OSAKA – Along the Sanyo trails shinkansen bullet train connecting Osaka and the southern island of Kyushu, high-quality sound microphones are strangely attached to the fence.
The mics are part of a monitoring system, developed by NTT Data and JR West, a train operator serving the Osaka area, designed to quickly identify potential problems with trains.
JR West, formally known as the West Japan Railway, in March last year ran a trial run of the program.
The system uses artificial intelligence to identify subwoofer noise in need of repair. If the AI makes an extreme noise, the information is passed to an investigator, who will decide whether to suspend activity.
The invention was inspired by a near miss event in 2017 in which a 14cm crack was found in the undercarriage of a Nozomi bullet train car on the Sanyo line. Although the train crew heard strange noises and smelled the burning smell, the train continued to operate. When the crack was last discovered by a team, it was just 3cm short of completely removing the train’s steel frame.
In addition to preventing a similar problem in the shinkansen line, JR West is looking to implement this new type of data analysis in the maintenance of trains and equipment ahead of other rail operators.
Umeda Osaka Metro Station: Historically, the Kansai area has been a center of innovation in the Japanese conservation railway industry. (Photo by Tomoki Mera)
The company operates nearly 5,000 km of railway in western Japan, making it necessary to improve maintenance accuracy to ensure safety while keeping costs down.
“We have maintained rail vehicles and equipment based on usage hours, but with data analysis, we can track them individually,” said Hideo Okuda, corporate officer at JR West.
Similar data analysis is being used to predict the breakdown of nearly 600 ticket machines and ticket gates in the Kobe area, west of Osaka. Each morning, servers collect data related to operating frequency along with the types of problems and when they occur. Movements based on a year’s worth of data help predict breakdowns for each individual ticket machine and gateway.
By predicting potential malfunctions in advance, repair workers can deal with problems in a timely manner, reducing work time and avoiding service interruptions.
Regular inspections, now performed once every one to three months, can be performed once every six months. JR West is considering introducing data-based maintenance as early as this spring.
In addition, the company is trying to set up a network to collect uniform data to manage via the internet all equipment used in its six service lines in more Osaka .
A data collection tool is installed every kilometer or so on the trails to monitor positions at signals and trails using sensors and cameras. This will allow JR West to stop relying on teamed inspections, which could reduce its staff shortage.
The rail operator also plans to save money with the new system, predicting a 500 million yen ($ 4.8 million) reduction in electricity costs alone across the region in fiscal 2030.
Competition among private rail operators in the larger Osaka area has led to innovation ahead of Tokyo.
Greater Osaka, which includes Kyoto, has historically been a hub of innovation in Japan’s conservation railway industry. Unlike Tokyo, where the government has a major influence in infrastructure development, the region’s rail industry is dominated by the private sector.
The area is home to Nankai Electric Railway, formerly Japan’s first privately funded railway in Japan. Many other private operators such as Hankyu, Hanshin Electric Railway and Kintetsu Railway also operate here.
Speed is crucial when it comes to adopting AI and other digital technologies. “We are taking an energetic approach” by conducting a number of experiments and centers over a short period of time, said JR West’s executive group.