Belt rip prevention and detection
The severity of conveyor belt longitudinal rips can be minimized by a dual approach involving both conveyor system design improvement and belt rip detection.
Rip detection systems supplied by the belt manufacturer usually consist of embedded loops or other implanted
devices. The application of rip detection loops is discussed in relation to alternative methods for lowering belt rip.
In some cases, instruments do not detect a belt rip and so the risk of expensive belt failure always exists. Belt rip
management may also involve modification to mechanical components along the flow path, specifically the shape
of the hoppers and chutes in relation to elongated tramp objects.
Conveyor belts of all types are prone to longitudinal rips from
impact by sharp objects in the flow path. Many different
types of tramp materials unintentionally become part of the bulk
material stream, causing belt damage and rips. For example, steel
plates from chutes and hopper liners can impact a belt causing
a rip. Any elongated object such as long sharp rocks from crushers,
metal roof bolts, elongated steel objects or wood props used
in mine operations can jam in transfers causing belt gouging and
Research shows that five separate aspects of the belt rip problem
are worthy of consideration :
1. Conveyor load-points need to be designed to prevent object
jamming and belt rips.
2. Conveyor belting can be designed to minimize the penetrating
effect of impacting objects.
3. Drive system monitoring to detect impulsive changes in
torque or drive slip factors.
4. Rip detection systems that mechanically or electronically
sense a belt intrusion or rip, and
5. Large magnets can effectively remove tramp steel.
Conveyors designed without a simultaneous consideration of
tramp objects in the bulk material are prone to belt rip damage.
A higher risk of belt rip exists when the flow path is constricted
in any way relative to the length of a tramp object. Increased risk
of belt rips occur when there is a long drop length for the bulk
material and a perpendicular angle of impact to the belt. The use
of sacrificial feeders and placement of tramp metal magnets are
design additions that will lower the risk of a belt rip. In general,
long overland steel cord belts and high tension coal mine drift
belts are good candidates for rip detection systems.
Belt rip detection can reduce the amount of damage from a
rip incident. When belting is supplied with built-in rip detection
loops or other sensing systems, the mine often “feels” more secure
that the belt investment is protected.
While many belt rips are prevented every year by rip detection,
many rips go undetected either due to poor sensing, bypassed
detectors, or damage at locations other than the loading/
discharge areas. Testimonials about belt rips that were detected
are often countered by instances where a rip was not detected
even when detection systems were in place. The paper examines
the problem of belt rip detection and prevention based on the
five points listed above.
Rip detection loop systems
The aim of belt rip detection is to stop the belt, once a rip is detected.
Detection is only as viable as the number and location of
sensors. Usually, a sensor is placed outby from the loading point
and inby from the discharge, since this is historically where rips
are known to be generated.
Placement of the sensors is critical and may depend on the
belt’s profile some distance after the load point or discharge pulley.
If a rip is initiated outby from a tail-end carry-side sensor, half
the belt length could rip before it is detected. More sensors along
the carry-side or return-side belt reduce the possible rip length.
Rip detection systems are supplied by various belt manufacturers.
Most systems consist of embedded metal loops or combinations
of loops with RFID encoding. A few systems are available
from non-belt manufacturers, however those that require embedding
into the belt as a retro-fit are often not considered viable by
Essentially, embedded metal loops are the industry standard
for belt rip detection. Loops are placed in the belt at manufacture,
with a spacing anywhere from 33 m to 150 m. Some of the
better-known belt supply companies that include rip detection in
their belting are: Veyance (SensorGuard), Fenner Dunlop (EyeQ)
and Phoenix (Phenocare SLT). Other belting suppliers such as
ContiTech use rip detection systems supplied by Coal Control
Sourcing belting in a hurry after a long rip is a potential problem.
A rip detection system supplied by one belt manufacturer is
not compatible with other manufacturers. Suppose a mine uses
Goodyear belting with SensorGuard loops, but can only quickly
source Fenner belting. The Fenner belting may have EyeQ loops
but the detection systems are different, and so the added belt is
not protected. Mines with long overland belts are often locked-in
to belt supply from a single source, which is not optimum.
In summary, belting compatibility in the event of a rip failure
is an issue. The mine would need to hold spare belting on site
(at some cost) just in case a rip incident occurs. One could argue
that development of an alternative reliable non-loop rip detection
method is well overdue.
Alternative rip detection methods
Other non-embedded rip monitoring systems appear in the marketplace
from time to time.
