Sometimes called "Preventative Maintenance", a Preventive Maintenance program tries to maximize equipment availability and minimize product defects by taking care of the equipment before the problems arise.
The costs of preventive maintenance include:
The benefits of preventive maintenance include:
Repair costs are often, but not always, higher than the cost of preventive maintenance. Neglecting to lubricate can result in a seized motor, for example. The extra costs are generally due to collateral damage when the machine fails, plus the need for more highly trained repair personnel, at higher rates, for longer periods of time.
The maintenance activity might prevent product defects caused by excessive wear on a saw blade, or gradual misalignment due to vibration. Therefore, a step such as recalibration might be included in the preventative maintenance schedule.
Safety may be an issue, depending on the equipment's failure mode. An automotive example is that neglecting tire pressure may lead to under-inflation, premature wear, and a blow-out. A factory example is a worn drive belt that might snap.
The words "schedule" and "plan" imply that the maintenance activity will be calendar-based: daily, weekly, monthly, etc. This is certainly one way to plan for maintenance activities.
Some machines should undergo maintenance based on other scheduling factors:
A saw blade, for example, gets dull with use. Assume it does not rust due to extended idleness. If it cuts the same type of material at the same pace every working day, then the calendar days, hours of operation, and number of units will all point to the same maintenance cycle. If this saw is only used for one type of product out of many in the factory, and sits idle for varying periods of time each day, then the "hours of operation" or "units produced" are more appropriate.
Some machines, especially trucks and airplanes, are inspected and maintained before each use.
In all these situations, the production schedule for a machine can predict its maintenance schedule, once the rate of wear-and-tear has been determined. The production planner, or the planning system, can raise the work orders as soon as it sets the schedule.
Here the plan is to perform preventive maintenance in reaction to conditions observed about the machine or its performance. Examples include:
This is planned maintenance, but it does not derive a schedule from the production schedule. It is "preventive maintenance" rather than "unexpected repair" because the machine is still functioning, and its outputs are still within specifications. The operator or quality inspector raises the work order.
Condition-based maintenance for a machine with no internal sensors relies on operator observations and measuring the products against specifications. A machine that "sounds the alarm" based on internal sensors is generally asking for repairs, rather than raising a flag for preventive maintenance.
Predictive maintenance supplies a longer lead time using computer analysis of the outputs from a number of internal sensors plus measuring the "drift" of products from specifications. The concept is similar to condition-based maintenance except for extending the lead time.
To gain better results by these methods requires a greater investment in sophisticated sensing devices and, generally, in better computer modeling of the machine and its failure modes.
Generally, an analysis of the history of repairs and condition-based maintenance for a machine can lead to near-optimal planned scheduled maintenance. Predictive maintenance can do better yet, because it reacts to actual deterioration but provides similar lead times. In either case, the longer the lead time, the better use of scheduled downtime for several machines.
The goal is to balance the costs and benefits of a preventative maintenance program. At one extreme, you could run every machine until it breaks down: your preventative maintenance cost is zero, but there are serious problems with repair costs and the factory schedule.
The other extreme would have 100% maintenance with no productivity. Examples of reasonably high maintenance are truck drivers who inspect their rigs daily, or factories where the final few minutes of each shift are dedicated to minor maintenance and clean-up.
A maintenance manager might tweak the schedule on an ongoing basis, stretching the time between maintenance until problems are noticed, and then cutting back. This approach might be limited by the maintenance schedule built into the warranty agreement for that machine.
Investing in sensors and computer systems for condition-based or predictive maintenance can be expensive. Some industries rely on high overall equipment effectiveness (OEE), so they need to minimize downtime. Or there may be serious problems in an assembly line if one machine fails. For example, if an item is cooling (or warming) as it travels so it will be at an optimum temperature for the next step, then delays will lead to incorrect temperatures and, presumably, product defects or re-work to bring the item back to proper temperature. An assembly line that can shut down without this kind of problem might not benefit from an expensive predictive maintenance system.
Start with the maintenance plan recommended by the manufacturer, especially if the warranty has any value.
Adjust for the historical data: what repairs have been made? When? Could they have been prevented by timely maintenance?
The Reliability Centered Maintenance (RCM) approach is rather onerous but provides a thorough way of determining what maintenance is required to prevent equipment failure. It is appropriate when the potential damage from a failure is too great to risk.
Set a schedule, with several goals in mind:
How can one improve on "Check and, if necessary, fix"? Consider "Check and, if necessary, raise a work order". That new work order can be prioritized and scheduled. Another common sense approach is that a very brief fix, such as "If the bolt is loose, tighten it" can be scheduled as "Tighten all loose bolts".
As noted earlier, some shops do light maintenance at the end of each shift: wipe down, add lubricants, check for wear and tear, and tidy the workplace. This may take only ten minutes, but benefits the next shift's workers and helps keep the machinery in good repair.
You may also want to read about Reliability Centered Maintenance.
By Oskar Olofsson
Succeeding With Lean Leadership
Succeeding With 5S
Succeeding With Standardized Work