International Journal of Solids and Structures 45 (2008) 5503–5520

Contents lists available at ScienceDirect

International Journal of Solids and Structures

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

Analytical procedure for modelling recursively and wire by wire

stranded ropes subjected to traction and torsion loads q
H. Usabiaga *, J.M. Pagalday
Ikerlan Koop Jose Maria, Arizmendiarreta Pasealekua, 2, Arrasate, Gipuzkoa, Spain

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

Article history: The aim of this article is to introduce a new theoretical procedure for modelling wire ropes
Received 16 January 2008 subjected simultaneously to tensile and torsional loads. The procedure is based upon the
Received in revised form 25 March 2008 beam assumption and takes account wire by wire of the double helical wires on the basis
Available online 26 April 2008
of general thin rod theory developed by [Love, A., 1944. Mathematical Theory of Elasticity.
Dover Publications, New York]. The proposed kinematics are based on the assumption that
wires are un-lubricated and therefore that no relative sliding between adjacent wires
Keywords:
happens.
Wire rope
Stranded rope
Ó 2008 Elsevier Ltd. All rights reserved.
Wires
Stress
Strain
General thin rod theory
Resistance of materials

1. Introduction

Tension resisting elements like wire ropes are very critical elements in countless applications ranging from lifting to civil
work among others. The main advantage of the wire rope lies in its capacity to support large tensile forces with compara-
tively small bending and torsional stiffness. This property becomes particularly useful for easing the storing and transpor-
tation of ropes and it is indispensable in applications like lifting or mining where largely tensioned elements have to be
bent respectively over sheaves or drums. The source of such a peculiar mechanical property of the rope has to be found
in the local relative movements between adjacent wires of the rope when the latter is subjected to tensile and bending
and/or torsion.
Because of its criticality, some applications do very often require a quantitative evaluation of the relevant mechanical
parameters of their wire rope. Empirical rules based upon an extensive number of experimental tests have been conducted
in particular applications to predict the mechanical response of the rope. However, many authors have developed theoretical
models with the objective of reducing the large amount of expensive tests that have to be carried out in order to determine
the rope response, and with the aim of improving the knowledge about the influence of the many rope parameters involved
in the final rope response.
In this article attention is focused on stranded rope modelling. In most of the cases, despite the outer strands support the
larger part of the applied tensile force, usually they are modelled by means of some kind of homogenisation hypothesis as if
they were wires helically bent around the core.

q
The author would like to acknowledge Orona Koop. for their support. The authors are grateful to Maider Usabiaga for reading carefully the article and
making very pertinent remarks.
* Corresponding author. Tel.: +34 943 71 24 00; fax: +34 943 79 69 44.
E-mail address: [email protected] (H. Usabiaga).

0020-7683/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijsolstr.2008.04.009
5504 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520

Fig. 1. 7  7 WSC rope.

In this article we retake the approach reported by Ashkenazi et al. (2003) with the purpose of modelling outer strands
more precisely using a wire by wire modelling approach on the basis of the general thin rod theory of Love (1944). In the
first part of the article a brief introduction to the literature of stranded rope modelling is reported, in the second the devel-
opment of this model is introduced and in the third some results of this model are reported, discussed and compared against
Velinsky et al. (1984) – Costello (1990), a popular homogenised model.

2. Bibliographical review

Many theoretical strand models have been reported in the past for characterising ropes subjected to traction and torsion
loads traction–torsion. The vast majority are analytical but, in the last decades, some numerical models, based upon the
finite element method (see Jiang et al., 1999; Durville, 1997; Nawrocki and Labrosse, 2000) have been reported as well.
Extensive literature reviews of analytical models of strands can be found, for instance, in Cardou and Jolicoeur (1997),
whereas some comparative analysis between diverse analytical models and experimental results carried out by Utting
and Jones (1987) are reported in Jolicoeur and Cardou (1991) and more recently in Ghoreishi et al. (2007b).
In this article attention is focused on theoretical approaches for modelling entire stranded ropes. As depicts Fig. 1, rope or
stranded rope is composed by outer strands twisted around a central straight element called core.
Hruska (1951) proposed a first approach for modelling the stranded ropes, and its principles would be followed by some
of the subsequent authors. Using a fibre hypothesis,1 it basically computes the deformation of the outer strand as if it was an
additional wire. Once calculated the axial strain to which the outer strands are subjected, he uses the same expression as for
straight strands, developed in the frame of strand models, in order to model the stresses of the wires of the outer strands.
Velinsky et al. (1984) extents the strand model of Costello et al.2 using a similar approach as Hruska (1951) but keeping the
beam hypothesis3 used previously by Costello et al. for modelling strands. A mixed approach that, on one hand, considers the
homogenisation of strands as being a wire, and, on the other, considers the straight strand expressions derived previously by
Costello et al. is used for computing outer strands contribution to the total rope tensile force and torque. This model takes ac-
count of the radial contraction of the wires due to the Poisson effect.

1
Hypothesis that neglects any wire internal effort except the tensile or axial one (see Cardou and Jolicoeur, 1997).
2
A good collection of the work carried out by Costello and its collaborators can be found in the monograph Costello (1990).
3
Hypothesis that considers remarkable, on the contrary of what fibre hypothesis do, the contribution of the remaining beam efforts of the wire in the general
rope mechanical response.
 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520 5505

On the basis of a previous work carried out by Benndorf (1904), Schiffner (1986) introduced a model very similar in prin-
ciples to the one reported by Hruska (1951). However, on the contrary of Hruska (1951), Schiffner (1986) includes the wire
radial contraction due to the Poisson effect. According to this model, the tensile stress of the wires of the outer strands is
constant along their respective centrelines.
The author did also estimate the maximum bending and shearing stresses of the outer wires of the outer strands taking
fully account of the trajectory of the double helical wires. For this derivation, the doubly helical parametric equations devel-
oped previously by Andorfer (1983) were used. The authors noticed that direct Frenet (1852) and Serret (1851)4 curvature
and torsion variation between the unstressed and stressed configurations is not appropriate for estimating the maximum bend-
ing and torsion stresses of double helical wires. In order to estimate correctly those stresses in the doubly helical wires, the
author proposed a corrective term based on the normal vector angle variation.
Ashkenazi et al. (2003) reported a new model that considered simultaneous tensile and torsion loads on the basis of a
beam assumption. The procedure dealt with double helical wires of the rope with the same rigorousness as for the simple
helical wires. This model, which is based upon the general rod thin theory developed by Love (1944) requires, beside the wire
centrelines in the unstressed and stressed configurations, some extra kinematics description in order to compute maximum
bending and torsion stresses in the wires. Those additional conditions are developed under the assumption that friction be-
tween wires is high enough in order to prevent any relative movement between neighbourhood wires. The model reports an
uneven tensile stress distribution along the centrelines of the wires, and ratifies as well, that maximum shear stress estima-
tion by Serret–Frenet torsion change between the unstressed and stressed configuration overestimates the correct torsion
shearing stresses in the doubly helical wires.
Finally, Elata et al. (2004) reported a new model based on the same parametric equation as the one reported by the pre-
vious authors. They consider two different boundary conditions, one called locked rope-level sieves in which an initial cross
section of the rope remains planar in the deformed position and, a second one, called unlocked rope-level sieve, in which rel-
ative sliding could happen along the path of the centrelines of the wires. The first boundary condition of the article of Elata
et al. (2004) derives in the same tensile stress of wires than the model reported previously by Ashkenazi et al. (2003).
In this article, we consider again the Ashkenazi et al.’s (2003) approach. Even though we keep Ashkenazi’s infinite friction
condition, the hypothesis of the additional kinematics conditions in order to compute curvature and torsion are quite differ-
ent. Beside this, in this article, a detailed description of the procedure employed for computing Love’s (1944) kinematic
parameter is reported.

