Railroads Turn to Predictive Analytics to Conquer the Last Frontier: Downtime

Dec. 12, 2017
Predictive analytics is increasing railroads' uptime through improved maintenance and increased safety.

Despite a modest uptick in carload traffic this year, railroads continue to face financial pressure. Running longer trains, shuttering switching yards and furloughing more workers have helped relieve some pressure. But now railroads are taking aim to eliminate their costliest problem in light of uncertainty: downtime.

To do it, they are turning to predictive analytics software that forecasts problems before they happen, so operations can be more productive, reliable, safe and secure.

Today’s modern locomotives generate thousands of signals, which provide insight into the locomotive’s status and physical location. In the past, most of that data was never harvested. Real-time readings from onboard gauges often sufficed. If data was stored, it was often only reviewed periodically, if at all.

Now, data can be transferred wirelessly for processing and stored inexpensively in the cloud. Insights derived from these data sets can be delivered directly to the technicians and engineers who need them when they need them. By empowering the workforce with prognostic, actionable insights, they are able to take action that results in more efficient and reliable operations without sacrificing safety or security.  

Across the globe, railroads are beginning to tap into their data to improve their operations. A major Class 1 in North America has been using predictive analytics software with real, valuable results. Through two months of use, the software has prevented more than 50 road failures from occurring, creating well over $1 million in value for this operator.

Railroads can measure these insights and the impact they create across the entire operation in three key areas. First, on the track. These insights help maintenance technicians make informed decisions based on a locomotive’s status, performance and shop history. Real-time prognostic alerts increase locomotive availability. Perhaps even more importantly, it prevents a failure that causes a locomotive from breaking down on the tracks disrupting the network and delaying tens of millions of goods from arriving on time.

Predictive analytics is also being used to reduce the time it takes to build trains in the yard (coupling of rail cars together with a locomotive). When locomotives aren’t moving, railroads aren’t making money. By knowing the locomotives health and readiness, a technician can match the locomotive to the right mission ensuring the highest priority trains are well served.

Another North American Class 1 railroad has been using the same software to eliminate costly manual tests in the shop. Through the use of the software, this railroad has saved more than 35,000 gallons of diesel and over 1500 man-hours over the course of six months at one shop

Predictive analytics solutions are proven to help major railroads become more reliable, efficient and safer by alerting users of critical failures before they occur. In 2016, there were 244 incidents or accidents caused by mechanical failure, according to Federal Railroad Administration data. Railroads hope to increase safety with the use of this new software.

Railroads have conquered many frontiers during the last 150 years; from opening the West to turning products into commodities unifying the United States along the way. With predictive analytics, railroads are well on their way to conquer the next frontier — downtime.

Joe Becker is the director of industrial analytics at Uptake, a leading predictive analytics software company in Chicago.

Uptake 59b94e44c44a1


Sept. 13, 2017