MRO Magazine

Feature

Using Predictive Maintenance Effectively

This article is part of an ongoing series. The introduction appeared in Machinery & Equipment MRO's February 2005 issue and the series has run in every issue since. Previous instalments are archiv...


This article is part of an ongoing series. The introduction appeared in Machinery & Equipment MRO’s February 2005 issue and the series has run in every issue since. Previous instalments are archived online at www.mromagazine.com. In this issue, we pick up where we left off in our June 2007 issue, as maintenance manager Bob Edwards and his team at the Plentya Paper Company plan on working out how to cost-justify the next steps of advanced predictive maintenance. We open with a meeting in Bob’s office.

The meeting to introduce and explain predictive maintenance had gone very well, I thought, and the guys were definitely buying in — even those who had been skeptical so far. John, the vibration analyst, had stayed behind so we could discuss the frequency of our inspections, so I asked Carol (the maintenance planner), and Ted and Ivan (both millwrights), to meet with us in my office.

“Okay, whatever frequency we come up with, I’d like Ted and Ivan to go around with John and start to understand even more about what he’s doing. So Carol — even if I forget — you should schedule them for that,” I said. Then I turned to John. “Well John, you’ve seen our equipment. How often do you think you should come in and measure?”

“Based on your run time, I think a good starting point would be every four weeks, the same frequency as your shutdowns. Like the last time, if I come in a week before the shutdown, you’ll have enough time to react, get your parts together and set your schedule. The rollers on the paper machine should take about half a day to measure, but I would need to have a day working on the gearboxes as they have so many different frequencies.”

“Okay, just arrange it with Carol and she’ll make sure the guys are available.”

“John, I’ve just thought of something,” offered Ivan. “We changed one of the gearboxes last year but the other hasn’t been changed at all. I think you’re going to get different readings, so how do you tell which is right?”

“There’s a few ways we could approach this,” John replied. “We could assume that the newer gearbox would have the more correct readings, but that would be foolish as we can’t tell what caused the gearbox to fail last year, or whether the root cause has been removed. In addition, incorrect installation can affect the performance — just like the main drive motor on the paper machine we’d discussed previously.

“The second approach is to take the standards around vibration and apply them. If we get readings way above the standard, we’ll take a deeper look. The last item is to monitor the rate of increase in vibration, as any sudden jump means we need to take action.”

“We can also take readings of the amperage for the roller motors and the gearbox drive motors in the respective control rooms,” suggested Carol.” Could we use those as indicators of change in condition, something like primary and secondary techniques?”

“Great idea!,” said John. “If you track those on a daily basis, call me if you see an increase of 10%. Be sure to take the readings under the same operating conditions or you may come to the wrong conclusion. Before I go, I want to give Ivan one of my ‘loaner’ vibration meters so he can set up the route for the PMs in stock prep. It takes overall readings so it won’t tell you what’s wrong, but it will tell you something has changed. You can use the table of standard readings that indicate good, fair and poor readings to help you.”

“Thanks,” Ivan replied. “I’ll start the set-up as soon as you give me the meter. I hope it’s not too complicated to use, as I’m not that good with these fancy new electronic gadgets. To be honest, that was one of the reasons I didn’t support this right away — I don’t understand this new technology very well. I know I’m good at fixing things, but this computerized stuff frightens me a little,” he admitted.

For the rest of the day, Ivan was true to his word and he worked on the route for the pump monitoring in stock prep. I decided I’d better bring Joe Davis, the plant manager, up to speed — especially with the costs we anticipated.

“Carol, I’m going to work on a simple business case for Joe. Could you bring up the costs for the suction roll failure last year and the gearbox earlier this year?”

“It’ll take me a little while, as I’ll have to go through the production reports and the purchase orders to get the values. This is another case where life would be simple if we used the CMMS fully. The only thing I can say for certain is that we changed those two paper-carrying felts at a cost of $40,000 each. If we add the overtime and the cost of the repair of the roller journal to the cost of the felts, then that is the added cost to the bearing change. That should be a good number for you.”