3. kinematics of the wires

3.1. Wires centreline descriptions

For the following, we will assume that rope global deformation can be governed by two parameters: the engineering uni-
tary deformation of the rope
L  0L
e0 ¼ ; ð1Þ
L
and the rope torsion, H, defined by virtue of
HL
H¼ 0L
; ð2Þ

where 0 L and L attends to the rope length in the unstressed and stressed configuration respectively and HL is the relative
torsion between the two ends of a L length rope around the centreline of the rope.
Beside the central wire, which is straight, as depicts Fig. 1, a conventional wire rope shows several helical and double heli-
cal twisted wires. Fig. 2 shows a space curve that represents the centreline of the helically twisted wires in either, the un-
stressed and stressed configurations.
The helical trajectory of a generic centreline of the wire obeys in its initial configuration to the following 0 r S parametric
equation:
0
xs ¼ RS cos hs ; ð3aÞ
0
ys ¼ Rs sin hs ; ð3bÞ
0
zs ¼ RS tan as h; ð3cÞ
0
r s ¼ f xs ys zs gT ; ð3dÞ
and, driven by the previously introduced global deformation parameters, the centreline of a helically laid wire in its de-
formed configuration obeys, if the radial contraction of the wire due to the Poisson effect is neglected, to the following r s
parametric equation:

4
See Kreyszig (1991) for a more recent reference.
5506 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520

Fig. 2. Centreline of a simple helical wire in the initial and deformed configurations.

xs ¼ Rs cosðhs þ Hzs Þ; ð4aÞ

ys ¼ Rs sinðhs þ Hzs Þ; ð4bÞ
zs ¼ 0 zs ð1 þ e0 Þ; ð4cÞ
r s ¼ f xs ys zs gT : ð4dÞ
In the previous two parametric curves hs ¼ h þ h0 . Note that, as it is depicted in Fig. 2, h is, in both cases, the free param-
eter of the curve and represents the angular position of a particle of the centreline of an helical wire around the centreline of
the rope. In the former expressions as obeys to the initial helix angle of the helical path and h0 represents what could be
called helix phase, so that this parameter enables to consider any helical wire of the strand.
The derivation of double helical wire trajectories for undeformed and deformed configurations can be developed using
the well known Frenet (1852) and Serret (1851) local axes f1 f; 2 f; 3 fg 2 R3 . Hereafter, the assumption is made that the reader
is sufficiently familiarised with this concept and its main parameters.

Fig. 3. Double helical centreline and its associated Love’s local reference system in the unstressed and stressed configurations.
 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520 5507

As it is depicted in Fig. 3, the double helical trajectory of the centreline of the wire in its initial configuration can be de-
rived making use of Serret (1851) and Frenet (1852) local axes associated to 0 rs helical curve.
A generating point is considered for that purpose. This point belongs to the normal plane of the former Frenet–Serret local
axes and turns with a Rw radius circular path around the tangent vector, 01 f s , of the helical wire centreline.5
The angle hw represents the location of the generating point with respect to the normal, 02 f , axis. This circular trajectory
expressed in the global frame f1 e; 2 e; 3 eg 2 R3 attends to:
89
>
<>
= 0
0
rw ¼ 0 rs þ R0 f s e Rw cos hw ; ð5Þ
>
: >
;
Rw sin hw
where
Rs 0 f s e ¼ ð 01 f s 0
2fs
0
3fs Þ:
A convenient ratio between hs and hw :
hw ¼ mhs þ hw0 ; ð6Þ
together with (5) leads to a double helical parametric curve.
The parameter m is a constant value that in terms of conventional rope making parameters can be estimated by virtue of
hs
m¼ ; ð7Þ
hw cos as
where hs and hw are respectively the lay lengths of the outer strands and of the outer wires of the outer strands. The ±1 value
depends respectively on whether the rope is regular or lang lay, hw0 is a strand phase angle parameter so that any double
helical wire of the outer strand can be considered by this equation and finally, f01 f s ; 02 f s ; 03 f s g denote the three unit vectors that
constitute Frenet–Serret local axis of the base helical curve.
Following this procedure the parametric equation of the double helical wires for the undeformed configuration obeys to6:
0
xw ¼ 0 xs  Rw cos hw cos hs þ Rw sin hw sin hs sin as ; ð8Þ
0 0
yw ¼ ys  Rw cos hw sin hs þ Rw cos hw sin hs sin as ; ð9Þ
0
zw ¼ 0 zs þ Rw sin hw cos hw ; ð10Þ
0 T
r w ¼ f 0 xw 0
yw 0
zw g : ð11Þ
For the development of the deformed double helical centreline equation, the assumption that the helical lay length ratio
between outer strand and outer wires of the outer strand, m, remains constant during loading is considered. This implies that
the double helical trajectory deforms proportionally to what its base path – a simple helix – does.
Therefore, the deformed helical trajectory can be derived using (5) if previously 0 rs is conveniently replaced by its
deformed expression, rs , and accordingly Serret–Frenet local reference system is referred to (4d) instead of (3d).
The expression derived with the former procedure results excessively long to be included in this article. If the proposed
procedure results laborious for the reader, an alternatively but analogue expression was also presented by Ashkenazi et al.
(2003). Note, however, that the former author’s expressions exhibit an arctangent function with whom working in further
derivations can result somehow annoying. The expression derived from the here proposed procedure does not involve this
kind of functions and can be easily manipulated for further derivations using symbolic computation software like MapleTM or
MatlabTM.

3.2. Kinematics parameters

According to the general thin rod theory of Love (1944), the kinematics of a generic rod and, hence, for this particular case
of a generic wire, are fully defined by four parameters:

 axial deformation ðeÞ

 2 curvatures ðjL ; j0L Þ
 twist ðsL Þ.

