically twisted around the central core. The strand is constituted of many wires helically wrapped
around the inner wire. This implies that the rope geometry can be complicated [1]. Wire ropes are
generally used to transmit tensile forces. They have high load-bearing capacities upon relatively small
dead weights [2]. Moreover, the rope structure enables the wire ropes to resume loads despite the
break of one or more wires [1]. Wire ropes are commonly used in civil and mechanical engineering,
e.g., in bridges, cranes, cable cars, mine hoists, elevators, prestressed concrete structures, road safety
equipment, and other similar applications [3–7].
Due to the increasing usages of ropes, numerous works investigating their performance have
been published. They present, among other aspects, experimental tests [8], numerical tests [9,10],
fatigue analysis [11], performances in various environments and conditions, and the behavior of partly
damaged [1] or corroded wire ropes [5]. Developing a mathematical model for predicting wire rope
responses is challenging due to the complex architecture of ropes, particularly if the effect of contact
and frictional forces between wires is to be considered [12]. Most mathematical models are limited
to a single, straight strand, where all wound wires are single helices [3,13,14]. However, most actual
ropes consist of several or more strands in which most wires are configured as double helices [3,15].
To analyze ropes with a complex topology and to include various factors in calculations, the Finite
Element Method (FEM) is increasingly being employed [16,17].
One analytical model was proposed by Costello (1997) [13]. It describes the equilibrium of
a single wire, as well as static responses of a single strand and wire rope. Argatov (2011) [3] studied
Materials 2020, 13, 3842; doi:10.3390/ma13173842 www.mdpi.com/journal/materials

x FOR PEER REVIEW
Materials 2020, 13, 3842 2 of 22
the response of a wire rope strand to axial and torsional loads. The refined discrete mathematical
the response
model of a wire rope
was developed strand toassess
to correctly axial and torsional loads.
its deformation, Theinto
taking refined discrete
account mathematical
interwire contacts. model
This
was developed to correctly assess its deformation, taking into account interwire
model enables local contact stresses to be obtained. Argatov et al. (2011) [4] studied the wear contacts. This model
enables local contact
degradation of wirestresses to be obtained.
ropes subjected Argatov
to cyclic et al.over
bending (2011)
a [4] studied
sheave by the wear degradation
applying the Archard– of
wire ropes subjected to cyclic bending over a sheave by applying the Archard–Kragelsky
Kragelsky wear law. The proposed mathematical model can be useful for estimating the fatigue life wear law.
The proposed
of wire ropes. mathematical model can be useful for estimating the fatigue life of wire ropes.
Nowadays,
Nowadays, numerousnumerousresearch researchemploying
employing numerical
numerical simulations
simulations is being
is being conducted.
conducted. FEM
FEM simulations make it possible to analyze the response of wire ropes in detail,
simulations make it possible to analyze the response of wire ropes in detail, even for a complicatedeven for a complicated
geometry.
geometry. The The most
most valuable
valuable simulations
simulations are are those
those compared
compared and and validated
validated against
against experiments.
experiments.
Generally, numerical models of wire ropes can be divided into 2D (simple) [2] and
Generally, numerical models of wire ropes can be divided into 2D (simple) [2] and 3D (more complex) 3D (more complex)
formulations [12,18]. 3D models are more common and universal. These
formulations [12,18]. 3D models are more common and universal. These models are usually models are usually developed
using solidusing
developed finite solid
elements
finite(FEs) or beam
elements (FEs)FEs [2]. The
or beam FEssimulations, where solid
[2]. The simulations, FEssolid
where are applied,
FEs are
are high-cost computational methods [2]. There are also so-called mixed models
applied, are high-cost computational methods [2]. There are also so-called mixed models utilizing utilizing solid and
beam FEs at the same time. For instance, the core can be modeled as solid FEs
solid and beam FEs at the same time. For instance, the core can be modeled as solid FEs and the wires and the wires as
beam
as beamFEsFEs [19]. Advanced
[19]. Advancedand anddetailed
detailed3D 3Dmodels
models(Figure
(Figure1a)1a)are
aretypically
typically used
used toto analyze
analyze short
short
sections of ropes because they require much work and computational resources.
sections of ropes because they require much work and computational resources. For analyses For analyses of a rope
of a
of a considerable
rope of a considerable length (e.g.,
length several
(e.g., dozen
several dozenmeters,
meters,Figure
Figure 1b),
1b),simplified
simplifiedmodels
modelsareare commonly
commonly
employed
employed [11,20].
[11,20].
(a) (b)
1. Wire rope models:
Figure 1. models: (a) Detailed 3D model (for short section); (b) simplified (for long rope
section, here used in
in a cable
cable barrier
barrier model).
model).
Previously,
Previously,numerical
numericalanalyses
analyseswere
werelimited
limitedtotosimple
simplestraight
straightstrands;
strands; currently,
currently, they
theyareare
used to
used
examine
to examine geometrically complex
geometrically ropes.ropes.
complex Jiang Jiang
et al. (2000) [16] developed
et al. (2000) a numerical
[16] developed model of model
a numerical a wire rope
of a
utilizing solid FEs. The elaborated model showed better agreement with the experimental
wire rope utilizing solid FEs. The elaborated model showed better agreement with the experimental results from
the literature
results thanliterature
from the the analytical model
than the proposed
analytical by Costello
model proposedin 1997. Elata etinal.1997.
by Costello (2004) [15]etdeveloped
Elata al. (2004)
a[15]
newdeveloped a new model for simulating the mechanical response of a wire rope subjectedload
model for simulating the mechanical response of a wire rope subjected to both an axial and
to both
an axial torque. The model was validated against experimental data collected from
an axial load and an axial torque. The model was validated against experimental data collected from the testing of two cables
and compared
the testing to Velinski’s
of two cables andandcompared
Costello’sto
model. Erdönmez
Velinski’s et al. (2009)
and Costello’s [21] numerically
model. Erdönmez studied the axial
et al. (2009) [21]
loading and bending of simple strand wire over a sheave. The results showed that
numerically studied the axial loading and bending of simple strand wire over a sheave. The results the maximum stress
occurred
showed thatover the
the upper
maximum sheave midpoint.
stress Judge
occurred et al.
over the(2011)
upper[22] conducted
sheave experimental
midpoint. Judge etand numerical
al. (2011) [22]
research on a cable consisting of 120 wires subjected to the impact of a 20-mm
conducted experimental and numerical research on a cable consisting of 120 wires subjected to the fragment with velocities
between
impact of200 and 1400
a 20-mm m/s. The
fragment numerical
with velocities simulation
between showed
200 and good agreement
1400 m/s. with the simulation
The numerical laboratory
tests.
showed Boroška et al. (2014)
good agreement [23]
with described
the laboratory thetests.
influence
Boroškaof various factors
et al. (2014) [23]on Young’sthe
described modulus
influenceof
of various factors on Young’s modulus of the wire ropes and gave some values. They also numerically
Materials 2020, 13, 3842 3 of 22
Materials 2020, 13, x FOR PEER REVIEW 3 of 22

