Revolutionising maintenance with RCM

Over the years, several myths and misunderstandings have arisen about reliability-centred maintenance (RCM): what it is; whether it consumes too much resource, whether it can be applied to all types of assets including structures; whether the ends justify the means.

Against a background of having to do more with less, RCM offers a proven and robust means for industry to obtain maintenance 'value for money' and, in parallel, improve operating safety, system reliability and platform availability.

The conditional probability of failure for equipment types was believed to increase around a specific number of operating periods.

The history and development of RCM

Recognition of the complex nature of aviation equipment failure culminated in a new approach to the development of maintenance programs, first trialled at Boeing in the late 1960s. This methodology, known as MSG-1, recognised that:

• Scheduled overhaul had little effect on overall reliability of a complex item unless there was a dominant age-related failure mode;
• The intrusive nature of the overhaul activity itself was the cause of unreliability; and
• There are many items and failures for which there is no effective form of scheduled preventive and/or predictive maintenance.

In the early 1980s, RCM was first applied in the South African mining industry. Through this pioneering work, in 1990, the industrial version of RCM known as RCM II which over the past decade or so has become the standard approach, was adopted throughout the world.

So what is RCM?

RCM is a process used to determine the maintenance requirements of any physical asset so that it fulfils its intended functions over the life cycle and in its operating context. The RCM process must therefore start by defining user requirements or 'functions'.

This in itself is usually something of a challenge for most organisations. Unless this user requirement is understood, it is hardly surprising that operators and maintainers have difficulty agreeing and communicating on when equipment failure has occurred.

RCM asks the following seven questions from which a comprehensive approach to failure for the asset can be developed:

• What are the functions of the asset in its present operating context?
• How can the asset fail to fulfil each function?
• What would cause each functional failure?
• What happens when each failure occurs?
• In what way does each failure matter?
• What can be done to predict or prevent each failure?
• What should be done if no suitable proactive task can be found?

The first four questions develop a functional failure modes and effects analysis (FMEA) and the last two define the appropriate failure management policy. The vitally important fifth question determines how we should react to the failure in relation to whether the failure is 'hidden' or 'evident' and whether safety, the environment or operations are affected. These seven questions can only sensibly be answered by people who know the asset best; this includes maintainers and operators, supplemented by representatives from OEMs. The group (a typical example of which is shown in Figure 2) is guided through the RCM process by a competent 'facilitator' who is an expert in the RCM process and its application rather than the system expert.

RCM falls squarely within the logistic support analysis (LSA) process and provides inputs which are needed for a rational approach to spares, tools and skills determination.

Application of RCM produces a 'safe minimum' maintenance program which includes:

• A comprehensive range of failure management tasks for maintenance and operations staff (incorporating predictive, preventive, detective maintenance as well as the foundation for the development of all likely corrective tasks);
• Mandated and recommended redesigns of either the asset or the way it is operated or maintained; and
• Recommendations for 'no scheduled maintenance' or run-to-failure which require the development of strategies to deal with such failures as they occur.

The process has been applied widely to mechanical, electrical and electronic systems as well as to platform structures.

The application of RCM in Australia and New Zealand

In the 1990s, many organisations experimented with RCM to address excessive maintenance manpower and resource costs and to develop a more defensible approach to risk management and operational reliability. Since RCM was a 'new' technology for industry, although well established in many other applications, success was not always guaranteed.

A number of early applications failed to achieve the full benefits. The causes included a lack of understanding of the process, failure to obtain strong commitment from senior management and a lack of understanding of the importance of training.

Many organisations believed the process would provide them with a magic-pill quick fix but, when these organisations started to apply the RCM process in a rigorous manner, it become evident that many of their maintenance strategies were potentially and sometimes profoundly wrong. Instead of realising that their previous thinking was out of date and flawed, many found this revelation to be quite unpalatable and proceeded to either condemn the RCM process or sought ways to streamline the process.