5
The nomenclature of local axes can be confusing. If we consider the axis designated by ij f k :
i
ðÞ denotes whether the axis refers to the unstressed or stressed configuration,
j ðÞ denotes the jth unit vector of the axes, and finally,
ðÞk denotes the curve type to which is associated the current axis: simple (s) or double (w) helical.
6
Equivalent equations were formerly formulated by Andorfer (1983).
5508 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520

Once the centreline position of single helical and doubly helical wire has fully been described in Section 3.1 for the un-
stressed and stressed configurations, the deformation of the wires can be directly computed for helical and double helical
wires by means of:
ds
e¼  1; ð12Þ
d0 s
where 0 ds and ds are infinitesimal arc length elements of the centreline of the wire in respectively the unstressed and
stressed configurations (Fig. 4).

3.3. Line element, curvatures and twist

According to Love, for the derivation of the rest of the kinematic parameters, a new orthonormal local frame system
f1 l; 2 l; 3 lg 2 R3 called principal torsion–flexure axes has to be defined for every cross section of the rod for both, the unstressed
and stressed configurations. The main difference of these new axes respect to the Serret–Frenet axes is that the formers keep
some kind of material relation with the cross section of the rod, whereas the firsts, (the ones related with Serret–Frenet),
they only keep relation with the centreline of the wires.
In the unstressed configuration, these frames are defined by means of the tangent unit vector of the centreline of the wire
and by means of two additional unit vectors that, together with, the former unit tangent vector, constitute the orthonormal
frame of the principal torsion and flexure axes. Further constraints or conditions for orienting the two additional unit vectors
are not restricted in the unstressed configuration. Hence, this lets some freedom in order to orient them in the unstressed
configuration according to the convenience of the developers.
To ease the derivation of curvature and twist values, the two vectors can be oriented collinear to the Frenet–Serret normal
and binormal vectors, making, in this manner, the principal torsion–flexure axes equivalent to the Frenet–Serret axes,
i li f; i ¼ 1; 3. If this is the case, the derivation of curvature and twist becomes simpler, since:

 the twistðsL Þ of the rod has the same value than Frenet–Serret torsion value ðsF Þ
 the first curvature of Love ðjL Þ is null
 the second ðjL Þ, is equivalent to the Frenet–Serret curvature value ðjF Þ.

According to Lee (1991), Serret–Frenet curvature can be calculated by virtue of

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðx_ €z  y €z_ Þ2 þ ðz_ €x  €zxÞ _ 2 þ ðx_ y €  €xyÞ _ 2
jF ¼ 3
; ð13Þ
ðx_ 2 þ y_ 2 þ z_ 2 Þ
whereas Frenet–Serret torsion can be computed by virtue of

Fig. 4. Description of the considered kinematics.

 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520 5509

Fig. 5. The kinematic assumption used by Ashkenazi et al. (2003) (from Ashkenazi et al., 2003).

 
 x_ y_ z_ 
 
 €x y € €z 

     ð14Þ
x y z
sF ¼ 2
:
ðy_ €z  y _ 2 þ ðx_ y
€z_ Þ þ ðz_ €x  €zxÞ _ 2
€  €xyÞ
  
In the former equations ðÞ, ðÞ and ðÞ denote, respectively the first, second and third derivative with respect to the free var-
iable, h, whereas x; y and z denote respectively the first second and third components of a space curve in global basis.
According to Love, in order to develop principal torsion–flexure axes in the deformed configuration, a tracking of some
materials points of the rod between unstressed and stressed configurations is required. In particular, according to this the-
ory, it must be determined the spatial location in the stressed configuration of those material points of the cross section of
the rod that in the unstressed configuration belonged to the ray defined by the vector 02 l. The vector that in the stressed con-
0
figuration defines the ray that contains those material points will be denoted in this article by 2 l .
0
Once 2 l is determined, the torsion–flexure axes in the deformed position are defined in the following manner:
the first axis, 1 l, is always collinear to the tangent vector of the centreline of the stressed rod. The third unit vector of the
axes, 3 l, is defined as
0
3l ¼ 1l  2l ; ð15Þ
and, finally, the second unit vector is derived taking into account that the local axes form an orthonormal basis. Therefore

2l ¼ 3 l  1 l: ð16Þ
According to the kinematic assumption made before, for simple helical wires, the principal torsion–flexure axes in the
deformed position, 2 l, will always remain collinear to the normal vector of the deformed path. This implies that Love’s axes
will always be equivalent to the Serret–Frenet’s ones. Therefore, for simple helical wires, curvature and torsion can be cal-
culated in the stressed configuration by means of (4d), (13) and (14).

3.3.1. Additional kinematic assumptions considered by Ashkenazi et al. (2003)

For the derivation of the principal torsion–flexure axes of the doubly helical wires in the stressed configuration, the
assumption that wires are slightly lubricated or even un-lubricated is made by Ashkenazi et al. (2003). Therefore, no relative
movement is allowed between the adjacent wires in the strands.7

7
This assumption was previously considered by Ashkenazi et al. (2003) but the materialisation of this assumption in wires kinematics is different in this
model.
5510 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520

For tracking the material particles defined by 02 lw along the stressed configuration, Ashkenazi et al. (2003) assume,
by virtue of infinite friction hypothesis, that these undergo a rotation around the centreline of the rope. According
to these authors, this rotation is equal to the rotation that the cross section of the wire associated with the con-
sidered material points undergoes around rope centreline between the unstressed and stressed configurations (see
Fig. 5).

