Diamond wire cutting: failure modes, risks for safety and workers’

protection
G. Tantussi & M. Lanzetta
Department of Mechanical, Nuclear and Production Engineering – University of Pisa, Italy
S. Gentile & G. Depperu
GDTech snc – Pisa

ABSTRACT: Despite the fast development and diffusion of diamond wire cutting, the intrinsic risk of death
caused by the ejection of diamond beads in case of wire breakage is still an open problem.
This project, sponsored by Italian Carrara manufacturers, aims to evaluate and improve the safety of diamond
wire saw machines. This paper describes the phenomena occurring immediately after wire breakage during
squaring operations of marble blocks. The results of a systematic series of full-scale experiments with
provoked breakage and of laboratory wire testing are provided.
Over 30 experiments have been documented by high-speed imaging and analyzed in two conditions: without
protections, to examine the behaviour of a free diamond wire (which has been also simulated by a numerical
model) and with various protection equipments, to assess their effectiveness in eliminating all or at least
drastically reducing risks for workers.
1 INTRODUCTION
The introduction of diamond wire cutting in the
early eighties has produced a tremendous increase of
productivity in quarrying (Tantussi, Lanzetta &
Romoli 2003). Unfortunately, this new extraction
method has also caused a crucial increase of risks for
workers. Risks are due to the presence of a diamond
wire looped partly around the drive pulley of the
wire saw machine and in a kerf in the material to be
cut. The wire loop itself, up to tens of meters long,
runs at high speed, up to 40 m/s, and presents a risk
of contact for workers. A much higher threat is
represented by the wire breakage (Chaplin 1995,
Huang & Xu 2006), for the projection of its active
elements (diamond beads) and parts (springs and/or
spacers) that are closer to the broken ends. In some
conditions, they can reach very high speeds, higher
than the speed of sound, causing even risk of death,
also for workers at tens of meters of distance from
the cutting machine.
To reduce risk factors, machine and diamond
wire manufacturers from the Italian leading Massa-
Carrara stone district have pointed out restricted

X
C
B
A
●
●
●
●
T1
T0
T1
T0
optional idle
deflection pulleys
fixed
protection
Kevlar ribbon
mobile protection
stone
block
diamond wire
wire saw
machine
drive
pulley
rail system
pulley cover
●
C’
●
B’


Figure 1 Configuration of the wire saw machine, forces in diamond wire cutting and protection equipment: drive pulley cover, fixed
and mobile protections.
areas around the machine, and propose the adoption
of fixed and mobile protections on the machine and
over the wire loop (Figure 1).
This work has been sponsored by a Consortium
of manufacturers from the same district in order to
understand the phenomena occurring immediately
after the wire breakage and to assess the
effectiveness of current protection measures.
2 PROBLEM STATEMENT
Modelling the “whip cracking” effect, which occurs
after breakage to an elastically tensioned wire free to
deform in space, has been approached by various
researchers (Goriely & McMillen 2002, Torkar &
Arzensek 2002). This work examines with an
eminently experimental approach, the specific case
of diamond wire breakage in order to consider the
element ejection problem too.
In this initial work, the common case of vertical
cutting in block squaring is examined. The sense of
rotation is the one generally used in quarry, with the
machine pulling the top wire segment. Tests have
been carried out in a closed environment with
screens and protections for the investigators’ and
equipments safety against the projected elements.
Two series of tests have been carried out
respectively with and without various protection
equipments, to assess their performance and to
measure the wire path (due to the whip cracking
effect).
Both traditional and gummed wire have been
used. With the former, beads and spacers are free to
rotate about the steel rope, while pressers generally
positioned at intervals of five diamond beads, limit
the number of projected elements in case of
breakage. With the latter, beads are locked and
spaced by high-pressure injected resin.
3 MODELLING OF DIAMOND WIRE CUTTING
Based on the analysis of the forces involved,
according to the belt transmission theory, breakage
may occur mainly in three segments (Tantussi 2008)
shown in Figure 1:
1. inside the block (segment C’-X);
2. in the upper tighter segment between block and
pulley (X-B);
3. over the pulley (B- B’).
4 EXPERIMENTAL CONDITIONS
The acquisition system used to monitor breakage
tests is based on three cameras positioned in order to
allow a complete view of the wire. High-speed
imaging is required because phenomena occur in a
fraction of a second. A high-speed camera
(1280x1024@240 Hz) is positioned perpendicularly
to the cutting plane, viewing the area in the back of
the machine as in Figure 2. Two high-resolution
(1024x768@60 Hz) digital matrix monochrome
cameras have also been used. One of them is
positioned with a view from the top to detect the
wire motion out of the cutting plane, if any.
Two 55 kW commercial cutting machines from
different manufacturers have been used. Both are
driven with inverter to adjust the cutting speed and
the moving back speed (to keep the wire tension
constant). A current limiter set at 120 A stops the
machine to limit the wire tension.
At machine setup, the diamond wire is looped
over the drive pulley and the block, and the two wire
ends are clamped.
The mechanical properties of both the steel rope
and of the (cupper or steel) end clamps have been
measured by laboratory tensile tests and are reported
in Table 1.
The rope and clamp resistance is two to eight
times higher than that, which can be exerted by the
machine before the intervention of the current
limiter.
To provoke the loop opening during cutting, two
methods have been experimented:
1. weakening the wire rope by cutting some of the
strands;
2. clamping the loop ends with a lower load than the
nominal one.
The first method has been quickly abandoned
because it required cutting almost all the external
strands and successive breakage did not occur as a
clean cut but as unthreading of the central strands
from the clamp. The second method has produced
some dispersion of the breakage tension values
caused by the dispersion in the weakened (manual)
clamping load, as shown in Table 1.
5 RESULTS
Over 30 full-scale experiments are documented in
detail in (Tantussi 2008).
Breakage occurred in all cases in the segments
with the highest tension, as predicted in § 3. The
most significant cases are discussed here. In figures,
the wire is enhanced in red.


