Pneumatic cylinders that break and cause a production stop is the most frequent hardware-related cause of downtime for NPB equipment. But with data at hand, they have been able to predict - and eliminate - the issue at one customer's production site.
High-frequency data capture across the entire line. The CATCH.AI data platform automatically structures and normalizes the data.
“We collect basically everything there is to look at in the machine: Sensors, steps of sequences, temperatures, whatever. CCTV footage goes in as well. And then we can look at every point of data we have collected at a given time” says Olof Blom.
NPB used the platform’s rule engine to generate events when certain issues occurred, enabling them to jump from the event to the relevant camera view for immediate visual confirmation.
They also built a wide range of custom dashboards ranging from a quick overview of potential hot spots with color-coded status and event charts to per-machine drilldowns used for detailed troubleshooting.
Based on their extensive knowledge of the machine, they created a rule that would track air pressure vs. stroke time. If the time deviates by more than 15% repeatedly (e.g. six times in a row), this indicates an impending failure.
“If the air pressure remains the same but the time fluctuates, that means that the mechanical part is probably going bad soon,” says Olof Blom.
With that knowledge, they could notify their customer to let them schedule maintenance instead of risking an emergency stop. It also gave the customer time to ensure that they had the right spare parts available.
Anders Meister, CCO
