Maintenance Mindset: How maintenance automation is reducing manual work on the plant floor
Welcome to Maintenance Mindset, our editors’ takes on things going on in the worlds of manufacturing and asset management that deserve some extra attention. This will appear regularly in the Member’s Only section of the site.
Earlier this year I had the opportunity to present at the RPM Symposium in Kalamazoo MI. The conference, organized by HECO, brings together about 500 maintenance and reliability professionals from the Great Lakes area, and many of the learning sessions focused this year on how to improve manual maintenance processes, everything from pump and shaft repair to bearing failure diagnosis and inventory management.
My session focused on changes that have taken place over the past 10 years in our industry, especially in maintenance and reliability. In the discussion afterwards, a bunch of us observed that the general mix of maintenance work (reactive / preventive / predictive) might not have changed much in 10 years but, due to the skills shortage, retirements, and the overnight rise of wireless networks on the plant floor, the way that the work gets done most definitely has changed.
If you can't find enough people to walk all the regular routes, what do you do? To paraphrase James Carville, it's automation, stupid! And this year it has become very clear that maintenance automation is taking place in a two-phase process: (1) hardware that reduces manual rounds, and (2) software that reduces manual data processing.
One example of game-changing hardware in this regard are automated precision lubrication systems, like UE Systems' On Trak, which integrate ultrasound, vibration, and temperature monitoring with an automated lubrication dispenser. These systems deliver a precise amount of lubrication only when the bearing needs it, which enables maintenance managers to eliminate unnecessary manual checks.
Keeping the focus on wireless remote monitoring, in a case study presented at the recent Maintenance Days conference in Sweden, Södra Cell showed how they are moving away from traditional hand-held vibration measurements to help eliminate route-based work. Currently 3,000 wireless vibration sensors are installed that offer continuous monitoring, and a team of just two engineers are able to handle alarms/alerts from the system.
However, Södra Cell also showed how they took the next step in the one-two maintenance automation punch: after deploying hardware to reduce manual rounds, they deployed software to reduce manual data analysis and alarm management. The Södra Cell team deployed Multiviz, an AI-powered tool, on 5% of their sensors to sort through the thousands of alarms to identify the few that needed manual intervention. Initial results revealed a significant reduction in workload, saving engineers 2–3 hours of analysis daily with just 5% coverage.
The Södra Cell case study also reminded me of the sessions at RPM Symposium where HECO demonstrated their Apollo IIoT solution, which combines physical remote monitoring sensors with predictive AI modeling to help teams identify and then act on abnormal asset behavior. The best part of these case studies is these solutions have moved beyond the hype cycle and are within reach of any plant team.
Maintenance automation through software was also the focus of the IFS Unleashed event in October, where CEO Mark Moffat kicked off the week by remarking that “the next industrial revolution is being powered by Industrial AI” with innovation happening at scale and at lightning speed. "We can watch this happen all around us, or we can run toward it and grab opportunity with both hands.”
Moffat added that IFS is currently working on 300 different AI use cases, with 60 of them planned for rollout in their software soon. And I have to admit, there was one AI maintenance application presented by IFS that immediately caught my attention, in terms of its ability to reduce manual work. What if you could use your CMMS to streamline the effort involved in a FMECA?
This is what a new IFS module is promising. From an IFS news release: "The new IFS.ai Copilot for FMECA (Failure Modes, Effects, Criticality Analysis) ... provides detailed analysis of how an asset might fail, the probability, and consequences of making or adjusting maintenance strategies. AI supports FMECA by unlocking insights from unstructured information such as manuals and maintenance reports to support and refine the analysis." At the event itself, IFS demonstrated how Exelon was already using the module to reduce manual data processing work, including being able to dig deeper on root cause analyses.
Now, there is definitely a place for teams that want to put in the manual work required by standards-based root cause analyses and failure modes and effects analyses. Manual effort is still the norm for these, and in fact Shon Isenhour and Brian Hronchek from Eruditio recorded three podcasts on this topic, offering tips and best practices for getting a handle on FMEAs and RCAs. These are time-intensive analyses requiring significant manual resources, so it's important to know how, when, and why to devote your resources to this kind of work.
However...software is clearly on the verge of reducing maintenance work that has traditionally been considered highly manual-intensive, either digesting unstructured asset data (both manuals and historical paper-based reports), probing possible root causes, automatically triggering work orders, updating FMECA analyses, and more. Imagine the time and work you could save, and/or the greater volume of RCAs and FMEAs that could be completed on your critical assets.
AI may be at the peak of its hype cycle, but only a few years ago that was also the case for wireless remote monitoring and IIoT technologies. This year more than ever, software is poised to combine with hardware and help you cut though manual work and scale the impact of your maintenance teams.