As an application engineer in heavy industry, Dr. Shubrajit Bhaumik had experienced the physical demands of inspecting bearings, gears, shafts and other parts of heavy machinery within a short period of time and submitting a report quickly.
“Tribologists examine the friction parts and the maintenance team uses their report to determine the maintenance plan,” he says. “During my time in the industry, I have seen how difficult this process can be. In the steel and cement industries, it is difficult for application engineers to provide an immediate report due to the high ambient temperatures and harsh operating conditions.”
Bhaumik, currently the Centre Director and Principal Researcher at the Tribology and Interactive Surfaces Research Laboratory (TRISUL) in the Department of Mechanical Engineering at Amrita Vishwa Vidyapeetham, has leveraged his industry experience to come up with a solution to the problem of on-site inspections – the BearingVisionAI app. Developed by the Tribology Research Group of TRISUL at the university’s Chennai campus, it makes on-site inspections much easier. BearingVision, an AI-integrated application, is the collaborative project of Bhaumik and his colleague Dr. R. Prasanna Kumar (Associate Professor, Co-Researcher, School of Computing), along with three third-year students from the Artificial Intelligence department of the School of Computing (M. Shree Prasad, G. Jeevan Sendur and G. Venkata Krishna Kumar) and a third-year mechanical engineering student (Manohar Reddy).
The app’s biggest unique selling point is its ease of on-site diagnosis. “Users simply take a photo of the surface to be examined and upload the image. The app recognizes the type of defect and quantifies the damage within seconds,” says Shree Prasad. For the students themselves, the AI/computer collaboration with mechanical engineering was a great learning experience.
“The experience of working on a multidisciplinary project is priceless,” says Venkata Krishna Kumar. “It was exciting to apply the concepts we learned in our programming classes to a core mechanical engineering problem in real time. It was a perfect blend of programming and core engineering,” adds Jeevan Sendur. For mechanical engineering student Manohar Reddy, the learning that comes with selecting different algorithms along with his AI teammates is what he values the most.
The team is also very pleased with how market-friendly the app is. “The app can be customized to detect faults in any mechanical element,” says Bhaumik.
The team worked on the project for eight months and has now filed a patent for the app. “Once we get in touch with industry, we need their requirements and can adjust the algorithm,” says Bhaumik. The team plans to bring this product to market as soon as possible. “We are now open for industry to use this technology directly and we plan to license this product,” he adds.
“We carry out many projects for industry,” says Bhaumik. “For example, we work on failure analysis. We are also working on developing self-lubricating products and coatings and it is against this background that the app was developed.”