the wire ropes and gave some values. They also numerically investigated the 100-mm long sample and
presented
investigated the the
stress distributions
100-mm within the
long sample andrope. Wu (2014)
presented the [10]stressdeveloped a numerical
distributions withinmodel the rope. of a wire
Wu
rope
(2014)subjected to tensile
[10] developed loads andmodel
a numerical compared the results
of a wire to theoretical
rope subjected and experimental
to tensile loads and compared data taken the
from
resultstheto literature.
theoretical andFoti experimental
et al. (2016) [24]datainvestigated
taken from the the behaviorFoti
literature. of aetstrand
al. (2016)under axial–torsional
[24] investigated the
loads using a new analytical formulation and FEM simulations. The
behavior of a strand under axial–torsional loads using a new analytical formulation and FEM simulations. results were validated against
experimental
The results were andvalidated
theoretical results
against from the literature.
experimental The effects
and theoretical of wire
results fromlay theangles and The
literature. the torsional
effects of
boundary were discussed. Karathanasopoulos et al. (2017) [2]
wire lay angles and the torsional boundary were discussed. Karathanasopoulos et al. (2017) [2]proposed a 2D elastoplastic FE proposed
model to
simulate the behavior
a 2D elastoplastic of a simple
FE model strand.the
to simulate The advantage
behavior of a of this simplified
simple strand. The planar modelof
advantage isthis
its robustness
simplified
and
planar model is its robustness and low computational cost. The results were compared withal.the
low computational cost. The results were compared with the 3D FE model of Foti et [24]3DandFE
amodel
very good agreement was obtained. Vukelic et al. (2017) [12] numerically
of Foti et al. [24] and a very good agreement was obtained. Vukelic et al. (2017) [12] numerically investigated the influence
of a reductionthe
investigated of influence
the cross-sectional area on
of a reduction of stress levels and thearea
the cross-sectional remaining
on stress fatigue
levelslifeand forthethree design
remaining
types of wire ropes. The results showed that the reductions in cross-sectional
fatigue life for three design types of wire ropes. The results showed that the reductions in cross-sectional areas below 90% of
the initial area cause a substantial increase in stress levels and significant
areas below 90% of the initial area cause a substantial increase in stress levels and significant decreases in decreases in the remaining
life
the of the wirelife
remaining ropes.
of theWang
wireetropes.
al. (2017)
Wang [17]
et developed
al. (2017) [17] an developed
FE model of anwire rope and
FE model of wireshowed rope that
and
stresses are unevenly distributed within the rope. They also conducted
showed that stresses are unevenly distributed within the rope. They also conducted experimental tensile experimental tensile tests and
compared the resultsthe
tests and compared with the simulation.
results The calculations
with the simulation. showed good
The calculations showed agreement with physical
good agreement with
tests.
physical tests. Smyth (2018) [11] analyzed a wire rope used in an overhead crane and, becausethe
Smyth (2018) [11] analyzed a wire rope used in an overhead crane and, because theactual
actual
bending
bending stiffness
stiffness properties
properties werewere not
not known,
known, the the author
author applied
applied data data from
from another
another wire wire rope
rope typetype and
and
appropriately
appropriately scaled scaledititusing
usingthethe relative
relative stiffness
stiffness ratiosratios
betweenbetween the ropes.
the ropes. The papers
The papers cited above citedindicate
above
indicate
that boththat both theoretical
theoretical and numerical
and numerical modelsmodels
still need stilltoneed to be developed,
be developed, tested,tested,
and then and refined.
then refined.It is
It is noteworthy that publications dealing with the bending properties
noteworthy that publications dealing with the bending properties of wire ropes are still limited. of wire ropes are still limited.
In
In this
this paper,
paper, the the subject
subject of of the
the analysis
analysis is is aa 33 ×× 77 19-mm
19-mm wire wire rope,
rope, which
which is is commonly
commonly used used inin
road cable barrier systems (Figure 2) [25–27]. Descriptions of the cable
road cable barrier systems (Figure 2) [25–27]. Descriptions of the cable barrier systems and examples barrier systems and examples of
their use are
of their usepresented
are presented in previous reports reports
in previous [28–32]. [28–32].
Knowledge about thisabout
Knowledge wire ropethis andwire itsrope
properties
and its is
crucial, helping to better understand and analyze vehicular impacts
properties is crucial, helping to better understand and analyze vehicular impacts on cable barriers.on cable barriers. So far, research has
mainly focused on
So far, research hasevaluations
mainly focused of theon performance
evaluationsofofthe thewhole safety system
performance of the[33,34].
whole safetyTo the systembest of
the author’s
[33,34]. To the knowledge,
best of the the only studies
author’s knowledge, on the theproperties
only studies of the × 7properties
on3the 19-mm wire rope
of the 3 ×have been
7 19-mm
published by the team led by Reid [20,35,36]. To determine the
wire rope have been published by the team led by Reid [20,35,36]. To determine the bending bending properties, they tested the wire
rope in a cantilever
properties, they testedposition. The longest
the wire rope intested rope was
a cantilever 1.359 m The
position. long,longest
and thetested
maximum rope applied
was 1.359 point
m
load was approximately 36.5 N. The physical tests were supplemented
long, and the maximum applied point load was approximately 36.5 N. The physical tests were with numerical simulations.
The rope was modeled
supplemented in a simplified
with numerical manner,
simulations. Theusing
rope was onlymodeled
one beaminFE in the wiremanner,
a simplified rope cross-section,
using only
because these studies were concentrated on the development
one beam FE in the wire rope cross-section, because these studies were concentrated on of a wire rope computational modelthe
which can be used in applications of a considerable length. It should
development of a wire rope computational model which can be used in applications of a considerable be mentioned that in the literature,
there
length. areItno studies
should beon four-pointthat
mentioned bending
in theofliterature,
this rope. there Due to arethe nosmall amount
studies of experimental
on four-point bending data
of
and the lack of numerical analyses using advanced 3D models,
this rope. Due to the small amount of experimental data and the lack of numerical analyses using further research on 3 × 7 19-mm wire
rope is desirable.
advanced 3D models, further research on 3 × 7 19-mm wire rope is desirable.
Figure 2.
Figure 2. Examples
Examples of
of safety
safety barriers
barriers using
using 33 ×
× 77 19-mm
19-mm wire
wire rope.
rope.
This work
This workaimsaimstotodetermine
determinethethe bending
bending properties
properties of 3the
of the × 7319-mm
× 7 19-mm wire rope.
wire rope. The result
The main main
result
of of the research
the research is the determination
is the determination of the moment–curvature
of the moment–curvature relationship.
relationship. The objectives
The other other objectives
are to
are to analyze
analyze therope
the wire wirebehavior
rope behavior
under aunder a bending
bending condition,
condition, to investigate
to investigate the interwire
the interwire contactcontact
and to
and to test whether the prestretching affects the rope bending performance. The practical aspects of
the article are to determine and explicitly deliver equations for recreating the actual 3D geometry of
the rope, to present the methodology of developing advanced numerical rope models, and to give
suggestions concerning the use of beam or solid FEs. To achieve the assumed objectives, 19
Materials 2020, 13, 3842 4 of 22
test whether the prestretching affects the rope bending performance. The practical aspects of the article
are to determine and explicitly deliver equations for recreating the actual 3D geometry of the rope,
to present
Materials 2020,the13,methodology
x FOR PEER REVIEW of developing advanced numerical rope models, and to give suggestions 4 of 22
concerning the use of beam or solid FEs. To achieve the assumed objectives, 19 experimental tests
experimental
were carried out, tests werewere
which carried out, which
supported were
by four supported
nonlinear FEM byanalyses.
four nonlinear FEM analyses.
With reference With
to published
referenceon
research to published
the 3 × 7 19-mmresearch on the
wire rope, 3 ×this
7 19-mm
studywire rope,the
presents thisresults
study of presents the results
four-point bending of four-
tests
pointthe
and bending
effect of tests and the effect
prestretching of prestretching
on the rope bendingon the rope and
properties bending properties
delivers and delivers
the equations enabling the
equations enabling one to recreate the 3D 3 × 7 wire rope geometry.
one to recreate the 3D 3 × 7 wire rope geometry. Furthermore, conducted numerical simulations will Furthermore, conducted
numerical simulations
supplement knowledge will
onsupplement
the contact knowledge
stress values onandthe their
contact stress values
distribution and their
within distribution
the rope, as well
within
as the rope,ofasthe
the influence well as thecoefficient
friction influencebetweenof the friction
the wires coefficient
on the ropebetween the wiresThe
performance. on previous
the rope
performance.
models of thisThe previous
rope, collected models of this rope,
and published collected
in [36], do notand published
allow for theinanalysis
[36], doof nottheallow for the
mentioned
analysisasofthey
issues, the were
mentioned
developedissues,
foras they were
a different developed
purpose. Thisfor a different
research purpose.because
is important This research
the wire is
important
rope analyzed because the in
is used wire rope
road analyzed
cable barriers is used
and itsin properties
road cable affect
barriers
theand its properties
effectiveness affect
of the wholethe
effectiveness
safety system,ofthus the whole
affect safety system,
the safety thus affect
of vehicle the safety
occupants of vehicle
during occupants
an accident. duringinto
Taking an accident.
account
Taking
the smallinto account
amount of the small amount
available experimental of available experimental
and numerical data,andit isnumerical
advisabledata, it is advisable
to conduct further
to conduct further studies allowing for a better understanding of the
studies allowing for a better understanding of the wire rope behavior. This data could potentially wire rope behavior. This data
couldsuggestions
give potentially for givethesuggestions
optimization forof the
theoptimization
3 × 7 wire rope,of thefor3 instance,
× 7 wire rope,
in termsfor of
instance, in terms
the dimensions
of the
and dimensions
geometry. Theand geometry.
study will also Thegive
study will also give
suggestions suggestions experimental
for conducting for conducting experimental
and numerical
and numerical
bending tests forbending
wire ropes.testsThe
for remainder
wire ropes.ofThe this remainder of this paper
paper is organized is organized
as follows: Section 2asdescribes
follows:
Section
the 2 describesand
characteristics the the
characteristics
geometry ofand thethe
wiregeometry of the 3wire
rope; Section rope;the
presents Section 3 presents the
moment–curvature
relationship from the Bernoulli–Euler beam theory; Section 4 deals with the four-point with
moment–curvature relationship from the Bernoulli–Euler beam theory; Section 4 deals the four-
experimental
point experimental
bending bendingthe
tests, introducing tests,
testintroducing the test stand
stand and specimens and and specimens
presenting theand presenting
results; Section the5 results;
shows
Section
the 5 shows the
computational computational
models developedmodelsand andeveloped
analysis ofand an analysis
numerical of numerical
simulations; simulations;
in Section 6, the final in
Section 6, the final moment–curvature relationship is proposed and
moment–curvature relationship is proposed and the research is discussed; and Section 7 contains the research is discussed; and
Section
the summary7 contains the summary and conclusions.
and conclusions.
× 77 19-mm
2. The 3 × 19-mm Wire
Wire Rope
Rope
2.1. Characteristics of
2.1. Characteristics of the
the Wire
Wire Rope
Rope
The
The subject
subject of
of this
this study
study is steel 33 ×
is aa steel × 77 wire
wire rope
rope (Figure
(Figure 3),
3), which
which is
is commonly
commonly used
used in
in road
road
cable barriers. The wire rope consists of three strands and each strand consists
cable barriers. The wire rope consists of three strands and each strand consists of seven 3 mmof seven 3 mm diameter
steel wires;steel
diameter thus,wires;
the whole
thus,wire
the rope
whole consists
wire of 21 wires
rope in total.
consists of 21The nominal
wires ropeThe
in total. diameter is 19rope
nominal mm
(0.75 in.). is
diameter The 19 wire rope geometry
mm (0.75 is described
in.). The wire in Section
rope geometry 2.2.
is described in Section 2.2.
Figure 3.
Figure 3. 3 ×
× 77 19-mm
19-mm wire
wire rope
rope used in road cable barriers.
The author’s
author’s laboratory
laboratorytests
testshave
haveshown
shownthat
thatthetherupture
ruptureload of of
load oneone
wire is approximately
wire is approximately10 kN
10
and for one strand (i.e., seven wires), is approximately 70 kN. Based on these values, it
kN and for one strand (i.e., seven wires), is approximately 70 kN. Based on these values, it can be can be concluded
that the breaking
concluded load
that the of the wire
breaking loadrope should
of the wirebe approximately
rope 210 kN. This value
should be approximately 210 can
kN.be confirmed
This in
value can
the literature, in
be confirmed e.g.,
thethe publications
literature, [20,35,37]
e.g., the show that
publications the load-carrying
[20,35,37] capacity
show that the × 7 19-mm
of the 3 capacity
load-carrying of
wire rope obtained from quasi-static tensile tests is approximately 200 kN. The
the 3 × 7 19-mm wire rope obtained from quasi-static tensile tests is approximately 200 kN. The Young’s modulus of
the non-prestretched
Young’s modulus ofand theprestretched wire rope
non-prestretched andis 79.9 and 116.1 wire
prestretched GPa, respectively
rope is 79.9[35].
andThe author’s
116.1 GPa,
respectively [35]. The author’s experimental tests confirmed these values. Moreover, the report [29]
states that an effective modulus of non-prestretched rope is approximately 76–90 MPa and that the
effective modulus of the wire rope in safety barriers can increase by almost 50% after two years of
installation due to cyclic loading from temperature variations and proper field monitoring of the
system. The cross-sectional area of the wire rope is 148.4 mm2. The material density is equal to 7948
3
Materials 2020, 13, 3842 5 of 22
experimental tests confirmed these values. Moreover, the report [29] states that an effective modulus
of non-prestretched rope is approximately 76–90 MPa and that the effective modulus of the wire rope
in safety barriers can increase by almost 50% after two years of installation due to cyclic loading from
temperature variations and proper field monitoring of the system. The cross-sectional area of the wire
rope is 148.4 mm2 . The material density is equal to 7948 kg/m3 [35].
Observations from real road accidents have revealed that cable barriers, whose main longitudinal
structural elements are wire ropes, are capable of containing vehicles exceeding design loads of the cable
barrier system [29] and rope fracture rarely occurs in cable barrier systems [35]. The values of the forces
in wire ropes during a road accident mainly depend on the speed, impact angle, mass, and geometry of
the vehicle. The maximum wire rope tensile load during a full-scale crash test, when a 2034-kg Chevrolet
C2500 impacted a 3-cable barrier system at 98.1 km/h and 26.2◦ , was 108 kN [20,35]. The maximum
load recorded during an experimental test of the perpendicular impact of a 772-kg boogie vehicle on
a single wire rope at the speed of 24 km/h was 165 kN [20]. Numerical simulations of the impact of
a 1500-kg passenger car on a 4-cable barrier at the speed of 110 km/h and at the angle of 20◦ revealed
that the forces in the rope can reach 76 kN [38]. The striking of a Heavy Goods Vehicle (HGV) weighing
38 tons into a 3-cable barrier at the speed of 65 km/h and at the impact angle of 7◦ can exert forces on
one wire rope of up to 50 kN [39]. These values are below the rupture load of the wire rope.
2.2. Geometry of the Wire Rope

