When maintenance becomes a data business
Industry context
When a critical machine fails without warning, the repair is often the smallest part of the bill. The line stops and orders slip. Crews get called in at night and spare parts are shipped at premium prices. In food and beverage, spoiled product comes on top of that. Siemens, one of the major suppliers of factory automation, has put a number on it. According to its True Cost of Downtime 2024 report, unplanned downtime costs the world’s 500 largest companies $1.4 trillion a year, around 11% of their revenue. One hour of lost production costs about $36,000 in consumer goods. On an automotive line, it can reach more than $2.3 million.
Maintenance strategy
Every factory must decide how to look after its machines. Some let a machine run until it breaks and fix it afterwards, which is known as reactive maintenance. Others service their machines on a fixed schedule even when nothing is wrong, which is called preventive maintenance. It’s safer than waiting for a breakdown, but parts often get replaced while they still have plenty of life in them. But there is a third way, called predictive maintenance. Sensors are installed on the machines and monitor them all the time, picking up early warning signs like unusual vibrations or a bearing running hotter than normal. This often happens weeks or even months before anything breaks, so the team can plan the repair for when it suits the plant. And as the chart below shows, this type of maintenance lets you reach an “optimum”, where the total cost is lowest, because you avoid both the unnecessary servicing and the expensive breakdowns. If you want to go further, this IVC article explains the three types of maintenance in more detail.
AssetWatch overview
AssetWatch, based in the US, offers predictive maintenance as a full service. Its small wireless sensors, like the one pictured below, are fixed directly onto a plant’s critical machines. Vibration, temperature and oil analysis all come together in one platform. AI flags anything unusual, then AssetWatch’s certified engineers review it and tell the plant team which problems to fix first. Customers pay a single monthly subscription covering the sensors, the network, the software and the experts. There is no capital expense, and installation takes one to two days without turning into an IT project. The sensors don’t even need the plant’s wifi, since their data goes through AssetWatch’s own cellular hubs. A plant can start with a handful of critical machines and add more over time, as the Kerry case below shows.
Case study: Kerry
AssetWatch’s latest webinar looks at Kerry, the global taste and nutrition group, and its plant in Rome, in the US state of Georgia, where it runs six production lines. The difficulty was that the lines depended on each other, so when one machine failed, the problem often spread across the plant. The maintenance team had been asked to move from fixing breakdowns to anticipating them. But with urgent repairs to deal with every day, that was hard to do, especially on dryers running in high heat and on equipment washed down constantly for food safety. Rather than instrument everything, Kerry and AssetWatch started small, with around 20 assets and roughly 100 sensors. They focused on machines that were either critical to production or known to fail often, such as fluidizing dryers and pneumatic blowers. The results came quickly. In the first two months, AssetWatch caught three major issues early. The plant puts the avoided costs at about $43,000, and calls that a conservative estimate. On one fluidizing dryer, rising bearing temperatures gave the team time to schedule the repair and bundle several other fixes into the same planned shutdown. Kerry has since added 30 more sensors at the site, with plans for around 100 more.
You can watch the full webinar here.
Congratulations to Brian Graham (CEO), Will Zell (founding CEO and now Chairman) and the team for delivering year after year and making AssetWatch a key component of any successful industrial strategy.