A good preventive maintenance program can save your organization significant expense. It’s just as important, however, to also eliminate preventable maintenance.
Preventable maintenance can be defined as any maintenance task that doesn’t add value. These activities suck up time and resources but do not accomplish anything in the way of preventing failures. Since that’s the only reason we perform maintenance in the first place, it’s in our best interest to identify and eliminate as many of these activities as possible.
Your preventive maintenance programs are the most fruitful hunting ground for unnecessary maintenance tasks. This is simply because other maintenance tactics don’t have any “spare” tasks.
When reactive maintenance comes up, you’re going to do whatever it takes to rectify the situation. The most common way to reduce reactive maintenance is to avoid it in the first place, usually by building preventive measures into our PM programs. Overall, preventive maintenance costs less than reactive maintenance.
In the case of predictive and prescriptive maintenance, the fat has already been trimmed. The entire point of those tactics is to only perform maintenance when it’s necessary to prevent a failure.
In fact, if we eliminate all the unnecessary maintenance tasks from our PMS, we’ll likely find that we’re now practicing predictive maintenance. This is because reducing those tasks down to zero will probably require condition-based monitoring (CBM), either through sensors or inspections. This is an ideal state but running CBM on all your assets probably can’t be justified. A full predictive program is usually only worthwhile for production-critical assets that cannot abide failures.
This leaves us with preventive maintenance to search for efficiencies. Broadly speaking, there are two different ways we can make preventive maintenance more efficient and eliminate preventable tasks. The first is to stretch out the cycles (either time or usage) that we’re using to schedule maintenance. This is appropriate in some situations but not in others. The other is to eliminate the task completely.
Next, we’ll look at some of the techniques we can use to reduce waste in our PM programs.
For a very long time, conventional wisdom told us that the best way to avoid failures was a rigorous program of time-based preventive maintenance. Conventional wisdom was wrong. In fact, this idea was so far off the mark that it was proven incorrect twice, by groups working completely independently from each other.
Stanley Nowlan and Howard Heap released their groundbreaking report, Reliability-Centered Maintenance, in 1978. Produced under commission from United Airlines, the Nowlan and Heap report proved that the two central assumptions underlying maintenance at that time were not correct:
Nowlan and Heap showed that both of those assumptions are wrong. The principle that failures could be prevented by a stringent enough overhaul schedule was a fundamental aspect of most maintenance up until that point. Unfortunately, it is fundamentally untrue. To quote Nowlan and Heap directly, “Over the years, however, it was found that many types of failures could not be prevented no matter how intensive the maintenance activities.”
Their report further noted that “…only 2% of aircraft components have failures that are predominantly age-related…and that 68% have failures that are primarily infant mortality.”
What Nowlan and Heap showed, above all else, was that a maintenance program based exclusively on operating age would not have a significant impact on the failure rate, regardless of how frequently you set the intervals.
What Nowlan and Heap couldn’t know was that they were essentially replicating work that had been done decades before. British scientist C.H. Waddington made the same discoveries in 1943, again while investigating the reliability of aircraft. Waddington oversaw a group of scientists during World War II that had been charged with advising the military on more effectively combatting Germany’s submarine fleet. The group made several recommendations that defied conventional military wisdom at the time, such as painting bombers white instead of black and setting depth charges to explode at 25 feet instead of 100.
These recommendations later proved their efficacy, despite howls of protest from military commanders. Those objections were relatively mild, however, compared to the outrage that followed from their recommendations regarding maintenance.
Briefly, British Liberator bombers were spending too much time in the shop and too little time hunting subs.
When Waddington and his team started looking at the numbers, they discovered something shocking. Not only was scheduled maintenance not preventing failures, it seemed to be driving them upward! Each bomber was given scheduled maintenance after 50 hours. What Waddington discovered was that the number of unscheduled repairs increased sharply immediately following these scheduled maintenance sessions. Unscheduled repairs then declined steadily until the next time the aircraft had received its scheduled maintenance, at which point they would spike back up.
This wasn’t an isolated occurrence limited to a single bomber. The pattern repeated itself throughout the fleet. The solution was counterintuitive but inescapable: the scheduled maintenance sessions were increasing breakdowns. As Waddington put it, “…it is doing positive harm by disturbing a relatively satisfactory state of affairs.”
This work by Waddington and his team would remain classified until well into the 1970s, which is why Nowlan and Heap had to rediscover it independently. It’s a fun bit of trivia that the essence of reliability-centered maintenance (RCM) was once considered a military secret.
In brief, RCM is a process that ensures assets operate at optimal levels and determining the most effective maintenance methods for that equipment. Every RCM process rests on stages that must be followed in order, as each stage relies on information gathered in the previous stage.
1. Functions. This first stage identifies equipment and assets not by what they are, but by what they do.
2. Functional Failures. This stage documents how the equipment can fail. This must include both total failure (the equipment doesn’t work at all) as well as partial failure (equipment not delivering expected performance.