One of the basic ropes is a straight strand (1 + 6), which was analyzed, for instance, in [1,2,21,24].
This simple strand consists of a straight center wire surrounded by six helical wires. An example of
this strand is shown in Figure 4a. The straight core is marked in a yellow color, and the outer wires are
depicted in red (red color in Figure 4 means that the wire geometry is described by circular helices).
The analyzed 3 × 7 wire rope has a more complicated geometry and is made of three aforementioned
strands spirally wrapped around itself, as can be seen in Figure 4b. Therefore, the centerline of the inner
wire of each strand is defined through single helices (red color), in the same way as the outer wires of
the 1 + 6 strand. Around each inner core wire lays a layer of six outer wires described by a double
helix geometry marked in a gray color.
In the description of the wire rope geometry, it was assumed that in a plane perpendicular to the
wire rope, the wire has a circular cross-section with the diameter Dw . However, it has to be noted
that the actual cross-section of the wire is elliptical in a plane normal to the rope, see, e.g., [3,13,24].
Herein, differences resulting from this simplification should be negligible and a similar assumption
can be found in the paper [35]. In the first step, the geometry of a single strand was determined in
a three-dimensional x’y’z’ coordinate system, where the z’-axis coincides with the centroidal wire
(Figure 4c). The analyzed wire rope consists of three such strands, so, in order to determine the final
geometry, a new xyz coordinate system was assumed, in which the z-axis is in the longitudinal direction
of the wire rope. In this coordinate system, the angle between each strand is 120 degrees. To determine
the final coordinates, the points x’, y’, and z’ have to be transformed from the x’y’z’ systems of each
strand into the xyz global system, as shown in the cross-section in Figure 4d. In this figure, for better
clarity, the strands are separated from each other.
Materials 2020, 13, 3842 6 of 22
Figure 4. Geometry of the wire rope: (a) Simple 1 + 6 strand; (b) 3 × 7 wire rope structure; (c) single
Figure 4. Geometry of the wire rope: (a) Simple 1 + 6 strand; (b) 3 × 7 wire rope structure; (c) single
strand cross-section; (d) cross-section of a wire rope consisting of three strands.
strand cross-section; (d) cross-section of a wire rope consisting of three strands.
The final coordinates x, y, and z of the points of the helicoidal centerlines of the inner wire for
The final
each strand (N coordinates x, y, and on
= 1, 2, 3), depending z ofthe
the points oft,the
parameter arehelicoidal centerlines
given by the of expressions:
following the inner wire for
each strand (N = 1, 2, 3), depending on the parameter t, are given by the following expressions:
xN (t) = R1 cos(βN (t)) (1)
xN  t   R1 cos   N  t   (1)
yN (t) = R1 sin(βN (t)) (2)
y N  t   R1 sin   N  t   (2)
zN ( t ) = t (3)
The final coordinates x, y, and z of the points  t the

z Nof   tdouble-helical centerlines for each of the outer
(3)
wires (n = 1, 2, 3, 4, 5, 6) of each strand (N = 1, 2, 3), depending on the parameter t, are defined as
The final coordinates x, y, and z of the points of the double-helical centerlines for each of the outer
5, (6)
wires (n = 1, 2, 3, x4,N,n t)of (R2 cos
=each (αn ((N
strand t))=+1,R2,1 )3),
cos (βN (t)) − on
depending R2 sin αn (t)) sin(t,
the(parameter βNare
(t))
defined as (4)
x (t) t=
xN,n
N ,n 2 
cos(αn (t))
 (RR2 cos n   
t  + R
R1 ) cos
1
βN (tt))
cos(

 −RR2sin
N 2  n  
sin(αn(tt))sin
sin(βN(tt))
N  (5)
(4)
zN,n (t) = t (6)
xN ,n  t    R cos   t    R  cos    t    R sin   t   sin    t  
2 n 1 N 2 n N (5)
The angles in the above formulas are defined by the following expressions:
z N ,n  t   t (6)
π 2π
αn (t) = (n − 1) + t + ϕ2 (7)
3 P2
The angles in the above formulas are defined by the following expressions:
2π 2π
βN ( t ) = (N − 1) + t + ϕ1 (8)
 
 n  t    n  1  1t  2
3 2P
(7)
3 P2
Materials 2020, 13, 3842 7 of 22
In the above equations, N is the number of strands in the rope; n indicates the number of wires
within the strand; R1 is the distance from the origin of the xyz global system to the center of the inner
wire of the strand (initial radius of the helix of the inner wire of the strand); R2 is the distance from
the center of the inner wire to the center of the outer wire in the strand (initial radius of the helix
of the outside wire in the strand); ϕ1 denotes the initial rotation angle of the strand; ϕ2 stands for
the initial rotation angle of the outer wire in the strand; and P1 and P2 are the pitch of the strand and
the pitch of the single wire, respectively. The symbols are also shown in Figure 4c,d. The equations
were used to construct the geometry of a straight section of the wire rope (see, e.g., the 3D model
in Figure 1a).
The presented geometry description applies to a straight section of the 3 × 7 wire rope. In fact,
the wire rope has an additional initial curvature, because of, e.g., its storage in rolls. The initial
curvature is taken into account in experimental and numerical investigations.
3. Moment–Curvature Relationship
The moment–curvature relationship in beam theory is described in many books of mechanics
of materials, e.g., [40,41]. Bernoulli’s assumption of flat sections is often taken into account since,
in engineering applications, it is commonly fulfilled. Nevertheless, it cannot be applied to the analysis
of a wire rope because the relative motion of wires and strands is possible and the strain field in
a cross-section is not a smooth function.
It is assumed that some function of strains is defined over the cross-section domain εx = εx ( y, z)
so that the integral of this function over the domain can be calculated. A strain is directly related to
stress by Hooke’s law, which can be expressed mathematically as
1
εx ( y, z) = σx ( y, z) (9)
E
where E is Young’s modulus.
From equations of statics it can be obtained that the moment resulting from the normal stresses σx
acting over the cross-section is equal to the bending moment M:
x
σx zdA = M (10)
A
The substitution of expression (9) into (10) allows the bending moment M to be defined as
a function of the material parameter and strain function:
x x E
M= σx zdA = E εx zdA = I (11)
ρ
A A
s
where I = ρ εx zdA and ρ is the beam’s curvature radius. I can be denoted as the moment of inertia of
A
the cross-sectional area if Bernoulli’s assumption is fulfilled. In the current research, I denotes a certain
constant expressed in mm4 determined experimentally.
Taking into account that the curvature κ is defined as the reciprocal of the radius of curvature ρ,
Equation (11) can be rewritten as
1 M
κ= = (12)
ρ EI
Under a constant bending moment, the deflection of the beam is described by the equation of
the circle. The exact equation for the curvature of the beam, which expresses curvature through
Materials 2020, 13, 3842 8 of 22
derivatives of the transverse displacement function u (the prime mark represents the derivative with
respect to axial coordinate x), is
1 u00
κ= = h (13)
ρ i3
1 + (u0 )2 2
4. Experimental Bending Testing
4.1. Test Specimens

Four-point bending tests were conducted for four 1.0 m long 3 × 7 wire ropes. The ropes were
denoted as WR1, WR2, WR3, and WR4, respectively. Five bending tests were carried out for WR1,
WR3, and WR4 wire ropes, and four tests for the WR2 rope, so the total number of tests was 19.
As a consequence of this, the tests were denoted as follows: WRX-TY, where X refers to the wire rope
number and Y indicates the test number.
The following specimens WR1, WR2, and WR4 were not prestretched before the bending test.
The WR3 rope was initially prestretched to the force of 162 kN. The initial radius of curvatures prior to
testing for WR1, WR2, WR3, and WR4 samples was 3693.5, 3145.0, 15,629, and 2723.3 mm, respectively.
The large radius of curvature for the WR3 sample is due to the prestretching. In [20], in which a similar
rope was tested, the initial radius of curvature was 1708.4 mm.
4.2. Test Stand

The testing equipment and experimental setup for the 4-point bending of wire rope are shown
in Figures 5 and 6. The bending test was conducted using a 400-kN ZwickRoell Universal Testing
Machine (UTM). In the conducted studies, two test stands were applied, as depicted in Figure 6. In the
first test configuration (TC1), the distance between the supports equaled A = 80 cm and the distance
between the two loading points was half of the distance between the supports and equal to B = 40 cm
(Figure 6a). For the second configuration (TC2), the distances were as follows: A = 57 cm and B = 17 cm,
respectively (Figure 6b). The tests were performed under constant deformation control with the speed
of 20 mm/min. The diameters of the loading pins and the supporting pins were the same and equal to
80 mm.2020, 13, x FOR PEER REVIEW
Materials 9 of 22
(a) (b)
Test stand: (a)
Figure 5. Test (a) General
General view;
view; (b)
(b) displacement
displacement transducer.
transducer.
(a) (b)
Materials 2020, 13, 3842 9 of 22
Figure 5. Test stand: (a) General view; (b) displacement transducer.
(a) (b)
Figure 5. Test stand: (a) General view; (b) displacement transducer.
Figure6.6.Test
Figure Testconfigurations:
configurations:(a) TC1:AA==80
(a)TC1: cm,BB== 40 cm; (b) TC2: A =
80cm, cm, BB ==17
57 cm,
= 57 17cm.
cm.
The wire ropes were equipped with a displacement transducer in the middle of the span.
The inductive transducer was fixed to a spreader beam and moved with it during the test. In this
way, the deflection of the rope in the middle of the span relative to the loading pins was measured.
The method of the displacement measurement is shown in Figure 7. The displacements were calculated
as uf − ui . This measurement was used to determine wire rope curvatures, assuming that, between the
loading points, the curvature was described by the equation of the circle. Because of the irregular
geometry of the wire rope, the transducer had to be laid on the specimen via a stabilizing pipe,
which prevented
Figure 6.the transducer’s
Test head
configurations: (a) from
TC1: Asliding onB the
= 80 cm, = 40rope (Figure
cm; (b) TC2: A5b).
= 57 cm, B = 17 cm.
Figure 7. Measurement with a displacement transducer.
4.3. Test Results

