When attempting to maintain a large industrial facility we face a complex challenge. We search yet rarely discover a single silver bullet capable of subduing the beast. The fiend has a hundred thousand spikes with which to retaliate. Tens of thousands of failure modes stalk our physical assets. To win against so tenacious a foe we require no less than a disciplined and fully developed military-like strategy. Know thy enemy – his relentless patience, his cunning, and his ability to call upon every law of physics to ambush our plans and attack our profit margin.
The unobtrusive name “Reliability Centered Maintenance” or RCM provides hardly a hint of its power to enable an organization to attain its maintenance objectives. Those maintenance professionals who have not yet wielded that power, easily recline to the position that considers RCM yet another of many maintenance philosophies. They are dangerously mistaken and in peril of allowing months and years to slide by before the inevitable realization that will drive them to embrace RCM in their own organizations.
RCM is a fundamental technology whose practice exposes the basic principles governing harmonious long-term human-machine interaction. Yet its hidden strength can be easily missed in a cursory reading of the textbooks on the subject.
John Moubray intelligently packaged the required training and tools that an organization needs before it can deliver the logical set of maintenance tasks to accomplish its corporate mission. One may, using RCM worksheet documents, trace each task executed within the RCM developed maintenance program to a particular failure mode, its effects and its consequences in order to ascertain that the right job is to be performed.
RCM thinking, though absolutely scientific and logical, is in many respects, counter-intuitive. Therefore maintenance professionals, tradesmen, and managers, in order to benefit, must shed their long held views and habitual reactions regarding maintenance. They accomplish this difficult feat through phased education and training. RCM education uncovers the realization that the right maintenance activity addresses the preservation of function. An obvious proclamation, yet, to his astonishment, the maintenance professional discovers that the functions of the machinery under his control were inadequately or incompletely identified. Consequently the failures of those functions and their causes have by-in-large escaped his conscious effort to deal with them. That reality has excluded him from transacting directly in the currency of maintenance – the failure mode, the cause of a specific failed state.
RCM transports our attention to previously ignored functions, their functional failures, and their failure modes. Then it reconstructs our entire philosophical framework using a precise language and methodology of logical inference. Old practices, familiar experiences and intuitive notions begin to coalesce around this new skeleton and the eminent appropriateness of RCM to our own situation dawns upon us. The education of RCM is an experience of self-discovery.
The astonishing power of RCM lies in its ability to unleash previously unused strata of intelligence and unravel the combined knowledge of years of observation by operating and maintaining staff. The vehicle for accomplishing the miracle, the facilitated review group, is formed around a master plan aimed at a set of specific corporate improvement objectives.
RCM illuminates maintenance concepts and principles that, heretofore, have been relegated to the back of the drawer. With RCM one revisits the idea of equipment lifetime. A landmark study by Nowlan and Heap (Ref. 2) revealed that every physical item is subject to one or more failure modes each of which exhibit 1 of six distinct patterns of failure (Fig 1).
Fig. 1 The six failure patterns
A revealing way of looking at equipment lifetime is to plot a (random) variable known as the conditional probability of failure (called H in Fig. 1, see also the article Time to Failure). The conditional probability of failure is the probability that an as yet unfailed item will fail in the upcoming short time interval. The maintenance planner or manager has a decision to make – whether to preventively intervene and renew an asset or to leave it alone until the next (relatively short) interval has elapsed. The gravity of the possible consequences of failure and the conditional probability of failure, H, are to be considered for the decision.
The conditional probability of failure and the 6 patterns of failure discovered by Nowlan and Heap make the notion of ‘useful life’ glaringly apparent. The latter three (patterns D, E, and F) show that the component (with a single dominant failure mode) is not more likely[1] to fail at a later time than at the current time.[2] Components governed by these failure patterns have no definable “useful life”. Hence preventively renewing or restoring such a component is unproductive. And in the case of failure pattern F a maintenance intervention is counter-productive, actually introducing a higher risk of failure (i.e. ‘infant mortality’ resulting from installation errors or material defects).
By analyzing behavior at the failure mode level, RCM takes unique advantage of the 6 distinct patterns of risk behavior. It channels maintenance effort to where it will do the most good. A centrifugal pump, for example, may fail due to one of several causes (modes). The mode “impeller wear”, however, adheres undeniably to failure pattern B (Fig. 1) because it experiences a long running period during which the chances of failure is relatively small followed by a short period (known as wear out) where the probability of failure increases rapidly with age. Knowledge of the failure pattern is of immense help in the next logical step in determining a maintenance tactic. Knowing the failure pattern (one of A to F), the effects, and consequences of the failure mode one proceeds to select a maintenance task (preferably a condition monitoring task if one can be found which is both technically feasible and worthwhile, or an age based renewal in the case of failure pattern B) or an operational or physical modification that will eliminate, reduce, or otherwise mitigate the consequences emanating from that failure mode.