3.3.2. Additional kinematic assumptions considered in this model

Some authors assume in their models what is sometimes called the Euler–Bernoulli hypothesis (see Cardou and
Jolicoeur, 1997). In rope modelling, this hypothesis denotes that if we consider in the initial configuration of the strand
any plane normal to the centreline of the strand and we materially track any centroid of the wires contained on that
plane along the deformed spatial configuration, all the initially considered centroids still remain within a plane. How-
ever, the Euler–Bernoulli hypothesis hasn’t got the same meaning as in the beam case since, locally, each wire of the
cross section bents independently to the rest of the cross-section of the wires of the same strand along its particular
neutral axis.
Due to the assumption that outer wires of the outer strand deforms proportionally to what the central wire of the
strand does, the Euler–Bernoulli hypothesis is not strictly kept in Ashkenazi et al.’s (2003) model, nor in this new model.
The assumption is completely fulfilled by the rope if the outer strand is considered as an homogenised medium. More-
over, it is still fulfilled by all the strands of the rope independently one by one. But those Euler–Bernoulli compliant
planes are always oriented normal to the tangent vector of the centreline of the strand. As the tangent vector of the
centrelines of the outer strands and of the central strand are not parallel, their respective Euler–Bernoulli compliant
planes do not conform the same plane. Accordingly, a generalised Euler–Bernoulli hypothesis does not exist along the
whole rope.
Even if infinite friction is considered, a generalised Euler–Bernoulli behaviour seems difficult to happen in prac-
tice along the whole rope if any outer strand helix-angle variation is assumed between the unstressed and stressed
configurations. Note that the previously introduced material line element of double helical wires does not lie in the
plane normal to the centreline of the rope. A generalised Euler–Bernoulli hypothesis along the rope assumes that
this material line element does not undergo any angle variation respecting to that normal plane during loading.
However, this assumption is unlikely to happen in practice, furthermore when doubly helical wires should not be
attached by virtue of infinite friction hypothesis to the centreline of the rope but to the centreline of the outer
strands. As we are assuming that the outer strands undergo some angle variation respecting to the plane normal
to the centreline of the rope, the material line element should accordingly also undergo some angle variation
respecting that plane. Therefore, extending the relaxing of a generalised Euler–Bernoulli hypothesis to the additional
wire kinematics to the double helical wires strands may lead to a more reasonable deformation of the outer wires
of the outer strand.
In this context, the additional kinematics reported by Ashkenazi et al. (2003) assume that the tracked material line ele-
ments of the doubly helical wires between the unstressed and stressed configurations are rigidly attached to the central wire
of the rope. This hypothesis would be very appropriate if a generalised Euler–Bernoulli hypothesis would be considered
along all the wires of the rope. But according to the relaxation applied to the centreline of the double helical wires, it seems
to be more reasonable to attach those material points of the sections to the principal torsion and flexure axis of the central
wire of the outer strand, instead of to the central wire of the rope.
0
Hence, in order to compute 2 l ðhÞ of the double helical wires, its 02 lw ðhÞ is first calculated in terms of the principal flexure
and torsion axes of the central wire of the outer strand:

rw ðhÞe  0 f s ¼ ðR0 f s e ðhÞÞ1  02 lw ðhÞ: ð17Þ

Then, the material line element in the principal flexure and torsion axes is restitute to the global frame in the stressed
configuration assuming that it is described in terms of the Love’s local axes of the central wire of the outer strands as is de-
noted by the following expression:
0
2 lw ðhÞ ¼ Rf s e ðhÞ  ðR0 f s e ðhÞÞ1  02 lw ðhÞ; ð18Þ
where
Rf s e ¼ ð 1 f s 2fs 3fs Þ: ð19Þ
After fully determining 2 lw 0 placement in the stressed configuration, the unit vectors of the deformed configuration are
computed by virtue of the tangent unit vector and Eqs. (15) and (16).
Once fully determined the Love’s frame of the double helical wires jL ; j0L and sL can be computed by virtue
8
of :

8
Note that the derivative with respect to s can be computed using the parametric curve equation given before and the so called chain-rule in the following
manner: dðds Þ ¼ dðdhÞ dh
ds
.
 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520 5511

dlð2;1Þ dlð2;2Þ dlð2;3Þ

jL ¼ lð3;1Þ þ lð3;2Þ 0 þ lð3;3Þ 0 ; ð20Þ
d0 s d s d s
dlð3;1Þ dlð3;2Þ dl
jL ¼ lð1;1Þ 0 þ lð1;2Þ m 0 þ lð1;3Þ ð3;3Þ
0
; ð21Þ
d s d s d0 s
dl dlð1;2Þ dlð1;3Þ
sL ¼ lð2;1Þ ð1;1Þ þ lð2;2Þ þ lð2;3Þ 0 ; ð22Þ
d0 s ds0 d s
where lði;jÞ is the jth component of ith l unit vector of Love.
Ashkenazi et al. (2003) used alternative expressions to the one showed here in order to estimate twist values. The pro-
cedure used for the curvature estimation was not specified in their article.

4. Local forces and equilibrium of the wires

Fig. 6 shows the components of rod inner forces and moment to which is subjected Love’s rod.
Along the rod the following forces and moments may be presented:

 Three forces: t, axial tension, and n and n0 shearing forces

 Three moments: twisting moment, h, and bending moments g and g0
 Three distributed forces v, tand f
 Three distributed moments x, x0 and n.

As it is depicted in Fig. 6, these loads are oriented according to the Love local axes in the rod.
Once the kinematics of the wires are fully defined for unstressed and stressed configurations, the axial tensile forces of the
wires are calculated by virtue of
t ¼ EAel1 ; ð23Þ
and the applied bending and torsion moments can be calculated from:

g ¼ EI1 ðjL 0 jL Þl2 ; ð24Þ

g0 ¼ EI2 ðj0L 0 j0L Þl3 ; ð25Þ
h ¼ CðsL 0 sL Þl1 : ð26Þ
In the preceding expression E denotes Young modulus, I1 and I2 denote inertia moments with respect to the axis defined
respectively by 2 l and 3 l vectors, and, finally, C denotes torsional rigidity.
According to Love, the equilibrium of thin rods is expressed by virtue of the following force and moment balance
equations:

Fig. 6. Inner forces and moments.

5512 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520

dg
 n0 sL þ t j0L þ v ¼ 0; ð27aÞ
ds
dn0
 t jL þ nsL þ t ¼ 0; ð27bÞ
ds
dt
 nj0 þ n0 jL þ f ¼ 0; ð27cÞ
ds
dg
 g 0 sL þ hj0  n0 þ x ¼ 0; ð27dÞ
ds
dg 0
 hjL þ g sL þ n þ x0 ¼ 0; ð27eÞ
ds
dh
 g j0L þ g 0 jL þ n ¼ 0: ð27fÞ
ds
Shearing forces can be directly calculated from (27d) and (27e) if x and x0 are neglected. These parameters were previ-
ously neglected among others by Costello (1990) and by Ashkenazi et al. (2003). The rest of non-zero distributed loads (v; t; f
and n) which can be interpreted as contact interactions force between adjacent wires may be also calculated from (27a),
(27b), (27c) and (27f).
Once calculated the load on each particular wire, the contribution of kth wire of jth layer of ith strand to the total tensile
and torque reaction induced at the ends of the rope can be calculated according to
T i;j;k ¼ ðti;j;k  1 l þ ni;j;k  2 l þ n0i;j;k  3 lÞ  3 e; ð28Þ
Mi;j;k ¼ ðq  ti;j;k  1 l þ q  ni;j;k  2 l þ q  n0i;j;k  3 l þ g þ g þ hÞ  3 e: ð29Þ

where q is the vector that denotes the minimum distance between the considered wire centroid and the rope centreline.
Finally the induced total reaction at the ends of the rope can be computed by virtue of

X nl ðiÞ nX
ns X w ði;jÞ

T¼ T i;j;k ; ð30Þ
i¼0 j¼0 k¼1

X nl ðiÞ nX
ns X w ði;jÞ

M¼ Mi;j;k ; ð31Þ
i¼0 j¼0 k¼1

where ns denotes the number of strands in the rope, nl , the number of layers in the ith strand and, finally, nw , the number of
wires in the jth layer of ith strand.