Displacement-force curves were obtained from the testing machine. Deflections at the wire rope
mid-span relative to the points where the load was applied were measured by the transducer. The
results for all specimens are presented in Figures 8–11. Considering the first TC1 configuration (WR1-
WR3), the forces corresponding to the largest displacements are within the range of approximately
200–290 N. Wire rope deflections obtained from the displacement transducer are approximately 30–
35 mm.
Figure
Figure 7.
For the TC2 configuration Measurement
Measurement
7.(WR4), with
with aacorresponding
the forces displacement
displacement transducer.
transducer.
to the largest displacements are
approximately 400–460 N and the displacements from the transducer do not exceed 16 mm. The first
4.3. Test
For Results
the TC1 configuration, WR1, WR2, and WR3 ropes were tested, and for the TC2 configuration,
test WR4-T1 was conducted for the settings as in the TC1 tests, where the maximum displacement of
WR4 rope was tested. It was decided that the three test samples would be used for tests in the first
Displacement-force curves were obtained from the testing machine. Deflections at the wire rope
configuration to properly observe the initial bending-curvature dependence, in addition to one sample
mid-span relative to the points where the load was applied were measured by the transducer. The
for the second configuration, so as to examine the bending response for larger rope curvatures. In order
results for all specimens are presented in Figures 8–11. Considering the first TC1 configuration (WR1-
not to damage the transducer, the displacement of the UTM traverse was limited to 80 mm for
WR3), the forces corresponding to the largest displacements are within the range of approximately
WR1-WR3 ropes and 110 mm for WR4 rope.
200–290 N. Wire rope deflections obtained from the displacement transducer are approximately 30–
35 mm.
4.3. Test Results
For the TC2 configuration (WR4), the forces corresponding to the largest displacements are
Displacement-force
approximately 400–460 Ncurves
and thewere obtained from
displacements fromthethe
testing machine.
transducer Deflections
do not at mm.
exceed 16 the wire
Therope
first
mid-span
test WR4-T1relative to the pointsfor
was conducted where the loadas
the settings wasin applied
the TC1were
tests,measured
where thebymaximum
the transducer. The results
displacement of
for all specimens are presented in Figures 8–11. Considering the first TC1 configuration (WR1-WR3),
the forces corresponding to the largest displacements are within the range of approximately 200–290 N.
Wire rope deflections obtained from the displacement transducer are approximately 30–35 mm.
WR3, and WR4 wire ropes were 2862.9, 2575.5, 4047.8, and 2079.7 mm, respectively. The bending
response of the prestretched WR3 rope is smoother than the response of the non-prestretched ropes,
but the values of the maximum forces are comparable to the forces obtained in the non-prestretched
WR1 rope test. The displacements measured at the midspan for WR1, WR2, and WR3 samples are
similar. Because of the small differences between the non-prestretched and prestretched samples, it
similar. Because
Materials 2020, of the small differences between the non-prestretched and prestretched samples,
13, 3842 it
10 of 22
was decided to determine one final moment–curvature relationship applicable to both ropes.
was decided to determine one final moment–curvature relationship applicable to both ropes.
(a) (b)
(a) (b)
Figure 8. WR1: Load vs. displacement curve from a Universal Testing Machine (UTM) (a), and
WR1:Load
Figure 8. WR1: Loadvs.vs. displacement
displacement curve
curve fromfrom a Universal
a Universal Testing
Testing Machine
Machine (UTM)(UTM) (a),
(a), and
displacements from a transducer (b).
and displacements
displacements fromfrom a transducer
a transducer (b). (b).
(a) (b)
(a) (b)
Figure 9.9. WR2:
WR2: Load
Load vs.
vs. displacement
displacementcurve
curvefrom
froma aUTM
UTM(a),(a),
and displacements
and from
displacements a transducer
from (b).
a transducer
Figure 9. WR2: Load vs. displacement curve from a UTM (a), and displacements from a transducer
(b).
(b).
(a) (b)
Figure 10. WR3:
WR3: Load vs. displacement
displacementcurve
curvefrom
froma aUTM
UTM(a),
(a), and
and displacements
displacements from
from a transducer
a transducer (b).
(b).
(a) (b)
(a) (b)
Figure
Materials 2020,10.
13, WR3:
3842 Load vs. displacement curve from a UTM (a), and displacements from a transducer
11 of 22
(b).
(a) (b)
Figure 11. WR4: Load vs.
vs. displacement
displacementcurve
curvefrom
froma aUTM
UTM(a),
(a), and
and displacements
displacements from
from a transducer
a transducer (b).
(b).
For the TC2 configuration (WR4), the forces corresponding to the largest displacements are
approximately
5. 400–460Testing
Numerical Bending N and the displacements from the transducer do not exceed 16 mm. The first
test WR4-T1 was conducted for the settings as in the TC1 tests, where the maximum displacement of
In this study, two test configurations were considered (TC1 and TC2, see Figure 6) and for each
the traverse was set to 80 mm. Due to the significant safety margin of the transducer, in the second test,
configuration, two numerical models were developed. In the first, the wire rope was modeled using
WR4-T2, the traverse displacement was increased to 120 mm; however, for the last three tests (WR4-T3
beam FEs (B) and in the second, the rope was made of solid FEs (S). Therefore, the numerical tests
to WR4-T5), it was decided to set the maximum displacement of the traverse to 110 mm.
are denoted as follows: TC1-B (configuration 1: A = 80 cm, B = 40 cm, the wire rope modeled with the
Based on the acquired displacement curves, one can notice that there is a linear relationship for
use of beam FEs), TC1-S (configuration 1, the wire rope modelled using solid FEs), and TC2-B
displacements at the mid-span. A linear relationship can also be observed between the force and
(configuration 2: A = 57 cm, B = 17 cm, beam FEs), TC2-S (configuration 2, solid FEs). The numerical
the displacement measured by UTM, especially if considering the range up to 70% of the maximum
simulations in which the wire rope was modeled using beam FEs will be abbreviated as “beam
displacement. The final radii of curvatures, after conducting all of the bending tests, for WR1, WR2,
model” and those where solid FEs were used will be abbreviated as “solid model”, respectively.
The calculations were conducted using LS-DYNA software [42–44] (MPP double-precision
R10.1.0) on the supercomputer Tryton, managed by the Academic Computer Centre (CI TASK) in
Gdańsk, Poland. The keyword names are used in accordance with LS-DYNA documentation [42–44].
similar. Because of the small differences between the non-prestretched and prestretched samples, it was
5.1. Numerical Model
decided to determine one final moment–curvature relationship applicable to both ropes.
The numerical model of 4-point bending consists of half-round supports and loading pins, the
5. Numerical
traverse, Bending
and the Testing
wire rope (Figure 12). Depending on the model, the following elements are varied:
The spacing of thetwo
In this study, supports and loading points
test configurations (depending(TC1
were considered on the
andconfiguration TC1
TC2, see Figure 6) or
andTC2), and
for each
the wire rope model (beam or solid model).
configuration, two numerical models were developed. In the first, the wire rope was modeled using
beam FEs (B) and in the second, the rope was made of solid FEs (S). Therefore, the numerical tests
are denoted as follows: TC1-B (configuration 1: A = 80 cm, B = 40 cm, the wire rope modeled with
the use of beam FEs), TC1-S (configuration 1, the wire rope modelled using solid FEs), and TC2-B
(configuration 2: A = 57 cm, B = 17 cm, beam FEs), TC2-S (configuration 2, solid FEs). The numerical
simulations in which the wire rope was modeled using beam FEs will be abbreviated as “beam model”
and those where solid FEs were used will be abbreviated as “solid model”, respectively.
The calculations were conducted using LS-DYNA software [42–44] (MPP double-precision R10.1.0)
on the supercomputer Tryton, managed by the Academic Computer Centre (CI TASK) in Gdańsk,
Poland. The keyword names are used in accordance with LS-DYNA documentation [42–44].
5.1. Numerical Model

The numerical model of 4-point bending consists of half-round supports and loading pins,
the traverse, and the wire rope (Figure 12). Depending on the model, the following elements are
varied: The spacing of the supports and loading points (depending on the configuration TC1 or TC2),
and the wire rope model (beam or solid model).
Materials 2020, 13, 3842 12 of 22
12. General
Figure Figure view of the numerical test of 4-point bending.
12. General view of the numerical test of 4-point bending.
Contact between the adjacent

