The Journey: Make the successful transition from reactive maintenance
Raise your hand if you would like to go from reactive to proactive maintenance. Going from reactive to proactive is like yearning for the shiny red bike in the toy store window before Christmas. All the kids going by want one, but only a few of those kids’ parents will spring for one. Proactive maintenance is similar — all the smart maintenance professionals want it, but few companies (management) will actually step up to the plate and pay. After all, the usual way to achieve proactive maintenance is an arduous multi-year journey that needs money (sometimes a good deal) and top management support.
There has got to be a better way. If you believe what you read, all it takes is a good solid preventive maintenance (PM) and predictive maintenance (PdM) system. There’s a logical problem with merely installing a system, however, unless you get extra and specific resources for the duration of the effort. Let’s look at this point. PM will get you to proactive maintenance, but without extra resources to support it. Especially in the beginning, you’ll need a very thick skin, a great deal of patience and a decades-long attention span.
To see why this is so, let’s take a closer look at reactive maintenance. What does it mean to run a reactive maintenance effort? If we look at the condition of equipment in a typical reactive organization, we’d find that most have suffered deterioration. In fact, the majority of equipment is in various states of impending catastrophic failure. There’s also no effective effort to identify deterioration, which will lead to failure and prioritize and deal with it before the machine breaks down.
If that’s the state of a reactive shop, why not use a PM system to turn the ship around? What’s the logical problem mentioned earlier? Let’s assume there’s a reasonably level of resources for maintenance. Your people are already working on breakdowns most of the time. Who is available to do the additional PM workload? Well, usually the PM is done during periods of downtime when there’s nothing else to do. This is called "bootstrapping" the PM system.
Bootstrapping might have worked in the days when your personnel weren’t already stretched so thin, but it actually never worked that well because of the logic of PM. Half the effectiveness of PM comes from basic maintenance tasks, such as cleaning, lubricating and tightening bolts. These lengthen the life of equipment already in good shape (no critical deterioration to cause breakdowns). The other key to PM effectiveness in reducing breakdowns is that when inspectors find deterioration-it’s fixed before it turns into a breakdown (called corrective maintenance).
When you decide to add a PM activity, visualize that your shop is stretched thin from cutbacks. On top of this, add additional corrective activity. What do you think will happen? Will you have the discipline to do corrective maintenance work on the equipment that might fail instead of working on equipment that has already failed? Remember there are no extra resources.
This is the contradiction of trying to install PM systems without adding more resources. Since PM has demonstrated it takes a year or more to start to take effect (when done fully and correctly), many more installations are abandoned during their first year than succeed. This is the reason why most PM attempts to move from reactive to proactive maintenance fail (usually not with a bang but with a whimper).
What else can be done? Another strategy is to spend some time (say a year) doing two activities at once. The first activity is getting better at reactive maintenance. Have the crew figure out what should be done to reduce the time it takes to react and repair a breakdown. Use people in teams and create a structured technique to capture ideas and propose experiments. More on the specific technique later, but consider root-cause analysis (RCA) to reduce the time it takes to be reactive.
The second activity is to look at all your repetitive failures plus any repetitive requests for maintenance help. Using established RCA techniques, try to fix them forever. Not all repetitive failures will be solvable with your teams and within your limited budget, but if you pick well-these will free up some time to start more RCA projects.
This dual strategy requires some bootstrapping of labour for the first few months and some limited additional materials. As a year goes on, maybe 20 to 30 percent of the repetitive failures can be resolved through RCA. As these problems get fixed forever, labour is freed up. Use this time to start a PM program.
Start the program small and be sure you install PM in a rigorous manner. Begin with the most critical machines. Go over any machine destined for PM and fix everything your inspection turns up before letting it become an asset in the PM system. One by one, add assets to the PM list. Now this is the way to bootstrap a PM system.
What does RCA involve? Simply speaking, RCA is a structured process to solve problems. It can be applied to everything from accidents, plant fires and explosions to small, localized situations. Even small interruptions to production can be analyzed with RCA. Basic RCA techniques can be learned in two days.
It’s best to conduct RCA with small teams of people (three to six) with different backgrounds. At a minimum, someone from operations, an expert in the process and someone from maintenance should be present. Also smart people from other parts of the company are useful to ask other questions.
The first step is to figure out what the problem is. This is tougher than it sounds. Some problems are concealed; other problems are the result of misunderstanding, misdirection (both intentional and unintentional) or mishandling.
Once the problem is defined, the consequences are spelled out and, most important, ways to measure results are chosen. The second big step is to gather data and brainstorm causes. You may develop a handful of potential causes. Some are primary and others are secondary. Causes themselves have causes. These causes form trees (called "cause trees"). These cause trees should be checked both practically and logically to determine actual causation. The completed set of cause trees for an event should model the event closely.
The last step is looking for an intervention to stop the chain of events that caused the failure or accident. You’re trying to interrupt the cause tree. Once you have a likely intervention, you can test it to see if it indeed interrupts the problem in the "real" world. Then look closely to see if there are any unintended consequences of the intervention.
As part of the last step, you have to make the change stick by rewriting standard operating procedures (SOPs); changing parts procured; communicating the changes to everyone; adding the change to drawings, computerized maintenance management software (CMMS) tasks; or wherever will make it happen and keep it happening, etc. That’s the little secret of going from reactive to proactive by your bootstraps!
Joel Levitt is a management consultant and teacher with U.S.-based Springfield Resources. You can reach him by email: firstname.lastname@example.org. He’s also writing a book funded by RCA RT, a leading provider of RCA training and software. RCA RT is located in Australia. For more information, visit www.rcart.com.au.