John,

The idea behind CBM frequency is this:

CBM reveals a warning sign that a functional failure will occur.

How often you look for that warning sign depends on two things.

1. How much warning does it give you?

For example, if Thermography gives you a year of warning about
functional failures, you should do Thermography at least yearly.

2. How much time do you need to respond to the warning, once
you've discovered it?

To continue the example, if the failure you discover through that
Thermography requires three months of lead time to repair, you
should do Thermography at least every nine months (so that you
will have time to set up for the repair job, before the onset of the
functional failure that the Thermography will discover).

(It should be obvious that all of these numbers are given only for the
purpose of illustrating these concepts.)

In practice, the answers to each of these questions will vary from
one site to another, and from one asset (within a site) to another.
So you can't depend on a single generic frequency to work for *all*
of *your* assets at *your* site. You have to ask these two
questions of each asset that you wish to monitor (by whatever
method). (While it may be possible to handle as a group some
identical assets that are in identical operating contexts, you have
to take care to ensure that the assets *are* identical, and that the
operating contexts *are* identical.)

A few other thoughts:

All this presupposes that you *will* do the monitoring. Sometimes
that decision is a no-brainer, but sometimes it isn't.

First, it is possible that the failure might not really give any warning
signs. If it gives warning signs, it is possible that it might not give
enough warning to permit any effective response.

It is not unusual to find that CBM (that is, using some sort of
equipment to detect the onset of failures) is technically feasible for
no more than 20% of the failure modes (failure causes) that are
reasonably likely to affect your assets.

John, I know you said that you have been "quite successful in
capturing potential problems" with your CBM program. That
certainly sounds encouraging. :-) I'm saying this to the wider
audience who might be "in love" with monitoring technology.

Second, the more often you monitor, the more expensive
monitoring will be, over a period of time. Likewise, a more
sophisticated technique might give you a longer warning time
(permitting a lower frequency of monitoring), but it might be more
expensive (it might require expensive tools, or expensive labor (or
both)).

If the most important consequence of failure is economic, it may be
worthwhile to compare the cost of doing the monitoring, over some
period of time, to the cost you expect to have if you don't do the
monitoring (over the same period of time). For failures that don't
affect operations, this comparison might show that there is no form
of monitoring that is worth doing.

Even for failures that do affect operations, monitoring *might* not be
worth doing *if* the frequency is very high or if the technique is very
expensive. Shifting to a different technique might do the trick.
However, if no form of monitoring is worth doing for a functional
failure that affects operations, it is possible that the only real
solution might be to change the asset's design in some way
(through changes to hardware/software, to procedures, and/or to
training).

If the most important consequence of the failure is
safety/environmental, or if the failure is hidden until some other
thing goes wrong, then other "worth doing" criteria apply.

Again, it's not unusual to find that less than half of the CBM tasks
that are technically feasible are also worth doing. Clearly, the only
tasks that should be put into a maintenance program are those
that are *both* technically feasible *and* worth doing.

That's not to say don't use CBM technology -- it just means, be
judicious about using it. :-) (And again, John, this comment is for
the wider audience.)

All this discussion is one small portion of the wider topic known as
Reliability-Centered Maintenance (RCM), which offers a strategic
view of your assets' maintenance requirements so that you can
identify the best ways to manage the consequences of their
failures.

John, I know this doesn't give you the one-number cookbook
answer you hoped for ... but I hope it points you in a direction that
you'll find both useful and profitable. :-)

Yours,

Dana Netherton
President, Athos Corporation
netherton@...

On 16 Jul 2001, at 17:57, John Isaac wrote:

> Dear group,

> A question for the experts.

> We are embarking on an improved system of CBM. Utilising Vibration
> monitoring, Oil analysis and Thermography. We are a multi product
> Food factory with 24/7 production. At the moment we do Vibration
> Monitoring every 2 months. Up until now it has proved quite
> successful at capturing potential problems. We do Thermography once
> a year. The question is. Is there a "Bench Mark" for the frequency
> of CBM? It has been suggested that we should do Vibration Checks
> every month and Thermography every 3 months. Your thoughts and
> opinions are welcomed.
>



If you ever need to get in contact with the owner of the list, (if you have
trouble unsubscribing, or have questions about the list itself) send email
to <owner-plantmaint@...> .