Condition monitoring is, in essence, monitoring the characteristic performance parameters of a machine, recording them in a time series to reveal trends, and then marking the limits that separate normal from abnormal performance. Trends give an indication of how the machine is performing, and highlight the opportunity to fix a problem in a planned shutdown, rather than a costly unplanned shutdown.
For bearings, a key parameter of condition is vibration. The raw waveform of a vibration can reveal numerous aspects of condition. For example, low-frequency events of one or two times rotational speed are indicative of imbalance on a shaft: a missing gear tooth, for instance.
The emergence of low-cost digital microelectronics, sensors and processing power in the mid-2000s led to dramatic breakthroughs in condition monitoring, and an explosion in the number of embedded sensors. These devices are powerful enough to monitor a vast range of parameters, and today, alongside the well-established techniques of vibration analysis, many more measurements are recorded and assessed. A flood of data is now emerging from rotating equipment and the bearings within.
Paul Deighton, rotating equipment performance specialist (global) at bearing manufacturer SKF, explains: “Now vibration data is collected, but so are operating parameters such as temperature, flow pressures, motor current usage. As opposed to focusing in on one particular area, we now get a bigger picture of how a pump, for example, is operating, as well as its maintenance constraints.”
Deighton continues: “People are changing their approach to condition monitoring. They’re now getting more access to information that makes decision-making more complete. Obviously the more information we have, the better our prognostics are going to be.”
MAKING SENSE OF IT ALL
It is the interpretation of data that is central to maximising the value in terms of condition monitoring. Dieter Charle, business development manager for portable reliability solutions at Emerson Automation Solutions, says: “You can have a lot of features on your measurement device, but the smarter you make your devices, the smarter you need to make your operator as well.”
The need for better interpretation and analysis of data has led to new approaches in condition monitoring, as Ian Pledger, field service engineer at Schaeffler UK, notes: “We can do a lot more with computing than we could before; we’ve got faster chips and we’ve got huge amounts of memory available. Sensors for monitoring bearings now are not only relatively cheap, the analysis and interpretation of this data is more sophisticated as well.”
Effective condition monitoring should enable businesses to switch away from a scheduled maintenance strategy to a predictive maintenance programme in which issues are addressed as they arise.
Turning to how this relates to the specific case of bearings, the ISO standard for the calculation of bearing life defines basic life (L10) as the lifetime which 90% of a large group of identical bearings can be expected to reach. It is a function of the load, capacity and the speed of the bearing among other parameters, but cannot account for the random nature of failures. As Pledger says: “Most failures are random. The bearing life for a particular application is calculated from many factors, including the efficacy of the lubricant and how it might be hampered by temperature, contaminants within it or age. If you look at the [ISO] Adjusted or Expanded Adjusted calculation, depending on how well the lubrication does its job (or not), you could get a factor of 5 or 10 times (or 0.5 or 0.1 times) the basic calculated life.”
Not only is the concept of operational life fraught with variability, but also other engineering errors can further reduce life. So instead of watching lifetime, companies are using condition monitoring to extend maintenance intervals and therefore save costs, reduce the number of opportunities to introduce errors, as well as gather indications of when a bearing is in distress and needs attention. ”Condition monitoring allows you to look at the data that you’re gathering, look at the trends of these data and make decisions based on the condition of the part rather than some notional time that it’s going to last,” says Pledger.
He continues: “We use big data: we pour all the information in from all the sensors that we’ve got into this big pond and we take out various combinations of the results for the information that people want. You can find hidden trends from that, or even find hidden associations that perhaps you weren’t aware of in the past.”
While more challenging issues require human intervention and assessment, simpler and more easily diagnosed issues are already falling into the domain of computers.
CM as a service
In the meantime, condition monitoring has become a significant part of the commodity services trend for industry. Charle explains: “It is not just having the data but correctly interpreting that data to provide appropriate services. That’s something we’ve seen since the mid-2000s, especially in oil and gas where, in the past, equipment was purchased to do condition monitoring programmes by their own people. Now they are outsourcing it to specialist service companies.”
Observes Deighton: “Service providers literally put engineers on site to collect data to analyse and report on asset health. One of the areas that we’re looking at now is more embedded technology and remote online monitoring systems that are communicating with the cloud. We can bring all that data into central, regionally-based remote prognostic centres. Within those centres we can have highly-skilled analysts able to interpret that data and report upon it.”
He adds: “We also have some dedicated industry sectors. For example, in Germany, one centre focuses purely on the wind turbine industry. Because all they’re monitoring is wind turbines, they’re able to build up a real picture around different models, different experiences and a real specialist skill.” Indeed, this is a key mechanism for condition monitoring companies to differentiate themselves from the competition, for instance offering a complete service based on operational certainty.
Charle at Emerson outlines how this process works. He says: “First we analyse their biggest issues in terms of operational certainty, with a complete audit and criticality ranking. Based on that we look at the opportunities in terms of condition monitoring. Then, together with the customer, we scope solutions, perhaps integrating these with the CMMS [central maintenance management system] where they track all the alarms which come from the condition monitoring system and create notifications or even work orders.”
Concludes Deighton: “Industry is looking for something that’s going to change and improve the business. We’re having to react to that to improve reliability of machines, or deliver a reduction in cost or maybe energy.”