5. Recursive procedure

The generality of the presented procedure should be highlighted since it is completely recurrent, as it is also recurrent the
manufacturing process of the rope. Thanks to its recursion, higher twisting levels, e.g. stranded rope wrapped again around a
core, can be considered by the here considered model.
Based upon the Serret–Frenet frame and a rotating generating point, the presented procedure does not involve any the-
oretical limitation for developing parametric equations of curves of higher twisting levels. The proposed wire kinematics nei-
ther do involve any theoretical limitation when line element position in the stressed configuration is estimated, and,
accordingly, when each wire force and torque contribution is computed.
However the determination of an appropriate relation between different generating point turning angles requires some pre-
caution. Interesting discussions about this can be found in Schiffner (1986) for the case in which the rope is bent over the sheave.

6. Results and discussion

A basic 18 mm-7  7-WSC stranded rope was selected for a first evaluation of the developed model. Both, regular and lang
lay, were simultaneously considered in the analysis. Table 1 collects the main parameter of the considered ropes.
For all the wires of both ropes an elastic modulus ðEÞ of 197:9  109 Pa was considered.

6.1. Model verification

Some basic verifications of models consistency were considered first. In this context, note that, in this model, independent
equilibrium of individual wires is considered without further considerations regarding rope global equilibrium. A consistent
kinematic and the assumed hypothesis should however lead to a good global equilibrium balance. One way to measure the
global consistency of the equilibrium of the model consist on considering the resultant tensile force and torque of any cross
section of the rope along its centreline. These resultant tensile force and torque should always be equal to the reaction gen-
erated at rope ends.
 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520 5513

Table 1
Main rope parameters considered in this study

Parameter Value
R1,1 (m) 1.970  103
R1,2 (m) 1.865  103
R2,1 (m) 1.600  103
R2,2 (m) 1.500  103
p1,1 (m) –
p1,1 (m) 0.070
p2,1 (m) 0.193
p2,2 (m) 0.070

Tables 2 and 3 show in their respective first column the maximum variation in percentage registered for regular rope case
and three loading conditions. In all the cases, the registered differences of the resultant tensile and torque along rope centr-
eline is negligible, which, in turn, demonstrates that rope, beside presenting consistent individual equilibrium in its wires,
presents, as well, a consistent global equilibrium.
Some uncertainties arose in the past about whether the beam hypothesis was required in order to estimate rope stiffness
or not (see Cardou and Jolicoeur, 1997). Tables 2 and 3 summarise the contribution of tensile forces and of the additional
beam forces of the wire, that is, shearing forces and bending and twisting moments, in the global rope tensile force and tor-
que. For the here considered kinematics, rope construction and loads, beam forces seem not to have significant influence in
restrained rotation. However, when twist is applied in the rope, the contribution of additional beam forces may arrive to be
significant in torque (about 10%), and therefore, it may be convenient to take these additional forces into account.
As the model is fully linear respect to the kinematics and material law, as other linear models like Velinsky et al. (1984),
Ashkenazi et al. (2003) or Elata et al. (2004) in linear relation between external applied forces becomes also linear to the
selected deformation parameters:
      
T  T TH 
¼K ¼ : ð32Þ
M H M MH H
As was previously discussed by Ramsey (1988), those models should be fully compliant with the Betti reciprocity theo-
rem. Some authors like for example Ashkenazi et al. (2003) or Elata et al. (2004) do verify the consistentcy of the model by
verifying the symmetry level of the stiffness matrix K. In Table 4 the elements of the stiffness matrix are displayed for the
characterised regular and lang lay ropes.
The results shown in Table 4 ratify the consistency of the model since the asymmetry displayed by both rope construc-
tions is not significant.

6.2. Rope stiffness analysis

In Figs. 7 and 8 results of the model for rotation restrained tests and forced twist under constant axial deformation tests
are illustrated for regular and lang lay. Principally, due to its popularity in the rope sector, results of the model of Velinsky
et al. (1984) also known as Costello’s model have been considered as well. Note in this sense, that the new proposed model
should be, being rigorous, compared to the model of Costello et al. considering m ¼ 0, since wire radial contraction has not
been taken into account in the here presented model. However m ¼ 0:3 results have been included in the figures since they
are part of the original model.

Table 2
(Column 1) maximum variation of the resultant tensile force along the centreline of the rope, (column 2) mean resultant tensile force of the rope and (columns
3-4), respectively mean fibre, TF, and additional beam loads, TB, contribution to the total rope tensile force
max DT z TF TB
Cases/parameters T (%)a T (N) T (%)b T (%)b
3 5
e = 0.006, H = 0 2  10 3.7  10 99.98 0.02
e = 0.001, H = 2  104 0.16 2.96  104 100.32 0.32
e = 0.001,H = 1  103 0.13 7.33  104 100.10 0.10

( RL )
TðzÞdz
max TðzÞ Rz¼0L
dz
a
More precisely: RL z¼0
.
TðzÞdz
Rz¼0L
dz
z¼0

RL
T X ðzÞdz
RL
z¼0
dz
b
More precisely: R L z¼0
.
TðzÞdz
Rz¼0L
dz
z¼0
5514 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520

Table 3
(Column 1) maximum variation of the resultant torque of the rope along its centreline, (column 2) mean resultant torque of the rope and (columns 3–4) and,
respectively, mean fibre, MF, and additional beam load, MB, contribution to the total torque
max DM z MF MB
Cases/parameters M (%)a M (N mm) M (%)a M (%)a
e = 0.006, H = 0 0.32 9.6  105 99.84 0.16
e = 0.001, H = 2  104 0.65 7.05  104 109.07 9.07
e = 0.001, H = 1  103 0.33 2.04  105 92.10 7.90
a
Same expression as used in Table 2 substituting accordingly T by M.

Table 4
Stiffness of the regular and lang lay ropes considered in this study

Regular lay Lang lay

7
T (N) 6.2079  10 5.727  107
TH (N mm) 1.6582  105 1.9731  105
M (N mm) 1.6195  105 1.9574  105
MH (N mm2) 4.8189  105 1.0623  106
T H M 
TH ð%Þ 2.33 0.79

5 Regular lay 5 Regular lay

x 10 x 10
4 10
Usabiaga and Pagalday Usabiaga and Pagalday
Costello (1990) ν=0.3 Costello (1990) ν=0.3
Tensile force (T) [N]

Torque (H) [Nmm]

8
3 Costelo (1990) ν=0 Costello ν=0
6
2
4

1
2

0 0
0 1 2 3 4 5 6 0 1 2 3 4 5 6
–3
Rope deformation (ε) –3
x 10 Rope deformation (ε) x 10

5 Lang lay 5 Lang Lay

x 10 x 10
4 14
Usabiaga and Pagalday
Usabiaga and Pagalday 12 Costello(1990) ν=0.3
Tensile force (T) [N]

Costello(1990) ν=0.3
Torque (H) [Nmm]

3 10 Costello(1990) ν=0
Costello (1990) ν=0
8
2 6
4
1 2
0
0 –2
0 1 2 3 4 5 6 0 1 2 3 4 5 6
Rope deformation (ε) –3
x 10 Rope deformation (ε) –3
x 10

Fig. 7. Simulation of restrained rotation tensile test for regular and lang 7  7 rope: e = 0,0.006; H = 0.