Many lessons have been learnt since the early applications of RCM in Australia and New Zealand but what remains unchanged is the dramatic and sustainable impact true RCM has on maintenance strategies.

The standard for RCM

The benefits that are now regularly obtained from correctly applied RCM are usually so spectacular that it is incredible that some still choose to abandon the proven RCM process and modify it in an effort to achieve the same apparent result with less effort. While certainly a noble goal, every improvement to date has compromised the outcomes achievable with RCM, some so much as to make the output worthless.

Since its birth in the early 1960s, the RCM process has suffered a long list of modifications that have seen steps in the process deleted, the logic modified to make it fit a flawed understanding of equipment failure and software written to try to make the actual RCM decisions based on incomplete or flawed data.

In August 1999, a definitive RCM standard to clearly define an RCM process was released.

While this standard, SAE JA 1011, has achieved a slowing of the proliferation of RCM clones and a differentiation between RCM offerings, there are still enough flawed processes being applied, mostly out of ignorance, to cause considerable anxiety and to make our society less safe than it could be.

What is of particular concern is the effort that is applied to complete a non-complying RCM process is usually as much, or in some cases more, than the effort required to apply rigorous RCM.

A comparison trial of RCM II

The largest comparison trial of RCM II was conducted in the late 1990s on a group of large and complex assets. The comparison was between a maintenance regime developed using RCM II with the existing regime which had evolved over the life of the asset.

Of 15,000 failure modes examined in the sample analysis group, it was found that just 20 per cent responded to some form of condition-based maintenance, while 'traditional' scheduled restoration or discard was only appropriate for just 6 per cent of all failure modes. A further 18 per cent of the failure modes were associated with protective devices which required detective maintenance or 'failure finding' as this is known in RCM. Two-thirds of the failure modes did not respond to any form of proactive maintenance activity. These findings closely aligned with the initial experiences of the aviation industry.

Results from the implementation of RCM derived maintenance strategies were divided into three categories: operational costs, availability and impact on support functions.

• Operating costs for the RCM control group showed a 19 per cent reduction, despite a stores supply problem which pushed up spares cost for the control group. Discounting this effect, savings in maintenance effort for the control group were 33 per cent lower.
• Availabilities of the control group and the non-RCM assets were comparable, although the control group was negatively influenced by the stores problem above. Control group availabilities have improved as stores processes have been improved.
• The most significant changes have been a shift in maintenance effort from offline to online maintenance, and the removal of large work packages associated with major shutdowns. This has required the development of a more dynamic relationship between operations and maintenance. The associated review of the maintenance cycle has recommended slightly more frequent, but shorter maintenance stops - with an associated increase in availability.

Overall costs for running the RCM program amounted to just under AU$5 million. Across the assets, savings of around AU$12 million per annum were achieved, providing a return on investment measurable in months.

From this trial, confidence in the RCM analysis and implementation led to a further RCM analysis which was completed in mid-2002. In this study, over 31,000 failure modes were examined over 377 separate systems with, once again, a major shift towards condition-based maintenance with only 14 per cent of failure modes responding to some form of scheduled restoration or discard (overhaul) activity. Benefits obtained so far include:

• Removal of maintenance with no value;
• Fewer overhauls and reduced requirement for maintenance shutdowns when compared with the former maintenance cycle.
• Reduced time needed for testing and tuning;
• Much improved online fault diagnostics available through the RCM analysis and associated database and maintenance management software packages;
• The acquisition of accurate failure rate data for specific equipment failure modes which is better enabling the determination of spares requirements and supply chain location.
• An overall maintenance cost reduction for this asset group estimated at AU$35 million.

Additional benefits are anticipated with availabilities expected to increase by around 10-15 per cent. From a commercial perspective, the RCM project has provided a return of about 80 per cent over this time period at current interest rates.

Not surprisingly, faced with this type of saving, many disbelievers in the RCM process argue that reductions in the level of proactive maintenance will be quickly followed by increases in reactive maintenance. Experience in Australian and New Zealand industry confirms the improvement sustainability.