I have been working in the field of Maintenance and Reliability for over 30 years now, and I have to say the business has changed significantly in these three decades.
Improvements in technology have not only automated a large percentage of today’s manufacturing equipment, but we can also use various tools, technologies and equipment to detect defects and mitigate failure modes that years ago would have surely ended with a catastrophic failure. Asset Management, formerly known as Maintenance, is becoming more high tech by the decade as companies become introduced the technologies such as Vibration Analysis, Airborne Ultrasound, Motor Circuit Analysis, Infrared Thermography and Process Verification. These technologies, when used as part of a complete Maintenance Strategy or Equipment Maintenance Plan, have helped companies make significant gains in equipment reliability and have also generated some much needed interest from today’s youth and future Maintenance Technicians.
In 30 years, one thing hasn’t changed: Preventive Maintenance (PM). The PM programs of today’s manufacturing facilities are the same mess they were 30 years ago. In fact, some might even be worse.
In the last 15 years, I have worked with clients at well over 300 different sites, and while an impressive number had an above average Predictive Maintenance (PdM) program, not one of these companies had a good PM program.
Now don’t get me wrong, they all had PM programs. But each and every one was a mess in one way or another. As a result, I felt it was important to identify the 10 mistakes that will mess up what could be a good PM program.
- Trying to apply PM to anything and everything – Quite simply, Preventive Maintenance only applies to wear-based failure modes and components that fit the conditional probability of failure distributions A, B, and C (Figure 1). When we try to apply the concept to failure modes and components that fit the other distributions, it will not reduce the number of failures you experience with these components and it will result in increased maintenance costs.
This is one of the most common abuses of PM and also one of the easiest to address. Simply stop trying to inspect or replace components whose failures are random based. When it comes to these components, if we can’t detect potential failures with a PdM task and the frequency of failure doesn’t warrant redesign, it’s time to run the component to failure and look at consequence reduction tasks.
Four years ago, I worked with a pharmaceutical company that put in place a PM to replace 1 input/output (I/O) card for a variable speed drive because it failed once and they lost a very expensive batch of product. In the next two years, they replaced the I/O card as scheduled, and in their minds, the PM seemed to be working very effectively. In the third year, they performed the PM and in the middle of the very first batch, the brand new I/O card failed and, once again, they lost a batch. A new card was installed and two batches later, that card failed again, costing them yet another batch. While replacing the third card in less than two weeks, the technician noticed a loose connection that was causing the cards to fail. How did the connection get loose? From performing the PM, of course!
- PMs are vague – This is the mistake that drives me crazy, and I would also say it has to be the most common. Somewhere along the line in generations since PM was first put into place, it became an unwritten rule that one should never clearly explain what to measure, what condition to look for, or what one should do if they indeed find an unsatisfactory condition while performing a PM. Instead, it must have been decided at some point that all maintenance people know exactly what it means when a PM reads “check the pump”. Let’s be honest, the words “check this” or “inspect this” mean different things to different people. If you want the job done right, you need to be specific. What does the person need to inspect, measure or check, and what needs to be done if they find an unsatisfactory condition?
- The PM Interval is not the correct frequency – As I work with companies facilitating RCM Blitz® and the team determines a specific failure mode is a good candidate for a Preventive Maintenance task, the next step is to determine the task frequency. Nearly 90% of the time, someone will speak up and say, “Well, the MTBF (Mean Time Between Failure) is 12 months, so if we do the PM just short of that, we should be okay.” One important thing to understand, PM frequency should be based on the useful life of the component, NOT MTBF!
- The PM doesn’t appear to be important – If you’re going to take the time to set up a valid PM, make sure that very early on you explain why the PM is important, the failure modes it addresses, and the consequence of not performing the PM as described. If you forget this important step, expect that at some point someone will push the PM aside or delete it all together.
- Confusing Failure Finding tasks with PMs – Without hesitation, I can say that 90% of the people who work in a manufacturing plant don’t know the difference between Failure Finding tasks and PM tasks. As a result, the PM tasks and Failure Finding tasks are all lumped together and we just call them PMs. Failure Finding tasks are time-based inspections of a component that look for failures that have already occurred but are not evident to those who work with the equipment, such as an e-stop. Your equipment will operate perfectly with the item in a failed state, but the consequences of these failures could be catastrophic. Failure Finding tasks are NOT PMs. While they are time based, they rarely have anything to do with wear and belong in a category all their own. (To learn more on Hidden Failures and Failure Finding tasks, I suggest you read my book, Reliability Centered Maintenance using RCM Blitz®.)