Restrained rotation tests show a very similar performance in tensile force for either regular and lang lay. The generated
resultant torque at the ends of a lang lay rope is slightly higher than for regular lay. Regarding to the forced twist perfor-
mance under constant deformation, regular and lang lay have similar behaviour respecting tensile force, however, torque
is significantly different for regular and lang lay. It seems that lang lay is stiffer regarding the twist-rope torque relation than
a regular lay rope.
Regarding the comparison between Costello et al. and the here reported model, a very similar performance between both
models, either for regular and lang lay is achieved in restrained rotation tensile test. Forced twist test shows bigger differ-
ences, above all, for the lang lay case.
Beside this, note that with the model of Costello the influence of Poisson wire contraction in rope stiffness can be eval-
uated. On the basis of the results attained in this analysis and on the previous work carried out by Jolicoeur and Cardou
(1991) the assumption of neglecting wire radial contraction seems to be reasonable.
 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520 5515

5 Regular lay 5 Regular lay

x 10 x 10
2.5 7
Usabiaga and Pagalday
Usabiaga and Pagalday

Tensile force (T) [N]

6 Costello (1990) ν =0.3
Costello (1990) ν=0.3

Torque (H) [Nmm]

2 Costello (1990) ν=0
Costello (1990) ν=0 5
1.5 4

3
1
2
0.5
1

0 0
–2 0 2 4 6 8 10 –2 0 2 4 6 8 10
–4 –4
Rope rotation (Θ) [rad/mm] x 10 Rope rotation (Θ) [rad/mm] x 10

5 Lang lay 5 Lang lay

x 10 x 10
3 14
Usabiaga and Pagalday Usabiaga and Pagalday
12
2.5 Costello (1990) ν=0.3 Costello(1990) ν=0.3
Tensile force (T) [N]

Torque (H) [Nmm]

Costello (1990) ν=0 10 Costello(1990) ν=0
2 8
1.5 6
4
1
2
0.5
0
0 –2
–2 0 2 4 6 8 10 –2 0 2 4 6 8 10
–4 –4
Rope rotation (Θ) [rad/mm] x 10 Rope rotation (Θ) [rad/mm] x 10

Fig. 8. Simulation of forced torsion under constant deformation tests for regular and lang 7  7 rope e = 0.001; H = 2  104, 0.001.

6.3. Stress in the wires

Stresses in the wires are one of the most relevant parameters in rope endurance analysis when rope is under fluctuating
tension and torsion loads. In the following, some stress estimations attained with this new model are presented.

1200 1200

1150 1150

1100 1100
Wire axial stress (σ) [MPa]
Wire axial stress (σ) [MPa]

1050 1050

1000 1000

950 950
wire 1 wire 1
wire 2 (centroide) wire 2 (centroide)
wire 2 (maximum) wire 2 (maximum)
900 wire 8 (centroide) 900 wire 8 (centroide)
wire 8 (maximum) wire 8 (maximum)
wire 15 (centroide) wire 15 (centroide)
wire 15 (maximum) wire 15 (maximum)
850 850
0 50 100 150 0 50 100 150
Rope axis (z) [mm] Rope axis (z) [mm]

Fig. 9. Maximum and centroid axial stress on wires along the centreline of the rope for the restrained rotation tensile tests: e = 0.006, H = 0.
5516 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520

In Figs. 9 and 10 stresses of:

1. rope central wire: wire 1,

2. core outer layer wire: wire 2 (see Fig. 1)
3. outer strand central wire: wire 8
4. outer strand outer wire: wire 15

are shown. The stress of the wire in its centroid obeys to:
j ti j
ðrc Þi ¼ ð33Þ
Ai
Maximum axial stress in the wire (assuming that the cross section of the wire is round and therefore that I1 = I2=I is cal-
culated from
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
j gj22 þ j g 0 j22 Ri
ðrmax Þi ¼ ðrc Þi þ ð34Þ
Ii
Finally torsion stress is calculated from
2hi
ðrh Þi ¼ ð35Þ
pR3i
Regarding stresses, on the basis of the assumptions of this model, the only difference between regular and lang lay con-
struction lies in the stresses of the double helical wires (see Fig. 9). In order to show the former statement, in Fig. 9, the reg-
ular and lang lay are each plotted in separated figures, however, for the rest of the figures, the regular and lang lay cases are
brought together. Note as well that only the stresses of wire 1, 2, 8, and 15 are presented in this article. The stresses of the
rest of the wires of the outer layer of the core and of the central wire of the outer strands are equal to the stress levels shown
in wire 2 and 8, respectively. In case of doubly helical wires, the stress distribution along the centreline of the wire is the
same among them, but showing some ‘‘phase” offset.
According to the attained results, for the restrained rotation case, the 7  7 regular and lang models show a very similar
stress level in the wires. Core wires usually show superior stress levels than outer strands wires. Note as well that the bend-
ing stresses are more significant in lang lay, whereas stresses due to the axial deformation are on the contrary more impor-
tant in the regular lay construction.

1400
200

1200

150
1000
Wire axial stress (σ) [MPa]
Wire axial stress (σ) [MPa]

100 800

600
50

400 wire 1
wire 1
wire 2 (centroide) wire 2 (centroide)
wire 2 (maximum) wire 2 (maximum)
0 wire 8 (centroide) wire 8 (centroide)
wire 8 (maximum) 200 wire 8 (maximum)
wire 15 lang (centroide) wire 15 lang (centroide)
wire 15 lang (maximum) wire 15 lang (maximum)
wire 15 regular (centroide) wire 15 regular (centroide)
wire 15 regular (maximum) wire 15 regular (maximum)
–50 0
0 50 100 150 0 50 100 150
Rope axis (z) [mm] Rope axis (z) [mm]

Fig. 10. Maximum and centroid axial stress on wires along the centreline of the rope for twisted rope under constant tensile deformation.
 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520 5517

As depicts Fig. 10, when torsion is considered, the wires of the outer strands, specially the outer wires of the outer strands,
undergo larger stress increasing or decreasing with respect to its equivalent restrained rotation test.
Note that, according to Fig. 8, for negative H (that is, for torsioning the rope in opposite sense in relation to the outer
strand lay sense) the wires of the outer layer of the outer strands are, for the depicted case, hardly loosed. Note that, as de-
picts Fig. 8, rope under this load condition is very near from a free-rotation tensile test condition.
For positive torsion, the highly stressed wires are located in the outer layers of the outer strands. This might indicate that
in case the rope is subjected to severe torsion loads, the wires of the outer layer of the outer strand are subjected to the most
severe stresses.
According to the results, if a positive torsion is applied in the rope, the lang lay rope construction shows a lower stress
level than the regular lay rope. On the contrary, when negative torsion is considered, the results show the opposite situation:
regular lay construction shows a lower stress level than the lang lay construction.
Finally, regarding torsional stresses, as depicts Fig. 11, regular lay seems to lead to higher torsional stress values than the
lang lay for the outer wires of the outer strands, at least for all cases considered here.

