It is slightly worrying that, if asked to name plant condition monitoring (CM) techniques, a high proportion of plant managers, engineers and technicians would probably only think of vibration monitoring, thermography (infrared cameras) and possibly oil analysis. It is equally concerning that many in management seem to believe that it is enough simply to purchase basic CM tools of one type or another, and maybe (but maybe not) some training, to get predictive maintenance working.
Talk to anyone in the contracting business specialising in MRO (maintenance, repair and overhaul) and they'll nod wryly. They'll also add that the likelihood of anyone in management giving serious thought to any kind of CM strategy – how, where, with whom and, most important, why to implement equipment, processes and systems – is relatively slim. That's despite what should be an obvious requirement for at least a quick and dirty cost-benefit analysis, taking into account: plant criticality and real downtime costs; equipment choices and skills needs; spares stocking or contracted lead time requirements; and any potential for plant uptime and product quality improvement.
Why doesn't CM get the attention it deserves? Because in far too many plants and factories it is still the case that maintenance is regarded as a cost – a necessary evil – not a potential profit centre. So the function, the people, the equipment and any related projects simply aren't taken very seriously. And that's despite the advent of the ISO 55000:2014 series of asset management standards, published late last year (Plant Engineer, March/April 2015, page 16).
David Manning-Ohren, Eriks' business unit manager for CM, says he sees it all the time. "All too often condition monitoring is implemented in a haphazard, rather than strategic, fashion – frequently targeting the wrong plant and using the wrong equipment. Then management are surprised when it has little impact on uptime or productivity." And the result: the historic prejudice around maintenance engineering is cemented in: it's a vicious circle.
"I was with a potential customer recently who said he just needed our technicians to carry out inspection and maintenance duties for a couple of days a week," he recalls. "But he hadn't considered which plant was critical, the types of machinery, or how the failure modes might present themselves. Yet all of that has implications for where to focus, the kind of tools you need, the level of diagnostics and whether to go for fixed equipment or periodic inspection with portable tools. Also, some plant may take five years to fail, while other equipment is quite dynamic, meaning its condition could change in a couple of weeks. You need to know all that to make any CM programme work well."
For Manning-Ohren, there are three key steps to consider well before determining budgets, selecting equipment or thinking about technicians' time. "First, you need to know where your plant is today, in terms of reliability, efficiency and uptime. Second, you must understand the criticality of plant and its repair/replacement lead times and downtime implications and costs – not the item price. Then third, you need to get a handle on the failure modes of the plant that matters most. Knowing those three you can pull together a strategy for the CM methodologies and frequencies – and hence the kit or services to buy, the training required and so on."
Putting meat on those bones, if a machine's expected failure mode is imbalance, then you might need measurements every day to detect small changes and eventually the ramp ahead of failure. In that case, a set of fixed vibration sensors – picking up the acceleration envelope and overall velocity, and linked to a PC in the engineering office – is a sensible investment, probably augmented by weekly or monthly inspections. But even then, the devil is in the detail: if the machine is critical and its failure expensive, you might well argue for better analytical information that comes from higher-level diagnostic packages capable of identifying degradation with greater precision.
As Manning-Ohren says: "If you've got, say, a 750kW gearbox, you might want to see spectral and waveform data to assess which gear or pinion is failing. You might also use an endoscope. But if it's a 0.7kW machine, then you just want to know it's failing so you can replace it in time." Equally, if your machine is in a very noisy environment, then configuring a system to reliably detect spectral changes would be expensive, and thermal imaging wouldn't necessarily help either. So you might want to consider acoustic emissions sensing and/or oil analysis – even a simple magnetic oil contamination sensor as an initial alarm.
Strengths and weaknesses
Clearly, one size does not fit all, so what are the strengths and weaknesses of the major techniques?
Manning-Ohren suggests that inexpensive low-end vibration analysis is adequate for monitoring the condition of motors and gearboxes on, for example, conveyors in food and beverage plants. Hand held diagnostic kit, involving low-end vibration pens costing less than £2,000, is ideal for trained fitters and electricians on their rounds, and the results can be stored in a database to reveal trends.
