How do you improve operations with IoT and predictive maintenance?

To answer that, we should first understand that a large portion of maintenance costs end up being wasted. While regular preventive maintenance tasks can help you keep certain failures from happening, it is a bit imprecise.

The problem with PM is it’s based on the assumption that equipment failures occur on a schedule. The reality is that only 18% of all assets fail based on age. The rest of equipment failures seem to occur at random.

The end result is a lot of PM tasks either achieve nothing or might even cause an unbalance within assets that are just fine. A lot of hours and resources are wasted, and in the worst case scenario, it can result in extra costs as equipment fails due to excessive PM.

This is where IoT (Internet of Things) comes in. Using various sensors—ultrasound, vibration, infrared, oil analysis—your equipment can “talk” to you and let you know when something’s out of balance. Using that data, you can schedule maintenance when something’s off, essentially moving you from an inefficient time-based system to a data-driven approach.

As you use IoT with predictive maintenance (PdM), you’ll start to see:

• Fewer resources going toward work that doesn’t need to be done
• Less time being spent preventing failures that won’t impact your processes
• Drastically fewer breakdowns and unplanned outages
• Lower preventive maintenance hours overall

Of course, to achieve all that, you need to use it properly. Some companies will actually implement IoT technology, but never use the data, which kind of defeats the purpose.

You need to connect your IoT data with a condition-based monitoring (CBM) system. That system can track IoT data and create a work order whenever something is off.

I’d also recommend implementing PdM one asset at a time if you’re just getting started. Pick one of your most critical assets, install the necessary IoT equipment, and begin tracking the data. Set parameters on when to perform maintenance checks and make adjustments as you go. Once you’ve got the system down, move on to the next until you have a fully-fledged PdM program.