5.1 Path of a free diamond wire
The example shown in Figure 2 represents a
significant phenomenon occurred in the same way
during 11 breakage tests under the same conditions.
With reference to Figure 2, immediately after the
breakage, the tension T
1
in the upper segment drops
to zero, while the potential elastic energy released
accelerates that wire segment in the right direction.
An energy balance has been estimated using
experimental parameters (tensile tests reported in
Table 1 for T
1
), the traction of the machine
calculated from the absorbed current, and the belt
transmission theory for the trend of T
1
on the
segments C-X and B-A. This analysis has shown
that the kinetic energy is about twenty times higher
than the potential elastic energy produced by the
wire tension. This latter is estimated as about
2800 N from Table 1.
The wire detaching from the pulley cancels the
friction that produces the wire traction. The lower
wire segment stays in touch with the pulley in the
tangency point A of Figure 1. It remains under strain
because of the unthreading resistance exerted by the
part of wire trapped within the block on one side,
and by the traction exerted by the upper wire
segment continuing to move towards the right, along
the original trajectory.
The limited wire flexibility and its displacement
towards the right (C and subsequent of Figure 2)
determines the formation of a U-shaped curve
starting from ¤.
From the observation of the top camera and for
the absence of forces in the transverse direction,
motion takes place mostly in the cutting plane.
The wire segment between the breakage point and
the U-shaped curve gets progressively shorter (C to
C) and the free end increases its speed. When the
length of this wire segment becomes zero, the whip
cracking effect takes place: the curved segment
pivots extremely fast about the lowest point of the
U-shaped curve, the wire stretches completely and
the curved segment finally completes its pivoting
toward the bottom, hitting the floor (C to ¹).
During pivoting, the elements closest to the free end
that are not perfectly secured to the wire are
projected “fanwise” at high speed.
Figure 2 High-speed images. A typical example of free (without protections) diamond wire path after (provoked) breakage between
block and drive pulley.



Table 1. Results of laboratory tensile tests on rope samples of
493 mm.

