Predictive maintenance is the science of knowing when to apply maintenance techniques to your equipment, based not on guesswork or set time intervals but careful measurements made with monitoring tools.
In general, there are three kinds of maintenance:
- Reactive maintenance: Repair it when it breaks. (Example: a kid’s bike.)
- Preventive maintenance: Replace before we think it will break. (Example: automotive fluid changes)
- Predictive maintenance: Monitor, measure, and base decisions on actual performance and expected fail probability.
The problem with reactive maintenance is that typically things break when you don’t want them to, leading to a lot of downtime and production loss.
The problem with preventive maintenance is that you are wasting parts and doing unneeded replacements, while still running the risk of things breaking when they wear more than expected.
Example of Predictive Maintenance
Let’s consider a starter engine for a car. You could decide to replace this every 100,000 miles to avoid breakdown due to wear. But this isn’t necessarily the best way to handle it, because it doesn’t take usage into account. If the owner starts the car 100,000 times to drive one mile each time, the wear on the starter will be far greater than if that person drives 500 miles every time he starts the car. Things like heat, climate and other factors also influence wear. The policy of replacing a starter every 100,000 miles wouldn’t work well for many situations.
Predictive maintenance is far more scientific, because it is about measuring and comparing. Instead of replacing the starter at the 100,000 mile mark, you would instead perform a diagnostic test on it to see how it was working. If it hadn’t been used much, then there is no reason to replace it. This saves time and money.
When applying predictive maintenance measures on manufacturing or processing equipment, you will measure things like energy consumption, vibration, temperature, and heat to see how moving parts like bearings and gear boxes operate. You may also measure chain stretch to predict when a chain will be at the end of its service life.
There is a challenge inherent with predictive maintenance, and that is the switchover. Many companies may feel like they have to implement it the same way they would a major project. They will buy software, measure baselines, enter maintenance points, create work sheets, collect data, etc. But many maintenance departments are understaffed and are typically running from crisis to crisis, so they don’t have the resources to do this. Management may also not be willing to invest in this, because the payback is half a year down the line and hard to quantify.
Predictive maintenance can be phased in gradually, rather than all at once. It represents a new way of looking at things, and sometimes people can be slow to accept change.
Interflon is also developing a fourth regime, called Pragmatic Maintenance. Rather than trying to improve the maintenance of the whole plant, just start with the mission critical components and leave the rest as is.
You are the maintenance manager at a facility that cuts marble countertops. You have one cutting machine and 50 hand-held buffers to finish the countertops. You can have a predictive maintenance regime on the parts of the cutter that stop production if they fail, but still practice reactive maintenance on the buffers. These cost just a few hundred bucks, and by buying a few spare ones, you don’t need to spend a lot of time administering all kinds of things regarding them.
Our message to you about predictive maintenance is this: you can grow towards a predictive maintenance regime without a Big Bang that puts half your team into a burnout.
Have questions about how to implement predictive maintenance techniques at your business? Give us a call or drop us an email. We are only too happy to help you save money!