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

net/publication/318702498

DISCRETE ELEMENT MODELLING OF PUNCH TESTS WITH A DOUBLE-TWIST

HEXAGONAL WIRE MESH

Conference Paper · May 2017

CITATION READS

1 144

4 authors, including:

Antonio Pol Klaus Thoeni

University of Padova University of Newcastle
9 PUBLICATIONS   4 CITATIONS    78 PUBLICATIONS   552 CITATIONS   

SEE PROFILE SEE PROFILE

Nicola Mazzon
University of Padova
19 PUBLICATIONS   88 CITATIONS   

SEE PROFILE

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

Discrete Element Modelling of Wire Meshes View project

Predictive computational science for granular mechanics across scales View project

All content following this page was uploaded by Antonio Pol on 26 July 2017.

The user has requested enhancement of the downloaded file.

 t h
6 Interdisciplinary Workshop on Rockfall Protection
May 22 – 24, 2017, Barcelona, Spain

DISCRETE ELEMENT MODELLING OF PUNCH TESTS

WITH A DOUBLE-TWIST HEXAGONAL WIRE MESH
Antonio Pol1, Fabio Gabrieli1, Klaus Thoeni2, Nicola Mazzon3

Metallic wire meshes are commonly used for rockfall protection and rockfall mitigation.
Their design is mostly based on empirical results and experience. Due to the significant im-
portance of these structures their design needs to be more accurate and, therefore, numerical
methods are nowadays used for improving and optimising the design. The most common ap-
proaches are the Finite Element Method (FEM) and the Discrete Element Method (DEM).
The latter is used in this work as it is particularly well suited for studying discontinuous prob-
lems, including failure, with high accuracy. The study focuses on the validation of a numeri-
cal model of a double-twisted hexagonal wire mesh subject to punch tests. The numerical
model is represented by a set of spherical particles at the physical nodes of the mesh. Remote
interactions are implemented to represent the wires. The numerical predictions are compared
to the results of experimental tests to calibrate and validate the model. In particular, the influ-
ence of the stress-strain curves of the single wire and double-twist on the punch tests are in-
vestigated by using deterministic and stochastic models.

Key words: Discrete Element Method, Wire Meshes, Punch Test, Rockfall Protection

INTRODUCTION

The detachment of small rock fragments or blocks from slopes represents one of the most
significant hazard for people and infrastructures also in the context of quarries and mines. A
common way to reduce this hazard is the installation of rockfall protective systems as drapery
or cortical meshes and rockfall net barriers. In general, the design is based on engineering
work experience, but in the last years the behaviour of structural protections has also been
studied numerically and experimentally. Adequate numerical models to simulate wire meshes
are still in development and several approaches have been proposed in the literature. The most
common one is the Finite Element Method (FEM), where the wire meshes are modelled by
using truss elements, beam elements, shell finite elements and special purpose finite elements.
The FEM is well established for dynamic modelling of continuous problems with non-linear
geometries, complex mechanical behaviour, and various contact conditions. A big issue oc-
curs when the failure of the wire mesh needs to be considered, in this case computational de-
mands become very high. Hence, the Discrete Element Method (DEM) represents a good al-
ternative being particularly well suitable for dynamic problems that involve discontinuous
materials and failure. In the DEM the material is represented by a discrete number of rigid
particles which can overlap during collision. Particles interact with each other by a specific

1
ICEA dpt., Univ. of Padova, via Ognissanti 39, 35129, Padova, Italy. +390498277994, [email protected]
2
Centre for Geotechnical and Materials Modelling, University of Newcastle, Callaghan, NSW 2308, Australia.
3
Maccaferri Innovation Center Srl, Via Werner Von Siemens 19-Siemensstrasse 19, 39100 Bozen (BZ), Italy.

Rocexs: Rockfall Expert Network Page 1

 t h
6 Interdisciplinary Workshop on Rockfall Protection
May 22 – 24, 2017, Barcelona, Spain

contact or long-range interaction laws while contact detection algorithms and an explicit inte-
gration scheme of the equation of motion updates their position.
In this work, a numerical model of a double-twisted hexagonal wire mesh is used. Punch test
results from the manufacturer are used in order to calibrate the model with particular attention
on the choice of the constitutive law in the implementation of the wire mesh. The open-source
framework Yade [6] is used in this study.

DISCRETE ELEMENT MODEL OF THE WIRE MESH

This study utilises the Maccaferri double-twisted hexagonal wire mesh of the type 8x10 with
a wire diameter of 2.7 mm. The characteristics and size of the mesh considered in this study
are summarized in Fig.1a.
The hexagonal honeycomb-like structure increases the overall strength of the wire mesh and
the double-twist wires prevent that failure of a single wire compromise the entire panel.

Fig. 1 (a) Sketch of a double-twisted hexagonal wire mesh with definition of the geometrical parameters and the
corresponding remote interaction model. (b) Snapshots of punch test geometry at the beginning and (c) at failure.

