An Alternative and (More) Effective Methodology for Idler Condition Monitoring to Enhance

Conveyor Performance

Summary

This summary report provides the rationale behind the proposal to undertake trialling a new
methodology for monitoring the condition of conveyor idlers to enable more effective overall
performance via application of a more effective predictive maintenance technique for these
components.

Background

Idler failures on conveyors during production runs have historically caused much grief to WAIO
Operations and significant loss has and continues to be incurred due to idler failures and the
consequential unplanned replacement of these components. It is common knowledge that a great
number of “defective” idlers are in plants which frequently cause plant stoppages (i.e. unplanned
downtime).

Traditionally, defective idlers had been identified by personnel “walking the conveyor” and picking out
defective idlers which create some sort of audible noise. In most places this technique is still the norm.
The problem with this method is that if a lot of idlers are picked up as defective then the process of
replacing all of them becomes insurmountable if downtime is to be kept low.

The whole idea of a proactive maintenance regime is to optimise downtime and this methodology of
idler maintenance (if it is looked upon as Proactive maintenance) brings about its own problem – i.e.,
one would not stop a plant to replace one defective idler if the idler in question does not pose danger
to the plant. On the other hand, when the numbers of defective idlers build up, significant downtime is
incurred due to unplanned maintenance that often ensues in order to maintain plant integrity. (Note:
this methodology is actually run-to-failure)

In the recent times thermography has been relied upon for idler maintenance. Proponents of this
methodology state that this is an “effective” methodology for idler maintenance, albeit the methodology
is very time consuming, taking away valuable technician time from other tasks that could be more
important in the overall scheme of things. Given that there are so many defective idlers in the plant
causing continuous series of unplanned downtime, it is questionable whether this methodology, i.e.,
thermography, is an effective means for idler condition monitoring.

The quantum of losses suffered by WAIO Operations due to idler failures makes it imperative to find a
more effective means of condition monitoring idlers. The sole purpose of any proposed methodology
will be to monitor the condition of the rolling element bearings in idlers ensuring a significant and
effective lead time is achieved to plan the replacement of “failing” idlers in a proactive manner to
enable progress towards “Zero Unplanned Downtime due to Idler Failures”.

Rationale behind the Proposal

Any predictive maintenance methodology relying on “condition monitoring” of specific components in a

system relies on the ability to pick up some “signature” exhibited by the operation of the component
which changes as some attribute of that component deteriorates. The signature must be able to
indicate the onset of the certain aspect of the component’s attribute and must change over time to be
able to ascertain some form of “rate of deterioration” leading from an incipient failure (potential failure)
to where the component has either failed or is deemed to have failed because it is too risky to continue
operating the component. The interval between the incipient failure and “failure” (whether actual or
deemed) is referred to as the P-F interval. This interval (shown in Figure 1 below) is very important
and is discussed further.
 Figure 1 – The concept of P-F Interval

The P-F interval shown in the diagram above can be thought of as the “lead time to failure” implying
that the time interval can be used to plan your repair. Provided the interval is sufficient to enable
effective planning, efficient maintenance can be carried out, going a long way towards reducing plant
downtime.

When the concept is applied to conveyor idlers, where the component failing is essentially the rolling
element bearings, the idea is to use a predictive technology that is not only applicable but also most
effective. Note, the two criteria being “applicable and effective”, the “effective” leg having a very broad
meaning. Effective in this sense implies that the downtime due to idler failures is reduced as much as
possible and ideally to zero!

The two contending predictive technologies for idler condition monitoring in the present case are:

a: Thermography
b: Ultrasonics.

The essential component that is being monitored in this case is a rolling element bearing. It must be
noted that ultrasonics has been determined to be the most effective means to condition monitor rolling
element bearings (NASA studies). However, as can be noted both technologies are applicable for
condition monitoring purposes for this scenario as both are able to detect some form of the
progressive failure before the functional failure occurs; however, ultrasonics is picking up the onset of
failure whereas as thermography is picking up an event that is occurring well past the onset.

Figure 2

Figure 2 above outlines the differences between the application of thermography and ultrasonics for
rolling element bearing condition monitoring. In this case ultrasonics can pick the onset of a
failure in the bearing compared to thermography which is picking up an advanced stage of the failure.
As soon as the early, incipient fatigue cracks (or any other mechanical defect in the earliest of stages)
occur in bearings, the very high frequency, energy emissions are capable of beingP'picked up by
ultrasonic devices.

The onset of failure is the point “P” in the diagrams above giving you the “lead time to failure” as being
the “P-F Interval”. In the case of thermography, the onset of failure that ultrasonics would be ableP‘-Fto
Interval
pick up would not be able to be picked up. As thermography senses heat energy, it would only be able
to pick impending failure when sufficient heat is generated. In this case that stage will be at the point
P', giving a “lead time to failure” of “P'-F Interval” which is much shorter than that which can be
achieved by using ultrasonics. It is this “lead time to failure” interval that makes the use of ultrasonics
more “effective” for condition monitoring conveyor idlers and rollers.

In summary and as noted above, thermography for condition monitoring does not provide a P-F
interval which is long enough to enable effective and efficient maintenance to replace tagged idlers
and rollers. The setting of alarm levels when thermography is used does not provide much help
because once a bearing starts heating up significantly, failure can be very variable – being a few hours
to days.

The Proposal

Essentially this proposal is about using ultrasonics technology for condition monitoring of conveyor
idlers to ensure that the conveyor idlers do not cause as much unscheduled downtime as it does
presently in WAIO Operations.

It is worthy of mention that this technology has been tried on site previously, and in particular,
BamBino, an ultrasonic device specifically tailored for conveyor idler condition monitoring, was the
equipment that was used. BamBino apparently had some issues which presented its own set of
problems to provide the results that were anticipated. Presently BamBino is being further developed to
overcome those issues, however, it is debatable whether the expected performance would be
delivered through this development.

Ultraprobe is an ultrasonics device that has been very successfully used for other issues in industry
and is an equipment which has continuously been developed over numerous years. The sophisticated
suite of both hardware and software tools that come with Ultraprobe 10000 is expected to deliver the
results that is desired for condition monitoring of conveyor idlers and rollers on WAIO Operations. It is
recommended that the Ultraprobe 10000 series unit be acquired and trialled on a few sites to ascertain
its performance with regards to conveyor idler condition monitoring. It should be noted that the device
could be used for a whole raft of other activities such as corona discharge and arching in electrical
installations and/or equipment, detection of air leaks, bearing diagnosis and condition monitoring, and
other activities as outlined in the information material on the Ultraprobe.

The Ultrabrobe will essentially be tuned in to treat the good idlers on a conveyor as the “descriptor” of
a “normal” condition and an alarm level set, say 20-30dB above this level to indicate a defective idler.
Each time a CM run is made the same criteria could be used when a set of idlers are known to be in
good condition. The other option would be to set a benchmark for good idlers and trigger idlers as
being faulty once that threshold has been reached. With this latter option the condition of a set of idlers
could be trended over time.

Attachments

Enclosed are some literature on the Ultraprobe 10000.