RAILCAR inspections can be accelerated exponentially with artificial intelligence to check 120 cars in the time it takes to check one, according to the Canadian National Railway (CN).
CN showed off an example of this technology in advance of its investor day in Toronto, where it said such advanced technologies will save the railway as much as C$400 million (US$301 million) from 2020 to 2022.
The cost savings are part of CN's revised three-year financial target of low double-digit profit growth, exceeding its previous target of 10 per cent.
The checking portals are composed of metal arches placed over the tracks equipped with 36 stadium-quality lights and 360-degree cameras that capture continuous imagery of a train as it passes through.
The colour, high-resolution images are fed real-time into propriety algorithms that use machine learning to recognise defects. If a problem is spotted, the system flags a car and enables faster maintenance.
Trains don't even need to slow down, as the portals capture images at up to 60 miles per hour. Each portal cost about $3 to $4 million, executives said on the tour, adding the algorithms that recognise problems the human eye could miss are the 'secret sauce'..
CN has already installed four portals in Winnipeg and one north of Toronto, with plans to install two more in the US by the end of the year. By 2020, it will hit full deployment with 10 portals across the network, with regulatory approval expected by 2021.
Since the portals have not yet received regulatory approval, they have not yet replaced the traditional inspections that are still being conducted by 1,700 workers.
CN executives insisted the move isn't about replacing workers, who will spend more time actually fixing defects rather than spotting them. Rather, they highlighted the dramatic increase in efficiency and reliability that they believe will ultimately improve safety.
WORLD SHIPPING
CN showed off an example of this technology in advance of its investor day in Toronto, where it said such advanced technologies will save the railway as much as C$400 million (US$301 million) from 2020 to 2022.
The cost savings are part of CN's revised three-year financial target of low double-digit profit growth, exceeding its previous target of 10 per cent.
The checking portals are composed of metal arches placed over the tracks equipped with 36 stadium-quality lights and 360-degree cameras that capture continuous imagery of a train as it passes through.
The colour, high-resolution images are fed real-time into propriety algorithms that use machine learning to recognise defects. If a problem is spotted, the system flags a car and enables faster maintenance.
Trains don't even need to slow down, as the portals capture images at up to 60 miles per hour. Each portal cost about $3 to $4 million, executives said on the tour, adding the algorithms that recognise problems the human eye could miss are the 'secret sauce'..
CN has already installed four portals in Winnipeg and one north of Toronto, with plans to install two more in the US by the end of the year. By 2020, it will hit full deployment with 10 portals across the network, with regulatory approval expected by 2021.
Since the portals have not yet received regulatory approval, they have not yet replaced the traditional inspections that are still being conducted by 1,700 workers.
CN executives insisted the move isn't about replacing workers, who will spend more time actually fixing defects rather than spotting them. Rather, they highlighted the dramatic increase in efficiency and reliability that they believe will ultimately improve safety.
WORLD SHIPPING