A Good Day: Looking Ahead

In the near future, the data that’s generated by assets will transform our daily activities and provide a new level of efficiency and reliability. AI-generated image

When today’s technologies are properly integrated and implemented with a human-centric approach, firefighting disappears and the world of a maintenance manager can become an interesting and enjoyable place to work.

Much has changed since I first described a good day for a Maintenance Manager in Efficient Plant, Nov. 2016, and updated in June 2022, so it’s time to look further ahead.

As I arrive in the morning, my personalized digital device (attached to my wrist) provides a schedule of the day’s activities. Moving toward my desk activates my computer, turns it on, and a retinal scan confirms my identity. I see that one of my technicians is printing a 3-D temporary part and another is using a drone to do a roof and pipe inspection. This afternoon, a different drone will do a confined-space inspection and then go up four stories alongside a building for another inspection. For a needed repair, a special 4-D printed part has been ordered that will react to temperature to grow to a pre-engineered larger size, avoiding an expensive disassembly.

Multiple systems are now fully integrated, including production, maintenance, supply chain, and maintenance parts inventory, so a complete picture of asset health is possible. The Enterprise Management System (EMS) is also linked to plant floor mobile devices that have capabilities such as radio-frequency identification, built-in camera, global positioning, and links to parts data and drawings. Ordering parts is almost fully automated. With links to MRO (maintenance, repair, and operations stores), parts availability is easily assessed and updated. The highly automated data collection greatly improves data quality/accuracy (a long-time past issue). Now the data is directly used with machine learning to generate work orders as needed (prescriptive maintenance). AI (artificial intelligence)/ML (machine learning) continuously monitors asset health in real time. Digital twins for asset maintenance and production impact are used with the maintenance-management system to evaluate/predict/decide on fixes needed (what, how, why, when, and who).

Another technician is working on a repair with a cobot that works alongside to hold parts, provide tools, and answer any questions. The cobot collects data for machine learning analytics, prescriptive maintenance, and continual preventive maintenance optimization (PMO). She is using safety/training glasses that provide step-by-step visual directions. Augmented reality helps to clarify what to do next.

For work directly performed by the technician, an enhanced ergonomic glove is used to provide additional gripping strength to minimize carpal-tunnel injuries. She is also wearing her health-monitoring vest, which includes a movement-collection device for ergonomic analysis and muscle-memory training to minimize stresses. Safety issues are almost non-existent with minimal breakdowns. Robots/cobots perform intricate and complicated work tasks. They can hold heavy parts and tools and are essentially unaffected when working in hazardous/extreme environments for long hours.

From the Enterprise Management System (EMS), I obtain a quick overview of production, reliability, and maintenance. Weibull-analysis software generates a chart that shows how much can be gained by improving production efficiencies and reliability. This information is supplemented by a computer-generated verbal summary that can be used in place of, or with, the charts. The EMS data is integrated with a “learning system” that makes some decisions (within defined parameters) and reports on those decisions, along with the underlying reasons.

Production processes are statistically controlled. As soon as there is evidence of out-of-bounds parts, they are corralled for recycling. However, it doesn’t happen often. Any small deviations from cycle time are monitored for each piece of equipment to enable timely maintenance interventions and assure throughput requirements.

Maintenance schedules are dynamically integrated with production requirements to schedule optimal maintenance times. Since the machinery and equipment are purchased with significant “design-in” reliability and maintainability (R&M) specifications, MTTR (mean time to repair) times are optimized.

With fierce competition demanding high levels of uptime, predictive maintenance is now the norm. This is made possible with regular inspections done by mobile robots, cobots, and drones, plus more sensors tied to more condition monitoring. It wasn’t that long ago that about a third of preventive maintenance was done too frequently, repeat failures were too commonplace, and too many work orders were still printed on paper, along with many other wasteful, time-consuming, and disruptive unnecessary practices.

An R&M simulation calculates the potential impact of the level of Design for Maintainability (DfM) for major assets. Purchasing is responsible for life-cycle costing which is measured as asset life-cycle performance. Asset providers are selected based on best historical MTBF (mean time between failures) and reliability growth performance. Once a week, the global continuous-improvement team meets virtually (using 3-D imaging to view all participants) to make decisions using data and recommendations from the learning system.

The lack of available plant-floor technical skills, recruiting/onboarding/retention difficulties of new hires and ongoing retirees is accelerating the need for alternatives. This still contributes to the faster digitization of maintenance, with more and smarter automation-focused solutions. Maintenance has now evolved to high use of technologies. With greater connectivity and communications and trustworthy data, the EMS system can more easily support Industry 4.0/5.0 initiatives (minimize material and energy waste, improve resilience, reduce carbon-release activities, and improve sustainability and other planet-friendly choices).

The Maintenance Continuous Improvement Team is testing new applications such as Nano-Robotics to perform tasks (at microscopic levels) without machinery/equipment dis-assembly and related disruptions. Another group is experimenting with Swarm-Robotics applications (uses individual robot input and then combines the collective knowledge to optimize results) to enable faster inspections, handle complex repairs, and more—think of how an ant colony works. Drones already use this technology with algorithms for coordination and control. Humanoids (robots that resemble the human body) typically have a specific function that emulates human movement and are the next step.

The final step is having robots look like people (Androids). Intelligent robots with autonomous mobility and the ability to learn will be a critical point for humanity. Prof. Stephen Hawking (renowned Univ. of Cambridge physicist) stated in multiple interviews that, “The development of full artificial intelligence could spell the end of the human race.”

About “90% of the world’s data was created in the past two years. Every two years the volume of data across the world doubles in size,” (Data Statistics (2024)—How much data is there in the world? (rivery.io). It’s up to us to use this data and knowledge (mainly AI) in a way that maintains a human-centric approach so people have easy access/understanding/control to enhance human ingenuity and continually improve as individuals and as organizations. EP

Based in Knoxville, Dr. Klaus M. Blache is director of the Reliability & Maintainability Center at the Univ. of Tennessee, and a research professor in the College of Engineering. Contact him at kblache@utk.edu.