Preventing downtime with centralized reliability management

A leading producer of essential minerals and distributor of specialty chemicals, Omya’s operations span 50 countries and 160 plants. And each of those plants is different, varying in scale, in complexity, age, accessibility, their levels of automation and data maturity. Running an enterprise that varied, at that scale, across that many countries, is, among other things, a maintenance challenge. How do you standardize what has never been standard? To solve that problem, Omya partnered with AVEVA to build a unified predictive maintenance program that spans its entire global enterprise.

Using predictive analytics has helped us identify equipment issues almost three months ahead of time.

–Jonathan Vincent Regional Maintenance Manager for the Americas, Omya

From reactive to predictive

Like many global industrial organizations, Omya’s maintenance practices evolved locally, shaped by site-specific needs and circumstances. Without a standardized way to capture and share insights, plants operated in isolation from one another, each relying on their own strategies and mixes of traditional condition-based monitoring tools—like oil and vibration analysis, thermography, and visual inspections—to track equipment health.

Although these tools can be useful, they share a key limitation: they detect problems late in the failure curve, leaving maintenance teams only a narrow window of time in which to intervene. The narrower that window, the more difficult it is to avoid failures, and the more likely unplanned downtime becomes.

To widen that window, Omya needed to push detection further upstream. That’s when it turned to AVEVA Predictive Analytics. Now, rather than waiting for physical symptoms to appear, Omya’s predictive maintenance program monitors multiple variables in real time to identify anomalies far earlier than condition-based monitoring allows.

“It’s become a first line of defense for us as part of our maintenance strategy,” says Jonathan Vincent, Omya’s Regional Maintenance Manager for the Americas. “Using predictive analytics has helped us identify equipment issues almost three months ahead of time.”

From decentralized to unified

The real impact of AVEVA Predictive Analytics isn’t just what it can do for a single plant; it’s what it can do for 160 of them. Before the new program, individual plants identified and solved equipment problems every day, accumulating hard-won operational knowledge in the process. Those lessons, however, rarely travelled beyond the plants where they were learned. Insights remained local, siloed, and largely invisible to the broader enterprise.

To ensure those lessons could travel between plants, Omya shifted to an equipment-based modeling approach, defining predictive models by asset type. These models incorporate multiple variables—like temperature, vibration, power, pressure, flow, and feed rate—evaluating their relationships over time, establishing a baseline of normal behavior, and detecting any deviations.

Beyond a more holistic view of critical assets, equipment-based modeling gives maintenance teams a common language for describing how a given class of equipment behaves when it’s healthy and how it deviates when it’s not.

With this common modeling foundation in place, Omya’s next step was to build an organizational framework to match using the case management capabilities of AVEVA Predictive Analytics. Centralized teams met biweekly with individual plants, providing pattern recognition across a given region, while plantlevel teams provided on-the-ground context.

The resulted is a hybrid system of centralized oversight and local autonomy. “This is really what empowered our users to resolve actions in a timely manner while giving upper management an understanding of what’s going on with the equipment,” said Vincent.

Now, every finding, every intervention, every lesson learned gets logged into this one, centralized system, creating a shared, searchable record of how issues are identified and resolved. “So three years down the road,” said Vincent, “a plant with similar equipment can refer back to the case management of another plant and see, okay, this is what they did to reduce the vibration or to correct the potential failure.”

AVEVA Predictive Analytics has a fantastic case management tool, and this is really what empowered our users to resolve actions in a timely manner while giving upper management an understanding of what’s going on with the equipment.

–Jonathan Vincent Regional Maintenance Manager for the Americas, Omya

Other catches are less straightforward. At another plant, about three years after implementation, the system flagged an issue with a flash dryer: a suspicious drop in speed. The drop was sudden—a novel occurrence, not seen elsewhere in the flash dryer’s dataset. The vibration and motor power, however, remained steady. Past cases had taught the team that a drop or increase in one usually corresponds to a drop or increase in the other. They flagged the issue and followed up with the production team, which confirmed what the data suggested. The speed of the equipment was difficult to control. After tracing the issue back to the drive, they found the culprit: a faulty IGBT card.

The catch carried a particular significance. This same issue occurred at the same plant in 2017. Back then, there was no early detection, no intervention. The faulty card led to a critical failure. This time around, the new system identified the anomaly in time for maintenance to swap the card out before the equipment went down.

Within the first two years of its predictive maintenance program, Omya had already saved 15,000 hours of unplanned downtime. The business case, Vincent said, is very straightforward. “A single catch pays for the software for a year.”

A single catch pays for the software for the entire year.

–Jonathan Vincent Regional Maintenance Manager for the Americas, Omya

Today, Omya’s new predictive maintenance program is live at 48 plants with another 16 expected in a few more months. The roadmap ends ultimately at all 160. Each new plant that joins the predictive maintenance program gains immediate benefits: earlier issue detection, reduced downtime, rapid ROI.

But the benefits extend beyond the plant level. With every new deployment, Omya expands its repository of case history, diagnostics, and proven mitigation strategies, building an institutional memory that grows more valuable with every plant that contributes to it.