6.4. Comparison with Ashkenazi et al. (2003) kinematics

In order to compare both kinematic assumptions DjL ; Dj0L and DsL of double helical wires are depicted for regular and
lang lay rope.

500

400
Δτ [1/mm]

300

200

100

0
0 20 40 60 80 100 120 140 160
Rope axis (z) [mm]

40

wire 1
20 wire 2
Δτ [1/mm]

wire 8
0 wire 15 regular
wire 15 lang

–20

–40
0 20 40 60 80 100 120 140 160
Rope axis (z) [mm]

250

200
Δτ [1/mm]

150

100

50

0
0 20 40 60 80 100 120 140 160
Rope axis (z) [mm]

Fig. 11. Torsional stress in wires for different rotation conditions.

5518 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520

–4
x 10 x 10
1 5

Δκ [1/mm]
[1/mm]
0.6 2.5
0.2
0
–0.2
–0.6 –2.5

–1 –5
0 40 80 120 160 0 40 80 120 160
rope axis (z) [mm] rope axis (z) [mm]

–3
x 10 x 10
1 1
Usabiaga and Pagalday (2008)

Δκ’ [1/mm]
’ [1/mm]

0.6 Ashkeazi et al. (2003) 0.5

0.2
0
–0.2 Usabiaga and Pagalday (2008)
–0.6 –0.5
Ashkenazi et al. (2003)
–1 –1
0 40 80 120 160 0 40 80 120 160
rope axis (z) [mm] rope axis (z) [mm]

–3
x 10 x 10
1.5 1

Δτ [1/mm]
[1/mm]

1.25 0.5

1 0

0.75 –0.5

0.5 –1
0 40 80 120 160 0 40 80 120 160
rope axis (z) [mm] rope axis (z) [mm]
(a) Regular lay (b) Lang lay

Fig. 12. DjL ; Dj0L and DsL calculated for Ashkenazi et al. (2003) and the kinematics reported in this article for regular and lang lay and for  = 0.0001 and
H = 0.001.

The kinematic assumptions of Ashkenazi are materialised as if the tracking of material particles was made by applying a
rotation of value H0 zw along the centreline of the rope. Mathematically this can be written as
0
2 lw ¼ RH  02 lw ; ð36Þ
9
where, provided that the axis of the rope is aligned with e3 axis of the global frame,
0 1
cosðH  0 zw ðhÞÞ  sinðH  0 zw ðhÞÞ 0
B C
RH ¼ @ sinðH  0 zw ðhÞÞ cosðH  0 zw ðhÞÞ 0 A: ð37Þ
0 0 1
Once the material line tracking is fully determined, curvatures and torsion are estimated in both cases by the procedure
proposed in this article, (20)–(22).
Fig. 12 shows some representative results of curvature and twist variation in double helical wires for respectively regular
and lang lay cases. As this figure depicts, both kinematics lead to significant differences in twist variation either for regular
and lang lay cases. With regards to the curvatures, this differences seem not to be significant in Dj0 case. However, for lang
lay this difference is quite significant for Dj, whereas is not that significant for regular lay.
Thus, the represented kinematics parameters show that relaxing the kinematics imposed by Ashkenazi et al. (2003) by
attaching the material line elements to the Love’s frame of the central wire of the outer strand instead of Love’s frame of
the straight central wire of the core can imply a significant difference in the local wire stress estimation.

7. Conclusions

A new wire rope model considering the material fibre tracking of cross section has been developed by means of new wire
kinematics based on the assumption that friction between wires is high enough for preventing wire-sliding between neigh-
bourhood wires in the strands. On the contrary of most of the models reported in the literature, this procedure model doubly
helical wires with the same rigorousness than straight and helical wires. This new model has been developed on the basis of
Love’s general thin rod theory.

9
As it is considered in the development of the parametric equations of the centreline.
 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520 5519

The consistency of the model has been highlighted by the invariability of resultant tensile force and torque along the
centreline of the rope and by the reciprocity theorem of Betti.

7.1. Global rope response

Based upon the results of the here introduced model, the inclusion of additional beam forces is not essential when rope
tensile stiffness with minor torsion has to be computed. Ghoreishi et al. (2007a) highlight that the influence of additional
beam forces is dependent on the number of rope components. With a larger number of wires this influence seems to be less
significant. At least for the simulations carried out it is convenient to take the additional forces and moments of a rod into
account if the rope is subjected to more severe twist levels.
The comparison with Costello’s model shows good agreement for the tensile force stiffness estimation. However, some
disagreement arises in torque calculation when forced twist tests are considered. It seems that the source of this disagree-
ments comes from the fact that the here reported model takes account of fully double helical wires, whereas the case of Cos-
tello’s model is subjected to some simplifying procedures for deriving outer strand contributions to total tensile force and
torque.
Note as well that this analysis indicates that the wire contraction by means of the Poisson effect seems not to have sig-
nificant influence in the rope stiffness. This is concluded from the small variation that is observed in Costello’s model be-
tween m ¼ 0 and m ¼ 0:3 cases. Previous work carried out by Jolicoeur and Cardou (1991) reported the same conclusion.
Therefore, at least when the stiffness of the rope has to be computed, the assumption of neglecting Poisson effects seem
to be reasonable.