Failure mode oriented maintenance management contrasts strikingly with the usual, more intuitive practice of considering each component and attempting to arrive at a single appropriate proactive maintenance task – an impossibility given that a single unit under analysis suffers from several functional failures each in turn responding to several failure modes each with its own failure pattern. Hence the diametric shift in viewpoint – from component to function. As an example, one of the functions of a rail car suspension system is to isolate riders from impacts due to track irregularities while another is to minimize oscillations following such impacts. A variety of components contribute to one or both of these functions and a failure of either can be caused by a particular combination of events in one or more components (oil leak in a horizontal or vertical damper, a loosened damper mounting bolt, or a damper one-way valve failed open). Without doubt, starting the RCM process with a functional analysis broadens the perspective of the analysts by considering all the ways (failed states) that each function can fail to perform as required.
Working from this new perspective one moves systematically from function to failed state to failure mode. After resolving the first failure mode one then proceeds to a second that may exhibit an entirely different failure pattern. It may, indeed provoke entirely different effects and thereby engender entirely different consequences. Clearly, by basing our approach on failure modes derived from failure states and functions rather than on the component, one may systematically construct a comprehensive and defensible maintenance policy for handling each failure mode. When all the reasonably likely failure modes are thus dealt with, the asset management plan is complete.
In the world of RCM, a “comprehensive failure management policy” is developed first by systematically asking what the consequences of each failure mode are likely to be. The next steps are to determine what tasks could be done to anticipate, prevent (or in some cases, detect) each failure mode from a purely technical viewpoint, and then to ask whether each technically feasible task is able to reduce the consequences of the failure to an extent that makes the task worth doing. If a task cannot be found that is both technically feasible and worth doing, the final step is to determine what other “default” action (if any[3]) should be taken to deal with the failure mode. Default actions usually consist of redesign or changes to operating procedures.
Fig. 2 The RCM review group (Ref 1)
The RCM process lives or dies within the structure of “facilitated review group” meetings first advocated by Moubray in 1986 (Ref. 3). An unguided or misguided RCM project can drown in a thousand ways. If we neglect to expose all the functions of a complex system, if we fail to identify all the possible failed states for each function, if we do not pin down the reasonably likely causes or modes of each failed state, we will lose the helm and any hope of arriving at the appropriate failure management policy. Traps await us at each curve and bend of the typical RCM analysis. The RCM review group facilitator holds the map by which to circumnavigate the obstacles that would otherwise stall or even halt our progress.
Successful completion of an RCM analysis delivers a logical, rational, and defensible set of proactive and reactive maintenance tasks as well as appropriate design and procedural modifications. That desired result depends entirely upon the synergistic interaction of review group members. The qualified RCM facilitator effects and sustains constructive interplay among group members, heading off negative forces while guiding them past blind alleys and time killing forays into fruitless detail. The trained facilitator recognizes when emotional attachments to past viewpoints impede progress. He or she acts quickly and sensitively to such situations. The qualified facilitator knows when to intercede and when to retreat into the background while the group experience flushes out the technical details that the RCM process requires.
You should cultivate RCM facilitators from within your organization. The RCM program sponsor must ensure that RCM facilitators are properly selected, adequately trained, and mentored during their maiden review group sessions. Moubray (Ref 1) has identified 45 distinct skills that an RCM facilitator needs to acquire before he or she can be relied upon to conduct a safe and defensible RCM analysis. No facilitator should be left on their own to lead the application of RCM until he or she has been brought to an adequate level of competence in all 45 of these skill areas by a suitably qualified mentor.
Once you have arranged for your facilitators to be trained, you need to plan, resource, schedule, and execute the RCM analyses review group meetings. Except for the facilitator, individual review group members only need devote a few man-days to each RCM analysis, either all at once or spread over a few weeks. This minimizes organizational disruptions. Within weeks or months of implementing RCM you should begin to realize tangible benefits, sometimes justifying the entire cost of the RCM program even before the RCM recommendations are actually implemented. The latter occurs often as a direct result of revelations within the review group meetings leading to the elimination of some chronic problem which had never before been so fundamentally addressed.
RCM analysis on a company wide basis delivers not only a maintenance management policy which aptly targets corporate objectives, but equally important, a policy that is owned by the very people most directly responsible for the effectiveness of all physical assets – the maintainers and the operators.
References:
- Reliability-centred Maintenance, John Moubray, 2nd edition Butterworth Heinemann
- Reliability-centered Maintenance, Nowlan FS and Heap H, Springfield, Virginia. National Technical Information Service, United States Department of Commerce
- Reliability-centred maintenance, A Conference on Condition Monitoring, Gol, Norway, 2-4, November 1987
© 2011 – 2014, Murray Wiseman. All rights reserved.
- [1]in the conditional sense. (See next footnote)↩
- [2]That is to say, if we ask, “What is the probability of failure in the upcoming interval”, the answer will be the same as that when we ask the question at a future time. This is the case even though failure probability (as opposed to conditional failure probability) actually increases with age.↩
- [3]The analyst may legitimately decide that no task at all be recommended if the failure mode’s consequences and occurrence rate are acceptable.↩