“But why wouldn’t we include the cost of the bearing replacement?”

“Well, we would have had to change the bearing anyway, Bob. You need to remember that predictive maintenance doesn’t prevent failures. It detects them at a very early stage and that allows for planning — with no overtime, no secondary damage, no journal repairs and no new felts. So those are the numbers we should use for justification.”

“You’re right again! If you bring those numbers up, along with the cost of the paper we didn’t produce due to the reduced speed we ran because of the gearbox failure, then I’ll use those.”

John was going to cost us $1,000 a day or $1,500 a month, which translated into $18,000 a year, so I knew I was going to be able to prove the predictive program’s value very easily. When Carol arrived with total costs of $105,000 for the suction roll failure and $84,000 for the gearbox failure, I gave Joe a call. He said he could see me right away.

“Just thought I’d bring you up to speed with what we’re doing with the predictive maintenance, Joe.”

“I hope you’ve got some numbers for me this time, Bob. As I said, if I’m going to support this program, I need to take details to my boss.”

“I have some simple figures from last year that will help me explain,” I said as I showed him the costs of the failures. “So it seems obvious that if John had warned us about one of the impending failures, we would have saved a lot of money, and all he is going to cost is $18,000 a year, so far.”

“What do you mean, ‘so far’?”

I explained that we thought that there would be many more pieces of equipment that we could apply this logic to, but we just hadn’t done the analysis yet, and that was a separate program we needed to start.

“Just remember that if you’re going to spend money, you need to have a return, but based on what you’ve shown so far, carry on,” said Joe.

The next morning, Ivan was waiting for me when I got to the stock prep area. “I’ve mapped out the route,” he said. “I took readings with John’s meter and marked them down on the sheet. I’ll get Carol to put them into the computer. But before she changes the PMs, I’d be more confident if someone else walked the route, just in case I missed something. Rob said he’d like to do it.”

“Sounds like a good idea. Just let me know what happens.”

I saw a worried look on the faces of Ivan and Rob when they came to my office four hours later.

“I told you I was no good with this new technology,” blurted Ivan. “Rob did the route and came up with completely different readings!”

“The route itself was fine, but I couldn’t seem to get the same readings as Ivan.”

“Why don’t you both walk the route and take the readings again,” I suggested.

“Okay — we’ll be back in a couple of hours.”

And they were.

“We’ve solved the mystery, Bob. Rob was taking the readings on the top and I was taking them on the side of the housings.”

“But which one is right?”

“That’s the problem, but based on what John has told us so far, it’s important that we take the same reading each time. Remember his rate of change theory?”

“Hmm, let me run that one by Carol,” I answered.

Carol confirmed their idea. “It’s the tr
end we’re looking for, so consistency is essential.”

“Great,” said Rob, “and how do you suggest we get that?”

“I’ve got an idea,” suggested Ivan. “How about if we take a half-inch washer, place it on the housing and spray black paint through the hole. Then, whoever takes the reading places the meter on that point.”

“Brilliant! You may not be confident with this new technology but you’ve shown there’s no substitute for good old common sense! Okay, you and Rob have another job to do,” I said.

When I picked up my messages the next day, there was one from Ivan and one from Dave in plant services, who had been working on the seal problem on the wastewater pump.

Ivan was worried again, for when he and Rob had taken the readings for the third time at the marked locations, they had found that some of the readings didn’t even come close to the previous ones.

Dave was really happy, though. “Hi Bob, I met with the pump supplier and we think we’ve got a solution to the seal problem. If you come to the shop I’ll explain.”

At least the day was starting out 50% positive!

Cliff Williams is the maintenance manager at Wrigley Canada in Toronto, Ont., and a consultant with TMS Total Maintenance Solutions of Markham, Ont. He can be reached at williamscliff@rogers.com.

Key Points

* People’s reluctance to accept something new is often driven by a fear of the unknown.

* Even in the high-tech era, common sense prevails.

* More than one technology can be used to determine asset condition.