Run To Failure: That’s a Maintenance Strategy?

“What’s wrong with the car? What do you think’s wrong with the car? Damn thing’s broke.”  — Richard Pryor

When I was a poor college student, I wasn’t too poor to have an old beater of a car. It was paid for, but I could barely afford gas, much less proper maintenance. So, I tended to run-to-failure. I didn’t change tires unless a vehicle inspection said I had to, and then I bought used tires with just enough tread to pass inspection. I ran on them until I had a blowout because I wanted to get as much mileage out of them as possible. (I turns out you can run on a tire for quite a while, even when the steel belt is showing.) That meant that I had to keep a spare in the trunk and a firm grip on the steering wheel at all times.

The first time I had a blowout, it was on the front passenger side while I was driving on a relatively untraveled two-lane state highway with wide shoulders. It was easy enough to pull over, jack the car up and change the tire. I was back on the road in ten minutes and wasn’t in a hurry in any case.

The second time I had a blowout was on a long gravel driveway. It was in the back, but I don’t remember on what side. I didn’t even have to pull over to change the tire.

My third blowout, however, was on the driver’s side while I was driving on an urban interstate during evening rush hour, as the sun was setting and the shadows were long. Changing that tire was a frightening experience. For me, that was the end of run-to-failure as a deliberate maintenance strategy.

At least for my car tires.

Three Maintenance Strategies

The most commonly recommended maintenance strategy, especially in the context of operational safety, is preventative maintenance. Preventative maintenance (PM), also called planned maintenance or scheduled maintenance, is a strategy where components get replaced at regular intervals. The intervals can be based on the calendar or on some other measure. Changing oil every 5,000 miles, for instance, is a form of preventative maintenance that is not based on the calendar.

This is regardless of whether the component needs to be replaced or not.

A more sophisticated strategy is predictive maintenance, also called conditional maintenance. With predictive maintenance, some test result or other condition calls for maintenance based on the specific state of a component. It requires monitoring the component continuously, or at least much more frequently than the failure rate. So, while it requires a lot more attention, predictive maintenance allows for more of the useful life to be extracted from the component.

The oldest approach to maintenance is sometimes called corrective maintenance or reactive maintenance; colloquially it is known as “run-to-failure”. Run-to-failure (RTF) is just what it sounds like: the implementation of that saying popularized by Bert Lance, “If it ain’t broke, don’t fix it.” After a component fails, react to the failure by correcting it, usually with a repair or replacement.

From a Safety Perspective

Preventative maintenance is often touted as the safest approach to maintenance. This assumes that a failure is unsafe and that the maintenance is frequent enough to avoid failures. It doesn’t work if the interval is too long. “I bathe once a month, whether I need to or not,” is scheduled maintenance, but in the 21^st century, it would be considered inadequate. Even when the interval is just a fraction of the mean time to failure, it is the nature of random distributions that there is some probability, albeit small, that a failure will occur immediately following preventative maintenance.

Predictive maintenance, with its continuous, or frequent, inspection and monitoring, reduces the likelihood of being caught by a premature failure. And it reduces the costs associated with premature repair or replacement. It is especially useful when the cost of repair or replacement is very large when compared to the cost of monitoring. When the cost of monitoring is large in comparison to the cost of repair or replacement, or when there is no way to inspect or monitor, then predictive maintenance is not a viable option.

Sometimes, however, a failure is not unsafe. Sometimes a failure is just inconvenient, and sometimes the inconvenience is simply making the repair or replacement. In these situations, corrective maintenance – RTF – is a perfectly reasonable strategy.

Really.

Light Bulbs and Lubrication

In my home and my offices, I change the lights when they burn out. Not before. This is a classic example of RTF. Is it unsafe to work or walk around in the dark? Sure. But absolute darkness in my home or offices requires that not one, but every light fails simultaneously. That can happen if there is a power failure, but changing the lights on a regular schedule won’t prevent a power failure.

Some lighting benefits from preventative maintenance, however. In big box stores, with extremely high ceilings, they will change the lights on a schedule. The cost of setting up for the light change is so much that all of the lights will be changed at the same time, usually in the off hours. Do lights with remaining life get discarded? Absolutely. But the cost of changing is sufficiently high to make scheduled maintenance the preferred strategy.

As for conditional monitoring of lighting – I don’t even know how that would be done.

Lubrication calls for different maintenance strategies. In a car engine, running the lubricating oil to failure would mean running until the oil light came on, which means that engine is experiencing conditions that can damage it, or in the extreme, running until the engine seized. A seized engine is an expensive repair, so it makes more sense to change the oil before that happens.

For most of us, changing the oil means preventative maintenance – an oil change every 5,000 miles, or whatever the manufacturer recommends. Oil changes are much less costly than engine replacements, and we have some flexibility about when we schedule them, so they don’t have to be terribly inconvenient.

There are some engines, in heavy industrial use or in aircraft, for example, where an oil change is an expensive undertaking. So, instead of preventive maintenance, these situations call for oil condition monitoring systems. These OCMs tell maintenance when the condition of the oil has deteriorated to the point that it needs to be changed. This allows the oil to run its full life and also anticipates premature failure.

There are other applications for lubrication as well. The hinges on my doors are lubricated.  I don’t lubricate them, though, until they squeak. Another case of corrective maintenance.

No One Maintenance Strategy Is Best

Despite the scoffs that corrective maintenance elicits and despite the praise heaped on preventative maintenance, there is no one maintenance strategy that is “best.” It will depend on the nature of the maintenance, and the circumstances of the equipment. RTF can be an effective way to provide a safe working environment in some circumstances, and completely inappropriate in others. Likewise, predictive maintenance can result in large cost savings while assuring that equipment operates safely in some instances, and be wildly impractical in others.

Insisting on a single approach is a mistake. Corrective maintenance, preventative maintenance, and predictive maintenance are all tools that should be available to keep a workplace safe. They aren’t strategies, they’re tactics. And we all need a range of tactics at our disposal.