In past years, ultrasonic devices (The Little Ripper by CBM)
have been used to test the continuity of the rubber path between
a transmitter and sensor. The accuracy of an ultrasonic rip detection
system may depend on cable diameter in steel cord belts and
may not be very reliable on all fabric belts.
Laser profiling is being trialed in Australia (by CBM International)
as a way of detecting abnormalities to the rubber cover
(splits, bulges etc.). This is a non-contact indirect measurement,however from experience, optically-based devices are very prone
to moisture and dust interference, particularly in coal mines and
dusty outback areas.
As with all detection systems used in mining, direct and simple
detection methods are the most reliable. The embedded loop
system is certainly in wide use around the world, and one could
conclude that this type of system is the accepted reference at
present even though it is not foolproof.
With regard to the belt during a ripping incident, the question
arises as to how the belt behaves after the rip commences?
The answer is not self evident, but can determine the viability
of alternative non-embedded detection systems. With respect to
loops the answer is simple: a loop breaks on a belt rip and the
inductively coupled signal at a loop sensor disappears.
Research conducted in Australia in the 1980’s showed that
thick steel cord belts tend to close back together at a rip. This behaviour
made detection unreliable when using protrusion detectors,
smart-width detectors (that accommodate edge rubber damage),
trip wires and ultrasonic transmit/receive devices. Ripped
fabric belts behaved differently to steel cord belts. Depending on
the troughing idler tilt, fabric belts could be made to overlap or
separate at the rip, based on test rig observations at South Bulli
Colliery in NSW. The conclusion was that width detection would
be a good indicator of a rip when belt separation was induced.
Any alternative rip detection method in the future would be a
useful addition to the loop systems of today.
Belt fabrication and defects at loops
Most belt manufacturers can produce belting with an increased
rip resistance. In the case of steel cord belts, a layer of transverse
(weft) fabric may be added to help eject penetrating objects. A
strong weft layer can provide a cumulative force that will resist
ongoing penetration and in many cases can dislodge the damaging
object. Some manufacturers use small transverse steel cords
in close proximity to increase rip resistance.
Addition of rip-resisting layers in a belt may have secondary
consequences to reliable operation. For example, stiff weft layers
outside the neutral axis of bending may reduce the belt troughing
(AS 1333/4) and cause return belting to ride high in vee-returns.
Return belts that do not properly contact idlers can rapidly wear
the idler shells at the contact points, as well as preventing proper
tracking control. In the latter case, belts that mis-track are prone
to edge damage, which can be just as substantial as a belt rip.
Belts that contain a fabric rip-resisting layer are more common
in high-impact mining operations. A combination of highenergy
impacts and wet mining environments is conducive to
fabric damage and water wicking. For a steel cord belt, the presence
of a water-wicking fabric layer can cause corrosion spreading
and rapid belt failure.
Embedded loops in steel cord belts can also become problematical
if there is a belt stiffness change at the rubber interface
with the loop. If the rubber cracks transversely along the loop,
water ingress will result in corrosion across the belt width and
possible catastrophic belt parting. Embedded loop structures
can also displace fabric ply layers and cables in steel cord belts,
resulting in fraying or fatigue of tensile members beneath the
Non-destructive test (NDT) records of thousands of steel
cord belts since 1980 show regular instances where embedded
loops coincide with broken steel cables. The cables break in bending on pulleys when strains are induced in out-of-plane cables
that are slightly displaced by a loop during belt manufacture.
Many belts show no problems with loops after years of operation,
while others deteriorate to unsafe levels within the first year of
Figure 1 shows an example of a roll of steel cord belting
where cables have failed at approximately 66 m intervals, which
is the loop interval in this case. At least 6+ broken cables exist
beneath embedded loops 1, 2, 3, 6 and 7 in this example. Regular
remote belt scanning is used to monitor the growth of broken
cables at loop sites. This technology allows at-risk belts to be
flagged before failure occurs. It is noted that belting with loop
placement problems is not unique to any one belt manufacturer.
Belt rip outside transfer zones
Longer belts are at higher risk of rip from sources other than
loading and discharge impacts. Any larger rock spillage on takeup
carriages can eventually jam into a belt, causing a rip. The
rip detection sensor needs to be located past the take-up on the
return run to catch this rip. In the case of steel cord belts, a large
impact can break many cables and cause a knuckle to appear.
The knuckle and its cord ends can become snagged on scrapers
or other structure, inducing a rip. Placement of the loop sensors
should take this possibility into account, so that the sensor is
located after the take-up, plows and scrapers.
