was the discovery that there a large number of legitimate ways of managing failures.  These can be divided into five broad categories.

• The first category includes well-known routine overhauls, where equipment is taken to pieces and reassembled at fixed intervals, and cases where components are replaced at fixed intervals without checking their condition. This category is generally known as preventive maintenance, and is occasionally still appropriate, but far less often than first thought.

• The second category covers vast array of techniques known as predictive or condition-based maintenance.  This entails checking equipment regularly to find out if it is in the process of failing, and taking corrective action only if it is needed.  We do this when we check the oil level in our cars - if the oil level is OK we take no action, but if the level is low we add oil. In industry, if sufficient warning that a failure is about to occur can be obtained, then this method provides engineers with time to collect the necessary spares and plan remedial action in a way which minimises disruption.

• The third category covers a special but increasingly common class of equipment. These are protective devices which usually can fail in such a way that no-one knows that they have failed unless someone makes a special point of checking.  These checks differ from predictive maintenance because they entail checking whether an item has failed, rather than checking whether it is failing.  An apparently subtle difference  but one which profoundly affects both the nature of the check and the frequency with which it has to be done.   These checks are known as detective maintenance. Typical examples includes periodic checks on smoke detectors and burglar alarms to find out if they have failed.

• The fourth broad category of failure management techniques is known as corrective maintenance.  As the name suggests, this entails fixing items either when it is immediately evident that they have failed of their own accord (normally referred to as a 'breakdown'), or when they are found to be failing following a predictive maintenance check, or if they are found to be failed after a detective maintenance check.

• The fifth and final way of dealing with failures is to change of design of the asset in such a way that the failure no longer occurs, or if it does occur, to change the consequences of the failure in such a way that the failure no longer matters.  Redesign usually is the last resort.  This is because the engineer on duty today has to maintain the equipment as it exists today, and not what should be there or what might be there at some time in the future. Since redesigns take time, we only challenge the design when a cost-effective maintenance programmes cannot be found.

Once the early researchers had gained a clear understanding of the strengths and weaknesses of these different approaches, the next step was to develop a sensible basis of deciding when to use which one. The first point which became clear was that it is too simplistic - and too dangerous - to haphazardly choose any one of these approaches. Different approaches are needed for different machines, or even for different parts of the same machine.  Think again about a private car.  We check the brakes and tyres for wear on a regular basis - both forms of preventive maintenance, while checking periodically to see the hazard warning flashers are predictive maintenance.  On the other hand, changing the oil and the spark plugs at fixed intervals is still working is detective maintenance.  Then again, we would probably do not routine maintenance at all on (say) the cigarette lighter or the electric window winders , and only arrange for them to be repaired if they fail.

In trying to establish a sensible frame work for making these choices, the aviation industry realised that the reason why we worry about failures at all is because they have consequences.  Sometimes failures only cost the fund of  the repair job.  More often, they interfere with production or operations, in which case they usually cost rather more than the direct cost of repair.  In the most serious cases, they can lead to serious environmental incidents or kill people directly.

Clearly, the more serious the consequences of a failure, the more time and effort should be spent on trying to prevent it.  This led the airlines to use a formal evaluation of failure consequences as a basis for deciding what maintenance should be done. The decision-making framework which they developed to do this is at the heart of Reliability-centred Maintenance.

The application of RCM starts with an analysis of the functions of the equipment. This is done in close consultation with the equipment users.  Maintenance is all about preserving the function of physical assets, so a comprehensive review of equipment functions enables everyone to agree on what maintenance is trying to 'maintain’.  It is also surprising - sometimes very surprising - how much people learn about how the equipment is supposed to work.

Once the functions of the equipment have been agreed, the next step is to analyse all the ways in which it is reasonable likely to fail, and what would happen if each failure actually did occur.   This step is called failure modes and effects analysis, and it enables everyone to agree on what they are actually trying to prevent when they do 'preventive’ maintenance.

is to assess the consequences of each failure in a strict sequence. Depending on both the nature and the severity of the consequences, the final step is to select the most appropriate of the five types of maintenance listed above for dealing with the failure.

This process leads to much more tightly focused and far more effective maintenance programmes.  RCM has contributed to a massive decrease in the number of civil aircraft accidents over the past thirty years - from more than sixty crashes per million takeoffs in 1960 to less than two per million takeoffs in 1988.  Such results led to its widespread adoption in ground-based industries around the world. RCM has and is being used to initiate step changes in maintenance effectiveness - often in the space of a few months - by someone the world's leading companies in fields such as petrochemicals, food manufacture, pharmaceuticals, railways, electricity generation and distribution, mass housing, steel making, water distribution, military undertakings and auto-mobile manufacture.

Given the power of the RCM process and the speed with which it can produce results, the final questions concern how and by whom it should be applied. Most people tend to believe that the equipment supplier is in the best position to provide a viable maintenance programme for his equipment.  However, no supplier - indeed no outsider - can possibly appreciated all the unique features of the user's business which will affect the machine throughout its life, nor can they fully appreciate all the ways in which the failure of their equipment might affect the users business. This means that although the supplier can - and in the case of new equipment, should play a part in developing such a programme, there is no way that he can do it all.

In practice, the organisations which have derived the most value from the RCM process are those which have developed the capability to apply it themselves. They do so using multi-disciplinary teams of people trained to use RCM under the guidance of even more highly trained individuals known as RCM facilitators.  These teams usually include members of the operations department, because they understand most clearly what the equipment is to be used for and what the consequences are if it fails.  The equipment users are the 'customers’ of the maintenance department, and it is wholly in keeping with all the principles of Total Quality that the customers should be involved in specifying the kind service they expect.  The teams also include members of the maintenance function, because they tend to have the clearest understanding of what fails and what can be done to prevent or repair it.

Getting members of operations and maintenance departments to apply RCM together in this way not only ensures that the 'best' maintenance programme is drawn up with all the available information.  It also causes two notoriously hostile sections of the business to start working together as a team to fulfill common, clearly understood objectives. This feature of RCM  alone is causing it to be widely adopted by some of the world's largest multinational manufacturing organisations.

RCM is transforming the world of industrial maintenance from what is often seen as an expensive and erratic drain on resources, into truly proactive contributor to world-class manufacturing excellence.