Ultimate or Maximum
unthreading
load [N]
elongation
[mm]
Weakened cupper clamp 3207
3069
2194
2804
4272
6.0
6.5
3.7
5.3
6.3
Normal cupper clamp 4529
4704
4827
4776
4199
6.8
8.7
6.8
8.5
4.9
Normal steel clamp 10192 10246 7.0 6.1
Rope without clamp 19287 19389 17.1 18.4
The projection speed has been estimated from
images over 500 m/s and depends on the instant of
release during pivoting, and consequently also on
the actual elements position relatively to the wire
end. Because of the big threat of the whip cracking
effect, it must be prevented by interposing obstacles
(like fixed and mobile protections) before it occurs.
In the final phase, the wire end violently hits the
floor, with high deformations. The wire spring back
recalls the wire toward the machine and the
phenomenon is exhausted.
The whip cracking takes place in a few
milliseconds, while the whole phenomenon from the
breakage instant to the floor hitting can be estimated
in about 0.3 s.
5.2 Numerical simulation of whip cracking
The whip cracking phenomenon has also been
simulated using a commercial multibody software
(Solid Edge – MSC.ADAMS.solver). The diamond
wire has been modeled using a concentrated mass
system, with beads ∅ 10 mm, 5 mm thick and links
5×5 mm, 150 mm long. Constraints are of type
“hinge” without friction both for bead to link and
link to link.
A three dimensional contact has been defined
with the stone block, the drive pulley and the floor.
This has strongly affected the simulation processing
time.
No self generated elements or constraints have
been used.
The component material has been assigned in
order to fulfill the actual 0.3 kg/m wire density. The
wire length is 15 m. As required by the software, the
wire has been modeled as an open loop with
coincident ends.
Other simulation parameters: integrators GSTIFF,
time interval 0.0001 s (initial), 1 E-07 s (minimum),
0.01 s (maximum), accuracy 0.01. The number of
iterations depends on the breakage type and is in the
order of 2 E+05.
Forces and reactions at the wire ends are such to
drive all components at 40 m/s speed at the breakage
instant, e.g. when the wire ends start separating.
Gravity has been taken into account. The
evolution has not been conditioned, the only
approximation has been neglecting air drag.
Figure 4 shows the graph of the wire end speed.
The maximum simulated speed is of the same order
of that experimentally detected.
+30 ms +64 ms
+109 ms +121 ms +132 ms
+140 ms +148 ms +167 ms


Figure 3 Graphical output of the numerical simulation of the free (without protections) wire breakage between block and pulley
(closer to the block) in the case of vertical cut.



0.00 0.05 0.10 0.15 0.200.23
0
50000
100000
150000
200000
250000
300000
333868
time [s]
speed [mm/s]

Figure 4 Numerical simulation. Absolute speed of the wire end.
The simulated speed and the good agreement
between the sequence in Figure 3 and experiments
(Figure 2) validate the numerical model proposed.
5.3 Effect of protections
Among 20 experiments with at least one protection
type (drive pulley cover, fixed or mobile protection),
the 8 cases, where all protections were present as
shown in Figure 5, are discussed here. In the other
experiments individual protections have been
selectively removed to better understand their effect.
In some tests, two optional idle deflection pulleys
have been installed to increase the wire wrap angle
over the pulley (Figure 1 and Figure 5).
The following three cases have been observed.
1. Breakage inside the block. In this case the wire
end immediately afterwards the breakage point
moves up for conservation of linear momentum,
dragging up also the rest of the wire. In some cases
it hits the Kevlar protection over the block, which
dissipates the kinetic energy and prevents any whip
cracking risk.
2. Breakage between block and pulley. The wire
afterwards the breakage point continues its motion
along the same direction it had before breakage and
hits the drive pulley cover or other protections
nearby, dissipating its energy. As for the path of the
wire segment before the breakage point, two sub
cases are available: breakage point close to the block
or to the pulley. In the first case, the wire fragment
remains inside the block and may or may not hit the
upper Kevlar protection. In the second case, the
energy owned by the wire segment between the
block and the breakage point may determine a
partial or total unthreading of the wire from the
block, colliding the upper protection and even the
machine. The friction between wire and block
partially dissipates energy so collisions occur at
lower speed. With worn out diamond wires, because
of the lower friction between diamond beads and
block, the wire unthreading is increased.
3. Breakage over the pulley (more probable in the
presence of idle deflection pulleys causing rapid
path changes, 4 events on 13 experiments). In this
case (Figure 5), the lower wire segment continues its
motion towards the block and in some cases, it can
also rises the block. This motion is supported by the
traction of the top segment, which continues to move
towards the pulley, dragging also the wire segment
still inside the block. With idle pulleys, the wire end
between the top idle pulley and the breakage point
slides up hitting the fixed protections (O and O of
Figure 5) and falls in the space between the drive
pulley and its cover. The rest of the wire following
the breakage point, if able to slide up, continues its
motion towards the drive pulley, as shown by the
tangles in O and O. In O it slips in the gap between
drive pulley and its cover. In O the other end 0.7 s
after breakage hits with low energy the mobile
protection (partially rotated) and overcomes as in O.
Because of the chaotic wire movement and of
rotating pulleys, the wire can get entangled and
break further: a short wire segment is visible near
the fixed protection in O to O.
5.4 Benefits of protections
Rigid protections (drive pulley cover and fixed
protections) behind the drive pulley are necessary to
O
+0 ms
+200 ms
+700 ms
+67 ms
+333 ms
+1000 ms
O
O
O
O
O


