Vibration monitoring is finding wider use through the growing range of smart, affordable mobile devices, but accurate analysis remains a key factor for success.
See or hear a piece of rotating equipment vibrate abnormally, and the solution is obvious: shut it down, find the cause and fix it. In the pre-electronic era, this is all there was to vibration detection—just what human eyes and ears could detect.
When electronic sensors arrived, plant personnel could routinely monitor equipment vibrations and get a valuable perspective of potential failure trends before oddities became evident on the equipment. Now, wireless is expanding in-plant and remote-site opportunities for vibration monitoring, but brings a level of complexity its condition-monitoring cousins like infrared and ultrasound do not.
“The acquisition of vibration data is pretty much done the same as we’ve always done it,” says Trent Phillips, Condition Monitoring Manager at Ludeca, Inc., in Doral, FL, “but the data transmission aspect is more tricky. Transmission of process-related data over a distance usually requires short bursts of simple numbers and can be done very fast. With vibration monitoring, if you limit your data transmission to just overall values, it’s similar to temperature data: simple numbers. To do real vibration monitoring and analysis, you have to transmit the spectrum and the waveform. In that respect, we’re talking thousands of numbers that must be transmitted for every sample taken. Those numbers have to be transmitted error-free, so from that perspective, it gets more complicated.”
And as more equipment—both more types and in more locations—is monitored remotely for vibration, more benchmark data is needed. So is the specific background of each piece of equipment, according to Phillips. “In Europe, more than in the U.S., they rely on ISO values to determine what’s an alarm and what’s not,” he says. And while standards can be helpful, Phillips says he has “found that you have to let the machine tell you based on its own rate of change. That’s where you’ll really learn about the severity of a situation.” Machines are like people, he adds, noting that “one may be able to vibrate at a high level, which is normal for that machine. The machine next to it, however, may fail if the vibration reaches only 20% of that level. So you have to look at all machines differently, and you have to look for the rate of increase and how it is trending. That’s how a good analyst will determine severity of a problem.”
More ‘feet on the street’
Phillips and others note that while wireless opens the doors to more vibration monitoring, mobile units, in turn, expand the ranks of personnel available to obtain the data. “Everyone wants to become more mobile,” says Scott Brady, the Applications, Training and Documentation Manager for SKF USA’s Condition Monitoring Center in San Diego, CA. “This is the same trend we had moving from desktops to laptops to make the office worker more mobile. Now we’re looking at how to make the plant worker mobile, so we have people who didn’t necessarily take vibration readings [before]. We have more ‘feet on the street,’ you might say. It’s no longer just up to the vibration guy.” Operators and others in close proximity to the machines can now gather that data with wireless mobile devices. “We’re making it easier to get that data,” he says, “and as the age-wave hits, as well as for companies who are trying to save money and not hire when some retire, they can still be gathering this information.”
The lower price points of most mobile units remove yet one more barrier to bringing this technology into the plant. “A lot of these techniques have not been adopted in the past because of the high cost of the equipment and the training,” says John Bernet, Mechanical Application Specialist at Fluke Corp., in Everett, WA. “The trend today, which has been coming a long time, is to find a way to make proactive maintenance like vibration testing, thermography and others more attractive to the smaller and mid-sized companies.” The mobile tools Fluke and others offer are not only “less expensive than some that have been on the market for the last 30 years,” says Bernet, “they’re easier to use,” an important factor when placing the tools in less-experienced hands. And because many handheld vibration testers come pre-programmed to analyze vibration spectra, “training can also be short,” says Bernet. “Within a few hours they can be using it, and that information can be linked wirelessly over the Internet. It’s much easier to sign off on.”
Phillips adds that with non-maintenance personnel doing more monitoring, “the analyst will have more time to do machine assessments, which is where the real savings come in. This is when you find out what’s causing things to fail, you drive out the failure modes and you truly become more reliable.”
Keep your analyst happy
Getting there, of course, still hinges on an analyst’s ability to understand and interpret accumulated data. Phillips says there’s no substitute for experience, education and getting certified in the field. “It also takes a certain personality trait,” he says. “They have to be very proficient and detail-focused, and be passionate about what they do because—and this is true if you’re using any of the condition-monitoring technologies—most analysts will tell you they get rejected a lot. Management is looking to understand why, and they’ll usually come back to the vibration analyst and say, ‘Why didn’t you tell us?’ So you have to have your documentation to be able to show that you did tell them.”
But even this process is undergoing simplification. “Our customers are looking more for red/yellow/greens and go/no-go indications,” says Phillips. “They want the equipment to give them that sort of traffic light. I would argue the technology for this is still not quite there, but it’s much closer.”
SKF’s Brady points to the growing use of systems that “will send you an e-mail alert to your phone if you have a moderate outer-race bearing defect or if you have a severe misalignment problem” as examples of the trend toward simplification. “Today, this information comes from the office computer,” he adds, “but I see this moving closer and closer to the person out near the machine. This could be wireless or it could be intelligence embedded in the machines themselves.”
In five or 10 years, predicts Bernet, “We’ll have even more smaller and faster tools, and eventually each machine will have its own sensor. We’ll still need walkaround tools,” he says, “but its going to be more localized. So you’ll have wireless sensors all over a plant, and you’re going to have that information going to smartphones and the cloud, and 80% is going to be done at the machine instead of in an office. Think of your car,” he adds. “It has vibration sensors all over it, so why can’t we put those vibration and temperature sensors on other types of rotating machinery? Because it has been so hard to get to a lot of machinery, we’ve been limiting ourselves to the top 5% or 10% of the machines that are the most production-critical. As we start using smart tools,” he says, “we can encompass 80% of the plant, maybe even 90% or more.”