especially with respect to some of the more classic maintenance problems that we need to address.  It is recognized that the list of problems discussed here is not all-inclusive, nor are these problems necessarily common to everyone.  But they are thought to represent a mainstream of experiences that have been observed with sufficient frequency to warrant their attention here.

1.  Insufficient proactive maintenance.  This problem clearly heads the list simply because the largest expenditure of maintenance resources in plants typically occurs in the area of corrective maintenance.  Stated differently, the vast majority of plant maintenance personnel operate in a reactive mode.  A major contributor to the unit cost thus accrues from a combination of the high cost to restore plant equipment to an operable condition coupled with the penalty associated with lost production.  Since simple arithmetic readily demonstrates this fact, it is really quite surprising that this reactive environment continues to occur.  Yet, to one degree or another, it seems to be a commonplace situation.

2.  Frequent problem repetition.  This problem, or course, ties in rather directly with the preceding.  When the plant modus operandi is reactive, there is only time to restore operability.  But there is never enough time to know why the equipment failed, let alone enough time or information to know how to correct the deficiency permanently.  The result is that the same problem keeps coming up.  This repetitive failure problem is often discussed in terms of root cause analysis, or more appropriately the lack thereof.  Unless we understand why the equipment failed and act to remove the root cause, restoration to service may be a temporary measure at best, and the cycle not only continues but also is reinforced.

3.  Erroneous maintenance work.  Humans make mistakes, and errors will occur in maintenance activities (both preventive and corrective).  But what is a tolerable level of human error in a maintenance program?  Is it 1 error in 100 tasks, 1 in 1000, or 1 in 10?  The answer could depend on the consequence realized from the error.  Most plant managers wish to have that 1 in 1 million statistic, but seem to believe that reality is more like 1 in 100.

There is strong evidence, however, that human error which occurs during intrusive-type maintenance actions is the cause of about 50 percent of plant forced outages, and that some form of human error might be occurring in some locations in one of every two maintenance tasks.

4.  Sound maintenance practices not institutionalized.  One way to solve the human error problem is, for starters, to know the practices and procedures that can assure that mistakes are not made, and then to institutionalize them in the everyday work habits at the plant.  Collectively, industry has a great deal of knowledge and experience on how equipment should be handled (often called Best Practices) — how it should be removed from the plant, torn down, overhauled, reassembled, and reinstalled.  Individual plants are usually informed on only a small percentage of this collective picture.

5.  Unnecessary and conservative PM.  At first glance, one might feel that this problem is in conflict with item 1 above.  While the need for more PM coverage is clearly an appropriate issue, there is a parallel need to look at the PM that is currently performed in terms of "Is it right?" Historical evidence strongly suggests that some of our current PM activities are, in fact, not right.  A second form of this problem is where the PM task is right, but too conservative.  This problem is usually associated with task frequency (i.e., the frequency requires the PM action too often).  This seems to be especially true of major overhaul tasks where there is some substantial evidence to suggest that 50 percent or more of the PM overhaul actions are performed prematurely.  See Sec. 5.9 for additional insights on this issue.

6.  Sketchy rationale for PM actions.  Did you ever ask the maintenance manager why some PM task is being performed?  Did you receive a credible response?  Unfortunately, the absence of information on PM task origin or any documentation to clearly trace the basis for plant PM tasks is the rule and not the exception.  Perhaps one might suggest that this is not always unacceptable.

If maintenance costs (PM ? CM) were low and still decreasing, and if plant forced outages were virtually nonexistent, one might allow that this could be the case.  But neither of these factors is sufficient for us to ignore the issue of why we do a PM task, nor to forego the ability to record the basis for such actions in appropriate documentation.

7.  Blind acceptance of OEM inputs.  The original equipment manufacturer almost always provides some form of operations and maintenance manual with the delivered equipment.  From a PM point of view, two problems develop with this input.  First, the OEM has not necessarily thought through the question of PM for the equipment in a comprehensive and cost-effective fashion.  Second, the OEM sells equipment to several customers, and these customers operate that equipment in a variety of different applications.  The OEM recommendations are conservative and not necessarily applicable to the plant's operating profile.

8.  PM variability between like or similar units.  Within a given company, it is likely that multiple plant locations are involved in production and, in some instances, each plant may even have multiple units at each location.  The utility industry typifies the latter situation where two or more power generation units are frequently located at each plant site.  These multiple plant or unit situations are likewise composed of production facilities that are often virtually identical from site to site or, at the very least, contain a wide spectrum of equipment that is identical or highly similar.

Under these circumstances, it would seem reasonable to assume that their PM programs share this commonality in order to standardize procedures, training, spare inventories, etc. to capitalize on the obvious cost savings that can be achieved.  Unfortunately, this is not a good assumption; more often than not we find that each plant location tends to be its own separate entity with many of its O&M characteristics different from those of its sister plants within the company.

9.  Ineffective use of predictive maintenance technology.  This new area of maintenance technology has been evolving for several years under the name of predictive maintenance (PdM).  It is also described with titles such as condition monitoring, condition-based maintenance, monitoring and diagnostics, and performance monitoring.  All of these names describe a process whereby some parameter is measured in a nonintrusive manner and either trended over time or alarmed at some predetermined limit, the said parameter being one with a direct relationship to equipment health, or at least to some specific aspect of equipment health.

Clearly, this process has the potential for significant payoff when it can be used to tell us when it is necessary to perform some maintenance task, thus precluding both unnecessary as well as premature intrusive preventive maintenance actions that otherwise would occur.  To a large extent, much of this technology is being introduced into our plants and facilities.

But where a plant has a predictive maintenance program, more often than not its focus is on the deployment of the sophisticated, not the simple technologies, and not always directed at critical functions where the return on investment (ROI) would be significant.

10.  Failure to employ the 80/20 rule.  Our experience across several areas of U.S. industry has indicated that the majority of O&M managers and their staff do know about the 80/20 rule, and have a reasonable grasp of its meaning.

This rule states that 80% of an observed effect tends to reside in 20% of the available source.  For example, 80% of carpet wear is found in 20% of the available carpet area — because this is where the traffic occurs.  So, in a plant, 80% of the reactive maintenance and lost production costs are likely to be located in 20% of the plant's systems — the so-called bad-actor systems.  So it would seem rather logical to find that these managers would utilize this rule to allocate resources and to focus their priorities on the bad-actor 20%.  Surprisingly (to us), we rarely find this to be the case, and must conclude that there are lost opportunities for cost-effective decisions and actions occurring daily in our plants and facilities.

Printed with permission from Butterworth-Heinemann, a division of Elsevier, from RCM - Gateway to World Class Maintenance, by Anthony M. Smith, AMS Associates Inc. in California, and Glenn R. Hinchcliffe, Asset Performance Technologies Inc., Corrales, NM.  Copyright 2004.  For more information about this title and similar titles, please visit www.books.elsevier.com.