Contact between the adjacent wires wiresand andbetween
between the thewires
wires andand the UTM’s
the UTM’s parts wasparts was modeled
modeled
using a penalty-based mortar contact algorithm [42,45]. In
using a penalty-based mortar contact algorithm [42,45]. In the literature concerning simulations the literature concerning simulations of of wire
wire ropes, one can find that values of the friction coefficient (FC) are mostly from 0.1 to 0.2
ropes, one can find that values of the friction coefficient (FC) are mostly from 0.1 to 0.2 [10,11,16,17,21,22].
[10,11,16,17,21,22]. Considering the above, it was decided to assume that the friction coefficient was
Considering thetoabove,
equal 0.1. Theitimpact
was decided to assume
of the friction thatvalue
coefficient the friction coefficient
will be discussed was equal
in Section 5.2.2. to
The0.1.deadThe impact
of the friction
loadcoefficient
was considered value
in thewill be discussed
numerical simulations. in Section 5.2.2. The dead load was considered in the
Figure 12. General view of the numerical test of 4-point bending.
The nonlinear analyses were solved with the use of the full Newton solution method [42,44,46].
numerical simulations.
In theContact
numerical testing
between theofadjacent
wire ropes, wires in and
the TC1 configuration,
between thethewires the simulation
and the UTM’s time
partswas
was 240 s. For
modeled
The nonlinear analyses were solved with the use of full Newton solution method [42,44,46].
the
usingTC2 a configuration,
penalty-based the timecontact
mortar was equal to 360 s.
algorithm The time
[42,45]. stepliterature
In the for the simulations
concerning employing
simulationsthe of
In the numerical
beam testing
wire rope of wire ropes, in the TC1 configuration, the simulation time was 240 s.
wire ropes, onemodel was that
can find 0.5 s,values
and forofthe calculations
the using the (FC)
friction coefficient solid are
ropemostly
model,from
due to 0.1thetosize
0.2
For the TC2 of configuration,
the computational
[10,11,16,17,21,22]. the time
model, it the
Considering was equal
wasabove, to
1.0 s. itItwas 360
should s. The time
be mentioned
decided to assume thatstep for
that,the the
because simulations
frictionthese numerical
coefficient wasemploying
analyses
the beam wireequal to haveThe
rope
0.1. a static
model character,
was
impact s,theand
0.5friction
of the simulation
for thetime
coefficient does
willnot
calculations
value be correspond
using in
discussed toSection
the physical time
solid5.2.2.
rope Theand
model,
deadis due to
simply a proportional
loadcomputational load
was considered inmodel, multiplier.
the numerical Since the global displacement in simulations was considered
the size of the it wassimulations.
1.0 s. It should be mentioned that, because these numerical
to be The
large, the displacement
nonlinear analyses werenormsolved for thewithconvergence
the use oftestthe(DNORM)
full Newton was switched
solution to 1 (increment
method [42,44,46].
analyses have
vs. a static
displacement character, theofthe simulation time does not correspond to physical time and is simply
In the numerical overtesting current
wire ropes, step) inandthetheTC1 displacement
configuration, convergence
the simulationtolerance
time (DCTOL)
was 240 s.was For
a proportional
set
thetoTC2load
0.001 multiplier.
for TC1-B, TC1-S,
configuration, Since
andwas
the time the
TC2-S
equalglobal
tests 360displacement
to and 0.005
s. Thefor
timethestep
TC2-Bin
for simulations
the simulationswas
test. considered
employing the to be
large, the displacement
beam wire rope model normwas for0.5thes, andconvergence test (DNORM)
for the calculations using the solid wasropeswitched
model, dueto to 1the(increment
size vs.
5.1.1.
of theBeam Model
computational model, it was 1.0 s. It should be mentioned that, because these numerical
displacement over the current step) and the displacement convergence tolerance (DCTOL) was set to
analyses have geometry
TheTC1-S,
rope a static character,
was developed the simulation time does not correspond to
in physical
Section 3.time andthe is
0.001 for TC1-B, and TC2-S tests and based
0.005on for the
theequations
TC2-B presented
test. Then,
simplywas
model a proportional
bent to takeload
intomultiplier.
account the Since thegeometry
actual global displacement
of the ropesinfrom
simulations was considered
experimental tests. The
to be large,
length of thethe displacement
wire rope equaled norm1.0 for
m. the convergence
Discretization of test
the (DNORM)
single wirewaswasswitched
conductedto by
1 (increment
means of
5.1.1. Beamvs.Model
displacement over the current(ELFORM
step) and =the
Hughes–Liu beam-type elements 1) displacement
[42,44] with the convergence
use of the tolerance
2 × 2 Gauss (DCTOL)
quadraturewas
set tointegration
(four 0.001 for TC1-B,
pointsTC1-S,
in the and TC2-S testsThe
cross-section). andlength
0.005 of
forathe TC2-B
single test.
beam element was approximately
The rope geometry was developed based on the equations presented in Section 3. Then, the model
4 mm and the diameter corresponded to the actual diameter of the wire and equaled 3 mm. The wires
was bent to5.1.1.
takeBeam
intoModel
account the actual geometry of the ropes from experimental tests. The length of
were modeled using the elastoplastic constitutive law with the material properties acquired from
the wire rope equaled
tensile testing 1.0 m. out
carried
The rope geometry
Discretization
on samples
was developed
ofbased
cut the single
from the wire
rope.was
wireequations
on the
conducted
Forpresented
the beam in by means
model, the3.number
Section
of Hughes–Liu
of
Then, the
beam-typenodes
elements
model (NE) (ELFORM
was was
bent 5271
to take =the
andinto 1)account
[42,44]the
number ofwith
FEs thegeometry
was
actual use of
5250. Thethe
ofFEM 2×
the 2 Gauss
beam
ropes model
from quadrature in(four
can be seen tests.
experimental integration
Figure
The
13
points in the (note
length ofthat the visualization
cross-section).
the wire The length
rope equaled of
1.0cross-sections
m. of a single is displayed;
Discretization beam additionally,
element
of the single inapproximately
the box,bythe
wasconducted
wire was model
means 4 mm and
of
showing
Hughes–Liuthe beam FEs without
beam-type elements cross-section
(ELFORM visualization is also
= 1) [42,44] with thepresented).
use of the 2 × 2 Gauss quadrature
the diameter corresponded to the actual diameter of the wire and equaled 3 mm. The wires were
(four integration points in the cross-section). The length of a single beam element was approximately
modeled using the elastoplastic constitutive law with the material properties acquired from tensile
4 mm and the diameter corresponded to the actual diameter of the wire and equaled 3 mm. The wires
testing carried out on samples
were modeled using thecut from theconstitutive
elastoplastic wire rope.lawFor thethebeam
with model,
material the number
properties acquired of nodes (NE)
from
was 5271 and thetesting
tensile number of out
carried FEsonwas 5250.cutThe
samples from FEM beam
the wire model
rope. can
For the be model,
beam seen intheFigure
number13of(note that
nodes (NE)
the visualization was 5271 and theisnumber
of cross-sections of FEs was
displayed; 5250. The FEM
additionally, beam
in the model
box, thecan be seen
model in Figure the beam
showing
13 (note that the visualization of cross-sections is displayed; additionally, in the box, the model
FEs without cross-section visualization is also presented).
showing the beam FEs without cross-section visualization is also presented).
Figure 13. General view of wire rope made of beam finite elements (FEs) (visualization of
the cross-section of the beam FEs is turned on).
Figure
Materials 2020,13.
13, General
3842 view of wire rope made of beam finite elements (FEs) (visualization of the cross-
13 of 22
section of the beam FEs is turned on).
Solid Model
5.1.2. Solid
Similar to the beam model, the 3D geometry of the straight wire rope was constructed using the
equations introduced in Section 3. Next, Next, the
the model
model was
was bent
bent to
to recreate
recreate the geometry
geometry of the wire
ropes used in the experimental tests. The length of the wire rope was 1.0 m. The The wires
wires were modeled
using 88 node
nodeconstant
constantstress
stresssolid
solidelements
elements (ELFORM
(ELFORM = 1) [42,44]
= 1) with
[42,44] a regular
with shape.
a regular EachEach
shape. wirewire
had
32 solid
had 32 FEs
solidin FEs
the cross-section. The length
in the cross-section. Theof length
the sideofof the hexahedron
side of the elements
hexahedron was approximately
elements was
0.5 mm. The complete
approximately 0.5 mm.solid model of the
The complete solidwire ropeofisthe
model defined by NE
wire rope is = 1,722,861
defined by NEand=FE = 1,344,000.
1,722,861 and
The
FE =view of theThe
1,344,000. model
viewandof its
thedetails
modelare andshown in Figure
its details 14. Forinbetter
are shown Figurevisibility,
14. For due to visibility,
better the fine mesh,
due
the mesh is not displayed for the general view. The wires were assigned the same
to the fine mesh, the mesh is not displayed for the general view. The wires were assigned the sameelastoplastic material
model as for material
elastoplastic the beammodel
wire rope
as formodel.
the beam wire rope model.
Figure 14.
Figure General view
14. General view of
of wire
wire rope
rope made
made of
of solid
solid FEs.
FEs.
5.1.3. Parts of the Universal Testing Machine

5.1.3. Parts of the Universal Testing Machine
The numerical model of UTM includes half-round supports, half-round loading pins,
The numerical model of UTM includes half-round supports, half-round loading pins, and the
and the traverse (Figure 12). The diameter of loading and supporting pins is 80 mm. All of the UTM parts
traverse (Figure 12). The diameter of loading and supporting pins is 80 mm. All of the UTM parts
were modeled using constant stress solid elements (ELFORM = 1) [42,44]. In the traverse, the length of
were modeled using constant stress solid elements (ELFORM = 1) [42,44]. In the traverse, the length
the side of the solid elements equaled 10 mm, and for the support and loading pins, the side length
of the side of the solid elements equaled 10 mm, and for the support and loading pins, the side length
varied from approximately 2.8 to 6.3 mm. The modeled UTM parts include 91,536 nodes and 81,460 solid
varied from approximately 2.8 to 6.3 mm. The modeled UTM parts include 91,536 nodes and 81,460
FEs. The UTM was assigned the rigid material model [43] with properties corresponding to steel for
solid FEs. The UTM was assigned the rigid material model [43] with properties corresponding to steel
contact analysis purposes. The loading pins were constrained to the traverse. The boundary conditions
for contact analysis purposes. The loading pins were constrained to the traverse. The boundary
were defined through constraining all 6 degrees of freedom of bottom supports. For the remaining
conditions were defined through constraining all 6 degrees of freedom of bottom supports. For the
parts of the UTM (i.e., the loading pins and the traverse), vertical movement with a constant velocity
remaining parts of the UTM (i.e., the loading pins and the traverse), vertical movement with a
was solely possible. The wire rope lay on the supports and it could freely slide on them during the
constant velocity was solely possible. The wire rope lay on the supports and it could freely slide on
loading, and the same was true in the physical tests.
them during the loading, and the same was true in the physical tests.
5.2. Numerical Test Results
5.2. Numerical Test Results
5.2.1. Comparison with Experimental Results
5.2.1. Comparison with Experimental Results
A comparison of the results from numerical simulations with the experimental ones is shown
A comparison
in Figures 15 and 16.of the
Theresults from numerical
experimental results aresimulations
plotted inwith thecolor.
a gray experimental
The blueones
lineisrepresents
shown in
Figures
the beam 15model
and 16.(TC1/TC2-B),
The experimental resultsrepresents
and orange are plottedthe in solid
a graymodel
color. (TC1/TC2-S).
The blue line In
represents
Sections the5.2
beam model
and 5.2, (TC1/TC2-B),
T denotes time andand
DT orange represents the
is the displacement of solid model (TC1/TC2-S). In Sections 5.2.1 and
the traverse.
5.2.2,Considering
T denotes timetheand DT is theresults
numerical displacement
obtainedoffromthe traverse.
the TC1 configuration, the forces, as well as
Considering the
the displacements numerical
at the midspan,results obtained from
are approximately the TC1
linear. Theconfiguration, the forces,
forces from simulations areasinwell
the as the
lower
displacements
range of those at the midspan,
obtained are approximately
from experiments, linear.
i.e., in mostThe forces
cases, for from simulations
the same traverseare in the lower
displacement,
range of those
the forces fromobtained from experiments,
experimental i.e., than
tests are greater in most cases,
from for the tests.
numerical same traverse displacement,
The maximum the
force from
forces from experimental
the simulations testsand
for the beam aresolid
greater thanisfrom
model numerical
239 and 219 N,tests. The maximum
respectively. force from the
The displacements of
simulations formidspan
the rope at the the beamrelative
and solid model
to the is 239
loading pinsand
are219
alsoN,inrespectively.
the lower rangeTheofdisplacements of the
the results acquired
experimentally. The right-hand side graph in Figure 15 portrays an additional window presenting
numerical calculations is almost 30.9 mm. For the TC1 configuration, the beam model reflects the
numerical calculations is almost 30.9 mm. For the TC1 configuration, the beam model reflects the
experimental results a little more accurately.
experimental results a little more accurately.
By analyzing the second configuration, TC2, one can see that the force increases in a linear way to
By analyzing the second configuration, TC2, one can see that the force increases in a linear way to
the value of the traverse displacement DT of approximately 70 mm; this is ~60% of the analyzed range
the value of the traverse displacement DT of approximately 70 mm; this is ~60% of the analyzed range
(Figure 16a).
Materials 16a).
Then,
2020, 13,
the slope of the curve changes. The maximum force obtained from the beam model
3842 the slope of the curve changes. The maximum force obtained from the beam model
is
14 of 22
(Figure Then, is
489 N, whilst from the solid model, it is 440 N. The displacements at the midspan of the wire rope increase
489 N, whilst from the solid model, it is 440 N. The displacements at the midspan of the wire rope increase
linearly, similar to the TC1 configuration, and coincide well with experimental outcomes (Figure 16b). The
linearly, similar
of to the TC1 configuration, =and coincide
T=well with experimental outcomes (Figure obtained
16b). The
comparison
maximum displacements
displacement beamTand
for thefrom 234 s to
solid model 235 s. The
is 14.8 and maximum displacement
15.0 mm, respectively. For the TC2
maximum
from numericaldisplacement for the beam and solid model is 14.8 and 15.0 mm, respectively.
beam modelthe
For TC2
configuration, thecalculations
solid modelisshows
almostbetter
30.9 mm. For thewith
agreement TC1the
configuration,
experimentalthemeasurements, reflects
and the
configuration,
the experimental the solid
results model shows better agreement with the experimental measurements, and the
beam model appears to beaslightly
little more accurately.
more rigid.
beam model appears to be slightly more rigid.
(a) (b)
(a) (b)
Figure 15.
Figure Comparisonofofnumerical
15.Comparison numerical andand experimental
experimental results
results forTC1
for the the TC1 configuration
configuration (A = 80 = 80
(Acm, B =cm,
40
Figure 15. Comparison of numerical and experimental results for the TC1 configuration (A = 80 cm, B = 40
B = 40
cm): cm):vs.Load
Load vs. displacement
displacement of the curve
of the traverse traverse
(a),curve (a), anddisplacements
and relative relative displacements at the(b).
at the midspan midspan
cm): Load vs. displacement of the traverse curve (a), and relative displacements at the midspan (b).
(b).
(a) (b)
(a) (b)
Figure 16. Comparison of numerical and experimental results for the TC2 configuration (A = 57 cm, B = 17
Figure 16.
Figure Comparisonofof
16.Comparison numerical
numerical andand experimental
experimental results
results forTC2
for the the TC2 configuration
configuration (A = 57 = 57
(Acm, B =cm,
17
cm): Load vs. displacement of the traverse curve (a), and displacements from the transducer (b).
B = 17
cm): cm):vs.Load
Load vs. displacement
displacement of the traverse
of the traverse curve (a),curve (a), and displacements
and displacements from the transducer
from the transducer (b). (b).
For the two considered configurations, the displacements in the middle of the wire rope span,
By analyzing
For the secondconfigurations,
the two considered configuration, the
TC2,displacements
one can see that the middle
in the force increases in a rope
of the wire linearspan,
way
relative to the loading pins, obtained from the beam model and the solid model, are similar. The final
to the value
relative to theofloading
the traverse
pins, obtained from D
displacement T of
the approximately
beam 70 mm;
model and the solidthis is ~60%
model, of the analyzed
are similar. The final
deformations for all models are depicted in Figure 17. The responses of the beam model are more
range (Figure 16a). Then, the slope of the curve changes. The maximum force obtained
deformations for all models are depicted in Figure 17. The responses of the beam model are more from the beam
rigid. For the TC1 configuration, the beam model delivered slightly better results, and for the TC2
model is 489 N, whilst from the solid model, it is 440 N. The displacements at
rigid. For the TC1 configuration, the beam model delivered slightly better results, and for the TC2the midspan of
configuration the opposite holds true. Notwithstanding, when considering the TC2 configuration in
the wire rope increase
configuration linearly,
the opposite similar
holds true. to the TC1 configuration,
Notwithstanding, and coincidethe
when considering well
TC2with experimental
configuration in
the range corresponding to the TC1 (i.e., the forces up to 250 N), the curves from the beam model and
outcomes (Figure 16b). The maximum displacement for the beam and solid model is 14.8
the range corresponding to the TC1 (i.e., the forces up to 250 N), the curves from the beam model and and 15.0 mm,
the solid model are similar, and, for the force range greater than 250 N, the solid model better reflects
respectively.
the solid modelForare
thesimilar,
TC2 configuration,
and, for the the solid
force model
range shows
greater better
than agreement
250 N, the solidwith
modelthebetter
experimental
reflects
the real wire rope behavior.
measurements,
the real wire rope and the beam model appears to be slightly more rigid.
behavior.
For the two considered configurations, the displacements in the middle of the wire rope span,
relative to the loading pins, obtained from the beam model and the solid model, are similar. The final
deformations for all models are depicted in Figure 17. The responses of the beam model are more
rigid. For the TC1 configuration, the beam model delivered slightly better results, and for the TC2
configuration the opposite holds true. Notwithstanding, when considering the TC2 configuration in
the range corresponding to the TC1 (i.e., the forces up to 250 N), the curves from the beam model and
the solid model are similar, and, for the force range greater than 250 N, the solid model better reflects
the real wire rope behavior.
Materials 2020, 13, 3842 15 of 22
Figure 17. Wire

