When Preventive Maintenance Doesn’t Work

“Sometimes, preventive maintenance just doesn't work. That's when it becomes time to stop the collective insanity and start learning from others.”

For many years, I’ve performed and managed preventive maintenance (PM) on every type of equipment, never asking myself why the equipment still fails, even after I’ve performed PM. My compliance rates were always high, but so were the number of recurring equipment breakdowns.

Finally, I asked myself how it is possible that a maintenance professional could perform the same PMs on equipment that continues to fail.

I know now that my PM program was flawed because it was essentially a reactive maintenance program that relied mainly on time-based PM tasks following manufacturers’ suggestions and stuff we learned along the way. I had no technical justification for any task other than “we always do it this way,” or “it’s the latest predictive technology,” so “we can’t stop doing it now or we’ll risk more failures.” If that isn’t the definition of insanity, I don’t know what is.

The research that changed the way I think about failures and PM actually started more than 30 years ago, yet many plants are still falling apart today. It’s time to stop the collective insanity. If you face the same problems on a daily basis (sometimes with little hope in sight), then read on, because I found a solution -- and you can, too.

Research on equipment failures during the past 30-plus years has proven that more than 80% of failures aren’t related to equipment age or use. The implication of the finding is that less than 20% of our proactive maintenance tasks should be driven by time, equipment age or usage. The majority (more than 80%) should be predictive and detective forms of proactive maintenance. Predictive maintenance is the use of technology or some form of condition monitoring to predict equipment failure. Detective maintenance refers to work that determines whether a failure has already occurred, and applies well to hidden failures that aren’t (at least initially) evident when they occur.

With this new understanding of failures, I migrated my department from operating in reactive mode to operating in proactive mode. The key difference is that our programs now focused on monitoring asset health and letting that determine the maintenance work to be performed proactively.

The research further showed that, once we truly understand an asset’s failure modes "How something fails" (or causes), our program will look more like best-in-class. Here’s an example of a maintenance program that transformed itself from reactive, time-based PMs to a proactive maintenance program.

Before:

1. Clean the pump strainer once a month.
2. Take oil samples from the reservoir, which tells whether to replace or merely filter the oil.
3. Inspect pressure gauges to ensure the pump is developing sufficient head.

After:

1. Watch for early signs of specific failure modes (reservoir temperature or excessive pressure fluctuations).
2. Use electronic predictive checks to watch for early signs of specific failure modes (pressure and flow).
3. Use predictive technologies to catch early signs of specific failure modes (e.g. oil sampling for a specific particle types).

There’s a significant difference between these two maintenance programs. The new program produces far better asset reliability.

I’ve seen equipment failures reduced by 30%, 50% and more. The business impact of a well-defined proactive maintenance program is huge. You’ll increase equipment reliability, reduce capital replacement cost, achieve higher equipment availability and reduce maintenance costs. The soft benefits are a motivated workforce, a less-stressed management team, more time at home, and so on.

While the numbers will get management to support a project to prove the benefits on just one asset, once they see the size of the opportunity and the soft benefits, the next question will be, “What is your plan to roll this out on the rest of our critical assets?” Allow management to work with you to develop the plan. They’ll feel some ownership of the process.

After running a compliant PM program for years, I found I couldn’t rely on time-based maintenance alone. Research and experience in applying that research has proven that there’s a better way to run the business of maintenance. The properly balanced use of predictive, detective and time-based maintenance forms a successful proactive maintenance program.

With such a huge potential to improve business competitiveness, maintenance managers have a great vehicle for generating interest and support among senior management, all of whom are looking for rapid return.

Just my thoughts!