Deeply, an acoustic artificial intelligence solutions provider, said on Oct. 1 that it has supplied its "Listen AI" predictive maintenance solution for multi-axis transfer robots operated by a global advanced components manufacturer.
Listen AI uses acoustic analysis technology to monitor the various sounds generated on factory floors in real time and turn them into usable data. The solution diagnoses faults in the motors and gearboxes of multi-axis robots that inspect and transport advanced components. Microphones installed on the robots' drive units detect sounds that deviate from normal operating noise, score them and automatically flag abnormalities.
Continuous sound-based monitoring also allows early signs of trouble to be caught before a defect occurs, while building up related data. Previously, when a problem arose on a production line, every robot had to be inspected. Now only the line where an abnormal signal is detected needs to be checked, cutting maintenance staffing and costs.
Deeply plans to advance the technology beyond detecting and flagging abnormal sounds, to diagnosing which axis is likely to fail and predicting remaining useful life (RUL). On that basis, the company is also discussing expanding the number of lines covered by predictive maintenance and rolling out the solution to other plants.
"Automated processes using robots are increasing rapidly, but predictive maintenance for robots still requires more advanced technology," Deeply CEO Lee Su-ji said. "We will improve our analytical precision and lead the robot predictive maintenance market."
