IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

Predictive Analytics Aboard the London Underground?

And the rail system's online dashboard allows the company to remotely monitor equipment vibration, video cameras, HVAC equipment, temperature, humidity, system alerts and fault warnings.

As an internationally renowned rail system, the London Underground has been a trendsetter since it was first constructed in 1863. It’s gone from four miles of track to 249 miles, is the third largest subway in the world and, as of April 15, it will be one of the few underground rail systems to adopt a real-time monitoring system infused with predictive analytics.

The news was announced at a Microsoft event in San Francisco, where CEO Satya Nadella unveiled improvements to Microsoft’s Azure cloud platform and its SQL Server 2014, the latest update to the software that uses a speedier in-memory system to process large volume data from apps.

Steve Pears, the managing director of rail at Telent — an IT vendor for Transit for London, the organization that governs the Underground — said the system has modernized the rail with an array of innovations through Azure and support from the IT firm CGI. The system’s online dashboard now allows the company to monitor equipment vibration (a performance indicator), video cameras, HVAC equipment, temperature, humidity, system alerts and fault warnings remotely. In both the short and long term, Pears said the advantages of the innovations are anticipated to decrease laborsome on-site maintenance while streamlining the rail’s day-to-day operations.

“It’s all about looking at performance, seeing when it’s starting to degrade and being able to take action,” Pears said.



To facilitate this, the dashboard has a tool chest of data gathering features that includes its ability to cross reference the status and location data of both maintenance personnel and equipment assets. When problems arise and warnings surface — such as an elevator breakdown or an escalator jam —  closest staff can be sent for repairs almost instantaneously and with specific instructions.

In testing, the accuracy and quality of initial repair efforts has nearly doubled due to the improvements. Previously there was literally a less than 50 percent chance a mechanical or maintenance problem could be fixed in a first attempt, Pears said. Now estimates show a 70 percent probability repairs can be handled without a second visit.

Predictive analytics are influencing preventative care for the rail too. With the system, Talent has identified a list of key indicators linked with equipment failures. Pears put a finger to escalator repairs as an example and said, through predictive analytics, his team has identified mechanical parts that need to be replaced in pairs to avoid future failures. More forecasted maintenance tasks and predictive servicing models are still in development.

telent-steve-pears-jasonshueh.jpg
“What we’re really trying to do is get to the point where you can predict when it’s going to fail," said Pears, pictured at left with the dashboard. "So it’s not reacting to failure it’s about saying, ‘I can see that performance is starting to degrade and therefore I should take action before it becomes catastrophic.’”

With roughly 2 billion trips taken on the Underground every year, passenger mobility, that constant flow of human foot traffic, was the critical concern for London’s transit authorities in the decision-making process, Pears said, and ultimately officials reasoning for the upgrade.

According to a release from the Transport for London earlier this month, additional plans will follow alongside Telent’s work. The announcement reported that a total of 330 million pounds (about US$559.4 million) has been allocated for other modernization projects to 70 stations. The preparations are hoped to modernize and secure the Underground for the next decade.



Jason Shueh is a former staff writer for Government Technology magazine.