The wire mesh is represented by a set of spherical particles located at the physical nodes of
the mesh [1,2,3]. Physical contact between these particles is generally not allowed. The wires
between the nodes are not discretised explicitly but they are substituted by remote interactions
(Fig. 1a) with piece-wise linear force-displacement curves. The generation of the mesh fol-
lows the procedure outlined in [4]. Thereby, all particles have the same diameter which is
equal to four times the diameter of a wire, and their density is adjusted so that the numerical
mesh and the physical mesh have the same mass.
In this work, different constitutive laws are used in order to verify their validity, comparing
numerical predictions with experimental data. In the first model, presented in [3], the force-
displacement curve for a double-twisted wire is directly derived from the one of the single
wire by using two local parameters λk and λε: the first one acts on the initial stiffness of the
double-twist, while λε takes the strain reduction at failure into account.
The second model uses two distinct force-displacement curves for the single wire and the
double twist [5]. Finally, the stochastically distorted model presented in [5] is applied. The
latter takes into account the irregularity of the real wire mesh deriving from the difference of

Rocexs: Rockfall Expert Network Page 2

 t h
6 Interdisciplinary Workshop on Rockfall Protection
May 22 – 24, 2017, Barcelona, Spain

the real length of the wire between nodes and the numerical length. For this aim, two parame-
ters are introduced: the first parameter λu defines a horizontal shift for the force-displacement
curve, the second parameter λF modifies the stiffness of the wire in the shifted part. This kind
of irregularities are randomly distributed on the mesh panel according to a triangular distribu-
tion with the average value equal to 0.5λuL0, where L0 is the undistorted length of the wire.

CALIBRATION OF THE PUNCH TESTS

Experimental results of punch tests conform to the conditions and the geometry given in UNI
11437-2012 have been used for the model calibration. The shape of the punching element is
shown in Fig.1b. The mesh panel is 3 x 3 m with xyz displacement constraints at the four
edge boundaries. The same conditions were reproduced in the numerical tests using a fric-
tionless triangular mesh to describe the punching element, imposing the vertical displacement
and measuring the force on the punching element.
In order to be able to compare the various models, the same laboratory stress-strain curves for
wires are used with each model at the beginning. With the model 1 of [3], only the single wire
curve is relevant and the best fitting parameters for the calibration of the double-twist interac-
tions resulted λk = 0.10 and λε = 0.83. For the deterministic model 2 of [5], both curves from
experimental tensile tests (single wire and double-twist) are used. Instead, with the stochastic
model 3 of [5], using the previous curves, the best fit of experimental data is obtained using
λu = 0.02 and λF = 0.80. In Fig.2a the force-displacement curves obtained by using the above
mentioned models are compared. A good agreement with the experimental results can be ob-
tained using the different models.
With regards to the stochastic model, it is important to observe that the low value of λu, ob-
tained in this calibration, suggests a negligible distortion and the model approximately be-
haves in a deterministic way. In Fig.2b a parametric study of the effect of λu is presented. It is
possible to observe that δmax decreases with decreasing λu value, whereas Forcemax initially
decreases with decreasing λu and after a value of λu = 0.2 this parameter shows no more influ-
ence on the force at failure. In this case, neither the force nor the displacement values at fail-
ure are influenced by λF, due to the very low value of λu.

Fig. 2 (a) Comparison of experimental (1 test) and numerical (3 models) results of punch tests, the curves are
normalized with the mean of maximum experimental values of force F* and displacement δ*; (b) influence of
the stochastic parameter λu on force and displacement at failure (model 3).

Rocexs: Rockfall Expert Network Page 3

 t h
6 Interdisciplinary Workshop on Rockfall Protection
May 22 – 24, 2017, Barcelona, Spain

Finally, the influence of adopting three different stress-strain curves obtained from a set of
single wire and double twist laboratory tensile tests has been investigated. As shown in Fig.3,
the choice of the stress-strain curve influences the numerical results in a non-negligible way.
Moreover, using different stress-strain curves could be a way to assess the variability that is
typical of laboratory tests.

Fig. 3 (a) Mechanical response, curves are normalized as mentioned above, of punch tests (model 2) using
curves from three repetitions of single wire and double-twist tensile tests; (b) experimental stress-strain curves,
normalized with the mean of maximum experimental values of σ and ε, of single wire and double-twisted wire
tensile tests [5].

CONCLUSIONS

Experimental results of punch tests are relevant for the assessment of the mechanical behav-
iour of wire meshes. The good agreement between discrete element simulations and experi-
mental data confirms the effectiveness of this numerical tool to evaluate laboratory tests from
small strains to failure conditions of the mesh.
Intrinsic variability of materials and mesh geometry seem to be important and can be effec-
tively evaluated performing single wire and double-twist tensile tests on a set of samples and
calibrating a stochastic model.

REFERENCES
[1] NICOT F, CAMBOU B, MAZZOLENI G (2001) Design of rockfall restraining nets from a discrete element
modeling. Rock Mech Rock Eng 34, 99-118.
[2] BERTRAND D, NICOT F, GOTTELAND P, LAMBERT S (2005) Modelling a geo-composite cell using
discrete analysis. Comput Geotech 32(8), 564-577.
[3] BERTRAND D, NICOT F, GOTTELAND P, LAMBERT S (2008) Discrete element method (DEM) nu-
merical modelling of double-twist hexagonal mesh. Canadian Geotech J 45(8), 1104-1117.
[4] THOENI K, LAMBERT C, GIACOMINI A, SLOAN S.W (2011) Discrete modelling of a rockfall protective
system. Oñate E, Owens D, editors. Particle-based methods II: Fundamentals and applications. II International
conference on particle-based methods. CIMNE Internat Center for Numerical Methods in Engineering, 24-32.
[5] THOENI K, LAMBERT C, GIACOMINI A, SLOAN S.W (2013) Discrete modelling of hexagonal wire
meshes with a stochastically distorted contact model. Comput Geotech 49, 158-169.
[6] ŠMILAUER V, CATALANO E, CHAREYRE B, DOROFENKO S, DURIEZ J, GLADKY A, ET AL
(2010) Yade Documentation, 1st ed., The Yade Project <http://yade-dem.org/doc/>.

Rocexs: Rockfall Expert Network Page 4

View publication stats