Crystal ball gazing14 December 2013
Techniques such as vibration monitoring, acoustic emissions sensing and thermography are more accessible than ever, but what should you use where?
It's curious that in today's high-tech age, if equipment promises a lot and costs a little, once buyers have overcome their initial scepticism, they're likely to give it a go – believing it must also be easy to use. So it is with thermal cameras, and the results have been twofold. On the one hand, inexperienced and/or poorly trained technicians have made mistakes, potentially devaluing not only this technique, but also the whole plant condition monitoring industry. But on the other hand, it has led some buyers to believe they know more than they do – and to specify specialist diagnostic equipment costing a small fortune when a different approach might do the job better and far cheaper.
Talk to practitioners and they'll tell you that condition monitoring is not just about thermography and/or vibration monitoring, fixed or mobile. Other techniques invaluable for periodic data collection and analysis include acoustic emissions monitoring, shock wave analysis, ultrasonics, oil analysis and, for electric motors (and their driven equipment), automatic signal analysers. But it's horses for courses, and sometimes several or indeed none of these will provide the crystal ball you need.
In fact, the real deal is first, knowing the likely failure modes of your plant – and hence which technique(s), if any, will yield best results – but second, absolutely understanding the value of the equipment to be monitored. Not the purchase price, but the value to the operation and the risk it presents, in terms of the consequence of failure, measured as downtime and total cost, including to business reputation.
For David Manning Ohren, condition monitoring business manager at maintenance and repair specialist Eriks, that is certainly the key to getting the level and type of intervention right. It's also how engineers can ensure that condition monitoring does what it was always intended to do – fill in the gaps where 'conventional' planned and preventive maintenance inevitably fall short, but without costing the earth.
Early warning of machine problems
"If you understand how your equipment fails, then you can monitor certain parameters more closely than others," he explains. "Then you can get, say, a low-level indication of impending failure before bringing in higher-end diagnostics, where necessary."
And he points to modern, low-cost dual-output vibration sensors, which can provide early warning of machine problems verified by known secondary indicators, such as temperature, flow or pressure changes. "Simple equipment like this may be enough to detect changes and generate alarms through the existing plant PLCs. A lot of people are spending way too much on high-end stuff as first-line and they just don't need to."
Meanwhile, there's another message for those obsessed with lower-level condition monitoring, who could be accused of being penny wise but pound foolish. As Manning Ohren puts it: "What's the point of spending 10 minutes checking vibration on a standard motor gearbox every month when it's non-critical and only costs £400 to replace? Why not just install a plug-and-socket system so that, if it fails, you can quickly swap it out?"
Again, getting the level right is about understanding your plant and taking pragmatic steps to manage failure – in this case, ensuring ready access to standard stock and an air gun. That said, where regular condition monitoring is deemed necessary, management of the process need not require anything more glamorous than the existing CMMS (computerised maintenance management system). It can also be linked into maintenance technicians' machine lubrication routines, or operator-driven checks. "Visual checks help and there are plenty of relatively cheap and simple meters these days that give you a lot of information," says Manning Ohren.
Much more approachable
So far, so good, and condition monitoring should now be looking much more approachable for maintenance technicians. Indeed, the best implementations are those that specifically do not treat condition monitoring as a stand-alone exercise, but integrate it as part of their preventive maintenance strategies. However, be warned: just as with any branch of engineering, training remains key. And also, sometimes an expert or a second opinion is prudent.
Consider two extremes. First, if your condition monitoring was instrumental in reducing the plant insurance premium, you're going to catch a serious cold if there's a fire and the thermal imaging record shows features that should have been detected. And second, if technicians believe they see inexplicable hot spots, for example, in electrical systems, you may spend a lot of money investigating line harmonics only to find that it's a warm heat sink.
Which tools to use where
Vibration analysis is good for early detection of problems in rotating machinery ranging from gearboxes to fans, motors, compressors and pumps. It identifies component faults, such as defective bearings and misalignments, and enables maintenance action before an in-service failure.
However, David Manning Ohren of Eriks suggests that good practitioners are likely to use thermography first – certainly on hot, mucky plants, where it is likely to identify the obvious failure modes – but then back up findings with vibration analysis. He also argues that acoustic emissions tools work well, although care needs to be taken where inverters introduce high-frequency noise.
Thermography, using infrared cameras such as the latest Exx series from Flir, is strong when it comes to identifying faulty and overheating components in mechanical equipment. But it is also useful as a non-contact method for electrical panel safety inspections, highlighting fault conditions such as loose bus-bar connections and overloaded components. Additionally, the technique is effective in energy loss surveys and for ensuring steam traps are operating efficiently.
Acoustic emissions tools are useful for a wide range of plant condition monitoring – not just for slow moving or intermittent motion machines, as originally promulgated by Holroyd Instruments, now part of Parker Hannifin. Related systems are also good for compressed air and vacuum losses, as well as steam leak detection. "I'm a great fan of acoustic emissions as a front-end tool where the bearing or gearbox you're monitoring is the only source of noise, and on slow-moving plant," comments Manning Ohren. "But, again, I'd back it up with vibration diagnostics, certainly where expensive plant and remedial work is involved."
Oil analysis is effective in inferring problems in a wide range of machines, with tests ranging from viscosity to dilution, oil condition, appearance, ferrous debris, water, ISO cleanliness, wear metals and contamination metals.
Wayne Pheasey, associate director at specialist engineering firm PEME, agrees that none of these need be mutually exclusive. "We develop an integrated maintenance system tailored to the characteristics of a site... This approach is flexible and delivers production efficiency and cost benefits, as well as ensuring that any maintenance-related risks to health and safety fit the ALARP [as low as is reasonably practical] criteria."
PEME also offers its web-based PEME Condition Register (PCR) software tool, developed to ensure that information about developing failures and required remedial actions is communicated. Plant engineers get remote access with simple traffic light graphics indicating current severity. You can also navigate around your plant model to view the live condition status of connected assets.
Pheasey gives examples of recent successes: "Routine vibration analysis on an extruder indicated a fault on a main dc drive motor. The degradation was monitored and this allowed its replacement to be planned, avoiding an uncontrolled in-service failure. Then, at another site, oil analysis indicated that repeat gearbox failures, which showed no obvious failure trends, were due to refurbished units being supplied with the wrong grade of oil. And, on another plant, where mechanical inspections had failed to identify the cause of its recurring pump failures, our vibration analysis indicated baseplate flexing. Remedials were completed, leading to no further issues."
This material is protected by MA Business copyright
See Terms and Conditions.
One-off usage is permitted but bulk copying is not.
For multiple copies
contact the sales team.