Overcoming barriers to connected reliability: Is 2020 the year we make strides?
2020 is widely expected to be a year of progress for maintenance teams against their connected reliability goals, as they increase adoption of Industrial Internet of Things (IIoT) systems.
Why this year? In his December 2019 webinar with Fluke Reliability, technologist Oliver Sturrock explores some of the underlying conditions around technology adoption in maintenance operations and frames expectations for where connected reliability is headed.
Several significant advancements for reliability maintenance transpired in 2019. More facilities reported active connected maintenance pilots than ever before, even though full conversion to active connected reliability programs fell short of expectations, leaving many people disappointed and frustrated.
The good news: Reliability engineering has continued to gain respect and value in the operations organization, with more positions being created and more attention paid to uptime statistics.
“Plant managers start with wanting more data and then logically progress to needing people who understand what this data means and then needing the right kind of data to support further analysis,” Sturrock says. “As the asset data starts to arrive, managers ask more questions about how my plant is doing versus asset-level performance. That’s a change in thinking and it’s pushing reliability experts into the limelight.”
Still, the more valuable reliability maintenance becomes, the more difficult it has been to find skilled people to do the work. (A Michigan State University faculty leader reports that six job opportunities are available to every manufacturing sciences graduate!) Workforce shortages were regarded as a major hindrance to companies meeting connected reliability goals in 2019, according to poll-question data gathered during the webinar.
Many teams also reported not having enough management-level support or work hours for proactive maintenance work. “Too many maintenance organizations are still seen as a cost on the business,” says Sturrock. As a result, these maintenance teams waver between traditional downtime practices and break/fix centered thinking instead of prioritizing around preventing failures.In the end, even with the record number of pilot deployments, most sites did not meet their 2019 digital maintenance goals. According to Sturrock, some sites aren’t collecting enough meaningful asset health data, while others have too much data and no strategy for getting insights from it.
“The goal,” he says, “is for data to generate not just equipment status but to provide a better understanding of what we need to be working on at every point in time. If you knew which asset needed your attention, you’d have much higher chances of success. We don’t quite know that yet and we’re still surprised too often. More data and fewer surprises equal higher overall success within the organization.”Figure 2. As we collect more asset health data, we look for patterns, and we associate those patterns with equipment states. Once we see a data pattern 100 times and it lines up with a certain equipment problem, we can say with certainty this is what that data pattern means.
A substantial number of organizations are also unable to achieve widespread connectivity of their maintenance systems. If legacy EAM systems don’t accept data from newer wireless sensors, for example, teams struggle to build an overall picture of asset health.
In 2020, much is expected to change. As more maintenance system developers agree to open platforms and APIs, and as the operations workforce continues to add more data-centric thinkers, integrating systems should be less difficult.
Management support for data-driven reliability maintenance is forecasted to continue growing and more services are arising to help maintenance organizations convert to a connected-reliability way of operations.
In the webinar, Sturrock uses the diagram below to describe the technology pathway to building context for maintenance data: A substantial number of organizations are also unable to achieve widespread connectivity of their maintenance systems. If legacy EAM systems don’t accept data from newer wireless sensors, for example, teams struggle to build an overall picture of asset health.
In 2020, much is expected to change. As more maintenance system developers agree to open platforms and APIs, and as the operations workforce continues to add more data-centric thinkers, integrating systems should be less difficult.
Management support for data-driven reliability maintenance is forecasted to continue growing and more services are arising to help maintenance organizations convert to a connected-reliability way of operations.
In the webinar, Sturrock uses the diagram below to describe the technology pathway to building context for maintenance data:Figure 3: Technologists predict maintenance data analytics will progress from manual alarms all the way to prescriptive maintenance.
Many people still regard machine learning and artificial intelligence (AI) as buzzwords, rather than near-term solutions for maintenance operations. However, many newer connected-reliability systems are built to enable machine learning; organizations just need to get enough relevant data flowing into them. True AI-powered analytics still require more data and more processing power than most facilities have access to.
Happily, 2020 is seen as the year when in-plant connectivity starts to become easier. “Right now, we are getting data over corporate Wi-Fi, which creates security problems,” Sturrock acknowledges.
He advises looking at WAN (wide-area network) technologies and other alternatives to get connected. “Operating a separate network has significant security and ease of use advantages. WAN is low cost and can cover enormous facilities. The cost of setting up your own private network has already come down and will continue to,” he says, adding that 5G networks could be feasible in larger facilities.
Near term, Sturrock advices focusing on building the condition-monitoring mindset and collecting the right kind of indicative data. In his view, in the factory of the future, we will still rely on a combination of teams, systems, and data. Those three core elements won’t change but what each are doing might progressively look a little different.
To hear more predictions, tips, and advice from Sturrock, tune in to the entire webinar.
Additional resources:
