How artificial intelligence can predict parts failure before they fail


For the artificial-intelligence (AI) company Acerta Analytics Solutions Inc., the end of the production line has always been the stopping point.

But as the five-year-old industrial tech firm scales up, a new predictive-maintenance project with Nissan Motor Corp. will extend Acerta’s reach from the shop floor to vehicles on the open road.

The Kitchener, Ont., company has made a name for itself writing sophisticated algorithms to parse reams of production data for the likes of Dana Inc., ZF Friedrichshafen and BorgWarner Inc., helping the auto suppliers increase efficiency while improving part quality.

The new research partnership represents a different kind of “big-data problem,” said CEO Greta Cutulenco. But it’s one that Acerta is perfectly equipped to handle.

“In our solutions on the production floor, we’re monitoring hundreds of signals all the time and running these analyses against them. And that’s really our core capability, having this ability to monitor and launch machine-learning (ML) models against this big data that’s available from the manufacturers.”

The collaborative work with Nissan, Cutulenco said, simply exchanges the data points from manufacturing equipment for signals from up to 250 components beneath the hood of a Nissan vehicle.

FIX BEFORE PROBLEMS OCCUR

The $1.4-million project, announced June 22, focuses on predicting failures within the powertrain. Piggybacking on data already being collected by sensors embedded in engine components, Acerta is developing the algorithms to identify problems with individual parts before they become problems.

“The goal has been to look at that data and use ML and AI to start predicting the wear patterns of those components and ultimately the failure pattern so that we can help Nissan become much more proactive in how they’re identifying problems,” Cutulenco said.



Source link

Leave a Comment

Your email address will not be published.

Shopping Cart