Wire rope deformation in numerical bending tests.
5.2.2. Analysis
5.2.2. Analysis of
of Numerical
Numerical Results
Results
The stress
The stress distribution
distribution at at the
the midspan
midspan from from the
the solid
solid model
model is is shown
shown on on the
the top
top of
of Figure
Figure 18.
18.
The
The analysis
analysis of of the
the stresses
stresses reveals
reveals that
that each
each wire
wire works
works separately
separately and and under
under bending.
bending. TheThe material
material
of the bottom portion of each wire is stretched and the material of the top
of the bottom portion of each wire is stretched and the material of the top portion is compressed. The portion is compressed.
The stress
stress statestate of the
of the single
single wire wire approximately
approximately corresponds
corresponds totothe
thestate
stateofofpure
purebending.
bending. TheThestresses
stresses in
in
the middle
the middleofofthe thecross-section
cross-section height
heightareareclose to zero,
close which
to zero, corresponds
which to the to
corresponds position of the neutral
the position of the
neutral axis, and the stresses in the outmost fibers of the single wire have similar values, but opposite
axis, and the stresses in the outmost fibers of the single wire have similar values, but opposite
signs. By
signs. Byanalyzing
analyzingvon vonMises
Misesstresses
stresses(Figure
(Figure18, 18,bottom),
bottom), similar
similar observations
observations cancan also
also be be made.
made. In
In the middle part of the wires (red area in the figure), the stresses are
the middle part of the wires (red area in the figure), the stresses are significantly smaller than thesignificantly smaller than
the stresses
stresses in the
in the outmost
outmost fibers.
fibers. Hence,
Hence, the influence
the influence of axial
of the the axial
forcesforces is negligible
is negligible compared
compared to
to the
the effects
effects of bending.
of the the bending. In addition,
In addition, stress
stress analyses
analyses overover the section
the section between
between loading
loading pins pins showed
showed that
that
the the variability
variability
Materials 2020, 13,of
of the
the PEER
x FOR
extreme
extreme stresses is approximately
stresses is approximately ±5%.
REVIEW
±5%. 16 of 22
For maximum traverse displacement, in the TC2 configuration, the maximum stress is more than
twice as high as the maximum stress of the TC1 configuration. The numerical simulation of the TC1
configuration shows that the stresses do not exceed the yield strength of 1118.5 MPa. However, for
the TC2 configuration, the wire rope reached the plastic region. Maximum plastic deformations of
0.1‰ are located under loading pins. It should be highlighted that between the loading pins the wire
rope remains in the elastic region. In Figure 18, by comparing the deformed cross-sections of TC1 and
TC2 configurations, it can be seen that the strands move relative to each other during the bending
and the displacement between the strands is greater for larger curvatures.
Figure 18. Solid

Figure18. Solid model,
model, stress
stressdistribution—cross-section
distribution—cross-sectionat
atmid-span.
mid-span.
Similar conclusions can be stated based on the results obtained from the beam wire rope model.
Figure 19 shows the stresses at integration points (abbreviation: Int Pt) of the two beam FEs located
near to the midspan of the wire rope for the TC2 configuration. For each FE, tensile stresses were
obtained at two integration points and compressive stresses at the other two. These stresses have the
same values, but opposite signs.
Materials 2020, 13, 3842 16 of 22
For maximum traverse displacement, in the TC2 configuration, the maximum stress is more
than twice as high as the maximum stress of the TC1 configuration. The numerical simulation of the
TC1 configuration shows that the stresses do not exceed the yield strength of 1118.5 MPa. However,
for the TC2 configuration, the wire rope reached the plastic region. Maximum plastic deformations of
0.1% are located under loading pins. It should be highlighted that between the loading pins the wire
rope remains in the elastic region. In Figure 18, by comparing the deformed cross-sections of TC1 and
TC2 configurations,Figure
it can18.
beSolid
seen model,
that the strands
stress move relative to eachatother
distribution—cross-section during the bending and
mid-span.
the displacement between the strands is greater for larger curvatures.
based on
Similar conclusions can be stated based on the
the results
results obtained
obtained from
from the
the beam
beam wire
wire rope
rope model.
model.
Figure 19 shows the stresses at integration points (abbreviation: Int Int Pt) of the two beam FEs located
near to the midspan of the wire roperope for
for the
the TC2
TC2 configuration.
configuration. For each FE, tensile stresses were
at two
obtained at twointegration
integrationpoints
pointsand
andcompressive
compressivestresses
stresses
at at
thethe other
other two.
two. These
These stresses
stresses havehave
the
the same values, but opposite
same values, but opposite signs.signs.
Figure 19. Beam