In addition, belt side travel (due to edge damage, crooked
belting, non-straight splices or local idler misalignment) may
result in belt sliding into return-run steel structure, causing an
edge rip. Edge rips can propagate great distances and may even
run through a splice. If enough edge is removed, the belt capacity
and tracking becomes an operational problem.
A potential always exists for carry idler failure and dislodgement,
resulting in impact of an idler roll on the return belt run.
Idlers have been known to initiate a belt rip. Idler noise monitoring
helps to locate and replace failing idlers before a bigger
problem exists. With regard to rip monitoring, more loop sensors
need to be placed along the belt to properly minimize the extent
of a belt rip from sources other than the load/discharge points.
A 5 km long belt could have as many as 6 return and carry run
sensors to limit belt rip to a 1 km length if a rip occurs outside
the loading areas.
Conveyor design to minimise belt rip
Designing a conveyor to minimize the risk of belt rip requires a
multi-faceted approach. Transfer and load point geometry is usually
designed for optimum bulk material flow. However, a welldesigned
transfer would allow the passage of long objects that
otherwise would become jammed in the bends of a transfer. In
more confined spaces, sacrificial belts with metal removal magnets
may be a better solution.
Conveyor design therefore needs to consider transfer geometry
in relation to tramp material shape, and plan to use tramp
metal magnets and sacrificial belt feeders. Motor torque spike
monitoring should also be a considered option is design. The entire
problem of conveyor design to minimize belt rip is complicated
since it requires a specific knowledge of the individual
mine operation including the type of “long” objects that can find
their way into the flow path of the bulk material.
Crushers often generate long sharp objects (rocks) that can
cause a belt rip. The longest dimension of a rock produced by
crushers should be considered in the design of transfer chutes.
Roof bolts, timber props and other tramp metal dislodged from
equipment are other sources of potential belt rip. In all cases,
design of transfer geometry to eliminate hang-up of elongated
objects should be primary objectives in relation to preventing
belt rip in the first place.
For example, if an underground mine uses 3 to 4 m long timber
poles for temporary support, the transfer chute dimensions
should allow such an elongated object to pass without interference
or jamming on ledges and bends. A similar consideration is
required for steel roof bolts, however in this case a tramp metal
magnet above a discharging belt would most likely remove the
As a final and less direct way of preventing belt jams and rips,
motor control is available today that allows detection of sudden
jumps in motor current. Once a belt is running at full speed, the
current average can slowly change as loading or unloading occurs.
However, at no time should motor current (equivalent to
torque) rise in a step function. An increase in motor current relative
to average background current could indicate a belt rip in
progress, a jammed belt due to a rip, and even excessive returnrun
edge binding or edge rip.
Modern motor controllers can be programmed to trigger on
rapid or step changes in drive torque. While the average power,
current or torque is continually computed as a time-integral, a
high sample-rate is required to catch current glitches, torque
spikes and step shifts in motor power factor. Under a belt rip
condition, current steps can occur within less than 0.5 seconds,
and so comparison between motor current values needs to occur
at typically every 0.1 second. This should be a design goal.
A number of steps can be taken to minimize belt rip damage
to critical and expensive conveyor systems. The paper has reviewed
existing commercial rip detection systems using embedded
loop-based monitoring. Other less well know or industryadopted
methods are briefly mentioned. Belt rip detection is
reactive rather than pro-active, nevertheless any system that can
minimize belt rip damage is a valuable addition to protection of a
valuable asset : the conveyor.
Once a rip has occurred, the fix is usually expensive irrespective
of the length of belt rip. Sometimes the rip will exist until
an open-circuited loop passes by a sensor, in which case a great
deal of the belt could be destroyed. A better solution is to prevent
conditions leading to a rip in the first place. Design of chutes and
transfers is seen as a critical part of the goal.
An emphasis in the paper has been directed towards:
a/ Belting design and fabrication that will help eject ripping objects
from the belt,
b/ Transfer chute design to accommodate the longest possible
object expected to occur in the mining process,
c/ Removal of tramp metal objects from the bulk material flow
using powerful magnets,
d/ Monitoring for rip damage at all times using existing technology,
e/ Monitoring the belt for rips at locations other than the load
points and discharge areas, and
f/ Use of secondary or indirect indicators of a belt rip, such as
drive motor operation and current trend analysis that includes
*Alex Harrison, conjoint professor, Newcastle University, NSW,
and manager, Conveyor Technologies Ltd, Denver, USA.