- PMs are lumped into one BIG turnaround – I would hope that everyone who works in the field of Maintenance/Reliability/Asset Management knows that, at best, PM only applies to somewhere between 5% and 11% of your components. The reason for the range is that we can now use some type of PdM technology to better manage these components by first detecting failure and replacing or repairing the component before it fails. Yet, many of the major companies around the world still use turnarounds as the backbone of their maintenance strategy, which by the way, is the second most costly way to maintain assets. The most costly, of course, is to run everything to failure. The next best way to waste your company’s money is to plan these huge shutdowns where we believe we can repair and replace hundreds of components over a few weeks’ time, then restart the equipment and everything will run fantastic for the next 2 or 3 years. While this sounds like a grand idea, these turnarounds tend to be way overdone and include hundreds of jobs that have nothing to do with wear. In the end, it’s important to remember that every task included in the turnaround is subject to infant mortality.
- PM data/information is not recorded – It would seem to be common sense that if we have someone perform a PM where they are asked to measure wear that one would not only measure the wear but also record it, or provide comments about what they found when performing the inspection. Unfortunately, this is not the case. Worse than this is when companies use a PM format that doesn’t even allow for comment and then ask what they might be able to do to optimize their PM program. If you really want to optimize your PM program, you need to begin collecting data to make fact-based decisions.
- PMs are pencil whipped – A company I worked with recently told me they did a simple review three times each year to determine what percent of the PMs with time charged against them were pencil whipped (time was charged, but the task was not performed). To do this, they simply added a check box step to a random number of PMs that stated, “I didn’t really do this PM.” In doing this, they found that over 60% of the PMs they assigned were not completed, yet people charged time to them. If this is the case, why have the PM program at all?
- PMs are not completed when due – In order for a PM program to be effective, the PMs must be completed when they are assigned. Companies who struggle with this will also say in the same sentence that their PM program is ineffective. Well of course it is! If you can’t complete your PMs on time, expect that you will still have unexpected failures of these same components. They will also be some of the first companies to tell you they need to reassess or optimize their PM program. The reality is that you need to accredit your PM program. Over the past 20 years, Preventive Maintenance has gotten a bad name because people don’t understand the concept. PMs are still a very important part of a complete maintenance strategy, but you have to do the right things at the right time.
- Anyone is allowed to add a PM – While over 90% of the companies I work with have a formal Management of Change (MOC) process in place, they will allow nearly anyone to add a PM to the maintenance workload without question. The pharmaceutical companies I have worked with over the past decade have some of the worst PM programs I have seen because they will let anyone add a PM to the Computerized Maintenance Management System (CMMS). As a result, I have seen PMs to replace proximity switches and read switches that have absolutely no wear-based failure modes. So now not only do they perform a PM that has no impact on the conditional probability of failure of the component, but they have to shut the equipment down to do so, they replace a perfectly good component, and add in a high conditional probability for infant mortality. Nearly everything wrong you could do in regard to a PM and all because they have no standard process for managing PMs.
As I stated earlier, over the past couple of decades, the words “Preventive Maintenance” have gotten a bad name. It’s important to remember that a good PM program will help work to ensure the inherent designed reliability of your equipment. It’s also important to ensure the people who manage this program have a thorough understanding of where Preventive Maintenance should be applied.
Thanks to my friend Ron Moore for the Age Related Failure slide!
- Audit/ Track
- Failure Finding
- Failure Modes
- Maintenance Planning
- Maintenance Strategy
- Malaysia Flight 370
- Performance Quality
- Predictive Maintenance
- Predictive Techcnologies
- proactive maintenance
- RCM Analysis
- RCM Facilitation
- RCM Facilitator
- RCM Facilitator Training
- RCM Training
- Reactive Maintenance
- Reliability Centered Maintenance
- Reliability Centered Maintenancec
- Reliability Maintenance