Using AI Algorithms for Predictive Maintenance to Eliminate Unplanned Machine Downtime

In the image above (above the introduction), you can see that we’re tracking data from a reflow oven. In this specific case, the algorithm used by IIoT.Edge is analyzing data from both the sensors (temperature, etc.), and collected machine data using the Co-NECT platform (warning, status, condition changes, etc.). As you can see in figures 1 & 2, both data sources are updated in real-time for the algorithm to be as accurate as possible.

Based on the inputs from the machine, the AI algorithm looks for patterns to predict the equipment’s Remaining Useful Life (RUL). Simply put: how much time does the machine have left before it’s most likely to break down? Based on these calculations, it will recommend when maintenance should be performed to keep your lines up and running.

If you look at the image above, you can see that the algorithm issued a “RUL alert” as soon as the indicator went below a certain threshold, which is identified in red on the z-axis of the graph. When the indicator is under said threshold for a specific amount of time (the orange bar at the bottom of the graph), IIoT.Edge automatically issued a “maintenance alert”, (identified by the yellow diamond underneath the graph bar). Upon maintenance completion, we can see that the indicator went back to its original state.

IIoT.Edge offers multiple options when it comes to alerts – message or dashboard alerts being the most popular. It can even use an API call to any maintenance system to automatically schedule a maintenance activity. In other words, you can use this technology to predict when maintenance should be performed to avoid breakdowns and maximize the availability of your machines while never having to rely on an operator (which can be really hard to find these days with all the labor issues our industry is facing).

Now, what should you expect from such a solution? When used by one of our recent OEM clients, they experienced a significant increase in equipment availability, as well as a reduction of about 70% in machine breakdowns. It's important to note that even though our illustration here is for a reflow oven, predictive maintenance can work with any type of SMT machine. For example, it could also work for placement machine with vibration sensors, pressure sensors, vacuum sensors, and so forth.

When you think about all the breakdowns that could be avoided on all your SMT machines, it’s easy to understand how these AI tools could be critical in increasing your factory’s overall output. In a world where no one can afford to lose time, doing the most with what you currently have is key. And that’s exactly what IIoT.Edge is enabling.

Hungry for more use cases? Learn about another use of IIoT.Edge in this article about Statistical Process Control (SPC).