Beam model
model showing
showing axial stresses at integration points.
A contact
contactstress
stressanalysis
analysis was
wasconducted.
conducted.TheThe
author’s preliminary
author’s studies
preliminary revealed
studies that thethat
revealed friction
the
coefficient
friction betweenbetween
coefficient the rope the
andrope
UTM’s andparts
UTM’sFCUTC_WR
parts FC has more
UTC_WR of more
has an impact
of anon the results
impact on thethan the
results
friction coefficient between the wires
than the friction coefficient between theWR_WRFC . These preliminary tests revealed that
wires FCWR_WR. These preliminary tests revealed that if the FC if the
UTC_WR
is reduced
FC UTC_WR isfrom 0.1 to 0.02
reduced fromwithout
0.1 to changing FCWR_WR
0.02 without , the final
changing WR_WRF
FCforce the from
, UTC finalthe displacement-force
force FUTC from the
curve can decrease by
displacement-force up to
curve can6%. However,
decrease by upwhen onlyHowever,
to 6%. increasing FCWR_WR
when from 0.1 toFC
only increasing 0.3, the F
WR_WR from
UTC
force remains almost the same. If both coefficients,
0.1 to 0.3, the FUTC force remains almost the same. If UTC_WR FC and FC
both coefficients, , are increased
FCUTC_WR and FCWR_WR, are
WR_WR from 0.1
increased FUTC 0.1
to 0.3, the from force canthe
to 0.3, increase by 20%.
FUTC force This indicates
can increase by 20%.that
Thisthe FCUTC_WR
indicates thatcoefficient
the FCUTC_WR hascoefficient
a greater
effect on the results than FC
has a greater effect on the results . Finally, in the numerical models, one coefficient
WR_WR than FCWR_WR. Finally, in the numerical models, one coefficient of friction equal ofto
0.1 was assumed (see Section 5.1). To thoroughly investigate the impact of friction, a detailed contact
stress analysis was carried out. Figure 20 illustrates the normal contact stresses between the two
strands in the neighborhood of the middle of the wire rope (see location in Figure 20) obtained from
the solid model, with the TC2 configuration. Large stresses occur locally at points of contact between
the wires. Higher values are observed between strands (e.g., 11.5 and 8.0 MPa) than between the wires
within the single strand (e.g., 6.1 and 2.2 MPa). However, for most surfaces, the contact stresses equal
zero. This results from the geometry of the wire rope; most surfaces are not in contact with other
parts. A similar view of the contact stress distribution is shown in [17]. It is worth noting that the
contact stresses between the wires are smaller than between the wires and UTM’s parts. For instance,
the maximum contact compressive stress between the wire rope and loading pin is 65.4 and 153.9 MPa
for TC1 (T = 240 s, DT = 80 mm) and TC2 (T = 360 s, DT = 120 mm) configurations, respectively.
This also confirms that the friction coefficient between the wires and the UTM affects the results more
than the friction coefficient between the wires.
configurations, respectively. This also confirms that the friction coefficient between the wires and the
UTM affects the results more than the friction coefficient between the wires.
Numerical simulations correctly reflected the wire rope performance during experimental tests,
giving a reliable insight into the wire rope behavior. Despite minor differences, it can be stated that
both models
Materials provide comparable bending responses; therefore, the developed models can
2020, 13, 3842 be
17 of 22
considered as validated.
Figure
Figure20. Contactcompressive
20. Contact compressive stress
stress distribution.
distribution. Note:
Note: Maximum
Maximum of the of the level
fringe fringewas
level
set was
to 0.5set to
MPa.
0.5 MPa.
6. Results and Discussion
Numerical simulations correctly reflected the wire rope performance during experimental tests,
giving a reliable insight into
6.1. Moment–Curvature the wire for
Relationship ropethebehavior.
Wire RopeDespite minor differences, it can be stated that both
models provide comparable bending responses; therefore, the developed models can be considered
The bending moment was determined based on the forces acquired from UTM (Figures 8a–11a),
as validated.
and the curvature of the wire rope was calculated based on the measurements recorded by the
6.displacement
Results and transducer
Discussion(Figures 8b–11b).
The first step was to determine the bending moment. Because a symmetrically loaded simple
6.1.
beamMoment–Curvature
was considered,Relationship for thethat
it was assumed Wire Rope
the central region of the wire rope, between the loading
pins,The
underwent pure bending. The bending moment
bending moment was determined based on the forces M wasacquired
calculated from
from UTM the(Figure
following relation:
8a, Figure 9a,
Figure 10a, Figure 11a), and the curvature of the wire rope was calculated based on the measurements
FC
recorded by the displacement transducer (Figure  Figure
M8b, , 9b, Figure 10b, Figure 11b). (14)
2
The first step was to determine the bending moment. Because a symmetrically loaded simple
beam was considered, it was assumed that the central region of the wire rope, between the loading
pins, underwent pure bending. The bending moment M was calculated from the following relation:
FC
M= , (14)
2
where F is the force from UTM and C denotes the distance from the loading pin to the support
(see. Figure 6).
The second step was to determine the curvature of the bent wire rope. Between the loading
pins, the curvature was constant and the deflection curve was described by the equation of the circle
(Figure 21):
(x − xc )2 + ( y − yc )2 = ρ2 , (15)
where xc and yc are the center coordinates and ρ is the radius. Since the coordinates of three points
(two loading pins and the displacement at the mid-span) were known, the equation of the circle could
be determined.
theFor each laboratory
curvature test, the
was constant andbending moment
the deflection curveM acting on the wire
was described rope
by the was determined
equation from
of the circle
Equation
(Figure (14).
21): When determining the radius ρ from Equation (15), it was taken into account that the
arc of the circle coincided with the longitudinal axis of the curved rope (by adding half of the rope
thickness, see Figure 21) and that the xwire  x c  rope
  y had
yc  been
2 2
  2 , initially bent before the tests(15)(see ui
deflection
Materials 2020,in
13,Figure
3842 7). The sought curvature κ is the reciprocal of the radius ρ. In this way, 18 ofthe
22
moment–curvature
where xc and yc arerelationships were determined
the center coordinates and ρ is for theall experimental
radius. Since the tests, as shown
coordinates in Figure
of three points22.
(two loading pins and the displacement at the mid-span) were known, the equation of the circle could
be determined.
For each laboratory test, the bending moment M acting on the wire rope was determined from
Equation (14). When determining the radius ρ from Equation (15), it was taken into account that the
arc of the circle coincided with the longitudinal axis of the curved rope (by adding half of the rope
thickness, see Figure 21) and that the wire rope had been initially bent before the tests (see ui
deflection in Figure 7). The sought curvature κ is the reciprocal of the radius ρ. In this way, the
moment–curvature relationships were determined for all experimental tests, as shown in Figure 22.
Figure 21. Determination of the radius of curvature for the deformed wire rope.
rope.
For each laboratory test, the bending moment M acting on the wire rope was determined from
Equation (14). When determining the radius ρ from Equation (15), it was taken into account that
the arc of the circle coincided with the longitudinal axis of the curved rope (by adding half of the rope
thickness, see Figure 21) and that the wire rope had been initially bent before the tests (see ui deflection
in Figure 7). The sought curvature κ is the reciprocal of the radius ρ. In this way, the moment–curvature
relationships were determined
Figure for all experimental
21. Determination of the radius of tests, as shown
curvature in Figurewire
for the deformed 22. rope.
Figure 22. Moment–curvature relationships obtained from experimental tests.
To determine the final moment–curvature relationship, the curve was fitted to the set of the
experimental discrete points utilizing the method of least squares. It was assumed that this final curve
must be concave. The determined nonlinear elastic relationship is shown in Figure 23, in which it is
also compared in the analyzed range of curvatures to the curve proposed by Reid et al. [20].
Figure 22.Moment–curvature
Figure22. Moment–curvature relationships obtainedfrom
relationships obtained fromexperimental
experimentaltests.
tests.
ToTodetermine
determinethe the final
final moment–curvature relationship,
moment–curvature relationship, thethe curve
curve waswas
fittedfitted
to thetoset theof set
the of
theexperimental
experimental discrete
discrete pointspoints utilizing
utilizing the method
the method of least of least squares.
squares. It was
It was assumed assumed
that this finalthatcurve this
finalmust be concave.
curve The determined
must be concave. nonlinear
The determined elastic relationship
nonlinear is shownis in
elastic relationship Figure
shown in 23, in which
Figure 23, in it is
which
it isalso
alsocompared
Materials compared
2020, inin
13, x FOR the analyzed
the
PEER REVIEW range
analyzed range of of
curvatures
curvaturesto to
thethe
curve proposed
curve proposedby by
Reid et al.
Reid et [20].
al. [20]. 19 of 22
Moment–curvature relationship
Figure 23. Moment–curvature
Figure for 33 ××7719-mm
relationship for 19-mmwire
wire rope.
rope.
The curvature and the bending moment are related by Equation (12). Based on the determined
relationship, the flexural stiffness EI at point (0,0) for the wire rope is 13.7 × 106 Nmm2. Assuming that
the modulus of elasticity of the non-prestretched wire rope is 79.9 GPa [35], the moment of inertia of
the cross-sectional area of the wire rope I is equal to 172 mm4.
Materials 2020, 13, 3842 19 of 22
The curvature and the bending moment are related by Equation (12). Based on the determined
relationship, the flexural stiffness EI at point (0,0) for the wire rope is 13.7 × 106 Nmm2 . Assuming that
the modulus of elasticity of the non-prestretched wire rope is 79.9 GPa [35], the moment of inertia of
the cross-sectional area of the wire rope I is equal to 172 mm4 .
6.2. Discussion
The four-point bending tests of the 3 × 7 19-mm wire rope indicate that in the initial range
of curvatures, the wire rope behaves linearly. The bending response of the prestretched wire rope
is smoother than the response of non- prestretched rope. However, the force values are similar
for non-prestretched, as well as for prestretched, rope (compare Figures 8a and 10a). Moreover,
the displacements in the middle of the span are similar (see, e.g., Figures 8b and 10b). Therefore,
it was decided not to independently determine the bending characteristics for prestretched and
non-prestretched ropes, expecting that the discrepancies would be insignificant. With the potential
use of this curve to analyze cable barrier systems in numerical simulations, this approach can allow
considerable facilitation. Furthermore, as mentioned in Section 2.1, the properties of the rope can alter
over time under real road and environmental conditions, so this approach seems even more reasonable.
Based on the experimental tests, the final moment–curvature relationship in the range of curvatures
up to 0.004 mm−1 was determined (Figure 23). The obtained curvature corresponds to bending the wire
rope into a circle with a radius of 25 cm. This relationship is comparable to that proposed by
Reid et al. [20,35]; however, it should be noted that Reid’s relationship covers a higher range of
curvatures, up to 0.05 mm−1 . The curves, compared in Figure 23, differ in the initial range of
curvatures. The proposed curve is characterized by a smaller slope at the origin point and consequently,
the obtained flexural rigidity of 13.7 × 106 Nmm2 is lower than the value of 32.3 × 106 Nmm2 given
in [35]. Considering the cross-section of the wire rope as either 21 independent wires or as a single
solid cross-section, the moment of inertia is 83.5 and 2660 mm4 , respectively. The experimentally
determined moment of inertia I = 172 mm4 indicates that the wire rope works more as a set of 21
separated wires than as one solid section. A similar observation can be found in [35]. The reason for
this is that the wrap of the rope, contact surfaces, and friction forces between the wires do not provide
sufficient cross-sectional bonding to allow it to work as a single solid structure.
Detailed stress analysis demonstrates that each wire is under bending and that the influence of the
axial forces is insignificant compared to the effects of the bending. In the analyzed wire rope section,
between the loading points, the strains of the wires are in the elastic range. However, the strains of
the wires placed directly under the loading pins are locally in the plastic range. This fact may suggest
that even if the striking of the vehicle into the wire ropes does not exert substantial forces on the ropes,
and plastic strains may occur locally in the place of contact between the rope and the vehicle.
The contact stresses occur locally between the wires and this results from the architecture of
the wire rope. The values of contact pressure between the wires are small compared to the contact
stresses between the rope and the UTM’s parts. This indicates that the friction coefficient between
the adjacent wires influences the test result a little. It is due to the fact that the ratio between the sum of
contact surfaces and the sum of outside surfaces of the wires is small and the relative displacements
between wires are negligible. A similar conclusion can be found in the papers [12,24].
The rope modeled using beam FEs gives a more rigid bending response than the model utilizing
solid FEs. However, the differences are small and the overall behavior of the two considered models is
similar. It should be emphasized that the number of FEs of the beam rope model accounts for 0.4%
of the number of FEs used in the solid rope model. This directly influences the calculation time and
the necessary memory and disc resources. The outcomes of the numerical simulations are similar to
the experimental ones; consequently, the tests can be considered reliable.
Materials 2020, 13, 3842 20 of 22
7. Conclusions
Wire rope is a complicated structure that is commonly used in many fields of civil and mechanical
engineering. Previously, mainly analytical analyses were carried out. Nowadays, numerical simulations
are often employed in wire rope investigations. This paper presents research on the bending properties
of 3 × 7 19-mm wire rope used in road safety equipment. The most important achievements include
the following:
• Developing equations for a wire rope geometry;

• Developing two advanced nonlinear FE models of wire rope utilizing beam and solid
finite elements;
• Conducting an analysis of 19 experimental tests and four numerical simulations of four-point
bending. The simulations were validated against the experimental results;
• Detailed analysis of numerical results including both cross-sectional and contact stress analyses;
• Determination of the nonlinear elastic moment–curvature relation for the wire rope.
The aforementioned actions enabled an in-depth analysis of the wire rope work mechanism under
bending conditions and the formulation of conclusions. The most important discoveries and findings
are listed below:
• The responses of non-prestretched and prestretched wire rope in the range of curvatures up to
0.004 mm−1 (i.e., the radius of curvature equals 25 cm) are similar and one moment–curvature
relation is assumed for both prestretched and non-prestretched ropes;
• In the analyzed range of curvatures, the wire rope worked in the elastic range. Plastic strains in
wires appeared solely locally under the loading pins. This suggests that in real-life accidents,
the wire rope may work in the elastic range as well; however locally, in the vicinity of a point
where a vehicle impacts the barrier, plastic strain may emerge;
• The considered wire rope works more as a set of 21 separated wires than as one single solid section;
• The interwire friction coefficient does not substantially affect the results. This is due to the wire
rope geometry as most of the wire surfaces are not in contact with other wires.
Funding: This research received no external funding.

