China is using an AI system to process large amounts of data in real time, helping to quickly alert high-speed rail maintenance teams.
High-speed trains parked on the railway from above. Photo: ximushushu/iStock
China is using an artificial intelligence (AI) system in Beijing to operate and maintain its high-speed rail network, Interesting Engineering reported on March 12.
With a total length of about 45,000 km, China has the largest high-speed rail network on the planet.
The AI system processes large amounts of data across China in real time, and can alert maintenance teams of unusual situations within 40 minutes with an accuracy rate of up to 95%.
Liu Daoan, a senior engineer at the infrastructure inspection center of China State Railway Group, said the system works effectively. “The system helps on-site teams conduct re-inspections and repairs as quickly as possible,” Daoan said in a study published in the industry journal China Railway.
China's high-speed rail is the fastest in the world , with an operating speed of 350 km/h and expected to increase to 400 km/h next year. The country plans to continue expanding its rail network until every city with a population of over 500,000 is connected.
With AI, minor track faults on high-speed rail lines operating in China have dropped by 80% over the past year. None of these lines have had to issue deceleration warnings due to serious track problems. Research has also shown that the amplitude of track displacement due to strong winds has been significantly reduced thanks to AI.
The US rail network is facing predictable challenges, with poor maintenance leading to ongoing safety risks. Over the past 50 years, the average number of derailments has exceeded 2,800 per year.
More than a decade ago, countries like Germany and Switzerland first recognized the potential of AI in railway management. Both have tried to use AI to improve their rail networks. However, their networks are smaller than China’s.
To train the AI system, Chinese scientists had to collect a huge amount of raw data. They drew on records of train body movements, track vibrations, waveform values, and meteorological records. Previously, the maintenance center issued warnings once a week. Now, thanks to AI, reports are issued daily.
Thu Thao (According to Interesting Engineering )
Source link
Comment (0)