Figure 5. Sequence of images of a typical example of diamond wire path after (provoked) breakage on the drive pulley, with idle
deflection pulleys, drive pulley cover and with both fixed and mobile protections.


stop the wire motion towards the back of the
machine. The fixed protection must hang upward
and merge to the ribbon (mobile protection) not to
leave gaps.
The relatively tight gap between the pulley and its
cover may provoke wire clamping with machine
parts causing further breakage risks and consequent
projection of elements in addition to heavily
damaging the wire, which cannot be reused as is.
Drive pulley covers must be smooth, without
overhangs (e.g. sharp edges, brackets, bolts, etc.)
that may interfere with the normal wire trajectory
inside the pulley cover after breakage. In addition,
the pulley cover should be shaped in order to contain
the wire and dissipate its energy after breakage.
The top Kevlar protection is necessary to limit the
wire motion upward, particularly in the case of
breakage close to the block outlet. Such protection
should stay as close as possible to the wire and
should wrap the block back and be secured to the
floor to prevent the motion of the wire passing down
the block and proceeding behind it, as occurred
during some experiments.
6 CONCLUSIONS
This extensive experimental work represents an
initial systematic analysis of risks caused by
breakage in diamond wire cutting and the first
approach to investigate the effects of diamond wire
breakage and the ejection of elements.
The main failure modes have been documented
and reviewed in the paper.
The case of block squaring in vertical cutting,
both with free wire cutting and with standard
protections has been examined.
High-speed imaging has been used to characterize
the behavior of a diamond wire during the most
critical 0.3 s after breakage occurs.
Experiments have shown that
without protections, the ejection of elements
(diamond beads, springs and/or spacers) at
speeds up to 500 ms is caused by the whip
cracking effect;
traveling at 40 m/s, the elastic potential
energy of the wire is negligible with respect
to its kinetic energy;
after breakage, the wire remains in the same
plane.
The beneficial effect of protections in preventing
the whip cracking effect has been demonstrated.
Consequently, the projection of elements in the case
of gummed wire with protections is practically
absent because elements are secured. With
traditional diamond wire, projection occurs in the
space between machine and block or in small
neighbour lateral areas; the energy of the projected
elements is anyway very low because the
phenomenon takes plane mostly in the cutting plane.
In the presence of detrital rocks on the floor,
bounces in unpredictable directions are possible, so
some lateral safety limits should be defined. Larger
mobile and fixed protection would provide increased
safety.
A numerical model to simulate failures has been
described. An analytical model of the wire path and
of diamond bead bounce is under study.
The mechanical properties of the steel rope and
end clamps have been tested to be included in the
model and to design the experimental conditions.
Based on that and considering the risk of fatigue
cracking, minimizing the number of clamps on a
wire and wire assembly in controlled conditions is
strongly recommended.
Current experiments include: cutting in different
configurations, e.g. investigating the reverse rotation
(pulling the lower wire segment) and the case of
horizontal cut.
7 ACKNOWLEDGEMENTS
This work has been carried out by the Department of
Mechanical, Nuclear and Production Engineering
(DIMNP) and sponsored by the Consortium
“Macchine Marmo Sicure (Safe Marble Machines)”,
established in 2006 at the Massa Carrara Industrial
Association to promote safety by research.
Thanks are due to Mr. G. Moretti and F.
Antonelli of DIMNP for support during tests.
8 REFERENCES
Chaplin, C.R. 1995. Failure mechanisms in wire ropes,
Engineering Failure Analysis, 2 (1): 45-57.
Goriely, A. & McMillen, T. 2002. Shape of a Cracking Whip,
Phys. Rev. Lett., 88 (24): 244301.
Huang, G.Q. &. Xu, X.P. 2006. Analysis of the Breakage of
Diamond Wire Saws in Sawing of Stone, Key Engineering
Materials. 304-305: 123-126.
Tantussi, G., Lanzetta, M. & Romoli, V. 2003. Diamond Wire
Cutting of Marble: State of the Art, Modeling and
Experiments with a New Testing Machine, A.I.Te.M VI,
Proc. 6th Int. Conf. Italian Assoc. Mech. Tech., Ed. L.
Carrino, Gaeta (LT), Italy, Sep. 8-10, 2003: 113-126.
Tantussi, G. 2008. Studio del comportamento del filo
diamantato a seguito di rottura, Atti del Dipartimento di
Ingegneria Meccanica, Nucleare e della Produzione, 007
(2008), Università di Pisa.
Torkar, M. & Arzensek, B. 2002. Failure of crane wire rope,
Engineering Failure Analysis, 9 (2): 227-233.