Acoustic artificial intelligence (AI) solution provider DIPLI has begun supplying predictive maintenance solutions that diagnose anomalies in multi-axis robots through sound. The company is expanding its business into the robot predictive maintenance market based on acoustic data from manufacturing sites.
DIPLI announced on October 1 that it has supplied its predictive maintenance 'Listen AI' solution to a global advanced parts manufacturer’s multi-axis transfer robots. This marks the first time DIPLI has provided solutions in the field of robot predictive maintenance.
Listen AI is an acoustic AI solution that analyzes sounds generated in manufacturing environments in real-time to determine equipment anomalies. DIPLI has previously applied Listen AI to quality inspections of drive components such as motors, gears, and bearings, as well as to the inspection of fastening sounds from connectors and bolts.
Leveraging its experience analyzing over 70 types of sounds and more than 15 terabytes of self-collected manufacturing acoustic data, DIPLI is expanding its application into predictive maintenance. The company is currently conducting proof of concept (PoC) tests of Listen AI technology with several global firms, and has completed healthcare technology validation for the multi-axis transfer robots of a major global manufacturer, leading to actual supply.
The supplied Listen AI diagnoses anomalies in the motors and gearboxes of multi-axis robots that inspect and transfer advanced parts through sound. DIPLI analyzes the correlation between the load on the robots and the sounds produced, providing indicators to detect signs of anomalies early by tracking the degradation trends of components.
The system detects sounds that differ from normal sounds using microphones installed on the robot drive components, scoring them to automatically determine anomalies. It is designed to extract signs of anomalies even in a broadband noise environment where multiple robots operate simultaneously. Listen AI is also being applied to detect abnormal sounds in bearings of raw material transfer equipment for displays.
DIPLI plans to further enhance its sound-based predictive maintenance technology. The goal is to advance from merely providing alerts when abnormal sounds are detected to diagnosing the faulty axes and predicting the remaining useful life (RUL) of components. The company is also discussing plans to expand the application of predictive maintenance lines and broaden its reach to other factories.
Lee Su-ji, CEO of DIPLI, stated, “While the use of robots in automated processes is rapidly increasing, robot predictive maintenance still requires advanced technology. Listen AI has proven the potential and validity of sound-based predictive maintenance, leading to actual supply.”
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.