Latest posts by Blake Griffin (see all)
- Three Key Points in Understanding Why Predictive Maintenance is Growing Exponentially - May 4, 2020
- Predictive Maintenance in Motor Driven Systems – 2020 - May 1, 2020
- COVID-19 Could Force Quicker Adoption of New Predictive Maintenance Service Models - March 26, 2020
- Predictive Maintenance: Primed for Growth - March 6, 2020
Modern predictive maintenance systems provide a vast amount of analytical data, upon which critical actions can be carried out before they actually become critical. Here we outline some of the systems involved, and look to a future where the hardware itself becomes a very small part of the equation.
Predictive maintenance has been around longer than you think
Predictive maintenance, as a concept, isn’t actually all that new. If you think back to when an engineer could put his ear to the end of a screwdriver, touch the other end to a machine, and judge if a bearing was nearing the end of its life, then you have some idea of how predictive maintenance perhaps isn’t the revelatory new technology it’s sometimes pitched as.
What is undeniable, however, is the ways in which combining that engineering knack with modern methods of data capture has resulted in the creation of an entire marketplace dedicated to predicting – and therefore enabling the prevention of – failure. Even handheld condition monitoring devices, of the kind offered by Fluke, for example, have been offering a leg-up to engineers wishing to predict the life of a particular machine or its components. While this technology is still growing, new technologies are emerging which will undoubtedly enhance the prevalence of modern predictive maintenance. Smart sensors, which quietly sit and amass data on metrics like temperature, vibration and speed, help further colour the picture of the inner workings of a machine, and provide greater opportunity to monitor trends or patterns which could prove vital in predicting failure or maintenance requirements.
Could devices like smart sensors, which feed into a wider connected ecosystem containing drives, motors and PLCs, ever fully replace handheld monitoring devices? Or is there space in industry for both approaches?
The role of drives in predictive maintenance
Variable speed drives hold a certain ace up their sleeve when it comes to predictive maintenance. As the devices most closely linked to the motor, they are able to stockpile colossal amounts of data. Engineers can set parameters for such readings and will then receive notification if, for example, the current draw from a motor falls outside of the set parameters. This information on current draw alone is enough to get a strong view into the health of a motor.
However, as with any technology, some drives are more capable – or intelligent – than others. The level of intelligence housed within a drive will likely dictate how the drive will be deployed with regard to predictive maintenance. There are effectively three ways a drive can be used in predictive maintenance: as a sensor, as a data aggregation device, or as both. We expect that the most intelligent drives, like those found in the Automotive industry, will be able to both sense current draw from the motor, perform basic anomaly detection, and communicate that information through some designated channel.
Regardless of the drive’s capability, there are still questions. Namely, is the data that a drive can produce better than what a smart sensor could produce? Will smart sensor data be used in conjunction with current draw data? If so, how will this work? It’s also unclear if drives companies are aware of the significant value of the data they collect. The answers to these questions provide strong insight into the differing approaches of major drives vendors and, more broadly, the general direction of the drives market.
Hardware vs software
While the majority of conversations around predictive maintenance involve hardware – be that drives or sensors – there is a clear move towards understanding the importance of software in the equation. After all, with the amount of data collected only ever likely to rise, it becomes necessary for the curators of this data to find an efficient solution for its storage and subsequent analysis.
Here you will see the emergence of more familiar household names into the mix. Amazon and Microsoft, with their AWS and Azure platforms respectively, are already major players in the IIoT world, competing with more traditional industrial names like Siemens and GE. Indeed, GE saw the direction of travel in its recent rebrand, assuming the ‘Digital Industrial Company’ moniker and increasing its capability in software and cloud computing.
There’s an argument that the big software firms will limit their involvement in dedicated predictive maintenance software. Looking instead to be effectively large-scale data and analytics providers. However, a look at the potential revenues on offer to whoever gets predictive maintenance software right could prompt these large software firms to step up their participation in predictive maintenance analytics. Alternatively, and perhaps more likely, we might see a rise in more focused start-ups who form strategic partnerships with the large software and/or industrial houses, offering the hardware component of the system for free in order to benefit from the software as a service (SaaS) model.
Regardless of how the market evolves, there are certain truisms. Life-earned, tried-and-tested application knowledge from engineers out in the field is still important, and always will be. Predictive maintenance isn’t intended to phase out the skill, intuition or knack of an engineer with 40 years’ experience. What is clear, though, is that the engineer of the future will need to be as competent understanding analytical data as they will with a screwdriver in their hand. After all, what use is gigabyte upon gigabyte of data if you don’t know how best to interpret it, or reap the benefits?