Mike Monday and Predictive Maintenance
February 15, 2018
Why I am looking at predictive maintenance? Well, our department must justify every penny in the budget. So, when we talk about spending more money, our justifications should be solid. Part of that is doing the math and figuring how much money we lose by not practicing these methods. Another part is showing real improvements on actual machines.
The goal of predictive maintenance is to identify early indicators of change. There are pros and cons to predictive methods, several of which I will expand on. Practicing these methods takes time and training. We need people to know what to look for, but before they can do that they need to know how to look.
Below are some of the tools involved in predictive maintenance.
- Sensory testing (sight, touch, smell, hearing)
- Record keeping
- Infrared thermography analysis
- Ultrasonic analysis
- Vibration analysis
- Oil analysis
While some of these methods may be useful for one machine they might not apply to another. Some of them, like oil analysis, need involvement with an independent lab. There are ways to do field examinations, and that’s why training our people is so important.
Sometimes a mere sniffing of the air can tell you that a machine is burning oil. Sight glasses are great additions to machines because they allow you see both the level of the oil and the its color. If you hear a grinding sound on a machine with ball bearings you know that the ball bearings are likely in bad condition.
Smaller, but no less important, things to consider- where should people take the samples from the machines? How often should a check-up take place? Testing and check-ups require a proactive mindset to work. If people record the information but don’t examine or communicate it, then we’re not practicing predictive maintenance.
Early detection of condition changes reduces downtime. For some machines poor maintenance causes legitimate safety concerns.
Now, if we tried to change the whole plant I can predict what would happen- nothing. What I mean is, nothing would change because we would be changing too much at once, without enough support. Instead, we looked at three machines over the past month, while applying predictive practices.
Press Brake 117 is a clear example of why predictive maintenance can be useful.
I started with the oil in the reservoir. I used a dipstick to draw out some of the oil and the color was a deep brown-black color. I emptied out the tank and put in a new batch of oil.
When I looked in the reservoir two days later the oil was as dark again. I thought that the oil I used was bad, and changed the oil for the second time.
It occurred to me then that I had not looked at the filter. I’d focused so much on the oil as the culprit that I forgot about the filter. When I took the filter out the amount of dirt caked onto it was surprising. This old filter had not been changed in so long that it led to a build-up of particles and dirt in the pipe. By the time I changed it the oil had not been passing through the filter, but around it. The oil was not being cleaned and because of the build-up, it was becoming contaminated.
Following these two oil changes and filter replacement, I monitored both the machine and the reservoir for four days. Checking for excessive heating, watching for any shaking, using a dipstick in the reservoir, and taking oil samples for oil analysis. After four days the oil darkened again and I looked at the replacement filter and it had reached dirt holding capacity. I replaced the filter for a second time, and for good measure changed the oil.
Even though this was new oil, the filter still caught a significant amount of dirt and particles. The new oil was not clean. So, after changing the oil three times and replacing the filter twice, I discovered that the new oil I put into the machine was part of the problem.
If I changed the oil and filter once and not followed up I would not have discovered this. So, I took an old filter cart that was sitting in the lube room and unused for years, and ran some of the new oil through it then put that oil into the reservoir. The machine is running better now, and I will continue my check-ups to ensure it stays that way. Currently there is no major discoloration of the oil, but I’m taking samples to send to a lab.
In the next post I will dig deeper into predictive maintenance methods that we are testing.