Carrying spares for every component in a manufacturing plant is not economical. For some very expensive and critical components like power transformers and large gear reducers, the cost may be so high that you decide to accept the risk of operating without a spare, either installed or in the stores. Modern manufacturing methods, especially those supported by standards such as ISO 9000, have increased machine reliability to the point where carrying a spare is less necessary than it used to be. However, there is still an obvious risk if no spare is kept on hand. Several actions can be taken to minimize this risk.
Consider the three "P"s principle (prevent, predict and prepare) to reduce operating risk. You prevent failure by ensuring the component is as well-protected from damage and wear as is economically possible. You predict failure by using the best available predictive maintenance tools to provide the earliest possible warning of a failure and by examining the conditions under which the component is operating (its "operating context") to determine its probable life. You prepare by establishing procedures and processes that can reduce downtime in the event a failure does occur.
The three "P"s principle is primarily based on reliability-centered maintenance (RCM) fundamentals. RCM is the best way to ensure reliability and is universally employed in the design and maintenance of equipment where a failure may have disastrous results, such as commercial aircraft and submarines. However, RCM in industrial plants is very expensive to implement. A well-designed preventive maintenance program that is followed with discipline is a more economical approach for the general plant. Full RCM and more may well be justified for critical, unspared components.
Now let's look at the three "Ps" in more detail.
The first step is to determine what you want to prevent, predict or prepare for. To do this, the standard technique is to perform a failure mode and effects analysis (FMEA), which is a key component of RCM. FMEA involves examining each component in detail and determining the various ways it could fail. This is a formal analysis that should be carried out by a team of experienced operators, maintenance people and engineers who are thoroughly familiar with the equipment and its operating context.
For example, in an FMEA for a large gear reducer, each sub-component could be examined to identify the possible ways that it could fail, the effect of each failure mode, the probability of each failure mode occurring and the failure development period, which is the amount of warning that can be expected from the time it can be determined that something is starting to fail until a component can no longer perform its intended function.
For each failure mode, if the probability of occurrence is significant, then actions to predict or prevent each failure mode for each sub-component are identified. Obviously, for failure modes that have a very short failure development period, there will be no predictive actions.
Actions to prevent failure may include changing operating procedures (e.g., frequent cleaning, avoiding overloads, etc.), changing maintenance procedures (e.g., increasing oil change frequency, checking the torque of electrical connection fasteners, etc.) or redesigning (e.g., replacing lip seals with bearing isolators, changing lubricant type, installing shear pin couplings or installing cooling systems for electronic equipment).
Actions to predict failures include standard predictive maintenance methods such as vibration measurements, oil analysis, infrared inspections of loaded electrical equipment and other appropriate testing. For critical, unspared components, the frequency and extent of these measurements will normally be greater than for other operating equipment. Not all failure modes lend themselves to predictive maintenance.
The second type of prediction is to obtain an expert opinion on the expected life of the component in its actual operating context. For example, if a large gear reducer is operating in clean and dry conditions, is not subject to shock loading, and is generally loaded below its continuous rating, its life expectancy may be many times the average life for similar components. This knowledge will help to establish the predictive maintenance methods to employ and to assess the value of the various options for increasing protection. The equipment manufacturer is in the best position to assist with this prediction. If this is not possible, a local expert should be approached.
Even after the best preventive measures are in place and predictive maintenance is well-established, there will always remain some possibility of a failure. For critical, unspared equipment, it is worth having a contingency plan prepared and filed under the equipment location number so it can be used if a breakdown does occur.
For instance, when delivery times from the equipment manufacturer are long, one option is to have a local machine shop make a drawing of the component so the shop can quickly make a replacement of at least sufficient quality to allow the part to be ordered and delivered from the manufacturer in the event of a breakdown.
Also, be sure to develop a repair procedure for all likely failure modes and have an alternative component available which could be used long enough to purchase a replacement from the manufacturer. Finally, locate other companies that use the same component and come to an agreement on sharing spares.
At the very least, critical, unspared components should be clearly identified as such so those responsible for their operation and maintenance give them the care they deserve.