7.2. Local stresses in the wires

As Schiffner (1986) highlights, bending and torsion stresses may contribute significantly to the local stress level of the
wires. In Love’s theory the bending moments of the wire are directly proportional to the change of Love’s curvatures and
twist of the rod. Some authors like Hobbs and Nabijou (1995) estimate doubly helically laid wire stresses in an analogue
manner but using direct Frenet–Serret curvature change of wire centrelines between the unstressed and stressed configu-
rations. This, however, in strict sense, seems not to be correct since maximum twist and curvatures in the unstressed and
stressed configurations might not be located in the same material fibres of the wires. Love’s (1944) theory ease to develop
a procedure to take into account of this material relation.
Ashkenazi et al. (2003) highlight the importance of a rigorous material tracking and thereby Love’s theory is used for mak-
ing rope bending and torsion estimations. The authors propose their own kinematics and pay special attention to the fact
that twist values derived from Love’s (1944) theory are dissimilar from the geometrical twist change calculation of the dou-
ble helical wire centreline. However, it must be pointed out as well, that the same rigorous material tracking also involves
differences in bending stress calculations. A procedure for computing correctly bending stresses by Love’s curvature for dou-
ble helical wires is included in this article.
New kinematics are also reported in order to model, by means of Love’s theory, the mechanics of stranded ropes under the
assumption that the friction of the rope is high enough in order to prevent any slipping between the wires. In this sense the
additional kinematics assumptions proposed by the authors and by Ashkenazi et al. (2003) lead to quite significant differ-
ences in the calculation of local kinematics parameters along the double helical wires. The here proposed additional kine-
matics seem to be more reasonable in the frame of the assumption considered by either authors during the development
of centreline of the double helical deformed parametric equations (see Section 3.3.2).
The infinite friction hypothesis and the neglecting of Poisson radial contraction may have some influence in the calcula-
tion of the local stresses.
Concerning the infinite friction assumption, Elata et al. (2004) report significant local stress differences in double
helical wire whether the infinite or frictionless case is considered. The tensile force contribution computed by this
assumption can, in this model, be replaced by the tensile force contribution estimated in the frame of Elata et al.’s
(2004) investigations if it is presumed that its rope may be nearer from frictionless conditions rather than from the
infinite ones. However, local bending and torsion stresses of the wires have been apparently calculated in all the ref-
erences of the literature under the infinite friction assumption. Moreover, this assumption has been extensively used in
more debatable situations like for example in the modelling of global bending of the rope (see Schiffner, 1986; Knapp,
1988).
Concerning the Poisson effects, Feyrer (2006) reports that it is important to be very precise estimating Poisson modulus in
order to assess accurate results in local wire stresses. However, as he admits, in practice the Poisson modulus can be only
accurately estimated by empirical procedures. Therefore, if a estimation of local stresses is required and no local experimen-
tal test is available for estimating Poisson modulus, the here introduced method could result useful for making a first eval-
uation of local stresses. Moreover, at least for infinite friction assumption, it seems that it is not available any procedure in
the literature for considering the radial contraction due to the effect of Poisson contraction.
For concluding, as the stiffness prediction of this model is very similar to the one reported by Ashkenazi et al. (2003), and
as this former model has been validated experimentally for estimating the rope global behaviour, the here reported model
should in accordance lead to a correct global rope stiffness estimation.
5520 H. Usabiaga, J.M. Pagalday / International Journal of Solids and Structures 45 (2008) 5503–5520

However, local stress estimations between the two models lead to quite different results. It would be very interesting to
carry out some research in order to validate those models and verify the influence of considering infinite friction hypothesis
and neglecting Poisson modulus.
In this context, rigorous experimental work similar to the one reported for simple strands by Utting and Jones (1987)
seems to be complicated to carry out for stranded ropes. Moreover, experimental tests could not provide local stress infor-
mation of the inner wires, and accordingly, only partial information can be obtained empirically. Therefore, the author has
decided to contribute to clarify this point by comparing global and local values of this model with a second model based
upon the finite element method. In this finite element method based model a finite friction coefficient and Poisson radial
contraction are taken into account. The analysis is being carried out at present and results of this investigation will be hope-
fully soon reported.

References

Andorfer, K., 1983. Die zugkraftverteilung in schwingend beanspruchten geraden drahtseilen. Ph.D. thesis, Techn. Universität Graz.
Ashkenazi, R., Weiss, M.P., Elata, D., 2003. Torsion and bending stresses in wires of non-rotating tower crane ropes. In: Oipeec technical meeting:
experiences with ropes. OIPEEC, Lenzburg, Suiza, pp. 77–99.
Benndorf, H., 1904. Beiträge zur theorie der drahtseile. Zeitschr. d. öster-reichischen Ingenieur-u. Architektenvereins 56 (30), 433–437.
Cardou, A., Jolicoeur, C., 1997. Mechanical models of helical strands. Applied Mechanics Reviews 50, 1–14.
Costello, G., 1990. Theory of Wire Rope. Springer-Verlag, New York.
Durville, D., 1997. Modélisation du comportement mécanique de câbles métalliques. In: 2ème Colloque National en Calculs de Structures, vol. 1. Presses
Académiques de l’Ouest, pp. 139–144.
Elata, D., Ashkenazi, R., Weiss, M., 2004. The mechanical behaviour of wire rope with internal wire rope core. International Journal of Solids and Structures
(41), 1157–1172.
Feyrer, K., 2006. Wire Ropes: Tension Endurance Reliability, first ed. Springer.
Frenet, F., 1852. Sur les courbes à double courbure. Ph.D. thesis, Toulouse.
Ghoreishi, R., Cartraud, P., Davies, P., Messager, T., 2007a. Analytical modeling of synthetic fiber ropes subjected to axial loads. Part I: A new continuum
model for multilayered fibrous structures. International Journal of Solids and Structures 44 (9), 2924–2942.
Ghoreishi, S., Messager, T., Cartraud, P., Davies, P., 2007b. Validity and limitations of linear analytical models for steel wire strand under axial loading, using
a 3d fe model. International Journal of Mechanical Sciences 49 (11), 1251–1261.
Hobbs, R.E., Nabijou, S., 1995. Change in wire curvature as a wire is bent over a sheave. Journal of Strain Analysis 30 (4), 271–281.
Hruska, F.H., 1951. Calculation of stresses in wire ropes. Wire and Wire Products 26 (5), 766–767. 799–801.
Jiang, W., Yao, M., Walton, J., 1999. A concise finite element model for simple straight wire rope strand. International Journal of Mechanical Science (41),
143–161.
Jolicoeur, C., Cardou, A., 1991. A numerical comparison of current mathematical models of twisted wire cables under axisymmetric loads. ASME Journal of
Energy Resources Technology 113 (4), 241–249.
Knapp, R., 1988. Helical wire stresses in bent cables. Trans. ASME, J. Offshore Mechanics and Arctic Engineering (110), 55–61.
Kreyszig, E., 1991. Differential Geometry. Dover.
Lee, W.K., 1991. An insight into wire rope geometry. Journal of Solid and Structures 28 (4), 471–490.
Love, A., 1944. Mathematical Theory of Elasticity. Dover Pub., New York.
Nawrocki, A., Labrosse, M., 2000. A finite element model for simple straight wire rope strand. Computers and Structures. 77, 345–369.
Ramsey, H., 1988. Theory of thin rods with application to helical constituent wires in cables. International Journal of Mechanical Science 30 (8), 559–570.
Schiffner, G., 1986. Spannungen in laufenden drahtseilen. Ph.D. thesis, Universität Stuttgart.
Serret, J.A., 1851. Sur quelques formules relatives à la théorie des courbes à double courbure. Journal de Mathematiques 16.
Utting, W., Jones, N., 1987. Response of wire rope strands to axial tensile loads. Part I: Experimental results and theoretical predictions. International Journal
of Mechanical Science 29 (9), 605–619.
Velinsky, S.A., Andewn, G., Costello, G., 1984. Wire rope with complex cross sections. ASCE Journal of Engineering Mechanics Division 110 (3), 380–391.