The Internet of Rails

$40 billion IoT spending in transportation and logistics by 2020

Fuel savings of up to 17% with predictive maintenance

It’s a railroad operator’s nightmare: a 100-car train comes to a halt in the Canadian wilderness, hundreds of miles from the nearest repair depot, because of a blown piston or broken axle. Breakdowns like these have historically been addressed by dispatching people and equipment from the closest service areas, a long and expensive process. The longer the delay, the greater the risk of spoilage, missed connections and lost productivity for both people and rolling stock.

These risks are part and parcel of the railroad industry, which operates under unusually difficult conditions. Freight trains run through sparsely populated area at all hours and in all kinds of weather conditions. A seized axle or broken coupling on one car can bring the entire line to a halt. Because the cost of a breakdown is so high, trains must carry repair parts and engineers to effect repairs in the field. Diagnosis and repair on a cold winter night is both time-consuming and dangerous.

But thanks to the Internet of things (IoT) and big data analytics, these problems will soon become a rarity. What’s more, rail operators stand to gain substantial operating efficiencies from better understanding and optimizing the performance of their equipment in the field.

IoT leverages a network of smart connected devices to continually monitor the performance of critical components during their journey. Sensors track factors like temperature, vibration, pressure, fluid levels, fuel efficiency and emissions, sending a constant stream of updates to an on-board computer running analytical algorithms.

The local computer can look for anomalous conditions – such as a sudden drop in oil pressure – and alert local operators to an impending failure so that they can bring the train to a safe stop. The more dramatic impact of the IoT-big data combination, however, is the ability to perform predictive maintenance. That’s an analytic process that identifies equipment that’s at risk of failure in time to fix it while the train is in the repair yard.

Predictive maintenance combines historic data with streaming or recently recorded information and applies analytic algorithms to determine future maintenance needs. The process is complex because not all equipment wears out at the same rate, but there are big payoffs. For example, predictive maintenance enables equipment owners to customize maintenance schedules to individual assets based upon their condition, rather than servicing everything on the same schedule. This fact alone can add up to significant savings in cost avoidance and unnecessary use of spare parts.

IoT is also bringing new potential savings in track maintenance. The process of surveying track conditions has historically been a manual one, but autonomous drones will soon shoulder much of that burden. These low-cost flying devices can constantly watch from the air to look for hazards like fallen trees, rockslides and floods. They can also zoom down for close inspections to a degree that would be impractical with conventional aircraft. Tethered drones can even be operated from locomotives, flying several miles ahead to make sure the coast is clear.

These technologies aren’t science fiction, General Electric, which has been a leader in industrial IoT, sells a closed loop analytics system that rail operators can use to optimize efficiency, monitor, track conditions and test driver response times. The company’s Trip Optimizer can deliver fuel savings of up to 17%, while also improving safety and reducing maintenance costs.

Cisco Connected Rail is an end-to-end architectural framework that enables rail operators to monitor everything from on-board conditions to lighting and temperatures in railroad station waiting rooms. It also provides passenger-friendly features like onboard Wi-Fi, on-demand entertainment and up-to-date information services.

IoT services for railroads leverage the latest in distributed analytics processing. Because sending large data streams to a central facility or cloud is expensive – or even impossible in certain conditions – much of the stream processing is delegated to local onboard devices, which have just enough information to identify and react to urgent situations. Streaming data can be captured selectively or in its entirety and uploaded to powerful analytics servers when the train reaches its destination. Rail operators are thus able to perform detailed historical analyses to improve their predictive analytics capabilities.

The internet of things will give rail operators unprecedented visibility into their rolling stock, no matter where it’s located. The efficiencies and cost savings they can achieve will make rail transport safer, faster and more competitive with a growing variety of options.


What do you think?