That said, such devices won't pick up motor cooling fan blockages, for example. so there's also a strong case for periodic sweeps with budget thermal imaging cameras to look for hot spots. Indeed, as a general rule, vibration monitoring and thermal imaging can work well together across a range of rotating plant. As thermographer John Reynolds, of Flir Systems, puts it: "Thermal imaging provides a good immediate snapshot, while vibration analysis, when it's done well, is good at showing trends. It's not that you can't do that with thermography. You can, but vibration doesn't depend, for example, on the time of year."
What about larger-scale equipment and/or plant in aggressive environments involving exposure to heat or water – such as pumps and fans? Depending on scale, this is where better diagnostics come in, the objective being to characterise minor problems that can be resolved in-situ, but also to predict when maintenance needs to get involved, so that major interventions can be planned to minimise disruption. Again, remember that often it's not the value of the machinery that matters: it's the 10 hours of downtime caused by the breakdown of a £700 gearbox in its drivetrain. Best advice is to invest more in the CM kit (portable or probably fixed, if plant is critical and history shows a potential for rapid change) and/or to consider using a third party specialist CM resource.
Moving on to control panels, the best approach is almost invariably to use high-end thermal cameras with a minimum resolution of 320 by 240 pixels. These are capable of quantitative thermographic survey work. Prices have come down, but operators do need formal training. Look for a technician with at least a Level II thermographer qualification to avoid expensive errors around, for example, emissivity, pointing at the right stuff and even straightforward misinterpretation.
Beyond generalising though, there are times when you just can't beat years of CM experience. Manning-Ohren recalls a recent project involving a food factory filling machine. The challenge, he explains, was to eliminate frequent and expensive failures of its slewing ring, which root cause analysis revealed was due to intensive CIP (cleaning in place) activities washing out the bearings. "One of the problems was they didn't know how much fresh oil was reaching the bearings through the machine's lubrication system. So we recommended installing Holroyd [now owned by Parker Hannifin] slow-speed acoustic sensors linked to a PC, which would instantly show dry bearings." Three were installed to cover the axes and the installed and commissioned cost was £3,000. Compare that against £6,000 for a replacement slewing ring. No brainer.
Flying doctor
Factory and building inspections from the air are now being offered by Thermographic Consultancy, which is a CAA (Civil Aviation Authority) licensed condition monitoring contractor and uses Flir thermal imaging cameras mounted on drones.
Stuart Holland, a Level III thermographer with the company, says the approach allows for quick and fluid thermographic surveys, and provides the ideal capture angle for roofing inspections. "In the past we have had to rely on masts and poles for our cameras to combat reflectivity issues. Now we have the capability to view a large area with ease since we can operate at up to 120 meters altitude and fly within 500 meters of the 'pilot' in any direction."
The firm uses drones in two weight categories, 0—7kg and 7—20kg, determined by the camera type. One is Flir's standard commercial grade T600 camera (or similar), which can also be used on the ground for other condition monitoring purposes, but means a heavy cargo. The other is a camera core, such as the TAU2, which is a professional OEM product. The downsides: it can't be used anywhere else and it requires deeper knowledge of thermal imaging for configuration and set-up. The upside: this camera is much smaller and lighter.
Top ten rules
1 Never use condition monitoring on its own, or as a trending tool.
2 Spend your maintenance money on the most valuable assets.
3 Tap into the people using your machines every day. They are the ones who will notice the rattles, smells and drips that are out of the ordinary.
4 The ideal condition monitoring frequency depends on many variables – from the maintenance regime to the quality and age of the equipment and its criticality.
5 Keep a vibration database that includes trends, spectra and time waveforms.
6 Understand the plant and its operating conditions: changing temperatures and operating speeds matter to many sensing systems.
7 Condition monitoring certification that requires updating is preferable to qualifications that don't.
8 Condition monitoring tools are not enough: technicians need training and the workforce needs to be switched on.
9 Judge a condition monitoring provider on more than the day-rate. Check the scale of support on offer.
10 Remember that good maintenance is designed to delay plant failure – which saves money.