3. Failure Modes. Failure modes are the events that cause functional failures. Every functional failure will have at least one failure mode and could have dozens depending on the nature of the failure.
4. Failure Effects. This stage describes the impacts on the equipment itself and the larger operating context when a failure mode occurs. You must make sure to include how to gather evidence that a particular failure has occurred, threats to personal safety or the environment, impact on operations, damage or potential damage resulting from the failure, and the steps necessary to return the equipment to functioning stats. You will need all this information for the next stage.
5. Failure Consequences. This stage reveals how much the failure matters to the organization from a holistic perspective. In the simplest possible terms, this stage asks and answers the question “why does it matter?” Avoiding failure consequences could be said to be the entire basis of RCM.
RCM is a complex topic, but you can pick up a quick primer in “5 Key Concepts to Better Understand Reliability Centered Maintenance.”
The concepts of lean manufacturing can help direct our efforts. The essence of lean processes is to eliminate any part of the process that doesn’t add value to the customer. Maintenance and production use different processes and have different goals, but we can still use the lean methodology to improve maintenance.
In “The 7 Wastes of Enterprise Maintenance Departments,” we identified some of the most common wastes of time and budget found in asset management:
1. Not dealing with root causes of maintenance problems
2. Useless maintenance tasks
3. Lack of structured maintenance procedures
4. Time to repair/delays
5. Spare parts availability
6. Cost control
7. Get caught in day-to-day activities
There are plenty of efficiency gains to be made in each of these areas, but right now we’re going to concentrate on number two, “useless maintenance tasks.”
It might be better to say, “inappropriate maintenance tasks.” They are rarely useless, but instead are applied at the wrong time or to the wrong equipment. There are many different failure patterns, and time-based preventive maintenance is only effective at preventing a small percentage of these.
Reliability-centered maintenance takes a more holistic view by analyzing every asset function, its operating context, criticality to the production process, and related safety issues. Once this information is known, it becomes possible to select the appropriate maintenance tasks for every component, and the best intervals at which to conduct them.
The actual actions themselves vary wildly. In some cases, the very best approach is to literally do nothing until the component fails. At that point you replace it and go on with your life. In other cases, you may discover that there is no way to properly maintain the component, but nor can it be allowed to fail. Those cases call for a redesign.
Most cases, though, fall somewhere in the middle. Knowing the failure patterns will help you design plans to get the most life out of those components with the least expenditure of maintenance budget.
One way to slim down your preventive maintenance is to make use of our old friend, the Plan-Do-Check-Act (PDCA) cycle, also known as the Deming cycle and by a host of other names. Using PDCA techniques, you can determine more accurate usage-based schedules for your preventive maintenance tasks.
The simplest way to go about it is to target one task performed on one kind of asset. For example, perhaps we want to reduce the frequency of oil changes performed on our compressors. Let’s say we are currently performing these oil changes at 30-day intervals. We decide to extend that to 40 days.
We’ve gone from performing a little over 12 oil changes a year to just over nine per year. The savings practically compute themselves.
This puts us squarely in the middle of the PDCA cycle. We have planned, and we have done. Now we must check.
We let the new oil change program run for a certain amount of time, say six months. At this point, we check to see if the failure rates have increased, decreased, or stayed the same. This will give us the information we need to move onto the act phase.
If they’ve increased, and we can tie those failures back to a lack of oil changes, then we must decrease the interval we’re using. However, if they’ve decreased or stayed the same, we can probably extend the interval once again, this time to 50 days. And the cycle begins again.
You can use this methodology to fine-tune your preventive maintenance intervals on any asset, but it’s best suited for use on those assets that are not critical to production. For production-critical assets, we’d be better off turning to Condition-Based Monitoring (CBM).
CBM focuses on preventing equipment failure while also reducing or eliminating unnecessary maintenance tasks. CBM in maintenance is a key tool in moving from preventive maintenance to predictive maintenance, as it dictates that maintenance should only be performed when certain conditions are met that indicate a failure is likely (or inevitable).
CBM data is gathered by sensors or inspections, providing an ongoing picture of asset health. The goal is to spot an upcoming failure so maintenance can make sure it does not occur. Common types of CBM include oil analysis, temperature readings, and vibration monitoring and analysis.
This can be considered the holy grail of eliminating unnecessary maintenance. CBM allows you to deliver targeted maintenance to specific assets at exactly the right time, not before, and certainly not after when it’s too late. You get the maximum life out of your parts at a lower cost than you can possibly achieve with either time or usage-based cycles. By connecting a solution like Prometheus IIoT, you can even connect the events detected by your sensors to actions within your CMMS, such as automatically issuing work orders, material requests, and so on.
The downside is that CBM itself can be very expensive. It’s generally only suited for assets that are both production critical and too expensive to maintain a backup unit. For more on CBM, we suggest “The Importance of Condition-Based Monitoring” and “5 Ways to Identify the Status of Your Industrial Assets for Enterprise Maintenance Management.”