Acknowledgments: The author gratefully acknowledges Marcin Krajewski (Gdańsk University of Technology)
for technical support in the experimental tests. The author is thankful to Wojciech Witkowski and Stanisław
Burzyński (Gdańsk University of Technology) for their advice and constant interest. Calculations were carried out
at the Academic Computer Centre in Gdańsk, Gdańsk University of Technology, Poland.
Conflicts of Interest: The author declares no conflict of interest.
References
1. Mouhib, N.; Ouaomar, H.; Lahlou, M.; Ennaji, A.; El Ghorba, M. Tensile test of a strand with 2 broken wires
artificially damaged and life prediction. Int. J. Eng. Sci. 2015, 4, 11–16.
2. Karathanasopoulos, N.; Reda, H.; Ganghoffer, J. Francois finite element modeling of the elastoplastic
axial-torsional response of helical constructions to traction loads. Int. J. Mech. Sci. 2017, 133, 368–375.
[CrossRef]
3. Argatov, I. Response of a wire rope strand to axial and torsional loads: Asymptotic modeling of the effect of
interwire contact deformations. Int. J. Solids Struct. 2011, 48, 1413–1423. [CrossRef]
4. Argatov, I.I.; Gómez, X.; Tato, W.; Urchegui, M.A. Wear evolution in a stranded rope under cyclic bending:
Implications to fatigue life estimation. Wear 2011, 271, 2857–2867. [CrossRef]
5. Meknassi, M.; Barakat, M.; Tijani, A.; El Ghorba, M. Characterization of Residual Energy Loss and Damage
Prediction of wire Rope wires Exposed to Sulfuric Acid. Int. J. Eng. Res. Dev. 2017, 13, 94–99.
6. Zhou, H.; Kodur, V.K.R.; Nie, H.; Wang, Y.; Naser, M.Z. Behavior of prestressed stayed steel columns under
fire conditions. Int. J. Steel Struct. 2017, 17, 195–204. [CrossRef]
Materials 2020, 13, 3842 21 of 22
7. Li, D.; Ou, J.; Lan, C.; Li, H. Monitoring and failure analysis of corroded bridge cables under fatigue loading
using acoustic emission sensors. Sensors 2012, 12, 3901–3915. [CrossRef]
8. Utting, W.S.; Jones, N. The response of wire rope strands to axial tensile loads-Part I. Experimental results
and theoretical predictions. Int. J. Mech. Sci. 1987, 29, 605–619. [CrossRef]
9. Nawrocki, A.; Labrosse, M. Finite element model for simple straight wire rope strands. Comput. Struct. 2000,
77, 345–359. [CrossRef]
10. Wu, J. The finite element modeling of spiral ropes. Int. J. Coal Sci. Technol. 2014, 1, 346–355. [CrossRef]
11. Smyth, A. Simulation of Overhead Crane Wire Ropes Utilizing LS-DYNA. In Proceedings of the 15th
International LS-DYNA Users Conference, Detroit, MI, USA, 10–14 June 2018; pp. 1–12.
12. Vukelic, G.; Vizentin, G. Damage-induced stresses and remaining service life predictions of wire ropes.
Appl. Sci. 2017, 7, 107. [CrossRef]
13. Costello, G.A. Theory of Wire Ropes; Springer: New York, NY, USA, 1997.
14. Papailiou, K.O. On the bending stiffness of transmission line conductors. IEEE Power Eng. Rev. 1997, 17, 46.
[CrossRef]
15. Elata, D.; Eshkenazy, R.; Weiss, M.P. The mechanical behavior of a wire rope with an independent wire rope
core. Int. J. Solids Struct. 2004, 41, 1157–1172. [CrossRef]
16. Jiang, W.G.; Henshall, J.L.; Walton, J.M. Concise finite element model for three-layered straight wire rope
strand. Int. J. Mech. Sci. 2000, 42, 63–86. [CrossRef]
17. Wang, H.; Fu, C.; Cui, W.; Zhao, X.; Qie, S. Numerical simulation and experimental study on stress
deformation of braided wire rope. J. Strain Anal. Eng. Des. 2017, 52, 69–76. [CrossRef]
18. Yu, Y.; Wang, X.; Chen, Z. A simplified finite element model for structural cable bending mechanism. Int. J.
Mech. Sci. 2016, 113, 196–210. [CrossRef]
19. Kim, S.Y.; Lee, P.S. Modeling of helically stranded cables using multiple beam finite elements and its
application to torque balance design. Constr. Build. Mater. 2017, 151, 591–606. [CrossRef]
20. Reid, J.D.; Lechtenberg, K.; Stolle, C.S. Development of Advanced Finite Element Material Models for Cable Barrier
Wire Rope; MwRSF Research Report No. TRP-03-233-10; Mid-America Transportation Center: Lincoln, NE,
USA, 2010.
21. Erdönmez, C.; Imrak, C.E. Modeling and numerical analysis of the wire strand. J. Nav. Sci. Eng. 2009, 6,
30–38.
22. Judge, R.; Yang, Z. Numerical Simulation of Spiral-Strand Cables Subjected To High Velocity Fragment Impact.
In Proceedings of the 8th European LS-DYNA Users Conference, Strasbourg, France, 23–24 May 2011.
23. Boroška, J.; Pauliková, A.; Ivančo, V. Determination of elastic modulus of steel wire ropes for computer
simulation. Appl. Mech. Mater. 2014, 683, 22–27. [CrossRef]
24. Foti, F.; de Luca di Roseto, A. Analytical and finite element modelling of the elastic–plastic behaviour of
metallic strands under axial–torsional loads. Int. J. Mech. Sci. 2016, 115–116, 202–214. [CrossRef]
25. Ray, M.H.; Silvestri, C.; Conron, C.E.; Mongiardini, M. Experience with cable median barriers in the United
States: Design standards, policies, and performance. J. Transp. Eng. 2009, 711–720. [CrossRef]
26. Fang, H.; Wang, Q.; Weggel, D.C. Crash analysis and evaluation of cable median barriers on sloped medians
using an efficient finite element model. Adv. Eng. Softw. 2015, 82, 1–13. [CrossRef]
27. Budzyński, M.; Gobis, A.; Jamroz, K.; Jeliński, Ł.; Ostrowski, K. Road Restraint Systems as a Basis for
Roadside Safety Improvement. IOP Conf. Ser. Mater. Sci. Eng. 2019, 471, 1–10. [CrossRef]
28. Alberson, D.C.; Bligh, R.P.; Buth, C.E.; Bullard, D.L. A Review Of Cable/Wire Rope Barrier Design
Considerations. Transp. Res. Rec. 2003, 1851, 95–104. [CrossRef]
29. Cooner, S.A.; Rathod, Y.K.; Alberson, D.C.; Bligh, R.P.; Ranft, S.E.; Sun, D. Development of Guidelines for Cable
Median Barrier Systems in Texas; Texas Department of Transportation: Austin, TX, USA, 2009.
30. Kampschneider, L.; Homan, D.M.; Lechtenberg, K.A.; Faller, R.K.; Bielenberg, R.W.; Sicking, D.L.; Reid, J.D.;
Rosenbaugh, S.K. Evaluation Of A Non-Proprietary, High- Tension, Four-Cable Median Barrier On Level; MwRSF
Research Report No. TRP-03-258-12; Midwest States Regional Pooled Fund Program: Lincoln, NE, USA, 2012.
31. Savolainen, B.P.T.; Gates, T.J.; Russo, B.J.; Kay, J.J. Study of High-Tension Cable Barriers on Michigan Roadways;
MDOT ORBP Project Number: OR10-036, FINAL REPORT RC-1612; Michigan Department of Transportation:
Lansing, MI, USA, 2014.
32. Stolle, C.S.; Reid, J.D.; Lechtenberg, K.A. Update To Cable Barrier Literature Review; MwRSF Research Report
No. TRP-03-227-10; Midwest States Regional Pooled Fund Program: Lincoln, NE, USA, 2010.
Materials 2020, 13, 3842 22 of 22
33. Bruski, D.; Burzyński, S.; Chróścielewski, J.; Jamroz, K.; Pachocki, Ł.; Witkowski, W.; Wilde, K. Experimental
and numerical analysis of the modified TB32 crash tests of the cable barrier system. Eng. Fail. Anal. 2019,
104, 227–246. [CrossRef]
34. Bruski, D.; Witkowski, W. Numerical studies on the influence of selected construction features and road
conditions on the performance of road cable barriers. MATEC Web Conf. 2018, 1003, 1–8. [CrossRef]
35. Stolle, C.S.; Reid, J.D. Development of a wire rope model for cable guardrail simulation. Int. J. Crashworth.
2011, 16, 331–341. [CrossRef]
36. Stolle, C.S.; Reid, J.D. Modeling Wire Rope Used in Cable Barrier Systems. In Proceedings of the 11th
37. Report: Tensile Strength Test. 2016, pp. 1–3. Available online: http://www.swedwire.se/uploads/file/
14f878ba876f48f8864c29d45a34c9b1/Tensile_strength_test_on_ropes_2016_SP.pdf (accessed on 15 April 2020).
38. Wilde, K.; Bruski, D.; Budzyński, M.; Burzyński, S.; Chróścielewski, J.; Jamroz, K.; Pachocki, Ł.; Witkowski, W.
Numerical analysis of TB32 crash tests for 4-cable guardrail barrier system installed on the horizontal convex
curves of road. Int. J. Nonlinear Sci. Numer. Simul. 2019, 21, 65–81. [CrossRef]
39. Wilde, K.; Bruski, D.; Burzyński, S.; Chróścielewski, J.; Pachocki, Ł.; Witkowski, W. LS-DYNA simulations of
the impacts of a 38-ton Heavy Goods Vehicle into a road cable barrier. In Proceedings of the 12th European
LS-DYNA Conference, Koblenz, Germany, 14–16 May 2019; pp. 1–12.
40. Hibbeler, R.C. Mechanics of Materials; Pearson Prentice Hall: Upper Saddle River, NJ, USA, 2011.
41. Gere, J.M.; Goodno, B.J. Mechanics of Materials; Cengage Learning: Stamford, CT, USA, 2009.
42. LS-DYNA, Keyword User’s Manual. Volume I.; Livermore Software Technology Corporation (LSTC): Livermore,
CA, USA, 2015.
43. LS-DYNA, Keyword User’s Manual. Volume II. Material Models; Livermore Software Technology Corporation
(LSTC): Livermore, CA, USA, 2015.
44. Hallquist, J. LS-DYNA, Theory Manual; Livermore Software Technology Corporation (LSTC): Livermore, CA,
USA, 2006.
45. Borrvall, T. Mortar Contact Algorithm for Implicit Stamping Analyses in LS-DYNA®. In Proceedings of the
10th International LS-DYNA Users Conference, Detroit, MI, USA, 8–10 June 2008; pp. 1–10.
46. Pathy, S.; Borrvall, T. Quasi-Static Simulations using Implicit LS-DYNA. In Proceedings of the 14th
© 2020 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Materials 13 03842 v2

Uploaded by

Copyright:

Available Formats

Materials 13 03842 v2

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Materials 13 03842 v2

Uploaded by

Copyright:

Available Formats

materials

Keywords: 3 × 7 wire rope; cable barrier; moment–curvature relationship; bending properties;

Materials 2020, 13, 3842; doi:10.3390/ma13173842 www.mdpi.com/journal/materials

Materials 2020, 13, x FOR PEER REVIEW 3 of 22

2.2. Geometry of the Wire Rope

The final coordinates x, y, and z of the points  t the

4. Experimental Bending Testing

4.1. Test Specimens

4.2. Test Stand

Figure 7. Measurement with a displacement transducer.

4.3. Test Results

5.1. Numerical Model

Materials 2020, 13, x FOR PEER REVIEW 12 of 22

Contact between the adjacent

5.1.3. Parts of the Universal Testing Machine

Figure 17. Wire

Figure 18. Solid

Figure 19. Beam

Figure 22. Moment–curvature relationships obtained from experimental tests.

• Developing equations for a wire rope geometry;

Funding: This research received no external funding.

You might also like