Lenovo’s AI-based solution unearths new efficiency in coal mines
27 Jul 2023
The solution
combines low-light sensing and AI-based image recognition capabilities to make
underground mining a safer and more efficient process.
China’s coal industry relies heavily on
underground mining. As of 2022, this method is employed by 92% of the nation’s
coal mines, contributing to 82% of domestic coal production.
High demand for
coal has driven Chinese coal companies to leverage intelligent mining
technology. According to the National Mine Safety Supervision Bureau of China,
there are over 1,300 intelligent mines in
China.
One of these
intelligent mines is the Wangjialing coal mine. Located in Shanxi Province, it
produces approximately 16,000 tons of coal per day, roughly equivalent to an
annual output of six million tons. Conveyor belts known as scraper conveyors
are essential to transport coal to the surface. Each conveyor may stretch up to
20 kilometers, consisting of a series of chains that are three to four
kilometers long.
Under constant heavy loads, scraper conveyors
can deform or even break, affecting the mining process. Replacing a deformed
chain takes around 30 minutes, but if a chain breaks, dismantling and
reorganizing the entire conveyor system could take 20 hours or more. The
challenge lies in detecting deformations before they lead to breakdowns.
Previously, the
Wangjialing Coal Mine employed manual inspection methods, which required
approximately 200 specialized maintenance personnel to descend underground for
inspections. This approach was costly and posed significant risks to safety,
prompting the mine to seek alternatives.
It eventually
collaborated with Lenovo to develop a 3D visual recognition system for scraper
conveyors. By installing high-resolution 3D cameras atop each scraper conveyor,
images of the conveyor chains can be captured continuously and in real-time to
assess their condition. These cameras are also equipped with cleaning devices
to adapt to the dusty mine environment.
Equipped with low-light sensing and artificial
intelligence-based image recognition capabilities, Lenovo’s system can capture
and analyze 3D depth images in dark environments, allowing it to detect
abnormalities of conveyor chains up to a one-millimeter precision.
The accuracy,
efficiency, and safety of maintenance operations at the Wangjialing coal mine
have significantly improved since the adoption of Lenovo’s solution. Defect
detection coverage and daily maintenance coverage have risen above 95% and 99%
respectively. The time taken to respond to anomalies has also been cut by 80%,
and the average daily downtime caused by line stoppages has been reduced to two
minutes, boosting the mine’s productivity.
By automating the inspection process, the
mine’s maintenance team now spends 90% less time underground. Manual inspection
of scraper conveyors is no longer required, and maintenance workers only
descend underground when repairs are necessary.